2004SAEZoroufiFatemi2004-01-0628

7/31/2019 2004SAEZoroufiFatemi2004-01-0628

1/11

1

2004-01-0628

Fatigue Life Comparisons of Competing ManufacturingProcesses: A Study of Steering Knuckle

Mehrdad Zoroufi and Ali FatemiThe University of Toledo

Copyright 2003 SAE International

ABSTRACT

A vehicle steering knuckle undergoes time-varying

loadings during its service life. Fatigue behavior is,therefore, a key consideration in its design andperformance evaluation. This research program aimedto assess fatigue life and compare fatigue performanceof steering knuckles made from three materials of different manufacturing processes. These include forgedsteel, cast aluminum, and cast iron knuckles. In light of the high volume of forged steel vehicle components, theforging process was considered as base for investigation. Monotonic and strain-controlled fatiguetests of specimens machined from the three knuckleswere conducted. Static as well as baseline cyclicdeformation and fatigue properties were obtained andcompared. In addition, a number of load-controlledfatigue component tests were conducted for the forgedsteel and cast aluminum knuckles. Finite elementmodels of the steering knuckles were also analyzed toobtain stress distributions in each component. Based onthe results of component testing and finite elementanalysis, fatigue behaviors of the three materials andmanufacturing processes are then compared. It isconcluded that the forged steel knuckle exhibits superior fatigue behavior, compared to the cast iron and castaluminum knuckles.

INTRODUCTION

There has been a strong trend towards the adoption of optimum materials and components in automotiveindustry. Automotive designers have a wide range of materials and processes to select from. Steel forgingsare in competition with aluminum forgings and castings,cast iron, and sintered powder forgings. The competitionis particularly acute in the chassis, and it is not unusualto find a range of different materials and manufacturingtechnologies employed within modern chassiscomponents.

The steering knuckle, being a part of the vehiclessuspension system, has alternatives of forging andcasting as its base manufacturing process. Since it isconnected to the steering parts and strut assembly fromone side and the wheel hub assembly from the other, ithas complex restraint and constraint conditions andtolerates a combination of loads. In addition, parameterssuch as internal defects, stress concentrations andgradients, surface finish, and residual stresses can haveconsiderable influence while designing for fatigue.

A common practice of fatigue design consists of acombination of analysis and testing. A problem thatarises at the fatigue design stage of components is thetransferability of data from smooth specimens to thecomponent. The component geometry and surfacespecifications often deviate from that of the specimeninvestigated and neither a nominal stress nor a notchfactor can be defined in most cases. An advantage of component testing is that the effects of material,manufacturing process parameters, and geometry areinherently accounted for, even though synergistically.

Gunnarson et al. (1) investigated replacing conventionalforged quenched and tempered steel with precipitationhardened pearlitic-ferritic cast steels. They alsocompared toughness and machinability characteristics of forging versus casting components. They observedinsufficient toughness but higher machinability for castcomponents. Lee (2) evaluated fatigue strength for truck

axle housing, crankshaft, leaf spring, torque-rodassembly, and steering knuckle. For the case of thesteering knuckle, two sets of tests at constant loadamplitude were conducted. The load was applied to thewheel stud (spindle) and carrier tube (body), and theacceptance criteria were no crack initiation and nopermanent deformation until 2x10 5 cycles. Among fivetotal tests, none of the knuckles failed. Lee et al. (3

developed a methodology to quantitatively assessfatigue lives of automotive structures and to identifycritical and non-damaging areas for designenhancement and weight reduction. An MS-3760A cast


2/11

2

iron steering knuckle was the example component of thisstudy. The methodology combines load-time history filewith results from elastic FEA to estimate fatigue lives.Knuckle strain gage measurements were made for elastic as well as inelastic load ranges. The correlationsof maximum principal strains between the FEA and theexperimental results showed average errors of 23% and27% for lateral and fore/aft loads in elastic range,respectively. The differences between observed andpredicted lives in the inelastic range were found to befactors of 3.9 and 1.4 at the R50C50 life (the fatigue lifewith reliability of 50% and confidence level of 50%) for fore/aft and lateral loading tests, respectively.

Beranger et al. (4) assessed fatigue behavior of a forgedsuspension arm and investigated the fatigue strengthreduction of the "as-forged" surface resulting from thesurface roughness in the presence of residual stresses.FEA using shell elements and LEFM calculations wereused. For component testing, constant-amplitude load-controlled tests at 10 Hz frequency and with R = -0.5(typical braking/acceleration cycles measured onvehicle) were considered. Endurance limit was definedat 2x10 6 cycles. A multiaxial fatigue model based on acritical plane approach was implemented. The minimumvalue of the safety factor on the part was located in thearea where failure occurred on the test rig. Theyreported very good correlation between experimentaland fatigue life predictions.

