D3.2 Noise vibration and harshness

Towards a REgulatory FRamework for the usE of Structural new materials in railway passenger and freight

CarbOdyshells

Grant Agreement no.: 605633

WP 3.2 Noise vibration and harshness

Deliverable: D 3.2 Due date of deliverable: D[30 11 2014] Submission date: [30 11 2014] Version: [Issue 1]

Project co-funded by the European Commission within the Seventh Framework Programme

D3.2 – Proposal of NVH strategy for structural composite parts based on material properties

1

REFRESCO Deliverable D3.2 was produced by UTC and received contributions from

the following members of the consortium:

• [BT]

• [DLR]

This document should be referenced as:

“REFRESCO- Proposal of NVH strategy for structural composite parts based

on material properties,D3.2,version1"

QUALITY CONTROL INFORMATION

Issue Date Description Revising Authorship Draft 6 11 2014 Draft version of REFRESCO D…

for TMT COMMENT Sébastien Personne, Nicolas Dauchez,UTC

Final 30 11 2014 Submission of REFRESCO D… final version to the EC

Sébastien Personne, Nicolas Dauchez,UTC

DOCUMENT HISTORY

Issue Date Pages Comment

1 30/11/2014 67 Initial issue

2

3

4

5

DISSEMINATION LEVEL

PU Public x

PP Restricted to other programme participants (including the

Commission Services)

RE Restricted to a group specified by the consortium (including the


CO Confidential, only for members of the consortium (including the



2

TABLE OF CONTENTS

Executive summary ................................................................................................. 10

1. Introduction ...................................................................................................... 13

2. Composite for light weight car body state of the art .......................................... 14

2.1. Light weight structure in Railway ............................................................... 14

2.1.1. Composites ........................................................................................ 16

2.1.2. Sandwich structures ........................................................................... 17

2.1.3. Examples of composite carbody construction ..................................... 18

2.2. Carbody functions ..................................................................................... 20

2.2.1. Mechanical constraints ....................................................................... 21

2.2.2. Acoustical constraints ........................................................................ 22

2.2.3. Objectives of the report ...................................................................... 26

3. Vibro-acoustic behaviour of a panel, single and double layer ........................... 27

3.1. Radiation efficiency ................................................................................... 27

3.2. The transmission loss................................................................................ 28

3.2.1. Single wall panel ................................................................................ 28

3.2.2. Double wall ........................................................................................ 30

3.3. Sound insulation and excitation type. ........................................................ 31

3.3.1. Oblique wave and diffuse acoustic field .............................................. 32

3.3.2. Turbulent Boundary Layer .................................................................. 32

3.3.3. Rain on the roof ................................................................................. 33

3.4. Specificities of composite panels ............................................................... 33

3.5. Experimental analysis of composite structure ............................................ 35

4. Modelling for acoustical criteria ........................................................................ 37

4.1. Approaches for vibro-acoustic simulation .................................................. 37

4.1.1. Global approach ................................................................................. 37

4.1.2. Multilayer modelling ........................................................................... 39

4.1.3. Local model for layers ........................................................................ 40


3

5. Main drawbacks / advantages of using composite from vibroacoustic point of

view, and how to face the drawbacks ...................................................................... 41

5.1. Drawback and countermeasures ............................................................... 42

5.1.1. Airborne transmission: ....................................................................... 42

5.1.2. Structure borne transmission ............................................................. 49

5.2. Advantages ............................................................................................... 49

5.3. Cases study with TMM approach .............................................................. 51

5.4. Role of joints on vibro-acoustic design ...................................................... 54

6. Conclusion ....................................................................................................... 56

Bibliography ............................................................................................................ 57

Appendix I: Composite structure from literature review ............................................ 61

Appendix II: Datasheet for honeycomb .................................................................... 64

Appendix III: Parametric study ................................................................................. 65


4

LIST OF TABLES

TABLE 1 : WP2 MATERIAL BENCHMARKING FOR LIGHTWEIGHT STRUCTURE IN TRAIN ...................... 13 TABLE 2 : COMPOSITIONS OF COMPOSITE PANELS WITH HONEY COMB OR FOAM CORE AND FOR FLOOR

OR ROOF APPLICATIONS CASES. ......................................................................................... 22 TABLE 3 : CONTRIBUTION OF THE THREE MAJOR GLOBAL NOISE SOURCES IN % OF 𝑳𝑳𝑳𝑳,𝑨𝑨, 𝒆𝒆𝒆𝒆 AT

300KM/H (LEFT) AND 360KM/H (RIGHT) ............................................................................... 23 TABLE 4: EVALUATION OF PANEL VIBRO-ACOUSTIC BEHAVIOUR WITH NOVA.................................... 52 TABLE 5 : MATERIAL PROPERTIES FOR SIMULATION ...................................................................... 52 TABLE 6 : TYPICAL FREQUENCY FUNCTION OF PANEL’S STRUCTURE .............................................. 53

LIST OF FIGURES

FIGURE 1: A) MASS RATIO BY PASSENGER FOR DIFFERENT TRANSPORT MEANS (WENNHAGE 2005);

B) EVOLUTION OF CO2 EMISSIONS (TRANSPORT 2010) ........................................................ 14 FIGURE 2 : REPARTITION OF THE CARBODY WEIGHT (WENNBERG 2011) ........................................ 15 FIGURE 3 : SUBSTITUTION OF CORRUGATED STAINLESS STEEL BY COMPOSITE PANEL (WENNBERG,

STICHEL ET PER 2012) ...................................................................................................... 15 FIGURE 4: WEIGHT AND COST SAVING IN THE FUTURE USING COMPOSITES IN TRANSPORT (ROBINSON

ET CARRUTHERS 2005) ..................................................................................................... 15 FIGURE 5 : TYPES OF MATERIALS USED IN TRANSILIEN (PATRICK 2013)........................................ 16 FIGURE 6 : EXAMPLE OF A LAYER REINFORCE WITH UNIDIRECTIONAL FIBRES (WENNBERG 2011) .... 16 FIGURE 7 : COMPOSITION OF A LAMINATE PANEL. (WENNBERG 2011) ........................................... 17 FIGURE 8 : SANDWICH STRUCTURE (WENNBERG 2011) ................................................................ 17 FIGURE 9 : COMPARISON OF STRESS REPARTITION BETWEEN SANDWICH AND UNIFORM BEAM

(WENNBERG 2009) ........................................................................................................... 18 FIGURE 10 : MECHANICAL PROPERTIES FUNCTION OF CORE THICKNESS (ZENKERT 1997) .............. 18 FIGURE 11 : A) TTX, B)THE C20FICA BY BOMBARDIER (WENNBERG 2011). ................................. 18 FIGURE 12 : A) SANDWICH PANELS AND FRAME STRUCTURE (KIM, LEE ET SHIN 2007) , B)MODELLING

PROCESS FOR THE CARBODY (JUNG-SEOK ET JONG-CHEOL 2006) ...................................... 19 FIGURE 13 : STAINLESS STEEL FRAMEWORK FOR SIDE AND ROOF (JUNG-SEOK ET JONG-CHEOL

2006) AND FLOOR (JUNG-SEOK, JONG-CHEOL ET SANG-JIN 2007) ..................................... 19 FIGURE 14: SANDWICH STRUCTURE FOR THE C20 (WENNBERG 2011) .......................................... 20 FIGURE 15 : SCHEMATIC VIEW OF FUNCTIONS THAT THE CARBODY HAS TO FULFIL (WENNBERG 2013)

........................................................................................................................................ 20 FIGURE 16 : REPLACEMENT OF THE CORRUGATED PANEL FOR THE FLOOR BY A SANDWICH COMPOSITE

PANEL (WENNBERG, STICHEL ET WENNHAGE 2012) ........................................................... 22 FIGURE 17 : SPECTRUM FREQUENCY ASSOCIATED TO THE TREE MAIN SOURCES (POISSON, ET AL.

2011) ............................................................................................................................... 23


5

FIGURE 18 : AIR BORNE AND STRUCTURE BORNE CONTRIBUTIONS FUNCTION OF SPEED (POISSON, ET

AL. 2011) .......................................................................................................................... 24 FIGURE 19: MULTI-OBJECTIVES CONCEPTION PROCESS FOR A SANDWICH PANEL (WENNBERG 2013).

........................................................................................................................................ 24 FIGURE 20 : OPTIMIZATION PROCESS FOR STRENGTH AND ACOUSTIC INSULATION TARGET

(WENNHAGE 2005) ........................................................................................................... 25 FIGURE 21: DESIGN PROCESS FOR A COMPOSITE PANEL WITH AN ACOUSTIC CONSTRAINT (WENNBERG

ET STICHEL 2014) ............................................................................................................. 26 FIGURE 22: ACOUSTIC CONSTRAINT PERFORMED FOR A BROAD BAND SOUND REDUCTION INDEX LEVEL

(WENNBERG ET STICHEL 2014) ......................................................................................... 26 FIGURE 23 : EVOLUTION OF THE TRANSMISSION LOSS WITH FREQUENCY (GRIESE 2012)................ 29 FIGURE 24 : COINCIDENCE PHENOMENON A) PLATE AND SOUND WAVE NUMBERS AS FUNCTION OF

FREQUENCY; B) COINCIDENCE BETWEEN SOUND AND BENDING WAVELENGTHS. (CHAZOT 2006)

........................................................................................................................................ 30 FIGURE 25: TYPICAL TRANSMISSION LOSS OF A SIMPLE SUPPORTED DOUBLE PANEL (FAHY 1985) .. 31 FIGURE 26: TRANSMISSION LOSS OF AN INFINITE A) SINGLE WALL PARTITION B) DOUBLE WALL FOR

VARIOUS ANGLES OF INCIDENCE. (TEWES 2005) ................................................................. 32 FIGURE 27 : A) OBLIQUE PLANE WAVE EXCITATION, B) DIFFUSE ACOUSTIC FIELD (ESI 2014) .......... 32 FIGURE 28: WAVE VECTOR SPECTRA FOR DIFFERENT TYPES OF RANDOM EXCITATION AS A FUNCTION

OF THE STREAM-WISE WAVE NUMBER: TBL, DAF, RAIN ON THE ROOF

(FROM REFERENCE (MAURY, GARDONIO ET ELLIOT 2002)).................................................. 33 FIGURE 29 : IMPACT OF THE PANEL’S STIFFNESS IN THE RADIATION EFFICIENCY (FAHY 1985)......... 34 FIGURE 30: CONCEPTUAL SCHEMA FOR LIGHTWEIGHT STRUCTURE DESIGN (MOSCHINI 2014) ........ 35 FIGURE 31 : EXPERIMENTAL INVESTIGATION OF PANEL DYNAMICAL PROPERTIES (CHRONOPOULOS

2012) ............................................................................................................................... 36 FIGURE 32 : TEST SET-UP FOR TRANSMISSION LOSS MEASUREMENTS (VAN DER WAL ET NILSSON

2005) ............................................................................................................................... 37 FIGURE 33: DIFFERENT WAYS AND ACCURACY FOR MODELLING (ATALLA 2012) ............................. 37 FIGURE 34: TYPICAL FEM BEM APPROACH FOR TRANSMISSION LOSS CALCULATION (HUFENBACH, ET

