The Annual Review of Tire Materials and Tire Manufacturing Technology

Solutions for tomorrow’s problems, today.

Ergon recognized the need for safe products in 1996 and pioneered the

development of environmentally responsible base oil solutions. Our line

of aromatic replacement oils for the tire and rubber industries meet all

requirements found in the EU Directive 2005/69/EC.

Hyprene L2000Hyprene BO300Hyprene L1200Hyprene BO390

For more information, please email your regional contact.

Ergon North & South America – Jim Borthwick – jim.borthwick@ergon.com

Ergon Europe, Middle East & Africa – Per Klintstam – per.klintstam@ergon.com

Ergon Asia – Harry Chan – harry.chan@ergon.com

ergonlubes.com | ergoneurope.com | ergon.com

Ergon.Global.

1

10 Improving key tire performanceDr Joachim Neubauer, Michelin, France

14 Tire pressure maintenance: a statistical investigationNational Highway Traffic Safety Administration, USA

22 Partial replacement of silica with microcrystalline cellulose in rubber compositesWen Bai & Kaichang Li, Department of Wood Science and Engineering, Oregon State University, USA

32 Uniformity: a crucial attribute of tire/wheel assembliesMarion Pottinger, M’Engineering, USA

40 Comparison of tire models and their application for vehicle system dynamicsP. Lugner & M. Plöchl, University of Technology, Austria

48 Exceeding aircraft tire speed during take-offIngrid Wakefield, flight operations engineer & Chris Dubuque, service engineer, landing gear systems, The Boeing Company, USA

52 Estimating tire-road friction for chassis-control and driver-assistance systemsMarco Pesce, vehicle systems, vehicle dynamics and fuel economy, Centro Ricerche Fiat SCpA, Italy

54 Viscoelastic analysis of rolling tires using the finite element methodMir Hamid Reza Ghoreishy, Iran Polymer and Petrochemical Institute, Iran

59 Noise and rolling resistance: is there a conflict?Jerzy Ejsmont & Grzegorz Ronowski, Technical University of Gdansk, Poland

62 Misuse load cases and MBS simulationProf Dr-Ing Ch Oertel, FH Brandenburg (Brandenburg University of Applied Sciences), Germany

66 Tire-road contact information for driver assistance systems: an optical sensor approachAri J. Tuononen, Helsinki University of Technology (TKK), Finland

70 Tire-sand interaction research for lunar applicationsJaifeng ‘Jeff’ Ma & Professor Joshua Summers, Clemson University, USA

76 Europe’s new regulatory framework: requiring and inspiring new technology and innovationFazilet Cinaralp, secretary general, European Tyre & Rubber Manufacturers’ Association (ETRMA), Belgium

78 A universal flexometer and dynamic mechanical thermal spectrometerHorst Deckmann, Gabo Qualimeter Testanlagen GmbH, Germany

82 From golf balls to rolling resistanceDr Heike Kloppenburg & David Hardy, Performance Butadiene Rubbers, Lanxess Deutschland GmbH, Germany

88 Fatigue test machine for crack growth testingHugues Baurier, 01dB-Metravib, France

92 Final-finish tire testing: what is the true cost?Shaun M. Immel, Micro-Poise, USA

98 Getting a grip on future tiresMarika Rangstedt, Nynas, Sweden

100 Cutting-edge adhesion promoters for the tire industryMarcus Bayer & Elke Gebauer, EMS-Griltech, Switzerland

105 Engineered solutions for niche applicationsDr Berrin Yilmaz, Kordsa Global, Turkey

SECTION 1: DESIGN, MATERIALS, TESTING, AND FUTURE POSSIBILITIES

SECTION 2: PRODUCTION, QUALITY, AND RECYCLING110 Curing rubber compounds efficiently and cost-effectively

Ali Ansarifar, Department of Materials, Loughborough University, UK; Co-authors: Li Wang, Saeed Ostad Movahed and Farhan Saeed, Department of Materials, Loughborough University, UK; and K. Ansar Yasin, Department of Chemistry, The University of Azad Jammu and Kashmir, Muzaffarabad, Azad, Kashmir, Pakistan and S. Hameed, Department of Chemistry, Quaid-i-Azam University, Islamabad, Pakistan

116 Increased production rate through improved bladder compoundMaryam Mokhtarimehr & A.G. Moteshareie, laboratory manager and compounding manager respectively of Dena Tire, Iran

118 Rubber devulcanization: a successful projectAndrew James, Smithers Rapra Technology, UK

120 Tire industry innovations including co-extrusion heads and Ethernet line controlDr Tim Pohl, Troester GmbH & Co KG, Germany

124 Automatic bead apex stacking and loading into a tire building machineJ.K. Grashuis, VMI-Group, the Netherlands

128 Web width measuring systemsAndreas Flöter & Sabine Sladky, BST International GmbH, Germany

130 Tire molding: the final step in the tire production chainRainer Hilke, A-Z Formen- und Maschinenbau GmbH, Germany

132 The importance of checking the length, width, weight and profile of cut treadDr Hartwig Suhr, Dr Noll GmbH, Germany

Member of the AuditBureau of Circulations

2008 Average Net Circulation: 6,067 copies

Editor Adam GavineDeputy editor Jon LawsonAssociate editor Roger WilliamsEditorial assistant Bunny RichardsChief sub-editor Alex BradleySub-editor William BakerProduction manager Ian DonovanProduction team Joanna Coles, Carole Doran, Lewis Hopkins, Emma UwinsArt director Craig MarshallArt editor Andy BassDesign and production team Louise Adams, Anna Davie, Andrew Locke, Duncan Norton, James Sutcliffe, Nicola Turner, Julie Welby, Ben WhiteProofreaders Frank Millard, Christine VelardeSales director Colin ScottCirculation Adam FrostTranslation Feng Peng (f.peng@yahoo.com)Managing director Graham JohnsonCEO Tony Robinson

Tire Technology International ispublished quarterly together with an Annual Technology Review. The views expressed in the articles and technical papers are those of the authors and are not necessarily endorsed by the publisher. While every care has been taken during production, the publisher does not accept any liability for errors that may have occurred.

Subscriptions £50/US$105Four issues plus Annual Review

ISSN 1462-4729Tire Technology International

This publication is protected by copyright 2009.

Published by UKIP Media & Events Ltd, Abinger House, Church Street, Dorking, Surrey RH4 1DF, UK

tel: +44 1306 743744 fax: +44 1306 742525 editorial fax: +44 (0)1306 875824 email: tire@ukipme.com

Printed by William Gibbons, Willenhall, West Midlands, WV13 3XT, UK

Published by

2009

2

tire technolo

gy international 2009

•U

KIP Media &

EventsL

Holly Sundex® 8000 EU, the Oil of Choice™

for EU compliant tires, is the quick and

economical, high aromatic drop-in

replacement for tires and SBR Polymers.

Go ahead, drop it right in.

www.hollylubricants.com

Holly Refining & Marketing

Global tire production has exceeded one billion units annually for the past several years (excluding retreads and bicycle tires), despite the serious economic impact on the tire and auto industries over the last 12 months.

In 2008, the market value of worldwide production was about US$130 billion, with the center of gravity of production increasingly migrating east for business reasons in recent years. Production is equalled by the number of worn-out tires discarded each year, of which more than 85% in the developed economies are recovered for their energy content or geoengineering uses.

China recently emerged as the dominant country for manufacturing tires, and by 2006 produced more units than the traditional national leaders – the USA and Japan. As a region, Asia now produces a greater number of tires than North America and Europe combined, although the bulk of the units are entry-level products. Production of the petrochemical feedstocks required for synthetic rubber, mainly styrene and butadiene, is also shifting eastward, where natural rubber has always been readily available. These patterns will continue to persist as the majority of planned capacity increases for tires and their raw materials are scheduled for the developing markets. However, Bridgestone, Michelin, and Goodyear will most likely continue to supply 50% or more of global demand due to their geographically dispersed manufacturing, sales, and technical operations.

Stifling regulatory activity has steadily increased and remains firmly planted in Brussels and Washington despite the trend of shifting tire production to lower-cost countries and the semi-moribund state of the automotive sector. Tokyo, in contrast, has not been particularly pro-active with tire regulations at the governmental level, and seems to be taking a ‘wait and see’ approach.

The forthcoming EU environmental and safety directives mandate will have wide-ranging resonance. S-mark legislation limiting the sound levels emitted by tires is scheduled by the end of the 2009, followed by ‘clean oil’ legislation starting in 2010, with REACH chemical regulations ensuing later that year, and ultimately, rolling resistance, wet braking, and labeling requirements coming in 2012 – all with impact beyond the EU.

As early as 2005, before the recent proliferation of pending regulations, ETRTO tabulated 21 different global test methods required to obtain worldwide approval for the few procedures originally mandated in the now defunct FMVSS 109 (high speed, endurance, etc). Even countries with small markets have been issuing tire rules and regulations that seem to be an attempt to limit imports to protect domestic producers.

Pirelli Tyre’s CEO, Francesco Gori, stated that the world’s tire makers need to find ways to harmonize the growing number of national standards that threaten to swamp the tire industry in paperwork and added costs – and, I might add, in some cases with marginal benefits. Since Gori’s comments two years ago, the regulatory climate has only intensified.

In response to environmental concerns, the CEOs of the eight major tire manufacturers have initiated a multiyear research study to determine the impact of their products on human health. Reputable international risk-assessment organizations have received multimillion-dollar grants to determine the potential ecological hazards, if any, of tire wear particles. For example, each year about 2,000,000 tons of worn tread debris are released throughout the world, and questions have been raised for decades about the presence of such debris in the air we breathe and in roadside sediment that might leach into water supplies. Forthcoming reports should prove enlightening to all interested parties.

The pending EU ‘clean oil’ requirement affects extender oils rich in PAHs (polycyclic aromatic hydrocarbons) used for decades in tread compounds to facilitate processing, improve traction, and reduce cost. The tire industry has been addressing this issue since 1994 when the Swedish National Chemicals Inspectorate first publicized the toxicity of these materials. Such oils have already been phased out of winter and truck/bus tires. The ban covers oils containing more than 10 parts per million of listed PAHs. Non-listed replacements, such as naphthenic oils, are lighter and more highly refined. This means they are more expensive for tire producers and consumers, but should hopefully deliver comparable performance characteristics on the road.

REACH (Registration, Evaluation, Authorization, and Restriction of Chemicals) represents what is perhaps the most expansive and complex chemical regulation in the history of the tire industry and its many suppliers. It addresses the production and use of chemical substances, and their potential impact on human health and the environment. Tire raw materials are covered, with some exceptions.

Two important tire constituents are polymers and carbon black. Polymers per se are exempt, but their monomers are not. The seven major producers of carbon black reinforcing filler have recently formed a consortium and successfully submitted an appropriate REACH technical dossier and chemical safety report to ECHA (European Chemicals Agency). It now appears that all consortium companies will be fully registered well ahead of the November 2010 deadline. It would be interesting to unearth the cost of REACH compliance for just this one important segment of the tire manufacturing supply industry.

On labeling for tire performance, environmental concerns and highway safety, many non-harmonized national and international agendas are in play. The industry will have to contend with a hodgepodge of new regulatory schemes that frequently differ from one jurisdiction to another. Even similar measures often have varying implementation dates. In June 2009, the US NHTSA issued its 148-page NPRM (Notice of Proposed Rulemaking) for a national tire fuel efficiency rating system and consumer education program. In particular, a removable paper label must be affixed to every replacement passenger car tire at

ForewordThe regulatory climate intensifies

5

the point of sale, and information made available online, containing performance ratings for rolling resistance, tread wear, and wet traction. However, wear and traction are part of the existing UTQG system.

Once the new standard is implemented, NHTSA recommends retiring the current UTQG system – long overdue in my opinion. The defunct temperature grade (which is a surrogate for speed potential) would be replaced by a rolling resistance index, and the present tread wear and traction ratings would be retained. NHTSA estimates that it would take up to nine years to test every single passenger tire SKU for rolling resistance, at a cost to tire makers of US$18-50 million annually. It would enable manufacturers to self-certify one tire within a family (make/line/size range) and use that rating for the entire family. Winter and spare tires are excluded from the proposal. The California Energy Commission, with its perceived wisdom, has issued a competing, but different, proposal for tire fuel efficiency. Presumably, a federal mandate would pre-empt any state initiative.

On the other hand, the pending EU labeling regulation targets both tire performance and environmental measures: fuel efficiency, wet grip, and noise emissions. The EU legislation covers all passenger car, light truck and heavy truck/bus tires, and specifies maximum values for rolling resistance and noise, as well as a minimum value for the wet grip of passenger car tires. NHTSA prescribes the use of rolling resistance coefficients for rating tire efficiency, using a scale of 1-100, and the EU requires that the measured force of rolling friction be reported. The EU force value will vary directly with tire load, but the NHTSA coefficient is relatively load insensitive.

It is incongruous to expect that future production tires will have to be labeled with two different rolling resistance schemes. Somewhat fortuitously, the EU and NHTSA procedures are both based on the same pending ISO 28580 test method. The EU regulation also includes winter tire labeling – excluded by NHTSA. Other than highway noise generated by truck tires at elevated speeds, tire noise is not presently of much concern in the USA, and is unlikely to be further regulated by the EPA.

Heat, in the presence of oxygen, is the Achilles heel of all tires as it causes thermo-oxidative degradation of material properties that promotes tire aging. Over time, elevated operating temperatures support conditions that can cause tire failures that are exacerbated by underinflation, overload, and excessive speed.

NHTSA has recently estimated that as many as 600 highway fatalities in the USA may be attributable to tires – with a high number due to low pressure – a statistic leading to a previously issued TPMS safety standard for all new US cars and light trucks.

It is well known that flats and air loss caused by foreign object penetration are more likely to occur as tires wear over time because of reduced tread depth. However, the number of highway fatalities and crashes caused principally by tire aging, independent of wear, remains undetermined. Not surprisingly, field research shows that tires age and wear faster in regions with higher ambient temperatures. Several accelerated laboratory tire aging methods have been evaluated to determine their effectiveness in replicating natural aging characteristics. To date, the most encouraging results have been achieved with new tires inflated with oxygen-enriched gas exposed to elevated oven temperatures – followed by dynamometer testing of the tires. California, on the other hand, has been pursuing legislative mandates based solely on the chronological age of a tire,

as several automobile and tire manufacturers recommend not using PC tires past a certain limit, usually 6-10 years from the date of manufacture.

Some evidence seems to suggest that US trial lawyers are promoting

passenger tire age limits to engender potentially lucrative litigation. Interestingly, medium truck tires in many markets are designed to run one million miles or more using retreaded rubber on the original casing, and are often in service for over a decade. Even so, it’s doubtful that any legislator would have the temerity or political will to propose banning truck tires based solely on age – retreaded or not. Hopefully, logic would dictate that any federal tire aging standard based on controlled laboratory test procedures would trump any recommendations or regulations based on an artificially imposed chronological age limit.

Proliferating governmental policies seem to be today’s norm regarding tire regulations, even in a dour business climate. In economic terms, these actions raise barriers to new entrants into our mature industry. The large manufacturers, even with their present unremarkable financial results, will no doubt survive, perhaps prosper, and possibly absorb smaller competitors. However, as tire companies struggle with these additional non-productive administrative costs, I trust that important R&D programs are not compromised. tire

“It is incongruous to expect that future production tires will have to be labeled with two different rolling resistance schemes”

Dr Joseph D. WalterAdjunct professor, The University of Akron

6

www.tiretechnologyinternational.com

| News and exclusives | Supplier directory | Recruitment | Industry opinions | Image gallery | Read the latest issue online

goes digital!

Big things are happening in 2010.Tomorrow’s environmentally sound tyre oils are ready today.

Nynas is a different kind of oil company. We’re

focused on specialised oil applications – and in

this case that means tyre oils.

The European Union has stipulated that by

2010, tyres may no longer contain aromatic

extender oils. The replacement of these oils with

our environmentally sound alternatives will not

only meet European regulations, but also safe-

guard our environment. And keep your business

rolling. That’s what we call – Taking oil further.

www.nynas.com/tyreoils

Subject: “Curing presses”

harburg-freudenberger.compowered by commitment

More than just mechanical engineering.

Harburg-Freudenberger considers itself as a partner providing know-how and innovation for economic production solutions in line with mar-ket requirements – for example curing presses.

We developed the first hydraulic curing press and have strongly influenced current tire produc-tion processes with this tech-nology. We are still setting the standard – this also applies to our new column type curing press.

Contact us to find out more!

Section 1Design, materials, testing, and future possibilities

9

Predictions on future traffic policy suggest that the number of cars will double worldwide to 1.5 billion in the 30 years from 2000.

A corresponding increase in mileage is also expected.

As a consequence, innovative and sustainable solutions will have to be developed in the area of transport in particular to meet the economic and ecological challenges of the future.

The European Commission has developed harmonized regulations contained in COM2009/661,1 which have already been passed by the European Council and Parliament, and which guarantee a higher level of safety, while effectively protecting the environment from further emissions.

Starting on November 1, 2012, but depending on tire types, this regulation will define minimum requirements for wet adherence, maximum values for rolling resistance, and specified limits for noise emissions.

In a second step, the European Commission has proposed a supplementary regulation COM2009/348 of the European Parliament and of the

European Council on labeling of tires with respect to fuel efficiency and other essential parameters, aimed at providing the customer with information about the performance of selected tire criteria in the form of pictograms.

In addition to driver safety, the current, focus is on measures to reduce CO2emissions. As 25% of all CO2 emissions are generated by road traffic and approximately 20-30% of a vehicle’s energy consumption can be attributed to tires, it is clear that every endeavor must be made to reduce rolling resistance (RR).

This represents a challenge for tire manufacturers in particular. They have the task of creating a high-tech product that will combine the various criteria of the tire in an optimized way, despite existing interactions among the parameters.

Michelin’s tire development takes into account the fact that improving tire quality is part of an integrated concept. This is to ensure the optimization of the performance criteria mentioned above does not lead to a reduction in performance characteristics elsewhere.

Given these conditions, the use of innovative technologies and the

development and availability of new kinds of materials mean that the following tire performance criteria take on a key role in tire development: wear life, fuel economy, external noise, and grip (Figure 1).

Managing these kinds of complex relationships means granting sustainable mobility while providing protection to life and the environment. This results in improved road traffic safety and reduced noise and CO2 emissions. It is telling that in this context Michelin also mentions life mileage as an additional key criterion.

This is not simply because increased durability contributes considerably toward sustainable mobility – it also saves resources. The economic benefits are clear: the customer is given added value by the combination of fuel economy and wear life as a result of a reduction in the total cost of ownership (Figure 2).

Based on assumed driving conditions, the Energy Saver saves more than one-third in costs over other products (Figure 3). (The tire uses a rubber compound based on a new generation of silica and undergoes a specially developed mixing process that almost completely eliminates the need for carbon black. The design

Technological breakthroughs in the areas of wet adherence, rolling resistance and noise emissions will lead to improved safety and energy efficiency, and mileage

by Dr Joachim Neubauer, Michelin, France

Improving key tireperformance

Figure 1: Tire performance criteria with a key role in tire development Figure 2: Reduction in the total cost of ownership, i.e. tire price, fuel, and distance

10

helps to create a low RR of 20%, generating considerable fuel savings, resulting in a reduction in CO2emissions of nearly 4g/km.)

As a result, the tire and car industries are working together in this area, to maximize the cost-benefit effect from reduced fuel consumption and emissions on the one hand, and increased tire mileage on the other.

Let us put this statement into context: when a vehicle moves forward, the tires – determined by RR – account for up to 30% of the overall fuel consumption. What are the reasons for this?

Deformation appearing on tires, along with external affective parameters (such as stress, temperature, and pressure), have the greatest influence on RR (Figure 4).

The value of RR is determined in accordance with stipulation ISO 28580, which uses a specified procedure to guarantee a sufficiently precise reproduction of results. This is also a prerequisite for grading, which means that the determined RR levels can be divided into classes, and the result is made clear by labeling for all tire types according to COM2009/348. A calculator would finally provide customers with information about the degree to which the tire reduces fuel consumption and cuts CO2 emissions.

Michelin has, since as far back as 1992, advocated the development of the green tire and taken the reduction of RR very seriously, implementing it on the market. This has created economic as well as ecological savings: economic, by reducing fuel consumption and costs; and ecological, by reducing CO2 emissions.

The improved RR achieved by Michelin can be seen by the example of the Energy Saver, as shown in Figure 5.

Reference should be made in this context to the noticeable saving potential of tire’s RR, especially for electrically propelled vehicles. This effect is mainly seen in urban traffic, which accounts for approximately 33% of the overall energy consumption required to drive a vehicle forward.

What is common to both tire RR and the mileage of a tire is that resistance to wear is influenced essentially by the materials used, tire structure and manufacture, and tire design (Figure 6), and grip and noise characteristics are mainly dependent on the road structure and its surface structure, which again have a big influence on handling, both in longitudinal and in transversal effective direction (Figure 7).

Longitudinal and transversal grip are essentially determined by two different driving situations: a change in driving speed (longitudinal grip), and a change in driving direction (transversal grip).

When in direct contact with the roadway, the characteristics of the tread compound and the profile design of the tire are crucial for the adhesive mechanisms, such as interlocking and molecular grip.

Elastomers, such as those in the tread compound, have proved to be dependent on stress frequency and temperature with regard to module, and power loss (=hysteresis). With low excitation frequencies, the hysteresis is weak and the material shows more elastic qualities, whereas viscosity falls with increased frequency and the material can then hardly be shaped at all. This is the ideal area for tire grip, where hysteresis is at its greatest, equivalent to greater energy loss, which in turn results in increased tire RR.

Because the tire grip is also influenced by the structure of the road surface, changing friction coefficients on wet, damp, and dry road surfaces have to be considered.

The friction coefficient is smaller on damp layers than on dry surfaces and there is strong variance with the surface quality. A water film on the tread rubber inhibits the molecular grip, for instance, until the water film is broken. Under these conditions, only the interlocking remains in effect, which is why micro-rough road surfaces offer the best grip on damp roads. In this context, macro-roughness is of secondary importance.

As a consequence, tires with appropriate profile design, shape, and formation of the contact patch, siping, and the visco-elastic material characteristics of the profile blocks, have massive effects on grip and handling. Schematically, the design parameters and interaction parameters to be considered here are summarized in Figure 8.

This also explains why the results gained for the noise or wet-grip potential for a tire in outdoor tests are subject to considerable variation. This also applies to the tests to determine noise levels.

The tests conducted on different roads show that differences in noise levels can occur due to different surface compositions up to 10dB(A). In contrast, notable improvements to the tire, even on different road surfaces, are only registered up to 3dB(A) (see Figure 9).

As a consequence, and given the various consistency and surface composition of different road surfaces across the world, it is therefore very difficult to provide the end user with even approximately constant performance data for noise and braking distance.

Figure 3: Total cost of ownership: cost-benefit effect Figure 4: Deformation on tires and the effect on RR Figure 5: Improved RR achieved in the Energy Saver

Figure 7: Grip and noise are dependent on surface

Figure 6: Wear resistance is influenced by materials

11

Conversely, this also means that spatially limited problem zones can most certainly be solved by optimization measures, for instance where anti-skidding road surfaces assume a key role in eradicating accident black spots (wet safety), or noise-optimized surfaces become important in solving local noise problems, e.g. in residential areas.

As mentioned earlier, the development and implementation of new technology are a necessary precondition for considerable improvement in the performance of a tire, with the stipulation that different performance characteristics can be improved simultaneously and in the same direction.

With this in mind, the environmental benefit of longer-lasting tires quickly becomes clear: what is the point of a tire that is perfectly designed with regard to RR and wet braking, if it has to be replaced much earlier because of much lower mileage capacity? (Figure 10.)

And if this means that the energy and raw materials needed to produce the tire also have to be used far sooner? Economic aspects will also come to the fore here.

The ‘added value’ aspect gained by increasing mileage goes largely unreported in the media – which is all the more incredible considering the great interest in the conflict between tire RR and adherence.

In order to achieve low tire RR while maintaining the shortest possible braking distance in the wet and a high mileage, a fundamental understanding of tire’s rolling condition is absolutely necessary in order to influence the behavior of

the individual parameters in a targeted fashion.

Wet grip and RR are affected by hysteresis/energy loss, guided by completely different tire characteristics, as the phenomena also occurs in different frequency areas.

Deformations of the tire surface that generate the grip potential occur in excitation frequencies between 103Hzand 10**10Hz. Deformations in the tire structure, on the other hand, depend on every tire revolution: for a car tire, this is about 15 revolutions per second at a speed of 62mph (100km/h).

With the previously used materials of a standard ‘black tire’, the frequency

of energy absorption (=hysteresis) leads to a loss of grip when roll resistance is reduced, and vice versa.

For instance, the tires that Michelin first marketed in 1992 with a reference to ‘Green X-Technology’ already show a small energy loss in the low-frequency area relevant to RR, and a high grip potential for higher frequencies.

So, when using ‘intelligent materials’, the absorption curve (hysteresis) is characterized by combining a low RR with high grip potential; in the example shown, the crossing of the absorption curve takes place at 100-10,000Hz.

It becomes obvious that using innovative technologies in the area of development, design, and manufacture is the only way to optimize the key characteristics at the same time. A further example is provided by the optimization measures of wet-grip and wear, which will be explained below in light of the interaction mechanisms.

This is how Michelin implements innovations and combines safety with energy efficiency, simultaneously realizing an improved mileage. This benefits the end user in the form of added value. Other tire characteristics can also be improved and so there is a considerable expansion of the application and portfolio of the tire as a whole.

In this context, the fundamental realization is that any remaining problems can only be properly solved by cooperation with the vehicle and tire manufacturers, road building authorities, and companies all coordinating their optimization measures with one another.

By marketing the Energy Saver, Michelin has shown that it is possible to use a newly developed tire compound and matching process technology, to greatly improve fundamental performance characteristics, and to optimize them in the same direction.

And the optimization of these key indicators already mentioned will also serve as a spur to further innovation. This is an essential prerequisite for responding to the objectives of the European Commission’s initiatives with regard to the tire limits set in the vehicle safety regulation and the tire labeling on integrated performances, which aim to accelerate transformation toward even more efficient and safer tires. tire

1) Regulation of the European Parliament and of the Council concerning type-approval requirements for the general safety of motor vehicles

Figure 8: The design and interaction parameters to be considered to create a tire with good grip and handling

Figure 9: Improvements are registered up to 3dB(A)

Figure 10: Environmental benefit of long-lasting tires

12

Continental AGConti MachineryJädekamp 3030419 Hannover – Germany

Phone: +49 (0)511 976-6780Fax: +49 (0)511 976-6086

e-mail: cm@conti.deWeb: www.conti-cm.com

Excellent Technology To Meet Customers Targets

is offering products along the entire

production chain for passenger, light truck and truck tires.

14

For a vehicle to handle safely and use fuel economically, proper tire inflation, as recommended by the vehicle manufacturer, needs to

be maintained. Underinflation can cause high heat generation that in turn can cause rapid tire wear, tire blowout, and loss of vehicle control that may cause a crash. Tire pressure monitoring systems (TPMS) are believed to be an effective means to monitor the tire pressure – a claim that is statistically tested in this study. A comparison of vehicles equipped with TPMS and vehicles without TPMS can throw light on the effectiveness of this device.

Further, there are two types of TPMS: direct and indirect. Direct systems operate with a tire pressure sensor in each tire cavity, while indirect systems monitor tire pressure by comparing characteristics of tires, such as wheel speeds using the anti-lock braking system. Indirect systems do not distinguish between overinflation and underinflation. Therefore it is also of interest to assess whether direct TPMS is more effective than indirect TPMS.

In support of rule-making activities mandated by Section 13 of the TREAD Act, the National Highway Traffic Safety Administration’s National Center for Statistics and Analysis conducted the Tire Pressure Special Study (TPSS) and the Tire Pressure Monitoring System Study (TPMSS). The TPSS was designed to assess to what extent passenger vehicle operators are aware of the recommended tire pressures for their vehicles, the frequency and the means they use to measure their tire pressure, and how much the actual measured tire pressure deviated from the manufacturer’s recommended pressure. This study

is focused on the last aspect of tire pressure maintenance.

The TPMSS was designed to gather tire-pressure-related information on vehicles equipped with different kinds of tire pressure monitoring systems, so that their respective effectiveness could be evaluated. The main objective was to assess the effectiveness of TPMS in general, and investigate if direct TPMS is more effective than indirect TPMS. The present study conducts statistical analysis to estimate some tire-pressure-related statistics, as well as make inferences about the effectiveness of TPMS devices.

Data collection methodologyThe objective of TPMSS was to assess real-world tire pressure maintenance of vehicles in the United States. Accordingly, the data collection was planned to obtain a nationally representative sample of passenger vehicles, including passenger cars, light trucks, pickups, and sport utility vehicles. The target population for the survey consisted of two categories of vehicles: those equipped with the tire pressure monitoring system, to be referred to as the TPMS group, and the ones that were not equipped with TPMS,

to be referred to as the peer group. As mentioned earlier, there are two types of tire pressure monitoring systems: direct and indirect. Therefore care was taken to include both types of TPMS systems in the sampling frame. Finally, the vehicle age was also taken into account by including vehicles from model years 1997 through 2003 in the two groups.

The survey was conducted through the infrastructure of the National Automotive Sampling System (NASS) Crashworthiness Data System (CDS). As in the NASS-CDS data collection system, the TPMSS data were collected from 24 primary sampling units (PSU). The sample was selected from the state registration files and was comprised of vehicles from the TPMS group as well as from the peer group as determined with the assistance of the Alliance for Automobile Manufacturers.

The peer group was formed by including vehicles that were not equipped with TPMS, but were of the same model years and of similar body styles and price ranges as the vehicles selected in the TPMS group. A computer program randomly selected study vehicles (TPMS and peer) from the list of those that were eligible.

Case weights can be used to obtain nationally representative estimates of under and overinflated vehicles, with and without TPMS

by the National Highway Traffic Safety Administration, USA

Tire pressure maintenance:a statistical investigation

Table 1: Sample segmentation by vehicle groups and their subcategories

Planned sample size† (Actual sample size)*

Total subject vehicles 12,001† (2,316)*

TPMS-equipped (TPMS) 5,977† (1,259)*

Direct TPMS 1,261† (213)*

Indirect TPMS 4,716† (1,046)*

Without TPMS (Peer) 6,024† (1,057)*

Direct Peer 1211† (243)*

Indirect Peer 4,813† (814)*

15

Target informationInformation on several tire-pressure-related variables, such as actual pressure in all the tires, the recommended tire pressure levels for the vehicle, ambient temperature, age of the vehicle, and age and sex of the vehicle owner/operator was collected in this survey. To assess the effectiveness of monitoring devices, the information on the categorical variables: TPMS (presence or absence), TPMS type (direct or indirect) was documented as well. This study is focused on assessing the effectiveness of TPMS and uses the tire-pressure-related variables only.

Sample sizeOriginally the data were planned to be collected on a sample of 12,001 vehicles. Anticipating the response rate of the survey to be 60%, the data on 7,000 vehicles was the actual target. However, the survey was not completed and data was collected on 2,316 vehicles only. The allocation of the originally planned sample of 12,001 vehicles over peer and TPMS groups, as well as their subcategories, direct and indirect, is shown in Table 1. This table also shows (within parentheses) similar allocation of the final sample of 2,316 vehicles.

Given the reduced sample size, it was of interest in this study to assess how well the actual sample could represent the population. This is done by comparing the ratio Rplanned of planned sample size to the corresponding PSU size with the ratio Ractual of actual sample size to the corresponding PSU size. Figure 1 presents profiles of these two ratios over PSUs in which data were collected. The patterns of the two ratios in this figure show that the number of vehicles actually documented in the sample is consistently proportional to the planned PSU sample sizes. However, this does not completely make the sample nationally representative in certain other aspects of the sample design. Therefore, an adjustment needs to be made through case weights to achieve a reasonable level of national representation. The following section is devoted to developing these weights.

Sample design and case weightsA two-stage stratified sampling scheme was used to select the study vehicles. The first stage of sampling consisted of selecting geographic areas, such as counties, groups of counties, or cities forming PSUs. These PSUs are

a probability sample selected from a frame of all geographic areas in the continental United States. Twenty-four PSUs were selected for this survey from among the 1,195 determined by the NASS system. The selection is based on the number of motor vehicle traffic crashes occurring within these regions.

At the second stage, stratification was done with respect to the vehicle’s age (model year) and vehicle category (in terms of the presence or absence of TPMS) in each selected PSU.

Finally, from the selected PSU as in the first stage, a simple random sample of vehicles was selected from each stratum formed at the second stage.

The objective of the TPMSS was to collect a sample of vehicles with and without TPMS that is representative of the US fleet. A survey conducted according to a designed sample and the use of appropriate case weights would help achieve this objective. Although the sample actually selected in this survey has some affinity with the population as discussed earlier, due to early termination the probabilities of selection at different stages of the design are no longer valid. Consequently the case weights based on these probabilities would fail to make the sample representative of the population. This study proposes the methodology to compute case weights for the terminated survey that can be used for estimating the tire-pressure-related parameters of the target population.

Typically the case weights in a sample survey are based on the scheme used for the case selection: in the two-stage sample design used in this survey, it would be the inverse of the product of the probability of PSU selection and the probability of vehicle selection from a stratum. Originally the weight for the ith case selected randomly from the jth stratum of the kth selected PSU was defined as:

where Pk = Prob {kth PSU is selected}, Figure 1: Comparison between actual sample sizes and planned sample sizes for the 24 selected PSUs

16

Pij/k = Prob {ith vehicle is selected from jth stratum, given that kth PSU has been selected}.

Although the survey was terminated prior to completion of data collection based on the laid-down sample design, the PSU selection had already been made. This preserves the probability Pk of PSU selection.

However, due to early termination the selection of vehicles in the strata was disturbed: fewer than planned vehicles could be selected in the strata. This questions the validity of the corresponding selection probability Pij/k.To introduce a reasonable amount of national representation, revision of these probabilities is imperative. What is needed is to account for what was actually selected in a stratum as against what was supposed to be selected. This suggests that the posterior rather than the prior distribution should be used for computing the vehicle selection probability Pij/k to be used in computing weights as in (1). Based on this rationale, this study proposes a Bayesian methodology to compute case selection probabilities Pij/k, and hence the case weights Wij/k.

Consider the sampling design layout for an exemplar PSU as shown in Table 2, where nij is the sample size originally planned to be collected from the stratum (ij) of the PSU and mij is the sample size actually collected from the ijth stratum.

The case selection at each stage is basically a Bernoulli process with

probability of selection (success) determined by the ratio of the number to be selected to the number from which to be selected. Thus if in a stratum of Nijvehicles, nij are to be selected, then the probability of selection (success, speaking in terms of Bernoulli trials) is the ratio ( = nij/Nij). However, instead of nijvehicles, only mij (mij nij) could be selected by the time the survey was terminated. Speaking in terms of the Bernoulli trials, this amounts to having mij successes in nij trials with probability of success given by . It is this probability that needs to be revised using the prior

information about the sample size. Bayesian approach is used for this purpose as outlined in the appendix. The probability that maximizes the selection of ith vehicle from the jth stratum of the kth PSU, that is the mode of the posterior distribution [1], is given by:

Finally, in the two-stage selection process used in the survey, the probability of selection of a vehicle is the product of the probability of selection Pk of the kth PSU and the posterior probability given by (2), that is:

However, some adjustments are needed before this posterior probability of case selection could be used in calculating case weights. It should be noted that the strata in which the number of vehicles actually selected is equal to the number that was originally planed for a stratum do not require the Bayesian treatment. Thus the case weights are calculated using the formula (1) or (3), depending on whether the number of vehicles actually selected is equal to or less than the number of vehicles planned to be selected. Non-response in a survey is not uncommon. To account for this contingency, the weights need to be adjusted by using the response rates in the selected strata. Finally, the case selection probabilities and hence the case weights are calculated by:

Table 2: Stratification in an exemplar PSU

Vehicle Age Vehicle group

With TPMS (Study group)

Without TPMS(Peer group)

Less than three years n11 n12(New vehicles) m11 m12Three years or older n21 n22(Old vehicles) m21 m22

Figure 2: Percent frequency distributions of underinflated and overinflated vehicles

Table 3: Allocation of sampled vehicles and their estimates over vehicle groups

(Actual sample size)* [Weighted estimate]‡

Sampled number of vehicles and their weighted estimates

(2,316)* [654,817]‡

TPMS-equipped (TPMS) (1,259)* [332,046]‡

Direct TPMS (213)* [74,595]‡

Indirect TPMS (1,046)* [257,451]‡

Without TPMS (Peer) (1,057)* [322,771]‡

Direct Peer (243)* [71,693]‡

Indirect Peer (814)* [251,078]‡

17

where rj×100 (rj 1)% is the response rate in the jth stratum.

Using these weights, the number of vehicles in the peer and TPMS categories and their subcategories direct and indirect are estimated as shown in Table 3. The statistics presented in Table 3 show that in an estimated total number 654,817 of vehicles, 332,046 are estimated to be TPMS-equipped and 322,771 without TPMS. Of the TPMS-equipped, 74,595 vehicles are estimated to be equipped with direct TPMS, while an estimated 257,451 vehicles are equipped with indirect TPMS.

AnalysisThe measured tire pressure being below or above the recommended pressure level is all that matters.

There are two ways an improper tire inflation could be hazardous: overinflation and underinflation.

While overinflation can be dangerous for vehicle handling, underinflation can become a safety hazard not only due to blow-outs and tire destruction, but also due to the development of dangerous driving scenarios, such as directional loss of control. This brings out the importance of some means by which vehicle operators can determine if the tires need to be inflated or deflated. One of the devices that can be used for this purpose is TPMS. This study is focused on assessing the effectiveness of TPMS as a means to maintain tire inflation at the recommended levels, as well as studying the comparative effectiveness of direct and indirect TPMS. One of the ways this can be done is to compare the TPMS group of vehicles with the peer group and TPMS direct subgroup with TPMS indirect subgroup. This study makes these comparisons in terms of the frequencies of underinflated and overinflated vehicles for the two groups, as well as their average underinflation and overinflation. Statistical analysis is conducted to estimate some tire-inflation-related

parameters using the proposed weights and to test certain hypotheses to compare groups of vehicles with and without TPMS, as well as their respective subgroups, direct and indirect.

The tire-pressure-related parameters that can be used to determine underinflation and over-inflation are the manufacturer’s recommended tire inflation pressure and the one actually measured in the survey for each tire of a vehicle.

The difference between the two is used to determine the extent of underinflation and overinflation of a tire:

• Underinflation = Recommended pressure - Measured pressure ( 0, recommended exceeds measured)

• Overinflation = Measured pressure - Recommended pressure ( 0, measured exceeds recommended)

For analysis purposes, these are converted into percentages (percentage of the corresponding recommended pressure of the subject tire) for each tire of the vehicle.

With regard to the tire pressure status of a vehicle on the whole, an underinflated vehicle in this study refers to a vehicle that has at least one tire underinflated. The same criterion is used in the case of overinflation. It should be

noted that one vehicle can have underinflated and overinflated tires.

While the variables underinflation and overinflation are computed for all the tires of a vehicle, the minimum of all tire pressures in the case of underinflation and maximum of all tire pressures in the case of overinflation are used as measures of vehicle underinflation and overinflation, respectively.

Preliminary analysisThe profiles of two groups over underinflation and overinflation with threshold values 0, 5, 10, 15, 25, 35, 45, 75, and 100% (expressed as % of the recommended pressure) are presented in an Ogive. An Ogive is a line graph that depicts the percentage of cases (vehicles) that have values of the variable (underinflation or overinflation) less than or equal to a certain threshold value. In this study, to assess the effectiveness in terms of reduction in underinflation or overinflation, the percentage of cases greater than a certain threshold value is used. Hence, instead of Ogive, its image, to be referred to as reverse Ogive, is used. Obviously a reverse Ogive is a line graph depicting the number of vehicles greater than the threshold value. Figure 2(a) and

Figure 3: Percent vehicles with underinflation and overinflation exceeding threshold values, 0, 5, 10, 25, 35, 45%

“To assess the effectiveness in terms of reduction in underinflation or overinflation, the percentage of cases greater than a certain threshold value is used”

18

Figure 2(b), respectively, show underinflation and overinflation profiles of the two groups. Figure 2(a) shows that about 61% of the peer vehicles were found underinflated (Underinflation=0), while a much smaller percentage (48.2) of TPMS vehicles had one or more underinflated tires. For other levels of underinflation (5, 10, 15, 25, 35%), too, the TPMS-equipped vehicles demonstrate lower frequencies of underinflated vehicles than the peer vehicles. The overinflation profiles of the two groups over the threshold values 0, 5, 10, 15, 25, 35, 45, 75, and 100 presented in Figure 2(b) indicate lower percentages of peer overinflated vehicles as compared with the TPMS-equipped vehicles that were overinflated.

The performance of the TPMS group is compared with the peer group also over different ranges of underinflation. The results of this comparison are presented respectively in Figure 3(a) and Figure 3(b). Figure 3(a) shows that while the percentage of vehicles with zero underinflation is higher for the TPMS group, it is lower for other ranges of underinflation. In the case of overinflation, the frequency distributions in Figure 3(b) show that more of the peer vehicles had no overinflation as compared with the TPMS-equipped vehicles.

Comparison of vehicle groups and subcategoriesThe efficacy of a tire pressure monitoring device is reflected primarily in the level of underinflation it allows before illuminating the low pressure warning lamp. The following analysis is focused on three aspects of tire inflation – correct tire inflation, underinflation, and overinflation – and compares the TPMS and peer group of vehicles for each of these three tire conditions. Specifically, the TPMS and peer group of vehicles are compared with respect to Figure 3.

Correct pressureOne of the indicators of effectiveness of TPMS would be its contribution in aiding the operator of a vehicle to keep inflation of its tire at the correct level, which means none of its tires is under-inflated or overinflated with reference to the recommended tire inflations.

The estimates based on the TPMSS data show that among vehicles with correct pressure (in the above sense) about 57% are TPMS-equipped, which is much higher than 43% for the peer

vehicles. However, correct pressure on a vehicle is an ideal scenario. The following analysis determines how effective TPMS could be in warning the vehicle operator against underinflation or overinflation of a vehicle’s tires.

UnderinflationWith regard to TPMS function in warning against underinflation, the pressure characteristics considered in this analysis include: 0 (no underinflation), underinflation more than 25%, and underinflation more than 30%. The percentage frequency distribution of vehicles is presented in Table 4.

The statistics in this table show that 57% of TPMS vehicles have no underinflation as compared with 43% peer vehicles without an underinflated tire. Further, among vehicles with no underinflation, there are 7.6% peer (direct) and 12.8% TPMS (direct) vehicles. Similarly, only 35% of the vehicles are without an underinflated tire in the peer (indirect) group and about 45% in the TPMS (indirect) group. These statistics show that whether the TPMS device is direct or indirect type, it does aid in preventing vehicle underinflation to a large extent.

The estimates presented in Table 4 also project how the two groups compare in terms of underinflation (i.e. underinflation > 0). Among such vehicles, the percentage 44.9 of TPMS vehicles (9.5% direct and

35.4% indirect) is well below the percentage 55.1 of the peer vehicles that consists of 13.7% direct and 41.5% indirect type. These percentage differences between the underinflated TPMS and peer vehicles further support the claim, TPMS is effective in preventing underinflation.

The two groups are also compared for underinflation above threshold levels of 25 and 30%. Among vehicles with more than 25% underinflation, the percentage (46.2%) for the TPMS vehicles is much smaller than the percentage (53.8%) for the peer vehicles. Similarly, among vehicles with more than 30% underinflation, the percentage (43.8) of TPMS vehicles is much smaller than the percentage (56.2%) of the peer vehicles.

Statistical tests performed on these differences show (with 95% confidence level) that the data provide sufficient evidence in favor of the TPMS in tire pressure maintenance in terms of underinflation. However, no significant difference was observed between the two groups for underinflation greater than 35%. This is obvious, as underinflation to that extent cannot go unnoticed, whether or not the vehicle is equipped with the TPMS.

Additionally, another parameter of tire inflation has been looked at in this study: namely the average underinflation; the two groups are compared in terms of the average underinflation. As a word of caution, an average presented in this table

Table 4: Underinflation status by vehicle group and device type

VEHICLE GROUP DEVICE TYPE UNDERINFLATION

No Yes More than 25% More than 30%

Peer Direct 7.6 13.7 9.7 11.2

Indirect 35.0 41.5 44.1 45.0

Sub-total 42.6 55.2 53.8 56.2

TPMS Direct 12.8 9.5 8.0 10.9

Indirect 44.6 35.4 38.2 32.8

Sub-total 57.4 44.9 46.2 43.7

Table 5: Average underinflation (percent) by vehicle group and device type

VEHICLE GROUP DEVICE TYPE UNDERINFLATION

Average 95% Conf. Interval

Peer Direct 15.77 [14.02, 17.51]

Indirect 16.10 [14.34, 17.87]

Overall 16.01 [14.52, 17.50]

TPMS Direct 12.61 [10.16, 15.06]

Indirect 14.70 [13.15, 16.24]

Overall 14.30 [12.85, 15.75]

19

is an average over the vehicle group rather than the average over the vehicle. The estimates of averages presented in Table 5 are used for this purpose. These results show that on average, TPMS vehicles have only 14.3% underinflation as compared with 16% for the peer vehicles. The difference between 15.8% for the peer (direct) group and 12.6% for the TPMS (direct) group further shows that on average, direct type of TPMS contributes to a significant reduction in underinflation.