To perform fatigue analysis and implement the localstress-strain approach in complex vehicular structures,Conle and Chu (5) used strain-life results, simulated 3-Dstress-strain models and multiaxial deformation paths toassess fatigue damage. After the complex load historywas reduced to a uniaxial (elastic) stress history for each

critical element, a Neuber plasticity correction methodwas used to correct for plastic behavior. Elastic unit loadanalysis, using strength of material and an elastic FEAmodel combined with a superposition procedure of eachload point's service history was proposed. Savaidis (6) verified the local strain approach for durability evaluationof forged bus axle steering arm. It was concluded thatthe local strain approach, in conjunction with the Smith-Watson-Topper and J-Integral parameters, are able torepresent and estimate many influencing factorsexplicitly. These include mean stress effects, loadsequence effects above and below the endurance limit,and manufacturing process effects such as surface

roughness and residual stresses.

Sonsino et al. (7) discussed the procedure of specimenfatigue data transferability to real components using theexample of a randomly loaded truck stub axle. S-Ncurves under constant amplitude loading and strain-lifecurves under variable amplitude loading for unnotchedand notched specimens and components werecompared. It was concluded that the same failurecriterion (i.e. first detectable crack), accuratedetermination of the local equivalent stress or strain, andthe same maximum stressed/strained material volume

for both the specimens and the components, werepreconditions for the transferability of material dataobtained from specimens to the component. Themaximum stressed/strained material volume appearedto be suitable for taking into account the statistical andmechanical size effects in a relatively simple manner.

The objectives of the current study were to comparefatigue performance and assess fatigue life for steeringknuckle, a fatigue critical part, made from three materialsof different manufacturing processes. Knuckles of threevehicles were selected. These included forged steelSAE Grade 11V37 knuckle of the rear suspension of a4-cylinder sedan weighing 2.5 kg, cast aluminum ASTM

A356-T6 knuckle of front suspension of a 6-cylinder minivan weighing 2.4 kg, and cast iron ASTM A536Grade 65-45-12 knuckle of the front suspension of a 4-cylinder sedan weighing 4.7 kg. Only the forged steelknuckle included the spindle portion. Figure 1 shows thedigitized models of the three components.

Figure 1: From left to right the digitized models of theforged steel, cast aluminum and cast iron steeringknuckles.

In this paper, first the results of specimen testing of thethree materials are presented. Monotonic, cyclic andfatigue properties are compared. Then the procedure of finite element analysis is detailed, the methods to verifythe models are described, and the critical location of thecomponents with respect to the selected boundaryconditions are identified. Details of component testing of the forged steel and cast aluminum knuckles aredescribed, and fatigue lives for the three componentsbased on S-N and strain-life predictions are compared.

MATERIAL FATIGUE BEHAVIOR ANDCOMPARISONS

EXPERIMENTAL PROGRAM

Identical flat plate specimens with square cross sectionand uniform gage section length, as shown in Figure 2,were machined from the steering knuckles. Therelatively short gage section length was chosen toprevent buckling in compression. Specimens in threegeometrical orientations were taken from the forgedsteel knuckle to investigate the effect of directionality.For cast aluminum and cast iron knuckles, since thematerial properties are less dependent on geometricalorientation, the specimens were taken from the hub and


3/11

3

one of the arms, respectively. The degree to whichanisotropic behavior may exist depends on the specificcasting practice.

The monotonic and cyclic specimen tests wereperformed by a 50 kN closed-loop servo-hydraulicuniaxial testing machine with computer control andhydraulic-wedge grips. Total strain was controlled usingan extensometer rated as class B1, according to ASTMclassification. All tests were conducted at roomtemperature. Significant effort was put forth to align theload train and minimize bending. According to ASTMStandard E606 (8) , the maximum bending strains shouldnot exceed 5% of the minimum axial strain rangeimposed during any test program, which was fulfilled. Allmonotonic tension and constant amplitude fatigue testsin this study were performed using test methodsspecified by ASTM Standard E8 (9) and ASTM StandardE606, respectively.

Specimens of the three materials were tested at strainamplitudes ranging from 0.7% to 0.125%, resulting infatigue lives between about 10 2 and 10 7 (run-out)reversals. Strain amplitudes larger than 0.7% were notpossible due to specimen buckling limitation. Straincontrol was used in all tests, except for some long-lifeand run-out tests, which were conducted in load-controlmode. For the strain-controlled tests, the appliedfrequencies ranged from 0.1 Hz to 2 Hz. For the load-controlled tests including run-out tests, the frequencywas increased to up to 30 Hz in order to shorten theoverall test duration. All tests were conducted using a

triangular waveform. Test data were automaticallyrecorded at regular intervals throughout each test.