AL. 2010) .......................................................................................................................... 38 FIGURE 35: SEA REPRESENTATION OF THE DOUBLE WALL SUBSYSTEM (CAMPOLINA, ATALLA ET

DAUCHEZ 2012) ................................................................................................................ 39 FIGURE 36: TRANSFER MATRIX REPRESENTATION OF THE TRANSMISSION TROUGH DOUBLE-WALL

SUBSYSTEM (KUO, LIN ET WANG 2008) .............................................................................. 39 FIGURE 37: EXAMPLE OF POROUS MATERIALS (ATALLA ET PANNETON 2012)................................. 40 FIGURE 38: HONEYCOMB SCHEME AND ASSOCIATED ORTHOTROPIC DIRECTIONS (PLASCORE S.D.) . 41 FIGURE 39: NOISE AND VIBRATION TRANSMISSION PATHS IN A RAILWAY COMPARTMENT (CARLSSON

1997) ............................................................................................................................... 42 FIGURE 40: EVOLUTION OF THE WAVENUMBER OF A COMPOSITE PANEL AND ASYMPTOTIC BEHAVIOUR

OF SINGLE LAYERS (ATALLA 2012) ..................................................................................... 43 D3.2 – Proposal of NVH strategy for structural composite parts based on material properties

6

FIGURE 41 : SCHEMATIC IMPACT OF THE LIGHT STRUCTURE ON TRANSMISSION LOSS (DESMET,

WWW.FP7-ELIQUID.EU S.D.) ............................................................................................... 43 FIGURE 42 : TRANSMISSION LOSS FOR A SINGLE LEAF PANEL-IMPACT OF THE DAMPING .................. 44 FIGURE 43 : EXAMPLE OF DAMPING LAYER IN COMPOSITE PANEL (DESMET, WWW.FP7-ELIQUID.EU

S.D.) ................................................................................................................................. 44 FIGURE 44 : IMPACT OF THE DAMPING TREATMENT ON THE RADIATED POWER (ASSAF, GUERICH ET

CUVELIER 2010) (ASSAF, GUERICH ET CUVELIER 2010) ..................................................... 45 FIGURE 45 : IMPROVED TRANSMISSION LOSS WITH SPECIFIC HONEY COMB DESIGN (GROSVELD,

PALUMBO ET KLOS 2006) .................................................................................................. 46 FIGURE 46 :IN-PLANE HONEYCOMB CORE VERSUS OUT-OF-PLANE CORE (GRIESE 2012)................ 46 FIGURE 47 : VARIATION OF THE TRANSMISSION LOSS WITH VARIATION OF CORE GEOMETRY (GRIESE

2012) ............................................................................................................................... 47 FIGURE 48 : MASS SPRING SYSTEM FOR METAMATERIAL REALIZATION (DESMET, WWW.CAV.PSU.EDU

S.D.) ................................................................................................................................. 47 FIGURE 49 : META MATERIAL CONCEPT (DESMET, WWW.CAV.PSU.EDU S.D.) .................................. 48 FIGURE 50 : INCREASED TRANSMISSION LOSS NORMALIZED WITH RESPECT TO ADDED WEIGHT:

FOR A PLATE WITH PIEZOELECTRIC SENSOR AND ACTUATORS; FOR CONSTRAINED LAYER

DAMPING. (AHMADIAN ET JERIC 2001) ................................................................................ 49 FIGURE 51 : EVALUATION OF THE FIRST EIGEN FREQUENCIES FUNCTION OF THE PLY ANGLE

(WENNBERG 2013), A) SHIFT IN FREQUENCY B) MODAL DEFORMATIONS SHAPE ..................... 50 FIGURE 52: EVOLUTION OF THE TRANSMISSION LOSS WITH PROPERTIES OF THE COMPOSITE PANEL.

(WENNBERG, STICHEL ET WENNHAGE 2012) ...................................................................... 51 FIGURE 53 : PARAMETERS FOR FOAM WITH BIOT’S THEORY .......................................................... 52 FIGURE 54 GIVES THE RESULT FOR TRANSMISSION LOSS WITH THE DAF EXCITATION. AS EXPECTED

THE CRITICAL FREQUENCY FOR COMPOSITE PANELS SHIFTS DOWN COMPARED TO STEEL PANEL

AND THE DIPS ARE CLOSE TO THE MAXIMUM HEARING SENSIBILITY (SEE TABLE 6). ................. 52 FIGURE 55: TRANSMISSION LOSS FOR DAF EXCITATION COMPARISON BETWEEN STEEL AND

COMPOSITE PANELS ........................................................................................................... 53 FIGURE 56: TRANSMISSION LOSS FOR TBL EXCITATION COMPARISON BETWEEN STEEL AND

COMPOSITE PANELS. .......................................................................................................... 53 FIGURE 57: EXAMPLE OF ASSEMBLY WITH ADHESIVE AND BOLTED JUNCTIONS (CAIGNOT 2009) ...... 55 FIGURE 58: A)FIRST MODES SHAPES OF SINGLE LAP-JOINTED CANTILEVERED BEAM,B)NATURAL

FREQUENCY VERSUS YOUNG'S MODULUS OF ADHESIVE FOR POISSON'S RATIO ν=0.3 (HE ET

OYADIJI 2001) .................................................................................................................. 55 FIGURE 59: AN EXAMPLE OF VISCOELASTIC JOINT (HOUSE 2007) .................................................. 55 FIGURE 60 : A) EXAMPLE OF THE BOMBARDIER MATLAB TOOL TO EVALUATE THE INTERIOR NOISE

(LETH 2010), B) EXAMPLE OF MULTI-EXCITATION FOR AIRCRAFT INTERIOR NOISE EVALUATION.

(YUAN 2013) .................................................................................................................... 56


7

FIGURE 61: TRANSMISSION LOSS FOR DAF EXCITATION MONOLITHIC VERSUS SANDWICH STRUCTURE

........................................................................................................................................ 65 FIGURE 62: TRANSMISSION LOSS FOR DAF EXCITATION, MONOLITHIC CFRP PANEL 2-8MM ........... 66 FIGURE 63: TRANSMISSION LOSS FOR DAF EXCITATION, IMPACT OF SYMMETRIC CFRP FACES

THICKNESS 2-5MM WITH HONEYCOMB 60MM CORE .............................................................. 66 FIGURE 64: TRANSMISSION LOSS FOR DAF EXCITATION IMPACT OF THE HONEYCOMB CORE

THICKNESS 20-100MM WITH SYMMETRIC 3MM CFRP FACES................................................ 67


8

DEFINITIONS 𝜔𝜔 Angular frequency in 𝑟𝑟𝑟𝑟𝑟𝑟/𝑠𝑠

𝑓𝑓 Frequency in Hz

𝑘𝑘 Wave number in 𝑚𝑚−1

𝑓𝑓𝑐𝑐 Critical frequency in Hz

𝑐𝑐 Sound velocity in 𝑚𝑚/𝑠𝑠

𝜌𝜌 Density 𝑘𝑘𝑘𝑘/𝑚𝑚3

𝜌𝜌𝑆𝑆 Surface density 𝑘𝑘𝑘𝑘/𝑚𝑚2

𝐸𝐸 Young’s modulus in 𝑀𝑀𝑀𝑀𝑟𝑟

𝐺𝐺 Shear modulus in 𝑀𝑀𝑀𝑀𝑟𝑟

𝜈𝜈 Poisson ratio(-)

𝐿𝐿𝑝𝑝 Sound pressure level in dB

dB Used to measure sound level with 𝑀𝑀𝑟𝑟𝑃𝑃𝑓𝑓 = 2 ∗ 10−5𝑝𝑝𝑟𝑟 for sound pressure level

dB(A) Sound pressure level with A weighting to take into account human hearing

sensibility

𝑇𝑇𝐿𝐿 Transmission loss in dB

𝜎𝜎 Radiation efficiency

𝜂𝜂 Damping loss factor

CFRP Carbon fibre reinforced plastic

GFRP Glass fibre reinforced plastic

FEM Finite Element Method

BEM Boundary Element Method

SEA Statistical energy analysis

TMM Transfer matrix method

DAF Diffused acoustic field

ROR Rain on the roof

TBL Turbulent boundary layer


9

EXECUTIVE SUMMARY

This D3.2 deliverable deals with “Noise, vibration and harshness”: it gives the impact

of using composites for railway carbody shell on vibro-acoustic properties. The

approach is based on a review of the state of art and numerical simulations to

illustrate the main results.

Drawbacks Considering the airborne transmission, noise reduction gets worse in case of panels

with less weight and more stiffness as seen in Figure 1. First, the lower mass will

decrease the insulation in the mass law region. Secondly, the higher stiffness to

mass ratio will shift the coincidence dip (fg) towards low frequencies falling near the

maximum sensibility of the ear (around 1 kHz). This will also increase the sound

radiation due to a vibration excitation.

Figure 1 : schematic impact of the light structure on transmission loss (Desmet, www.fp7-

eliquid.eu s.d.)

Advantages Composite structure are characterised by a higher damping in comparison with

metallic structures due to the viscoelastic behaviour of its constituents (core,

adhesive, skin). Damping will play a role in two regions (Figure 2). First, damping will

reduce the modal response in the resonance zone. Secondly, damping will limit the

dip in the coincidence zone.


10

Figure 2 : transmission loss for a single leaf panel-impact of the damping (Lamancusa 2000,

chap 9)

Another advantage of a composite panel is the ability to fulfil several functions:

structural, thermal, acoustic, etc... It is possible to evaluate and optimize the

properties to reach multi objectives. To improve the optimization process with an

acoustic constraint, it is necessary to account for the various vibration and acoustic

excitations like rolling noise, turbulent boundary layer, dynamic load, ..., because the

structure will respond in a different manner to each type of excitation. Figure 3 shows

that GFRP foam is better than steel only in the higher frequency range for an

acoustic excitation, but it is better over the whole frequency range for turbulent

excitation. The optimization process should also be guided by a perceptive analysis

using psycho acoustic criteria.

Finally, the ability to adapt the geometry allows controlling the local stiffness, and will

offer a better use of vibration isolators or the possibility to improve the vibration

dissipation within viscoelastic joints.


11

Figure 3: Transmission loss for acoustic (DAF) and turbulent (TBL) excitations: comparison

between steel and composite panels.

Recommendations The disadvantage of composites, due to a lower mass and a higher stiffness to mass

ratio in comparison with metallic structures, can be faced by a higher damping, a

better design of joints and a choice among many parameters to design an optimal

solution. The latter should account for the different types of excitation and acoustic

perceptive criteria.