OverinflationThe statistics showing comparison between the two vehicle groups in terms of overinflation are presented in Table 6. These results show that, in general, the TPMS-equipped vehicles tend to be overinflated. For example, there are 55.2% peer vehicles and 44.8% TPMS vehicles among vehicles that have no overinflation (Overinflation=0). Similarly, among the over-inflated (Overinflation=0) vehicles, 46.5% are peer vehicles in con-trast with 53.5% TPMS vehicles.Also, among vehicles with more than 25% overinflation, TPMS vehicles have larger representation, 54% as compared with 46% for the peer vehicles. Similarly, among vehicles with more than 30% overinflation, TPMS vehicles have larger representation: 51.8% as compared with 48.2% for the peer vehicles.

In terms of the vehicle group averages, there is insignificant difference between the averages: 15.4% for the peer vehicles and 15.5% for the TPMS vehicles (Table 7). However, the comparison between the averages 12.3 and 13.9%, respectively, of the TPMS (direct) and peer (direct) groups shows that on average, direct type of TPMS aids in keeping the level of underinflation low. Thus, although a higher percentage of overinflated vehicles belong to TPMS, on average, the overinflation is lower for direct TPMS vehicles as compared with direct peer vehicles.

TPMS effectiveness analysisIn this section, the overall effectiveness of TPMS is studied by considering ranges of underinflation/overinflation, simultaneously. A simultaneous test is performed to test if TPMS retains its effectiveness as under/overinflation increases from 0% through 100%. Descriptive analysis of the data shows that estimates of the skewness coefficients (2.9 for the peer and 5.2 for the TPMS group) differ significantly from 0, which is the skewness of a normal distribution. This suggests using multi-category-based non-parametric analysis to compare the two groups with respect to TPMS effectiveness. Specifically, the non-parametric Wilcoxon Rank-Sum Test [2] is used to test the significance of the effect of TPMS, i.e. to test the hypothesis:

H0 : mtpms = mpeer against the alternative HA : mtpms < mpeer

where m stands for the average underinflation or overinflation, with subscript indicating the group. This test assumes that the models associated with the two samples differ in terms of location, the difference being attributable to the effect of TPMS.

Assuming that the distributions of Upeer and Utpms have the same shape and spread, the testing is done only in terms of shift in location. The test basically concerns testing the hypothesis H0: = 0 against the alternative HA: < 0. Due to large sample size, normal approximation of the rank sum statistic is used, i.e. the statistic:

where ntpms and npeer are the numbers of TPMS and peer vehicles, respectively, in the samples. The statistic W* has an asymptotic N(0, 1) distribution.

Controlling underinflationTo conduct the above analysis for underinflation, all subject vehicles are categorized depending on the level of underinflation. The analysis is conducted using the variable Effectiveness, defined in:

The Wilcoxon two-sample test statistic W*equals 9.817*1010 which is the sum of the Wilcoxon scores for the smaller group (peer). This sum is greater than the expected value 1.050*1011 underthe null hypothesis of no difference between the two groups.

This difference, with Z-value= -95.3894 and one-sided p-value < 0.0001 shows that TPMSis significantly effective in keeping the underinflation low for all ranges of underinflation.

Effectiveness – controlling overinflationTo conduct similar analysis for overinflation, all subject vehicles are categorized depending on the level of overinflation.

The analysis is conducted using the variable Effectiveness, defined in:

Table 6: Overinflation status by vehicle group and device type

VEHICLE GROUP DEVICE TYPE OVERINFLATION

No Yes More than 25% More than 30%

Peer Direct 13.6 9.6 8.1 11.3

Indirect 41.6 37.0 37.9 37.0

Sub-total 55.2 46.6 46.0 48.2

TPMS Direct 8.5 12.3 8.1 2.0

Indirect 36.3 41.2 45.9 49.7

Sub-total 44.8 53.5 54.0 51.8

Table 7: Average overinflation (percent) by vehicle group and device type

VEHICLE GROUP DEVICE TYPE OVERINFLATION

Average 95% Conf. Interval

Peer Direct 13.93 [11.08, 16.79]

Indirect 15.80 [14.26, 17.33]

Overall 15.37 [13.91, 16.83]

TPMS Direct 12.26 [10.54, 13.97]

Indirect 16.39 [14.81, 17.96]

Overall 15.52 [14.19, 16.85]

20

The Wilcoxon two-sample test statistic W*equals 1.108 * 1011, which is the sum of the Wilcoxon scores for the smaller group, peer. This sum is greater than the expected value 1.049* 1011

under the null hypothesis of no difference between the two groups. This difference, with positive Z-value = 80.6136 and the one sided p-value < 0.0001 shows that vehicles equipped with TPMS tend to be more overinflated as compared with those that are not.

Results and discussionAlthough the TPMS survey collected data only on a portion of the original sample size, the tire-pressure-related information is available for over 2,000 vehicles. This study found that when this sample was distributed over all PSUs, it resulted in almost the same proportions as the originally planned sample.

The actual sample consists of about 200 vehicles equipped with the direct TPMS and about 1,000 vehicles equipped with the indirect TPMS. Similar categories of peer vehicles are proportional to those in the TPMS group. However, the actual sample in hand at the time of termination could not conform to the second stage sample design, thereby lacking national representation – the case weights based on the original sample design could no longer be used for estimation purpose. Using a Bayesian approach, this study develops case weights that take into account the PSU size, strata sizes, planned sample sizes, as well as the actual sample sizes. The proposed weights are useful in making the best of the information that would otherwise be considered merely anecdotal. The data is analyzed by comparing the recommended pressures of each vehicle with its measured tire pressures, thus arriving at vehicle underinflation and overinflation. The proposed weights are used to obtain several estimates, such as percentages of underinflated and overinflated vehicles for TPMS and peer groups, as well as their subcategories, direct and indirect. The

significance of the differences in these percentages is statistically tested.

The analysis results show that the percentage of vehicles with correct pressure are much higher (57%) for the TPMS group as compared with 43% for the peer group. Regarding underinflation, while about 45% of the underinflated vehicles belong to TPMS group, a much higher percentage (55%) is attributed to the peer group. The analysis conducted for more than 25 and 30% underinflation shows similar differences. The differences between percentages are even more prominent when direct types of the two groups are compared with respect to no underinflation, and underinflation greater than 0% and 25%. In terms of the average underinflation, the TPMS-equipped vehicles have significantly lower average (14%) as compared with 16% for the peer vehicles. The difference between averages is more significant when direct type of TPMS and peer vehicles are compared, while the difference in the case of indirect types is insignificant.

Analysis conducted for overinflation shows that more TPMS-equipped vehicles are overinflated (53%) as compared with 47% peer vehicles that have at least one tire overinflated. The difference between percentages of the direct type TPMS and peer vehicles with overinflation more than 25% is not significant. The difference between 11% for the direct peer subgroup and 2% for the direct TPMS is highly significant. Comparison of averages for the two groups shows that while the overall difference is insignificant, the average overinflation – 12% for the direct TPMS – is significantly lower than 14% for the direct peer vehicles.

Analysis conducted to assess the effectiveness of TPMS shows that while this monitoring device is effective in aiding the operator to prevent significant underinflation, it is likely to result in overinflation, though within the safe limits supported by the above analyses.

Statistical analysis performed on the survey data provides evidence in favor of the TPMS, especially in the favor of direct TPMS. NHTSA recommends that vehicle operators check their tire pressures at least once a month and before long trips. The vehicle

manufacturer’s recommended tire pressure can be referenced for this purpose.

TPMS should be used as a supplement to regular tire maintenance. It should be noted that TPMS studied here may not be representative of the current designs.

AppendixThis appendix provides analytical details of the methodology used in revising case weights based on the rationale discussed previously. Speaking in terms of the Bernoulli trials, obtaining mij vehicles instead of planned number nij amounts to having mij successes in nij trials. This in turn means that the probability of selection of mij vehicles from nij vehicles in the jth stratum of the kth PSU is given by the Binomial distribution:

To use a Bayesian approach, a nearly non-informative prior for is used and is proportional to:

This further results into the posterior distribution which is proportional to:

After substitution of the appropriate normalizing constant, the corresponding posterior distribution for assumes the form of Beta distribution, i.e.:

p( /mij) in (8) is maximized at:

* given by (9) is the mode of the posterior probability distribution associated with the selection of ith

vehicle from the jth stratum in the kth

PSU. This is the probability that maximizes the selection of the ijth

case and can be used as the revised probability of selection of a vehicle. tire

References1) Box, G.E.P. Bayesian Statistical Analysis

(John Wiley & Sons, New York, 1980)2) Hollander, M. and Wolfe, D. A. Nonparametric

Statistical Methods (John Wiley & Sons, New York, 1972)

22

A lthough silica is not a petrochemical-based product, it consumes a lot of energy for its production. Along with

carbon black, the other filler mainly used in rubber tire applications, silica has high density, which has negative impacts on the fuel efficiency of vehicles. Therefore tire manufacturers are seeking new reinforcing fillers that are inexpensive, readily available, light in weight, and renewable.

Cellulose is the most abundant natural polymer; it makes up 40-44% of wood.1

Commercial cellulose fibers are mainly made from cotton and wood.

Short cellulose fibers have been studied as reinforcing fillers in rubber composites in the past 25 years.2-5 Cellulose fiber-reinforced rubber composites are mainly used for making ropes, hose, belt, mats, and insulation, and have yet to be used in tire applications.4, 6

Cellulose contains both amorphous and crystalline regions. Crystalline cellulose is much stronger and stiffer than amorphous cellulose and cellulose itself, and is supposed to be a better reinforcing filler than cellulose. Microcrystalline cellulose (MCC) is crystalline cellulose and is typically derived from the acid hydrolysis of cellulose.7

This is a low-cost process for production of MCC. The acid hydrolysis preferentially removes the amorphous regions of cellulose. When compared with glass fibers, silica and carbon black, MCC as a reinforcing filler in composites

has many advantages: low cost, low density, ease of processing, low abrasion to equipment, renewability, and biodegradability.8 There have been some studies on MCC as a reinforcing filler in plastic composites in recent years.9-11

However, very little has been reported about using MCC as a reinforcing filler in rubber composites.

In this study, MCC was used as a third filler to partially replace silica in styrene butadiene rubber (SBR) and polybutadiene rubber (BR) composites. The effects of MCC loadings in rubber

composites on energy consumption during compounding, rheological properties, mechanical properties, heat resistance, and dynamic mechanical properties were studied in detail.

ExperimentsTable 1 shows the materials used in this study, including code, trade name, manufacturer, and location of the manufacturer. The following two types of rubbers were used: styrene butadiene rubber (SBR) and polybutadiene rubber (PB). Three different fillers were carbon

The partial replacement of silica with microcrystalline cellulose reduces considerably the energy required for dispersion of fillers in a rubber matrix, and lowers the internal temperature during compounding

by Wen Bai & Kaichang Li, Department of Wood Science and Engineering, Oregon State University, USA

Partial replacement of silica with microcrystalline cellulose in rubber composites

An OSU doctoral student develops rubber composites using a new technology that incorporates microcrystalline

cellulose – an approach that may lead to tires that cost less, perform better and produce improved mileage

23

Table 1: Materials in this study

Code Trade name Manufacturer Location

Rubber SBR Duradene 751 Firestone Polymers LLC Akron, Ohio

PB Taktene 220 Lanxess Corp Orange, Texas

Filler CB N220 Side Richardson Carbon Company Akron, Ohio

Silica Ultrasil at GR7000 Degussa Corp Akron, Ohio

MCC Avicel PH 101 FMC Inc Philadelphia, Pennsylvania

Coupling agent Silane SCA985a Struktol Company of America Stow, Ohio

Process aid PA JB46F Struktol Company of America Stow, Ohio

Activator ZnO Zinc oxide Horsehead Corp Monaca, Pennsylvania

StAc Stearic acid Textile Rubber & Chemical Co Nurnberg, Germany

Antiozonant 6ppd 6PPDb Western reserve chemical Corp Stow, Ohio

Processing oil Sun Sundex 790 Sun Oil Company Aartselaar, Belgium

Accelerator DPG Perkacit DPGc Flexsys America LP Akron, Ohio

CBS Santocure CBSd Flexsys America LP Akron, Ohio

Vulcanizer S Rubbermakers sulfur R.E. Carroll Inc Akron, Ohio

a Bis(3-triethoxysilyipropyl)disulfide

b N-(1,3-dimethyl)N’-phenyl-p-phenylenediamine

c Diphenylguanidine

d N-cyclohexyl-2-benzothiazolesulfenamide

Table 2: Composition of rubber composites and compounding procedure

Ingredient Percent Batch 1 0MCC (phr) Batch 2 5.6MCC (phr) Batch 3 11.8MCC (phr) Batch 4 17.6MCC (phr) Batch 5 23.5MCC (phr)

1st passa

SBR 80.45 80.45 80.45 80.45 80.45

PB 19.45 19.45 19.45 19.45 19.45

CB 7.78 7.78 7.78 7.78 7.78

Silica 58.3 54.41 50.7 46.7 42.7

Silane 3.89 3.89 3.89 3.89 3.89

PA 2.33 2.33 2.33 2.33 2.33

ZnO 1.94 1.94 1.94 1.94 1.94

StAc 1.56 1.56 1.56 1.56 1.56

6ppd 1.56 1.56 1.56 1.56 1.56

Sun 3.89 3.89 3.89 3.89 3.89

MCC 0 3.89 7.70 11.6 15.6

2nd passb

DPG 1.07 1.07 1.07 1.07 1.07

CBS 1.30 1.30 1.30 1.30 1.30

S 0.92 0.92 0.92 0.92 0.92

Total 184.44 184.44 184.54 184.44 184.44

a Compounding procedure of 1st pass (master batch 1/MB1): 100rpm and 0.207 MPa (30 psi). (1) Added rubber and mixed it for 30 sec. (2) Added half silica, half MCC and silane, and then mixed them for 45 sec. (3) Added the left silica, the left MCC and Sun, and then mixed them for 50 sec. (4) Added CB, PA, ZnO, StAc, 6 ppd, and then mixed them for 25 sec. (5) Swept, auto ram lifted, and mixed for another 350 sec. 6. Discharged at 55rpm and 0.207 MPa (30 psi).

b Compounding procedure of 2nd pass (master batch 2/MB2): 77rpm and 0.207 MPa (30 psi). (1) Loaded 1/2 MB1, curatives, and ½ MB1. (2) Mixed for 30 sec and sweep. (3) Discharged either at 120 sec or 100°C.

24

black (CB), silica, and microcrystalline cellulose (MCC). The coupling agent was bis(3-triethyoxysilylpropyl)disulfide. Two types of activators were zinc oxide and stearic acid. The antiozonant was N-(1,3-dimethyl)-N’-phenyl-p-phenylenediamine(6PPD). The processing oil was a product with a trade name of Sundex790. There were two kinds of accelerators: diphenylguanidine (DPG) and N-cyclohexyl-2-benzothiazolesulfonamide(CBS). The vulcanizer was sulfur (S). The curing agents included DPG, CBS, and S.

Compounding recipe and procedureA Banbury 1600 internal mixer (Farrel Corp, Ansonia, Connecticut) was used for compounding a rubber mixture, which was called master batch. The temperature of the mixer was set and maintained at 65°C. There were five batches compounded according to the recipes and procedures in Table 2. The weights of all ingredients in the recipes were in units of per hundred rubber (phr). Batch 1 was the control, which did not contain MCC.

The percentages of silica replaced by MCC were 5.6%, 11.8%, 17.6%, and 23.5%, corresponding to Batch 2, Batch 3, Batch 4, and Batch 5. There were two passes to complete the compounding. After each pass, the master batches were made into sheets with the thickness of 2-4mm by an open rubber mill (Stewart Bolling Co, Cleveland, Ohio).

Evaluation of viscoelastic properties of unvulcanized rubber compositesMooney viscosity and scorch tests were both carried out with Monsanto Mooney Viscometer 2000 (TechPro Inc, Cuyahoga Falls, Ohio). Mooney viscosities were measured with a large rotor at an oscillating rate of 0.2rad/s. Initial viscosity was recorded immediately after preheating a rubber composite specimen for one minute at 100°C. Mooney viscosity at 100°C was recorded after preheating the specimen for one minute and the total testing time was four minutes according to ASTM D1646. The initial viscosities and Mooney viscosities of aged rubber composites were obtained after unvulcanized rubber composites were stored at room temperature for one week and two weeks, respectively.

The scorch test was carried out in accordance with ASTM D1646 as well. The testing temperature was 125°C and the total testing time was 60 minutes. The temperature was set at 125°C because the processing temperature in rubber manufacturing could ascend by incident. The times required for increasing Mooney viscosity above its minimum viscosity by 5, 10, and 35 units were designated as t

5,

t10

, and t35

. The total volume of two pieces of specimens for each test was around 25±3 cm3. Two pieces of specimens were

Table 3: Initial viscosity and Mooney viscosity (MV) for unaged and aged rubber composites at 100°C

Batch Unaged One-week aged Two-week aged

Initialviscositya

MVb c Initial viscosity

MV IV1c

MV1c Initial

viscosityMV IV2

cMV

2

1 99.3 76.4 -22.9 119.6 80.9 -38.7 20.3 4.5 127.6 83.3 -44.3 28.3 6.9

2 97.7 72.5 -25.2 116.8 76.6 -40.2 19.1 4.1 119.7 78.1 -41.6 22.0 5.6

3 98.8 68.4 -30.4 105.7 70.8 -34.9 6.9 2.4 112.1 71.9 -40.2 13.3 3.5

4 93.8 63.5 -30.3 98.5 65.9 -32.6 4.7 2.4 100.7 66.6 -34.1 6.9 3.1

5 89.6 62.3 -27.3 93.4 63.2 -30.2 3.8 0.9 94.9 64 -30.9 5.3 1.7

a Initial viscosity: the immediate reading of viscosity after one minute preheating.

b MV: Mooney viscosity after one minute preheating and four minute testing time.

c =MV initial viscosity, IV1 = initial viscosity(one-week aged) initial viscosity(unaged), DIV2 = initial viscosity(two-week aged) initial viscosity(unaged),

MV1 = MV(one-week aged) MV(unaged), MV2 = MV(two-week aged) MV(unaged).

Figure 1: Energy consumptions of Batch 1 (. . .), Batch 2 (-.-.-), Batch 3 (—), Batch 4 (- - -), and Batch 5 (-.. -..)

during internal rubber mill processing

Figure 2: Temperature profiles of Batch 1 (. . .), Batch 2 (-.-.-), Batch 3 (—), Batch 4 (- - -), and Batch 5 (-.. -..)

during internal rubber mill processing

25

used since they filled up the cavity of the test chamber better than a single piece of specimen.

Evaluation of apparent shear stress and apparent shear viscosityApparent shear stress and apparent shear viscosity of the rubber composites were obtained with a capillary rheometer, Rheo-testing 1000 (Goettfert, Germany) in accordance with ASTM D5099. The diameter of the die was 1.5mm and the length/diameter ratio of the die was 15. The entrance angle was 90º. The instrument was warmed up for 30 minutes at 100°C. The cycle of apparent extrusion speed was from 1,000 s-1, 500 s-1, 230 s-1, 115 s-1, 57.5 s-1, 23.0 s-1, to 11.5 s-1. The apparent shear stresses and apparent shear viscosities at all apparent shear velocities were obtained after three minutes’ preheating and the total testing time was 30 minutes.

Determination of curing characteristicsThe vulcanizing characteristics of each rubber composite were determined with an oscillating disk cure meter, Rheo-tech (TechPro Inc, Cuyahoga Falls, Ohio) in accordance with ASTM D2084. The curing temperature for SBR and PB rubber was 160°C. The oscillating frequency was 1.7±0.1Hz with amplitude of ±3.0º. The volume of a test specimen was between 3-5cm3. Total testing time was 30 minutes. Scorch time (T

s-2)

and cure time (Tc-90) were measured.

Evaluation of tensile strength of unaged and heat-aged rubber compositesThe tensile stresses of rubber composites were measured with an Instron 5500 R, model 4201 (Instron, Canton, Massachusetts) in accordance with ASTM D624. The crosshead speed was 50cm/min and the original gauge of the extensometer was 25mm. The thickness and width of dumbbell specimens were about 2mm and 6.25mm, respectively. Tensile strength and the modulus at 100%, 200%, and 300% elongation were calculated from the following equations:

The heat aging of dumbbell specimens was carried out in accordance with ASTM D573. The specimens were placed in a 100°C oven for 70 hours. The hardness (type A) of test specimens was measured by a Durometer type A (Instrument & MFG Co Inc, New York) in accordance with ASTM D2240. The thickness of test specimens for hardness test was about 6mm.

Evaluation of tear strength of rubber compositesThe tear strength was measured in accordance with ASTM D412. The die cutter for preparing tear-test specimens was Die C. The thickness of tear-test specimens was around 2mm. The tear strength of hot rubber composites was obtained after tear-test specimens were boiled in hot water for one hour. The specimens were tested when they were still hot.

Evaluation of heat resistance and heat blowout timeHeat build-up and heat blowout tests were performed with a Firestone Flexometer (Firestone Tire & Rubber Co, Akron, Ohio) in accordance with ASTM D623. The specimen was placed between two plates. One of the plates had an oscillating speed of 13.3Hz. For the heat build-up test, the pressure was fixed at 0.8MPa and the test time was 45 minutes. The inside temperature of the rubber composites was recorded after 45 minutes and used as an indicator for the heat-generating rate of rubber composites.

For the heat blowout test, the pressure was fixed at 1.7MPa and the moving amplitude of the oscillating plate was 8.2mm. The time was recorded when the deformation reached 8.2mm. The dimensions of Firestone specimen were: base 54.0 x 28.6mm, top 50.8 x 25.4mm, and altitude 38.1mm. The cure time for Firestone specimens was four times that for tensile specimens.

Determination of dynamic mechanical propertiesDynamic mechanical properties were determined with a mechanical energy resolver (MER) MER-1100B (Imass, Accord, Massachusetts) in compression and tension on a cylindrical specimen. The frequency of the MER test was 1Hz with a constant dynamic force of 0.7kg-RMS (RMS: root mean square). The MER test was done at two different temperatures of 23°C and 100°C. The diameter and length of MER specimens were about 17.8mm and 25mm, respectively. The cure time for MER specimens was double that of tensile specimens.

Statistical analysis of experimental dataAll experimental data with replicates were analyzed with a standard two-sample t-test using S-PLUS statistical software (Version 8.0, Insightful Corp, Seattle, Washington). All comparisons were based on a 95% confidence interval.

Compositions of rubber composites and compounding proceduresTable 2 shows the compositions of rubber composites from Batch 1 to Batch 5 and the compounding procedures for all batches. Batch 1 was the control rubber composites that did not contain any MCC filler. The percentages of silica filler that were replaced by the MCC filler were as follows: 5.6% in Batch 2, 11.8% (almost double Batch 2) in Batch 3, 17.65 (almost triple Batch 2) in Batch 4, and 23.5% (close to quadruple Batch 2) in Batch 5. The weight percentages of other components in rubber composites were the same for all batches. The compounding procedures included two passes. In the first pass, all the materials except the curatives of the rubber composites were included. The total mixing time in the first pass was 500 seconds. A high shear and a long mixing time were used in the first pass to allow the even dispersion of all fillers and other additives in the rubber matrix. The rubber composites obtained from the first pass were defined as Master Batch 1 (MB1). The curatives (DPG, CBS and S) could not be added in the first pass because the internal temperature could be high enough to cause the pre-cure of rubber composites. As a result, a second pass was used where the curatives were added to the rubber composites. The rubber composites obtained from the second pass

“A high shear and a long mixing time were used in the first pass to allow the even dispersion of all fillers”

26

were defined as Master Batch 2 (MB2). For preventing the pre-cure of rubber composites, MB2 was discharged either at 120 seconds of the mixing time or at 100ºC, whichever came first.

Compounding of the rubber composites in an internal rubber millFigure 1 shows the power required for the thorough mixing of the rubber composites in an internal rubber mill. Because all the components in the first pass were added into the mill during the first 240 seconds, the power required for thorough mixing was determined after 240 seconds (Figure 1). When the amount of silica replaced by MCC increased, the power required for compounding the rubber composites decreased. These results appeared to indicate that it took less energy to evenly disperse MCC into the rubber matrix than to evenly disperse the same amount of silica. The densities of MCC, silica, and carbon black are around 1.4212, 2.2013,and 1.8114 g/cm3, respectively.

MCC is lighter, i.e. bulkier, than silica. MCC particles had less tendency of aggregation among themselves than silica particles and carbon black particles. A part of energy required for the thorough mixing was used to break down the silica or carbon black aggregates. During the compounding, some MCC particles might be inserted between silica particles, which would reduce the re-aggregation of silica particles and facilitate the dispersion of silica particles into the rubber matrix, thus reducing the energy required for evenly dispersing the fillers into rubber matrix.

Figure 2 shows the temperature profiles of five batches during the compounding in the first pass. The temperature in all five batches linearly decreased as the mixing proceeded. The temperature was an indicator of the heat generated during the compounding.

Friction among fillers generated heat. When the compounding proceeded, more fillers were dispersed into the rubber matrix and less friction among fillers would occur, which might explain why the temperature linearly decreased along the mixing time (Figure 2). As the compounding proceeded, some filler particles would collide with each other to form aggregates again. When the aggregates were broken down, friction occurred and heat was generated. The higher aggregation tendency of silica and carbon black particles might explain why the temperature in Batch 1 decreased at a slower rate than that in other batches. From Batch 2 to Batch 5, the amount of silica replaced by MCC increased, and the temperature profile was also down-shifted in an almost parallel fashion. The reduction of filler-filler interactions by MCC might account for this temperature down-shift. Heat dissipation could be one of the reasons why temperature dropped down during the process, because the temperature of the mixer was maintained at 65°C and was lower than that of the rubber mixture.

The previous study has shown that the reaction of silane coupling agent

with silica occurs at a temperature above 150°C.15 The mixing temperature for each batch in this study was above 150°C (Figure 2), which indicated that the coupling reaction between the silane coupling agent and silica would occur.

Mooney viscosity of the rubber compositesTable 3 shows the initial viscosities and Mooney viscosities of unaged and aged rubber composites from each batch. The test temperature was set at 100°C because most industry processes such as milling, extrusion, and calendaring occur around this temperature. The numbers in the first three columns were the results of unaged rubber composites from each batch.The initial viscosities of the unaged rubber composites decreased from Batch 1 to Batch 2, increased from Batch 2 to Batch 3, and then decreased again from Batch 3 to Batch 5. Low initial viscosity resulted from less filler-filler interaction and better dispersion of fillers in the rubber matrix.

The Mooney viscosity for unaged rubber composites decreased with increase in MCC content from Batch 1 to Batch 5. The viscosity difference ( )between the Mooney viscosity and the initial viscosity is indicative of how easily the rubber composite is processed to form different products. The smaller the difference, the easier the process.

The viscosity difference for unaged rubber composites slightly increased from Batch 1 to Batch 3. The viscosity differences for Batch 3 and Batch 4 were the same. The viscosity difference slightly decreased from Batch 4 to Batch 5. The range of the viscosity differences for all batches was small. There will be little problem processing the rubber composites from all batches.

Table 4: Pre-vulcanization characteristics at 125ºC

Batch t5 (min)a t10 (min)b t35 (min)c

1 57.1 –d –

2 54.6 – –

3 52.0 58.8 –

4 43.5 48.2 55.4

5 45.2 49.7 56.2

a t5: The time required for increasing the viscosity above the minimum viscosity by 5 units

b t10: The time required for increasing the viscosity above the minimum viscosity by 10 units

c t35: The time required for increasing the viscosity above the minimum viscosity by 35 units

d (–): Data were not available because the viscosity did not increase by 10 units or 35 units when the test was done at 60 min.

Figure 3: The apparent shear stresses in each

batch at two different apparent shear velocities

500 s-1 ( ) and 1,000 s-1 ( )

Figure 4: The apparent shear viscosities in each

batch at two different apparent shear velocities

of 500 s-1 ( ) and 1,000 s-1 ( ).

27

The next three columns show the initial viscosity, the Mooney viscosity, and the viscosity differences ( ) for those rubber composites that have been aged for one week (Table 3). The one-week aging significantly increased the initial viscosity and the Mooney viscosity for all batches (Table 3). Both the initial viscosity and the Mooney viscosity for one-week-aged rubber composites gradually decreased from Batch 1 to Batch 5. In other words, the initial viscosity and the Mooney viscosity both decreased along with increasing the MCC content in the rubber composites.

The viscosity difference between the Mooney viscosity and the initial viscosity slightly increased from Batch 1 to Batch 2 and then gradually decreased from Batch 2 to Batch 5, which implied that addition of MCC facilitated the process of aged rubber composites into different products. For those rubber composites that had been aged for two weeks, the initial viscosity, the Mooney viscosity, and the viscosity difference between the Mooney viscosity and the initial viscosity all decreased when the MCC content increased from Batch 1 to Batch 5 (Table 3). These results demonstrated that the replacement of silica by MCC facilitated the processes of aged rubber composites.

For the rubber composites from each batch, the initial viscosity and the Mooney viscosity both increased when the aging time was increased (Table 3). During storage, i.e. aging, fillers tend to migrate and aggregate, which results in the viscosity increase. After the rubber composites were aged for either one week or two weeks, the net gain of the initial viscosity and the net gain of the Mooney viscosity for each batch decreased along with the increase in the MCC content from Batch 1 to Batch 5. We speculate that the following reasons may contribute the decrease in the net gains of the viscosities: 1) MCC has lower tendency to aggregate than silica and carbon black. The net gains of the viscosities would be smaller for the rubber composites with a higher MCC content even if all fillers have the same migration rates; 2) MCC slows down the migration of the other two fillers, silica and carbon black; 3) MCC interferes with the aggregation of silica and carbon black.

Pre-vulcanization characteristicsTable 4 shows the t

5, t

10 and t

35 of

all rubber composites. The t5 gradually

decreased when the MCC content increased from Batch 1 to Batch 4, and then slightly increased from Batch 4 to Batch 5. The t

10 was longer than 60

minutes because the viscosity did not increase by 10 units when the test was done at 60 minutes. The t

10 decreased

from Batch 3 to Batch 4 and then remained roughly the same from Batch 4 to Batch 5. The t

35 for the Batches 1-3

was longer than 60 minutes because the viscosity did not increase by 35 units when the test was done at 60 minutes. The t

35 for Batch 4 and Batch 5 was not

significantly different.Generally speaking, the t

5, t

10

and t35

indicate the pre-vulcanization characteristics during processing. If the t5, t

10 and t

35 are too low, the rubber

composites are vulcanized too fast, which means that the viscosity increases too fast to allow the rubber composites to take a shape of the resulting products. Therefore the t

5, t

10 and t

35 have to be sufficiently

high. Batch 1 is a typical recipe for rubber tire application. The pre-vulcanization characteristics of a new rubber composite are considered desirable if its t

5, t

10 and t

35

are close to those of Batch 1. The results from Table 4 suggested that MCC decreased the t

5, t

10 and t

35 to some extent.

As shown in Table 3, MCC decreased the initial viscosity and the Mooney viscosity of unaged and aged rubber composites. In other words, the rubber composites containing MCC have a lower viscosity than Batch 1 before any process. When compared with Batch 1, the lower initial viscosity will offset the faster increase in the viscosity (i.e. the lower t

5, t

10 and t

35)

during a vulcanization process. Therefore the lower t

5, t

10 and t

35 of the rubber

composites containing MCC would not negatively affect their vulcanization processes.

Rheological properties of the rubber compositesThe rheological properties of the rubber composites were obtained with the capillary rheometer at two apparent shear velocities of 500 s-1 and 1,000 s-1 because milling, extrusion and calendaring of rubber composites are usually carried out at 500 s-1 and the injection molding at 1,000 s-1. The apparent shear stress of each batch at 1,000 s-1 was higher than that at 500 s-1, which is consistent with the fact that a high force had to be applied to the rubber composites to achieve a high flow rate (Figure 3). The apparent shear stress decreased from Batch 1 to Batch

Figure 5: The scorch time (Ts-2) and cure time

(Tc-90) of each batch at 160ºC. Ts-2 ( ): the time

required for increasing the torque above its minimum

by 2 units. Tc-90 ( ): the time required for reaching

90% of the maximum torque

Figure 6: The tensile strength of unaged ( ) and aged

rubber composites ( ). Each data point is the mean

of three replicates, and the error bar represents one

standard deviation

Figure 7: The modulus at 100% (£), 200% (£l ), and

300% (£// ) elongation of unaged rubber composites

and the modulus at 100% ( ), 200% (£\\ ), and 300%

(£) elongation of aged rubber composites. Each data

point was the mean of three replicates, and the error

bar represents one standard deviation

28

4, and then maintained almost the same from Batch 4 to Batch 5. These results revealed that MCC reduced the force required for achieving the same flow rate.

With the increase in the MCC replacement from Batch 1 to Batch 4, the apparent dynamic shear viscosity

decreased at both apparent shear velocities (Figure 4). These results indicated that the MCC facilitated the process of the rubber composites and reduced the energy consumption for the process. The apparent dynamic shear viscosity for Batch 4 was almost the same as that for Batch 5, which the MCC effect on the reduction of the energy reduction flattened out when the MCC replacement was 17.6% or higher (Figure 4).

Cure characteristics of the rubber compositesThe scorch time (T

s-2) for all five batches

was about six minutes. This implied that the addition of MCC in rubber composites did not have significant influence on the time of initiating curing reactions (Figure 5). T

c-90 is usually used as the

cure time, i.e. the time required for full cure of the rubber composites. The cure time decreased from Batch 1 to Batch 2, remained the same from Batch 2 to Batch 3, decreased again from Batch 3 to Batch 4, and then kept almost the same from Batch 4 to Batch 5 (Figure 5). As discussed previously, MCC might facilitate the dispersion of silica. A good dispersion of silica (i.e. low aggregation of silica) means that more silane coupling agent-coated silica surfaces are available for vulcanization, thereby reducing the cure time. It is still poorly understood that the cure time did not linearly decrease along with increasing the MCC content. Because the T

c-90 was lower than 14 minutes

for all batches, the cure time of making tensile specimens for all the batches was set to be 14 minutes.

Tensile properties of unaged and heat-aged rubber compositesThe tensile strength significantly increased from Batch 1 to Batch 2 (Figure 6). The tensile strengths of rubber composites from Batch 3, Batch 4, and Batch 5 were comparable to that of control rubber composites from Batch 1 (Figure 6). For those heat-aged rubber composites, the tensile strength gradually decreased from Batch 1 to Batch 3 and then remained the same from Batch 3 to Batch 5 (Figure 6).

There are two opposite effects that can affect the tensile strengths. First, a good dispersion of fillers can improve the interactions between fillers and rubber matrix, thus improving the tensile strength. 4, 16 Second, MCC has a lower Young’s modulus than silica.17 MCC-rubber composites are inherently not as

strong as silica-rubber composites, which means that silica cannot be completely replaced by MCC. Results from Figure 6 demonstrate that tensile strengths of the rubber composites containing MCC were stronger than or comparable to the rubber composites without MCC (Batch 1) when the MCC content was below 23.5%. It is still poorly understood that the treatment of the rubber composites at 100°C for 70 hours slightly reduced the tensile strength of the rubber composites with MCC.

Figure 7 shows the effects of MCC replacement in rubber composites on modulus at 100%, 200%, and 300% elongations. The modulus at 100% elongation of unaged rubber composites from Batch 2, Batch 3, Batch 4, and Batch 5 were higher than that of unaged control rubber composites from Batch 1. The modulus at 100% elongation of unaged rubber composites gradually increased as the amount of MCC replacement increased from Batch 2 to Batch 5. The modulus at 200% elongation of unaged rubber composites from Batch 2, Batch 3, Batch 4, and Batch 5 was higher than that of unaged control rubber composites from Batch 1. The modulus at 200% elongation of unaged rubber composites slightly increased as the percentage of MCC replacement was increased from 0% to 11.8% (i.e. from Batch 1 to Batch 3), and then remained statistically the same from 11.8% to 23.5% of MCC replacement, i.e. from Batch 3 to Batch 5 (Figure 7). The modulus at 300% elongation of unaged rubber composites from Batch 2, Batch 3, Batch 4, and Batch 5 was higher than that of unaged control rubber composites from Batch 1. The modulus at 300% elongation of unaged rubber composites slightly increased as the percentage of MCC replacement was increased from 0% to 11.8%, i.e. from Batch 1 to Batch 3, remained statistically the same from 11.8% to 17.6% of MCC replacement, i.e. from Batch 3 to Batch 4, and then decreased from 17.6% to 23.5% of MCC replacement, i.e. from Batch 4 to Batch 5. The modulus at 100%, 200%, and 300% elongation for all aged specimens increased when compared with the corresponding modulus at 100%, 200% and 300% elongation for all unaged ones.

The modulus of aged specimens at 100% elongation with MCC was larger than that without MCC. The modulus of aged specimens at 200% elongation for Batch 1 had no difference from those for Batch 2 and Batch 4. The modulus of aged specimens at 200% elongation for Batch 1

Figure 10. The tear strength of rubber composites ( )

and hot rubber composites ( ). Each data point is the

mean of three replicates, and the error bar represents

one standard deviation

Figure 9. Hardness (type A) of unaged ( ) and aged

( ) rubber composites. The data point is the median

values of five measurements

Figure 8: The elongation at break of unaged ( )

and aged rubber composites (£\\ ). Each data point

is the mean of three replicates, and the error bar

represents one standard deviation

29

was larger than those for Batch 3 and Batch 5. Moreover, the modulus at 300% elongation of aged specimens from Batch 2 was comparable to Batch 1.

However, the modulus at 300% elongation of aged specimens from Batch 3, Batch 4, and Batch 5 were lower than Batch 1. The following mechanisms may explain why the modulus changed after aging. After vulcanization, the crosslinks in rubber composites were mostly polysulfide bonds. Polysulfide bonds were more flexible, i.e. less stiff than monosulfide bonds. As heat aging proceeded, some polysulfide bonds were broken down to monosulfide bonds, which resulted in the increase in the modulus.4

Figure 8 shows the effects of MCC replacement in rubber composites on elongation at break. The elongation at break of unaged rubber composites from Batch 2 was not significantly different from that of unaged control rubber composites from Batch 1. The elongation at break of unaged rubber composites from Batch 3, Batch 4, and Batch 5 were slightly shorter than that of unaged control rubber composites from Batch 1 (Figure 8). The elongation at break of unaged rubber composites from Batch 3 was almost the same as that from Batch 4 and Batch 5. The elongation at break for each aged specimen was shorter than the elongation at break for each unaged one. Polysulfide bonds are more stretchable than monosulfide bonds. Because some polysulfide bonds were transformed into monosulfide bonds as the heat aging proceeded, the elongation at break of the heat-aged composites became shorter than that of the unaged ones.4

The hardness of aged rubber composites was higher than that of corresponding unaged composites from the same batch (Figure 9). The hardness

of unaged rubber composites from Batch 1 to Batch 3 was almost the same, and then increased from Batch 3 to Batch 5. The hardness of aged rubber composites from Batches 1-4 was about the same and was higher than that from Batch 5.

Tear strengths of the rubber composites and hot rubber compositesThe tear strength of rubber composites from Batch 2, Batch 3, Batch 4, and Batch 5 was lower than that from Batch 1, respectively (Figure 10). The tear strengths of rubber composites were about the same from Batch 2 and Batch 3, decreased from Batch 3 to Batch 4, and remained the same from Batch 4 to Batch 5. Rubber composites without MCC filler had a higher homogeneity than those containing MCC, thus MCC-filled rubber composites had low tear strength. The tear strength of hot rubber composites remained the same from Batch 1 to Batch 3. The tear strength of hot rubber composites decreased from Batch 3 to Batch 4 and then remained the same from Batch 4 to Batch 5 (Figure 10).

Heat build-up and heat blowout time of the rubber compositesWhen rubber composites underwent a long-term dynamic distortion, the heat built up inside the composites (i.e. the temperature increased). The inside temperature after 45 minutes’ dynamic distortion and the time after the rubber composites deformed to 8.2mm were recorded and are shown in Figure 11. The temperature significantly decreased when the MCC content was increased from Batch 1 to Batch 5 (Figure 11). However, the blowout time slowly

increased from Batch 1 to Batch 3 and then rapidly increased from Batch 3 to Batch 5. The heat capacity of cellulose (1.3 kJ*kg-1*K-1) is larger than that of silica (0.7 kJ*kg-1*K-1).18 MCC could serve as heat sink and significantly delayed the blowout of the rubber composites, which were very desirable features for tire application.

There was smoke coming out of blown-out test specimens, which suggested that MCC was degraded to some extent. This is consistent with the fact that MCC is less heat resistant than silica and carbon black. The previous study on MCC with thermogravity analysis (TGA) revealed that there was little weight loss when the temperature was below 200°C.19 The temperature in some pocket areas of the rubber composites might exceed the degradation temperature (200°C) during the Firestone heat blowout test, which resulted in the smoke. The degradation of MCC requires energy, thus dissipating heat. A combination of the high heat capacity and a partial degradation of MCC may account for the low inside temperature and long blowout time. How the partial degradation may affect the actual usable life of rubber tires warrants further investigation.

Dynamic mechanical properties of the rubber compositesThe tan was obtained from each rubber composite at a low temperature (23°C) and a high temperature (100°C). For test specimens from the same batch, the tanat the low temperature was higher than that at the high temperature (Figure 12). At the low temperature, the tan values of Batch 2 and Batch 3 were statistically the same as that of Batch 1. At the low temperature, the tan decreased from Batch 3 to Batch 4 and then remained the same from Batch 4 to Batch 5. At the high temperature, the tan decreased from Batch 1 to Batch 2, remained the same from Batch 2 to Batch 3 and then rapidly decreased from Batch 3 to Batch 5 (Figure 12). A high tan of rubber composite means low friction between rubber tires and road, i.e. a high rolling resistance and low gas mileage. The tan from the low temperature has implications on the rubber tire application in the rainy season, when the temperature is low and the road is wet. In the rainy season the high rolling resistance between tires and the road is desirable.

Figure 12: Viscoelastic property (tan ) of each

batch at 23ºC ( ) and 100ºC ( ). Each data point

is the mean of three replicates, and the error bar

represents one standard deviation

Figure 11: Inside temperature of rubber composites

from the heat buildup test ( ) and the blowout time

of rubber composites from heat blowout test ( )

30

Therefore the higher the tan obtained from the low temperature, the better. Rubber composites with up to 11.8% of MCC replacement had the tan , i.e. the traction, comparable to control rubber composites at low temperature. Further increase in the MCC content in rubber composites decreased the tan , i.e. the wet traction at low temperature. The tan from the high temperature has implications on the rubber tire application in summer: when the temperature is high, the road is less slippery than in the rainy season, and the fuel efficiency of automobiles becomes an important consideration. In summer the low tanof rubber tires, i.e. the low rolling resistance between rubber tires and road, is preferred for obtaining a high fuel efficiency and good mileage for automobiles. The tan decreased from Batch 1 to Batch 2, flattened out from Batch 2 to Batch 3, and then rapidly decreased from Batch 3 to Batch 5, which implies that the partial replacement of silica by MCC in rubber composites would decrease the rolling resistance and increase the fuel efficiency of automobiles if the rubber tires are made with the rubber composites containing MCC.

Summary and conclusionsThe partial replacement of silica with MCC significantly reduced the energy required for thorough mixing of rubber, fillers, and other additives. MCC significantly reduced the initial viscosity and the Mooney viscosity of unvulcanized rubber composites, thus facilitating the formation of rubber composites with different shapes prior to vulcanization. Moreover, the addition of MCC reduced the difference in viscosities between unaged and aged rubber composites, thus facilitating the handling and processing of the rubber composites.

The rubber composites with MCC replacement had a higher pre-vulcanization rate than those without MCC. The low initial viscosity and low Mooney viscosity of rubber composites containing MCC would offset the negative impacts of a higher pre-vulcanization rate

on the process. MCC decreased the shear stress and shear viscosities, which further demonstrated that MCC facilitated the processes such as milling, calendaring, extrusion, and injection molding of rubber composites. The tensile strengths and the modulus at 100%, 200%, 300% elongation of rubber composites containing up to 17.6% of MCC were higher than those without MCC.

MCC increased the heat resistance of rubber composites. The partial replacement of silica with MCC slightly decreased the tear strength of the rubber composites and hot rubber composites. The partial replacement of silica with MCC did not decrease the tan at the MCC replacement of up to 11.8% at low temperature such as in the rainy season, and decreased the tan at high temperature such as in summer. The automobile tires made from rubber composites containing up to 11.8% MCC would have the same traction on the road in the winter as those without MCC, and would give the automobiles a higher fuel efficiency in summer than those without MCC. tire

AcknowledgementsFinancial support for this research was provided by Schill+Seilacher Struktol AG, Hamburg, Germany and Struktol Company of America, Stow, Ohio. The authors greatly appreciated the experimental support from P. Danilowicz, Z. Brdarski, B. Eikelberry, and K. Tracy at Struktol Company of America.

References1) Bowyer, J. L., Shmulsky, R., Haygreen, J. G.