Figure 2: Specimen geometry used for all specimentests. All dimensions are in mm.

EXPERIMENTAL RESULTS AND COMPARISONS

A summary of the monotonic properties for the threematerials is provided in Table 1, including the ratios of each property with respect to that of the forged steel.From Table 1 it can be seen that cast aluminum andcast iron reach 42% and 54% of the forged steel yieldstrength, respectively, and 37% and 57% of forged steelultimate tensile strength, respectively. The percentelongation, as a measure of ductility, for the castaluminum and cast iron are 24% and 48% of the forgedsteel, respectively.

Table 1: Summary of mechanical properties and their comparative ratios (forged steel is taken asthe base for ratio calculations).

Forged

Steel11V37

Cast

AluminumA356-T6

ratio Cast

Iron65-45-12

ratio

Monotonic Properties

Modulus of elasticity, E (GPa) 201 78 0.39 193 0.96

Yield strength (0.2% offset), YS (MPa) 556 232 0.42 300 0.54

Ultimate strength, S u (MPa) 821 302 0.37 471 0.57

Percent elongation, %EL (%) 21 5 0.24 10 0.48

Percent reduction in area, %RA (%) 37 10 0.27 25 0.68

Strength coefficient, K (MPa) 1,347 418 0.31 796 0.59

Strain hardening exponent, n 0.157 0.095 0.6 0.187 1.19

True fracture strength, f (MPa) 496 301 0.6 219 0.44

True fracture ductility, f (%) 47 10 0.23 28 0.59

Cyclic and Fatigue PropertiesCyclic modulus of elasticity, E (GPa) 195 73 0.38 169 0.87

Cyclic strength coefficient, K (MPa) 1,269 430 0.34 649 0.51

Cyclic strain hardening exponent, n 0.137 0.063 0.46 0.075 0.55

Cyclic yield strength, YS (MPa) 541 291 0.54 407 0.75

Fatigue strength coefficient, f (MPa) 1,157 666 0.58 761 0.66

Fatigue strength exponent, b -0.082 -0.117 1.42 -0.076 0.92

Fatigue ductility coefficient, f 3.032 0.094 0.03 0.864 0.28

Fatigue ductility exponent, c -0.791 -0.61 0.77 -0.771 0.97

Fatigue strength, S f @ 106 cycles (MPa) 352 122 0.35 253 0.72


4/11

4

Figure 3 represents the superimposed plots of monotonic and cyclic curves for all three materials. Ascan be seen in this figure, the forged steel has mixed-mode cyclic behavior. This material cyclically softens atlow amplitude, but slightly hardens at higher strainamplitudes larger than 0.5%. The cyclic deformationcurve of the forged steel was found to be independent of the geometrical direction (i.e. isotropic behavior) basedon the results of fatigue tests in three geometricaldirections. Cast aluminum and cast iron show cyclichardening behavior by about 20% and 30%,respectively. In addition, strain hardening is moreprominent for forged steel. The cyclic yield strengths of cast aluminum and cast iron were found to be 54% and75% of the forged steel, while the cyclic strain hardeningexponent of cast aluminum and cast iron were 46% and55% of the forged steel, respectively. These indicatehigher resistance of the forged steel to plasticdeformation.

0

100

200

300

400

500

600

700

0.0% 0.2% 0.4% 0.6% 0.8% 1.0%True Strain (%)

T r u e

S t r e s s

( M P a

)

Forged Steel 11V37Cast Iron 65-45-12Cast Aluminum A356-T6

Monotonic

Cyclic

Monotonic

Cyclic

Monotonic

Cyclic

Figure 3: Superimposed plots of cyclic and monotonicstress-strain curves for forged steel 11V37, castaluminum A356-T6, and cast iron 65-45-12.

Figure 4 shows a direct comparison of the threematerials with respect to S-N behavior. Comparison of long-life fatigue strength, S f , which is defined as thefatigue strength at 10 6 cycles, shows that fatiguestrength of cast aluminum and cast iron are 35% and72% of the forged steel, respectively. In addition, whilethe fatigue strength of forged steel at 10 6 cycles is

expected to remain about constant at longer lives,fatigue strength of cast aluminum and cast ironcontinuously drops with longer lives (i.e. beyond 10 6 cycles). Figure 4 indicates that at a given stressamplitude forged steel results in about two orders of magnitude longer life than cast iron, and more than four orders of magnitude longer life than cast aluminum. Withregards to anisotropy influence of the forged steel,fatigue test results indicated that both the long-life aswell as the short-life fatigue in the spindle centerlinedirection were longer by about a factor of two, ascompared with the other two directions. Since this

direction coincides with the primary stressing direction of the forged knuckle, it was selected as the basis for comparisons in Figure 4 and subsequent figures. Thefatigue data presented in Table 1 are also for thisdirection.