1.00E+01

2.00E+01

3.00E+01

4.00E+01

5.00E+01

6.00E+01

7.00E+01

1.00E+02 1.00E+03 1.00E+04

Tran

smiss

ion

loss

(dB)

Frequency (Hz)

Transmission loss for DAF excitation

CFRP foam DAF (dB)

CFRP hc DAF (dB)

GFRP foam DAF (dB)

GFRP hc DAF (dB)

steel DAF (dB)

4.00E+01

5.00E+01

6.00E+01

7.00E+01

8.00E+01

9.00E+01

1.00E+02

1.10E+02

1.00E+02 1.00E+03 1.00E+04

Tran

smiss

ion

loss

(dB)

Frequency (Hz)

Transmission loss for TBL excitation

CFRP foam TBL (dB)

CFRP hc TBL (dB)

GFRP foam TBL (dB)

GFRP hc TBL (dB)

steel TBL (dB)


12

1. INTRODUCTION

Significant reductions in emissions due to transport are being demanded by

governments and policy-makers, as well as by society. Reductions in the energy

consumption of railway rolling stock are therefore an important objective. It is in this

context of ‘lightweighting’ of rolling stock that the REFRESCO (full title: ‘Towards a

REgulatory FRamework for the usE of Structural new materials in railway passenger

and freight CarbOdyshells’) project has been conceived. New materials such as

composites and light metallic alloys encourage hopes of the construction of lighter

rolling stock, which will consume less energy and help reduce the emissions of rail

transport. While composite materials have already been used in the manufacture of

parts of rolling stock, there is currently no way to certify a rail vehicle built entirely or

in large part from non-metallic materials. The overall objective of REFRESCO is to

set the framework for the implementation of new materials in the railway sector

through the evolution of certification processes for rolling stock. REFRESCO will

generate recommendations and provide the information needed to adapt the

regulatory framework of railway carbody structures to the introduction of new

materials. (UNIFE 2013).

Thanks to the benchmarking done in work package 2 and presented in deliverable

2.4, a matrix of candidate material options for possible components has been

constructed. This matrix is presented Table 1 and can be used as starting point for

work package3.

Table 1 : WP2 material benchmarking for lightweight structure in train

Inside work package 3, task 3.2 deals with noise, vibration and harshness behaviour

of composite panels which is the objective of the present report. It will point out the

advantages and disadvantages of this kind of panels and highlights the main

„TOP“ Monolithic SandwichCFRP CFRP – Toplayer

Resin: Epoxy Resin: Epoxy UD top-layers: UD or quasi-isotropic

Foam: Airex T90Honeycomb: Aramid (alternative: Aluminium*)

GFRP GFRP – ToplayerResin: Epoxy Resin: Epoxy

quasi-isotropic top-layers: quasi-isotropicFoam: Airex T90

Honeycomb: Aluminium (alternative: Aramid)

1

2


13

parameters that have to be carefully taken into account to reach the best

compromise for carbody conception. While noise and vibration can be readily

measured, harshness is a subjective quality. This subjective quality should take into

account both vibration and acoustic fields. Some analytical tools are available

reflecting human subjective impressions and belong to the field known as

"psychoacoustics".

The report will mainly focus on objective criteria such as sound insulation, radiation

and damping. It is organized as follows. Chapter 2 is dedicated to a literature review

of uses of composites in railway industry. Then chapter 3 and 4 will focus on vibro-

acoustic behaviour of a panel and tools available for its modelling. Finally chapter 5

sums up the drawbacks and advantages of composites panels in a vibro acoustic

point of view.

2. COMPOSITE FOR LIGHT WEIGHT CAR BODY STATE OF THE ART

2.1. Light weight structure in Railway

In comparison with other transport industry (automotive, aerospace, etc…) the ratio

of weight by seat in railway is relatively high as shown in Figure 1-a,. showing that

railway transport is not a light transportation mean. Moreover Figure 1-b) shows that

CO2 emissions associated with transport sector tend to increase. That’s why

reducing the weight of vehicles is one way of transporting less, without actually

reducing the effective work performed by the transportation mode.

Figure 4: a) Mass ratio by passenger for different transport means (Wennhage 2005); b)

evolution of CO2 emissions (transport 2010)

For the common structure, 35 to 40% of the total weight is the main frame and within

this 40%, only 16% are due to panels, the rest being the framework of the structure

(Figure 2).

a

b


14

Figure 5 : repartition of the carbody weight (Wennberg 2011)

In his conference paper, Lallet (Patrick 2013) describes the evolution of materials

used for the carbody and the associated design. According to Lallet (Patrick 2013), to

continue the weight saving process, structures need to be developed incorporating

multi materials such as composite solutions. Wennberg (Wennberg 2011) in his

thesis works propose a composite sandwich structures optimized with respect to

mechanical stress. This kind of panel reduces the weight considerably (Figure 3).

In the automotive industry, the use of composites has shown the ability to reduce the

weight and maintaining the rigidity of the body using composite. Boeman et al have

shown weight reduction up to 60% of a body in white for composite against a classic

metal structure (Boeman et Johnson 2002). The use of composites can in addition

introduce cost saving as shown in Figure 4 .

Figure 6 : substitution of corrugated stainless steel by composite panel (Wennberg, Stichel et

Per 2012)

Figure 7: weight and cost saving in the future using composites in transport (Robinson et

Carruthers 2005)

16%

24%60%

Repartition of the carbody weightpanels beam/framework Non structural


15

2.1.1. Composites

Compared to other industries such as aerospace, marine and automotive composite

materials are rarely used in the manufacturing of a train. The graph in Figure 5

illustrates this point in the case of a French intercity train: less than 3% are

composed of composite (see other).

Figure 8 : types of materials used in Transilien (Patrick 2013)

The use of composites is not a new topic in the railway sector, the fibre reinforced

polymer materials were used before the 50s, like the doors of the UK underground

(Composites make tracks in railway engineering 1995).

The composites are referencing to a family of materials composed of at least two

phases which are non miscible. In the rail industry the most widely used composites

are fibre-reinforced plastic (FRP)where the reinforcement may be carbon or glass

fibres. The matrix incorporating these fibres is usually a polymer. Figure 6 shows a

typical composite layer, which is called lamina, with fibres oriented in just one

direction. In this direction stiffness and strength are higher than in the other two

orthogonal transverse direction. Materials with this type of mechanical response are

called orthotropic materials.

A fibre reinforced composite laminate is made of various laminae that may have

different fibre orientations according to the desired properties (see Figure 7).

Figure 9 : example of a layer reinforce with unidirectional fibres (Wennberg 2011)


16

Figure 10 : composition of a laminate panel.(Wennberg 2011)

2.1.2. Sandwich structures

Sandwich structures include three main components, a core at the centre and two

skins, one on each side, as shown in Figure 8. The components of the skin can be

metal sheets or fibre reinforced composites described in paragraph 2.1.1 .The skins

have properties of high mechanical strength whereas he core is composed of a

material having a density much lower than the skins: the core can be a porous

material or a honeycomb component.

Sandwich panels operate in the same way as an I-beam in bending. The skins

support the in plane load and moment while the transverse shear is supported by the

core (Zenkert 1997) (see Figure 9).

The sandwich structure highly increases the stiffness and flexural strength with

negligibly increasing the total mass in comparison to a homogeneous beam with the

same total skin thickness as shown in Figure 10.

Figure 11 : sandwich structure(Wennberg 2011)

four laminae

one laminate


17

Figure 12 : comparison of stress repartition between sandwich and uniform beam (Wennberg

2009)

.

Figure 13 : mechanical properties function of core thickness(Zenkert 1997)

2.1.3. Examples of composite carbody construction

Despite the use of composites and composite structures are usually limited to parts

of the cabin or interior equipment, we can cite two examples where composites have

been used to manufacture large structures.

Figure 14 : a) TTX, b)The C20FIca by Bombardier (Wennberg 2011)

The TTX (Korean Tilting Train eXpress) The carbody of 23 m length is made with composite panels. The panels are sandwich

structure with a carbon-epoxy layer of 5 mm for skins and aluminium honeycomb of

40 mm for the core material. The structure is made in one piece of a specially

designed autoclave (see Figure 12).

a) b)


18

Properties of material used can be found in Appendix I. This carbody keeps a

metallic frame to ensure the stiffness of the structure as shown in Figure 13.

According to Kim et al this type of structure allows a 40% weight saving compare to

the common aluminium carbody (Kim, et al. 2005).

Figure 15 : a) sandwich panels and frame structure (Kim, Lee et Shin 2007) , b)modelling

process for the carbody (Jung-Seok et Jong-Cheol 2006)

Figure 16 : stainless steel framework for side and roof (Jung-Seok et Jong-Cheol 2006) and

floor (Jung-Seok, Jong-Cheol et Sang-Jin 2007)

C20Fica The C20 is developed by Bombardier, the production started in 2003 in Stockholm.

The carbody integrates composite sandwich panels at the sides, roof and floor.

a) b)


19

This helped reduce the weight and increase the circulation space by 30%. The

panels are made of stainless steel skins and polymethacrylimide foam core (see

reference (Knutton 2003)). As the TTX, C20 has a metallic frame the composite

panels are fixed on top of this frame.

Figure 17: sandwich structure for the C20 (Wennberg 2011)

2.2. Carbody functions

The carbody for railway refers to the structure supporting the load and transferring

movement. Guillemot & al (Guillemot et Gruneval 2000) show the need for a new

approach to the definition of carbodies incorporating composite structures. The

starting point of the approach is to establish a functional specification of the shell.

Indeed, it must meet multiple objectives, from mechanical load and acoustic comfort

through the fire resistance. Figure 15 gathers the main functions a carbody has to

fulfil.

Figure 18 : schematic view of functions that the carbody has to fulfil (Wennberg 2013)

The following parts of this report will focus on the mechanical and acoustic

constraints associated with the function of the carbody.


20

2.2.1. Mechanical constraints

Dealing with mechanical properties, a carbody has to support running load with a

lighter structure. Considering the dynamic behaviour of this structure, specific care

has to be taken to avoid forced excitation and resonance due to contact wheel-rail: if

a resonance occurs at these frequencies this will induce bad feeling for travellers.

Strength and vibration comfort were identified in a European standard EN 12633

Railway applications - Structural requirements of railway vehicle bodies - Part 1:

railway vehicles other than freight wagons" (Illingworth 1996).

Multi-criteria optimization studies to reduce mass keeping the structure strength and

stiffness have already delivered results. Harte et al (Harte, McNamara et Roddy

2004) divided the body into area and per area, and an optimization process is

applied to the layer thickness 0 °, + / - 45 ° and 90 °. The mechanical stresses

applied to the structure are taken from (board 1994).

Kim et al proposed an expert approach for optimizing the design of laminated

composite structures in order to obtain an optimal stiffness (Kim, Kim et Han 2004).

The studies continue with the evaluation of the first natural frequency by a numerical

and experimental approach for the TTX design validation ((Jung-Seok et Jong-Cheol

2006) and (Jung-Seok, Jong-Cheol et Sang-Jin 2007)).

Zinno et al propose a multi-level approach to optimize the design phase by

combining numerical, analytical and experimental tools. This approach was used to

design a roof for railway material to satisfy the recommendations of the European

code EN12663 (Zinno, et al. 2010).

Finally Wennberg & al also studied the optimization of composite structures for

railway carbody under different mechanical loads ((Wennberg 2011),(Wennberg,

Stichel et Per, Substitution of corrugated sheets in a railway vehicle's body strucuture

by a multiple-requirement based selection process 2012),(Wennberg, Stichel et

Wennhage 2012)). In his work, Wennberg takes the mechanical properties of the

corrugated stainless steel shell for starting points: these are minimum target to reach

for composite. These studies have shown that it was necessary to use a relatively

expensive composite to observe similar mechanical properties to the corrugated

structure. However, economies of scale with increased production should make this

structure industrialized.