Composition and Structure of Wood Cells.In: Bowyer, J. L., Shmulsky, R., Haygreen, J. G., editors. Forest Products and Wood Science: An Introduction. 4th edn. Ames, Iowa: Iowa State Press (2003); p. 48-56

2) Ismail, H., Rosnah, N., Rozman, H. D. Curing characteristics and mechanical properties of short oil palm fiber reinforced rubber composites. Polymer (1997); 38(16):4059-4064

3) Joseph, S., Joseph, K., Thomas S. Green composites from natural rubber and oil palm fiber: physical and mechanical properties. Int J Polym Mater (2006); 55(11):925-945

4) Haghighat, M., Zadhoush, A., Khorasani, S. N. Physicomechanical properties of alpha-cellulose-filled styrene-butadiene rubber composites. J Appl Polym Sci (2005); 96(6):2203-2211

5) Derringer, G. C. Short fiber-elastomer composites. J Elastoplast 1971; 3(Oct.):230-248

6) Nunes, R. C. R., Visconte, L. L. Y. Natural fibers/elastomeric composites. Nat Polym Agrofibers Based Compos (2000); 135-157

7) Toschkov, T. S., Gospodinov, N. R., Vidimski, E. P., inventors; DSO “Pharmachim”, assignee. Method of producing microcrystalline cellulose. US patent 3954727 (1976); 19760504

8) Vigo, T. L., Kinzig, B. J. Composite Applications: The Role of Matrix, Fiber, And Interface. New York: VCH (1992)

9) Laka, M., Chernyavskaya, S., Maskavs, M. Cellulose-Containing Fillers for Polymer Composites. Mech Compos Mater (2003); 39(2):183-188

10) Reinsch, V. E., Kelley, S. S. Crystallization of poly(hydroxybutyrate-co-hydroxyvalerate) in wood fiber-reinforced composites. J Appl Polym Sci (1997); 64(9):1785-1796

11) Kubat, J., Klason, T. C. F., inventors; (Swed), assignee. Composites from cellulose or lignocellulosic materials and plastics. Application: WO (1983); 19821210

12) Sun, C. True density of microcrystalline cellulose. Journal of Pharmaceutical Sciences (2005); 94(10): 2132-2134

13) Perriot, A., Vandembroucq, D., Barthel, E., Martinez, V., Grosvalet, L., Martinet, C., et al. Raman Microspectroscopic Characterization of Amorphous Silica Plastic Behavior. Journal of the American Ceramic Society (2006); 89(2):596-601

14) Gillen, K. T., Celina, M., Clough, R. L. Density measurements as a condition monitoring approach for following the aging of nuclear power plant cable materials. Radiation Physics and Chemistry (1999); 56(4):429-447

15) Wagner, M. P. Reinforcing silicas and silicates. Rubber Chem Technol (1976); 49(3):703-774

16) Ishiaku, U. S., Chong, C. S., Ismail, H. Cure characteristics and vulcanizate properties of a natural rubber compound extended with convoluted rubber powder. Polym Test (2000); 19(5):507-521

17) Eichhorn, S. J., Young, R. J. The Young’s modulus of a microcrystalline cellulose. Cellulose (2001); 8(3):197-207

18) Ball, R., McIntosh, A. C., Brindley, J. The role of char-forming processes in the thermal decomposition of cellulose. Phys Chem Chem Phys (1999); 1(21):5035-5043

19) Ardizzone, S., Dioguardi, F. S., Mussini, T., Mussini, P. R., Rondinini, S., Vercelli, B., et al. Microcrystalline cellulose powders: structure, surface features and water sorption capability. Cellulose (1999); 6(1):57-69

Our drive is your performance.

Hägglunds Drives AB, SE-890 42 Mellansel, Sweden. Tel: +46 (0)660 870 00, E-mail Business Area Manager: uno.sundelin@se.hagglunds.com, www.hagglunds.com

Demanding applications?We’ve seen them all.

Let us know your demandss, we will meet them.

T he tire/wheel assembly (referred to here as ‘assembly’) is a critical component in providing good ride, a requirement for all road

vehicles, particularly automobiles and light trucks. The assembly effect is usually thought of in two ways. The first is in terms of transmission of vibration produced by the interaction of the tire with road surface irregularities – i.e. harshness. The second is in terms of tire structural irregularity generated, and energy generated and transmitted during operation on a smooth road – i.e. uniformity. Both types of energy share the dynamic properties of the tire structure, although they differ in source.

An assembly rolling on an absolutely smooth surface inputs rotationally keyed cyclical forces and moments to the spindle to which it is attached. There are three forces and three moments to consider (Figure 1).1 In a practical sense the variation of two forces (radial and fore-aft) and one moment (aligning moment) are responsible for most of the ride disturbances attributable to balance or non-uniformity. Each of the non-

uniformity inputs is a complex time-varying waveform keyed to the angular position of the assembly.

It is common engineering practice to use Fourier Analysis to decompose uniformity waveforms into a series of pure sine waves, either sine or cosine, whose frequency depends on the frequency of rotation of the assembly. Figure 22 is an example of a uniformity waveform and its decomposition into a series of sine waves. The decomposition simplifies analysis.

The basic sine wave, first harmonic, has a frequency equivalent to the frequency of rotation of the assembly. For example, if the assembly were rotating 10 times per second, the frequency of the first harmonic would be 10Hz (cycles per second) or 20radians per second. The second harmonic would have a frequency twice that of the first harmonic, 40 radians per second. The third would have a frequency three times that of the first harmonic, 60radians per second, and so on.

Each waveform varies with speed due to variation of the excitation and because tires, wheels, and assemblies,

like all physical objects, have characteristics called natural frequencies. Physical objects like to vibrate at these frequencies. Thus, if a harmonic reaches the natural frequency of a mode, the harmonic will grow very large (Figure 3).3

As a harmonic passes through the frequency range, the vehicle speeds up and slows down, and the system response is delayed with respect to the harmonic excitation as well as varying in amplitude. This delay is called ‘phase’ and is illustrated in Figure 4, where we see a shift between the start of excitation and the system’s response. Phase is usually expressed as an angle, as in:

Equation 1

Tuning forks are a good modal example. A tapped tuning fork emits a single note associated with its principle mode of vibration.

Complex objects such as a tire, a wheel, an assembly, a car, and humans

This review of force uniformity mainly focuses on disturbances at the wheel rotation rate, and also covers balance, which affects uniformity in well-balanced assemblies

by Marion Pottinger, M’Engineering, USA

Uniformity: a crucial attribute of tire/wheel assemblies

Figure 2: The waveform for each component is complexFigure 1: The uniformity axis system, specified in SAE J2047, is a 180° rotation of the ISO wheel axis system

Vector Summation1-2-3-4-5-6Harmonics

1stHarmonic

2ndHarmonic

3rdHarmonic

4thHarmonic

5thHarmonic

6thHarmonic

Road planeContact center

Wheel plane

Wheel center

Referenceradius

R

RF

RM

Projection of reference point on wheel plane

LMLF

LFM

FFF

Forward

32

have multiple natural frequencies, each of which has an associated mode shape. Figure 5 shows the most important modes for a radial tire with respect to ride vibration.

As noted in the equation above, the actual output from a rotating assembly into the vehicle spindle is composed of a first harmonic and higher harmonics. This is true for all the forces and moments shown in Figure 1, but most critically for the fore-aft and radial forces, plus, to a lesser extent, the aligning moment.5

Figure 6 illustrates which harmonics are associated with which types of smooth road ride problems.

All service adjustments to assemblies such as balancing and match mounting are based on adjusting first harmonic behavior. In a practical sense, the higher harmonic excitations are the province of the tire manufacturer and perhaps the wheel manufacturer, although this remains to be proved. The vehicle

structural response to all harmonics of each of the forces and moments is subject to modification by the vehicle manufacturer.

It is common to begin a discussion of smooth road-ride disturbances by providing a laundry list of potential sources of uniformity problems within the tires and/or the wheels. In a practical sense, from the standpoint of a company that wants to correct problems after the tires and wheel are manufactured, only a limited number of sources are relevant. A number of the sources that have been listed classically are no longer relevant because well-made tires manufactured with state-of-the-art equipment by top manufacturers either do not have these problems, or they have been reduced to a low level. Thus, problems such as improperly made splices, lumpy extrusions, poor belt-width control, and snaked belts will be largely ignored in what follows. Things that remain are

tire runout, rim runouts, rim-width variations, tire-bead area placement variations, and other bead-seating errors. These will be viewed in the context of radial tires.

Figure 7 is a wheel-plane schematic view of a radial tire mounted on a perfect wheel. The radial tire is basically a stiff belt and tread area hoop supported on a set of sidewall cords tensioned by inflation. It can be described as an odd relative of a bicycle wheel. The static balance heavy point of the tire alone is at the high point of the tire’s crown radial runout.5

Regardless of which radius is loaded against the road, the mounted radial tire’s radial stiffness is about the same. However, for a constant loaded radius the force exerted radially against the spindle varies over a revolution because the assembly’s inherent out-of-roundness causes the deflection of the carcass spring to vary. The following equation expresses

Figure 3: Waveforms are very speed dependent

Figure 5: The crucial radial tire modes for ride vibration

Figure 4: Example of phase delay, showing a shift between the start of excitation and the system’s response

Figure 6: Smooth road ride noise and vibration disturbances (reprinted with permission from STP 929 1986 ASTM3)

- Front suspension vertical acceleration on 30” wheel

• - BFG HisumX - Lab machine

M.G Pottinger, T.R. Wikand K.D. Marshall, BF Goodrich 1972

100

90

80

70

60

50

40

30

20

10

0

1.0

0.9

0.8

0.7

0.6

0.7

0.5

0.4

0.3

0.2

0.1

0

Radi

al fo

rce

(pea

k to

-pea

k- L

B)

Unsp

rung

mas

s ve

rtica

l acc

eler

atio

n (g

s)

10 20 30 40 50 60 70 80

Speed (mph)

Ampl

itude

3.0

2.0

1.0

0.0

-1.0

-2.0

-3.0

Angle (°)

0 90 180 270 360

Phaselag

Output or response

Radians

Input

12 3 12 2

Vertical Longitudinal

Rigid Body 18Hz Torsional 42Hz

Flexural 80Hz Flexural 88Hz

Frequency (Hz)

10l

100l

1,000l

Not part of this discussion

Higher harmonics

Tire inducedTire/pavement interaction noise

RoughnessShake

First harmonicRollingboom

Spring/mass damper Shell modes

Local modes, Airpumpingorgan pipes

Rigid body Structural modesAcoustic transmissability

(Rigid) suspension elements(local)

Chassisbeaming

Vehicleresponse

Tireresponse

Tireinduced

RoadinducedA

B

C

D

33

the radial force variation developed at the footprint.

Equation 2

Figure 8 illustrates the ‘well-known correlation between the first harmonic radial force variation, RFH1, and the first harmonic geometrical runout, H1. The slope of the linear fit represents the tangent stiffness of the tire-wheel assembly at the test load and inflation pressure. The obvious linearity of the data set is good experimental evidence that low-speed radial force non-uniformities are manifested as geometrical runout of tires, at least for the first harmonic.6

With the tire operating either against a test surface or moving over a road at a constant speed, the kinematics of rotation force the tread band to undergo angular acceleration and deceleration due to the presence of the fixed loaded radius. There is an associated cyclically varying fore-aft force that acts at the spindle

because the tire’s rolling radius is variable. This equation expresses the variable rolling radius.

Equation 3

Substituting the rolling radius expressed into this equation…

Equation 4

…and differentiating with respect to time, Walker and Reeves6 derivedan approximate expression for the angular acceleration of the belt/tread, which is:

Equation 5

Substitution of the angular acceleration into the equation describing motion about the spindle derived by drawing a free-body diagram of assembly leads finally to Equation 6, which explains the basic behavior of fore-aft or longitudinal force as a function of variation in the tire’s rolling radius.

Equation 6

Figure 9 shows the variation of radial force, fore-aft force, and lateral force with speed for tires tested by Walker and Reeves.5

Walker and Reeves were well aware of the interaction with the first longitudinal (torsional) mode shown as having a natural frequency of 42Hz for the tire design shown in Figure 5. They discussed this interaction primarily in terms of phase. Richards7 extended the work of Walker and Reeves to explicitly include the resonant effect using experiments on a larger tire with a first longitudinal (torsional) natural frequency of about 26Hz. Figures in Reeves’ work based on hundreds of tests demonstrate that the modal interaction conclusions are correct.

In a practical sense, Figures 10 and 11 approximate first harmonic force examples for a P225/60R16 tire, and demonstrate the effects. Note that the reason the radial force seems relatively speed independent in Figure 9 or 11 is that its excitation source is not speed dependent, and that the associated frequencies don’t get high enough to get near the first vertical flexural mode, which has a natural frequency of 80Hz in the example in Figure 5 and is about 60Hz for the P225/60R16 tire used to plot Figure 11. Looking back at Figure 3, there is an example of a higher harmonic of

Figure 7: Wheel plane schematic of mounted radial Figure 8: Correlation of first harmonics (runout & force) Figure 9: First harmonic variation vs speed for 40 tires

Figure 10: First harmonic force peak-to-peak fore-aft force for P225/60R16 tire assembly with 0.5mm amplitude

Spindle

Rim

Tire

Heavy point

Radi

al ru

nout

(mm

)

Radial force variation, H1 (daN)

0.2

0.1

0

-0.1

-0.2-4 -2 0 2 4Lateral

Longitudinal

Peak/PeakForce (LB)

Radial

100

10

1

100

10

1

100

10

110 50 100

Speed (km/h)

0 20 40 60 80 100 120 140

Neglecting 1st mode resonance

With 1st mode resonance

Peak

-to-

peak

fore

-aft

forc

e (N

)

100

80

60

40

20

0

34

Speed (mph)

radial force that did pass through the first vertical flexural mode. It had a sharp response to frequency that would seem to be a sharp-speed response to a driver or on a uniformity machine where data was reported as a function of speed.

The fore-aft force first harmonic, FFH1, lags RFH1 by 90° ( /2 radians) at low frequencies. This is expected from the fact that Equation 2 is a sine series and Equation 6 is a cosine series.

As the frequency of the first harmonics rises, the interaction with the first longitudinal mode causes FFH1 to lag further and further behind RFH1. Walker and Reeves5 observed this and Richards7

verified this. At this point it is worth making four

observations about radial force and fore-aft force. Although these observations are generalized, they apply very well to first harmonic disturbances, which are of interest in post-manufacturing tire correction as done prior to mounting the tires on vehicles.

First, the runout of the mounted tire’s crown with respect to the spindle is the major variable driving fore-aft and radial force first harmonic variations.

Second, radial force variations arise

due to flexing of the tire’s carcass spring due to crown runout. RF non-uniformity does not show a major change with speed, but does interact strongly with the first vertical flexural mode at frequencies above about one-half of the mode’s natural frequencies. This interaction is not relevant for RFH1 or RFH2, but can become important for higher harmonics at speeds driven in the USA. The interaction with the first and second harmonics would be considerable at autobahn speeds.

Third, fore-aft force variations show an increase with the square of driving speed. This is a major increase with speed. The FF variations arise from angular acceleration of the assembly about the spindle, due to the variation in rolling radius arising from the runout of the mounted tire’s crown with respect to the spindle. Due to the lower frequency of the first longitudinal (torsional) mode, all harmonics including FFH1 are affected by this mode at speeds driven in the USA. For FFH1 the interaction is not really relevant at speeds below 100km/h.

Next, fore-aft force variation lags the radial force variation by 90° ( /2 radians) until the angular frequency approaches

one-half of the torsional natural frequency, at which point the lag begins to increase as frequency increases.

Additionally, the literature makes an important point about the location of the heavy point and apparent location of the static unbalance.

Finally, the location of the heavy point of the tire alone is at the high point of the tire’s crown radial runout.

Correction of first harmonic non-uniformity forces in the post-manufacturing environment has taken two forms: match-mounting of the tire and wheel5,6,9,11 and post-cure, uniformity, grinding of tires.12 Weights have been considered as means of first harmonic non-uniformity force correction in a perturbation of balancing technology, but they do not work for this purpose,5 as discussed below.

Observing that out of roundness is the principle source of RFH1, Walker and Reeves5 explored the effect of match-mounting, which they called ‘selective fitting’ on high-speed uniformity for 1970s vintage radial tires. In match-mounting the wheel, first harmonic radial runout low spot is matched with the tire first harmonic radial force high spot,

Figure 11: First harmonic force peak-to-peak radial force for P225/60R16 tire assembly with 0.5mm amplitude Figure 12: Match-mounted assembly (selective fitting)

Figure 13: Assembly RFH1 randomly assembled improved and conventional wheels and tires

Figure 14: Assembly RFH1 for randomly assembled and match-mounted improved wheels and tires

Figure 15: Radial force first harmonic of 115 tires and wheels matched in accordance with Equation 7

Neglecting 1st mode resonance

With 1st mode resonance

Peak

-to-

peak

fore

-aft

forc

e (N

)100

80

60

40

20

0

Speed (km/h)

0 20 40 60 80 100 120 140

Wheel radial runout 1stharmonic low point

Tire radial force variation 1st harmonic high point

Cum

ulat

ive

perc

ent

Cum

ulat

ive

perc

ent

FRFV (LBS)

Conventional wheels and unground tires

Improved wheelsand ground tires

8 12 16 20 24 28

100

80

60

40

20

0

FRFV (LBS)8 12 16 20 24 28

100

80

60

40

20

0

Matched

Random

Peak

-to-

peak

– M

easu

red

(lbs)

50

40

30

20

10

0

200

150

100

50

0

Assembly rad. force1st harmonic

0 20 40 60 80 (N)

0 5 10 15 20

Peak-to-peak – Predicted (lb)

35

as shown in Figure 12. The result, on average, is a reduction in the overall first harmonic runout of the assembly. Walker and Reeves found that “the technique of selective fitting of tires to wheels to reduce low-speed radial force variation is, fortunately, fairly effective for limiting high-speed longitudinal, fore-aft force variations.” Note that they observed that things weren’t perfect, as represented by the words ‘fairly effective’.

A short time before Walker and Reeves published their work, Nedley and Gearig,9

who had only low-speed uniformity measurements and ride ratings for belted bias tires as data, published a report on the effect of improved component tires and wheels on ride. In the process they also covered match-mounting with respect to low-speed RFH1 behavior. The improved components were ‘Tru-centric’ wheels10 – well-piloted steel wheels manufactured using a process invented by Kelsey-Hayes, and uniformity ground tires, a concept introduced by General Tire.12 The improved components were reported to sharply reduce RFH1 for assemblies, even when randomly assembled (Figure 13). Match-mounting

of the improved components was even better (Figure 14).

Nedley and Gearig believed that you could take a set of RFH1 data for a group of tires measured on true wheels, and a set of H1 data from a group of wheels, multiply it by the tire spring rate, and then combine these data according to Equation 7 to predict the RFH1A, maximum amplitude data for the possible combinations of wheels and tires.

Equation 7

Thus, it is possible to quickly get the probable mounting results without experimentation, and using full assemblies. Many unfamiliar with the literature still believe this to be a true representation of match-mounting.

Walker and Reeves knew a bit later that the simple Nedley and Gearig model of match-mounting was not completely correct. In the early 1990s, Schuring11

showed explicitly that the simple assumption for the match-mounting

was in error (Figure 15). Thus, the conclusions by Nedley and Gearig need to be viewed with caution.

Furthermore, Schuring showed that the mounting uncertainty depended on the quality of the wheels used, but he was unable to provide an explanation for what was being observed. However, he did observe that the behavior of tires with respect to RFH1 was much more consistent on what he termed precision wheels when mounted multiple times on each wheel at different angular positions.

Rhyne, Gall, and Chang6 modeled the effect of rim precision on the location of the beads in terms of the wheel width variation and radius variation. The inclusion of the wheel width variation sharply improved the agreement between the predicted match-mounting behavior and the observed match-mounting behavior (Figure 16), although the results are still not perfect.

It is possible that had the effect of thickness variations in the bead region of the tire, due to factors such as splices, been included, then most of the rest of the deviation from the simple match-mounting relationship (Equation 7)

Figure 16: Vector difference between the measured

and predicted wheel contributions to the RFH1A

Figure 17: Effect of uniformity grinding (reprinted

with permission from SAE 700089 1970

Figure 18: Ground tire image as of 1969. Just grinding

on the shoulder ribs is no longer enough for tires

Figure 19: The relationship between longitudinal (fore-aft) and radial force variations of static unbalance

Measured Rim RH1 (daN) Radial and lateral model

+ radial only model

8 12 16 20 24 28

0 1 2 3 4

3

2.5

2

1.5

1

0.5

0

Mea

sure

d -

Pred

icte

d (d

aN)

Cum

ulat

ive

perc

ent

FRFV (LBS)

100

80

60

40

20

0 20 40 60 80

Ground

Unground

Radialforce

maximum+ve Long

forcemaximum

+ve

Maximum•

Co-axialcircle

High spot

Tire periphery

Tire out of balance

Fixed axle

Road wheel

1 23 4

Ground outer tread rows

“It is possible to quickly get the probable mounting results without experimentation,and using full assemblies”

36

would have been explained. Importantly, Rhyne and his associates showed that the response to wheel width variation is mean rim width dependent for a particular tire. Therefore, it is likely that every different tire specification yields a different response to wheel width variation. From a user’s viewpoint, the ideal wheel is one with constant bead radius and rim flange to rim flange width as a function of circumferential position. With this, match-mounting would work substantially well for RFH1 based on the simple match-mounting relationship.

The situation with respect to match-mounting can be summarized in three observations.

Match-mounting using the classical concept is a helpful procedure for reducing RFH1A provided that: the width between the wheel flanges is essentially constant over the wheel circumference; the wheel radial first harmonic runout multiplied by tire vertical spring rate and tire first radial harmonic are such that matching greatly reduces the assembly RH1; successful match-mounting substantially reduces FFH1A at highway speed; and the best results would be obtained if the tire first radial harmonic and wheel radial first harmonic runout multiplied by tire vertical spring rate were identical so that assembly RH1

could be made equal to zero by match-mounting.

It should not be forgotten that match-mounting results will still be somewhat imperfect even if all three observations are scrupulously adhered to, due to bead position lateral variation caused by thickness variations within the tire itself.

Earlier in this article, the process of uniformity grinding was briefly mentioned. This process, invented by Hofelt,12 involves selective alteration of the tire’s tread surface profile so as to reduce RFH1T. It was quite popular from around 1970 because of its effectiveness on tires of that era (Figure 17). But, its popularity has waned as tires have changed, so grinding on the shoulder ribs, as in Figure 18, is no longer enough. Indeed automotive OEMs often no longer accept ground tires for reasons of poor appearance and a tendency for the grinding to trigger uneven wear. In any case, other than some preliminary work by Caulfield and Higgins,13 uniformity grinding has been a tire factory, not a service, procedure.

Going back to Figure 7, if the assembly is rotating, the static unbalance assumes the positions indicated in Figure 19.

The magnitude of the static unbalance force vector is given in the following equation:

Equation 8

It is directed outward from the center of the spindle along a line through the heavy point; thus, its phase with respect to the assembly never changes.

The RFH1A, exclusive of unbalance, is a maximum at position 1 and a minimum at position 3 (the solid arrows). The FFH1A, exclusive of unbalance, is a maximum at position 2 and a minimum at position 4 (the dotted arrows). The unbalance force tends to make RFH1A smaller and FFH1A larger than they would be in its absence. For the balanced assembly (Figure 20), RFH1A is larger and FFH1A is smaller than in the unbalanced state because part of all of the static unbalance induced force has been canceled.

Noting this effect leads to the thought that maybe RFH1A and FFH1A can be canceled by a unique use of balance weights. Unfortunately, the phase for the weight required for correction is different for correcting fore-aft and radial forces. In the case of FFH1A, the phase of the uniformity component changes while that of the weight-induced balance component does not (Figure 21). Furthermore, the magnitude for RFH1A

Figure 20: Unbalance and balancing force vectors Figure 21: Effect of balance weight on longitudinal (fore-aft) force variation (reprinted with Tire Society permission)

Balanceweight

Unbalanceforce

Unbalance

Balancingforce

Phase lag

Locus of tire longitudinal vector

BALANCE WEIGHT EFFECT

ASSEMBLY REFERENCE

B

C

C

B

A

A

D

D

50 100

40

150

40

40

40

(LB)

100

100

100

“It is very likely that every different tire specification yields a different response to wheel width variation”

Assembly reference

Balance weight effect

C100

B100

A100

D100

D40

C40

B40

A40

Unbalanceforce

37

correction by use of a weight can only be correct at one speed. This all leads to an important observation first made by Walker and Reeves: “It is not possible to balance tire and wheel uniformity in a manner that satisfies radial and longitudinal fore-aft, directions over a wide speed range.”

In terms of pure balance correction, Ni14 showed that, besides the wheel’s own inherent imbalance and the effect of RH1W on imbalance by moving the wheel center of mass off the spindle center, wheel lateral runout also contributes to imbalance. By implication, if a hub has a face that is not perpendicular to the spindle it will create imbalance that is not accounted for in off-the-car balancing, because a good true running wheel will be forced to act in service as if it possesses lateral runout.

Neill and Kondo15 worked on balance right at the time that Tru-centric wheels appeared.9,10 They did not have the new ‘properly piloted’ wheels and concluded that on-the-car balancing was superior to off-the-car balancing. This points out the very major effect of having properly piloted wheels and truly by implication the value of properly centered hubs.

Recently, Ford discussed the possibility of doing away with the static balance specification on tires, and thereby reducing the number of tires rejected by their specifications.16 The analytical approach is rather unusual and also may not be valid.

Of more practical interest, Hunter Engineering’s Scribner17 pushed the concept that if the dynamic imbalance doesn’t exceed certain ride test determined levels, it is wasteful of materials and time, and is perhaps even counterproductive to do more than static balance. This appears to have validity.

There are three additional balance observations worth making: wheels must be well piloted for good balance results to be obtained in service; wheel-mounted lateral runout must be small to obtain good balance with a minimum amount of weight; and static imbalance is always important, but dynamic imbalance needs to reach a critical threshold before it is a notable problem.

Based on the observations made in this article, and the fact that service area ride treatment of tire-wheel assemblies is fundamentally confined to correcting RFH1 and FFH1 whether from a uniformity or balance source, there are four basic conclusions that can be made.

Reduction of assembly first harmonic radial runout is the most important thing that can be done to simultaneously reduce RFH1 and FFH1. Therefore, effective match-mounting is important.

For match-mounting to work effectively based on the standard match-mounting relationship, the wheel width between the rim flanges must be essentially constant within a few one-hundredths of a millimeter.

For good ride performance, particularly in balance terms, wheels must be well piloted and the wheel plus hub must suppress lateral runout.

Finally, static imbalance is predominantly important and dynamic imbalance can be ignored unless it is substantial.

In reality, neither tires nor wheels are bolted to hubs – assemblies are bolted to hubs. So the important thing is good assembly uniformity. Match-mounting offers help, but often the combination of wheel runout and tire uniformity in the components of an assembly cannot produce a nearly ideal situation in which the assembly’s first harmonic runout defined in force terms is effectively zero. But, there may be a way to resolve this

problem, first suggested by T. R. Wik in the early 1970s when he was involved in the tire industry. Suppose the wheel-center hole was larger than needed to pilot the assembly, and an element (ring) was superimposed between the hub and wheel, which was off-center just the right amount to make the radial runout of the assembly based on first harmonic radial force zero.

Figure 22 is a sketch of such an arrangement. Effectively perfect match-mounting could be induced. As an example, tap-in metal rings could be used to generate the effect or a quick set ring formed by an injectable polymer could be used. Obviously, there are many detail design considerations, from wheel bolting and insuring that the tire cannot move circumferentially with respect to the rim during operation to force measuring equipment to be considered. This idea and a good balance job could make first harmonic uniformity ride disturbances a thing of the past. tire© 2009 Tire Society, Inc. Used with permission

REFERENCES:1) Warrendale, P.A. Tire Performance Terminology,

J2047, SAE, 1998.2) Marshall, K. D. Tire Noise and Vibration,

Chapter 9, The Pneumatic Tire (edited by J. D. Walter and A. N. Gent), National Highway Traffic Safety Administration, Washington DC, 2005.

3) Pottinger, M. G. Marshall, K. D. Lawther, J. M. and Thrasher, D. B. A Review of Tire/Pavement Interaction Induced Noise and Vibration, TheTire Pavement Interface, ASTM STP 929, M.G. Pottinger and T.J. Yager, Eds., American Society for Testing Materials, Philadelphia, PA, 1986, pp. 183-287.

4) Marshall, K. D. Wik, T. R. Miller, R. F. and Iden, R. W. Tire Roughness – Which Tire Nonuniformities are Responsible, SAE 740066, Society of Automotive Engineers, Warrendale, PA, 1974.

5) Walker, J. C. and Reeves, N. H. Uniformity of Tires at Vehicle Operating Speeds, Tire Science and Technology, TSTCA, Vol. 2, No. 3, Aug.

Figure 22: Example assembly incorporating the

ability to deliberately offset the wheel center

Center of rotation

RingOffsetwheelcenter

Wheeldisc

Hub

The phased wheel offset from the center of rotation produces the ideal match mounting required

“Reduction of assembly first harmonic radial runout is the most important thing that can reduce RFH1 and FFH1”

38

1974, pp. 163-178.6) Rhyne, T. B. Gall, R. and Chang L. Y. Influence

of Rim Run-Out on the Non-uniformity of Tire-Wheel Assemblies, Tire Science and Technology, TSTCA, Vol. 22, No. 2, April-June. 1994, pp. 99-120.

7) Richards, T. L. The Relationship Between Angular Velocity Variations and Fore and Aft Non-uniformity Forces in Tires, SAE900761, Society of Automotive Engineers, Warrendale, PA, 1990.

8) Koutny, F. Analytical Comments on Radial Tire Nonuniformity, Tire Science and Technology, TSTCA, Vol. 24, No. 2, April-June, 1996, pp. 132-152.

9) Nedley, A. L. & Gearig, D. M. Radical Improvements in Tire and Wheel Manufacturing – Their Effects Upon Radial Force Variation of the Assembly, SAE 700089, Society of Automotive Engineers, Warrendale, PA, 1970.

10) MacIntyre, D. D. Advances in Wheel Uniformity, SAE 710087, Society of Automotive Engineers, Warrendale, PA, 1971.

11) Schuring, D. J. Uniformity of Tire-Wheel Assemblies, Tire Science and Technology, TSTCA, Vol. 19, No. 4, October – December, 1991, pp. 213-236.

12) Hofelt, C. Jr. Uniformity Control of Cured Tires, SAE 690076, Society of Automotive Engineers, Warrendale, PA, 1969.

13) Caulfield, R. J. and Higgins, R. J. On-Car Tire Grinder for Improved Tire Smoothness,

SAE720465, Society of Automotive Engineers, Warrendale, PA, 1972.

14) Ni, E. J. A Mathematical Model for Tire/Wheel Assembly Balance, Tire Science and Technology, TSTCA, Vol. 21, No. 4, October – December, 1993, pp. 220-231.

15) Neill, A. H. Jr. and Kondo, A. Correcting Vehicle Shake, Tire Science and Technology, TSTCA, Vol. 2, No. 3, August, 1974, pp. 179-194.

16) Tananko, D. Krivtsov, V. and Rohweder, D.Do We Really Need a Spec on Tire Static Balance? SAE 2003-01-0151, Society of Automotive Engineers, Warrendale, PA, 2003.

17) Scribner, D. New Dynamic Balancing Method Based Upon Absolute Force Reduction Algorithms, ITEC, Akron, Ohio, September 2006.

NOMENCLATURE:Symbol definitionAn Amplitude of the nth member

of a Fourier seriesF Fore-aft axisF(t) A force varying with timeFF Fore-aft, longitudinal, forceFFH1 Fore-aft 1st harmonic forceFM Moment about the fore-aft axis,

overturning momentH Wheel center height, loaded radiusH1 Wheel radial runout 1st harmonic

kR Global tire stiffness in the radial directionL Lateral Axis, coincident with spindle center

line hereinLF Lateral, forceLM Moment about the lateral axis, torquemUNBAL Static unbalance expressed as an equivalent

massn The index of the nth member of a Fourier

seriesrW Radius to the effective imbalance massR Radial axisR Mean rolling radiusRR Rolling radius

Rn Amplitude of the nth component or harmonic of radial runout

RF Radial, forceRFH1 Radial Force 1st harmonic forceRFH1A Assembly radial force 1st harmonic forceRFH1T Tire radial force 1st harmonic forceRFH1W Wheel radial force 1st harmonic forceRFH2 Radial force 2nd harmonic forceRM Moment about the radial axis, aligning

momentt TimeT Tire rotational periodUNBALFUnbalance forceV Forward velocity

Reference angle

n Phase angle of the nth member or harmonic

in a Fourier series assembly angular velocity assembly angular acceleration

EPLEXOR® DMTS/DMTAThe well-known Dynamic Mechanical Thermal Analyser to ascertain visco-elastic properties (Young’s Modulus, glass transition, frequency and temperature behaviour)

GABOMETER® Flexometer/Fatigue TesterThe newly designed Goodrich Flexometer for heat build-up and blow out tests

GABOTACK® Quality ControlThe tackiness tester for “Green Tires”, epoxy resins and much more

QUALIMETER® Quality ControlProvides online quality control during production processes (static and dynamic spring rate, hardness and visco-elastic properties)

The dynamic testing instruments of GABO QUALIMETER are used very successfully in the material development and for quality control since several decades.

Testing Instruments from Testing Experts

I N T E L L I G E N T M A T E R I A L T E S T I N G

tire technology Expo 2010Booth C4, FoyerKöln, Germany9 - 11 February 2010 www.tiretechnology-expo.com

GABO QUALIMETER Testanlagen GmbHSchulstr. 6 . 29693 Ahlden/Germany Phone: +49 5164 8019-0 Fax: +49 5164 8019-30e-mail: info@gabo.cominternet: www.gabo.com

39

40

For modern development of new or improved cars or car components, the dynamic simulation of the behavior of the system is an

essential advantage that saves time and money. Because the main force transfer between the car and the environment occurs via the tires, the description/modeling of the tire characteristics is of great importance for a useful and valid simulation of the vehicle dynamics.

On the one hand the modeling of tire behavior should not be too complicated and should be properly geared toward the goal of the investigation. On the other hand it should satisfactorily map the measurements of the tire characteristics. To obtain reliable simulation results as a substitute for tests and experiments, it is especially important that the tire behavior is correct.

In 2002 a group of experts started the Tire Model Performance Test (TMPT),1

for commercially-available passenger car tire models offering the possibility of getting objective information about their efficiency. Consequently, the TMPT also includes a comparison of different models.

Figure 1 indicates that the group of experts, establishing the test, first had to define the test specifications attuned

for tire handling properties, and for the high-frequency behavior of the tires. Special measurements were the basis for the parameterization of the tire models.

The tire characteristics of the chosen tire models were investigated by MBS-software suppliers using a special test rig (Figure 3). Results management and the general organization was done by the university, independently of all other participating groups. By creating this structure the intention was to guarantee an objective evaluation.

The majority of specifications for the test maneuvers to be performed applying the virtual test rig are summed

It is especially important that tire behavior is correct to obtain reliable simulation results as substitutes for tests and experimentsby P. Lugner & M. Plöchl, University of Technology, Vienna, Austria

Comparison of tire models and their application for vehicle system dynamics

Figure 2: Specifications for the tire performance test

Figure 1: Structure of the Tire Model Performance Test (TMPT)

Establishing Group: Experts of tire manufacturers,Tire model providersMBS software providers,Industry (application)University

MeasurementsContinental, Michelin

Tire models

TMPT specification and definition

MBS simulation (test rig)ADAMS, DADS, SIMPACK

Evaluation and comparison of results Inst. of Mechanics and Mechatranics Vienna University of Technology

Figure 3: Virtual test rig and possible input quantities for the simulation of the performance of the tire model.

Revolute joint between wheel guidance and sledge: steering angle δ

Wheel guidance device

Wheel carrier and coordinate system

Inclination of wheel to wheel carrier: camber

Revolute joint between wheel and wheel carrier: angular velocity of wheel

Vertical load to press wheel on ground surface

Translational joint between wheel guidance and wheel carrier: vertical motion

Translational joint for longitudial motion of the sledge: velocity

Handling.Parking: Steering at standstill

Cornering: steering angle δ(t)=kt, k=2º/s (steering angle δ is equal to the side-slip angle of the wheel for the test rig used

Turn around steering at v=10m/s until =180º or failure of simulation

Changing driving conditions (with =0.75 max): load step changes, -step changes

Longitudinal behavior ( =0): change of inflation pressure, -step change with braking

Combined steady-state characteristics Fy(Fx, ), M2(Fx, ), v=80km/h

High-frequency range.Dynamic brake cycle: braking torque applied corresponding to an ABS-system

Vertical excitation sweep: amplitude 1cm up to 40Hz, different running speeds, a=0 and a=0.75amax

Steering sweep: amplitude 2º to 40Hz: mean values m=0 and m=0.75 max

Running over a cleat: different speeds, =0 and a=0.75amax, 90º and oblique cleat

Sequence of potholes, step up and step down, with and without brake torque

•

•

•

•

•

•

•

•

•

•

•

41

up in Figure 2. The difference between ‘handling’ and ‘higher frequency range (hfr)’ should make an evaluation of the tire models clearer, having in mind the different complexity of the tire models.

The data for a great variety of working conditions to parameterize the tire models were measured by Michelin and Continental using a special test tire. Also its cross section and material properties were determined.

For the dynamic properties the first eigenmodes for radial and transversal vibrations were investigated. With a cleat fixed on a drum the transient tire characteristic were experimentally measured for different cleat heights and velocities. As a consequence an extensive amount of information was available for the tire model providers.

All seven tire models,1 that participated in the TMPT, were either already commercially available as software package or extensively applied for simulations in industry.

The F tire structural tire model2,3

uses the belt to rim suspension and a description of belt deformations to be able to take into account hfr problems and local road surface contours. Local contact elements describe the sticking and slipping phenomena, (Figure 4).

A further development of the well-known Magic Formula tire model,4

the Swift model, scans the road surface using cams that represent the tire contour (Figure 5). The belt-to-rim suspension, including the local contact element, extends the range of application into the hfr.

The CD tire model also utilizes the substructuring of the belt to be able to operate in the hfr (Figure 65,6). The effective tire-road surface contact is also described by local contact elements including elasticity and damping.

The TMeasy tire model is mainly aimed at handling problems and provides a mathematical approximation of measurements of the longitudinal and lateral tire characteristics. By a reasonable combination of this information any combined loading of the tire can be described (Figure 77).A basic description for the transient behavior is added.

Based on the Magic Formula,4 the PAC 2002 handling model calculates for pure and combined longitudinal and lateral slip valuing the corresponding

tire characteristics based on the approximation of measurements. Relaxation lengths in both directions are employed to describe the transient behavior.8

The semi-physical Uni tire model takes into account carcass flexibility in lateral and longitudinal directions, carcass twist stiffness, combined slip characteristics

and a non-isotropic friction concept.9,10

Although thereby non-steady-state tire behavior could be taken into account, the tire model is mainly aimed at studying handling maneuvers.

The Hankook-semi-physical-tire-model concept emphasizes handling behavior. It includes a transient behavior description based on visco-elastic properties.11,12

Belt flexibility stiffnesses

Radical force elements between each belt node and rim

Contact element tangential displacement model

Figure 4: Main structural features of the F tire structural model which describes sticking and slipping phenomena

wheel plane

wheel rimbelt

residualsprings

contact

oblique step

cams

effective road plane

C

W

_

sc

φ

_V

scc _

V

Mz

Fy

Figure 5: Structure of the Swift tire model, which scans the road surface using cams that represent the tire contour

Figure 6: Structures of tire model CD tire 30 (with one belt ring) and model CD tire 40 (with four belt rings) with

flexible connected structural elements. The CD tire model uses substructuring of the belt to operate in the hfr

EUROPE’S LEADING INTERNATIONAL TIRE DESIGN AND TIRE MANUFACTURING EXHIBITION AND CONFERENCEwww.tiretechnology-expo.com

9, 10 & 11 February 2010, Köln (Cologne) Messe, Germany

Design I Technology I Materials I Manufacturing

www.tiretechnology-expo.com

43

Camber, aligning torque and combined slip values are represented in considering the tire-patch structure.

The aim of the simulations was, on one hand, to show the range of application of the tire models by capability tests, and on the other, to check the quality of the results by validation tests in comparison with measurements, including measurements not available for the tire model provider.

The following results are marked with symbols: higher frequency models (hfr): A, B, E; handling models: C, D, F, G; MBS software: I, II, III; producing available combinations: IA, IB, IC; IIE; IIIA, IIIB, IIID, IIIF, IIIG.

Only a small selection of results are shown and for more information and for related names of tire models and software packages see Lugner & Plöchl, 2009.

One of the capability tests is turning the tire at standstill, typically when parking (Figure 9). Not all tire models were able to simulate this manoeuvre. The obvious differences between the models and softwares are relatively small.

Most essential for vehicle handling is the tire behavior during cornering, the lateral force and the aligning moment induced by the tire side slip angle (Figure 9). Besides the deviations for the less important moment, all tire models show a good performance, which was also validated by comparison with the measurements.

A typical example of transient tire properties can be considered for changing friction conditions at constant side-slip angle (Figure 10). Here the hfr-models, ABE encounter vibrations although the handling models show, more or less, rapid and relatively smooth transitions.

An example of the influence of the applied MBS-software I, III (caused by different integration methods, internal representation of the test rig and test, but same tire models) can be noticed by the longitudinal force behavior of a dynamic, step-wise brake moment increase as it may occur when ABS-braking (Figure 11). The same limit values with a locked-wheel (about) are achieved, but the time histories show significant differences.

An illustration of a validation test in Figure 12 shows the crossing of the hfr-tire models over a cleat. Despite obvious differences, the tire behavior is satisfactorily reproduced.

Figure 8: Turning at standstill: illustrations of the aligning moment Mz as a function of steering angle

Figure 7: Combination of lateral Fy and longitudinal F

x force measurements describing combined loading of tire

FxM

φ

Sy

400

300

200

100

0

-100

-200

-300

-400

F yG

FxM

Fx

dFx0

S xM

S xG

Sx

Fy

dF0

FMFG

Sy

SM

SG

Sx

S

S yG

SyM

dFy0

F yM F y

Fx

F (s)

mom

ent M

z [N

m]

-40 -30 -20 -10 0 10 20 30 40

delta [deg]-40 -30 -20 -10 0 10 20 30 40

delta [deg]

IIIAIIBIIF

IAIBIC

Figure 9: Lateral tire force Fy and aligning moment M

z as a function of the tire side-slip angle during cornering

6000

4000

2000

0

-2000

-4000

-6000

-20 -15 -10 -5 0 5 10 15 20

alpha [deg]

80

60

40

20

0

-20

-40

-60

-80

mom

ent M

z [N

m]

Forc

e Fy

[N]

IIDIIFIIIG

IIDIIFIIIG

IAIBICIIE

-20 -15 -10 -5 0 5 10 15 20

alpha [deg]

44

Considering the uncertainties of real road surface structures, such simulation results can be used for comfort investigations.

Through the TMPT investigation it becomes obvious that a careful handling and checking of the sequence tire measurements – tire modeling – dynamic simulation of tire characteristics is necessary.

It is necessary to be aware, that different tire models, even if already used previously, can show different results within the same range of application.

A testing of the tire model to be used with a virtual test rig, and a validation of the respective simulation results will finally guarantee effective vehicle dynamics simulation with statements and consequences that can speed-up, and reduce, costs in vehicle development. tire

References1) Lugner P., Plöchl, M. Tire Model Performance

Test (TMPT) Supplement to the International Journal of Vehicle System Dynamics (VSD), Volume 45 (2007)

2) Gipser, M. FTire software: advances in modelization and data supply. Proceedings of Tire Society Meeting, Akron (2006)

3) Gipser, M. FTire info & download. Available online at: URL: http://www.ftire.com

4) Pacejka H. B. Tyre and Vehicle Dynamics,Elsevier, Second Edition (2005)

5) LMS International LMS Comfort and Durability Tire User Manual (2005)

6) Gallrein, A., DeCuyper, J., Dehandschutter, W., Bäcker, M. Parameter identification for LMS CDTire, 3rd International tyre colloquium, tyre models for vehicle dynamics analysis proceedings. Vehicle System Dynamics, 43(Suppl.) (2005)

7) Hirschberg, W., Rill, G., Weinfurter, H. User-appropriate tire-modeling for vehicle dynamics in standard and limit situations,Vehicle System Dynamics, 38(2) (2002)

8) Adams/Tire Manual, MSC Software Corporation ( 2005)

9) Guo, K. H., Ren, L. A unified semi-empirical tire model with higher accuracy and less parameters, SAE Technical Paper Series, 1999-01-0785, pp. 37-44 (1999)

10) Guo, K.H., Lu, D., Ren, L. A unified non-steady non-linear tire model under complex wheel motion input including extreme operating conditions, JSAE Review, 22(4), pp. 396-402 (2001)

11) Gim, G., Choi, Y., Kim, S. A semi-physical tire model for vehicle dynamics analysis of handling and braking, Vehicle System Dynamics, 43(Suppl.) (2005)

12) Guo, K., Lu, D., Chen, S., Lin, W. C., Lu, X. The Unityre model: a nonlinear and non-steady-state tire model for vehicle dynamics simulation, Vehicle System Dynamics, 43(Suppl.) (2005)

Figure 10: Changing -conditions at a constant tire side-slip angle (nominal value n 0,9), =80km/h, 7,5°

Figure 11: Longitudinal force behavior of dynamic, stepwise brake moment increase under ABS braking, =20km/h

Figure 12: Validation test illustration. Behavior of hfr tire models crossing a 20x20mm cleat, =60km/h

9000

8000

7000

6000

5000

4000

3000

20000 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16

time [s]

9000

8000

7000

6000

5000

4000

3000

2000

IIIAIIIBIIE

IIIAIIIBIIE

measurement measurement

0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16time [s]

6000

4000

2000

0

-2000

-4000

-6000

6000

4000

2000

0

-2000

-4000

-6000

Forc

e Fx

[N]

Forc

e Fx

[N]

Forc

e Fz

[N]

Forc

e Fz

[N]

ICIIIF

Forc

e Fx

[N]

1000

0

-1000

-2000

-3000

-4000

-5000

1000

0

-1000

-2000

-3000

-4000

-5000

Forc

e Fx

[N]

IAIB

0.2 0.15 -0.1 -0.05 0 0.05longitudinal slip []

0.2 0.15 -0.1 -0.05 0 0.05longitudinal slip []

IIIGIIE

v=20km/h

IIIAIIIB

5 10 15 20 25 30distance [m]

5 10 15 20 25 30distance [m]

IIIAIIIBIIE

ICIIDIIIFIIG

IAIB1.2*

Forc

e Fy

[N]

2000

1000

0

-1000

-2000

-3000

-4000

-5000

-6000

Forc

e Fy

[N]

n

n

[}

0.3

19, 20, 21, February 2008Köln Messe, Köln, Germanywww.tiretechnology-expo.com

“I think it’s the most exciting show. We have all the suppliers for

the tire industry in one place. It’s always special to come here and see the latest developments. It’s

also a great location for us”Joerg Nohl, head of engineering,

Continental Tires

UKIP Media & Events Ltd l Abinger House l Church Street l Dorking l Surrey l RH4 1DF l UKTel: +44 (0)1306 743744 Fax: +44 (0)1306 742525 Email: colinscott@ukinpres.com

EUROPE’S LEADING INTERNATIONAL TIRE DESIGN AND TIRE MANUFACTURING EXHIBITION AND CONFERENCEwww.tiretechnology-expo.com

9, 10 & 11 February 2010, Köln (Cologne) Messe, Germany

Design I Technology I Materials I Manufacturing

www.tiretechnology-expo.com

“I think it’s the most exciting show. We have all the suppliers for

the tire industry in one place. It’s always special to come here and see the latest developments. It’s

also a great location for us”Joerg Nohl, head of engineering,

Continental Tires

EUROPE’S LEADING INTERNATIONAL TIRE DESIGN AND TIRE MANUFACTURING EXHIBITION AND CONFERENCEwww.tiretechnology-expo.com

Design I Technology I Materials I Manufacturing

9, 10 & 11 February 2010, Köln (Cologne) Messe, Germany

UKIP Media & Events Ltd l Abinger House l Church Street l Dorking l Surrey l RH4 1DF l UK Tel: +44 (0)1306 743744 Fax: +44 (0)1306 877411 Email: colinscott@ukintpress.com

Web: www.tiretechnology-expo.com

www.tiretechnology-expo.com

48

Airplane take-off speeds are designed to ensure the lift-off speed does not exceed the tire speed rating. But what factors can lead to a tire speed exceedance during take-off, and how can such events be prevented?

by Ingrid Wakefield, flight operations engineer & Chris Dubuque, service engineer, landing gear systems, Boeing, USA

Exceeding aircraft tire speed rating during take-off

A irplane tires are designed to withstand a wide range of operating conditions, including carrying very high loads and

operating at very high speeds. It is common for a jet airplane tire to carry loads as heavy as 60,000 lb while operating at ground speeds up to 235mph. To accommodate these operational conditions, each tire has specific load and speed ratings. Tires are carefully designed and tested to withstand operation up to, but not necessarily beyond, these ratings.