Figure 4: Superimposed true stress amplitude versusreversals to failure for forged steel 11V37, castaluminum A356-T6, and cast iron 65-45-12.

In automotive design, cyclic ductility is a major consideration when designing components subjected tooccasional overloads, particularly for notchedcomponents, where plastic deformation can occur. Thisis typical of suspension components, such as steeringknuckle. Figure 5 presents a comparison of the plasticstrain amplitude versus fatigue life behavior for the threematerials. Forged steel was found to be superior to castaluminum and cast iron with respect to low-cycle fatigue(i.e. cyclic ductility), as could be seen in Figure 5.

0.001%

0.010%

0.100%

1.000%

1E+2 1E+3 1E+4 1E+5 1E+6

Reversals to Failure, 2N f

T r u e

P l a s t i c

S t r a

i n A m p

l i t u d e

( % )


Figure 5: Superimposed true plastic strain amplitudeversus reversals to failure for forged steel 11V37, castaluminum A356-T6, and cast iron 65-45-12.


5/11

5

Comparisons of strain-life fatigue behavior of the threematerials, as shown in Figure 6, demonstrates that theforged steel provides about a factor of 20 and 4 longer lives in the short-life regime compared to the castaluminum and cast iron, respectively. In the high-cycleregime, forged steel results in about an order of magnitude longer life than the cast iron, and about afactor of 3 longer life than the cast aluminum.

0.10%

1.00%

1E+2 1E+3 1E+4 1E+5 1E+6 1E+7 1E+8


T r u e

S t r a

i n A m p

l i t u d e

( %

Forged Steel 11V37

Cast Aluminum A356Cast Iron 65-45-12

Figure 6: Superimposed true strain amplitude versusreversals to failure for forged steel 11V37, castaluminum A356-T6, and cast iron 65-45-12.

The product of strain and stress amplitudes versus life,known as Neuber plot, is shown in Figure 7. This plot isuseful when analyzing component geometries withstress concentrations, where the notch root fatiguebehavior is a function of both local stress and strain.Therefore, rather than considering the individual effects

of stress amplitude (Figure 4) or strain amplitude (Figure6), a Neuber plot considers the combined effects of bothstress and strain amplitudes.

0.1

1.0

10.0

1E+2 1E+3 1E+4 1E+5 1E+6 1E+7 1E+8


a

a (

M P a

)


Figure 7: Neuber curves for forged steel 11V37, castaluminum A356-T6, and cast iron 65-45-12.

FINITE ELEMENT ANALYSIS

Linear and nonlinear static finite element analysesemploying IDEAS-8 software were conducted on eachknuckle. Nonlinear analysis was necessary due to localyielding in most cases, as well as gross yielding in somecases, as mentioned previously. In order to generateprecise geometries of the three steering knuckles, aCoordinate Measuring Machine (CMM) was used, with

the resulting digitized models as presented in Figure 1.Material cyclic properties were used for the analysis. vonMises yield criterion and a kinematic hardening rule thatused a bilinear stress-strain curve, adequatelyrepresenting the material cyclic deformation behavior,were assumed for the nonlinear analysis.

The boundary conditions and loading were selected torepresent the component service and testing conditions.For the forged steel knuckle, the primary loading wasapplied to the spindle, and the four suspension and strutholes were restrained. The analysis showed thatchanging the location in the spindle length at which theload is applied does not affect the location andmagnitude of the stresses at the critical point. To verifythe model, other alternatives were analyzed by switchingthe loading and boundary conditions, and also byreleasing any one of the fixed points, to ensure thecritical locations remained the same. For the castaluminum knuckle in service, while the loading is appliedto the strut joints through struts, the four hub bolt holesare connected to the wheel assembly. Several trials for boundary conditions were analyzed, including fixing thewhole area of the four hub bolt holes, fixing thecenterline of the hub bolt holes, only fixing the pair of bolt holes away from the load application point, andfixing two points at the middle area of the hub. It wasfound that except for the case of fixing the bolt holes, for which the value of stress was lower and the criticallocation was different, all the other three cases providedapproximately similar results. Therefore, the choice of fixing the hub bolt hole-centerlines was selected. For thecast iron knuckle, where the geometry and serviceconditions are close to the cast aluminum knuckle,similar loading and boundary conditions were applied.