21

Figure 19 : replacement of the corrugated panel for the floor by a sandwich composite panel

(Wennberg, Stichel et Wennhage 2012)

The table describes the structures of composites panels from Wennberg study. In the

table two kind of composite can be observed with honey comb and foam core. Two

cases of application are presented for floor and roof panels.

Table 2 : compositions of composite panels with honey comb or foam core and for floor or

roof applications cases

2.2.2. Acoustical constraints

As vibration comfort is a constraint for sizing the carbody, acoustic comfort is one of

the functions that composite panels must complete giving enough insulation to filter

exterior noise. Interior comfort is one of the challenges for acoustic research in

railway (Gautier, Poisson et Foy s.d.). In a 2011 report, CETIM reports a different

way to reduce noise and vibration in the railway industry (CETIM 2011).

2.2.2.1. Sources and transfer path analysis

To effectively reduce the noise, it is necessary to consider the problem as a whole.

The first step is to understand the mechanisms of noise generation. The goal is to

attempt to reduce the source level at his root. However, this approach has limitations

such as it is possible to reduce the sources but it is not possible to cancel it. Then the

second step of the vibro-acoustic analysis is to quantify the path.


22

A measurement campaign conducted in 2008 by SNCF (Poisson, et al. 2011)

enabled to quantify the main sources of noise and to identify the pathways. The

contribution of the three major global noise sources is given in Table 3.

Table 3 : contribution of the three major global noise sources in % of 𝑳𝑳𝑳𝑳,𝑨𝑨,𝒆𝒆𝒆𝒆 at 300km/h (left)

and 360km/h (right)

The frequency content of each source is given for the lower part of the carbody in

Figure 17.

Figure 20 : spectrum frequency associated to the tree main sources(Poisson, et al. 2011)

Knowledge of the source type, level and frequency content can then be integrated

into models to assess the impact of a design modification. To complete the process

of silent vibro-acoustic design, it is also necessary to work on the pathways by

integrating absorbing and damping materials that will dissipate transmitted vibro-

acoustic energy.

The study presented in (Poisson, et al. 2011) dealt with characterization and transfer

paths in the driver cabin. From the separate air-borne and structure-borne transfer

characterization, it is then possible to rank the contributions of different pathways

during phases of operation. The results of the source separation for different speed

are shown in Figure 18.


23

Figure 21 : air borne and structure borne contributions function of speed (Poisson, et al. 2011)

Difference between sound pressure measured in cab and sum of these two

contributions is called “remainder”. This remainder represents an unidentified

contribution.

For high speed operation, the airborne path is the main contribution, so the panel has

to offer an efficient transmission loss.

2.2.2.2. Acoustic as a constraint in the end of the design

The simplest design of a composite panel is to add a layer to perform a function. The

first step is to ensure that the mechanical stresses are supported by the composite

structure. The second step is to add acoustic material like foam to dissipate

acoustical energy. Then there is a compromise to find. On one hand we look for a

panel with minimum weight, on the other hand we add material to achieve sufficient

acoustic performance.

The work on the Shikansen 700 (Matsumoto, Masai et Wajima 1999) and the work

presented in (s.d.) illustrate this approach.

2.2.2.3. Acoustic as constraint on the beginning of the design process

Figure 22: multi-objectives conception process for a sandwich panel (Wennberg 2013). D3.2 – Proposal of NVH strategy for structural composite parts based on material properties

24

An alternative to the design process is to think the panel as a global part that has to

fulfil multiple objectives. A multilayer panel is then replacing by a sandwich panel.

In an automotive context, Cameron (C. J. Cameron 2011) showed the possibility of

reconciling the apparently antagonistic aspects of reducing mass keeping a stiffness

equivalent to a metallic panel with sufficient sound insulation for cabin comfort

performance. This objective was achieved by addressing the panel as a system and

proceeding iteratively. At each iteration all the performances are evaluated and the

optimum is found jointly and not for a parameter, then another etc...

This kind of design process with multi objectives has already been done in the

railway industry.

Acoustic criteria in the design phase have been studied by Wennhage in his thesis

(Wennhage 2005). The acoustic constraint consideration relates to a level of

transmission loss through a flat panel, the different layers of the sandwich panel

being regarded as homogeneous and isotropic. The writer includes a constraint on

the acoustic to a local part of the carbody, while the mechanical strength is evaluated

with a global model of the carbody, particularly the first Eigen frequencies are

evaluated.

Figure 23 : optimization process for strength and acoustic insulation target (Wennhage 2005)

In the perspectives of this study, Wennhage recalls the potential of fibre reinforced

plastics with the aim to reduce structural weight.

The use of fibre reinforced materials was approached by Wennberg with the

objective to replace the corrugated panels. The structure optimization was initially

designed to reach the objectives of mechanical stresses presented in the previous

paragraph (Wennberg, Stichel et Per 2012). The author then integrated acoustic

constraint in the design process of the composite panels (Wennberg et Stichel 2014).

The entire optimization process is shown in Figure 21.


25

Figure 24: design process for a composite panel with an acoustic constraint (Wennberg et

Stichel 2014)

The process includes an evaluation of the overall 3D Finite Element structure in

regards to mechanical strength and a 1D acoustic constraint with a target on the

transmission loss (Figure 22). The author models the porous layer with a formulation

of Biot (Allard et Attala 2009) and the other layers are considered as solid elastic. To

account for the anisotropy of fibre-reinforced materials, Wennberg calculates at each

step of the optimization process two coefficients for the transmission loss and retains

the best case.

Figure 25: acoustic constraint performed for a broad band sound reduction index

level(Wennberg et Stichel 2014)

2.2.3. Objectives of the report

The objective of this report is to review the general laws governing the vibro-acoustic

behaviour of the panels and in particular single or multi walled composite panels.

The next section of this report will first describe the main physical phenomena

responsible for the sound transmission trough a panel. Specificities of composite will

be highlighted and then the report will presents the modelling procedure of this kind


26

of panels. Efficient models are of great importance to enable computing the vibro

acoustic behaviour of composite panels to be taken into account in the early design

stage.

3. VIBRO-ACOUSTIC BEHAVIOUR OF A PANEL, SINGLE AND DOUBLE LAYER

Before describing the main calculation tools available to predict the behaviour of

composite panels, it is important to recall the basics of vibro-acoustic for panel from

radiation efficiency to transmission loss for a single and double wall panels. The

vibro-acoustic behaviour is a topic that has widely been studied and one can refer to

books like (Fahy 1985)(Lesueur 1988).

In the two first paragraphs, the basic concepts are presented for isotropic panel. This

will enable to stress the main physical parameters and how they influence the

efficiency of the panel with objectives of the optimal sound insulation and weight

saving. The third part is dedicated to composite panel and what are their specificities

for sound insulation.

3.1. Radiation efficiency

The radiation is the ability of a panel to convert vibration into sound. There are a lot

of works dealing with the radiation of a simple panel that. First studies focus on the

infinite panel and enabled to highlight the critical frequency which corresponds to a

double coincidence in space and frequency domain between the bending waves in

panel and the acoustical waves. This frequency is defined by:

Equation 1: 𝑓𝑓𝑐𝑐 = 𝑐𝑐2

2∗𝜋𝜋 �𝜌𝜌𝑆𝑆𝐷𝐷

with 𝑐𝑐 the sound velocity in the fluid domain (air), 𝐷𝐷 the bending stiffness define by

𝐷𝐷 = 𝐸𝐸ℎ3

12(1−𝜐𝜐2) , 𝐸𝐸 is the Young modulus and 𝜈𝜈 is the Poisson ratio.

At this specific frequency, the bending wave length in the panel and the acoustic

wave length are the same. At this frequency, there is a strong spacial coupling and

the radiation of the panel is maximum.

Wallace(Wallace 1972) studied the radiation of simply supported plate and defined

modal radiation factor as follows:


27

Equation 2: 𝜎𝜎𝑚𝑚𝑚𝑚 = 𝑊𝑊𝑊𝑊𝑚𝑚𝑚𝑚𝜌𝜌𝑐𝑐𝑆𝑆⟨𝑉𝑉𝑚𝑚𝑚𝑚

2 ⟩

This factor is the ratio of the radiated power of the mn mode over the power radiated

by a piston of the same area. For each mode there exists a critical frequency so as

the mode radiates little power under the critical frequency and radiates all after the

critical frequency. As for infinite panel this frequency corresponds to a coincidence

bending wave of the mode mn and the acoustic wave.

Equation 3 :𝑓𝑓𝑐𝑐 = 𝑐𝑐2𝜋𝜋��𝑚𝑚𝜋𝜋

𝐿𝐿𝑥𝑥�2

+ �𝑚𝑚𝜋𝜋𝐿𝐿𝑦𝑦�2

with, 𝐿𝐿𝑥𝑥 and 𝐿𝐿𝑦𝑦 the dimensions of the plate.

3.2. The transmission loss

The most used criteria to quantify the sound transmission trough a panel is the

transmission loss. It is frequency depend and defined by:

Equation 4: 𝑇𝑇𝐿𝐿(𝑓𝑓) = 𝛱𝛱𝑖𝑖𝑚𝑚𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑚𝑚𝑖𝑖𝛱𝛱𝑖𝑖𝑡𝑡𝑡𝑡𝑚𝑚𝑡𝑡𝑚𝑚𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖

The transmission loss is given by the ratio between the incident and the transmitted

sound power, so the higher the transmission loss is the better the acoustic insulation

is.

3.2.1. Single wall panel

Here we consider a single wall panel like monolithic, metallic or composite plate. For

a finite size panel, Figure 23 is a typical curve for the transmission loss evolution with

frequency. One can see four regions associated with a specific kind control path.


28

Figure 26 : evolution of the transmission loss with frequency (Griese 2012)

Stiffness controlled: At frequencies well below the first natural frequency, the stiffness of the panel

dominates its sound transmission characteristics. In this region there is a -6

dB/octave slope of TL. This range ends with the appearance of the first resonance

frequency of the panel.

Resonance controlled: In this region, the modal behaviour of the panel influences the transmission. The

resonant frequencies depend on the material, size, shape, and mounting parameters

of the panel. At these driving frequencies, due to the high level of vibration of the

panel, high amounts of sound energy are transferred to the transmitted side and

there are noticeable dips.

Damping contributes to reduce this effect.

Mass controlled: In this region, the panel mass per unit area controls the transmission loss.

Considering an acoustic wave with incidence 𝜃𝜃 , travelling in a fluid of density 𝜌𝜌 and

with a speed 𝑐𝑐, the transmission loss is given by

Equation 5: 𝑇𝑇𝐿𝐿𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝐿𝐿𝑚𝑚𝑚𝑚(𝑓𝑓,𝜃𝜃) = 20𝑙𝑙𝑙𝑙𝑘𝑘 (𝜋𝜋𝜋𝜋𝜌𝜌𝑡𝑡𝜌𝜌𝑐𝑐

𝑐𝑐𝑙𝑙𝑠𝑠(𝜃𝜃))


29

Coincidence controlled: At higher frequencies, bending waves can result in what is known as the coincidence

effect. At coincidence the acoustical and the forced structural wavelengths are equal

and the panel becomes completely transparent to sound transmission.