It is uncommon to exceed the load rating of tires during normal airline operation because the weight and center of gravity position of the airplane are well controlled and understood. However, on occasion the speed rating of tires can be inadvertently exceeded during take-off.

Boeing is receiving an increasing number of operator inquiries about tire speed limits being exceeded during take-off. This does not appear to be a new issue. Rather, advanced data acquisition tools on modern airplanes have made operators more aware of tire speed exceedance events.

In most cases, the speed exceedance is small, only a few knots. Boeing is not aware of any of these overspeed events resulting in thrown treads, which suggests that airplane

tires in good condition can withstand these small speed exceedances without damage. However, it is important to remember that at high speeds, heat is generated within the tire structure.

This heat, combined with extreme centrifugal forces from high rotational speeds, creates the potential for tread loss. Ensuring that tires are operated within their speed ratings will help prevent possible tread losses and the potential for airplane damage.

When dispatching an airplane in compliance with the certified Airplane Flight Manual, the airplane take-off speeds are designed to ensure that the lift-off speed does not exceed the tire speed rating. Although rotation and lift-off speeds are generally expressed in knots indicated airspeed, the tire speed limit is the ground speed, which is usually expressed in statute miles per hour. This means that a tire rated at 235mph is designed for a maximum ground speed at lift-off of 204kts.

A number of factors can lead to a tire speed-limit exceedance during take-off. Typically, this occurs when an airplane is dispatched at or near the tire speed-limit weight and: the airplane rotation rate is slower than the Boeing-recommended rotation rate; and/or there is a late rotation;

and/or the tailwind is higher than anticipated.

Dispatch at or near the airplane’s tire speed limit is most likely to occur during take-offs from airports at high altitudes on warm days, because these conditions tend to drive the ground speed at lift-off of the airplane closer to the tire speed limit. However, tire speed limits can be encountered during take-off in less severe environmental conditions, such as when scheduling an improved climb take-off.

Crosswinds can aggravate the situation by unexpectedly shifting into a tailwind, which may further increase the ground speed at lift-off. An unexpected (and therefore unaccounted for) tailwind component will directly add to the ground speed at lift-off.

Take-off procedures Boeing publishes a recommended all-engine normal take-off procedure in the Flight Crew Training Manual (FCTM) for 727, 737 Classic, and Next-Generation 737, 747, 757, 767, and 777 models, and in the Flight Crew Operations Manual for 717, MD, and DC models. In order to avoid tire speed-limit exceedance during take-off, Boeing stresses adhering to the recommended average all-engine take-off rotation rate of 2-3° per second, which provides adequate tail clearance margins with a target lift-off attitude reached after approximately 3-4 seconds (Figure 1).

Tail clearance margins for all 7-series models except the 717 are also outlined in the FCTM. Tail clearance and tail strike concerns are often the reason flight crews give for opting to use a slower rotation rate than recommended by Boeing.

When dispatching at or near the tire speed-limit weight, which is most likely to occur at hot temperatures and high elevations, a slower rotation than the Boeing-recommended 2-3°-per-second average may increase the actual

49

50

51

groundspeed at lift-off beyond the certified tire speed limit. In addition, a slow rotation or under-rotation could significantly increase the runway distance required to reach the 35ft point, which is another important reason for adhering to the Boeing-recommended rotation procedure.

Wind accountabilityThe certified tire speed-limit weight does not contain any margin for wind accountability. For instance, the FAA-certified take-off field-length-limit weight typically contains a conservative factor for wind accountability of 1.5 times the tailwind and 0.5 times the headwind. In comparison, the tire speed-limit weight lacks any such conservative wind factor. Because of this, an unexpected tailwind component not accounted for in the take-off analysis, occurring during a take-off at or near the tire speed-limit weight, may increase the true ground speed at lift-off beyond the tire speed rating.

To avoid a tire speed-limit exceedance, Boeing recommends to conservatively account for the tailwind component when dispatching at or near the tire speed-limit weight in a crosswind situation. General guidelines for crosswind take-offs are outlined in the FCTM. These guidelines include the recommendation to use a higher thrust setting than the minimum required in order to minimize airplane exposure to gusty conditions during rotation, lift-off, and initial climb.

747-400 case studyA case study of the 747-400 helps illustrate this point. The operator sporadically exceeded the tire speed limit even though the take-off analyses showed a notable buffer between the tire speed-limit weight and the actual dispatch weight. The airline approached Boeing for assistance.

The study was performed at two different dispatch weights: 805,000 lb and 825,000 lb. There was a 40,000 lb and a 30,000 lb margin between scheduled dispatch weight and the tire speed-limit weight. These weight margins, which appear relatively large, only resulted in speed margins of 8kts and 5kts between the associated ground speeds at lift-off and the tire speed rating (Figure 2).

This case study shows the relationship of a tire speed-weight margin to the associated speed margin for a four-engine airplane. Under similar dispatch conditions on a two-engine airplane, a similarly large weight margin can be expected to result in an even lower speed margin, due to the higher all-engine acceleration.

The same case study showed that a rotation rate that is 1° per second slower than normal can result in a lift-off speed increase of 4-5kts. This is in addition to the increase in all-engine take-off distance associated with the slow take-off rotation (Figure 3).

This illustrates how a slower-than-normal rotation rate can easily use up what may seem like a large tire speed-limit margin, especially if paired with a higher tailwind component than accounted for in the take-off analysis used for dispatch.

Maintenance actionsAlthough tire speed-limit exceedance events during take-off are not a new phenomenon, widespread recognition of these overspeed events is relatively new because of advances in flight data-recorder technology that enables easier data acquisition. Airplane manufacturers, tire suppliers, and regulators have not yet developed an industry-accepted set of maintenance instructions following a tire speed-limit exceedance event during take-off.

One maintenance suggestion would be that all wheel/tire assemblies be removed from the airplane before further flight after such an event occurs. In practice, however, replacing all of the wheel/tire assemblies on an airplane represents a major logistical problem and likely results in flight cancellations and/or dispatch delays.

It would be difficult to locate and ship 18 wheel/tire assemblies to a 747 at a remote location following one of these events! Additionally, if the overspeed was very small (say, 2-3kts over the tires’ speed limit), it is unlikely that the tires would have suffered any damage.

Some operators have elected to simply examine the tires after an overspeed take-off event using the normal tire inspection criteria in Chapter 32 of the Airplane Maintenance Manual. If no damage is found, the airplanes are dispatched normally and no further maintenance actions are performed.

Based on many years of service experience, this approach seems to have worked well because very few, if any, tire tread losses have been attributed to an overspeed event.

Based on this service experience, Boeing has typically not objected to this practice even though there is no overspeed take-off capability specifically designed into the tire.

If an operator has any questions about the integrity of the tires, the wheel/tire

assemblies should be replaced before further flight.

Additional information on tire maintenance procedures can be found in the airplane maintenance manuals and in the following documents:• FAA Advisory Circular 20-97B, Aircraft Tire Maintenance and Operational Practices,US Department of Transportation (April 18, 2005)• Aircraft Tire Care And Maintenance,Goodyear Aviation, 10/04, www.goodyearaviation.com/img/pdf/aircraftmanual.pdf.• Bridgestone Aircraft Tires, Tire Care, and Maintenance, http://ap.bridgestone.co.jp/pdf/Care_and_Maintenance.pdf.• Bridgestone Aircraft Tires, Examination, and Recommended Action, http://ap.bridgestone.co.jp/candm/recommendedaction.html.• Aircraft Tire Care & Service Manual,Michelin, www.airmichelin.com/pdfs/Care_and_Service_manual.pdf.• General practices manual for aircraft tyres and tubes, Dunlop Aircraft Tyres Limited, 01/08, www.dunlopaircrafttyres.com/tech_support/dm1172/DM1172.pdf.

SummaryAlthough it is uncommon to exceed the load rating of tires during normal airline operation, Boeing is receiving an increasing number of operator inquiries about tire speed limits being exceeded during take-off.

There is no industry consensus on the maintenance actions that should be taken following tire speed-limit exceedance during take-off. At this time, operators, in conjunction with their regulatory agency, must determine the most appropriate maintenance action based on the tire speed-limit exceedance event.

The best approach is to try to avoid overspeed take-offs altogether. By taking the following steps, flight operations personnel can reduce the possibility of tire speed-limit exceedance during take-off: Follow the Boeing-recommended rotation procedure; when dispatching at or near the tire speed-limit weight in a crosswind situation, consider conservatively, accounting for the tailwind component; when dispatching at or near the tire speed limit in gusty wind and strong crosswind conditions, use a higher thrust setting than the minimum required. tire

For more information, contact Boeing Flight Operations Engineering at flightops.engineering@boeing.com.

This article originally appeared in AERO magazine

and is reprinted courtesy of The Boeing Company.

52

Friction has a crucial role in driving. All forces acting on a vehicle are put into action via the friction forces between the tire and the

road, except for aerodynamic forces and gravity. A human driver can detect road conditions, but most vehicle dynamics control systems do not, until the moment the vehicle begins to slide.

Unfortunately, drivers often fail to estimate friction correctly, and attempt unrealistic maneuvers. Road accident statistics show that up to 40% of accidents occur in bad road conditions, and overestimation of the friction coefficient is often among the major causes.

To reduce the number and the severity of road accidents, there is a strong demand to integrate more safety ‘skills’ into vehicles. This lead to the widespread use of ABS and ESC on small vehicle platforms and of more complex ADAS on premium, and even some compact, cars.

However, along with driver behavior and alertness, friction is one of the remaining key unknowns in the control strategies of many active safety systems. The EU-funded project Fricti@n (FP6-2004-IST-4) has focused on developing,

demonstrating and verifying a system that provides continuous sensing of friction for vehicle applications.

The goal is to offer new information for the vehicle systems to enable them to operate more accurately, especially in unusual road conditions. The main attention has been on the friction conditions of paved roads constructed with surfaces such as asphalt or concrete. The surface conditions researched included: dry, wet, snowy, and icy.

In general, most vehicle applications will benefit from more accurate and continuous friction sensing. Cooperative driving applications could provide drivers with more complete information about road conditions, and well-established systems such as ABS and ESP could benefit from an optimal tuning for every level of friction, reducing the compromise between high and low friction. ADAS

such as collision mitigation, collision avoidance, and curve speed warning could adapt thresholds and intervention according to the actual friction, preserving a constant safety margin in every condition to reduce false activation. For example, if a collision mitigation system assumes high friction, it will have poor performance on snow, since it will brake too late. However, if the system is tuned to be effective on low friction, a lot of unnecessary intervention will occur on a dry road, causing disappointment to the driver, who will switch the system off!

Sensor information is central to the estimation of friction, but for cost reasons the project emphasis was on making use of existing or planned sensors in a new way, rather than developing new sensors. Three main sensor clusters are used in the system. Firstly, the chassis sensors were employed including the basic sensors of ABS/ESC, steering-wheel angle, steering torque, wheel velocities, body accelerations, yaw rate, and pitch and roll. Secondly, the environmental sensors were used including air and road temperature, xy ground velocity, laser scanner, radar, camera, and laser spectroscopy. Thirdly, a tire sensor was included, currently a specific tire-deformation sensor from the EU Apollo project (IST-2001-34372).

The Fricti@n estimation system was conceived with a modular architecture, taking into account two factors. The first factor was fault tolerance: if one sensor or sensor cluster is damaged or missing, the whole Fricti@n system does not cease to function, but its estimation accuracy decreases.

Estimating tire-road friction for chassis-control and driver-assistance systems

The EU-funded Fricti@n project showed great potential road-safety benefits, including better driver information and improved ADAS performanceby Marco Pesce, vehicle systems, vehicle dynamics and fuel economy, Centro Ricerche Fiat SCpA, Italy

eSafety systems

Driver assistance systems (DAS)

Cooperative safety systems

Navigation Maneuvering Stabilization Pre-crash

IVIS ADAS Active safety Advanced protective safety

Passive safety

Rescueservices

e.g. navigation systems

e.g. adaptive cruise control (ACC)

e.g. electronic stability program (ESP)

e.g. seatbelt pretensioner

e.g. airbag, structuralmeasures

e.g. eCall

t= ~10 sec t= ~1 sec t= ~0.1 sec t= 0 sec t= ~ -0.1 sec

IVIS = In-vehicle information systemsADAS = Advanced driver assistance systems

Fricti@n cooperative applications

Fricti@n onboard applications

Figure 2: The context of intervention of automotive eSafety systems and the impact of friction information

Figure 1: Members of the Fricti@n consortium

53

The second factor was scalability: low-cost systems may include only a subset of sensors such as ESC sensors only, again with reduced capability in comparison to the full system.

Each sensor cluster provides friction estimation tagged with quality/validity information, and a fusion module provides the final friction estimation and probability. A ‘learning’ feature is implemented too, comparing environmental data and vehicle dynamics data, and updating an internal database used for estimation.

The main output of the Fricti@n system are two friction values: friction used and friction potential. Friction used shows how much friction a driving maneuver requires. Friction potential is the maximum friction that the tires can achieve on the particular road pavement. The degree of certainty with which the system ‘believes’ given friction values is also provided. This enables other systems to judge how much they want to utilize the information.

The Fricti@n system can also provide an estimate of wheel slip and road and weather conditions in terms of dry, wet, snowy, or icy. If tire sensors are present,

then information on tire forces and the detection of the early stages of hydroplaning can be offered additionally.

The research was carried out through simulation models and driving tests in winter and summer conditions using three vehicles: a Volvo FH12 truck, a Fiat Stilo, and an Audi A6. Both the Volvo FH12 and the Fiat Stilo are equipped with ADAS developed in concurrent EU projects such as Safespot.

The full Fricti@n system demonstrated a near-continuous estimation of friction potential in changing road conditions, using sensor fusion and learning features. Careful checking was undertaken to assess data validity, and changes in conditions and driving provided a reasonably valid estimate in most cases.

Even a basic system, including ESC sensors only, provides good friction information, if a certain amount of friction is used. In this system a double approach is used: the first method is based on lateral dynamics (yaw rate), and the second relies mainly on the steering-torque information.

The first approach is quite robust at medium and high driving dynamics, when saturation of the tire lateral force

characteristic begins; the second is more effective at lower slip angles, since it may detect the variation of self-aligning torque. A local fusion module based on the level of driving dynamics provides the fused estimation, tagged with a validity flag. This basic configuration of the Fricti@n system proved effective in the detection of friction variation on a wet or icy surface, although it is not able to guarantee an almost continuous estimation as would the full system.

The effectiveness of the system and its benefits for ADAS performance was proved by simulation and experimental tests on demonstrator vehicles. The project studied the benefits of friction estimation using collision mitigation system prototypes developed in the EU project Safespot, on the truck Volvo FH12 and on the Fiat Stilo. The distances of driver warning and brake activation were compared with, and without, friction information. Experimental tests confirmed the reduction of crash energy in different driving situations and road conditions.

Scenarios with a standing obstacle or moving obstacle in front of the car were considered at different speeds and on different surfaces with friction coefficients ranging from 0.2 (low friction) to 1 (high friction). The crash energy index (CEI) shows the reduction of crash energy achieved with the friction-enhanced system compared to the original system.

In conclusion, the Fricti@n project showed fairly positive results and interesting perspectives on friction estimation. It also showed potential benefits for road safety, starting from driver information to improvement of ADAS performance. Further R&D activity is needed to increase accuracy and reliability in everyday driving conditions, to optimize cost benefits and system integration, and to achieve a full integration of Fricti@n information in all vehicle control systems and ADAS. tire

current slip critical slip

Friction available

tire slip

Current operating point of the tire

Friction potential

=Fy/Fz

Friction used

slip angle

FyMz

Mz

Fy

Crash energy index – Standing obstacle

100

90

80

70

60

50

40

30

20

10

0

CEI [

%]

0.50 0.20Road friction

30km/h

40km/h

50km/h

Figure 3: Explanation of friction potential, the friction used, and the friction available Figure 4: Basic principle of the double approach used with a chassis sensor cluster

Figure 5: The crash energy index shows the reduction of crash energy achieved with the friction-enhanced system

54

Finite element analysis (FEA) of pneumatic tires plays a vital role during the design cycle of these complicated structures. The

accuracy and reliability of the results obtained by this approximate method is very much dependent on the constitutive models selected for the description of the mechanical behavior of the tire components. The unique molecular structure of the polymers (natural and synthetic rubbers) plus organic fibers used in a tire makes their mechanical behavior viscoelastic.

This means that they not only show elastic (or hyperelastic) behavior, but they tend to dissipate energy upon the release of the applied load. Time is an additional variable that should be considered when the strain is to be related to stress. It is generally found that about 90% of the total energy loss in a tire is due to internal hysteresis of polymer molecules. The others go through friction between tire, road, and air resistance (rolling resistance). To reduce the difference

between the predicted properties of the virtual tire obtained through mathematical modeling and the real one, and also to predict the energy dissipation or rolling resistance, viscoelastic models must be used in the FE model.

Many researchers have tried to develop robust methods to more accurately model tires under different loading conditions.1

However, most of the developed models and reported results are based on the assumption of either an elastic- or hyperelastic-mechanical model for the rubber and the reinforcing parts in the tires. This is mainly because development of robust mathematical models is difficult, obtaining the converged and stable results is not always guaranteed, more computational resources are required, and determination of accurate and repeatable material parameters is not straightforward.

In this numerical study, an FEA for a steel-belted tire is presented. This model was developed based on the numerical capabilities implemented in the ABAQUS code.2 The analysis was performed under

steady-state rolling condition. The difference between the results obtained by the inclusion of the viscoelasticity and the no-viscoelastic model using different values of the model parameters and tire velocities were numerically investigated.

A viscoelastic constitutive model not only relates the stress to strain, but also the variation of the stress (and strain) with time is also included in the mathematical formula that describes the mechanical behavior of the material. The current form of the viscoelastic model implemented in the ABAQUS is based on the Prony series given as (Equation 1):

Where gR(t) is the dimensionless shear relaxation modulus [g(t)/g(0)], gi

P is the material constant and i

G is the relaxation time, respectively. This equation gives the variation of the shear modulus with time. Figure 1 shows the variation of the gR(t)with time for two different values of the relaxation times (gi

P =0.3, =0.1, 0.025 second). As can be seen, reducing the relaxation time leads to a sharper decrease of mechanical property and also the lower value at the fully relaxed state.

Figure 2 shows the same variation for

“A viscoelastic constitutive model not only relates the stress to strain, but also the variation of the stress with time”

Finite element analysis shows that the material behavior during the loading/unloading cycle plays an important part in contact-zone behavior

by Mir Hamid Reza Ghoreishy, Iran Polymer and Petrochemical Institute, Iran

Viscoelastic analysis of rolling tires using the finite element method

Figure 1: Variation of shear relaxation modulus

with time (t= 0.1, 0.025s).

Figure 2: Variation of shear relaxation modulus

with time (gi= 0.3 , 0.5)

Figure 3: Variation of tan with frequency

(t= 0.1, 0.025s)

55

two values of the material parameter giP

(giP =0.5, 0.3, =0.1). In this case,

increasing the giP also means a lower

value at the fully relaxed state. In both cases, it is assumed that a 1-term Prony series model was selected (N=1).

Equation 1 can be easily extended to take the variation of the material parameters of hyperelastic models. Therefore, for a material-model parameter such as Cij in a hyperelastic model (such as Mooney-Rivlin), its variation with time is expressed by Equation 2:

The Prony series parameters can be related to the well-known dynamic properties of the rubbers G’, G” and tan via the following relations:

Where is the frequency of the cyclic loading on the material. Figures 3 and 4

show the variation of the tan with frequency for the two previously mentioned sets of viscoelastic parameters. As it is easier to measure the dynamic properties using a dynamic mechanical analyzer (DMA), these relations (3 and 4) can be used to calculate the Prony series parameters for a given rubber compound.

A 175/70 R14 steel-belted radial tire was selected in this work, which has already been analyzed in previous works3-5 under different loading conditions, without taking the viscoelasticity into account. The FE model consisted of 10,402 nodes and 12,321 elements, respectively, as shown in Figure 5.

The mechanical behavior of the rubber components was assumed to be described by the well-known Mooney-Rivlin hyperelastic model in conjunction with a Prony series model (see equation 2) to include the effect of the material history (viscoelasticity) into the model. The internal (inflation) pressure and the vertical load were assumed to be equal to 0.248MPa and 4,900N (500kg), respectively. Three linear velocities, namely, 6, 60, and 120km/h were selected. Two computational strategies were adopted in this work. In the first method, which is called the traditional approach, a constant linear velocity with various values of the rotational speed ( )is used while in the second technique, a fixed value for the rotational speed is chosen and the linear velocity was

changed within a pre-specified range to complete the analysis. Figures 6 and 7 show the flow diagrams of the traditional and the authors’ selected methods, respectively. The other specification of the model was exactly the same as reported in references 3-5.

The ABAQUS/Standard was used to carry out the FE calculation under rolling condition using the three mentioned linear velocities and two sets of the Prony parameters given in Figures 1 and 2. The analysis was first performed without taking the viscoelasticity into account using the traditional approach shown in Figure 6. The computed value of the rotational speed was then fixed and the corresponding analyses using the viscoelasticity were carried out using the new algorithm given in Figure 7. The linear velocity in this case was selected to be 60km/h. Figure 8 shows the variation of the longitudinal force versus linear velocity for three different cases (no viscoelasticity, viscoelastic analyses with relaxation time = 0.1, 0.25).

As can be seen, the Force versus Velocity curves associated to viscoelastic analyses do not pass through zero point, and show a positive value for the computed longitudinal force. These forces correspond to the rolling resistance of the tire and can therefore be used to assess the energy-dissipation characteristics of the tire, rolling in steady-state condition. Figures 9 and 10 show the distribution of contact pressure for no viscoelastic and

Figure 4: Variation of tan with frequency (gi= 0.3, 0.5) Figure 5: Finite element mesh of the tire Figure 6: Old steady-state rolling analysis approach

Figure 7: New steady-state rolling analysis approach

Figure 8: Variation of longitudinal force vs ground velocity for no viscoelastic and viscoelastic models.

The right hand curve is a zoomed view of the left curves around zero force

56

viscoelastic analyses, respectively. The relaxation time and linear velocity were taken to be 0.1 second and 60km/h, respectively. It was discovered that the viscoelastic analysis predicts higher values for the contact pressure than the analysis in which the effect of material history is not taken into consideration.

During the no viscoelastic analysis, the simulation was performed based on the fully relaxed material properties that are the lowest value for the hyperelastic parameters (Figures 1 and 2). Therefore, the computed values of the contact pressure in this case (fully relaxed or no viscoelastic) is lower than when the viscoelasticity is included in the model. In other words, in the viscoelastic analysis the materials that pass through the contact region do not have enough time to be completely relaxed, and so the period of a tire rotational cycle, does not permit the rubber molecules to completely find their final state. The contact pressure distribution for the viscoelastic analysis with identical relaxation time (0.1 second) and lower selected linear velocities (6km/h) is shown in Figure 11. Having compared the results in this state with those shown in Figure 10, it can be seen that, by reducing the linear velocity, the required time for the materials to be relaxed on contact release will be

increased. This means that increasing the velocity is equal to reducing the time and therefore computed contact pressure is higher for the linear velocity of 60km/h compared to when it is equal to 6km/h.

In another analysis, the linear velocity was kept constant (6km/h) and the relaxation time was reduced from 0.1 second to 0.025 second. The contact pressure corresponding to the latter case is shown in Figure 12. The reduction in the relaxation time (0.1 second to 0.025 second) causes the relaxation rate to increase and thus the behavior of the model will be closer to the no-viscoelastic case. Consequently, reducing the relaxation time makes the predicted values of the contact pressure lower than in the first case, in which the relaxation time is four times greater than the current value.

It is also worth noting here that the predicted contact pressure for viscoelastic models is asymmetric about the two-axis (axis perpendicular to the traveling direction of the tire). This is due to the time-dependent behavior of the material that passes through the footprint area. The mechanical properties of a material point that enters the contact zone changes continuously with time, and when it leaves the footprint zone, its material parameter is different to those at the entry

to the contact region. Consequently, the rear and front zones of the tire undergo different contact pressure.

In addition to contact pressure, to show the effect of viscoelasticity on the wearability of the tire, a wear index has been defined, which is based on the Archard’s model for the description of material wear. The Archard’s model is given by:

Where q is the rate of material (rubber) loss, k is a parameter, H is the material hardness, p is the contact pressure and is the slip rate. The wear index has been defined by the product of the magnitude of the slip rate vector and magnitude of the tangential force at the contact zone, as in the equation:

Figure 13 shows the wear index for four different analyses. In the first analysis that corresponds to no-viscoelastic analysis (upper-left contour), the value of the wear index has been predicted to be lower than

Figure 9: Distribution contact pressure for no viscoelastic model (V=60 km/h) Figure 10: Distribution contact pressure for viscoelastic model (V=60 km/h, t=0.1s)

Figure 11: Distribution contact pressure for viscoelastic model (V=6 km/h, t=0.1s) Figure 12: Distribution contact pressure for viscoelastic model (V=6 km/h, t=0.025s

57

the viscoelastic analysis (upper-right contour) with the same linear velocity (V=60km/h).

This is due the lower value of the contact pressure that has been obtained for this simulation (see Figure 9). On the other hand, by increasing the linear velocity, the slip rate and contact pressure also increase, which accordingly cause the wear index to be increased. Consequently, it is expected that the viscoelastic model predicts that by increasing the linear velocity, the tire will have a higher wear rate.

Using FEA, a series of steady-state rolling analysis was performed using the viscoelastic constitutive equations to take the effect of material history on the rolling behavior of a steel-belted radial tire at the contact zone. This was found via numerical study, which makes it possible to take the material history behavior during the loading/unloading cycle into account. The numerical simulation results also showed that the tire behavior, especially in the contact zone, is dependent on the history effect. The method in its current form, however, cannot cope with the relatively higher values of the Prony series parameters.

Therefore, more robust numerical techniques than those implemented in the current version of ABAQUS should be developed.

Although we have used this approach for the parametric study of a limited number of variables and conditions (contact area), the technique can be used to investigate other structural parameters in tire bulk construction. tire

X-Pander: Uniform cord distribution at the edges and throughout the entire web width

Erhardt+Leimer your reliable partner for innovative technology in the Tire and Rubber industry

Complexor: Precise complexing of secondary to master strip

Elsmart:Slatted Roll Guide

for gentle alignmentof rubber strips and fabric liners

Erhardt+Leimer GmbHPostfach 10 15 4086136 Augsburg, GermanyPhone ++49 (0)821 24 35-531Fax ++49 (0)821 24 35-320info@erhardt-leimer.comwww.erhardt-leimer.com

Quality first, no compromise on tires

Figure 13: Distribution of the wear index (Eq. 7) at the contact zone for no viscoelastic and viscoelastic analyses

“It is expected that the viscoelastic model predicts that by increasing linear velocity, the tire will have a higher wear rate”

References1) Ghoreishy, M. H. R. A state of the art review

of the finite element modelling of rolling tyres, Iranian Polymer Journal, 2008, 17, (8), 571-597

2) ABAQUS, v 6.8, Documentation, 20083) Ghoreishy, M. H. R. Finite element analysis

of the steel-belted radial tyre with tread pattern under contact load, Iranian Polymer Journal (English Edition) 2006, 15, (8), 667-674

4) Ghoreishy, M. H. R. Finite element analysis of steady rolling tyre with slip angle: Effect of belt angle, Plastics, Rubber and Composites 2006, 35, (2), 83-90

5) Ghoreishy, M. H. R. Steady state rolling analysis of a radial tyre: Comparison with experimental results, Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering 2006, 220, (6), 713-721

FREESUBSCRIPTION/READER ENQUIRY SERVICE

Request your free copy of Tire Technology Internationalmagazine and learn more about innovative, industry-leading suppliers @ www.tiretechnologyinternational.com

OCTOBER 2009INNOVATION INITIATIVE

Jean-Claude Kihn, Goodyear’s CTO, has

undertaken a historic innovation strategyTIRE LABELING

How will the forthcoming legislation benefit customers?

TOMORROW’S WORLD

Michelin’s view of future standards

and regulations, and how to meet them

Lunar impactHow the moon will help generate

possible next-generation tires

INTERVIEWSDr Joachim Neubauer

Standards and regulation, Michelin

Nina RenshawPolicy officer, T&EDr Stuart Cook

Director of research, TARRC

Professor Joshua Summers

Clemson UniversityBen MichellDunlop Motorsport

59

L ow rolling resistance and low tire/road noise are requirements imposed on modern tires for environmental and economic

considerations. Tire/road noise, in most driving conditions, is the largest contributor to overall vehicle noise. Social and other surveys have indicated that traffic noise is probably the type of environmental pollution that affects more people than any other type of pollution. Rolling resistance also directly influences the fuel consumption of the vehicle, which leads to a similar influence on CO

2 and other exhaust emissions.

Although it is often estimated that a rolling resistance change of 10% leads to a fuel consumption change of 2-3% (the so-called ‘return factor’), the truth is that there is no single and universal conversion from the tire rolling resistance coefficient to overall fuel consumption of any vehicle.

The specific influence depends on several factors. Low tire rolling resistance seems to be especially important for medium-speed roads with relatively free-flowing traffic (like in suburban areas or urban highways). When traffic is often interrupted and slow, most of the energy is lost on frequent acceleration and braking. On the other hand, on high-speed highways, the aerodynamic drag is the dominant factor responsible for energy losses. However, at constant low or moderate speed, tire rolling resistance plays a very important role.

Tire/road noise, tire rolling resistance and other important tire parameters like friction, handling or wear resistance depend on many tire and road properties, in addition to driving conditions. This leads to the inevitable question as to whether there is a compromise in other areas when tires are designed for lower rolling resistance. Is there a conflict between the requirements of low rolling resistance and low tire/road noise?

One way to investigate possible conflicts between low noise and low rolling resistance is to make parametric studies changing one tire or road parameter at a time and evaluating tire road noise and tire rolling resistance variations due to those changes. This method gives very interesting results. However, it is very expensive and

difficult to perform for an independent research center without tire manufacturing capabilities.

The Technical University of Gdansk (TUG) sometimes uses this method if cooperation with a tire manufacturer is possible to provide an appropriate set of tires. A few years ago a set of five tires that differed in reinforcement were tested both for tire/road noise and for rolling resistance. The results are presented in Figure 1, where the rolling resistance coefficient (CR) is plotted against the A-weighted sound level.

Rolling resistance was tested on a safety-walk surface, and noise was tested on a replica of dense asphalt concrete. The conclusion was that tire reinforcement can be optimized for rolling resistance without an increase in exterior noise. For interior noise, the situation may differ, since ‘stiff’ tires would transmit more noise and vibration to the suspension and structure.

Another approach is to make statistical tests comparing results of tire/road noise and rolling resistance for many different tires. If comparisons show that there are

Noise and rolling resistance:is there a conflict?Is there any conflict between low noise and low rolling resistance of tires?

by Jerzy Ejsmont and Grzegorz Ronowski, Technical University of Gdansk, Poland

Rolling resistance coefficient [-]

0.0132 0.0134 0.0136 0.0138 0.0140 0.0142

Replica of asphalt concrete Safety walk

A-W

eigh

ted

SPL

[dB]

106

104

102

100

98

96

94

92

90

Figure 1: Sound levels for five 175/70R13 tires with different carcasses plotted against rolling resistance coefficient

Figure 2: Drum with ASP-4 (left) and safety walk (right)

60

tires that have low noise levels and low rolling resistance coefficients, this would indicate that there is no conflict between those two parameters. Of course in contrast to the first method, the statistical approach is not able to show directly what makes the tire silent and energy saving (the same applies to road surfaces). However, if such tires are selected, it would be possible to closely investigate their construction and identify the parameters responsible for such results.

TUG tests numerous tires every year, both in relation to their noise and in relation to rolling resistance, so statistical evaluations are possible. Rolling resistance is tested according to different methods, including methods described in ISO and SAE standards, both on the drum facility (Figure 2) and using a specially built trailer R2 (Figure 3). The drum facility shown in Figure 2 is also used for close-proximity noise measurements together with a smaller drum that is equipped with a replica of ISO 10844 reference road surface (Figure 4).

Data presented in this paper was obtained for typical passenger car tires, intended for summer and winter use. Tires were loaded to 4120N and inflation pressure was adjusted according to their load index. This paper presents the results for speed of 80km/h that was judged to be the most representative for the majority of traffic conditions, where rolling resistance is of high importance.

Figure 5 presents the comparison between rolling resistance coefficients measured on a very smooth surface, safety walk, and tire/road noise measured on a rather smooth surface – a replica of the ISO 10844 reference surface. A similar comparison obtained from measurements performed on an extremely rough surface APS-4 (both rolling resistance and noise) is presented in Figure 6.

Both Figures 5 and 6 clearly indicate that the correlation between rolling resistance and tire/road noise is very weak, to the point of being practically non-existent. Indeed, it may be stated that it is no more probable that a given

tire has low noise and low rolling resistance than that it has low rolling resistance but is noisy.

Experiments performed in different conditions (different speeds, loads and inflations) show a similar lack of significant correlation between rolling resistance and the noise of passenger car tires, both for conventional smooth and rough road surfaces.

Unfortunately the number of rolling resistance road measurements performed so far is still not enough to formulate conclusions that would cover all common road surfaces and experimental designs, such as, for instance, poroelastic road pavements. Initial observations, however, indicate that the conflict between low noise and rolling resistance is unlikely to be present for them either.

In general terms, for modern passenger car tires, within existing tire and road pavement technologies tested so far, there does not appear to be any conflict between low noise requirements and low rolling resistance. tire

A-W

eigh

ted

SPL

[dB]

Rolling resistance coefficient [-]

0.006 0.007 0.008 0.009 0.010 0.011 0.012 0.013 0.014 0.015

105

100

95

90

V=80 km/hRR tested on safety walkSPL tested on the replica of ISO surface

R2 = 0.0805

A-W

eigh

ted

SPL

[dB]

110

105

100

95

R2 = 0.1242V=80 km/hRR tested on APS-4SPL tested on APS-4

0.010 0.011 0.012 0.013 0.014 0.015 0.016 0.017 0.018 0.019Rolling resistance coefficient [-]

Figure 3: Trailer for rolling resistance testing “R2” manufactured and used by TUG Figure 4: Drum for testing tire/road noise on replica of ISO 10844 reference surface

Figure 5: Comparison of rolling resistance coefficients tested on Safety Walk and close proximity tire/road noise tested on replica of smooth road surface

Figure 6: Comparison of rolling resistance coefficients tested and close proximity tire/road noise tested on a very rough surface approximating surface dressing

MACHINERY FOR THE MANUFACTURE OF TIRESAdditionalBrochures:

3400 MASSILLON RD., AKRON, OH 44312 U.S.A.TEL: +1 330-896-2387 FAX: +1 330-896-3282

E-MAIL: sales@rjscorp.com WEBSITE: www.rjscorp.com

MANUFACTURED BY

Ask For Bulletin QT-997

Ask For Bulletins RJS-CM476-477-06and RJS-476-07

Ask For Bulletin RP-284

Ask For Bulletin TC-207

Ask For Bulletin RJS-493-107

Ask For Bulletin SP-448-09

Be prepared for the next regulation!2012 - Labelling of rolling resistance, wet grip and abrasion required

Suitable Process Oilsfrom H&R Group

TDAE, MES, RAE, NAPRubber & Tire Industry

H&R Group US, Inc. • H&R Groupwww.hur.com

H&R Group US, Inc. • H&R Group • 7900 E. Union Avenue • Suite 1010 • Denver, CO 80237 USAsales.america@hur.com • Phone +1 (720) 235 4900 • Fax +1 (720) 275 4902

62

F ull vehicle simulations cover a wide range of excitations, which may be generated by the driver, control devices, or road

disturbances.The wavelength and magnitude of road

disturbances can be used to characterize different conditions. Handling simulations, for instance, are often done on local undisturbed ground surfaces such as measured race tracks and are based on GPS data. As a consequence, the very low excitation frequencies mean tire models based on steady-state conditions can be used. In the case of ride, discrete obstacles or road profiles from local surface measurement are taken into account – small wavelengths with small magnitudes will be found. Ride is limited in the frequency area by approximately 50Hz, but the road profiles require tire models that are able to compute the contact area’s shape and pressure distribution, and modeling of tire structure flexibility is needed.

Load cases involving vehicle suspension strength and durability uncover special road profiles with small wavelengths and larger magnitudes, as well as discrete obstacles, among them misuse obstacles. The impact-like character of a tire running into a pothole excites higher frequencies, and the dynamics of the tire structure have to be taken into account.

Some questions arise when related prototype tests are planned. One of them is the determination of the critical velocity. At low vehicle speed, the small amount of kinetic energy cannot cause larger contact

forces but at high velocity, the spring preload is not able to accelerate the wheel into the pothole during the obstacle crossing time. Therefore, between smaller and larger velocity, there must be a condition of maximum suspension load. The range of velocities between 30km/h and 60km/h is divided into 10 steps and two tire-rim combinations, two inflations, and two obstacles. The number of runs is 80. This limits the complexity in terms of computational effort for the vehicle and tire model. A multibody dynamics system (mbs) is often used for vehicle modeling.

For tire modeling, the load path of a typical passenger car tire is based on the inflation pressure, which pre-loads the structure. The load is carried by belt and sidewall as tensile elements and not by spokes. In normal rolling conditions, contact occurs between the tire and road surface. To handle misuse deformations, the contact between the inner surface of the tire and the rim has to be added to the model. A second load path is established

if inner contact between rim and tire or two tire surfaces occurs, as in the runflat and in the misuse cases.

Looking at structure dynamics involving contact and friction, finite element models (FEMs) are usually the first choice. Because of the large computational effort, the number of runs to determine the above-mentioned critical velocity is limited. Some simplifications may help to overcome the limitations, especially if the rim forces are of greater interest over the stresses and strains inside the tire.

A starting point in modeling a tire with FEM could be a structure representation using rebar elements for belt and carcass together with different linear or nonlinear rubber matrix material models. This leads to a FEM model close to the physical nature of the tire but containing a large number of nodes and elements. Classical reduction techniques such as condensation are able to reduce the number of nodes and therefore the number of degrees of freedom. An additional way is to replace rebar elements by elements with anisotropy. However, the resulting model is still too expensive in terms of computational effort in mbs full vehicle dynamics, and with respect to the required number of runs.

Going a step further, the load carrying structure is assembled in quad 4 elements, where the rebars are condensed analytically to the four nodes, avoiding any numerical integration. A typical mesh of the belt is shown in Figure 1. A curved contact layer at the outer tire surface is connected to the belt grid by an elastic foundation. The contact layer has a different node density that normally will be of higher density than that of the belt. The contact or gap nodes are loaded by the footprint forces in the contact area, if the gap to the road surface is closed, from

Misuse load cases and MBS simulationSimulating full vehicle misuse behavior enables OEMs in an early phase of the design process to obtain fundamental information on misuse-relevant force levels and the damage chain with minimal extra parameters or computational effort

by Prof Dr-Ing. Ch. Oertel, FH Brandenburg (Brandenburg University of Applied Sciences), Germany

“The wavelength and magnitude of road disturbances can be used to characterize different conditions”

Figure 1: A typical FEM mesh model of a tire belt

63

which the contact area’s shape and the normal stress distribution is known. The contact nodes transfer the load to the belt and carcass. Some additional discrete elements are used to represent the bending stiffness of the rubber matrix.

Rotating the cross section generates belt mesh, sidewall mesh and rim nodes. Rim nodes move together with the mbs part rim of the vehicle model. This simplified flexible structure dynamics model is able to follow the load path idea, based on inflation pressure and preloads in the structure elements. Together with contact algorithm and friction model, the tire model RMOD-K FB is built. The computational effort in this model is a linear function of the number of nodes because of the explicit time integration method. This ensures that the peak loads are not influenced by numerical damping.

The misuse load case requires additional model elements, which are able to generate contact forces between rim and inner tire surface. The basic idea before starting the development of the RMOD-K FB extensions was to add only a few new parameters and to keep the computational effort as small as possible. Looking at radial displacement, the rim is represented by a nonlinear stiffness in every rim node. At the structure nodes of the belt grid, corresponding nonlinear stiffness elements exists.

The combination of both represents the inner contact mechanism, shown in Figure 2. The upper nonlinear stiffness represents the rim part of the inner contact and the lower ones are symbols of the inner surface. If the gap between rim and belt nodes closes, unilateral stiffness elements begin to work, expressed by the gap indicators. In consequence, only one additional nonlinear equation per rim node must be solved in every integration

step, and only if contact occurs in that node, indicated by at least one gap indicator greater than zero.

Five new parameters describe the different materials. The rim related parameter is an elastic-plastic approach with constant slopes cE and cP. The border between elastic and plastic behavior is given by the force FE. Due to the elastic-plastic behavior of the rim, an additional dissipation mechanism is included in the model. The rubber related stiffness relation is a simple nonlinear function with two parameters. If results from material tests are available, any other non-linear function can be implemented. In combination, only five new values are added to the tire model data set.

The model extension based on unilateral contact elements between rim and inner surface of the tire is completed by the initial gap information from the tire’s cross section. It enables the model to deal with obstacles like cleats or potholes, which are of the same dimension (height or depth) as the tire sidewall. With a typical passenger car tire, wheel load peaks in impact situations such as misuse tests may reach more than 70,000.

Testing rim and tire local material properties or measuring the vertical

stiffness of tire and rim are two possible methods of parameter determination. The first approach was to measure the vertical stiffness of the combination of tire and rim. The results where used for determination of the five new parameters. Comparison between measurement and simulation in Figure 3 shows sufficient accuracy. Looking at suspension loads, the vertical force is of interest, and the torque generated by the asymmetry of the rim.

A test rig was built, able to measure the rim stiffness at both sides separately as well as the combined stiffness of tire and rim. This leads to different rim stiffness functions on left and right rim sides, based on three measurements: rim left and right load deflection relation and the tire and rim combination with the results in eight parameters.

In Figure 4, showing up to 60mm vertical displacement, the second or inner load path with contact between rim and tire is not active. In consequence, the torque remains zero and the vertical stiffness is a linear function of displacement. At greater displacements, the second load path becomes active and the torque goes up to 900Nm and above.

Using the model extension, the critical velocity of a simple mbs vehicle model is investigated. Two versions of the vehicle model have been built, one using joints between suspension and body and the other using bushings. The suspension geometry at the front axle differs from that at the rear to show the influence of the suspension design. The vehicle runs through a pothole obstacle and is at equilibrium with constant velocity before.

The front axle results are shown in upper part of Figure 5. The vertical and the longitudinal forces correlate with the critical velocity, which is to be found between 30 and 40km/h, around 35km/h.