While defining a solid mesh for the components, freemeshing feature of the software was employed since ithas no geometry restrictions and it could be defined oncomplicated volumes. The free mesh generator used an

algorithm that minimizes element distortion. 3-D linear tetrahedral solid elements, with global element sizes of 3.81 mm for the forged steel and cast iron knuckles and5.08 mm for the cast aluminum knuckle were used.Local element sizes of 0.254 mm for the forged steeland 0.635 mm for the cast aluminum and cast ironknuckles were considered at the critical locations (i.e.spindle fillets for forged steel knuckle, and hub bolt holesfor cast aluminum and cast iron knuckles). These meshsizes were obtained based on the convergence of stressand strain energy at certain geometry locations.


6/11

6

A potential source of significant error in fatigue analysisis inaccuracy of stress and strain predictions. Thereforevalidating the FEA models was critical to this study. Tovalidate the models, values of strains as measured bystrain gages in component testing and as predictedusing the finite element analysis were compared and arelisted in Table 2. The strain gages for the forged steelknuckle were positioned at the vicinity of the spindle rootand the first step fillets, and for the cast aluminumknuckle two gages were positioned at the goose neck of the strut arm and two at the hub bolt holes where thecrack initiation was observed during component testing.These locations are identified in Table 2. Depending onthe location of the gage, the proper component of thestrain obtained from the FE analysis was selected for comparison. The component testing was only conductedfor forged steel and cast aluminum knuckles; thereforeno data is presented for the cast iron knuckle. Thedifferences between measured and predicted strainsobtained for the two knuckles were consideredreasonable for the complex knuckle geometries. For theforged steel knuckle, which has a relatively simpler geometry, results of strain calculations from analyticalmechanics of materials equation are also listed in Table2. As can be seen, these results are mostly in betweenthe measured and FEA-predicted strains. The results of the finite element analysis were also checked withregards to symmetry and linearity of the loading in theelastic range. It should be noted that the position of thestrain gages and the magnitudes of the applied loadswere such that all measured strains were in the elasticrange.

The equivalent von Mises stress contours and criticallocations for a typical load value are presented for thethree models in Figure 8. The spindle 1 st step fillet areafor the forged steel knuckle, the hub bolt holes for thecast aluminum knuckle, and the strut arm root and hubbolt hole for the cast iron knuckle, were found to be theareas of high stresses. von Mises equivalent stressesand strains are used for subsequent fatigue life analysisand comparisons. For the forged steel knuckle, at thehighest experimental load level yielding occurred both

gross (in the spindle) and locally (at the fillet), whereasfor the cast aluminum knuckle at the highestexperimental load only local yielding (at the hub bolthole) occurred.

Table 2: Measured and predicted strain values at 2.2 kNstatic load. Locations of the gages are also shown.

GageNumber

MeasuredStrain

( strain)

Mc

A

P +

( strain)

PredictedStrain from

FEA( strain)

Diff.Meas.

and FEA(%)

Forged Steel

1 542 575 583 8

2 -527 -557 -546 4

3 1561 1571 1716 10

4 -1489 -1536 -1590 7

Cast Aluminum

1 455 - 434 5

2 534 - 470 12

3 228 - 268 18

4 289 - 320 11

Figure 8: Contours of von Mises stress showing the critical locations for the forged steel (left), cast aluminum (middle),and cast iron (right) knuckles. The darker areas (from left to right, spindle 1 st step for the forged steel, hub bolt hole for thecast aluminum, and strut arm root and hub bolt hole for cast iron knuckle) are the highest stressed locations.


7/11

7

COMPONENT FATIGUE TESTS

To obtain stress-life behavior of the components and tobe able to compare the fatigue behavior of the knuckles,constant-amplitude load-control fatigue tests wereperformed for forged steel and cast aluminum knuckles.

EXPERIMENTAL PROGRAM

The suspension system of each vehicle that thecomponent belongs to was identified and the loadingand attachment conditions of the knuckle in each vehiclewere investigated. A typical drawing for the suspensionsystem of the vehicle with the forged steel knuckle isshown in Figure 9. The strut mounts on one side of thiscomponent and on the other side the spindle attaches tothe wheel hub assembly. These attachments wereconsidered as the primary restraint and loadingconditions of the test, respectively. Similar procedurewas followed for the aluminum steering knuckle. Thecritical points of highest stress in the component wereobtained from the stress analysis, as describedpreviously. Accordingly, specific test fixtures for eachone of the two knuckles were designed and machined.