For infinite, homogeneous plates, the coincidence frequency for a given incidence

angle 𝜃𝜃 can be evaluated in terms of plate structural parameters:

Equation 6: 𝑓𝑓𝑐𝑐 = 𝑐𝑐2

2𝜋𝜋𝑚𝑚𝜋𝜋𝑚𝑚𝜋𝜋 �𝜌𝜌𝑡𝑡𝐷𝐷

The minimum coincidence frequency is called critical frequency and is given for 𝜃𝜃 =

0. At coincidence, the amplitude of the dips is controlled by the damping ratio of the

structure.

Figure 27 : coincidence phenomenon a) plate and sound wave numbers as function of

frequency; b) coincidence between sound and bending wavelengths (Chazot 2006)

3.2.2. Double wall

Here we consider double wall panel like sandwich or two plates separated by an air

gap or by a foam or honeycomb layer. Works of London on infinite double panel

(London 1950) show the efficiency of this kind of structure to increase the

transmission loss compared to the single wall structure. This is due to fluid phase

between the two panels that uncouples the panels at medium-high frequency.

However at the double panel resonance frequency, the breathing phenomenon

deteriorates the sound insulation of the double panel.

At low frequencies, the double wall can be seen as a simple wall with mass 𝑚𝑚 =

𝑚𝑚1 + 𝑚𝑚2. When the frequency increases, the structure can be simulated by a simple

b

a


30

mass-spring-mass system. The rigidity of the spring is given by 𝐾𝐾 = 𝜌𝜌𝑐𝑐2

𝑑𝑑𝑐𝑐𝑐𝑐𝑚𝑚𝜋𝜋2 , where d

is the inner distance between panels and 𝜃𝜃 the angle of incidence of plane wave. If

the rigidity and the damping of the panel are neglected, the frequency for the double

panel resonance is given by:

Equation 7: 𝑓𝑓2𝑃𝑃 = 12𝜋𝜋 �𝐾𝐾

𝑚𝑚1+𝑚𝑚2𝑚𝑚1𝑚𝑚2

If the frequency of the sound wave excited the panel is higher than 𝑓𝑓2𝑃𝑃, the air gap

avoids the vibration transmission to the second panel.

Another phenomenon can happen in the air cavity. Successive reflections may occur

inside the air layer, generating stationary waves, this phenomenon occurs when the

thickness is equal to half the wavelength and multiples:

Equation 8: 𝑟𝑟 = 𝑛𝑛 ∗ 12𝑐𝑐𝜋𝜋

The phenomena described before can be observed by dips in the transmission loss

as shown in Figure 25. With composite material having a poro-elastic material (foam)

as core, the impact of the resonant cavity can be damped thanks to dissipation of

acoustic energy through visco-thermal effects.

Figure 28: typical transmission loss of a simple supported double panel (Fahy 1985)

3.3. Sound insulation and excitation type.

For the transmission loss, the kind of excitation will be of great importance to

highlight phenomena previously mentioned.


31

3.3.1. Oblique wave and diffuse acoustic field

From paragraph 3.2 , the transmission loss depends on the angle of incidence of the

plane wave. As can be seen in Figure 26, this angle impacts the critical frequency

and the double wall resonant phenomena.

The diffuse field is defined as an infinite number of randomly distributed

incoherent plane waves. It is approximated as a superposition of incoherent plane

waves whose amplitudes are equal (it is assumed that each plane wave has an

amplitude per unit solid angle equals to one), and whose angles of incidence vary

equiprobably between two limit angles, Theta min and Theta max.

Figure 29: transmission loss of an infinite a) single wall partition b) double wall for various

angles of incidence (Tewes 2005)

Figure 30 : a) oblique plane wave excitation, b) diffuse acoustic field (ESI 2014)

3.3.2. Turbulent Boundary Layer

Turbulent Boundary Layer (TBL) excitation represents boundary layer flow loaded

on the surface of a structure. For this excitation, the spatial distribution of the

source governs the total acoustic force acting on the structure, and its

consequent vibration. The requirement for the analysis and prediction of

boundary layer induced noise is a description of the pressure distribution on the

b

a


32

surface which is then linked to an acoustic/structural model that determines the

sound levels which are transmitted through the structure.

3.3.3. Rain on the roof

The rain on the roof is a point force or point load excitation where all locations are

equally probable. It is defined by amplitude and phase, which can be constant or

associated to a spectrum.

A study (Maury, Gardonio et Elliot 2002) compared the diffuse acoustic field, the

turbulent boundary layer and the rain on the roof excitation using a wave number-

frequency approach. Figure 28 shows for a given frequency the wave number

spectra for the three random excitation fields associated with the same mean

square wall pressure. For the DAF the energy content is limited in the wave

number domain to �− 𝜔𝜔𝑐𝑐0

, 𝜔𝜔𝑐𝑐0� , with 𝜔𝜔 the excitation frequency. For the TBL

excitation the main energy in the wave number spectrum is concentrated near

𝑘𝑘𝑦𝑦~ 𝜔𝜔𝑈𝑈𝑖𝑖

, with 𝑈𝑈𝑐𝑐the convective flow velocity. For the rain to the roof excitation, the

level is very low compared to others since it is distributed over the entire

wavenumber frequency range.

Figure 31: Wave vector spectra for different types of random excitation as a function of the

stream-wise wave number: TBL, DAF, Rain on the roof (from

reference (Maury, Gardonio et Elliot 2002)

3.4. Specificities of composite panels

The major components that affect the sound transmission capabilities of a panel are

its stiffness, weight, damping, and resonant properties.


33

Due to the high stiffness of composite panels the resonance controlled region is

shifted in the medium frequency range and the panel can easily vibrate and radiate

noise. Figure 29 shows how the stiffness impacts the radiation efficiency of a typical

panel.

Figure 32 : impact of the panel’s stiffness in the radiation efficiency (Fahy 1985)

Due to the high ratio stiffness over mass the critical frequency shifts to lower

frequency in comparison with metallic panel. The coincidence zone is wider due to

different kind of wave propagating in the skins and the core materials.

Finally composite structure is a lightweight kind of panel: then the common principle

(Figure 30) of the heaviest structure to reach a good insulation cannot be applied.

However, given the application field and the frequency range of interest it is possible

to combine core and skin materials and microstructure to obtain the best structure

performance. A key role in the design phase is thus played by the numerical

modelling tools. A reliable model allows the investigation of different panel

parameters combination and the evaluation of the global panel performance. In the

vibro-acoustic framework, the numerical model should be able to predict both

dynamic and acoustic behaviour as well as the coupling of them.


34

Figure 33: conceptual schema for lightweight structure design (Moschini 2014)

3.5. Experimental analysis of composite structure

Dealing with lightweight structures makes the experimental activity more challenging,

especially in terms of dynamic characterization.

This section describes usual measurements that could be done on composite panel.

They are the radiation efficiency, the damping loss factor and the transmission loss.

Radiation efficiency

The radiation efficiency 𝜎𝜎 is defined as the proportionality between radiated sound

power Π and the surface normal velocity averaged ⟨𝑣𝑣2⟩ and radiating surface:

Equation 9: Π = σ. ρ0cS⟨v2⟩. The panel is excited with an electro mechanical shaker; the velocity is measured and

averaged over the radiated face. The sound power is measured with an intensity

probe by scanning over the surface.

Damping loss factor The damping loss factor of the panels placed on the measurement window can be

measured using the decay rate method (DRM). The excitation is performed using an

electro-mechanical shaker and results are averaged over N random excitation

locations and M randomly located points over the panel surface. Figure 31 shows the

experimental test bench with the electro-mechanical shaker. The response of the

panel is record thanks to a laser vibrometer or accelerometers.

DRM is based on the decay of vibration signals when the excitation is turned off. Two

assumptions are made: damping follows an exponential decay and all modes in a


35

third-octave band present the same damping loss factor. The damping loss factor is

given by

Equation 10: ηDRM = DR27.3 f

DR is the slope of the decay in units dB/second, f is the central frequency of a given

one-third octave band.

Figure 34 : experimental investigation of panel dynamical properties (Chronopoulos 2012)

When the radiation of the panel is significant, as it is the case for most composite

structures, the damping added by the radiation efficiency have to be subtracted to get

the in vacuo damping loss factor. This can be achieved by two measurements

(acoustic and vibratory excitations) to solve a two dimension SEA problem.

Transmission loss There are two main procedures for transmission loss measurement ISO-3140 and

ISO 15186-1:2000. The difference depends on the kind of receiving room reverberant

or anechoic so to measure the transmitted acoustic power. Next paragraph gives a

focus on ISO 15186 measurement.

The structure is fixed between a reverberant and an anechoic room using a mounting

frame. A white-noise in the frequency range of 100 Hz to 10 kHz is generated in the

reverberant room. The transmission loss of the structure is given by:

Equation 11: TL = Lp − LI − 6 − 10 ∗ log �SmS�.

Lp is the average sound pressure level in the reverberant room. LI is the averaged

intensity level over the measurement surface in the receiving room. Sm is the

measuring surface and S the area of the test specimen.


36

Figure 32 give the experimental set up for the transmission loss measurement with a

reverberant room and a semi-anechoic room.

The frequency range of measurements depends on the rooms typical dimensions.

Gagliardini et al (Gagliardini, Rolland et Guyader 1991) showed the influence of the

source and the room dimension for transmission loss measurement in low frequency.

Figure 35 : test set-up for transmission loss measurements (Van der Wal et Nilsson 2005)

4. MODELLING FOR ACOUSTICAL CRITERIA

4.1. Approaches for vibro-acoustic simulation

As seen in the previous chapter, several parameters can influence the vibro-acoustic

behaviour of a composite panel. Their influence may be investigated by modelling at

different accuracy levels (Figure 33). This section presents a review of simulation

tools dedicated to composite panel.

Figure 36: different ways and accuracy for modelling (Atalla 2012)

4.1.1. Global approach


37

Figure 37: typical FEM BEM approach for transmission loss calculation (Hufenbach, et al.

2010)

In the low frequency range, the vibro-acoustic problem may be solved by Finite

Element Method (FEM) and Boundary Element Method (BEM) (Figure 35). FEM is

dedicated to the structure and acoustic cavities, using a modal approach. For

acoustic radiation, BEM is preferred since it requires only a surface mesh. However

these techniques need extensive computational resources in the case of large

structure like carbody and high frequency.

In mid and high frequency, it is possible to use energetic approaches like SEA.

In the SEA context, a complex vibro-acoustic system is represented as an assembly

of coupled subsystems that can receive, store, dissipate and transmit energy. The

vibrational state is expressed in terms of vibrational energies of individual

components; the applied excitations are expressed in terms of input powers and the

coupling between.

Each subsystem is characterised by its energy 𝐸𝐸𝜋𝜋, the modal density 𝑛𝑛𝜋𝜋, the coupling

loss factor 𝜂𝜂𝜋𝜋𝑖𝑖, and the damping loss factor 𝜂𝜂𝜋𝜋𝜋𝜋. Figure 35 gives the SEA scheme for a

double wall panel.