F z[

kN]

0 10 20 30 70 80 90 100

FzMx

linear

Mx[

kNm

m]

70

60

50

20

10

0

1000

900

800

700

300

200

100

0

F z[

kN]

dz [mm]0 10 20 40 50 60

70

60

50

30

20

10

0

simulation

measurement

f 1 ( s 1 )

s 1

g 1 2

g 1 3

( s 3 )f 3

f 2 ( s 2 )

s 3

s 3s ( 2 )

s ( 3 )

Figure 2: Model of the inner contact mechanism

Figure 3: Comparison between measurement and simulation shows suitable accuracy Figure 4: Torque from inner contact, with up to 60mm vertical displacement shown

dz [mm]

64

Only at 30km/h (the red lines), the wheel reaches the bottom of the pothole. At any other velocities, the contact takes place at the positive inclined surface of the obstacle, and at the last edge, the maximum peak load is reached. The bushing version of the vehicle model generates fewer tire forces. This can be explained by the relative motion of the wheel in the joint version, where the wheel is able to avoid larger force levels by moving round the obstacle. The front axle runs into the hole based on a steady vehicle state. At the rear axle, effects from the pitch motion, influenced by the wheelbase for instance, build some interaction between the impact situation and vehicle properties.

The correlation between vertical and longitudinal forces vanishes at the rear axle (lower part of Figure 5) and the bushing version generates a similar force level compared with the joint version. The critical velocity at the rear axle is higher than at the front axle when looking at the results in the vertical force direction. One reason is the different suspension geometry at the rear. Another reason for differences between front and rear tire

forces results from pitch angle excitation when the front axle crosses the obstacle.

The impact situation – expressed by the time history of the rear wheel load – depends strongly on the pitch angle of the vehicle. During front impact, the rear axle is loaded and the pitch angle motion leads to vertical wheel travel. Reaching the first edge of the obstacle, the rear wheel moves up at lower velocity. At higher vehicle velocity, the wheel moves down when reaching the first edge, and the impact loads will be higher. This explains the differences in the critical velocity between front and rear axle.

Simulation results are usually compared with measurements. To do this, a test vehicle with a complete passenger car rear axle including differential gear box and equipped with several sensors and cameras is built for indoor tests. An mbs simulation model of the car is used for comparison. Two different obstacles are built – a cleat and a pothole. Only the rear axle crosses the obstacles because of the different track at front and rear axles.

The tests were run by IAT mbH in Berlin. Through the simulation model instead of the real vehicle, it was possible

to minimize the number of critical velocity tests. The measurement results identified the same range of maximum suspension loads as predicted from the simulation. Comparison between measurement and simulation showed some limitations of the mbs model such as the assumption of a massless spring, or the formulation of bump-stop forces.

The possibility of simulating full vehicle misuse behavior enables OEMs to obtain fundamental information regarding the misuse-relevant force level and damage chain in an early phase of the design process. Extensions of the tire model RMOD-K FB were made in order to run such simulations.

A major aim of these extensions was to add only a small additional computational effort and a small number of parameters, measured on a new test rig. The simulation approach to assess a vehicle’s overall misuse behavior helps reduce the number of prototype tests necessary which involve, at a very late stage of the development process, the risk of additional, extremely cost-intensive and time-consuming modifications and validation loops. tire

0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0t [s]

35.0

30.0

25.0

20.0

5.0

0.0

-5.0

-10.0

F [k

N]

60 [km/h]

50

45 40

35

30

vx_30_jointvx_35_jointvx_40_jointvx_45_jointvx_50_jointvx_60_jointvx_30_bushvx_35_bushvx_40_bushvx_45_bushvx_50_bushvx_60_bush

Fx Rearvx_30_jointvx_35_jointvx_40_jointvx_45_jointvx_50_jointvx_60_jointvx_30_bushvx_35_bushvx_40_bushvx_45_bushvx_50_bushvx_60_bush

30

35

4045

50

60 [km/h]

F [k

N]

0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0t [s]

60.0

50.0

40.0

10.0

0.0

-10.0

Fz Rear

vx_30_jointvx_35_jointvx_40_jointvx_45_jointvx_50_jointvx_60_jointvx_30_bushvx_35_bushvx_40_bushvx_45_bushvx_50_bushvx_60_bush60 [km/h]

50

45

4035

30

F [k

N]

60.0

50.0

40.0

10.0

0.0

-10.00.1 0.25 0.4 0.55 0.7t [s]

Fz Front

vx_30_jointvx_35_jointvx_40_jointvx_45_jointvx_50_jointvx_60_jointvx_30_bushvx_35_bushvx_40_bushvx_45_bushvx_50_bushvx_60_bush

3035

40

45

50

60 [km/h]

Fx Front

F [k

N]

35.0

30.0

25.0

20.0

5.0

0.0

-5.0

-10.00.1 0.25 0.4 0.55 0.7t [s]

Figure 5: The front axle results are shown in the upper two graphs, and the correlation between vertical and longitudinal forces vanishes at the rear axle in the lower graphs

Compounding Resins for Adhesion and

Reinforcement in Tires and Industrial Belts

Penacolite® Resins B-16, B-19-S and B-18-S -low fuming resins for wire adhesion

Penacolite® Resins B-21-S and B-20-S - non-fuming resins with low hygroscopicity

Penacolite® Resin I-868-C - resorcinol-blocked isocyanate resin for carcass compound and adhesive activated polyester

ULTI-PRO® 100 Processing Aid - reduces compound viscosity and increases hardness

Dipping Resins for Synthetic Industrial Fabrics,

including Aramid, Polyester and Glass

Penacolite® Resins R-2170, R-2200 and R-50 -industry standards for RFL formulations in a range of viscosity options

Resorcinol-Blocked

Isocyanates for Polyester

Penacolite® Resin I-168-L - liquid resin for adhesive activated polyester in single-step dip systemsPenacolite® Resin I-868-D - stable dispersion with

unblocking temperature for pre-coat for non-adhesive activated polyester

is a wholly-owned subsidiary of Occidental Petroleum Corporation

Resorcinol Bonding Agents to Enhance

Adhesion, Reliability and Performance

Global Sales Technical Center

INDSPEC Chemical Corp.1010 William Pitt WayPittsburgh, PA 15238 U.S.A.Phone: (+1)-412-826-3666Fax: (+1)-412-826-3699

INDSPEC Chemical B.V. Phone: (+31)-10-275-6060Gebouw de Goudesesingel Fax: (+31)-10-414-3035Kipstraat 8-103011 RT Rotterdam U.S. and Canada SalesThe Netherlands Toll Free (+1)-800-782-8134

66

An EC-funded Fricti@n project (2006-2008) developed sensor fusion technologies for friction potential estimation. In addition,

tire sensors were studied in depth. A predecessor project of Fricti@n, the Apollo project,1 concentrated only on tire sensors and resulted in a special 3-in-1 test tire. The tire possessed three different sensors: a MEMS (micro-electro-mechanical system) acceleration sensor, a piezoelectric strain sensor, and an optical position detection sensor. Since tire sensors had a minor role in the Fricti@n project, it seemed evident that there were no resources to develop all these sensor types further.

The acceleration sensor was extremely interesting, but it is being extensively studied at the moment. Some interesting results based on strain sensors were presented.2 However, the piezoelectric strain sensor was not durable enough even for research purposes and was the easiest to ignore.

The optical position detection sensor, selected to be the tire-based sensor in the Fricti@n project, showed excellent

performance in tire force estimation. Based on the 2008 Aachen Colloquium,3

the intention is to explain how the optical tire sensor is being further developed toward real-time tire force estimation, and its capability to detect hydroplaning.

The optical tire sensor principle is shown in Figure 1. The core of the optical

tire sensor is a two-dimensional position-sensitive detector (PSD) that utilizes photodiode surface resistance.4 The PSD is located on the rim and can detect the movement of a light-emitting diode (LED) that is glued into the innerliner of the tire. The intensity of the LED is not constant versus angular displacement. For example, a 10° angular displacement means approximately 2% lower intensity. A plano-convex (PCX) lens with anti-reflection coating focuses the light onto the sensor. The effective focal length is 9mm, roughly the distance of the lens from the sensor. The sensor setup has been installed into a special dividable rim. The tire in these tests was a winter tire without studs (friction tire, 225/60R16). A Li-ion battery was used as a power supply.5

The wireless data transfer system was the same as that used in the Apollo project.6 The resolution of each channel was 12 bits and the sampling rate was approximately 5,100Hz. The data was transformed into CAN message

Tire-road contact information for driver assistance systems: an optical sensor approachOptical tire sensor technology is an outstanding research and development tool for use in bringing the intelligent tire closer to a market reality

by Ari J. Tuononen, Helsinki University of Technology (TKK), Finland

Wheel rim

PSDsensor

Lens

Intensity of LED vs angular displacement

LED

Z

Y

Figure 1: The optical tire sensor, based on a 2D position-sensitive detector that uses photodiode surface resistance

Figure 2: Development of the optical tire sensor, starting from Apollo until the completion of the Fricti@n project

67

format to be ready for the vehicle network.

Figure 2 shows the lifecycle of the optical tire sensor, starting from Apollo until the end of Fricti@n. The first modification to the tire sensor was to embed the displacement calculation in the tire. In Apollo this calculation was done in post-processing, and thus the new sensor version saved a lot of processor time in real-time operation. The resolution of the data improved as a result of the analog signal processing before digitizing. The improved electronics were located in the alloy housing, which made possible the exact parallel installation of the optical components (Figure 3).

With intensive testing completed on the first Fricti@n prototype sensor, there was great enthusiasm to test the tire sensor on a truck. The same sensor electronics were installed into a separate sensor module. The sensor module is easily installed without removing the tire (Figure 4), and the flange joint was welded to the truck wheel rim.

The tire forces can be calculated only once per rotation. Thus it is essential to know the exact position of the sensor. An additional magnetic pick-up sensor was used to synchronize the rotation

angle of the optical tire sensor exactly. The sensor was installed in the rim and aligned with the optical sensor. A magnet was also installed into the suspension or test rig to indicate the upright position of the sensor. A new tire force estimate is calculated and a new rotation started when the sensor passes by the magnet.

Vertical signalThe vertical movement of the LED can be calculated from the intensity I

psd:

Figure 5 shows a comparison of the tire sensor measurement and FEM simulation (run by Nokian Tires). The peak value is found to be linear to the vertical force. An area of movement z is calculated:

where is the rotation angle. The area is also linear to the vertical force and more independent of any noise or errors during the contact.

Even if the vertical movement correlates well with the vertical force,

in the final version of the algorithms it is ignored. This is due to the fact that LED intensity depends on temperature, supply voltage, and LED orientation angle. The final vertical force estimation is based on the longitudinal movement signal, because it is almost independent of those environmental factors.

The longitudinal signal carries a lot of information (Figure 6). The amplitude is proportional to the contact length and it can be further exploited to estimate vertical force. The amplitude is calculated:

where

consists of all values of the x-signal for one rotation.

The longitudinal force shifts the longitudinal signal up- or downward, but amplitude is maintained almost completely (Figure 7). The recursive mean value of the longitudinal signal correlates with the longitudinal force. The mean value is calculated:

Figure 3: The first Fricti@n tire sensor prototype, comprising an alloy housing with circuit board and lens (on top) Figure 4: An optical tire sensor module for a truck rim

0.035

0.03

0.025

0.02

0.015

0.01

0.005

0

Verti

cal m

ovem

ent

LED position angle [º]

Wheel load

5700N

4900N

4100N

3250N

-150 -100 -50 0 50 100 150 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 6000Position

50004000

16

14

12

10

8

6

4

2

24

Load

Figure 5: Tire sensor vertical movement with different wheel loads (measurement shown on left, FEM on the right)

“The amplitude is proportional to the contact length and it can be further exploited to estimate vertical force”

68

and the longitudinal force is calculated after a completed rotation:

where c is the respective parameter defined from a calibration run. Figure 7 shows the estimated tire force comparison for the test rig measurement during braking. The tire force is slightly underestimated in some sections and values have some offset around zero forces, where only rolling resistance exists.

The vertical force is then calculated:

where the parabolic term is needed because of the inverse square relation of intensity and displacement. The x, c

z,x gain, and c

z,x offset terms are needed

to compensate the vertical force estimate under longitudinal force, which slightly increases signal amplitude.

The tire sensor estimate and test rig measurement are compared in Figure 8. The tire sensor can estimate vertical forces very accurately if no simultaneous longitudinal force exists. The influence of the compensation term can be seen on the right in Figure 8, where the braking sequence is the same as in Figure 7 (right). The additional term in Equation 6 compensates for the increased amplitude during braking.

The lateral force is calculated from the lateral movement signal. Equationally to the vertical signal (Equation 2), the lateral signal peak value and area of one rotation are proportional to the lateral force. However, the area of movement depends on rotational velocity and it has to be compensated. In real-time calculation this was a major problem because of the jitter and resulted in completely wrong lateral forces. Thus, the recursive mean value calculation (cf Equation 4)

was implemented:

and lateral force estimate:

The lateral movement signal is shown in Figure 9 on the left and the tire sensor estimate and test rig measurement are compared on the right.

Hydroplaning severely hampers interaction between the tire and the road. In hydroplaning the tire contact patch can be roughly divided into three sections7

(Figure 10). In zone A, the inertial effect of the water dominates and no contact between the tire and the road surface exists. In zone B, some rubber-road contact exists, but the viscous effect of the water squeezing out from the contact area limits this area. Zone C represents full wet road contact.

Figure 11 shows the tire sensor measurement in transition from dry tarmac to an 8mm water reservoir. The left- and right-hand figures relate to the same data; only the view is different.

The elevation of the front part of the

patch can be seen from the increased distance between the LED and the unloaded radius. The signal also shifts slightly towards smaller rotation angles, which can be seen on the right in the figure. The drop in the peak values just before hydroplaning reveals the descent of the tire into the water reservoir and means reduced vertical tire force.

A more detailed analysis of optical tire sensor behavior in hydroplaning has been carried out.8 The most difficult task of real-time hydroplaning detection is to distinguish wheel load deviation from the hydroplaning phenomenon. This can be done, for example, by calculating the shifting of the area covered by the signal. However, this is influenced by the longitudinal force, assumed to be minor for a hydroplaning tire.

The optical tire sensor can very accurately estimate tire forces as long as the tire inflation pressure is known. In the previous results the inflation pressure was always the same as during calibration measurement in the test rig. The inflation pressure increases enough in normal driving to introduce a bias to the tire force estimate. However, if there is ever a tire force sensor based on carcass

4

3

2

1

0

-1

-2

-3

-4

-5

Long

itudi

nal m

ovem

ent

Wheel load

5700N

4900N

4100N

3250N

-150 -100 -50 0 50 100 1505000

4000

Load 6000 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

-5

-4

-3

-2

-1

0

1

2

Disp

lace

men

t [m

m]

x 103

LED position angle [º]

Figure 6: Tire sensor longitudinal movement with different wheel loads (measurement on left, FEM on right)

“The most difficult task of real-time hydroplaning detection is to distinguish wheel-load deviation from the hydroplaning phenomenon”

Figure 7: Longitudinal signal in braking (left) and tire sensor force estimate compared to test rig measurement (right)

1 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09Time [s] x105

Long

itudi

nal m

ovem

ent [

lsb]

2240

2220

2200

2180

2160

2140

2120

2100

2080Braking applied

Tire sensor estimate

Test rig measurement

Time [s]

Long

itudi

nal f

orce

[N]

1800

1600

1400

1200

1000

800

600

400

200

0

-2000 20 40 60 80 100 120 140 160 180

69

movements in production, the direct TPMS will certainly be standard equipment at that time.

The longitudinal movement signal has been found to carry most of the information. It reveals the contact length and the longitudinal forces. Furthermore, it can detect similar differences in hydroplaning for the vertical movement. The longitudinal forces cannot be accurately measured during ABS braking, since the update rate is limited to one sample per rotation.

Vertical movement signal (intensity) is disturbed by LED alignment, temperature etc. Despite the facts, these limitations (especially LED alignment) are characteristic only for an optical sensor, which requires a light source in the innerliner. There are many optical devices that can measure distance without a light source in the object. This concept would be more production-oriented than the sensor considered in this paper. The reliable measurement of innerliner vertical movement would certainly allow at least dynamic wheel load estimation and possibly also the detection of hydroplaning.

The lateral tire force can be accurately estimated from the lateral movement signal. It makes it possible to estimate the vehicle slip angle accurately and rapidly, because the tire force is acting on the chassis before accelerations and rotational velocities are generated. The side slip estimator based on tire sensor forces is one of the further activities.

The lateral signal could also provide indications about the aligning moment of the tire, but no systematic studies have been done on this subject. In some tests, it was observed that the lateral signal peak was shifted forwards (towards the trailing edge) in the tire contact patch at small slip angles, which could be influenced by the aligning moment. The aligning moment would be interesting to estimate, because it describes very nicely the tire operating state, together with the lateral tire force.9

The optical tire sensor has been an excellent research tool to study dynamic tire behavior. It is not even intended to be a product, but resources have been allocated to study what information is available if the deflections of the tire carcass are known. It can also be used to validate complex physical tire models such as FEM much progress has been

achieved in embedding the cyclic data analysis algorithms into standard low-cost MCUs. tire

“The optical tire sensor has been an excellent research tool to study dynamic tire behavior”

Figure 8: Comparison of tire sensor estimate and test rig measurement results. The influence of the

compensation term can be seen on the right, where the braking sequence is the same as in Figure 7 (right)

Figure 9: Lateral movement signal is shown left, and tire sensor estimate and test rig measurement on the right Figure 10: The three-zone concept of hydroplaning

Time [s]

Late

ral f

orce

[N]

Late

ral d

ispl

acem

ent [

m]

0.02

0.018

0.016

0.014

0.012

0.01

0.008

0.006

0.004

0.0002

0

Sensor position angle [˚]

4500

4000

3500

3000

2500

1500

1000

500

00 5 10 15 20 25 30

Test rig measurementOptical tire sensor

Slip angle-Fy8°-3526 N6°-3360 N4°-3118 N2°-1982 N

2°

4°

6° 8°

Number of rotation [-] Number of rotation [-]0 100 200 300 400 500 600 700 0 50 100 150 200 250

Test rig measurementTire force estimateTire force estimatew/ox amp compensation

5000

4500

4000

3500

3000

2500

2000

1500

1000

500

0

Tire sensor estimateTest rig measurement

Verti

cal f

orce

[N]

4500

4000

3500

3000

2500

2000

1500

1000

500

0

-500Ve

rtica

l for

ce [N

]

References1) Apollo final report, 7.7. 2008 Available online at:

www.vtt.fi/apollo2) Morinaga, Hiroshi, 2006, The Possibility of

Intelligent Tire (Technology of Contact Area Information Sensing), Yokohama Fisita World Automotive Congress.

3) Tuononen, Ari J. et al. (2008) Tire sensing approach for friction estimation in Fricti@n project, Aachen: Aachen Colloquium, 6-8.10.2008.

4) Hamamatsu, PSD S5991-01 data sheet.5) Tuononen, Ari J. (2008) Optical position detection

to measure tire carcass deflections, Vehicle System Dynamics, 46:6, 471 — 481, Abingdon: Taylor & Francis.

6) Nepote, Andrea et al. (2004) The intelligent tire: A new challenge for automotive electronics, Barcelona: Fisita 2004 World Automotive Congress.

7) Browne, A. L, Cheng, H. and Kistler, A. (1972) Dynamic hydroplaning of pneumatic tires, Wear 20, pp 1–28.

8) Tuononen, Ari J, Hartikainen Lassi (2008) Optical Position Detection Sensor to Measure Tire Carcass Deflections in Hydroplaning, Vehicle Systems Modelling and Testing, Vol 3 Nr 3, Inderscience.

9) Pasterkamp, W. R. (1997) The Tire as Sensor to Estimate Friction, Delft University Press.

70

NASA’s renewed interest in manned lunar flight has resulted in research to create new mobile platforms for lunar exploration.

Unfortunately, current tire technologies cannot withstand the extremes of a lunar environment, where temperatures range from +250°F to -250°F between sunlight and shadow.

With the support of Michelin Inc, one Clemson University faculty member spent two weeks at NASA’s Jet Propulsion Laboratory (JPL) to research new methods to replace the polyurethane of the terrestrial Tweel. Developed by JPL, the Tweel, which has performance akin to a pneumatic tire, was an appealing option for use in the JPL ATHLETE (All Terrain Hex Limbed Extra-Terrestrial Explorer) project, seen in Figure 1, as it is not vulnerable to the pressures of expansion and contraction caused by extreme lunar temperatures that would make the rubber of conventional tires very brittle.

Timothy Rhyne and Steven Cron, two researchers at Michelin’s North American R&D department, are the men behind the non-pneumatic Tweel, which comprises three major components: a critical shear beam (or shear band), two inextensible membranes, and deformable spokes. In the Tweel, the shear beam is sandwiched

between two inextensible membranes that restrict the motion of the internal geometry, allowing shear only.

Unlike traditional balloon tires, the Tweel’s unique top-loading weight distribution shifts the vehicle weight from the hub, up through the spokes, to the arch of the inextensible membranes and the shear beam sandwich. The shear beam then carries the load to the ground, where it deflects to form a relatively uniform

contact patch (Figure 2). The Tweel also replicates four important pneumatic tire properties: low contact pressure; low stiffness; high load-carrying efficiency; and low energy loss from obstacle impact, which, with its unique configuration, make the Tweel an ideal candidate for use in a harsh lunar environment.

As part of their senior design project, mechanical engineering students at Clemson University designed three

NASA, Michelin, and South Carolina’s Clemson University have joined forces to research and test tread designs that may one day be used on lunar vehicles

by Jaifeng ‘Jeff’ Ma & Professor Joshua Summers, Clemson University, USA

Tire-sand interaction research for lunar applications

Figure 1: The All Terrain Hex Limbed Extra Terrestrial Explorer (ATHLETE) Rover undergoing testing on rough ground

Figure 2: Main components of the non-pneumatic wheel

Figure 3 (right): 2D FEM model replicating the prototype

71

non-pneumatic Tweels, all of which had a different construction encompassing bristle design, the segmented cylinder design, and the helical coil design. In subsequent research projects, the Clemson University research team created computational simulations of the interaction of each Tweel design and used the results, presented here, to improve Tweel traction performance. The modeling these students undertook plays an important role in predicting optimum parameters to improve Tweel capabilities and performance in all future physical prototypes. Their computational modeling of Tweel-soil interaction can be used for improving current tire designs used in both construction equipment and off-road vehicles.

The team has also developed computational models to provide rapid simulation and analysis tools for use in modeling interaction between tires and sand to inform improved military tire designs for use in the arduous confines of both Afghanistan and Iraq. From this work, the team uses new mission requirements provided by the US Army (temperature cycles, speeds, loads) to create new sand-tire interaction models for use in new tread designs with superior traction performance.

Simulating sand-tire interactions to accelerate and evaluate non-pneumatic tires and tread design requires the latest technology available. In their design work, the Clemson team has used the finite element method (FEM), the discrete element method (DEM), the meshless integral method (MIM) – as instantiated in ABAQUS and PFC2D commercial software – and in-house MIM codes to conduct FEM, DEM, and MIM modeling and simulation. The team also developed Ant Farm and Merry-Go-Round systems to test the endurance of their developed tire and tread systems, in addition to several other experimental systems. Eleven prototypes of different shapes and materials were also built and tested.

Although the Tweel is unique, more work is needed to make it viable.

“Developing high-traction and high-wear-resistant solutions that will last for 50,000 miles in temperatures ranging from -380°F to +260°F is a daunting challenge,” says Professor Joshua Summers of Clemson University.

One alternative to traditional experimentation that involves computational simulation of tire-soil interaction has been most useful for improving tire designs used in heavy construction equipment, off-road vehicles,

and in advancing understanding of the soil compaction.

“To improve our understanding of the Tweel and to optimize its traction performance in our project, we will have to run a sizeable number of computational simulations of Tweel-soil interactions,” Summers continues. “This computational modeling plays an important role for predicting optimum parameters before building or improving physical prototypes and conducting experiments.”

Compared with existing analytical and empirical methods, computational analysis can provide detailed models of tire-soil interaction and predict deformations of the tire and the soil within an acceptable accuracy once the constitutive behavior of both materials is well represented. The geometry of the contact area between the tire and the soil cannot be prescribed beforehand; it is a complex interplay resulting from the deformation of the tire and the soil.

Using the stress distributions in the tire and the soil, and the interface between the two, Summers and his team can predict and compare these experimental results for validation. They are currently using three computational modeling tools in their soil analysis: FEM, DEM, and MIM. The purpose is to use these models to develop a fully computational model that can accurately model tire-sand interaction. They can then use this model to search for novel tread concepts to improve the performance of tire-on-sand (traction), and to avoid slip.

FEM has been one of the most widely used computational methods for determining the interaction between the tire and the soil. It requires less computational effort and provides acceptable levels of accuracy. Summers and his team treat their soil samples as an elastoplastic solid, meaning that they can use a single equation to determine the mechanics of the soil behavior as it is affected by their Tweels.

Figure 4: Contour plot of the maximum logarithmic strain using Lade’s single hardening model of sand

Figure 5: Contour plot of the maximum principal plastic strain between Tweel and soil Figure 6: Contour plot of the maximum principal plastic strain between Tweel and soil

72

Figure 3 shows the contour plot of this plastic strain (PE12) simulation of Tweel-sand interaction using sand from Lebanon, New Hampshire, USA. The graphic shows the low contact pressure and high-load-carrying efficiency of the Tweel, with a low energy loss from obstacle impact. Because the wheel is non-pneumatic, it can be made of materials that can withstand extreme temperatures on both terrains.

Figure 4 also shows the boundary conditions and contour plot of the maximum logarithmic strain of a plane strain test using Lade’s single-hardening model of sand. This model was used as a user subroutine (UMAT) in ABAQUS. From the contour plot of the logarithmic strain, it is evident that this sophisticated, 11-parameter sand model captures failure zone/shear band.

Figures 5 and 6 show the contours of the maximum principal plastic strains of the two instants of the interaction model between 3D Tweel and soil (t=40 and t=80). The soil is divided into three layers, representing the change in bulk density and shear strength with depth. Each layer has a modified Drucker-Prager material

model with cap plasticity. These figures show the 3D Tweel/tire model.

The Clemson team also used DEM to model the flow of granular soil and sand under and away from their experimental Tweel designs.

“The DEM is a numerical technique that keeps track of the location of each particle, its velocity, and the forces acting on it, and calculates the resulting acceleration of the particle and its new position, velocity, and forces,” says Summers. “In FEM analysis, the granular material is treated as continuum by averaging the physics across many particles.”

The result is a macroscale system that inherits its properties from the properties of microscale particles. Figure 7 shows a DEM simulation of a direct shear test of sand, containing 33,000 grains. The box lid applies a constant average pressure to the top surface of the sand as the entire top half of the box moves slowly to the right, shearing the specimen. The force required to shear it is recorded for plotting as a function of displacement.

The team is also developing the meshless method that is different from

FEM to analyze the performance of tire-on-sand (traction) for eventual use on NASA moon rovers. This method, which can handle the moving boundaries and changing geometry of soil-tread interactions to avoid possible traction slip, consists of a distribution of nodes over the problem domain with the solution derived from these nodes. Consequently, nodal connectivity in the meshless method becomes simpler and more flexible. Each node, through the use of a localized weight function, is always connected to nearby nodes, which means that the number of interconnected nodes is unlimited.

The Clemson team developed its own meshless integral method to determine linear elasticity, and elastoplasticity for small and large deformations of tread ware. Jianfeng Ma, a post-doctoral researcher in Summers’ group, believes this method to be vastly superior to conventional tire-soil interaction modeling.

“This innovative idea is an improvement over existing methods in that essential boundary conditions are enforced directly, efficiently, and

Figure 7: DEM simulation of a direct shear test of sand Figures 8 (a) (left) and (b) (right): Deformation and von Mises stress distribution along bottom line for punch test

Figure 9: General test setup of the Ant Farm Figure 10: Overview of the lunar testing system setup Figure 11: The lunar drivetrain design subsystem

73

accurately,” says Ma. “The implementation is straightforward and more efficient than indirect methods.”

He went on to say that this meshless method needs no background mesh for integration, and it is more stable and robust than meshless methods, and than the finite point method. It is also more accurate than FEM for the same number of nodes.

Figures 8a and 8b show deformation and von Mises stress (in MPa) distribution along the bottom edge for a punch test. The model geometry had 561 nodes, and the spline weight function and linear monomial basis were used. The Young’s modulus is E=203 GPa, =0.3. After yielding, bilinear stress-strain behavior was assumed with a yield stress of 260MPa, and an elastoplastic tangent modulus of 1GPa. Von Mises yielding criterion was used. The prescribed displacement of the left half of the top edge was U2=-0.08m. The two figures indicate that the meshless results are in good agreement with the FEM solutions.

An undergraduate team of students also generated the novel concept of the Ant Farm, which aids the calibration of

the DEM analysis to the sand used in the endurance testing of the lunar wheels.

“The purpose of the Ant Farm is to verify a software simulation made using an FEM program and a DEM program,” says Summers. “We want to use the Ant Farm to lower an object (known as a punch) into sand at a constant rate for a specified period of time, to see how the sand deforms around the punch.”

The Ant Farm system includes a box made of fine-grain plywood, with a plexiglas front and aluminum punches, and was entirely designed and built by two students: an entering freshman and a sophomore mechanical engineering student. There are two punch shapes: a semicircle and a rectangle with rounded corners. It also has a sand box, a motor support, motor, cam, cam follower, and a punch. Figure 9 shows the general test setup of the Ant Farm.

The punch rests on the sand and is pushed into the sand by the cam follower. The punch force is approximated by the FEM and DEM simulations to run to 110N of force. The key information collected is the vertical force exerted on the punch versus time and also versus

displacement with secondary information being the visual displacement of the layers of sand.

The NASA Tweel project simulates the rotation of a lunar wheel, which must closely mimic the moon’s gravity and terrain for an appropriate testing environment. The Merry-Go-Round system for the team’s Tweel endurance tests includes a turnable subsystem, drive train subsystem, and sensory and DAQ system (Figure 10).

The turntable has a 20-sided polygon shaped trough with a circle of radius 6m and one of 7m combined to form the path of the trough. A framework underneath prevents deformation of the trough bottom. To stabilize the center column and trough area, the team set steel cables at the bottom of the frame and another set to the top of the center column. Cable tension was calculated for the amount of weight supported. Because the trough is suspended, the casters act as a safety measure, which can decelerate the turntable should any cables slip.

The drivetrain (Figure 11) has a 5bhp three-phase motor, a gearbox reducer, shaft, and suspension. A PTO shaft and

Figure 12: Testing location on a 25° constant incline Figure 13: Testing vehicle with prototype installed Figure 14: Hall effect counter and display on ATV axle

Figure 15(b): C&M wire tread concept

Figure 15(a): Coir rug tread concept

Figure 16(b): Carpet concept prototype tread

Figure 16(a): Astro-Turf prototype tread

Figure 17: Paddle wheel prototype tire tread

74

clutch provide movement in only one direction, with a four-link suspension to guide the lunar wheel on the correct path.

Sensors around the MGR measure the wheel slip, wheel deflection, wheel surface penetration, and wheel fatigue. Pictures of each are taken at different stages until the lunar wheel physically fails to ensure the greatest possible input of information.

Determining the most effective tread design for obtaining traction in sand occurred over a five-week period. The eleven prototypes built and tested integrate concepts used in sand racing, and snow and mud driving. The team analyzed test results to determine which components were the most effective in producing traction. The tests were performed on a hill with a constant incline of approximately 25° (Figure 12). The dry sand testbed was 45ft long and 8ft wide, with a sand depth of 6-8in. A flat 22ft entrance was created in front of the hill to enable the test vehicle to a reach steady state speed of 5mph before reaching the incline.

A Kawasaki Brute Force ATV was used to test each tread concept (Figure 13). The front tires were replaced with racing slicks mounted to a 13in wheel so that each experimental tread would encounter fresh, unpacked sand. Tire pressure was set at 25psi. The team applied an experimental tread while the other

was ‘left slick’ for each run. The ATV was kept in four-wheel drive mode and a low gear for all tests to measure tread performance in driving situations. A throttle hard stop governed the 5mph speed before reaching the incline. A camera mounted under the ATV recorded all tread-sand interaction.

The same driver (weighing 160 lb) conducted all tests, and the sand was raked before testing began and between each run. Tread performance was quantified from the distance the ATV traveled up the hill before encountering full slip, and the percentage of slippage during the run. The team measured the distance from hill bottom to the front tire of the ATV, allowing the tape to follow the contour of the hill. Sensors and steel wire strips tied to the ATV axle counted the number of experimental tread revolutions (Figure 14).

This counter and the circumference of the tire (20in) determined the distance the ATV should have traveled if no slip occurred. They compared this theoretical distance to the actual (measured) distance and calculated the percentage of slippage accordingly.

The 11 concept treads designed and tested by the team are: coir rug, C&M wire tread, Astro-Turf prototype, carpet concept, paddle wheel prototype, V-shaped grouser prototype, aluminum wire mesh roll prototype, concave tread

prototype, traction tape prototype, deformable adhesive tread concept prior to testing, and snow chain prototype (Figures 15 to 23).

Of the designs tested, the team found that five exhibited superior traction capability: the carpet, V-shaped grousers (both orientations), aluminum circumferential mesh, and concave treads.

“We found that these were the most impressive by far,” says Summers. “Our computational modeling capabilities using the FEM, the DEM, and MIM will be used to bear this out and find more optimal configurations using these concepts as seeds.”

So what’s next? Summers and his team will continue to refine their work to develop treads that are more robust for use in extremes of heat and cold.

“Our group’s philosophy is to push state-of-the-art design,” he concludes. “The work that we’re doing here will help to create the next generation of engineers who can look at seemingly unsolvable problems, find their core essence, createa solution, and communicate the solution to the customer.” tire

References1) The ATHLETE Rover, http://www.nasa.gov/

images/content/168147main_image_feature_748_ ys_full.jpg .

2) Rhyne, T. B. and Steven, S.M. Development of a non-pneumatic wheel, Conference of Tire Society (2005)

Figure 18: V-shaped grouser prototype tread Figure 19: Aluminum wire mesh roll prototype tread Figure 20: Concave tread prototype

Figure 21: Traction tape prototype tread Figure 22: Deformable adhesive tread concept prior to testing Figure 23: Snow chain prototype

SINCE 1998

The 9th International Exhibition on Rubber Technology

The 9th International Exhibition on Rubber Technology

Shanghai New International Expo Center,

Pudong, Shanghai, China

Tel: +86-10-58650277, Fax: +86-10-58650288, Email: rubbertech@chrubber.com

-- running parallel with Reifen China,the 3rd Asian Essen Tire Show

www.rubbertech.com.cnthe official show website

Exhibit Categories

Rubber Machinery: All kinds of rubber machinery for rubber processing and analysis equipment for rubber products, etc.Rubber Raw Materials: Nature rubber, synthetic rubber, carcass materials, thermoplastic elastomer, reclaimed rubber, rubber powder, and rubber compound, etc.Rubber Chemicals: Rubber accelerators, vulcanising agents, antioxidants, antiscorching agents, carbon black, fillers, etc.Non-tire Rubber Products: Rubber belts & hoses, engeering rubber products, and other parts for auto and mine, etc.

Regis

ter Now

Organizer

China United Rubber Corporation

Co-organizer

China Rubber Industry AssociationChina Synthetic Rubber Industry Association

Special Supportor

National Tech. Center of RP Industry

11, 12, 13 November 2009

76

Europe’s new regulatory framework: requiring and inspiring new technology and innovationEuropean tire development is being driven by the market requirements arising from the new European regulations for increased road safety and environmental protection, and the future of raw materials sourced from crude oil

by Fazilet Cinaralp, secretary general, European Tyre & Rubber Manufacturers’ Association (ETRMA), Belgium

Europe is the home of the tire according to every definition of the term. The pneumatic tire was invented here and significant

developments in the tire industry often happened first in Europe.

Europe has maintained its technological hold on the world’s tire industry, and it is in Europe that the tire industry fights its technological battles. Customers in Europe are the most demanding in terms of performance, lifetime, service and innovation. Any company that is successful in Europe has the technological tools to compete everywhere else in the world.

The EU is also the most segmented of any global market. Within a couple of years, it is also likely to be the most tightly regulated market in the world!

Current EU directives set standards for tires, including: tire integrity, dimensions and markings; minimum tread depth; tire/road noise limits; a landfill ban for used tires; polycyclic aromatic hydrocarbons restrictions in extender oils for tires; and chemicals policy (REACh).

In early 2008, the European Commission (EC) put forward a proposal for a regulation on the general safety of motor vehicles (COM(2008)316), laying down harmonized rules on the construction of motor vehicles and minimum safety and environmental performance requirements for tires.

In February 2007 the EC had adopted a comprehensive strategy, “A Competitive

Automotive Regulatory Framework for the 21st Century” for the European car industry, to keep the manufacturing of motorcars viable on a long-term basis, at prices affordable to consumers. The strategy covers a variety of areas, (including reduction of administrative burden, environmental sustainability, and road safety). In this document, the EC welcomed the recommendation to simplify the current whole-vehicle type-approval regulatory framework, published in the CARS 21 report.

The technical requirements for the type-approval of motor vehicles with regard to numerous safety and

environmental elements have been harmonized at EU-level to avoid requirements that differ from one member state to another, and to ensure a high level of road safety and environmental protection. New technologies now available can dramatically improve vehicle safety (such as ESC), or reduce CO

2

emissions (such as low rolling resistance tires). Research has shown that there would be significant benefits if such technologies were introduced on new vehicles as standard.

In the ‘explanatory’ section of the EC proposal (COM(2008)316), there is the following: “Concerning tires, the proposal introduces requirements on the following: rolling noise limit values, rolling resistance limit values, wet grip and tire pressure monitoring systems (TPMS).

For each of the above issues, the following options were generally

considered: do nothing, voluntary/market solution, mandatory solution with lower technical difficulty, and mandatory solution with higher technical difficulty.’

The option of, “mandatory solution with higher technical difficulty” was selected for all the issues.

The statement then continues: “In the case of rolling resistance, wet grip, and TPMS, research and product development are already at a quite advanced stage so that implementing the more stringent technical requirement in a relatively short timetable is considered to be feasible.

“Concerning TPMS, setting a higher standard would maximize safety benefits, and ensure the maximum likelihood of reaching CO

2 reduction targets.”

Finally, on the question of rolling noise, the statement says: “Concerning rolling noise, the option of ‘mandatory solution with higher technical difficulty’ has also been retained because this is the only one that would ensure the fulfillment of the environmental objectives of the proposal. However, since requirements on noise are more challenging than the other requirements, a longer implementation period has been foreseen.”

There is general agreement that three areas have to be addressed together to improve road safety and reduce CO

2

emissions by any meaningful amount: vehicle technology, driver behavior, and road infrastructure.

The challenge is to get the statutory bodies across the EU, and at various levels, to cooperate in this integrated approach. Safety and environmental features in vehicle technology are either introduced by industry (commercial introduction or voluntary measures), or by EU legislation. Driver behavior is largely the responsibility of member states

“Any company which is successful in Europe has the technological tools to compete everywhere else in the world”

77

through training, licensing requirements, enforcement and information campaigns, and road-safety and user organizations. Infrastructure is the competence of national, regional, and local authorities.

Safeguarding a common level of road safety across the borders in Europe (the right of citizens to road safety wherever they travel) can only be achieved by a joint effort between the various authorities plus industry and user organizations.

An additional level of integration in approach is required to improve tire performance. In tire design, rolling resistance is inextricably linked to wet grip and tire life. Reduce rolling resistance and the risk is making the tire less safe, or making a tire which does not last anywhere near as long. Wet grip relates directly to safety. Better wet grip means shorter stopping distances and reduced tendency to skid or slide in poor weather.

Being the only vehicle element in contact with the road, the tire helps grip, handling, and fuel efficiency of the vehicle. The proposed law recognizes the importance of an integrated approach to tire performance for consumers by considering safety and environmental performance in the same legislative act.

In particular, the legislation sees the mandatory fitment of TPMS on all passenger cars from 2012. Correct tire inflation pressure is essential for delivering tire performance and every effort must be deployed to guarantee proper setting and maintenance of the inflation pressure, which has an unquestionable influence on tire rolling resistance. A permanent 25% underinflation increases the tire rolling resistance by 10%, which, in turn, results in about 2% more fuel consumption. Industry studies have shown that 40% of drivers drive regularly with tires underinflated by at least -0.5 bar.

Also maximum rolling resistance thresholds will be introduced from October 2012, followed by a second stage from 2016. Lower rolling resistant tires, together with TPMS fitted as standard equipment to the vehicles on the European roads, will contribute to reducing fuel consumption, save 5-7 million tons of CO

2 annually, or

reducing CO2 emissions by approximately

5g/km driven.For the first time, by also setting

lower rolling resistance requirements for truck tires, the EU is anticipating CO

2

provisions beyond the current European automotive policy targets.

For minimum wet grip requirements for passenger car tires from October 2012, the legislation considered that voluntary options were not appropriate since wet grip was an important safety issue. The ‘no action’ option might lead to a reduction in wet grip standards due to the pressure on manufacturers to meet noise and rolling resistance requirements. Therefore, compulsory wet grip approval was considered the best option (Tire Safety Regulation, SEC (2008)1999).

For a second time (after the European directive 2001/43 introduced a series of external noise limits to tires), the legislation pushes the technological limits of the tire industry with the new limits

imposed, bringing a further 3-5 dB(A) reduction to tire rolling noise from 2012.

Tire/road noise in real life is generated on a wide range of different road surfaces. Improving the quality of roads is a key factor in achieving external noise reductions, and making a positive impact on decreasing CO

2 emissions. The EC

should continue assessing EU roads according to their noise levels with a view to setting maximum noise limits and laying down road surface specifications.

As Andreas Schwab, European Parliament rapporteur for global safety regulation, stated, “With improved tire design and stricter requirements for rolling resistance and stricter noise limits, the external noise of motor vehicles is reduced. It is clear that member states should endorse the associated investments in the medium-term to reduce noise emissions in the same way in all modes of transport, and therefore also provide noise protection in road construction by using improved quieter road surface coverings.”

Mandatory tire labeling and more consumer information is a further initiative from the EC, released in November 2008, enabling consumers to purchase tires which surpass the limits set in the safety regulation (COM(2008)316) for tire rolling resistance and wet grip performance. It aims at accelerating market transformation towards even more fuel efficient and safer tires.

Information on tire ranking for PC (class C1), LT (class C2) and CV (class

C3) tires is to be provided in the sequential order of the fuel efficiency class (A to G), the wet grip class (A to G), and the measured rolling noise value in dB.

These complementary regulations show the potential of an integrated approach. It is about setting challenging environmental and safety objectives, and providing visibility and regulatory certainty to industry. Only such certainty enables it to plan investments to develop product offerings with ever-increasing safety, reducing environmental impact over the next decade.

The stipulations in the new regulations will be proposed at the level of UNECE, and are to be integrated into UNECE

regulation 117, and will be directly applicable in all 27 member states. Tires sold in the EU market will have to demonstrate their compliance with these new requirements via type-approval certification and markings on the sidewall.

The challenge will be for the 27 enforcement authorities to ensure consistent and effective market surveillance. If importers are not rigorously held to the same standards, then EU manufacturers will be put at a competitive disadvantage, and the consumers’ safety may be at serious risk.

A Fraunhofer Institute study on the 2001 implementation of the ‘Fuel consumption labeling for cars in Germany’ illustrated that self-certification measures are often abused due to a lack of inspection mechanisms and sanctions. More than 50% of the electric appliances checked in 320 German stores were incorrectly labeled or not labeled at all. Clearly that must not happen with tires.

ETRMA is concerned about the risk of inadequate implementation and policing, to the disadvantage of the consumer and the European tire industry. ETRMA is insisting on the need for member states to set up robust verification mechanisms to protect industry and consumers from the less scrupulous producers and importers. Equally important, there must be severe penalties that punish enough to be truly dissuasive, and they must be rigorously applied when products do not comply with the legislative requirements. tire

“ETRMA is concerned about the risk of inadequate implementation and policing, to the disadvantage of the consumer, and of the European tire industry”

78

More than five years ago Gabo Qualimeter introduced the Gabometer – a modified universal Goodrich

Flexometer (Goodrich Flexometer according to ASTM D 623/ISO 4666). This type of flexometer is used to analyze the ‘warming up’ or ‘heat build-up’ of rubber specimens during mechanical oscillations at high frequencies (30-100Hz or above) and at high dynamic deformation amplitudes (up to ±6mm).

The Gabometer works basically in the compression mode and can be equipped with up to three independent temperature measurement units in order to determine the heat build-up of the test specimen. A needle-type thermocouple allows measurement of the temperature inside the samples.

Nowadays this flexometer can be rearmed to a complete high-power DMA/DMTS system named Eplexor 2000 or 4000, providing in addition tensile, shear, bending, creep and fatigue test and universal test procedures. Additional fields of application are material testing, thermal analysis, components testing and quality control, as well as relaxation/retardation tests. Moisture required? No problem. This DMA/DMTS can be equipped with a humidity generator (Hygromator) to create environmental conditions between 5% and 95% rH.

In other words: flexometer and DMA/DMTS tests with one instrument.

Nevertheless, like the Gabometer and the table-desk Eplexor series, the new Eplexor 4000 models can be equipped with the robot system ASSS – a fully automatic sample feeding system.

The technique of DMA/DMTS is mainly being applied in rubber and polymer research, in production and in the processing industry. Dynamic mechanical spectroscopy is an off-resonance method working with forced oscillations. It is being used for dynamic compression, tension, bending or shear applications. Dynamic material testing provides precise information about the viscoelastic mechanical properties

(e.g. elastic modulus |E*| and viscoelastic damping tan ). In the case of pure elastic materials (e.g. stainless steel spring) the sample response to a constant sinusoidal dynamic deformation is in phase ( = 0°) with the external excitation. Viscous materials show a time-delayed response to an external excitation. In this case a phase shift of 90° (‘out of phase’ component) can be observed ( = 90°, see Figure 1).