.Figure 9: The forged steel knuckle within the suspensionsystem. This arrangement drawing was used todetermine loading and boundary conditions.

As shown in Figure 10 for the forged steel knuckle, thespindle was fixed by a 2-piece block where threadedrods tightened the block to the spindle. A pair of L-shaped moment arms transferred the load from thetesting machine loading actuator to the spindle blocks inthe form of a bending load. The strut and suspensionconnections on the knuckle body were fixed to the benchusing round and square blocks. For the cast aluminumknuckle, a two-strut-attachment test was conducted, asshown in Figure 11. In this arrangement, the strutattachment of the arm was connected from both sides to

Figure 10: Schematic drawing for forged steel knuckletest arrangement.

Figure 11: Cast aluminum knuckle test arrangementshowing details of fixturing, schematic (top), and armfixturing close up (bottom).


8/11

8

a pair of moment arms. The moment arms transferredthe bending load from the loading actuator to theknuckle. The four hub bolt holes were fixed to the bench.

A closed-loop servo-controlled hydraulic 100 kN axialload frame was used to conduct the tests. Thecalibration of the system was verified prior to thebeginning of the tests. A rod end bearing joint was usedto apply the load from the actuator to the moment arms,in order to avoid any out of plane bending. Due torelative rigidity of the fixtures, the effect of horizontalfriction force was found to be significant at the joint-fixture contact point. Therefore, a needle roller bearingwas installed on each side of the pin of the bearing,allowing the moment arm to roll horizontally to minimizefriction force. Care was taken to ensure symmetry of thebending load transferred from the two moment arms. Allfixture bolts and nuts were tightened with identicaltorque values to maintain consistency.

The load levels were determined based on stressanalysis results and the true stress-true strain curve of the materials. A minimum load of 220 N was used in alltests, corresponding to an R -ratio ( P min /P max ) of less than0.07. A total of 7 tests at four load levels with amplitudesbetween 1100 N and 2350 N for the forged steelknuckle, and a total of 6 tests at four load levels withamplitudes between 1550 N and 3000 N for the castaluminum knuckle were conducted. The frequency of thetests ranged from 0.5 Hz for higher load levels, to 5 Hzfor lower load levels. The load levels chosen resulted infatigue lives between 10 4 and 2x10 6 cycles.

EXPERIMENTAL RESULTS

Displacement amplitude versus cycle data of the

component during each test was monitored in order torecord macro-crack nucleation (i.e. a crack on the order of several mm), growth, and fracture stages. Due to thenature of the loading and restraints on both knuckles,the locations of crack initiation could not be reached toenable detecting crack nucleation. Therefore, a markeddisplacement amplitude increase during the test wasconsidered as the crack nucleation point, and a suddenincrease as the final fracture.

Variations of displacement amplitude versus cycles for two typical tests of forged steel and cast aluminumknuckles are shown in Figure 12. As can be observed

from this figure, for the forged steel knuckle thedisplacement amplitude was nearly constant until aboutthe end of the test. This indicates that the time lagbetween macro-crack nucleation and fracture was asmall fraction of the total life. On the other hand, for thecast aluminum knuckle, the crack growth portion of thelife was significant. The crack lengths of the castaluminum knuckles were also visually observed andrecorded. For the typical cast aluminum knuckle data inFigure 12, the crack lengths were 8 mm, 13 mm, 20 mmand 27 mm at N/N f equal to 0.3, 0.5, 0.7 and 0.9,respectively, where crack grew with an approximately

linear trend versus number of cycles. The lives to failureused in latter comparisons for the cast aluminumknuckle were considered to be those of macro-cracknucleation.

0.5

1.5

0 0.2 0.4 0.6 0.8 1

Normalized Number of Cycles, N/N f

D i s p

l a c e m e n

t A m p

l i t u d e

( m m

) Forged Steel KnuckleCast Aluminum Knuckle

crack nucleates

Figure 12: Displacement amplitude versus normalizedcycles for typical forged steel and cast aluminumknuckles.

Figure 13: Superimposed stress amplitude versus lifecurves for forged steel and cast aluminum knuckles.

The stress amplitude versus life curves of the twoknuckles are superimposed in Figure 13. For the castaluminum knuckle S-N lines based on failure defined aseither macro-crack nucleation or fracture are shown. Ascan be seen, on the average, about 50% of the castaluminum knuckle life is spent on macro-crack growth.This figure also shows that the forged steel knuckle


9/11

9

results in about two orders of magnitude longer life thanthe cast aluminum knuckle, for the same stressamplitude level. This occurs at both short as well as longlives. Note that that the difference can be even larger atlong lives, due to the run-out data points for the forgedsteel knuckle. It could also be seen from this figure thatthe highest load levels provided life in the range of 10 4 to5x10 4 cycles. Load levels corresponding to this liferange are considered to be representative of overloadconditions for suspension components, such as asteering knuckle in service.