SEA is very efficient since the number of degrees of freedom is the number of

subdomains. The difficulty is to determine the input parameters such as damping and

coupling loss factors. Some may be determined by experiments and some by

calculation on a subsystem by FEM or simplified approaches such as the Transfer

Matrix Method (TMM), which will be presented next section.


38

Figure 38: SEA representation of the double wall subsystem (Campolina, Atalla et Dauchez

2012)

4.1.2. Multilayer modelling

Multilayer structures are encountered in many applications, where a structure is

coupled with thermophonic insulators or a trim panel. The transfer Matrix Method is

currently used to predict the propagation of a plane wave trough a multilayer

structure. The basic concepts and procedure are presented in (Allard et Attala 2009,

240).

This method considers infinite flat panels. Each layer can be elastic solid, porous, or

fluid. A layer is represented by its transfer matrix relating dynamic variables on both

sides of the layer.

The propagating waves are described by their angle of incidence, their wave number

and the amplitudes of the incident and reflexive waves.

Figure 39: Transfer matrix representation of the transmission trough double-wall subsystem

(Kuo, Lin et Wang 2008)

The principle of this method is to use the expression of continuity for displacement

and constraint at a given position of an interface layer.

To overcome some limitation due to the assumption of infinite flat panel, Attala et al

extended the application of TMM to deal with finite dimension effects occurring at low


39

frequency (Rhazi et Atalla 2010). There are other works which take into account the

curvature of the panel as well (Mejdi, Sgard et Atalla 2013).

With the use of General laminate representation of a composite panel and the TMM,

acoustic indicators such as transmission loss can be accurately predicted at a low

computational cost. This method offers a good way to optimize the sound package at

an early phase of the composite panel design.

4.1.3. Local model for layers

In this section, we describe how the constitutive materials of a composite structure

are modelled either in FEM and TMM.

Skin The skin is an elastic or viscoelastic layer, either isotropic or orthotropic. Plate

or shell modelling may be used in some circumstances. Poroelastic foam core Due to the porosity of the foam core, a poroelastic model has to be used. Poroelastic

materials gather a large number and types of materials such as foam with open and

closed cells, fibrous or granular materials. Figure 37 shows different kind of porous

material. They are widely used as sound absorbing materials. For composite with foam core, Biot’s formulation is well adapted. The Biot model

considers the fully coupled vibro-acoustic behaviour of the poroelastic material. In

this case the solid phase is elastic and the vibration of the skeleton and the

interaction between solid and fluid phase have to be described. The theory takes into

account three different types of propagating waves: one shear wave and two

compressional waves (Biot 1956). In the literature, one can find different ways to

consider the interaction between phases (Atalla, Panneton et Debergue 1998) .

Figure 40: example of porous materials (Atalla et Panneton 2012)


40

Honeycomb material models A honeycomb core material will be homogenised as a solid elastic material. Attempts

to model the individual cells of the honeycomb should be avoided for normal

engineering analyses. So as to integrate this kind of material into a global model of

panel, it is necessary to process with a phase of homogenization and finds equivalent

elastic modules for this layer.

When defining the properties of honeycomb core, the following points should be

taken into consideration. These points enable to take into account the orthotropic

nature of honeycomb material.

The properties of honeycomb can be found in the suppliers documentation, see for

example Appendix II.

Figure 41: honeycomb scheme and associated orthotropic directions(Plascore s.d.)

5. MAIN DRAWBACKS / ADVANTAGES OF USING COMPOSITE FROM VIBROACOUSTIC POINT OF VIEW, AND HOW TO FACE THE DRAWBACKS

Considering the vibro-acoustic impact of using composite panel in lightweight

carbody conception is necessary to take into account the vibrational behaviour of the

carbody shell and the whole transfer path as shown in Figure 39. The following

paragraph will sum up the advantages and drawbacks of such a material for the NVH

issue. For drawbacks, we will list some potential countermeasures.


41

Figure 42: Noise and vibration transmission paths in a railway compartment (Carlsson 1997)

5.1. Drawback and countermeasures

5.1.1. Airborne transmission:

Drawback

Due to high stiffness to mass ratio of composites, the critical frequency decrease in

frequency compare to stainless steel panel. This can bring a dips in the transmission

loss for audible frequencies. Furthermore the bandwidth where the coincidence

phenomena occur is wider due to the orthotropic properties of the skin and because

the evolution of the wave number for the whole panel depends on the kind of wave

propagating in the panel (see Figure 40).

To sum up the drawback of composite panel considering the airborne transmission,

noise reduction gets worse in case of panels with less weight and more stiffness. The

impact of using lightweight structure on transmission loss can be visualised in Figure

41.


42

Figure 43: evolution of the wavenumber of a composite panel and asymptotic behaviour of

single layers (Atalla 2012)

Figure 44 : schematic impact of the light structure on transmission loss (Desmet, www.fp7-

eliquid.eu s.d.)

Solutions

Researchers have attempted to improve the transmission loss performance by

optimizing the mechanical properties of the panels or by increasing the structural

damping.


43

Dissipate vibro-acoustic energy by damping effect. Damping properties play a vital role in the vibroacoustic behaviour of composite

structures, especially at frequencies above the coincidence frequency, where an

increase of damping raises the sound insulation to a large extent.

Figure 45 : transmission loss for a single leaf panel-impact of the damping(Lamancusa 2000,

chap 9)

Compared with monolithic or stiffened structures, sandwich structures have a higher

damping loss factor because the viscoelastic core has a high inherent damping

capacity. When the beam or plate undergoes flexural vibration, the damped core is

constrained to have a shear deformation. The shear deformation causes the vibration

energy to be dissipated. Additionally, the normal-to-shear coupling between the core

and face sheets reduces the vibration amplitude.

Figure 46 : example of damping layer in composite panel(Desmet, www.fp7-eliquid.eu s.d.)


44

It is well known that the radiated/transmitted sound power is not sensitive to the

structure damping for the forced transmission. The damping is only effective when

the radiated/transmitted sound power is resonance dominant.

In Figure 44, one can see the impact of damping in the peaks corresponding to a

resonant frequency of the panel.

Figure 47 : impact of the damping treatment on the radiated power (Assaf, Guerich et

Cuvelier 2010)

Work on core properties and shape

Change the core material repartition Authors in (Grosveld, Palumbo et Klos 2006) propose a honeycomb sandwich

structures featuring areas where the core is removed from the radiating face sheet

disrupting the supersonic flexural and shear wave speeds that exists in the baseline

panel. These honeycomb panel structures exhibit improved transmission loss for a

pre-defined diffuse sound field excitation.


45

Figure 48 : improved transmission loss with specific honey comb design (Grosveld, Palumbo

et Klos 2006)

Change the geometry of unit cell in honeycomb Current description of the honeycomb elastic properties is done to use this core in

out of plane direction. In his thesis, Griese (Griese 2012) proposes an alternative use

of honeycomb core by changing the orientation of cells. Figure 47 shows how the

shape of the honeycomb cell impact the transmission loss in broadband frequency.

Figure 49 :In-plane honeycomb core versus out-of-plane core (Griese 2012)


46

Figure 50 : variation of the transmission loss with variation of core geometry (Griese 2012)

In the same way in (Cameron, Nordgren et Wennhage 2014) authors worked on

acoustic foam in sandwich panels and study the influence of the topology and the

impact of air gap introduction between foam and one of the skin panel. The air gap

introduces changes in impedance and discontinuities in vibration transfer paths.

Use resonant metamaterial The principle of the metamaterial is to tune a resonant network (for instance, mass-

spring) acting as a stop band in a bandwidth frequency of interest. The material

concept is presented in Figure 48.

Figure 51 : mass spring system for metamaterial realization (Desmet, www.cav.psu.edu s.d.)

The concept of mass-spring subsystem can then be inserted within the unit cell of the

core material to build the metamaterial wall.


47

Figure 52 : meta material concept (Desmet, www.cav.psu.edu s.d.)

This kind of material has been designed and tested with adding up to 15dB additional

noise reduction with negligible additional mass.

Smart structure The acoustic transmission loss can be improved by passive methods such as mass

addition or the use of sound absorbing materials. Most of these techniques generally

involve an increase in mass or volume to provide a good insulation particularly for

low and mid frequency range. However this is not compatible with lightweight

structure. Active control can complete the passive methods and reduce the sound

radiated by structure.

An active structure is an engineering structure containing sensors and actuators that,

when active, modify the response of the structure to its environment. Active

structures have great potential in the design of light-weight and high-strength

structures that are widely used in areas such as aerospace (RG 1997) and

automotive industries. There are a lot of mechanisms and strategy used to reduce

vibration of a structure: reader can refer to (Sinan 2011) for an overview.

Authors in (Ahmadian et Jeric 2001) measure the transmission loss of a plate with

active control system and compare them to the results with an undamped plate and a

plate with constrained-layer damping materials. The test results indicate that

piezoelectric actuators can increase sound transmission loss by approximately 7 dB

for a broad band input of 10 Hz -10 000 Hz.

If the active control seems to be efficient with little additive weight, one can remind

the actuator need supply power unit that could impact the global weight. Another

drawback of this technology is that active control system will impact the maintenance

process and increase the cost.


48

Figure 53 : increased transmission loss normalized with respect to added weight: for a

plate with piezoelectric sensor and actuators; for constrained layer damping. (Ahmadian

et Jeric 2001)

5.1.2. Structure borne transmission

Drawback

Noise radiation and reduction in carbon fibre reinforced plastic is significantly worse

than traditional steel body panels, even if the surface weight is kept at an equal level.

Solution

Again damping may contribute to reduce the noise radiation, but not on the entire

frequency range.

However, due to the high tensile stiffness and freedom to design the material, it is

possible to achieve very good mounting conditions for isolators. Important is to

design the mounting consoles to maximize the use of the carbon fibre structure.

5.2. Advantages

As we explain in paragraph 2.2.2, the main advantage of a composite panel, except

the low weight, is the ability to fulfil several functions: structural, thermal, acoustic,

etc... It is possible to evaluate and optimize the properties to reach multi objectives.

For example, the angle of orientation of the layers of fibre reinforced composite can

be optimized so as to increase the first Eigen frequency (Figure 51) that is directly in

relation with ride comfort.

For sandwich structure, there are a lot of combinations of skins and core properties

which can be tested to get the best sound insulation. For example, Figure 52 shows

a parametric study of the transmission loss as function of the thickness of the layers,

the mass density, the Young’s modulus and the damping loss ratio.


49

Figure 54 : evaluation of the first Eigen frequencies function of the ply angle (Wennberg

2013), a) shift in frequency b) modal deformations shape

For sandwich structure, there are again a lot of combinations of skins and core

properties which can be tested to reach enough sound insulation.

Figure 52 presents a parametric study of the transmission loss evolution taken from

(Wennberg 2013). The two studied parameters are the mass density and the

damping of the layers. It can be seen that increasing the core damping increases the

transmission loss above the critical frequency (which is around 200 Hz). Increasing

the mass will increase the critical frequency, with an effect below it (mass law) but

merely no effect above.

a

b


50

Figure 55: evolution of the transmission loss with properties of the composite

panel.(Wennberg, Stichel et Wennhage 2012)

5.3. Cases study with TMM approach

In this section, we propose some simulations to illustrate the difference between

metallic and composite structures. The simulation is based on the TMM using the

software Nova that accounts also for the finite size of the panel. This tool enables to

perform different kinds of excitation like diffuse acoustic field (DAF), turbulent

boundary layer (TBL), and rain on roof (ROR).