An intelligent universal flexometer offers the advantages of a standard flexometer and the benefits of DMA/DMTA testing in one state-of-the-art instrumentby Horst Deckmann, Gabo Qualimeter Testanlagen GmbH, Germany

A universal flexometerand dynamic mechanical thermal spectrometer

Figure 1: Mechanical sample excitation and response Figure 2: Complex modulus in the complex diagram

Figure 3a: Heat build-up tests and thermal set (standard)Figure 3b: Heat build-up tests, thermal set and viscoelastic data (advanced)

79

Rubber materials demonstrate an intermediate behavior showing a mixture of elastic and viscous properties.

Figure 2 shows the complex modulus |E*| consisting of the two components E’ (storage modulus) and E’’ (loss modulus). E’ and E’’ describe the elastic and viscous mechanical properties of the material. E’’ corresponds to the loss energy dissipating into the sample. The ratio of E’’/E’ (tan )describes the phase shift between sample excitation and response. A metal spring contributes only elastic, an oil-only viscous components toward E*. Most polymers demonstrate a mixture of these properties. They are viscoelastic.

A quite different approach to dynamic sample loading was introduced by Goodrich. In the last century Goodrich developed the Goodrich Flexometer. This instrument is used to heat up rubber specimens by a sinusoidal mechanical deformation. Cylindrical rubber specimens with a diameter of approximately 17.8mm and a height of approximately 25mm are loaded with a constant static load between 1 and 2 MPa. Additionally, a sinusoidal dynamic load at a test frequency of 30Hz with a peak to peak amplitude of up to 6.6mm (±3.3mm) is superimposed. Due to intrinsic friction processes caused by the cyclic dynamic deformation, the test specimens start to heat up. The heat build-up of the samples is recorded by a contact thermo sensor on the surface of the sample (see Figure 3). In addition, due to the thermo-mechanical dynamic aging process, most samples start to reduce their sample height during the test (thermal set). Thermal set will also be recorded.

Standard test instruments (Figure 3a) are restricted for heat build-up and thermal set only, whereas the Eplexor 2000 and 4000 with flexometer function are able to record viscoelastic data (modulus, tan delta, etc – Figure 3b).

The missing step – a combination of both technologies – was introduced by Gabo Qualimeter Testanlagen GmbH. The combination of both methods allows the monitoring of changes in modulus and tan depending on the sample temperature during the heat build-up process.

The instrumentsFigure 4 shows the Eplexor 4000N system. The total load of this instrument is up to 4,000N. The total dynamic strain amplitude for the basic model depends on the configuration. Up to two dynamic strain transducers within the range from ±1.5mm up to +/-11mm can be installed simultaneously. These technical features enable the user to easily operate either in the DMA mode with the request for small deformations (+/-1.5mm) and a very high resolution (resolution <5Nm) or in the Flexometer mode using large deformations (heat build-up and blow-out test). Additionally, as for universal tensile testers, due to a large set of exchangeable force transducers (from 10N to 5,000N capacity) the sensitivity of the instrument can be adjusted to the requirements of the test. Just plug and play. Exchange can be made within a few minutes.

All instruments can be equipped with oven systems operating between -150°C and 500°C. These temperature chambers can be equipped with up to three thermocouples. The first thermo sensor controls the oven temperature whereas sensor number two determines the sample temperature at the sample surface contact area.

Thermocouple number three is designed as a needle-type thermocouple (optional). Again two versions are

available. A horizontal needle-type can be penetrated into the sample during the test. This needle has to be adjusted manually. The second needle operates vertically. A pneumatic needle drive with a position control mechanism penetrates the needle automatically into the core of the sample at the end of the test. Both needle-type thermo sensors record the intrinsic sample temperature (Figure 5).

The test resultsDynamic heat build-up tests provide a better understanding of the thermal aging properties of the corresponding elastomers.

Of course, due to homogeneity fluctuations within different batches of test specimen the most important requirement for a reliable flexometer test is the reproducibility of the test results. Figure 6 shows such a reproducibility test.

Two test specimens (same batch: cylindrical samples for compression load) were tested at 30Hz using identical static and dynamic load conditions. The tests were performed at room temperature. In both, the heat build-up within the center of the sample (measurement was carried out with the horizontal needle thermocouple) as well as the temperature on the ‘skin’ of the sample (corresponding to standard ASTM D 623, DIN 53 333) were recorded.

Electronic cabinet

Eplexor 4000

Liquidnitrogencontainer

PC Table with screen

Figure 4: The high-power Eplexor 4000 N can also be fitted with ASSS, a fully automatic sample feeding system

80

The results show the temperature rise for both samples are in a satisfactory accordance.

Even the tan delta (material damping) measurement shows an excellent reproducibility.

Tests with practical aspects – what is the benefit of an additional temperature sensor (needle thermocouple)?Nowadays, heat build-up tests performed with Goodrich Flexometers are quite common. However, occasionally with conventional flexometer tests there is a lack of information due to the nature of the basic test principle. Flexometer tests normally record only the skin temperature and thermal set. Changes of the viscoelastic properties related to the dynamic ‘warming-up’ test are not recorded.

Figure 7 shows a heat build-up test for two different compounds, A and B.

The temperature rise of samples A and B is more or less identical for the skin temperature. The skin temperature sensor records the temperature on the surface of the sample at the cross-sectional area on top of the test cylinder. Unfortunately this sensor does not supply the information concerning the real temperature in the core of the sample. To obtain this information, the horizontal needle-type thermocouple was used for the measurement of the core temperature in the center of the sample during the test.

Sample B shows a remarkably higher warming up within the center compared with sample A. Due to this behavior the lifecycle of sample B can be expected to be much shorter than that of sample A.

Indeed, practical tests show the expected behavior. Sample A has a longer lifecycle than sample B. The conventional heat build-up test still shows the same heat build-up behavior for sample A and B. It seems the recording of the skin temperature is not sensitive enough to show any difference.

Only with a second thermocouple – in this case the needle-type sensor located in the center of the test specimen – can additional information be provided.

But what is the reason for this difference? The basic compounds used for samples A and B are identical. In fact different carbon black fillers are used for system samples A and B. The heat created in the core of sample A during the heat build-up test can be transferred much more quickly to the surface of the sample compared with sample B. The heat conductivity of polymer A seems to be much better than that of compound B. The core temperature of sample A is reduced. Consequently the

duration of the lifecycle could be improved.

The lifecycle of compounds is not the only major topic for the tire industry: the lifetime of tire cords and the friction of the tire cords within the rubber are also of considerable interest.

Friction behavior of tire cords – determination of failure limitsFor the reinforcement of rubber components such as car tires, conveyor belts or v-belts, application-optimized tire cords are used. Usually during the curing process the tire cords inside the rubber components will be mechanically and thermally stressed.

Therefore the internal friction between the cords and the surrounding rubber is influenced by the curing process. With the DMA/DMTS system a straightforward test procedure for testing this kind of interaction between cord and rubber can be carried out.

Knowledge about the friction behavior of the tire cord within the rubber matrix is essential, in order to develop products with reliable and durable dynamic properties. The most convenient test setup is the H-test or T-test geometry (Figure 8). For both test specimens the tire cords are cured within the rubber matrix.

As shown in Figure 8, a T-test sample is clamped within the special holder. For showing the test procedure, two different cord rubber systems were investigated.

The target of this test is the determination of the dynamic mechanical stress limits. For this reason different load levels were applied on the samples. The static as well as the dynamic deformation levels were increased stepwise. On each

Figure 5: Compression test configuration with needle thermocouple

Figure 6: Reproducibility of test results in a heat build-up test Figure 7: Comparison of two compounds – Samples A and B

upper compression plate - with athermocouple in the centre of the upper plate

Horizontal needle thermocouple

lower compression plate

ASTM test specimen

81

deformation level a mechanical oscillation with test frequency of 10Hz was applied for about 250 seconds. The ratio between the static and dynamic deformation was maintained as constant on each deformation level. As expected, with increasing deformation the total force

increases too. After reaching a certain deformation level, for one sample (red curve) the cord starts to be pulled out from the rubber matrix, whereas the second sample (blue curve) stays stable up to the end of the test. All tests are made at room temperature (Figure 9).

Dynamic aging close to the mechanical failure limitOf considerable interest is the lifecycle behavior of both samples in the case of cyclic dynamic ageing. The applied load conditions should be very close to the mechanical failure limit. The above two cord rubber systems are investigated again with a fatigue test setup at a test frequency of 50Hz.

Figure 10 displays the behavior of the complex modulus |E*| and tan delta. Again two new samples of both the discussed cord rubber systems were investigated. The total load applied for these tests was approximately 90% of the ‘pull-out’ load conditions found in the previous test (see Figure 8).

Again, for the first sample (red curves) after a duration of above 2,000 seconds the destruction of the T-test sample takes place, whereas the second sample (blue curves) stays stable until the end of the test procedure. These tests were made at room temperature.

SummaryNot only conventional DMA/DMTS tests can be carried out with the new Eplexor 2000 and 4000 system line of Gabo Qualimeter Testanlagen GmbH. The combination of fatigue test, universal tensile or compression test applications, heat build-up and blow-out tests, and the complete application range of thermal analysis (temperature sweeps, master curves (time temperature superposition according to William Landel Ferry), frequency and strain sweeps, etc) and much more is the key feature of this instrument series. tire

Figure 8: H-Test or T-Test for dynamic testing of tire cords, with the T-Test sample clamped within the special holder

Figure 9: Static dynamic sweep for two different cord rubber-matrix systems Figure 10: Fatigue test of cord rubber matrices, showing complex modulus and tan delta

82

P olybutadiene rubber (BR) belongs, beside styrene butadiene rubber (SBR), to the ‘general purpose’ butadiene-

based synthetic rubbers, which contribute 24.5% of the synthetic rubber volume produced worldwide. Typical applications are in tires, high-impact polystyrene (HIPS), conveyor belts, caterpillar tread blocks, footwear soles, golf balls, and V-belts.

The major application for polybutadiene rubbers is in tires. More than 70% of the BR produced worldwide is consumed in this application. Figure 1 shows the average BR content in different elastomer blends of a typical radial car tire with two steel belts and a nylon cap-ply.

Polybutadiene rubbers are highly elastic, have a very high degree of resistance to dynamic stress, and retain these properties even at extremely low temperatures. Blending natural rubber or styrene butadiene rubber with polybutadiene rubber improves the abrasion resistance, crack resistance, and heat build-up under dynamic load.

Butadiene rubbers are produced by the solution polymerization process with the aid of organometallic catalysts. Commercially, Ziegler-Natta type catalyst systems are used. Due to an activation reaction between the catalyst components in a separate aging step or in situ, the real active complexes are formed to be a single site system or as a mixture of different catalytic active sites.

The catalyst system, which is characterized by the particular metal (neodymium, cobalt, nickel, titanium, or lithium), and the way to activate the different Ziegler-Natta type catalysts influences the microstructure and macrostructure of the polymers and therefore their properties (Table 1).

Unique micro- and macrostructured high-cis polybutadiene rubbers can be used to improve the dynamic performance of tires

by Dr Heike Kloppenburg & David Hardy, Performance Butadiene Rubbers, Lanxess Deutschland GmbH

From golf balls to rolling resistance

“Active complexes are formed to be a single site system or as a mixture of different catalytic active sites”

Table 1: Summary of polymer properties of different Polybutadiene Rubber grades

Catalyst Nd (1)

LXSNd (2) Co (1) Ni (2) Ti (2) Li (1)

Microstructure

1.4-cis BR [%] 98 97 - 98 97 97 93 38

1.2-vinyl BR [%] 0.5 0.5 – 0.9 1.3 – 2.5 1.2 – 2.0 5 11

Glass transition temperature Tg [°C]

-109 -109 -107 -107 -104 -93

Branching (%) < 5 < 5 10 – 50 5 - 40 15 < 5

Polydispersity (Mw / Mn) 2.1 2.4 – 4 3 - 4 4 - 6 3 - 4 2.0

(1) Product range of Lanxess

(2) Product variety not including Lanxess products

Figure 1: There are several areas within the tire where BR is applied to improve the tire properties

Carcass (2)

Innerliner

Carcass (1)

Apex (-25% BR)

Bead wire

Chafer

(50-100% BR)

Hump strip

Sidewall (30-50% BR)

Shoulder strip

(30-50% BR)

Tread (70% SBR

& 30% BR))

Belts

3. Nylon

2. Steel

1. Steel

Radial car tire

Two steel belts/nylon cap ply

Figures are showing the

average BR content in

the elastomer blend

83

The high-cis BR grades based on Nd, Co, and Ni catalysts differ only slightly in 1.4-cis content. More important is the difference in the 1.2-vinyl content, which for Co and Ni catalysts is up to five times higher than for Nd catalysts. This leads

directly to differences in the low-temperature crystallization behavior and in the glass transition temperature. Ti-BR contains slightly less 1.4-cis than the other high-cis BR grades and therefore has a higher crystallite melting temperature

and slightly higher glass transition temperature. Where these microstructural properties are concerned, the Co- and Ni-BR grades are intermediate between the Nd and Ti grades.

The macrostructure is distinguished by the molecular weight distribution (polydispersity) and the degree of branching. Figure 2 compares the molecular weight distribution of different high-cis BR grades, all having the same Mooney viscosity of 44. Figure 3 shows the GPC curves of NdBR grades of Lanxess with the same polydispersity and different Mooney viscosities.

For better physical properties, such as lower hysteresis, low heat build-up and finally lower rolling resistance, the number of free polymer chain ends should be limited. The number of polymer chains is best described by the numeric average of the molecular weight, (Mn) meaning that a higher Mn will produce fewer chain ends per unit volume, whereas the weight average (Mw) is more related to the polymer viscosity.

The higher the Mooney viscosity, the higher the Mn and Mw. Figure 4 shows the dependence of Mn and Mw on the Mooney viscosity for different linear NdBR grades with the polydispersity of 2.1 (red) and 3.5 (purple). The lower the polydispersity, the higher the Mn value for the same Mooney viscosity. One needs an increase in Mooney viscosity of approximately 20 points in order to get the same Mn value for the polymer with the high polydispersity compared to the

Table 2: Analytic data for the different BR grades

Catalyst system Ni (4) Ni (4) Ni (4) Co (3) Co (3) Nd (5) Nd (3) Nd (5) Nd (3) Nd (3)

Compound number 1 2 3 4 5 6 7 8 9 10

High-cis BR grade NiBR - 1 NiBR - 2 NiBR - 3 TAKTENE 220

TAKTENE 1203 G1

BUNACB Nd40

BUNACB 24

BUNACB Nd60

BUNACB 22

BUNAPBR 4007

1.2-vinyl BR [%] 1,2 1,5 1,7 2,5 1,3 0,9 0,6 1,0 0,5 0,5

ML 1+4 (100) [MU] 46 45 43 39 44 44 43 63 64 75

MSR (6) [MU/s] 0,34 0,69 0,41 0,40 0,40 0,59 0,66 0,54 0,68 0,72

Relax 30 [%] 19,7 5,4 14,4 14,5 13,8 7,2 4,9 9,2 5,5 5,3

Solution viscosity (7) [mPas] 98 312 177 54 76 433 174 812 350 585

Mn [kg/mol] 79 118 83 91 116 125 153 146 208 216

Mw [kg/mol] 458 453 460 327 360 446 300 512 434 456

PDI 5,3 5,3 3,6 3,2 3,6 2,8 3,6 2,0 3,5 2,1 2,0

(1) Product range of Lanxess

(2) Competitor product

(3) Product range of Lanxess SA (ex Petroflex)

(4) MSR: defined as absolute numbers (without negative sign)

(5) SV: 5,43 w-% of polymer in toluene at 20°C

Figure 2: GPC curves of different high-cis BR grades with Mooney viscosity of 44

Figure 3: GPC curves of NdBR grades of Lanxess with the same polydispersity and different Mooney viscosities

84

low polydispersity material. For CoBR grades (blue) with a higher degree of polymer branching, the Mn value for the same Mooney viscosity drops even more than for linear polymers with the same polydispersity.

To calculate the number of polymer chain ends, the number of polymer chain ends per molecule is important, which is two for a linear polymer and three or even more for a branched molecule.

There are different methods to analyze the polymer branching. The GPC-MALLS curve in Figure 2 shows with the RMS graph a higher branching degree of the high molecular weight polymer fraction of CoBR and NiBR compared to the linear NdBR. Another analytical tool is the branching index, which is defined as the ratio of Mooney viscosity to solution viscosity, where the solution viscosity is influenced by the molecule contraction in solution due to long chain branching and is exponentially dependent on the molecular weight. The Mooney Stress Relaxation (MSR) is a bulk property and depends on chemical branching and physical entanglements of the polymer chains. The higher the branching index and the lower the MSR values, the higher the polymer branching.

Table 2 contains analytical data for the different BR grades used in subsequent tire compound tests.

Figure 5 summarizes the branching analysis of these polymer grades. Both CoBR grades are highly branched, whereas the NdBR grades are more linear. The NiBR grades are taken to demonstrate the different branching degrees from highly branched (NiBR 1) up to a rather linear grade (NiBR 2).

Whereas the branching index depends on the solution viscosity, the MSR value as a bulk property seems to be more related to final compound properties.

Another differentiation of the polymers is the polydispersity (Figure 6). All NdBR grades ex Lanxess have a very low polydispersity of only two, followed by a group between three and four, with different CoBR, NiBR, and NdBR grades, and end up at a very high polydispersity of more than five with one NiBR.

As described earlier, the Mn value of the molecular weight distribution can best be used to indicate the number of polymer chain ends in the material: the higher the Mn the fewer chain ends are present.

To include the different branching degrees of the polymers, a polymer

Figure 4: Dependence of Mn and Mw on the Mooney viscosity for linear NdBR and branched CoBR grades with different polydispersities

Figure 5: Branching analysis of the polymer grades

Figure 6: Polydispersity and numeric average of the molecular weight distribution of the compound samples

Figure 7: Polymer characterization factor (PCF) as factor for free polymer chain ends and vinyl-1.2 content of the compound samples

85

characterization factor was defined as the product of Mn and the MSR value. Figure 7 shows this polymer characterization factor (PCF) as a factor for free polymer chain ends for the various BRs against the respective vinyl-1.2 content as an indicator of very short side groups. The lower the vinyl content and the higher the PCF, the better are the dynamic physical properties.

The compound test formulations used were as follows: 100phr high-cis BR; 50phr Corax N 326; 4phr Vivatec 500; 3phr Edenor C 18 98-100; 2phr Vulkanox 4020/LG; 3phr Vulkanox HS/LG; 2phr Zinkweiss Rotsiegel; 3phr Resin SP-1068;

Table 3: Compound results

Catalyst system Ni (4) Ni (4) Ni (4) Co (3) Co (3) Nd (5) Nd (3) Nd (5) Nd (3) Nd (3)

Compound number 1 2 3 4 5 6 7 8 9 10

High-cis BR grades NiBR - 1 NiBR - 2 NiBR - 3 TAKTENE220

TAKTENE1203 G1

BUNACB Nd40

BUNACB 24

BUNACB Nd60

BUNACB 22

BUNAPBR 4007

Mooney ML1+4 (100) @ final compound

ML 1+4 Compound [MU] 46 52 47 42 47 53 56 70 71 80

MSR [MU/s] 0,34 0,51 0,40 0,34 0,38 0,45 0,51 0,42 0,52 0,51

Monsanto – MDR @ 160°C, 30 min

S’ minimum [dNm] 1,70 2,00 1,77 1,61 1,69 2,03 1,95 2,59 2,63 2,94

S’ maximum [dNm] 14,4 17,7 15,8 15,4 17,6 17,9 19,2 18,7 19,5 20,1

S’ end [dNm] 11,7 14,8 13,1 12,7 14,6 15,5 16,6 16,4 17,3 17,5

Delta S’ [dNm] 12,7 15,7 14,0 13,8 15,9 15,9 17,2 16,1 16,9 17,1

TS1 [s] 146 129 137 136 136 116 133 100 108 100

TS2 [s] 172 156 163 163 164 149 168 139 145 140

Conversion 10% [s] 158 148 152 151 156 140 162 129 138 133

Conversion 50% [s] 211 198 205 206 208 197 220 189 199 194

Conversion 90% [s] 315 296 310 320 315 306 333 294 307 288

Conversion 95% [s] 360 340 357 371 363 357 385 342 358 333

Vulcanisate properties

Hardness @ 23°C [Sh A] 56,8 60,7 58,3 57,2 59,7 60,9 61,9 61,5 62,4 63,1

Hardness @ 60°C [Sh A] 53,3 58,2 56,3 55,4 57,2 58,8 58,5 60,6 61,2 60,8

Rebound @ 23°C [%] 54,3 56,2 55,2 54,1 57,7 58,7 59,3 61,5 62,8 62,1

Rebound @ 60°C [%] 57,0 59,8 58,8 56,6 60,5 61,1 62,1 63,1 65,2 67,2

Tensile strength @ 23°C

S10 [MPa] 0,4 0,5 0,5 0,5 0,5 0,5 0,5 0,6 0,6 0,5

S100 [MPa] 1,4 1,6 1,6 1,4 1,7 1,6 1,6 1,7 1,7 1,8

S300 [MPa] 5,3 5,5 5,4 4,9 6,2 5,3 5,3 6,3 5,4 6,1

D Median [%] 478 482 401 497 497 517 482 462 500 436

F Median [MPa] 11,5 11,7 8,7 11,2 14 12,9 11,6 12,8 12,7 11,6

MTS amplitude sweep @ 60°C, 1Hz

G* (0,5%) [MPa] 1,72 2,04 1,72 1,85 1,93 2,08 2,16 1,90 1,97 1,73

G* (15%) [MPa] 0,87 1,01 0,90 0,91 1,01 1,04 1,09 1,03 1,07 1,02

tan d (max.) 0,182 0,179 0,181 0,190 0,164 0,161 0,163 0,145 0,145 0,138

(3) Product range of Lanxess

(4) Competitor product

(5) Product range of Lanxess SA (ex Petroflex)

Figure 8: Vulcanization curve and the polymer characterization factor (PCF) as factor for free polymer chain ends correlated to delta torque

86

2phr Antilux 654; 1.4phr Vulkacit CZ/EGC; 2.36phr Rhenogran IS 60-75. The compounds were mixed in a lab Banbury (1.5-liter size) using a one-pass mix.

The rubber mix was left overnight to cool and then passed through a mill three times. Vulcanization was carried out at 160°C for cure times previously established in the rheometer (t90+5 minutes for tensile test samples and t90+10 minutes for thicker dynamic test

samples). Test samples for the tests were cut from the resulting vulcanized sheets.

Table 3 contains the compound results. For the polymer grades #1 to # 7 with a Mooney viscosity range of 39 to 46, the compound Mooney viscosities differ between 42 and 56.

The higher the MSR value, the more linear is the polymer grade and the Mooney viscosity of the compound increases.

The MDR analysis at 160°C (see Figure 8) gives a characteristic curve, where the Delta torque value (Delta S’) can be correlated to the polymer characterization factor as indirect number for the free polymer chain ends.

The influence of the strain amplitude sweep on the loss factor (tan ) is shown in Figure 9. The highest tan maximum is given by the highly branched CoBR grade Taktene 220, followed by all three NiBR grades. What becomes evident is the lower maximum tan and therefore reduced energy dissipation for the NdBR grades with a higher Mooney viscosity.

Another tool to predict rolling resistance is the rebound measurement at 60°C. In Figure 10, the correlation of rebound 60°C to the maximum loss factor measured on the MTS shows the polymers in the same sequence that is known from the polymer characterization factor PCF (Figure 11). The higher the PCF, the lower the number of free polymer chain ends and the lower the energy dissipation of the polymer.

In summary, the neodymium (Nd) catalyst system employed at Lanxess gives a polymer structure that cannot be obtained with other catalyst systems and which enables marked improvements in vulcanisate performance to be achieved. Lanxess Nd-BR grades are high-cis-1.4 grades with the lowest vinyl-1.2 content, a very narrow molecular weight distribution, and a low degree of branching.

This results in an NdBR series exhibiting reduced energy dissipation relative to all other BR types and can be used to reduce the rolling resistance of tires. It is hypothesized that the higher linearity and narrower MWD of NdBR results in less free chain ends, which in turn lowers the hysteresis of the compound and thus will reduce its rolling resistance. In addition, the reduced energy dissipation will also increase the rebound value.

The golf ball industry has taken advantage of NdBR’s superior resilience for the manufacture of high-performance golf balls over the past decade. Reducing the rolling resistance of tires will become increasingly important under the context of EU directives (COM 2008 0779) requiring tire labels in Europe from 2012. This label will describe the fuel efficiency of the corresponding tires and NdBR’s will have an important contribution to make in achieving a good rating. tire

Figure 9: Tangent delta dependence on strain at 60°C

Figure 10: Correlation of resilience to tangent delta maximum

Figure 11: Tangent delta maximum versus PCF

88

In terms of reliability and durability, resistance to tearing is a key factor in the performance of tire rolling tread.

Although many efforts have focused on gaining a better understanding of the occurrence and development of cracks in elastomer compounds, there is a lack of advanced testing methods dedicated to the study of crack propagation, beyond simple fracture tests.

01dB-Metravib, the world-renowned dynamic mechanical analysis specialist, introduces a brand-new fatigue test machine: the DMA+300, specifically designed and dimensioned for fatigue tests on elastomers and crack growth tests.

The principle of the crack growth test consists of initiating a crack on one side of the elastomer film specimen, and following up the growth of the crack using an optical system during a dynamic mechanical fatigue test.

Unlike servohydraulic machines, which are often used for fatigue tests on components, DMA+300 does not require any hydraulic station – just a connection to the electrical line and to the compressed air network.

The DMA+300’ s electrodynamic actuator does not require any specific maintenance (no more than an audio loudspeaker) and offers reliability and durability in full adequacy with the very nature of the endurance test that requires one million excitation cycles.

The 300N excitation force is dimensioned so as to offer the power required to cover the stress and strain

A new testing machine for the study of crack growth in elastomer compounds can compare or sort compounds according to their composition, and analyze influential parameters of mechanical excitation

by Hugues Baurier, 01dB-Metravib, France

Fatigue test machinefor crack growth testing

The DMA+300 testing machine

89

domains necessary for the analysis of elastomer compounds.

The specimen geometry is particularly well suited to the requirements of the tire industry as it enables testing not only of compounds that are currently being developed but also compound specimens sampled from new or used tires.

Furthermore, thermal regulation and gas conditioning can be easily optimized on a reduced volume around a specimen of relatively small dimensions.

The test machineThe DMA+300 consists of a floor-standing mechanical frame, control and acquisition electronics, and a computer station equipped with dedicated software.

The dynamic mechanical excitation is delivered by an original electrodynamic actuator developed by 01dB-Metravib, which includes a high-performance guiding and anti-rotation system maintaining the excitation within the specimen plane. Its 300N force range was defined based on the elastomer specimen geometry in order to offer an optimum analysis range.

The dynamic and static components of the specimen strain, as well as the applied force, are measured respectively at each end of the specimen.

The excitation signal is programmed using a new software program, devoted to fatigue and crack growth tests: Multitest.

The excitation signal is controlled by controlling the strain or the stress. The floor-standing mechanical test frame was designed so as to grant the operator easy and comfortable access to the specimen, to the cutting system, and to the crack growth follow-up optical system.

Specimen mounting deviceThe specimen is mounted on a specific mounting template according to a strict protocol, ensuring the jaws are accurately positioned on the same plane and parallel to the specimen plane. Once the specimen is pre-mounted on its jaws and the mounting template, it can be placed on the test machine with no risk of degradation or unwanted stress.

Thermal chamberThermal conditioning is ensured by a dedicated thermal chamber operating through forced convection.

The chamber was designed for an operating temperature range well beyond the requirements commonly encountered for elastomer analysis, i.e., from -150°C to 600°C.

The thermal chamber includes a large window that enables the entire surface of the specimen to be observed. The optical quality of this window allows for highly accurate control and optical analysis of the specimen.

Cutting systemIn order to ensure cutting control and to avoid any risk for the operator when handling cutting tools, 01dB-Metravib

has developed a surrounded cutting system with external manual controls.

The cutting system is integrated inside the front door of the thermal chamber. A cut can then be made in the specimen to initiate a crack, very precisely and repeatedly, without opening the chamber, hence without disturbing the temperature regulation or the gas mix, this operating mode ensures the safety of the operator.

The cutting system includes two blades, located on the left and right of the specimen. The crack can therefore be initiated on either side of the specimen.

Each blade is fixed onto a mobile blade holder moving along the horizontal axis (X) and the vertical axis (Y).

The displacement of each blade is controlled using two thumb wheels, one for horizontal displacement (X axis) and the other one for vertical displacement (Y axis).

Access to and replacement of the blades is quickly achieved by triggering a specific command.

Gas conditioning The test can be carried out by setting the specimen to a specific oxygen rate. The gas conditioning system is used to achieve an air/nitrogen gas mixture and to set the nitrogen rate for the mix.

The air/nitrogen mix is sent to a thermal exchanger and brought to the temperature required for the fatigue test.The specimen mounting device is highly accurate

Multitest software offers multiple excitation waveforms, as well as the import of operator-customized signals.

90

The oxygen rate is measured at the entrance of the thermal chamber using a dedicated probe.

Optical measurement systemThe optical measurement system includes a binocular microscope, a ring light system, a mechanical translation system driving the motions of the microscope along X, Y, and Z axes, and driving electronics interfaced with test software.

Micrometric translation plates driven by a step-by-step engine are used to drive and control the microscope displacements very precisely.

The entire optical system is placed on the right side of the fatigue test machine. It can then instantly be revolved in “crack growth” position or be quickly retracted to give access to the thermal chamber’s opening and to the specimen.

When the crack growth test is started, the microscope is placed in front of the specimen.

After the cutting phase, the operator takes a geometric reference shot that will represent the origin of the crack growth follow-up.

The microscope remains in this position during the entire crack growth test to allow for as many crack

propagation measurements as necessary during the test.

Multitest softwareTest driving and acquisition are achieved by Multitest, which is a brand-new fatigue software program developed by 01dB-Metravib engineers in close collaboration with research engineers from the rubber and tire industries.

Regardless of the user’s profile –

engineering specialist or skilled operator – the Multitest architecture offers optimum user-friendliness to access all functions and capabilities of the machine.

Multitest offers multiple excitation modes (waveforms): sine, haversine, pulse, triangle, square, etc., as well as the import of operator-customized signals.

The excitation frequency can be programmed from 2Hz to 1kHz.

The test can be controlled using different parameters: displacement amplitude, strain, force amplitude, stress, excitation frequency or strain rate.

Multitest enables fatigue test campaigns and crack growth tests to be achieved.

Fatigue test campaignsEach test campaign can consist of different tests chained together. A test campaign can include as many tests as desired.

Beyond fatigue tension tests achieved on elastomer films, other fatigue tests can be achieved with other test modes using other specimen holders and other specimen geometries: tension plates, shear jaws, tension jaws for rectangular specimens, etc.

Crack growth test campaignsCrack growth test campaigns must be composed of different type sequences that are specific to this type of analysis: accommodation, characterization, and crack growth.

The accommodation sequence consists of applying a certain number of excitation cycles to the specimen in order to stabilize it. This operation is only a conditioning phase for the specimen and does not yield any analysis results. The accommodation sequence always occurs before the characterization sequence.

The characterization sequence consists of applying a strain sweep (with imposed frequency or strain rate) to the specimen, and getting the tearing energy (G, in J/m2), i.e. the energy delivered by area unit.

“Multitest enables fatigue test campaigns and crack growth test campaigns to be achieved”

A test specimen exhibiting left and right cracks

An operator using the DMA+300. Note how all the features of the machine fall easily within the operator’s reach

91

The plot of G versus the amplitude of the imposed displacement is particularly interesting to characterize, as it enables comparison of the crack growth behavior of different materials under similar conditions, i.e. at identical tearing energies.

In a crack growth test campaign, the dynamic excitation amplitude setpoint to be applied during a crack growth sequence is automatically calculated and programmed based on the G curve obtained during the characterization sequence, and enables the test conditions corresponding to the targeted tearing energies to be defined.

Cutting stepAs soon as the crack growth sequence is started, the specimen is cut. Cutting the specimen consists of making a cut

of a few millimeters on one side of the specimen (left or right side, or both) to initiate the crack.

The specimen is first subjected to a pre-tension using Multitest, which controls the displacement applied to the specimen, as well as the corresponding force.

The crack growth sequence consists

of following up the crack growth, under given stress conditions. To do so, cut measurements are carried out periodically.

The cut measurement consists of stopping the dynamic test and using the microscope to read the geometric coordinates of the crack tip.

During the crack growth sequence, the user can perform a cut measurement at any time, and stop the dynamic test with a simple mouse click.

Successive acquisitions of the crack tip coordinates throughout the test enable the acquisition of the crack growth curve (length in mm, crack growth rate in mm/number of cycles).

In order to meet the requirements of testing laboratories, Multitest integrates new functions that enable test campaigns (accommodation, characterization, and crack growth) to be defined, starting from an internal work request in Excel format. These functions considerably shorten the time required to define the test on the machine, and decrease the risk of potential input errors by the operator.

Multitest can be used to edit results according to different modes: graphic representations and numerical data, reports on the different types of tests, etc.

The “Analysis View” function is particularly useful to navigate through the test campaigns and the tests themselves to display the results quickly.

Data export as ASCII files enables customized external analysis of all test results.

ConclusionDMA+300 is an innovative testing machine for the study of crack growth in elastomer compounds, either to compare or sort compounds according to their composition, or to analyze influential parameters of mechanical excitation.

Beyond this specificity, DMA+300 is a very versatile test machine that can be used to gather advanced knowledge of the elastomers used to produce tires, combining particularly relevant characteristics and possibilities such as achievement of fatigue tests (tension, compression, shear); achievement of crack growth tests on elastomer films; extended frequency range; high measurement precision; integration of real excitation signals; analysis at controlled temperature; analysis at controlled oxygen rate; and analysis of heat build-up. tire

Curves of both crack growth from left and right cracks presented with the Multitest fatigue software program

Tearing energy

The specimen is cut when the crack sequence starts,

92

If a tire manufacturing company were to be asked if it was interested in saving money on final-finish tire testing equipment, the answer

would be a definite yes. However, the opportunities to avoid extensive lifecycle costs and increase the value that the equipment investment provides are often ignored or overlooked. There are ways to minimize expenditure on final-finish tire testing and measurement, but they require a focus on the ‘right things’ during the supplier evaluation process, the purchasing process, and throughout the operational life of the equipment.

Investment in a single final-finish line to perform tire uniformity, dynamic balance and geometric runout testing and measurements can exceed US$1 million – a major purchase indeed. In actuality, this initial purchasing cost only represents a small portion of the expense that the final-finish operation will incur over the life of the equipment. Typically these costs are less than 20% of the total outlay over a 10-year lifecycle. By focusing on the ‘right things’, and through comprehensive analysis of potential suppliers and equipment options available in the marketplace, a better purchasing decision can be made during the equipment acquisition phase. This results in the minimization of costs to the organization for providing this critical capability.

Recent developments in high-quality, highly integrated tire testing systems have also enhanced a tire company’s ability to address this difficult situation. These new systems are optimized to deliver high customer value throughout the equipment’s lifetime, while adding other features that make the tire testing process easier to manage. By shifting a company’s sole focus away from initial purchase costs to total lifecycle costs, and through careful consideration of newly developed integrated final-finishing systems, a tire

company can reduce its overall required investment in tire testing equipment.

Piece (per tire) costThe most popular metric utilized to evaluate the overall value of purchased equipment is Piece Cost. This value is defined as the total cost per tire required to evaluate force variation uniformity, dynamic unbalance and geometric runout characteristics, for the purpose of final quality screening before delivering that tire to a customer. All costs required to purchase, install, maintain, and operate the equipment on an annual basis are rolled up and divided by the anticipated yearly throughput of the equipment.

Piece Cost = Annualized CostsAnnual Throughput

The tire producer’s goal is to minimize Piece Cost, which can be accomplished by ensuring the highest equipment throughput possible while minimizing annualized costs. To accomplish this, the correct equipment must be chosen during the acquisition phase and complementary strategies must be implemented that will sustain the competitive advantages over

the life of the equipment. The goals of throughput maximization and annualized cost minimization should be discussed. The analysis will also include a review of a rarely discussed or quantified cost component associated with machine repeatability-related measurement performance.

Throughput maximizationEquipment throughput is simply defined as the anticipated number of tires tested per year. This key metric is a direct function of the equipment utilization, typically expressed in units of seconds per year, and the total system cycle time, typically expressed in units of seconds per tire. The annual equipment throughput of the machine is calculated by dividing these two individual quantities.

[ ]EquipmentThroughput

tiresyear

=Equipment Utilization [seconds/year]

Total System Cycle Time [seconds/tire]

As a result, throughput can be maximized by purchasing and adequately maintaining the ‘most productive’ equipment, ensuring the possibility of maximum utilization of machines operating at minimal cycle times.

The only way to guarantee the greatest delivered value over the life of a tire testing machine is to look at the annual costs along with specific throughput information

by Dr Shaun M. Immel, Micro-Poise, USA

Final-finish tire testing:what is the true cost?

Table 1: Daily machine throughput (tires/day) based on the factors of system cycle time and system reliability

93

Two notable items deserve attention when trying to ensure superior equipment utilization: production availability of the equipment, and system reliability of the equipment.

The scheduled operation of the machine, along with daily scheduled downtime, defines the production availability of the equipment. This quantity depends solely on production scheduling and is the same regardless of the final finish equipment used. The other major contributor to machine utilization is system reliability. This factor represents the percentage of time that the machine will stay running in an automatic mode. For automated equipment in a tire plant, both the frequency of breakdowns as well as the time to recover from each instance should be taken into account.

The calculated effects of small differences in system cycle time and system reliability on daily machine throughput are summarized in Table 1. The equipment is assumed to be available for production for 90% of a 24-hour day. The throughput increase resulting from even a small improvement in cycle time or system reliability is quite dramatic.

In the typical equipment operational range for a final-finish testing process, highlighted in Table 1, one second of cycle time improvement can provide more than 200 extra tires per day (that’s 68,000 tires per year or more than 11% additional testing capacity). And one percent of system reliability improvement can provide more than 40 tires per day (that’s 14,000 tires per year or more than 2% additional testing capacity).

Total system cycle time and system reliability are two very important factors that deserve attention and evaluation when selecting a supplier for tire testing equipment that will maximize throughput. Selecting equipment that achieves high availability through excellent system reliability performance while operating at sustainable total system cycle times as low as possible will ensure the best return on investment for that

equipment. A trade-off exists when designing equipment that will operate at low cycle times and concurrently maintain high system reliability. For a given mechanical design, if mechanical motions are programmed to execute and complete rapidly in order to minimize cycle time, wear and tear on components will occur and typically cause more breakdowns and reduce system reliability. The reader is cautioned to recognize and consider this negative interaction when evaluating equipment for purchase. This phenomenon must be understood and accounted for when the supplier is designing the equipment.

Annualized cost minimizationAs previously mentioned, there are several components that go into determining the annualized costs for a system of equipment. These factors include but are not limited to the acquisition costs, installation costs, maintenance costs, and operating costs of the equipment. All these factors are important, but in reality have a very different impact on the total annualized cost of the equipment.

The installation costs are undoubtedly important, but dwarf in comparison with the large amount of capital outlay required to initially purchase such a testing system. With production tire testing equipment, installation costs and other costs associated with ‘start-up time’ are similar among different suppliers. The equipment is heavily automated and can be well tested before leaving a supplier’s facility, so it typically can be made production-ready in a short period of time. Installation materials and labor are also minimal, particularly when compared to other tire manufacturing equipment. Hence this is typically not an area where testing equipment suppliers can differentiate themselves.

The most widely discussed component of total cost is the up-front acquisition costs for procuring testing equipment. Examples of these costs include the complete system purchase price, taxes,

duties, and shipping. Clearly this component is an extremely important consideration when purchasing equipment, but it is often incorrectly used as the primary driver in making purchasing decisions. When total lifecycle costs are analyzed for this type of equipment, the acquisition cost is typically less than 20% of the overall equipment life costs when compared with the other cost categories.

The single largest contributor to annualized cost is unquestionably the operational costs for the equipment. There are a huge number of items worth considering in this category. These include, but are not limited to: operating labor, electric power, compressed air, marking tape, bead lubrication material, hydraulic fluid, grease, various maintenance costs, training costs, and spare parts. There is great potential for equipment suppliers to differentiate themselves in all these areas. A small amount of design intent can go a long way toward reducing costs in each of these categories. And, since the costs are incurred repeatedly over the life of the equipment, they can become extensive.

Unfamiliar cost contributorAn additional high-impact operating cost component, rarely discussed today, is related to the measurement performance of the testing equipment in operation. Not only can insufficient quality of measurement performance directly increase annualized costs by forcing incorrect screening decisions for a given tire, it can indirectly increase costs through the erosion of tire consumer confidence. Direct costs may be accumulated by incorrectly scrapping or downgrading a tire that is truly below the screening limit and considered acceptable. In addition, in some instances increased tire adjustment costs may be incurred as a result of mistakenly shipping an unacceptable tire to a customer.

Consider the example of collecting multiple uniformity measurements of

Figure 1: Normal distribution of uniformity (s=0.75 units) Figure 2: Normal distribution of uniformity (s=0.75 units) Figure 3: Alternate normal distribution of uniformity

Dist

ribut

ion

of In

divi

dual

Mea

sure

men

ts

1.2

1.0

0.8

0.6

0.4

0.2

0.0 Dist

ribut

ion

of In

divi

dual

Mea

sure

men

ts

1.2

1.0

0.8

0.6

0.4

0.2

0.0

Amount of Force Variation in Units of Force

good machinerepeatability

poor machinerepeatability

acceptable tire classified as scrap

screeninglimit

misclassified tire percentage

acceptable tire classified as scrap

screeninglimit

misclassified tire percentage

Amount of Force Variation in Units of ForceAmount of Force Variation in Units of Force

misclassified tire percentage

screeninglimit

Dist

ribut

ion

of In

divi

dual

Mea

sure

men

ts

1.2

1.0

0.8

0.6

0.4

0.2

0.0

Amount of Force Variation in Units of Force7 8 9 10 11 12 13 14 15 7 8 9 10 11 12 13 14 15 9 10 11 12 13 14 15 16 17

poor machinerepeatabilityrejectable tire classified as acceptable

94

an arbitrarily chosen tire that is known to have a ‘true’ uniformity value of 11 units with a machine that has a repeatability standard deviation specification limit of 0.38 units. Under a reasonable assumption of a normal distribution of measurements, the resulting distribution of measurement values may look something like that shown in Figure 1. The solid vertical red line in Figure 1 represents a potential screening limit for that particular tire. During a small percentage of the time, the tire will measure above the specification limit and be rejected unnecessarily. This percentage of time is represented by the black shaded region. This tire would, at a minimum, get measured a second time – requiring extra handling, labor and time. In the worst case, the tire would be downgraded or scrapped unnecessarily. The closer the tire’s true value to the grading limit, the greater percentage of the time a potentially devastating decision may be made on the disposition of the tire.

Now consider a similar scenario, with the same tire, where the tire is measured by a machine with a comparatively worse repeatability standard deviation specification limit of 0.75 units. The resulting distribution of measurement values is shown in Figure 2. When measuring this same tire through a machine with inferior measurement repeatability, a poor decision will be made a much greater percentage of the time, thus increasing the number of tires over a given year that are unnecessarily downgraded or scrapped. The number of tires unnecessarily scrapped over the course of the year depends on the distribution of tires as produced in relation to the specification limit. Under typical plant production conditions, the percentage of tires that may be misclassified in this way can exceed 1.5% of tire production. Imagine a tire uniformity checking system testing over 1,000,000 tires per year and unnecessarily scrapping over 15,000 tires that year. At

an estimated cost of US$30 per tire, that’s over US$450,000 of unnecessary scrap.

There is a similar but different mode of incorrectly assessing the disposition of the tire that relates to the machine’s measurement performance. Consider now measuring another arbitrarily chosen tire with a known ‘true’ uniformity value of 13 units with a machine that has a repeatability standard deviation specification limit of 0.75 units. The possible distribution of measurement values would look something like that shown in Figure 3. In a similar fashion, one can observe that during a percentage of the time that the tire is measured, the tire would have measured below the specification limit and would have been incorrectly classified as an acceptable tire. This percentage of time is again represented by the black shaded region. Not every one of these tires escaping the production facility will turn into a tire adjustment or customer ill-will; however, a small percentage of them may have that unintended consequence. These incidents are fewer in number but more costly than an incorrectly scrapped tire. There are real tire processing costs and replacement tire costs, not to mention the indirect cost of customer and consumer disappointment. Under typical plant production conditions, the percentage of tires that may be misclassified in this way can exceed 1.0% of tire production. If even 10% of these tires are returned at an adjustment cost of US$150 per tire, the overall costs can exceed US$150,000 a year. This decision-making failure mechanism can also occur in a similar fashion during dynamic balance and geometric runout screening.

There are a vast number of factors that affect the annualized cost for such an investment in tire-testing equipment, including installation costs, acquisition costs, and operating costs. One very important consideration when procuring equipment is the acquisition cost, but it should not be incorrectly used as the

primary driver of the purchasing decision, as it represents a relatively small portion of true lifecycle costs. Operating costs are the dominant expense – one example of which is the costs associated with repeatability-related rejects and adjustments. Measurement performance and the other components of operating costs should be a major consideration in the purchasing decision-making process when procuring tire-testing systems.