COMPARISONS OF COMPONENT FATIGUEBEHAVIORS AND LIFE PREDICTIONS

Manufacturing process, such as forging or casting,generally determines the strength level and scatter of mechanical properties, but the geometry can suppressthe influence of the material, as indicated by Berger etal. (10) . According to Sonsino et al. (7) , for a complexcomponent geometry where no notch factors could bedefined, transferability of material test data could beperformed only through local equivalent stresses or strains in the failure critical areas. In this study, the localequivalent stresses and strains corresponding to theexperimental loading conditions were obtained byapplying equivalent loads to the simulated finite elementmodels. Since the tests were conducted with a meanload, the modified Goodman equation was used toaccount for the effect of mean stress:

1=+u

m

Nf

a

S

(1)

where a , Nf , m and S u are alternating stress in thepresence of mean stress, alternating stress for equivalent completely reversed loading, mean stress,and ultimate tensile strength, respectively. The Basquinequation was then used to obtain the fatigue life usingthe material properties listed in Table 1:

b f f Nf N )2( = (2)

Normally, a surface finish reduction factor is applied tothe fatigue strength of a component. However, the filletof the forged steel knuckle was machined and polishedand, therefore, no surface finish factor was applied. For the two cast knuckles, due to the nature of the castingmaterials and the fact that the defects of a castingmaterial is uniform internally and externally, no surfacefinish factor was implemented either.

In the strain-life approach, the local values of stress andstrain at the critical location were used to find fatigue life,according to the Smith-Watson-Topper (SWT)parameter that considers the mean stress effect:

cb f f f

b f f a N E N E

++= )2()2()( 22max (3)

where max is the maximum stress ( max = a + m) and a is the strain amplitude. The strain-life properties in thisequation are defined and their values for each materialare listed in Table 1.

Superimposed stress amplitude versus life curves basedon stress-life approach, and SWT parameter versus lifebased on the strain-life approach for the three knucklesare presented in Figures 14 and 15, respectively. Toobtain stress unit in Figure 15, square root of the leftside of Equation 3 is plotted as the SWT parameter.Component test data for the forged steel and castaluminum knuckles are also superimposed in thesefigures for comparison with predictions. Figure 14indicates that predictions based on the S-N approachare overly conservative for both the forged steel andcast aluminum knuckles. This is partly due to theconservative nature of the modified Goodman equation.The predictions based on the SWT parameter are closer to the experimental results, as shown in Figure 15.Comparison of the forged steel and cast iron knuckleprediction curves in Figure 15 demonstrates that theforged steel knuckle offers more than an order of magnitude longer life than the cast iron knuckle, at bothshort as well as long lives. As compared with the castaluminum knuckle, the predicted lives for the forgedsteel knuckle are longer by about three orders of magnitude.

100

1000

1E+3 1E+4 1E+5 1E+6 1E+7

Cycles to Failure, N f

S t r e s s

A m p

l i t u d e

( M P a

)

Forged Stee l Knuckle - Test DataCast Aluminum Knuckle -Test DataForged Stee l Knuckle - Prediction

Cast Iron Knuckle - PredictionCast Aluminum Knuckle - Prediction

Figure 14: Superimposed stress amplitude versus lifecurves based on the stress-life approach for the forgedsteel, cast aluminum and cast iron knuckles.

In the forging process, hot working refines grain patternand imparts high strength and ductility, therefore forged


10/11

10

components have lower possibility of internal defects,whereas castings are weaker in this respect. In addition,lower ductility of castings limits their capacity for cyclicplastic deformation which often occurs at stressconcentrations and at overloads, and thereforeshortening their fatigue lives. Residual stresses at thecritical locations of the component generated during themanufacturing process could be a significant source of strengthening (if compressive) or weakening (if tensile),in terms of fatigue life.

Figure 15: Superimposed SWT parameter versus lifecurves based on the strain-life approach for the forgedsteel, cast aluminum and cast iron knuckles.

CONCLUSIONS

Specimen tests, finite element analyses, and componenttests were conducted in this study to assess andcompare fatigue behavior of forged steel, castaluminum, and cast iron steering knuckles. Based on theexperimental results and analyses presented, thefollowing conclusions can be made:

1. From tensile tests and monotonic deformationcurves it was concluded that cast aluminum and cast

iron reached 37% and 57% of forged steel ultimatetensile strength, respectively. The percentelongation of cast aluminum and cast iron werefound to be 24% and 48% of the forged steel,respectively.