The panel properties will use recommendation from task 2 and literature review to

have typical range of thickness, core density etc…Table 4 sums up the propositions

of panel and parameters to test for the vibro-acoustic behaviour evaluation. Table 5

and Figure 53 give the material’s properties used for simulation.


51

Table 4: evaluation of panel vibro-acoustic behaviour with Nova

Table 5 : material properties for simulation

Figure 56 : parameters for foam with Biot’s theory

Figure 57 gives the result for transmission loss with the DAF excitation. As expected

the critical frequency for composite panels shifts down compared to steel panel and

the dips are close to the maximum hearing sensibility (see Table 6).

For composite panel made with GFRP and foam core, the breathing mode can be

observed around 1840 Hz because of the higher mass of the skin. After this

frequency the transmission loss is better than for steel panel.

Sandwich Excitation Evaluation

CFRP – Toplayer

Resin: Epoxy

top-layers: UD or quasi-isotropic

Foam: Airex T90

Honeycomb: Aramid (alternative: Aluminium*)

GFRP – Toplayer

Resin: Epoxy

top-layers: quasi-isotropic

Foam: Airex T90

Honeycomb: Aluminium (alternative: Aramid)

DAF,TBLtransmission loss,

radiated power, modal density

density (kg/m^3) damping thickness(m)E1 E2 E3 G12 G23 G13 v12 v23 v13

CFRP 46.6 44.3 44.3 4.4 4.4 4.4 0.15 0.15 0.15 1725 0.1 0.0039GFRP 25.5 22.9 22.9 3.41 3.41 3.41 0.1 0.1 0.1 1950 0.1 0.0039

HC 0.17 0.17 0 0.042 1.48 1.48 0.996 0.3 0 55 0.01 0.06steel 200 200 200 75 75 75 0.32 0.32 0.32 7841 0.007 0.0025

Young modulus (Gpa) Shear modulus (Gpa) Poisson's ratiomaterial


52

Table 6 : typical frequency function of panel’s structure

Figure 56 shows that for the TBL excitation, the impact of the internal damping of

composite is observed on the whole frequency range. The performance of the

composite panels for sound insulation can reach the steel panel level. For this kind of

excitation, the critical frequency has less impact on TL but the double panel

frequency stay clearly visible.

Figure 58: transmission loss for DAF excitation comparison between steel and composite

panels

Figure 59: transmission loss for TBL excitation comparison between steel and composite

panels.

skin core skinPhysical

phenomena frequency HzCFRP Foam CFRP Fc 2100CFRP HC CFRP Fc 1200GFRP Foam GFRP F2p 1840GFRP HC GFRP Fc 1390

Fc 5700Steel

1.00E+01

2.00E+01

3.00E+01

4.00E+01

5.00E+01

6.00E+01

7.00E+01

1.00E+02 1.00E+03 1.00E+04

Tran

smiss

ion

loss

(dB)

Frequency (Hz)


CFRP foam DAF (dB)

CFRP hc DAF (dB)

GFRP foam DAF (dB)

GFRP hc DAF (dB)

steel DAF (dB)

4.00E+01

5.00E+01

6.00E+01

7.00E+01

8.00E+01

9.00E+01

1.00E+02

1.10E+02

1.00E+02 1.00E+03 1.00E+04

Tran

smiss

ion

loss

(dB)

Frequency (Hz)

Transmission loss for TBL excitation

CFRP foam TBL (dB)

CFRP hc TBL (dB)

GFRP foam TBL (dB)

GFRP hc TBL (dB)

steel TBL (dB)


53

5.4. Role of joints on vibro-acoustic design

There are different ways to connect substructures so as to create a complex assembly like a carbody. Structures can be riveted, bolted or assembled by adhesive connections (see Figure 57). Joints are primarily designed to undergo static load, but the dynamic behavior of the joint is a topic of interest when dealing with damping or vibration decoupling. For the latter, composite materials offer the possibility to work on the local stiffness, allowing a better optimization of silent blocks (Campolina 2012, 57). Moreover vibration insulation has to be complemented by damping since the excitations are various and broadened (turbulent layer, impacts,…). Several studies show that joints may contribute significantly to the overall damping of the structure. Festjens in his thesis (Festjens 2014) focuses on riveted and bolted assembled structures. For this kind of assembly, the damping is due to friction between interfaces. He provides practical solutions for the identification and the design of reduced order models for the dynamics of assembled structures including the modeling of joints damping. Concerning composites, there has been a significant increase in use of structural adhesive joints in place of the classical mechanical junctions because they are much lighter and spread the stresses more uniformly. He et al (He et Oyadiji 2001) propose a numerical approach to deal with the influence of the adhesive properties on the free vibrations response on lap-jointed cantilevered beams (see Figure 58). They showed that depending on the mode shape and the position of joint, if the joint are relatively stiff, it can correspond to a local stiffening effect and can suffer from higher strength and fatigue effects. Finally, House (House 1997) (House 2007) points out the influence of designing junctions so as to make matching progressively the impedance between composite and viscoelastic joints (see Figure 59, we can see smooth evolution of joint's shape from the plate to the beam). This enables a gradual transfer of strain energy from composite to polymer where it can be dissipated by the viscoelastic nature of the polymer.


54

Figure 60: example of assembly with adhesive and bolted junctions (Caignot 2009)

Figure 61: a)first modes shapes of single lap-jointed cantilevered beam,b)Natural frequency

versus Young's modulus of adhesive for Poisson's ratio ν=0.3 (He et Oyadiji 2001)

Figure 62: an example of viscoelastic joint (House 2007)

a b

beam

plate joint


55

6. CONCLUSION

This report is an overview of the use of composites for railway carbody shell and the

impact for vibrations and acoustics. The first part was dedicated to a literature review

of what has already been done in the railway industry. We saw that composite for

lightweight structure as been largely studied from the component characterization to

the vibro-acoustic behaviour of the whole panel structure.

The lightweight combined with a high stiffness leads to the main drawback of

composite materials that is an efficient acoustic radiation. However, damping of

composite material, adapted geometry and ability to fulfil several functions give

opportunities to face this drawback. Some attempts to optimize the vibroacoustic

behaviour together with other criteria, such as thermal insulation, strength, … have

been done. They are still limited to computation of partial criteria, like transmission

loss taking into account only the acoustic part of the excitations.

To improve the optimization process with an acoustic constraint, we propose two

ways. The first is to consider the structure borne part of the interior noise: the

radiation efficiency must be accounted for in the optimization process, turbulent

boundary layer, rain on the roof excitation could be another way to excite this path.

The second way is to evaluate the harshness with psycho acoustic criteria [64](

Figure 60-a). This process of multi objective optimization has already been done for

the composite structure of aircraft (Figure 60-b). In the railway industry, one can refer

to work done on high speed train (Yang, Men et Wang 2014), (Hu, et al. 2014) to

explore acoustic indicator to optimize the ride comfort.

Figure 63 : a) example of the Bombardier Matlab tool to evaluate the interior noise (Leth

2010), b) example of multi-excitation for aircraft interior noise evaluation (Yuan 2013)

a

b


56

BIBLIOGRAPHY

Ahmadian, M, et K.M Jeric. «on the application of the shunted piezoceramics for

increasing acoustic transmission loss in structures.» Journal of sound and vibration 243, n° 2 (2001): 347-359.

Allard, JF, et N Attala. Propagation of sound in porous media. Wiley, 2009. Assaf, S, M Guerich, et P Cuvelier. «Vibration and acoustic response of damped

sandwich plates immersed in a light or heavy fluid.» Computerd ans structures 88 (2010): 870-878.

Atalla, N. «Modeling the vibroacoustics response of sandwich-composite structures with attached noise control materials.» DYNACOMP. 2012.

Atalla, N, et R Panneton. «Noise Control Materials: characterization and modeling.» inter noise 2012. New york, 2012.

Atalla, N, P Panneton, et P Debergue. «A mixed displacemnt-pressure formulation for poroelastic materials.» Journal of Acoustical Society Of America 104, n° 3 (1998): 1444-1452.

Biot, Ma. «theory of propagation of elastic waves in a fluid-filled-saturated porous solid, I Low frequency range.» Journal of Acoustical Society of America 28 (1956): 168-191.

board, British railways. Structural requierements for rail vehicle bodyshell. GM/TT0265, 1994.

Boeman, Raymon G, et Nancy Johnson. «Development of a cost competitive, composite intensive, body-in-white.» 01 2002. http://web.ornl.gov/~webworks/cppr/y2001/pres/113371.pdf

Caignot, A. «Prédiction par essais virtuels de l'amortissement dans les structures spatiales.» Cachan, 2009.

Cameron, Christopher J. «Design of multifunctional body panels for conflicting structural and acoustic requirements in automotive applications.» Stockolme, 2011.

Cameron, J.C, E.L Nordgren, et P Wennhage. «On the balancing of structural and acoustic performance of a sandwich panel based on topology, property and size optimization.» Journal of sound and vibration 333 (2014): 2677-2698.

Campolina. «Modélisation vibro-acoustique de double parois aeronautiques avec liens mécaniques par la méthode de l'analyse statistique énergétique.» doctoral thesis, 2012.

Campolina, B, N Atalla, et N Dauchez. «Assessment of the validity of statistical energy analysis and transfer matrix method for the prediction of sound transmsission loss trough aircraft double-walls.» Acoustics 2012. Nantes, 2012.

Carlsson, U. «Acoustical and vibrational aspects of future railway vehicles,technical report.» 1997.

CETIM. «Réduction des bruit et vibrations:Note de veille WCRR2011.» Senlis, 2011. Chazot, J.D. «Transparence acoustique de doubles parois remplies de matériaux

granulaires, PhD thesis.» Lyon, 2006. Chronopoulos, Dimitrios. «prediction of the vibroacoustic response of aerospace

composite structures in broadband frequency range.» Lyon, 2012.


57

«Composites make tracks in railway engineering.» Reinforced Plastics 39, n° 11 (1995): 20-26.

«Countermeasures for noise from the lower car body with sound absorberIdo,Atsushi;Kurita,Takeshi;Horiuchi,Masahiko.» East Japan Railway Technical review,Specila edition 7: 23-26.

Desmet, Wim. «www.cav.psu.edu.» http://www.cav.psu.edu/workshops/2014/Desmet

Desmet, Wim. «www.fp7-eliquid.eu.» https://www.fp7-eliquid.eu/members/wp-documents/2013-leuven-marie-curie-graduate-school-on-vehicle-dynamics-and-mechatronics/Lecture, 2010

ESI. Nova users guide. ESI group, 2014 Fahy, F. Sound and structural vibration:Radiation Transmssion and Response.

Londre: Acadelic, 1985. Festjens, Hugho. «Contribution à la caractérisation et à la modélisation du

comportement dynamique des structures assemblées.» Doctoral thesis, 2014. Gagliardini, L, J Rolland, et J.L Guyader. «The use of a functionnal basis to calculate

the acoustic transmission between rooms.» journal of sound and vibration 145, n° 3 (1991): 457-478.