The bottom lineWhen confronted with the daunting task of acquiring several lines of final-finish tire testing equipment, the equipment supplier and the specific equipment should both be chosen with the goal of minimizing piece cost in mind. There are suppliers on the market with tools available to assist in the evaluation of the important factors that drive piece cost. Financial models can be used to compare equipment from two or more suppliers, enabling buyers to analyze and compare the important considerations presented in this paper. An example of final output from such a financial model, developed for adding 20,000 tires per day of tire testing capacity, is shown in Table 2.

In this example, Supplier A’s equipment is approximately US$50,000 more expensive per individual system than Supplier B’s equipment and completes a measurement cycle four seconds faster. This results in a requirement of seven Supplier A systems, as opposed to eight Supplier B systems, to provide the 20,000 tires per day of testing capacity. Even though Supplier A’s equipment is initially more expensive per system, selection of this supplier results in an approximately US$600,000 lower initial equipment investment because of the need for fewer systems. In addition, because of the characteristics of Supplier B’s equipment, the annualized process costs are more than 70% higher for Supplier B when compared with Supplier A’s equipment. The bottom line is an almost doubled piece cost if Supplier B’s equipment is chosen. With an in-depth look into the annualized costs along with specific throughput information, it becomes clear which equipment will ensure the greatest delivered value over the life of the machine. Equipment suppliers should be challenged to provide this critical process cost information, and chosen based not on their initial system acquisition price but on their ability to provide the best performance at minimal process cost. tire

Table 2: Example of a financial model output comparing equipment from two suppliers (20,000 tires per day)

SHORT COURSE ON

THE BEHAVIOR OF RUBBER MATERIALS8 & 9 FEBRUARY 2010

Queen Mary, University of London (QMUL) and Tun Abdul Razak Research Centre (TARRC)

LIMITED PLACES – BOOK NOW TO GUARANTEE YOUR PLACE!DOWNLOAD THE FULL PROGRAM NOW!visit: www.tiretechnology-expo.com

MONDAY 8 FEBRUARY

• Polymers, elastomers and rubbers – what are they?• Thermodynamics of elastomers• Physics of rubber elasticity• Finite strain elasticity theory• Inelastic behavior• Strength and fatigue of elastomers

TUESDAY 9 FEBRUARY

• Design with rubber materials• Failure of elastomer products• Friction and abrasion • Vibration and shock control: isolation and damping• Testing rubber materials and components

To be held in conjunction with theTire Technology Expo and Conference

The Behavior of Rubber Materials Short Course will be held concurrently with Tire Technology Expo 2010 in Cologne, Germany on 8 and 9 February 2010

The performance requirements for engineered rubber products

such as tires have continuously increased over the last few

decades and are expected to continue to increase.

The confl icting demands of weight reduction and reduced rolling

resistance, coupled with increases in abrasion resistance and wet

and dry friction performance, make the tire designer’s life diffi cult.

All rubber components have to be designed and manufactured

using robust engineering principles to ensure that they comply

with the expected performance and lifetime requirements.

TIRE TECHNOLOGY EXPO 9, 10, 11 FEBRUARY 2010COLOGNE, GERMANY

“I think it’s the most exciting show. We have all the suppliers for

the tire industry in one place. It’s always special to come here and see the latest developments. It’s

also a great location for us”Joerg Nohl, head of engineering,

Continental Tires

EUROPE’S LEADING INTERNATIONAL TIRE DESIGN AND TIRE MANUFACTURING EXHIBITION AND CONFERENCE www.tiretechnology-expo.com

Tire Technology Expo 2010 Conference

The Tire Technology Expo Conference is Europe’s leading tire design and manufacturing conference, with attendees coming from over 45 countries.

Over three days delegates will have the opportunity to see and hear presentations from speakers picked from the world’s leading experts on tire manufacturing. Presentations will cover the very latest developments in the tire manufacturing industry.

This high-profi le conference, coupled with one of the industry’s largest exhibitions, makes this not only a prestigious event to attend, but also a priority for all involved in the tire-making process.

UKIP Media & Events Ltd l Abinger House l Church Street l Dorking l Surrey l RH4 1DF l UK Tel: +44 1306 743744 Fax: +44 1306 877411 Email: m.fenner@ukintpress.com

Web: www.tiretechnology-expo.com

www.tiretechnology-expo.com

*Program subject to change

9, 10 & 11 February 2010, Köln (Cologne) Messe, Germany

Tire rubber made with naphthenic oil displays good quality and performance yet also adheres to new EU health, safety, and environmental legislation

by Marika Rangstedt, Nynas, Sweden

Getting a gripon future tires

Figure 1: Comparison of processing properties and properties indicative of strength for Nytex 4700, DAE and TDAE.

T he race is on. As of January 1, 2010, tires containing extender oils not compliant with EU legislation on the content of

carcinogenic compounds cannot be placed on the European market. Although the legislation is issued by the European Commission (Directive 2005/69/EC) and applies to the European market, its repercussions will be felt on a global scale, as Europe is the end market for products from many tire manufacturers and polymer producers operating outside the European Union.

Due to their high concentration of polyaromatic hydrocarbons (PAH) and other polycyclic aromatic (PCA) compounds, the main products traditionally used as extender oil in tire rubbers – aromatic extracts (DAE) – do not comply with the new legislation and hence need to be replaced by less harmful alternatives.

Several different oils comply with the legislated health, safety and environmental (HSE) criteria and can be used to serve this purpose, among them treated distillate aromatic extract (TDAE), mild extraction solvate (MES) and high-viscosity naphthenic oils.

TDAE is a DAE that is further refined to meet the HSE criteria. This makes it the alternative that most closely resembles DAE from a chemical perspective. MES is mildly refined paraffinic base oil, sufficiently refined to meet HSE requirements, but not as closely related to DAE as TDAE. It therefore doesn’t display characteristics and performances as similar to DAE as TDAE does.

The third product category that has shown good characteristics when used as extender oils in tire rubbers is highly refined high-viscosity naphthenic oils. As these oils also comply with the HSE criteria, they have been shown to constitute

an excellent choice in the quest to replace DAE as extender oils for the tire industry.

Studies carried out by researchers at Nynas have shown that there are a number of naphthenic oils that can be used to replace aromatic extracts as extender oils in tire rubber compounds. Several studies on various different naphthenic oils covering a relatively wide range of viscosity and other physical properties have been conducted.

Inside a given compound, the performance of some of these naphthenic extender oils, when used in rubber compounds based on styrene-butadiene-rubber (SBR), is similar to that of the TDAE extender oil. Other qualities display performance more akin to MES.

Earlier studies by Nynas and others

have also noted significant variations between different rubber formulations. This has led us to conclude that the effect of the whole formulation, especially regarding the choice of polymer as well as that of the filler, has far greater impact on the resulting mechanical and dynamic mechanical performance than any of the included extender oils.

The most recent studies have focused on Nytex 4700, the newest member in the Nynas tire oil family. This product displays a performance at least in line with that of TDAE in a comparative study using a tire rubber formulation based on an emulsion of polymerized SBR as the polymer and carbon black as the filler. DAE was also included in the study as a reference (see Table 1). The effects of

Abraded volume (DIN)

ML(1+4) 100˚C

Delta torque

Tear strength Tc90

Elongation at break Hardness

Tensile strength

Nytex 4700DAETDAE

80%

90%

100%

110%

120%

98

the extender oil on the compounding and vulcanization processes as well as on the material properties of the resulting rubber were investigated.

The processing properties of the naphthenic oil, TDAE and DAE can be seen in Figure 1. The Mooney viscosities of Nytex 4700 and DAE compounds are very similar. The degree of vulcanization displayed by the delta torque shows the Nytex 4700 compound reaching a little higher crosslink density than the other two compounds. The time required to reach 90% of complete vulcanization is shown to be a little longer for the compounds based on HSE-compliant oil. The hardness is very similar for all of them. The variation of the values can in general be said to be low, which indicates that the performance of the naphthenic oil

included in this study could parallel that of the other oils under investigation.

Furthermore, properties indicative of the mechanical strength of the resulting three materials are brought together in Figure 1. These show the tensile properties of the materials, where the results for the Nytex 4700 compound were almost identical to those of the

DAE-based reference compound. The elongation at break indicates a slightly higher value for the naphthenic oil compared with TDAE and DAE. Finally, the tear strength indicates an advantage for the naphthenic oil compound compared with TDAE, as the result is very similar to that of the DAE compound.

The effect of the extender oils on the glass transition temperature (Tg) of the final rubber compounds can be seen in Figure 2, where the Tg of the rubber is compared side by side with the Tg of the extender oil with which it has been extended. It was clearly shown that the spread in the extender oil Tg – approximately 15°C – is far greater than the variation in the Tg of the resulting compounds, which is less than 4°C.

The comparative study shows that the Nytex 4700 naphthenic oil presents a high-performance alternative, matching at least that of TDAE, in the task of replacing DAE. It is clear that Nytex 4700 offers a good solution, both technically and economically, to manufacturers of tires and producers of tire polymers. The study illustrates that it is possible to combine properties of high quality with compliance with more stringent health, safety and environmental legislation. Tires can get a grip on health, safety and environmental issues as well as on roads. tire

“There are a number of naphthenic oils that can be used to replace aromatic extracts as extender oils in compounds”

Figure 2: Glass transition temperatures of the extender oils alongside those of the resulting rubber compounds.The values of the Tg compound should be Nytex, DAE, TDAE: -41, -39, -40ºC

0

-10

-20

-30

-40

-50

Tg (º

C)

Nytex 4700 DAE TDAE

Tg Oil (ºC) Tg Compound (ºC)

Physical properties of the extender oils used in the study Nytex 4700 TDAE DAE

Density (kg/m3) 940 950 990

Viscosity (cSt) 700 410 1240

Flash point (°C) 220 240 230

Tg (°C) -53 -49 -38

HSE classification OK OK Not OK

99

For most car drivers, tires are simply made of rubber and they have little knowledge about how sophisticated wheels have become.

Since John Boyd Dunlop patented the first pneumatic tire in 1888,1 tires have evolved, but the requirements of the pioneer days for comfort, cushioning, and safety are still valid today.

The development of the car industry from building inconvenient motorized coaches to producing the high-speed, high-technology vehicles equipped with all possible convenience and safety features, means the demands on the interface of vehicle-to-road are increasing fast. During this evolution, the make-up of the tire and the raw materials have been improved continuously.

In terms of the body reinforcement, natural fibers such as cotton have been replaced with semi-synthetic fibers such as rayon, or man-made fibers such as nylon (polyamide, PA) and polyester (polyethylene terephthalate, PET).

With the invention of radial tires, carcass fibers with high strength and limited dimensional change were needed. For PET tire cord the breakthrough came with the introduction of high-modulus low-shrinkage (HMLS) PET fibers. The low raw material cost is an extra benefit encouraging the penetration of PET into the tire market, and the usage of PET fibers is increasing steadily.

Development of pre-treatmentPET tire cord has a low reactive surface, which makes the bonding of PET to rubber difficult. The bonding strength is essential to the performance of a tire, as it ensures that the radial forces are transmitted to the reinforcing fiber. Any weakness would reduce the tire durability during use. With the introduction of the PET cord, it became obvious that a modification of the existing bonding system for PA and rayon2

(resorcinol formaldehyde latex, RFL) was necessary. A lot of effort was put into the

development of a pre-activation for the PET fiber.

The alternativeOver the years, several pre-treatment technologies have been developed to improve the adhesion of PET to rubber. It was found that a combination of an epoxy and an isocyanate gave the best results.3

The isocyanate has to be blocked in order to not react immediately with the water present in the mixture. The blocking of isocyanates can be done with various chemicals. Some of the chemicals in use release hazardous decomposition products such as phenol. To protect operators and the environment against hazardous chemicals, a safer alternative has been found in caprolactam. A caprolactam blocked methylene diphenyl diisocyanate (MDI) has been commercialized for nearly 40 years under the trade name Grilbond® IL-6. The deblocking reaction mechanism of the isocyanate is shown in Figure 1.

Today combinations of epoxy and

Advances are being made in adhesives for the treatment of polyester cord and fabric for rubber applications. Recent product developments include high-solid dispersions and a single-dip solution for non-activated polyester cordby Marcus Bayer & Elke Gebauer, EMS-Griltech, Switzerland

Adhesion promoters:cutting-edge solutions for the tire industry

100

caprolactam blocked MDI are by far the leading pre-treatment systems for PET tire cord fabric in the industry. The form of delivery of the caprolactam blocked MDI has changed since its introduction. In the beginning, it was supplied as powder and had to be ground on-site before use. This production step was not particularly controlled. It resulted in inconsistent dispersion qualities and therefore varying adhesion values.

With the introduction of Grilbond IL-6 50%, EMS-Griltech solved this problem. This product was a ready to use dispersion with specified particle size.It provided consistent quality in dip mixing and consistent particle pick-up for a uniform adhesion. It was applied in a two-step dipping process by the converters.

AdvantagesDevelopments are triggered by the demands of the market: combining best technical performance with the best economical aspects. EMS-Griltech, as the leader in the global market for tire cord adhesive systems, shaped developments considerably.

Products have been optimized in respect of chemical consumption to improve the effectiveness of the used chemicals on the dipped cord. With optimized production technologies, the company was able to reach an effective isocyanate content (functional NCO groups) in practice, corresponding to the theoretical value. The content of

the reactive groups is higher than with alternative materials, reducing the consumption by up to 15% at the converters.4

Not only the content but also the distribution of the groups on the surface of the fibers is essential. Key is having a stable dispersion, combined with a fine particle size. For this reason, the Grilbond IL-6 50% F was introduced, which fulfilled these requirements exactly.

In addition, it offers a reduction of the sedimentation in the dip tank and the dip bath and gum-up on the rolls, resulting in reduced chemical loss. More important often is the reduction

of cleaning efforts and production-line downtime. As a consequence, a massive reduction in dipping costs can be achieved by the on-site cooperation of the technical service from EMS-Griltech and the converters.

High solid dispersionsThe recent development of a high solid dispersion is the next benchmark and there is a future generation of blocked isocyanate-based adhesion promoters. The newly introduced Grilbond IL-6 60% dispersions offer additional improvements in terms of foaming, sedimentation, and gum-up behavior.

Single dip for na PETThe latest development aims at dipping non-activated PET in one step, providing the best adhesion properties and fabric quality in one simple process step. This enables the converter to switch from adhesive-activated PET yarn to non-activated PET yarn, with significant cost savings on the fiber side. For converters using a two-step dipping process on non-activated PET, it denotes a simplification of the process, offering cost savings on the production side. This latest development is commercialized under the trade name Grilbond EasyDip®.

A comparison of the one-step and the two-step dipping process is shown in Figure 2.

Figure 1: The deblocking reaction mechanism of the Grilbond IL-6 isocyanate (Caprolactam blocked MDI)

Figure 2: Comparison of the one-step and the two-step dipping process. The one-step process is much simpler

101

Bonding systemTo optimize the adhesion strength, the bonding mechanism of all incorporated bonding partners must be known very well. To understand the difficult situation to bond a relatively non-polar substrate like PET to rubber, the structure of this polymer must be analyzed carefully. The main chain sections of the PET are non-polar like polyethylene, if compared to nylon (for example, PA 66). This makes it nearly impossible to bond any material to PET without pretreatment. The only dipolar structure in the chain is the carbonyl group, which offers only little polarity. The polar surface tension5 of PET is 10.4mN/m, and of PA 66 is 18.5mN/m.

Another option is to bond the polymer end group with the RFL to the rubber.3

Depending on the polymerization parameters, there are more or less carboxyl and hydroxyl end-groups in the PET available for chemical bonding. A typical amount of carboxyl end groups is 15-25μeq/g. In combination with a typical average molecular weight of Mn = 23,000 Daltons for PET, the number of active end groups is very limited. Only

0.05% of the PET could be chemically bonded to the RFL, which is used as interface to the rubber. This can be done during the spinning process with an epoxy or silane compound, which is applied with the spin finish.

The chemical reaction takes time and leads to a slow down of the manufacturing process. The chemicals used form aerosols in the spin-process, which can be harmful when inhaled and special precautions have to be taken to protect the workers. Both these features are counterproductive to the economics of the pretreated fiber, and in addition, the adhesion strength is limited.4

To improve the adhesion strength, a nylon-like reactive surface must be implemented. The insertion of hydroxyl and amide groups is, in praxis, done by a mixture of a blocked isocyanate and an epoxy, which is the first dip of a double-dip system. The mixture of unblocked isocyanate and epoxy on the cord is reacting in the heat setting phase of the dipping process to a urethane.6

Due to the comparable solubility parameters7 of this urethane and PET,

the reaction product dissolves in the PET, leaving hydroxyl and amide groups of the urethane exposed on the surface of the PET. These active sites form a stable covalent bonding with the RFL.4

First dip-solutionThe most important factor to achieve the nylon-like reactive surface is that the epoxy and the isocyanate are readily available for the reaction. Both chemicals have to be distributed on the surface of the fiber very uniformly in order to get the maximum adhesion levels.

EpoxyThe standard products are poly-functional aliphatic glycidyl ethers with low molecular weight. Essential for these epoxy compounds in this application is the quick and complete water solubility, preventing fabric defects such as white spots. A leading product for the application is Grilbond G1701. It combines the instant water solubility and high reactivity in the dipping process for shortest processing times.

IsocyanateIn contrast to the epoxy, the ‘raw’ isocyanate is solid. To ensure optimal and safe application onto the PET, the blocked isocyanate is delivered as a very fine water-based dispersion. Due to the nature of the blocking agent, and the deblocking temperatures way above the boiling point of water, there is no risk for the converter of free isocyanate.

The fine dispersed quality of Grilbond IL-6, together with the epoxy Grilbond G1701, is best for high-quality cord and fabric and easy, safe, and economic processing. tire

References1. Heissing/Ersoy, ‘Fahrwerkhandbuch’, Vieweg Verlag

(2007)2. Porter, Norman K., ‘Some major variables in RFL

formulations and their effect on dipped cord properties’ Journal of coated Fabrics, Volume 23 (July 1993)

3. Durairaj, Raj B., ‘Resorcinol: Chemistry, Technology and Applications’ Springer (2005)

4. Kurz, Günther, Presentation on ISIFM-meeting (2005)

5. http://www.igb.fraunhofer.de/www/gf/GrenzflMem/gf-physik/dt/GFphys-PolymOberfl.html

6. Hartz, Roy E., ‘Reaction during cure of a blocked isocyanate-epoxy resin adhesive’ Journal of Applied Polymer Science Volume 19 Issue 3, p735-746, 09 (March 2003)

7. Iyengar Y. and Erickson, D. E., Journal of Applied Polymer Science, 11, 2311 (1967)

Figure 3: Illustration of the bonding mechanism of all incorporated bonding partners for optimal adhesion

“Essential for these epoxy compounds in this application is the quick and complete water solubility, preventing fabric defects”

102

JOIN THE WINNERSwith EMS bonding agents

www.emsgriltech.com

for environmentally sound, simplified manufactureand enhanced product quality at reduced cost.

01dB-Metravib

www.dma-instruments.com

DMA

Payne effect

Simultaneous analysiscapabilities

Crack growth test

Viscoelastic properties E*, G*, Tan

Fatigue

Heat build-up

Mullins effectE*SimultaneousSimul

ca

Materials Characterizationat Its Best!

tanCracCra

Viscoelastic proViscoelastic pro

k growth growth

perties E*erties E

Creep/stress relation

Robotized analysis

VulcanizationGlass transition

Robotized analyd analy

VulcanizatiVulcan

ROB 100

DMA 450

EEG

EE

SHORT COURSE ON

THE BEHAVIOR OF RUBBER MATERIALS8 & 9 FEBRUARY 2010

Queen Mary, University of London (QMUL) and Tun Abdul Razak Research Centre (TARRC)

LIMITED PLACES – BOOK NOW TO GUARANTEE YOUR PLACE!DOWNLOAD THE FULL PROGRAM NOW!visit: www.tiretechnology-expo.com

MONDAY 8 FEBRUARY

• Polymers, elastomers and rubbers – what are they?• Thermodynamics of elastomers• Physics of rubber elasticity• Finite strain elasticity theory• Inelastic behavior• Strength and fatigue of elastomers

TUESDAY 9 FEBRUARY

• Design with rubber materials• Failure of elastomer products• Friction and abrasion • Vibration and shock control: isolation and damping• Testing rubber materials and components

To be held in conjunction with theTire Technology Expo and Conference

The Behavior of Rubber Materials Short Course will be held concurrently with Tire Technology Expo 2010 in Cologne, Germany on 8 and 9 February 2010

The performance requirements for engineered rubber products

such as tires have continuously increased over the last few

decades and are expected to continue to increase.

The confl icting demands of weight reduction and reduced rolling

resistance, coupled with increases in abrasion resistance and wet

and dry friction performance, make the tire designer’s life diffi cult.

All rubber components have to be designed and manufactured

using robust engineering principles to ensure that they comply

with the expected performance and lifetime requirements.

TIRE TECHNOLOGY EXPO 9, 10, 11 FEBRUARY 2010COLOGNE, GERMANY

105

Discovering or developing a new fiber or material for industrial designs and manufacturing is a fascinating exercise, but brings

some inconveniences: the process is long, and the R&D efforts require huge budgets. Yet needs are constantly increasing and we are being challenged by new expectations. Therefore, finding innovative ways to combine existing product capabilities is a more promising field, which Kordsa Global believes should be explored.

In the past, fibers used in industrial technical textile fabric reinforcements were limited to polyamide 6.6, polyamide 6, rayon, and polyester, which all have bulk tonnages in production and application. Aramid (AR), polyethylene naphthalate (PEN), lyocell, etc, are relatively new specialty fibers that have appeared in the market lately.

Customers using these industrial yarns are looking for new and superior

properties from industrial yarns and, as a consequence, superior properties in the end product.

Yarn producers are seeking new solutions either by intensifying their

efforts in new yarn production or optimizing the properties of the yarn by chemical or physical modifications. However it is well known that inventing and launching a new product requires huge resources, especially in terms of time and cost. The modification of fibers either

through new spinning techniques or chemical additives is another approach – it requires relatively fewer resources, and yet the developments can be for high-performance applications.

In addition to new fibers or fiber modification, there are opportunities to use a traditional fiber spectrum for the optimization of the final product. Each material within the stress-strain spectrum offers attractive properties belonging to its chemical nature. Polyamide 6.6 is an attractive material in terms of its high energy absorption, making it the fiber fatigue-resistant champion. Polyester has a high range of dimensional stability. Rayon (RY) is a high-modulus fiber and easy to adhere to the rubber. Aramid is a high-modulus and high-strength material. The unique properties of each fiber can be partially transferred to new cord structures by making proper combinations of them, known as hybrid cord structures. It is possible to engineer the mechanical and thermal properties of hybrid cords by making proper combinations of fiber materials compatible with final applications.

Various hybrid structures are possible, such as hybrids of AR with NY and PET, hybrids of PET with PEN and NY, hybrids of RY with nylon (NY), or PET, and so on. In recent years AR-NY hybrids have found more potential in high-performance tire applications.

“As the percentage of aramid increases, the hybrid breaking strength increases, while elongation decreases”

Engineered solutionsfor niche applicationsDeveloping high-tenacity polyamide 6.6 and HMLS polyester industrial yarns for the tire reinforcement and mechanical rubber markets is an ongoing process

by Dr Berrin Yilmaz, Kordsa Global, Turkey

B

Figure 1: Typical stress-strain curve of twisted cords made of Nylon 6.6 (curve A), Aramid (curve B), Hybrid (curve C)

Table 1: Typical examples of hybrid cord in comparison with the reference aramid and the reference NY cords

AR weight

%

NY weight

%

Twist ratio

of AR/NY

Twist ratio of

AR/Cable

Twist ratio

of NY/Cable

Breaking

strength, N

Elongation

at 44 N, %

Elongation

at 66 N, %

Elongation

at break, %

100 - - 1.0 - 690 0.49 0.68 4.44

70 30 1.0 1.0 1.0 586 1.06 1.50 5.40

70 30 1.6 1.0 0.6 543 2.22 2.99 9.06

37 63 1.5 1.0 0.7 337 3.34 4.23 8.76

- 100 - - 1.0 321 7.19 9.74 28.90

C

A

106

Aramid and nylon 6.6 hybrid cords are the combination of high-tenacity and high-modulus synthetic fiber with high-elongation and low-modulus fiber, respectively. The mechanical properties of hybrid cords are maintained and shaped between the gap of aramid and nylon 6.6 cords as reference (see Figure 1). The number of plies of each material, and the single and cable twist levels, have a pronounced effect on cord mechanical properties and performance. The shape and position of the hybrids can be engineered within this gap by proper adjustment of the above-mentioned combinations.

In Figure 1 a typical stress-strain behavior comparison of AR, NY, and hybrid cords is schematized. Curve A is NY cord, and curve B is AR cord, both having the same twist levels. Curve C is a hybrid cord example. The background of the figure shows the microscopic appearance of cords: from left to right is AR cord; hybrid cord with two-ply AR, one-ply NY; and one-ply AR, two-ply NY; and finally NY cord. If the nylon 6.6 content in the hybrid is increased, the hybrid curve shape shifts to the nylon 6.6 direction, whereas if the aramid content is increased in the hybrid composition, the hybrid curve shape shifts to the aramid direction. As mentioned above, since the twist level has a pronounced effect on cord properties, as the twist level increases, the shape of the curve shifts to the nylon 6.6 direction. This shows the flexibility of designing the hybrid cord structures.

Some examples of hybrid cords with different AR and NY composition and twist levels are given in Table 1. This table emphasizes the property variation with different cord structures. It is obvious that the aramid composition and twist level have great influence on mechanical property: as the percentage

of aramid increases, the hybrid breaking strength increases, while the elongation decreases and the modulus increases. However, the contribution of NY component twist level should not be ignored to optimize the hybrid properties.

AR is a tensile member of AR-NY

hybrids. It has a very high breaking strength and tenacity such as 20-22 g/dtex, whereas NY has 8.0-8.2 g/dtex tenacity. The degree of ply twisting of AR in particular has great influence on the breaking strength of hybrid. Examples can be given regarding the twist level versus tenacity change. For instance, as 1100 dtex AR retains its tenacity by approximately 80% at around 300tpm twist level, the same dtex fiber retains only approximately 70% at around 400tpm of its initial tenacity value. Although the breaking strength and tenacity are two of the indispensable material requirements for tire design, their optimization has critical importance for fatigue requirement.

Figure 2 shows a radar diagram of one-to-one ply AR-NY. In this radar diagram, AR and nylon cords have been used as

a reference. K728 has been utilized as NY component of hybrid. NY cord properties are taken as reference; the other cord properties are factorized based on this reference. The breaking strength and the tenacity of the hybrid seem to be close to the middle point of the reference cords,

i.e. higher than NY and lower than AR cords. The higher breaking strength of hybrids compared with NY can offer an opportunity for ply reduction in some tire types. The hybrids retain their breaking strength after curing in the rubber, and do not show any significant difference to that of AR and nylon cord.

Hybrids also enable designing the strain behavior of the cord. Although AR has very low elongation at break, such as approximately 3.5% in fiber form, hybrids of AR may have higher elongation at break values as high as 9% and even higher. The hybrid cord design parameters have a big influence on the elongation behavior of the materials. Regarding the partial load elongations (EASL) of the hybrids, Figure 2 shows that hybrids apparently have higher elongations compared with AR cord. The partial-load elongations can also give a clue about the modulus behavior of the cords. AR is a high-modulus material – it loses its high-level modulus when twisted with NY. Regarding Figure 1, some hybrid cord constructions have high elongations or low modulus values at low load levels similar to NY behavior. The same cord changes its stress-strain behavior by resembling AR cord behavior. The variation of modulus along the stress-strain curve gives some advantage in tire building and tire performance while running.

K728 fiber has around 5% thermal shrinkage at 177ºC with 0.045 g/dtex

Strength

Tenacity

Easl

Shrinkage

Fatigue

Nylon 6.6

Aramid

AR-NY

Adhesion

Weight

Dimensional Stability

Cure in rubber

2.5

2.0

1.5

1.0

0.5

0.0

2

“The variation of modulus along the stress-strain curve gives some advantage in tire building and tire performance”

107

The staff and management of Standards Testing Laboratoriesgratefully acknowledge fellow tire professionals, engineersand academics for recognizing us as the

2009 Tire Industry Supplier of the Year.We were honored for the breakthroughs achieved in our independent,confidential tire and wheel test lab – available to the entire industry – whichmonitors tire rolling resistance by force and torque methods, and by steady-state and step-wise coast-down procedures. What’s more, our machinerydivision provides the latest tire test machinery technology at exceptional value.

From innovations in test laboratories and machine design, to service on industrytesting committees and boards, our staff has set a high standard of excellencefor nearly 40 years. We are gratefulfor the recognition by our peers.

There is no higher honor.TOUGH ON TESTINGSM

SM

SM

1845 Harsh Ave. S.E., P.O. Box 758 � Massillon, Ohio 44648-0758 � USAPhone: 330.833.8548 � 1.800.833.8547 � Fax: 330.833.7902 � www.stllabs.com

� TRANSPORTATION TECHNOLOGIES

� TIRES, WHEELS & VEHICLE COMPONENTS

� TRANSPORTATION TESTING

� TESTING SYSTEMS FOR INDUSTRY

pre-tension. AR has very low thermal shrinkage: maximum 0.2% shrinkage under the same thermal testing conditions. After combining two yarns by twisting and then dipping, the resultant hybrid has shrinkage values around maximum 2.5%. This intermediate shrinkage level gives better tire uniformity compared with NY cords.

The term ‘dimensional stability’ is derived from shrinkage and EASL values of hybrids. The lower the dimensional stable value shows, the better the dimensional stability. AR is certainly a dimensional stable material. Figure 2 shows how the dimensional stability of the cord is improved by incorporating AR in the cord structure compared with NY cord. The dimensional stability of hybrid gives advantages to the carcass and cap ply of the tire for better tire growth and flat spotting performance.

It is well known that NY adheres easily to RFL with conventional techniques. However, AR does not give sufficient adhesion with the same type of dipping, due to its chemical structure. Therefore two-step dipping is usually applied for

good adhesion. Hybrids generally require the same method of dipping to achieve at least the same adhesion performance of AR. Figure 2 shows better adhesion of hybrids compared with AR cord.

AR-NY hybrids offer improved fatigue resistance compared with reference AR cords. An increase of nylon 6.6 content in the hybrid structure positively affects the fatigue resistance of the hybrid cord. An increase of AR cord twist level also positively affects the fatigue performance of the hybrid. In order to have improved fatigue performance of hybrid cords, optimization of AR and NY single and cable twist levels must be undertaken. A one-to-one ply hybrid example shows the improved fatigue resistance of hybrid cord compared with AR cord in Figure 2.

Considering the one-to-one ply hybrid cord example in Figure 2, the hybrid is a lighter material compared with AR cords. Hybrids can also supply weight reduction compared with metallic cords. Usage of lighter materials decreases tire weight, which has an influence on rolling resistance and fuel consumption.

AR is a costly material. Hybrids may

offer a cost advantage compared with AR cord when considering the whole cycle of the tire.

Hybrids can be utilized both on cap ply and carcass of tires, and each component requires different material properties. High fatigue resistances, adjustable modulus at low and high deformations, dimensional stability, thermal shrinkage, and adhesion are some examples of the required parameters. Hybrids can give design flexibility of modifying properties to some extent. Tires with AR-NY hybrid cord constructions are available on the market, and they may still have a potential use for new tires and non-tire applications.

Designing appropriate combinations of hybrid plies for optimized cord structures, twisting and weaving technologies and practices of hybrids, treatment in dipping lines to achieve desired adhesion levels, and setting the tensile and thermal properties are all critical steps in hybrid cord production. In this respect, Kordsa has built up design and manufacturing knowledge of hybrid cords that forms the company’s leadership and success. tire

Section 2Production, quality, and recycling

109

110

Curing rubber compounds efficiently and cost-effectively

Rubber compounders and tire manufacturers could benefit from reducing or eliminating harmful and expensive chemicals from the curing process

by Ali Ansarifar, Department of Materials, Loughborough University, UK; Co-authors: Li Wang, Saeed Ostad Movahed and Farhan Saeed, Department of Materials, Loughborough University, UK; and K Ansar Yasin, Department of Chemistry, The University of Azad Jammu and Kashmir, Muzaffarabad, Azad, Kashmir, Pakistan and S. Hameed, Department of Chemistry, Quaid-i-Azam University, Islamabad, Pakistan

T he implementation of the European Union Regulation REACH, and a need to use materials and resources more

efficiently to reduce waste and pollution, have imposed a considerable burden on rubber compounders and tire manufacturers.

The methods which are widely used to cure silica-filled rubber compounds with sulfur for green tire applications, do not take into account the exact requirements for the curing chemicals and are therefore inefficient, too expensive, and harmful to health and to the environment.

Rubber compounds used to manufacture tires contain several ingredients. They include fillers, curing agents, antidegradants and processing aids1. Traditionally, fillers and curing chemicals have performed two distinct functions in rubber compounds.

Reinforcing fillers for example colloidal carbon blacks and synthetic silicas with large surface areas ranging from 150 to 400m2/g, are very effective in improving the mechanical properties of rubber such as hardness, tensile modulus and abrasion resistance2.

Table 1: Recipe and ODR test results for the SBR, BR, NR and IR rubber compounds Ingredients Compound no.

1 2 3 4

SBR 100 - - -

High cis BR - 100 - -

NR - - 100 -

IR - - - 100

Silanised silica 60 60 60 60

TBBS 3 7.5 6 7

ZnO 0.5 0 0.3 1

Santoflex 13 1 1 1 1

Processing oil 5 0 0 0

Cure system 3.5 7.5 6.3 8

ODR results

Minimum torque (dN m) 18 37.5 26 26

Maximum torque (dN m) 56 129 107 137

torque (dN m) 38 91.5 81 111

Scorch time, ts2 (min) 16 8 9 8

Optimum cure time, t95 (min) 80 83 27 34

Cure rate index (min-1) 1.6 1.3 5.6 3.9

111

Curing agents, for instance sulfur, in combination with accelerators and activators, produce stable chemical crosslinks between the rubber chains in unsaturated elastomers. The cure systems in tire-tread rubber compounds often consist of primary and secondary accelerators, primary and secondary activators and elemental sulfur, which add up to 11 parts per hundred rubber by weight (phr)3. Reducing the use of these harmful and expensive chemicals, or eliminating them altogether, will be greatly beneficial to rubber compounders and tire manufacturers.

Styrene-butadiene rubber (SBR) (23.5 wt % styrene, Intol 1712, Polimeri Europa UK Ltd, Hythe, UK), high cis polybutadiene rubber (BR) (Buna CB 24, Bayer; not oil-extended, Newbury, UK) with a minimum 96 wt % cis 1-4 content, natural rubber (NR) (standard Malaysian natural rubber grade L) with 98 wt % cis 1-4 content, and synthetic polyisoprene (IR) with a minimum 96 wt % cis-1-4 content (Kraton IR-307, Kraton Polymers) were used in this study.

These elastomers are used in the manufacture of passenger car and truck tires. The reinforcing filler was Coupsil 8113 (Evonik Industries AG of Germany),

which is a precipitated amorphous white silica-type Ultrasil VN3 – the surfaces of which had been pre-treated with bis (3-triethoxysilylpropyl)tetrasulphane (TESPT)4. It has 11.3% by weight TESPT, 2.5% by weight sulfur (included in TESPT), 175m2/g surface area (measured by N2 adsorption), and a 20-54 nm particle size. This filler is known as a ‘crosslinking filler’.

The chemical bonding between the tetrasulphane groups of TESPT and the rubbers was maximized by adding N-t-butyl-2-benzothiazole sulphenamide (Santocure TBBS; a safe-processing delayed-action accelerator), ZnO (activator) and stearic acid (activator).

A heavy paraffinic distillate solvent extract aromatic-processing oil (Enerflex 74) was added to the SBR compound to reduce its viscosity. To protect the rubbers against environmental ageing, N-(1,3-dimethylbutyl)-N´-phenyl-p-phenylenediamine (Santoflex 13, antidegradant) was also included in the rubber compounds.

Four rubber compounds each having 60phr of Coupsil 8113 were prepared in a Haake Rheocord 90, a small-size laboratory mixer with counter-rotating Banbury rotors. The rotors and mixing chamber were maintained at 23°C for making the SBR, BR and IR compounds and 48°C for making the NR compound. The rotor speed was 45rpm. The volume of the mixing chamber was 78cm3, and it was 55% full during mixing. The filler particles were dispersed well in the rubbers by increasing the mixing time to 17 minutes. Full details of the mixing

Torq

ue (d

N m

)

torque = Tmax - Tmin

Tmin

Tmax

Time (min)

e

d

c

b

a

Torq

ue (d

N m

)

Time (min)

0 50 100

120

60

0

Figure 1: Typical torque-versus-time traces by

ODR for the filled IR rubber with (a) 0.4phr TBBS,

(b) 3phr TBBS, (c) 5phr TBBS, (d) 7phr TBBS, (e)

9phr TBBS

torq

ue (d

Nm)

TBBS loading (phr)

0 2 4 6 8 10 12

BR/60 phr silica• NR/60 phr silica

SBR/60 phr silica

100

80

60

40

20

0

Figure 2: torque versus TBBS loading for the filled SBR, BR and NR rubbers

112

conditions of these compounds were reported previously5-7.

The rubber compounds were subsequently tested by an oscillating disc rheometer curemeter (ODR) at 140°C to produce cure traces from which torque was calculated8. torque is the difference between the maximum and minimum torque values on the cure traces of the rubbers and is an indication of crosslink-density changes (Figure 1). torque was subsequently plotted against the loading of TBBS, ZnO and stearic acid.

Optimizing the chemical bonding between the filler and rubber via the

tetrasulphane groups of TESPT is a very efficient method for crosslinking and reinforcing rubber compounds.

Figure 2 shows torque versus TBBS loading for the filled NR, BR and SBR rubbers. For the SBR, torque increased from 8 to 22dN m when 3phr TBBS was added. Thereafter, torque increased at a much slower rate to 26dN m as the loading of TBBS was raised to 9phr. It seemed that 3phr TBBS was sufficient to start the reaction between the tetrasulphane groups of TESPT and the rubber. Similarly, for the BR, torque increased to 87dN m when the loading

of TBBS was raised to 7.5phr. Further increases in the amount of TBBS had little effect on torque, which remained at about 90 dN m. For the NR, torque increased sharply from 9 to 55 dN m as the loading of TBBS was raised to 6phr, and it continued rising at a much slower rate to about 61 dN m when the loading of TBBS reached 10 phr. For the IR,

torque increased to about 60 dN m as the loading of TBBS reached 7phr and remained at this level thereafter, when the amount of TBBS was raised to 10phr. The TBBS requirement for this rubber was slightly higher than that of the NR (Figure 3).

The filled SBR, NR, BR and IR rubbers needed 3, 6, 7.5 and 7phr TBBS, respectively, to fully react the tetrasulphane groups of TESPT with the rubber chains.

ZnO was incorporated into the rubbers to increase the efficiency of TBBS. For the SBR with 3phr TBBS, torque increased from 22 to 57 dN m as the loading of ZnO was raised to 0.5phr. The rate of increase of torque slowed down considerably with torque rising from 57 to 64 dN m as the amount of ZnO reached 2.5phr. For the BR with 7.5phr TBBS, torque increased to 131 dN m when 0.5phr ZnO was added and showed no further improvement thereafter when an extra 1phr ZnO was incorporated into the rubber. For the NR with 6phr TBBS,

torque increased sharply to 91 dN m when 0.3phr ZnO was included, and it continued rising at a much slower rate to 112 dN m when the loading of ZnO reached 2phr (Figure 4). The IR with 7phr TBBS needed 1phr ZnO to increase the efficiency of cure. For this rubber, torque increased from 60 to 109 dN m as the loading of ZnO was raised to 1phr.

torque subsequently rose to 125 dN m when the amount of ZnO reached 2.0phr. It was interesting that the IR, which is the synthetic analog of NR and is chemically and structurally similar to it, needed an extra 1phr TBBS and 0.7phr ZnO to fully cure compared with the NR (Figure 5).

The filled SBR rubber with 3phr TBBS, filled NR rubber with 6phr TBBS, filled BR rubber with 7.5phr TBBS, and filled IR with 7phr TBBS required 0.5, 0.3, 0.5, and 1phr ZnO, respectively, to optimize the chemical bonding between the rubber and filler and produce a fully efficient cure.

Stearic acid is a fatty acid that is added

Table 2: Mechanical properties of the SBR, BR, NR and IR rubber vulcanisates

Compound no.

1 2 3 4

Hardness (Shore A) 62.5 72 75 80

Tensile strength (MPa) 26 17 37 17

Elongation at break (%) 1308 606 837 404

Stored energy density at Break (mJ/m3)

140 49 137 33

T (kJ/m2) 75 30 58 17

Range of values 71-89 10-103 46-95 10-23

Relative volume loss in the abrasion tests, v (mm3/mg)

126 15.5 - -

Modulus (MPa) at 100% strain amplitude

0.73 2.2 2.2 3.0

Modulus (MPa) at 200% strain amplitude

0.93 2.2 3.2 3.7

Modulus (MPa) at 300% strain amplitude

1.17 2.6 4.2 4.3

Cyclic fatigue life (kc) 777->1000 40->1000 34-99 16-39

Figure 3: torque versus TBBS loading for the filled NR and IR rubbers

IR/60 phr silica• NR/60 phr silica

torq

ue (d

Nm)

0 2 4 6 8 10 12

80

60

40

20

0

TBBS loading (phr)

113

to improve the solubility of ZnO in rubber. The loading of stearic acid in the rubbers with TBBS and ZnO was increased to 2.5phr to measure the amount needed to optimize the efficiency of TBBS and cure. For the SBR with 3phr TBBS and 0.5phr ZnO, torque decreased from 57 to 46 dN m as the loading of stearic acid reached 2.5phr. For the BR with 7.5phr TBBS and 0.5phr ZnO, initially torque increased from 131 to 134 dN m with up to 1phr stearic acid, and then it dropped to approximately 113 dN m as the loading of stearic acid was increased progressively to 2.5 phr. However, the BR was fully cured with 7.5 phr TBBS and required no ZnO. For the NR with 6 phr TBBS and 0.3 phr ZnO,

torque decreased from 91 to about 83 dN m when up to 2 phr stearic acid was added (Figure 6). However, for the IR with 7phr TBBS and 1 phr ZnO, torque increased from 109 to 121 dN m when 1phr stearic acid was incorporated in the rubber and then, it returned to its original value when the loading of stearic acid was increased to 2phr. Notably, the torque values measured for the NR were noticeably lower than those calculated for the IR (Figure 7).

Stearic acid had no beneficial effect on the chemical bonding between the filler and rubber and in fact, it was detrimental to torque and therefore to the crosslink density and cure of the rubbers.

After the TBBS, ZnO and stearic acid requirements were measured for the rubbers, four formulations were produced

(Table 1). The rubber compounds were cured at 140°C in a compression mold to produce sheets approximately 2.4mm thick and cylindrical samples 15.6mm in diameter and 9.5mm in height for further work. The hardness, tensile strength, elongation at break, stored energy at break, tensile modulus, tearing energy, and abrasion resistance of the rubber vulcanisates were determined using the procedures described in the British Standards 9039-12 (Table 2). The cyclic fatigue life of the rubbers (number of cycles recorded when the samples fractured) was measured in uniaxial tension with dumbbell test pieces at a constant maximum deflection of 100% and a test frequency of 1.42Hz13.

The test temperature was 22°C and the strain on each test piece was relaxed to zero at the end of each cycle. For each rubber, eight test pieces were cycled to failure and tests were stopped whenever the fatigue life exceeded 1,000kc.

Good mechanical properties are essential for the performance, durability and life of tire compounds in service. As the results in Table 2 show, the rubber vulcanisates possessed good properties in spite of substantially reducing the curing

Table 3: Sulfur (in TESPT), TBBS and ZnO requirements based on the actual weights of the chemicals measured in the filled SBR, BR, NR and IR rubber compounds (see Table 1)Compound S/TBBS (g/g) ratio S/TBBS/ZnO (g/g/g) ratio

SBR - - 0.45/0.9/0.15 3/6/1

BR 0.45/2.25 1/5 - -

NR - - 0.45/1.8/0.09 5/20/1

IR - - 0.45/2.1/0.3 1.5/7/1

Filled BR/7.5 phr TBBS• Filled NR/6phr TBBS

Filled SBR/3 phr TBBS

torq

ue (d

Nm)

0 0.5 1 1.5 2 2.5 3

160

120

80

40

0

ZnO loading (phr)

Figure 4: torque versus zinc oxide loading for the filled SBR with

3phr TBBS, filled BR with 7.5phr TBBS and filled NR with 6phr TBBS

Figure 5: torque versus ZnO loading for the filled NR with 6 phr TBBS and filled IR with 7phr TBBS

Filled IR/7 phr TBBS• Filled NR/6 phr TBBS

torq

ue (d

Nm)

160

120

80

40

0

Zn0 loading (phr)

0 0.5 1 1.5 2 2.5

114

chemicals. The hardness was somewhere between 62.5 and 80 Shore A, and the tensile modulus between 0.73-4.3MPa at strain amplitudes from 100-300%.

The properties related to fracture were also impressive. For example, the tensile strength was 17-37MPa, elongation at break 404-1308% and stored energy density at break 33-140mJ/m3.

The tearing energy, which is of a major importance to the impact behavior of tires in service, was in the region of 17-75kJ/m2. It is worth noting that the relative volume loss in the abrasion tests, was very low for the BR and below the average for the SBR.

The shortest and longest cyclic fatigue lives were recorded for the IR and SBR, respectively. For the SBR and BR, seven and four test pieces, respectively, lasted longer than 1,000kc. The IR had the shortest fatigue life, followed by the NR.

The mechanical properties of the rubber vulcanisates were impressive in spite of reducing the curing chemicals substantially in the rubbers.

As mentioned earlier, Coupsil 8113 has 2.5% by weight sulfur included in TESPT. The actual weights of sulfur, TBBS and ZnO in the rubber compounds are calculated and summarized in Table 3. As the results show, at a given loading of sulfur, to optimize the chemical bonding between the rubber and filler and to achieve the most efficient cure, the requirements for TBBS and ZnO were totally different. This was because of the dissimilar composition of the rubbers.

The formation of stable covalent filler-TESPT bonds is essential in rubber

reinforcement14, and a large improvement in the rubber properties was seen, as indicated in Table 2. Interestingly, stearic acid was not at all essential to cure the rubbers.

The methods used at present to cure silica-filled tire compounds with sulfur, TBBS and ZnO do not take into account these different requirements and hence cannot be efficient. It is likely that a similar problem exists with the use of these curing chemicals in rubber compounds filled with carbon black.

In conclusion, a proper use of sulfur, TBBS and ZnO for curing SBR, BR, NR and IR rubber compounds for tire applications must take into account the composition of rubber at a given loading of sulfur. This will increase the efficiency

of sulfur-curing without compromising the good mechanical properties of the rubber vulcanisates. Other benefits include improvement in health and safety, less damage to the environment and cheaper tires. tire

AcknowledgementWe are grateful to Evonik Industries AG of Germany for supplying the silica filler.

References1) The Natural Rubber Formulary and Properties

Index (1984) EUR053, Archives, Tun Abdul Razak Research Centre, Malaysian Rubber Producers’ Research Association. Brickendonbury, Hertford, UK.

2) Warrick, E., Pierce, R., Polmanteer, E., Saam, J. C. (1979) Rubber Chemical Technology, 52, 437.

3) Byers, J. T. (2002) Rubber Chemical Technology, 75, 527.

4) Wolff, S., Görl, U., Wang, M. J., Wolff, (1994) European Rubber Journal, 16, 16.

5) Ansarifar, A. Wang, L. Ellis, R. J. Kirtley, S. P., Riyazuddin, N. J. (2007) Journal of Applied Polymer Science, 105, 322.

6) Ansarifar, A., Wang, L., Ellis, R. J., Haile-Meskel. Y. J. (2007 ) Journal of Applied Polymer Science, 106, 1135.

7) Ostad Movahed, S., Ansar Yasin, K., Ansarifar, A., Song, M., Hameed, S. J. (2008) Journal of Applied Polymer Science, 109, 869.

8) British Standards Institution (1977) BS 1673, Part 10, London, UK.

9) British Standards Institution (1995) BS 903, Part A26, London, UK.

10) British Standards Institution (1995) BS 903, Part A2, London, UK.

11) British Standards Institution (1995) BS 903, Part A3, London, UK.

12) British Standards Institution (1995) BS 903, Part A9: Method A.1, London, UK.

13) British Standards Institution (1986) BS 903, Part A51, London, UK.

14) Wolff, S. (1996) Rubber Chemical Technology, 69, 325.

Stearic acid loading (phr)

Filled IR/7 phr TBBS/1 phr Zn0• Filled NR/6 phr TBBS/0.3 phr Zn0

0 0.5 1 1.5 2 2.5

torq

ue (d

Nm)

140

120

100

80

60

40

Figure 7: torque versus stearic acid loading for the filled NR with 6phr TBBS and 0.3phr ZnO,

and filled IR with 7phr TBBS and 1phr ZnO

Figure 6: torque versus stearic acid loading for the filled SBR with 3phr TBBS and 0.5phr ZnO,

filled BR with 7.5phr TBBS and 0.5phr ZnO, and filled NR with 6phr TBBS and 0.3phr ZnO160

120

80

40

0

torq

ue (d

Nm)

Stearic acid loading (phr)0 0.5 1 1.5 2 2.5 3

Filled BR/7.5phr TBBS/ 0.5phr ZnO • Filled NR/6phr TBBS/0.3phr ZnO

Filled SBR/3phr TBBS/0.5phr ZnO

Proud of the technology we put

into every tire

With the combinedresources and experienceof Bartell, RMS and

Steelastic, the Pettibone TireEquipment Group continues to be at the forefront of innovativenew tire technology. New technolo-gy that continues to automate production, while saving valuableresources, labor and materials.And new technology that allows us to continue engineering evengreater quality and reliability intoour systems.

Proud of the technology we put into every tire

Bartell Machinery LTDTelford, England:44-1-952-201291Fax 44-1-952-201083sales@bartellmachinery.co.uk

Bartell Machinery Systems LLCRome NY 13440315-336-7600 Fax 315-336-0947sales@bartellmachinery.comwww.bartellmachinery.com

RMS Equipment CompanyKitchener, Ontario, Canada N2G 4J4519-749-4634Fax 519-749-4613rmssales@rmsequip.comwww.rms-ca.com

The Steelastic Company Akron OH 44310330-633-0505Fax 330-633-0527sales@steelastic.comwww.steelastic.com

The Pettibone Tire Equipment GroupAsia Representative Office4-21, Shin-Nishikata,Kuwana Mie, Japan 511-0863tisobe@steelastic.comwww.steelastic.com

RMS is well known as a designerand manufacturer of complex rubber extrusion equipment androller die heads, as well as first andsecond stage radial tire buildingmachines. In addition, Steelasticand Bartell offer one-stop shoppingfor many types of tire component-making machinery systems, includ-ing bead lines, automatic apex systems, automatic bias and plycutter systems, gum edgers, beadevaluation stations and cap stripproduction systems.

Look to the Pettibone TireEquipment Group to offer unparalleled technology and service to the tire industry. Withthousands of systems installed and operational world wide...Bartell, RMS and Steelastic togeth-er offer an even greater commit-ment to the development of quali-ty tire component machinery, aswell as tire building machinery,at substantial cost savings to our customers.

116

Sealed flexible rubber bags are inflated inside uncured tires during the vulcanization process. When the mold closes, air and

vapor or hot water circulate automatically into the bladder. Under heat and pressure conditions, the expandable bladder inflates into the inner surface of tire. After the green tire is shaped against the outer mold surface, the bladder deflates.

Curing bladders are one of the most severe applications for rubber in terms of heat and flexing resistance. The bladder composition, as usual, may contain butyl rubber, polychloroprene rubber, reinforcing carbon black, oil, resin and other usual ingredients.

Since bladder compounds have a resin cure and are used in high temperatures, increasing thermal stability is one of the most important subjects for them.

The resin-cure system does not use any sulfur and is used when high thermal resistance and low pressure set is important. The curing reaction in this system is slow, so it may need halogenated activators. The advantages of this curing system is that it can lead to obtaining ozone resistant compounds and

eliminating the tendency to scorch under high temperatures.

Heat produced during the service causes decomposition of the crosslinks’ network in the compound and reduces modulus. In Butyl compounds 1,3-bis (citra conimido methyl benzene) can fix modulus and keep crosslink density by forming stable carbon-carbon links.

In addition, it can control heat build-up, increase thermal resistance, resistance in dynamic flexes, reversion resistance, flex-cracking resistance and flexibility. It also improves compression set, and

the tensile and vulcanized physical properties.

Experimental workThe bladders discussed here contain 1,3-bis (citra conimido methyl benzene), the amount depending on the type of polymer and the cure system used. The formulation is shown in Table 1.

The 1,3-bis (citra conimido methyl benzene) is thermally stable at normal processing temperature and has a melting point below 90ºC. Thus its solubility and dispersion will not be a problem. The

Increased production rate through improved bladder compoundIncreasing the thermal stability of bladder compounds, while preserving other processing properties, can improve cracking resistance and flexibility, and reduce bladder failures

by Maryam Mokhtarimehr and A.G. Moteshareie, laboratory manager and compounding manager, respectively, of Dena Tire, Iran

Table 1: Formulation of compounds

Compound Ingredients Control Trial

Butyl rubber 100 100

Neoprene 5 5

Carbon black 50 50

Oil 7.5 7.5

Fatty acid 1.5 1.5

Zinc oxide 5 5

Resin 10 10

1,3-bis (citra conimido methyl benzene)

- 0.75

Figure 1: The average of cure cycles by Control and Trial bladders in size 16A

400

350

300

250

200

150

100

50

0deformed

319

197236

360337

212

0 0

blown out bared creased

Type of failure

Figure 2: The average of cure cycles by Control and Trial bladders in size 16B

deformed blown out bared creased

Type of failure

350

300

250

200

150

100

50

0

159

224

290

233

286

317

222

268

117

solubility is easy and the dispersion occurs well into an internal mixer. It resists decomposition up to 240ºC.

To obtain better dispersion in a multiple-stage mixing, it should be added at the masterbatch stage. The dispersion may not be optimized if the mixing temperature is below 90ºC. However, to ensure easier and more effective dispersion, it should be added at the first (non-productive) mixing stage.

In this research, bladders were tested in truck sizes 24A and 24RIB and light-truck sizes 16A and 16B.

Discussion and resultsSubject to constant curing conditions, the average cure cycles of four sizes of trial

bladders – 16A, 24A, 24RIB, and 16B – are more than in the control. Failure through blow-outs is reduced in these bladders (Figures 1-4).

Most of the trial bladders are scrapped because of their creased and deformed shape. These two failures are slightly better than the other scrapped bladder failures. In 16B, these two failures are observed with less than the average cure cycle (Figure 5).

ConclusionThe average cure cycles undergone by the trial bladders is more than that of the control bladder (Figure 6). These bladders have less fracturing and longer life. Most of the bladders suffered

deformation and abrasion and finally appear deformed (Figure 7). It seems the quality of compound in these bladders is improved and they have longer life and better efficiency compared to the control bladder. Increased life duration is observed in all the sizes (Figure 8). tire

deformed blown out bared

Type of failure

7194

Figure 3: Average of cure cycles by Control and Trial bladders in size 24RIB

250

200

150

100

50

0

187

239

88

0

Figure 4: Average of cure cycles by Control and Trial bladders in size 24A

244

159

255 246300

250

200

150

100

50

0 deformed blown out

Type of failure

Figure 6: Total cure cycles with all failures by Control and Trial bladders

350

300

250

200

150

100

50

016A 16B 24A 24RIB

Bladder size

245284

98

214

312 324

253

111

16A

Figure 8: Increase in bladder life compared to Control

16B 24A 24RIBBladder size

25

20

15

10

5

0

21.5

21.5

Figure 7: The average of scrapped Control and Trial bladders in all sizes

deformed blown out baredType of failure

400

350

300

250

200

150

100

50

0

369

312

309

6578

3 15

creased

11.5

15

deformed blown out baredcreased

Type of failure

350

300

250

200

150

100

50

0

300334

166153

316

268

175 158

Figure 5: Total number of cure cycles by Control and Trial bladders in all sizes

References1) Patitsas, P. Self-releasing curing bladders,

December 2001.2) Rubber Industries Engineering and Research Co,

Principle of Rubber Compounding and Technology Samar Publisher.

3) New material general information and guidelines for optimal use, Flexsys, BV Technical Bulletin, 1996.

4) Anti riversion agent, an overview, Flexsys, BV 1996.

118

Rubber devulcanization: a successful project

DevulCO2 shows the way to successful rubber devulcanization, and its commercial application could revolutionize the use of the huge amount of waste tire rubber now that it can no longer be disposed of in landfill sites

by Andrew James, Smithers Rapra Technology, UK

DevulCO2, the Technology Strategy Board (TSB) funded research project, has achieved its aims of developing a continuous

and effective devulcanization system for the production of devulcanized rubber from waste tires. The project concluded on April 30, 2009, having successfully achieved all its goals and objectives.

The DevulCO2 project was instigated to provide an answer to the EU ban on disposing of waste tire rubber in landfill sites that came into force on July 1, 2006. The project results have now made the 300,000 tons of waste rubber destined for landfill sites in the UK a viable commodity. With a further two million tons of waste tires available within the rest of the EU, the technology developed from the project will be in demand.

The steady increase in commodity prices due to increasing global demand has ensured major interest from compounders in DevulCO2 and finding a technically competent, economically viable, and environmentally friendly solution to the waste problem. It is clear that enabling manufacturers to reuse this large quantity of rubber has great potential benefits, both to them and the environment, but prior to the DevulCO2 project, efforts to substitute recycled rubber for virgin rubber had been largely unsuccessful. Past failures were due to the poor mechanical properties achieved when large volumes were used to produce final products, processibility problems, or economic considerations.

The continuing existence of this potentially exciting and lucrative market opportunity led to a number of different devulcanization techniques being developed aiming to regenerate rubber for the manufacture of high-quality rubber products. DevulCO2 has developed the technology to exploit this market.

The DevulCO2 project ran from November 2006 and was funded by the TSB within its Design and Manufacture of Sustainable Products program. Projects within this program are designed to develop eco-friendly and sustainable processes that will benefit SMEs based

in the UK. More than US$689,000 of direct financial assistance was provided by the TSB, with the project partners matching the amount between them. In addition to Smithers Rapra, the consortium partners were PJH Partnership, Martins Rubber Company, BD Technical Polymer, J. Allcock & Sons, and Charles Lawrence International.

The project satisfied its two key aims. The first aim was to develop a novel, effective and commercially competitive devulcanization system for the production of devulcanized rubber from waste tire crumb. The second aim was to evaluate this devulcanized rubber as a replacement for virgin rubber in the manufacture of a variety of high-added-value rubber products within the general rubber goods sector.

The project has been extremely successful in achieving both of these

“The first aim was to develop a novel, effective and commercially competitive devulcanization system”

The DevulCO2 project is a positive step forward for new products made from devulcanized rubber from waste tires

119

goals, and the new technology developed has been proven within a commercial environment. The devulcanized rubber can be easily processed via a range of standard rubber processing operations, such as extrusion and calendaring, into good quality blanks for re-vulcanizing. The re-vulcanizing step can then be achieved using techniques such as injection molding, compression molding, and transfer molding. This flexibility enables a wide range of high-added-value, high-quality products to be produced.

The physical testing results obtained impressive results on products manufactured from 100% of the devulcanized rubber showing tensile-strength values in excess of 18MPa combined with an elongation-at-break of 350%. The novel process is capable of recovering up to an amazing 80% of the physical properties of original virgin compounds.

The proven process utilizes standard processing equipment with minimal modification and will be easily scaled up to commercial levels of production. A key benefit of this process breakthrough is that it does not present any additional health and safety or environmental concerns to the rubber industry.

The project has also demonstrated that it is possible to blend the devulcanized rubber with either virgin rubber or virgin rubber compounds in a rubber intermix to increase the range of products that can be manufactured. The economic and commercial assessments carried out on the process have shown it to be

a cost-effective solution to the waste tire problem and provide excellent business opportunities.

The technology is now ready to be scaled up to commercial levels of production, it does not present any additional health and safety or environmental concerns, and the process is extremely cost-effective. The consortium is now able to take the products to market.

The success of DevulCO2 has secured further funding from the TSB for the ReMould project, which will transfer the devulcanization technology into other high-tonnage waste rubber products such as EPDM weather strip, further refine the processibility to enable profile extrusions to be produced, and extend the range of final products that can be manufactured such as the retreading of truck tires. This project will commence in the summer of 2009, will run for two years, and will include two new partners: Kingpin Tyres and London Metropolitan University. tire

“It is possible to blend the devulcanized rubber with either virgin rubber or virgin rubber compounds in a rubber intermix”

High-quality injection, compression or transfer-molded products can be created using the re-vulcanizing process

The technology is ready for commercial production levels

A number of technological innovations – including a quick-cleaning gear pump and a co-extrusion head – are creating benefits for the tire industry

by Dr Tim Pohl, Troester GmbH & Co KG, Hannover, Germany

Tire industry innovationsincluding co-extrusion heads and Ethernet line control

Figure 1: Co-extrusion Y-head

120

T roester is looking back on a successful year and heading optimistically into the future. At the same time the company has

implemented technological innovations that will benefit the tire industry.

Co-extrusion head Of particular note is the development of a new co-extrusion head with a special clamping system (Figure 1). The concept has already proved successful in practice for the manufacture of sidewall and apex profiles. The compact design allows space-saving integration into flexible manufacturing lines with short profile running times.

The movable head parts (upper and lower part) are connected by means of lateral Y-joint rods, and are pulled toward the fixed-head middle part by only one hydraulic cylinder each. The clamping force of the lateral tie rods acts vertically on the sealing faces of the flow channels.

The flow channels are matched to customers’ individual requirements and their products with the use of finite-element flow simulation. The streaming

history for the rubber in the head is analyzed and adopted in order to achieve a constant swelling behavior at the head outlet, and a straight flow. The general goal is a material-independent flow channel construction in order to allow an easier design of the subsequent flow segments.

In addition to an appropriate forming of the extrudate, the rubber compound guiding parts should be designed with the lowest possible pressure consumption. This helps reduce the compound temperature at the outlet and increase the extrusion speed.

The features as a whole meet the demand of the tire industry for flexible manufacturing equipment to achieve high productivity of the line and an increased variety of products with the best price-performance ratio.

Quick-cleaning gear pumpThe use of a gear pump connected to the extruder is state of the art for high-pressure applications such as straining or wire coating. The gear pump is fed by the extruder and allows a volumetric conveying characteristic of the compound. The product quality is enhanced by the stable discharge and by low tolerance deviations. Using a gear pump enables stable process conditions to be achieved quickly, which is advantageous when using such a system in cap strip, bead, and wire coating lines.

However a gear pump is not self-cleaning. In case of compound changes or longer production stops, the system is dismounted and disassembled. Therefore Troester developed a quick-cleaning gear pump for reducing the cleaning cycle.

Without using any tools, all parts are

“The flow channels are matched to customers’ individual requirements and their products with use of FE simulation”

Troester’s headquarters in Hannover, Germany

121

individually removed from the pump and can be quickly cleaned by the operator. The gears and bearings are part of a gear package set, which is hydraulically pushed out of the gearbox housing and can be replaced by a second clean set (Figure 2). All other parts are easily accessible and cleaned manually.

This new device enables a broader application field for gear pumps, and increases the acceptance of this machine type in the tire industry.

Line control via EthernetThe extrusion line control systems are part and parcel of Troester’s equipment portfolio. The extrusion unit works with the downstream equipment as a functional entity, enabling constant product dimensions in the various production stages. The control is developed, designed and programmed in-house using the latest available electrical components. Specification and preferred subsuppliers are considered

according to the individual needs of the tire manufacturer. Typically the line control consists of a PLC-PC architecture.

In the field area all electrical units such as drives, sensor, measuring equipment, identification, and marking system are controlled by PLC. For many years the most well-known field bus systems with the distributed I/O stations have been used. The advantage is reduced cabling on customers’ sites and more efficient checking of machine groups before delivery.

In a further development step, Troester recently installed complex tire production lines by using Ethernet instead of typical field bus systems. More than 80 units, including drives, distributed I/O stations, measuring devices, PLCs and line-PC were linked together.

The line-PC is used for visualization and process trending, recipe handling, and the long-term production and quality protocol. For each production run, the production parameters as well as the quality results will be stored. Such

production results are summed up in statistically measured parameters such as CPK values (statistical process ratio). As well as providing long-term information about the production history, the data can also be used for labeling the manufactured goods.

There is a tendency toward connecting the line-PC to the plant network to receive the pre-selected daily production schedule of the line from a host, and to provide the production data automatically from the line via network to a plant server. State of the art is a configuration with the PC as a redundant system to ensure that quality documentation and production reports are provided without any time lag.

The thorough Ethernet concept supports teleservice up to each end-connected unit. This way Troester engineers can support the operators on site in real time from their headquarters in Hannover, Germany. Ethernet supports the tendency of merging the company network with the entire line control.

Tire manufacturers around the world face daily challenges that require a variety of new manufacturing approaches. By using Troester’s engineering capacity, many ideas result in innovations and end up increasing the bottom-line results. tire

“This new device enables a broader application field for gear pumps, and increases acceptance of this machine”

Figure 2: Quick-cleaning gear pump

122

w w w . t r o e s t e r . d e

The tire industry is striving to

constantly improve its products

and manufacturing processes.

Our engineers develop inno-

vative production lines together

with our customers to meet

these requirements.

The TROESTER development

team uses state-of-the-art

computer-aided methods which

enable quick individual design-

ing of the line components.

Material fl ows, compound

temperatures and line speeds

are adjusted to the customer’s

requirements. Additionally, our

customers can use the exten-

sive test equipment in our

well equipped TROESTER

technology center to conduct

experiments and directly imple-

ment knowledge gained in

development.

TROESTER turns innovative

ideas into reality.

TROESTER GmbH & Co. KG has been develop-

ing innovative machines and lines for the

rubber and plastics processing industry since

1892. If you wish to know what innovations

we can provide, then send us an e-mail

to I nnovat ions@t roes ter.de

The VMI MAXX tire building machine can be equipped with a bead-apex handling and green-tire removal system, based on a robot.

Automating bead loading and green-tire removal disengages the operator from a short-cyclic task in front of the machine and gives him (or her) time to handle components on one or two VMI MAXX machines and reduces the total amount of required operators.

Each bead apex assembly (BA) lies on a specially designed VMI carrier. The carriers are stacked on a cart and fed into the VMI MAXX. To take full advantage of this system, VMI developed a stacking module for its bead apex assembly machines. This module, in combination with the new robotic system for bead loading in the VMI MAXX, creates further possibilities for full hands-off production.

In a nutshell the system consists of an automatic loading unit at the bead

apex assembly machine, carriers for storage and internal transportation, and an automatic unloading unit at the tire building machine. There are several requirements for such a system.

Automatic stacking of the BAs in the bead apex-assembly machine requires automatic quality monitoring of each BA produced, automatic placement of the BAs onto carriers, and controlled orientation of the BA splice position on the carrier.

Storage and internal transportation requires the optimal quality of produced BAs to be maintained while in intermediate storage. The carrier must be of a sturdy design (long lifetime) and

A stacking system has been developed for bead apex assembly machines that, combined with a robotic loading system, enhances hands-off tire production

by J.K. Grashuis, VMI-Group, the Netherlands

Automatic bead apex stackingand loading into a tire building machine

Figure 1: The robot gripper is used to pick up the green tire. The gripper is servo-driven and can handle a BA range from 13-24in, without any manual adjustments required by the machine operator, aiding efficiency

124

Figure 2: The VMI MAXX machine can be equipped with a robotic bead apex handling and green tire removal system

the stack of carriers must be stable and self-supporting. Special carts for easy handling around the machines are needed and minimum floor space for intermediate storage is required. There should also be the possibility of handling stacks in an automatic warehouse and of RFID identification on carts.

Automatic loading of the BAs in the tire building machine requires that there be no sticking of bead and apex compounds to the carrier, reliable loading of bead-apex assemblies, and controlled orientation of the BA splice in the green tire, as well as continuous production of approximately 50 minutes before reloading.

The carrierBased on the requirements listed above, a special carrier was developed and protected by a patent. The carrier is injection-molded and made of high-quality plastic. The carriers are optimized for storage of the BA. The BA lies on a conical surface with an angle of 10°, to support the shape of the initially hot apex.

The carrier has an outside rim to keep the BA sufficiently centered and to protect the apex and it has a fixed orientation in the stack. This facilitates control over the apex-splice in the tire. The carrier comes in two sizes: one up to 20in and the other up to 24in.

The open shape of the carrier prevents accumulation of heat when the fresh BA is stored on the carrier. The carrier’s upper surface has a texture to prevent the BA sticking.

For internal transportation the carriers, which have a vertical pitch of 24mm, can be stacked up to a height of 56 pieces (± 1.4m). The stack is stable and self-supporting. VMI supplies special carts, equipped with wheels for easy handling.

For intermediate storage, with a stack height of 56 pieces, 118 small carriers (max 20in) or 90 big carriers (max 24in) can be stored per m2. The storage capacity can be multiplied by storing the stacks in a rack.

To aid logistic control, each cart can be equipped with an RFID chip for logistical identification of the stack and for track-and-trace functionality.

BA stacking moduleThe BA stacking module that is added to a bead-apex assembly machine places the BA directly onto a carrier and stacks the loaded carriers onto a cart. After adding the BA stacking unit, it is still possible to use the bead apex assembly machine in the conventional way. Implementing this system means no operator checks or even touches the BA during the whole process, and an automatic VMI vision system monitors the quality of the BA.

Tire building machineThe central part of the module that is added to the VMI MAXX tire building machine is a robot that is located between breaker and tread, and carcass servicer. Up to three stacks of carriers with BAs can be loaded into the system. This equals 168 BAs, or 84 tires.

The robot gripper picks the BA from the carrier by centering and actively separating it. Also the separation of the carriers is done actively, resulting

Table 1: Requirements for a stacking systemAutomatic stacking of the bead apex assemblies in the bead apex assembly machine

Automatic quality monitoring of the produced bead apex assemblies

Placement of the bead apex assemblies into carriers

Controlled orientation of the bead apex (splice position) in the carrier

Avoidance of distortions of produced bead apex assemblies while in intermediate storage

Carrier must be of a sturdy design (long lifetime)

Stack of carriers must be stable and self supporting

Minimum floor space for intermediate storage

Placement of stack in automatic warehouse must be possible

RFID identification on stack of carriers must be possible

No sticking of bead and apex compounds to the carrier

Easy loading of bead apex assemblies into tire building machine

Autonomy of approximately 50 minutes before reloading in the tire building machine

Figure 3: The robot is mounted on a solid portal, and contains the conveying system and safety light screens

125

126

in a reliable process. The empty carrier is placed onto a stack for empty carriers, of which the system can also contain three. The BA is placed into the VMI MAXX offline bead loader, which assures an optimal positioning accuracy of the BA in the bead setter of the tire building machine.

After adding the robot system it will – if necessary – still be possible to load BAs into the tire building machine in the conventional way.

The gripper is servo-driven and can handle 13-24in without manual adjustment. The same gripper is used to pick up the green tire. Again no manual adjustments are required.

The robot can place the green tire onto a customer-specific unloading position, such as a conveyor system. All robot tasks are performed within the tire-building machine cycle time, which is less than 40 seconds.

The robot is mounted on a solid, static portal, which is firmly connected to the floor. The portal also contains the infeed- and outfeed system for the stacks of carriers.

The operator easily rolls the stacks on their carts in and out of the machine. Figure 4: With a stack height of 56 pieces, 118 small carriers of 90 big carriers can be stored on the rack per m2

Figure 5: Top view of the carrier. The robot handles the carrier from the center. The kernel is the same for both sizes

The robot system will continue without interruption during these operations. The safety of the machine operator remains guaranteed, with a safety system that complies with the new machine regulations and also to the CE-PL safety norm.

The system described for automatic bead-apex stacking and loading into a tire building machine offers a comprehensive logistic solution, starting at the production of the BA and ending at the moment the BA is built into a tire.

The robot takes over two tasks from the operator in front of the tire building machine, enabling a smaller workforce. VMI offers several automatic monitoring systems to perform the inspection, that the operator used to perform.

Also, where operators can be distracted by other tasks, the robot makes sure that the BAs and green tires will be handled every 40 seconds, resulting in a stable and continuous output. tire

1624 Englewood Ave. | Akron, OH USA 44305 | +1-330-784-1251 | www.micropoise.com

Tire Measurement Problems?

Micro-Poise® Modular Tire Measurement Systems

(MTMS) are specifically designed to eliminate tire

measurement problems with our streamlined systems,

optimizing the tire measurement process for uniformity,

dynamic balance and geometry measurement.

Contact us today to let Micro-Poise eliminate your tire measurement problems with our MTMS solutions.

FREESUBSCRIPTION/READER ENQUIRY SERVICE

Request your free copy of Tire Technology Internationalmagazine and learn more about innovative, industry-leading suppliers @ www.tiretechnologyinternational.com

OCTOBER 2009INNOVATION INITIATIVE

Jean-Claude Kihn, Goodyear’s CTO, has

undertaken a historic innovation strategyTIRE LABELING

How will the forthcoming legislation benefit customers?

TOMORROW’S WORLD

Michelin’s view of future standards

and regulations, and how to meet them

Lunar impactHow the moon will help generate

possible next-generation tires

INTERVIEWSDr Joachim Neubauer

Standards and regulation, Michelin

Nina RenshawPolicy officer, T&EDr Stuart Cook

Director of research, TARRC

Professor Joshua Summers

Clemson UniversityBen MichellDunlop Motorsport

128

T he tire market demands excellent product quality at competitive prices. To provide a high-quality product at an economically

efficient cost, manufacturers have to keep scrap and system complexity to a minimum, but cost cutting must never compromise product quality.

How can these requirements be aligned in one single application? Width measuring systems are a possible approach to this issue. The systems consist of sensors, analog sensors or CCD cameras, for data logging, and a controller to process the incoming data. Web-width measurements can be used simply as a display instrument to provide the user with material width data, or integrated into the machine controls so as to be actively involved in the production process by stopping the machine in the event of an incorrect width.

Web-width measurements are perfect for high-quality tire production for a variety of reasons and offer a huge savings potential. The systems provide precise information on the relationship between defined reference width and current material width during production. Likewise, they notify the user of deviations. Web-width measurements are suitable for measuring the position deviation of a material strip on a carrier material.

Measuring systems, used correctly during the production of raw materials, such as the innerliner, cushion, textile cord, tread, sidewall, can avoid the production of defective material and scrap by preventing these materials from being fed into the next production step.

Although material production cannot be influenced in downstream processes, in tire building machines the systems may well be used to determine further use. Once defective materials have been fed into the next production step and reach the end-product stage, ‘green tire’, they become special waste. The disposal

of this special waste is very costly, as the product mix of the various materials,includingrubber, textile, and steel, cannot easily be split into its components – only by applying a complex process.

Theapplications of width measuring systems are quite versatile, from a simple display of current material status to systems that produce detailed reports for use in downstream machines (workflow). These reports store all product data for each millimeter of material. The reports can be used to set up a quality control system or be integrated into existing systems.

In the case of the latter, only raw data is transmitted to the customer, who performs his own data evaluation. This enables the customer to use his own criteria to assess product quality.

The quality of a web-width measurement system is determined by the selection of suitable components. Ease of operation, along with an appropriate sensor system, is critical.

In the tire industry, CCD cameras are frequently used, which can even compensate height or thickness of the material to be measured, based on a BST International calibration method for which a patent has been filed.

Where space is limited, the best choice may be analog sensors, which can be positioned using fully automatic sensor adjusters. Just like a measuring system

with CCD cameras, these sensors are calibrated and provide measured values with a precision of 0.1mm or better.

When opting for a system made by BST International, users may choose from a broad range of products, from basic analog sensor variants with manual adjustment, to fully automatic sensor positioning and CCD cameras. A uniform ‘look and feel’ of component operation is a critical factor as this ensures that users can quickly and reliably operate systems regardless of the machine they are working on and what type of width measurement they are actually using.

If a BST International web-guiding system is already in use, the sensor system may be used for the control process and for web-width measurement at the same time. This gives additional synergies.

The use of web-width measurement systems provides a range of possibilities to enhance cost efficiency and environmental compatibility in production by saving on material, while ensuring highest end-product quality. tire

Tire manufacturers must minimize scrap and system complexity, without compromising product quality. Width measuring systems can help by Andreas Flöter & Sabine Sladky, BST International GmbH, Germany

Web width measuring systems

The EKR500 Edition X commander unit for a BST web-width measuring system

BST International GmbH | Heidsieker Heide 53 | 33739 Bielefeld | Germany | Phone: +49 5206 999-0 | Fax: +49 5206 999-759www.bst-international.com | automation@bst-international.com | A member of the group

Web Guiding, Web Width Measurement and Web Width Control systems for:

• Innerliner• Extrusion Lines• Calender Lines (Textile- and Steelcord)• Assembly Lines

One step ahead with BST International

Always Hitting Your Targets

• Cutting Lines (Textile- and Steelcord)• Tire Building Machines• Doubling Stations• Special Applications

KONŠTRUKTA-Industry, a.s. TRADITION IN TYRE MACHINERY

KONŠTRUKTA-Industry, a.s.K výstavisku 13912 50 TrenčínSlovak Republice-mail: sales@kotaind.sk

CUTTING TECHNOLOGY• Steel Cord Cutting Lines

- for Breaker Cutting & Splicing- for Body Ply Cutting & Splicing

• Steel Cord Cutting Lines combined- for Body Ply and Breaker Cutting & Splicing

• Textile Cord Cutting Lines- for Body Ply Cutting & Splicing

• Roller-Head Extrusion Lines- for Inner Liner Production- for Squeegee Production- for Cushion Production- for Gum Strip Production

• Cold & Hot Feed Extruders- for Bead Wire Rubberizing- for Apex Production- for Feeding Systems

www.konstrukta.sk

EXTRUSION TECHNOLOGY• Multiple Extrusion

Lines- for Apex Production- for Sidewall Production- for Tread Production

130

During the last 20 years there have been enormous improvements in materials and construction, and in the manufacturing process of

car tires. However, improvements in the tire curing process have been marginal.

Mechanical curing presses have been replaced by hydraulic operating systems, and improvements have been made in the control and handling systems. Radial-opening mold systems for PCR and TBR sizes are now common.

The tire mold shapes the final product, and the end user has a large number of different tire brands to choose from. In addition to the quality of the tire, promotions by test magazines, along with pricing and appearance (especially for PCR sizes), should not be underestimated.

As an independent tire mold producer, which sells to international tire companies, A-Z Formen has seen customers investing in new tread designs to improve the general appearance and upgrade the overall performance of their products.

Tire customers look for a modern, state-of-the-art design. Tire designers try to create up-to-the-minute designs with variable groove configurations, particularly for winter designs with a large number of sipes. The removal of air during the curing process requires a large number of vent holes (up to 5,000 per mold) and creates stubbles, which must be cut off. This additional production step creates an undesirable waste of rubber.

By using valve vents that automatically close after air removal, the curing press unloads the tire with a smooth, flawless surface. Spring vents are already being

used by many customers and can be installed into any kind of tire mold. However, care must be taken when cleaning the mold by blasting it with dry ice, to avoid damage to sensitive valves.

The continuation of the tread pattern due to wear during the lifetime of a tire is not just an optical aspect. Trimming the sipes is necessary to avoid damage to the sipes on the segment split area. In this case, due to normal tread wear, the tread picture will be interrupted, and sipe

grooves will disappear. The slightly different rubber volume in the split areas might also influence the general performance.

To avoid cutting through sipes, the split between the individual segments is designed to follow the bent sipes (S-split) and no sipe needs to be trimmed. The tread pattern remains and the overall performance will be unchanged during the life of the tire.

In addition to the use of high-precision segments, accurate mold-closing systems are essential.

A-Z Formen helps its customers meet their technical demands and also helps them maintain the quality of their molds.

A-Z’s in-house technology and different production possibilities for the tread pattern, combined with the well-established A-Z container system, ensure that the company remains one of the leading independent suppliers of tire molding products in the world. tire

State-of-the-art production technology makes all the difference when choosing a tire mold

by Rainer Hilke, A-Z Formen- und Maschinenbau GmbH, Munich, Germany

Tire molding: the final stepin the tire production chain

Tire molded with spring vents. Note the smooth surface Tire molded with small hole vents, with stubble

S-split segment at parting requires no sipe trimming

THETIREMOLDING COMPANY SINCE 40 YEARS

LINE OF PRODUCTS

Molds for tire curing in all sizes

In 2 piece or radial construction

Mold closing systems

AZ Container systems (WORLD PATENTS)

Machines for mold production

IN HOUSE PRODUCTION FACILITIES

Precision aluminium foundry for tread pattern

Tread pattern production by engraving in AL & STEEL

Complete production process by CAD - CAM equipment

A-Z FORMEN- UND MASCHINENBAU GMBHDessauerstraße 9 · 80992 München · GermanyPhone + 49-89 -14 9817- 0 · Fax + 49-89 -14 9817- 90info@a-z-gmbh.de · www.a-z-gmbh.de

Y.MTIS

Y.MTIS ■ the standard in tire inspection ■ over 100 units sold worldwide

The solution offers ■ superior image quality ■ modular design■ easy and intuitive operation■ extreme reliability

Further information ■ Please contact our sales organization or visit us at www.yxlon.com

YXLON. The reason whyYXLON International GmbH, a company of the COMET Group. Essener Bogen 15, D-22419 Hamburg,T: +49 40 527 29 - 0, F: +49 40 527 29-170, E-mail: yxlon@hbg.yxlon.com, www.yxlon.com

Modular.Precise.Fast

Standard in Tire Inspection

MTIS – Modular Tire Inspection Systems

132

In the tread extrusion process the treads are wound, cut, or directly fed to the tire building machine at the end of the extrusion line.

When cutting treads it is vital to check the important tread parameters continuously to guarantee high-quality treads. The basic parameters which can be measured are width, length and weight. Additionally, in modern extrusion lines, profile measuring systems check the cross-section profile of the cut treads.

The line control is able to gather all these important values for each produced single tread. Treads out of tolerance can be sorted out of the production process manually or automatically, to avoid the production of poorer-quality tires.

Although weight measurement with a final scale and width measurement with CCD cameras have been used for a long time, modern camera-based length

measuring systems and laser-based profile measuring systems are helpful new tools for judging cut treads.

The practice of length measurement began with simple rollers that have been used for decades, before moving to light barrier and line-scan camera-based systems, and then to modern CCD camera-based units.

Figure 1 shows the definition of tread length. The complicated situation results from the angled cutting area and the profiled structure of the tread.

Old-style length measuring systems lead to large instabilities and are influenced by many problematic characteristics: even the smallest inhomogeneity of the tread at the cutting position results in incorrect measurements when measured at only one point by laser light barriers along the wide cutting line; movements or vibrations perpendicular to

the transport direction result in significant errors when point sensors are used to detect the edges; applying rollers for length detection results in failures due to slippage of the transport equipment. The wear of the roller and dirt on the roller also influence the results, and some systems need a split conveyor belt, due to the backlight illumination necessary for line-scan cameras.

All these disadvantages are overcome by a CCD area camera-based measuring system. Within several microseconds, two high-speed cameras take images of the tread’s ends simultaneously while running along the final scale’s conveyor belt. The tread can run through the measuring position with high speed without influencing the measuring accuracy. Due to the used-area cameras, the length is not only checked at one point of the tread: in the field of view

Controlling the tread cutting process and monitoring tread parameters improves the results of further production steps, and improves product quality

by Dr Hartwig Suhr, Dr Noll GmbH, Germany

The importance of checkingthe length, width, weight and profile of cut tread

Figure 1: Schematic view of a length measuring system. The complicated situation results from the angled cutting area and the profiled structure of the tread

133

of the camera up to 700 lines are taken to check the length. Special filtering by software and special illumination procedures enable the system to find exactly the cutting lines inside the product. An overall accuracy of 0.5mm or better can be achieved over many years, independent of the length of the tread. Because external influences are strongly suppressed by this new method, the system’s performance will be stable once installed and calibrated.

Figure 1 shows a schematic view of the length measuring unit and the operator’s menu is shown in Figure 2.

The image of the detected ends of the tread are shown on the monitor together with the resultant length and detected edges. This facilitates the operator’s regular check to ensure that the system is measuring correctly. Furthermore, the operator is supported by the additional trend display.

The length system can be applied for tread with the profile lying either upside up or upside down. Figure 3 shows a length measuring system placed above the final scale’s conveyor belt directly after the skiver. In this case the tread is running upside down.

“Special filtering by software and special illumination procedures enable the system to find exactly the cutting lines”

Figure 3: Inline system with scale and length measuring system. The system is placed above the final scale’s conveyor belt directly after the skiver. The tread runs upside down

Figure 2: The length measurement system’s operator menu displays a combination of weight, width and length

Fixed camera Linear guidance Illumination

Scale

Skiver

Motor

Camera

The measured length value is not only important for sorting and qualifying the produced tread; it is also possible to give feedback to the skiver to correct the cut length for the following treads and to provide early information if the skiver’s knife is running into problems.

For best results, the length measuring system should be installed at the final scale together with a width measuring system before the scale. A compact system controlled by an industrial PC collects the most relevant tread parameters. Figure 2 shows a system menu with the measured length, width and weight of each tread.

All data (length, width and weight) for each tread can be stored in the industrial PC and can be used for further quality evaluation. In addition to length measuring for cut tread, this system is also available for cut sidewalls.

For a long time width measurement has been done by CCD cameras. Nowadays cameras with up to 5,000 pixels provide, by theoretical calculations, best results for tread width measurement. However results in practice are often worse due to underestimated influences of the applied optics, installation, layout and evaluation process of the measuring

system. Only careful selection of these components by experienced engineers, together with correct alignment of the camera, backlight illumination and the measured product in combination with software that calculates the statistical mean of a great number of measured width values of each single tread, give the best-possible results that can be obtained with the newest techniques.

Weight measurement by the final scale can achieve very high accuracies. Although the techniques of the weight cells have been stable for a long time, the scale outline to reduce the required total length of a scale has been improved in recent years. This is very important due to the increasing speed of the treads when passing the final scale.

For accurate length measuring the tread has to lie flat on an even surface; therefore the conveyor belt of the final scale is the best solution for the position of this measurement. A combination of both measurement systems is a good solution for this reason.

Now the only thing missing from the complete tread examination is the profile determination.

Figure 4 shows a profile measuring system which is installed at the end of the extrusion line after the skiver. It is based on scanning laser point sensors. Cross-section, shoulder width, symmetry, conicity and the control of important thickness points are some of the results of this unit. As an alternative to the scanning measuring system, a system with light intersection cameras (see Figure 5) can be placed before the skiver. Although this system is less accurate than the scanning system, sometimes this principle is the better solution, if fast measurement is required and if the budget is limited. This system has to be placed before the skiver, because the required opening in the transport equipment is too large for the cut treads. The scanning laser point system requires a much smaller gap.

New developments in measurement techniques have resulted in compact units, which make it possible to combine the measurement of all the important tread parameters at the end of the production process.

The subsequent production steps to the final tire are more safe and there is less scrap and less wasted production time from the building process until curing and final test results. Fewer tires are rejected in the final tests, if the early production steps are carefully monitored. tire

“For best results, the length measuring system should be installed at the final scale, together with a width measuring system”

Figure 4: Scanning laser profile measurement. One sensor from top, one sensor from the bottom scan the product

Figure 5: Profile measurement by light intersection

134

136

FREE SUBSCRIPTION/READER ENQUIRY SERVICE

Request your free copy of Tire Technology Internationalmagazine and learn more about innovative, industry-leading suppliers @ www.tiretechnologyinternational.com

OCTOBER 2009INNOVATION INITIATIVE

Jean-Claude Kihn, Goodyear’s CTO, has

undertaken a historic innovation strategyTIRE LABELING

How will the forthcoming legislation benefit customers?

TOMORROW’S WORLD

Michelin’s view of future standards

and regulations, and how to meet them

Lunar impactHow the moon will help generate

possible next-generation tires

INTERVIEWSDr Joachim Neubauer

Standards and regulation, Michelin

Nina RenshawPolicy officer, T&EDr Stuart Cook

Director of research, TARRC

Professor Joshua Summers

Clemson UniversityBen MichellDunlop Motorsport

01dB-Metravib........................................................ 103A-Z Formen-und Maschinenbau GmbH ................... 131Automotive Testing China Expo 2010 ....................... IBCBST International GmbH .......................................... 129China United Rubber (Group) Corporation .................. 75Continental Matador Rubber s.r.o .............................. 13Dr Noll GmbH .......................................................... 135EMS-Chemie AG ..................................................... 103Ergon Inc ................................................................. IFCErhardt+Leimer GmbH .............................................. 57Gabo Qualimeter Testanlagen GmbH ......................... 39H&R Oelwerke Schindler GmbH ................................ 61Hagglunds Drives AB ................................................ 31Harburg-Freudenberger Maschinenbau GmbH ............ 8Holly Corporation ........................................................ 4Indspec Chemical BV ................................................ 65Intralox ...................................................................OBCKonstrukta-Industry a.s ........................................... 129

Kordsa Global Endustriyel Iplik ve Kord Bezi ................ 3Lanxess Deutschland GmbH ..................................... 87MicroPoise Measurement Systems LLC ................... 127Nynas Naphinenics AB ................................................ 7Polyamide High Performance GmbH .......................... 58RJS Corporation ........................................................ 61Samson Machinery Inc ............................................. 21Standards Testing Labs ........................................... 107The Pettibone Tire Equipment Group ....................... 115The Shepherd Chemical Company ............................ 65Tire Technology Expo 2010 ........ 42, 45, 46, 95, 96, 104Tire Technology International Online Reader Enquiry Service ......................6, 58, 127, 136TROESTER GmbH & Co KG ...................................... 123TSM ........................................................................ 135VMI EPE Holland BV ................................................ 108Yxlon International GmbH ........................................ 131

Index to Advertisers

AUTOMOTIVE TESTING EXPO CHINA 2010Abinger House, Church Street,Dorking, Surrey, RH4 1DF, UKTel: +44 (0) 1306 743744 Fax: +44 (0) 1306 877411 email: expo@ukintpress.com

Our business partner in China is:Mr. Wilko FongReliable International Exhibition Services Co LtdRm.1702, Bldg.6, SOHO New Town, No. 88 Jianguo Road, Chaoyang District, Beijing 100022, P.R. ChinaTel: +86 10 85898181 • Fax: + 86 10 85898180 • Email: wilko_fong@reliable.org.cn

www.testing-expochina.com

China’s ONLY automotive test, evaluation and quality engineering trade show

September 14-16, 2010The CIAE Exhibition Center, Chaoyang District, Beijing, China

Let your conveyor be a source of profit.

MAINTENANCECOSTS

PRODUCTQUALITY

EFFICIENCY

At Intralox, we use innovative problem-solving to optimize conveyance in your plant and deliver solutions proven to reduce maintenance and unscheduled downtime. Our commitment to the industry is to bring meaningful economic value to each customer and provide the global support and guarantees to ensure complete success.

Please contact the tire team to calculate the savings in your conveyors.www.intralox.com.