Gautier, Pierre-Etienne, Franck Poisson, et Florence Foy. «Les enjeux actuels des recherches en acoustique et vibrations dans le domaine ferroviare.» Acoustique&Techniques 37: 6-11.

Griese, D. Finite element modeling and design of honeycomb sandwich panels for Acoustic performance. Clemson: Clemson University, 2012.

Grosveld, F.W, D.L Palumbo, et J Klos. «Finite element development of honeycomb panel configurations with improved transmission loss.» inter noise 2006. honolulu, 2006.

Guillemot, Jean-Michel, et Yves-Henri Gruneval. «Les composites pour les strucutres ferroviaires: Nécessite d'une nouvelle approche.» Mécanique &industrie 1 (2000): 223-234.

Harte, A.M, J.F McNamara, et I.D Roddy. «A multilevel appraoch to the optimisation of a composite light rail vehicle bodyshell.» Composites Structures 63 (2004): 447-453.

He, X, et S.O Oyadiji. «Influence of adhesive characteristics on the transverse free vibration of single lap-jointed cantilevered beams.» Materials Processing Technology, 2001: 366-373.

House, J.R. «Energy absorbing composite joints and their application to noise reduced sonar domes.» Eleventh international conference on composite materials. Murray L.Scott, 1997. 74-82.

—. «The role of polymeric materials for effective structural damping.» Polylers in Defence and aerospace Applications. Toulouse, 2007. paper20.

Hu, K, Y Wang, H Guo, et H Chen. «sound quality evaluation and optimization for interior noise of rail vehicle.» advances in mechanical engineering 2014 (2014): 1155-1165.

Hufenbach, Werner A, Denis Blanchet, Maik Gude, Martin Dannemann, et Frank Kolbe. «Integrated Vibro-acoustic Calculation Procedure of Composites structure using coupled Finite and Boundary element method.» ISMA. 2010.

Illingworth, R. «Structural Requirements for railway vehicles.» Railtrack PLC, 1996.


58

Jung-Seok, Kim, et Jeong Jong-Cheol. «Natural frquency evaluation of a composite train carbody with lenght of 23m.» Composites Science and Technology 66 (2006): 2272-2283.

Jung-Seok, Kim, Jeong Jong-Cheol, et Lee Sang-Jin. «Numerical and experimental studies on the deformational behaviora composite train carbody of the Korean tilting train.» Composite Strucutres 81 (2007): 168-175.

Kim, J.S, S.J Lee, et K.B Shin. «Manufacturing and structural safety evaluation of a composite train carbody.» Composite structure 78 (2007): 468-476.

Kim, Jung-Seoj, Nam-Po Kim, et Seong-Ho Han. «Optimal design of composite laminates for a train carbody by an expert system and enumeration method.» Composites Strucutures 68 (2004): 147-156.

Kim, S-H, S-G Kang, C-G Kim, et K-B Shin. «Analysis of the composite strucuture of tilting train express.» international conference on composite. Durban, 2005.

Knutton, M. «Bombardier develops' "game changing" technology.» International railway journal 43, n° 5 (2003): 28.

Kuo, Y-M, H-J Lin, et C-N Wang. «Sound transmission across orthotropic laminates with a 3D model.» Applied Acoustics 69 (2008): 951-959.

Lamancusa, J.S. Noise control. Penn State, 2000. Lesueur, C. Rayonnement acoustique des structures:vibro-acoustique,interactuion

fluide-structure. Paris: Eyrolles, 1988. Leth, S. «Silent trains for the future-an industrial approach for low noise design.»

ISMA. 2010. London, A. «Transmission of reverberant sound through double walls.» Journal of

Acoustical society of America 22 (1950): 270-279. Matsumoto, Masakazu, Kentaro Masai, et Takenori Wajima. «New technologie for

railways trains.» Hitachi review 48, n° 3 (1999). Maury, C, P Gardonio, et S.J Elliot. «A wavenumber approach to modelling the

response of randomly excited panel,PartI:General theory.» Journal of sound and vabriation, 2002: 83-113.

Mejdi, A, F Sgard, et N Atalla. «Validity of the transfer matrix method of the transmsission loss of curved panels.» ICA. Montreal, 2013.

Moschini, S. «Vibroacoustic modeling of sandwich foam core panels.» Milan, 2014. Patrick, Lallet. «Des voies dévolution pour les matériaux utilisés dans le matériel

roulant.» journées annuelles de la SF2M. Villeneuve d'ascq, 2013. Plascore. «http://www.plascore.com/pn1-nomex-aramid-fiber-honeycomb.php.»

http://www.plascore.com/pdf/Plascore_PN1.pdf Poisson, Franck, Olivier Coste, Sandrine Bouvet, et Nicolas Vincent. «Transfer path

analysis within the TGV Duplex driver cab, platforms, upper and lower rooms from 150 to 360 km/h.» WCRR Lille 2011. Lille, 2011.

RG, Loewy. «Recent developpments in smart structures with aeronautical applications.» smart materials and structures 6, n° 5 (1997): 11-42.

Rhazi, D, et N Atalla. «Transfer matrix modeling of the vibroacoustic response of multi-materials structures under mechanical excitation.» Journal of sound and vibration 329 (2010): 2532-2546.

Robinson, Mark, et Joe Carruthers. «Lightweighting of Rail Vehicles Using Composite Materials.» 2005. http://www.compositesuk.co.uk/


59

Sinan, Korkmaz. «a review of active structural control: challenges for engineering informatics.» computers and structures 89 (2011): 2113-2132.

Tewes, Stephan. «Active Trim Panel Attachments for Control of Sound transmission trough aircraft structure.» Munich, 2005.

transport, directorate-general for energy and. Eu energy in figures 2010 co2 emissions by sector, technical report. European comission, 2010.

UNIFE. «Annual report 2013.» 2013. Van der Wal, H.M.M, et A.C Nilsson. sound transmission measurement on

composite and metal fuselage panels for different boudary conditions. Stocholm: National aerospace laboratory, 2005.

Wallace, C.E. «Radiation resistance of a rectangulare plate.» Journal of Acoustical society of America 51, n° 3 (1972): 946-950.

Wennberg, David. «A light Weight Car body for High-Speed Trains, literature study.» Stockolme, 2009.

Wennberg, David. «Light-Weight Methodology in Rail Vehicle Design through Introduction of Load Carrying Sandwich Panels;these de licence.» Stockolme, 2011.

Wennberg, David. «Multi-functional composite Design Concepts for Rail Vehicle Car bodies, these de doctorat.» Stockolme, 2013.

Wennberg, David, et Sebastian Stichel. «Multi-functional design of a composite high-speed train body structure.» Structural Muldisciplinary Optimization 50 (2014): 475-488.

Wennberg, David, Sebastian Stichel, et Per Wennhage. «Optimisation of a sandwich sandwich panels for the load carrying structure of high speed rail vehicles.» International Journal of Aerospace ligth structure 2, n° 1 (2012).

Wennberg, David, Sebastian Stichel, et Wennhage Per. «Substitution of corrugated sheets in a railway vehicle's body strucuture by a multiple-requirement based selection process.» the Institution of Mechanical Engineers,Part F:Journal of Rail and Rapid Transit. 2012.

Wennhage, Per. «Structural Acoustic Optimization of Sandwich Panels;these de docotorat.» Stockolme, 2005.

Yang, L, F Men, et T Wang. «Sound quality evaluation analysis on the interior noise of high speed train.» adavnces in applied acoustics 3 (2014): 9-14.

Yuan, Chongxin. «Mechanical and vibro-acopustic aspects of composite sandwich cylinders.» Beihang, 2013.

Zenkert, D. The handbook of Sandwich constructions. EMAS publishing, 1997. Zinno, A, E Fusco, A Prota, et G Manfredi. «Multiscale approach for the design of a

composite sandwich structures for train application.» Composites Structures 92 (2010): 2208-2219.


60

APPENDIX I: COMPOSITE STRUCTURE FROM LITERATURE REVIEW

2.1.3 Composition of the TTX

2.1.3 composition of the C20 FICA


61

From Harte&al(Harte, McNamara et Roddy 2004)

From Zinno&al (Zinno, et al. 2010)


62

From Wennberg (Wennberg, Stichel et Per 2012)


63

APPENDIX II: DATASHEET FOR HONEYCOMB

Data coming from (Plascore s.d.)


64

APPENDIX III: PARAMETRIC STUDY

Figure 61 to Figure 64 show an example of a parametric study for the transmission

loss and a diffused acoustic field excitation.

Figure 61 shows how a sandwich structure adds stiffness compare to monolithic

configuration as explain in 2.1.2.

For sandwich structures, Figure 63 shows that the face thickness improves

transmission loss in the mass control area and in Figure 64 we show that the core

thickness impact the global stiffness of the panel and the coincidence frequency

decrease but there is few modification in the mass control area.

Figure 64: transmission loss for DAF excitation monolithic versus sandwich structure

1.00E+01

2.00E+01

3.00E+01

4.00E+01

5.00E+01

6.00E+01

7.00E+01

1.00E+02 1.00E+03 1.00E+04

Tran

smiss

ion

loss

(dB)

Frequency (Hz)


CFRP foam DAF (dB)

CFRP hc DAF (dB)

GFRP foam DAF (dB)

GFRP hc DAF (dB)

steel DAF (dB)

monolithic GFRP DAF (dB)

monolithic CFRP DAF (dB)


65

Figure 65: transmission loss for DAF excitation, monolithic CFRP panel 2-8mm

Figure 66: transmission loss for DAF excitation, impact of symmetric CFRP faces thickness 2-

5mm with honeycomb 60mm core

1.00E+01

2.00E+01

3.00E+01

4.00E+01

5.00E+01

6.00E+01

7.00E+01

1.00E+02 1.00E+03 1.00E+04

Tran

smiss

ion

loss

(dB)

Frequency (Hz)

Transmission loss for DAF excitation: CFRP monolithic panel

CFRP2mmDAF (dB)

CFRP6mm DAF (dB)

CFRP8mm DAF (dB)

10

20

30

40

50

60

70

1.00E+02 1.00E+03 1.00E+04

Tran

smiss

ion

loss

(dB)

Frequency (Hz)

Transmission loss for DAF excitation: impact of the face thickness

CFRP2 HC DAF (dB)

CFRP3 HC DAF (dB)

CFRP4 HC DAF (dB)

CFRP5 HC DAF (dB)


66

Figure 67: transmission loss for DAF excitation impact of the honeycomb core thickness 20-

100mm with symmetric 3mm CFRP faces

1.00E+01

2.00E+01

3.00E+01

4.00E+01

5.00E+01

6.00E+01

7.00E+01

1.00E+02 1.00E+03 1.00E+04

Tran

smiss

ion

loss

(dB)

Frequency (Hz)

Transmission loss for DAF excitation: impact of the core thickness

CFRP3 HC20 DAF (dB)

CFRP3 HC40 DAF (dB)

CFRP3 HC60 DAF (dB)

CFRP3 HC80 DAF (dB)

CFRP3 HC100 DAF (dB)


67

D3.2 Noise vibration and harshness

Documents

Transcript of D3.2 Noise vibration and harshness