2. The cyclic yield strength of cast aluminum and castiron were found to be 54% and 75% of forged steel,respectively, while the cyclic strain hardeningexponent of cast aluminum and cast iron were 46%and 55% of the forged steel, respectively. These

indicate higher resistance of the forged steel tocyclic plastic deformation.

3. Better S-N fatigue resistance of the forged steel wasobserved, as compared with the two cast materials.Long-life fatigue strengths of cast aluminum andcast iron are only 35% and 72% of the forged steel,respectively.

4. Comparison of the plastic strain amplitude versusfatigue life behavior for the three materials showedhigher capacity of the forged steel for cyclic plasticdeformation, and therefore better low cycle fatiguebehavior, as compared with cast aluminum and castiron.

5. Comparisons of strain-life fatigue behavior of thethree materials indicated that the forged steelprovides about a factor of 20 and 4 longer lives inthe short-life regime compared to the cast aluminumand cast iron, respectively. In the high-cycle regime,forged steel resulted in about an order of magnitudelonger life than the cast iron, and about a factor of 3longer life, compared to the cast aluminum.

6. The differences between measured and FEA-predicted strains obtained for the forged steel andcast aluminum knuckles were found to bereasonable for the complex knuckle geometriesconsidered.

7. Based on the component testing observations, crackgrowth life was found to be a significant portion of the cast aluminum knuckle fatigue life, while crackgrowth life was an insignificant portion of the forgedsteel knuckle fatigue life.

8. Component testing results showed the forged steelknuckle to have about two orders of magnitudelonger life than the cast aluminum knuckle, for thesame stress amplitude level. This occurred at bothshort as well as long lives.

9. The S-N predictions were overly conservative,whereas strain-life predictions were relatively closeto component experimental results. Comparison of the strain-life prediction curves of the componentsdemonstrated that the forged steel knuckle offersmore than an order of magnitude longer life than the

cast iron knuckle.

REFERENCES

1. Gunnarson, S., Ravenshorst, H., and Bergstorm, C.M., Experience with Forged AutomotiveComponents in Precipitation Hardened Pearlitic-Ferritic Steels, Fundamentals of Microalloying Forging Steels , Proceedings , Metallurgical Societyof AIME, 1987, pp. 325-338.


11/11

11

2. Lee, S. B., Structural Fatigue Tests of AutomobileComponents under Constant Amplitude Loadings,Fatigue Life Analysis and Prediction, Proceedings ,International Conference and Exposition on Fatigue,Goel, V. S., Ed., American Society of Metals, 1986,pp. 177-186.

3. Lee, Y. L., Raymond, M. N., and Villaire, M. A.,Durability Design Process of a Vehicle SuspensionComponent, Journal of Testing and Evaluation , Vol.

23, 1995, pp. 354-363.4. Beranger, A. S., Berard, J. Y., and Vittori, J. F., AFatigue Life Assessment Methodology for

Automotive Components, Fatigue Design of Components , ESIS Publication 22, Proceedings of the Second International Symposium on FatigueDesign , FD95, 5-8 September, 1995, Helsinki,Finland, Marquis, G., Solin, J., Eds., 1997, pp. 17-25.

5. Conle, F. A. and Chu, C. C., Fatigue Analysis andthe Local Stress-Strain Approach in ComplexVehicular Structures, International Journal of Fatigue , Vol. 19, No. 1, 1997, pp. S317-S323.

6. Savaidis, G., Analysis of Fatigue Behavior of aVehicle Axle Steering Arm Based on Local Stressesand Strains, Material wissenschaft und Werkstoff technik , Vol. 32, No. 4, 2001, pp. 362, 368.

7. Sonsino, C. M., Kaufmann, H., and Grubisic, V.,Transferability of Material Data for the Example of aRandomly Loaded Forged Truck Stub Axle, SAETechnical Paper No. 970708 in SAE PT-67 , Recent Developments in Fatigue Technology, Chernenkoff,R. A., Bonnen, J. J., Eds., Society of AutomotiveEngineers, 1997.

8. ASTM Standard E606-92, "Standard Practice for Strain-Controlled Fatigue Testing," Annual Book of

ASTM Standards, Vol. 03.01, 1998.9. ASTM Standard E8-03, "Standard Test Methods for

Tension Testing of Metallic Materials," Annual Bookof ASTM Standards, Vol. 03.01, 2003.

10. Berger, C., Eulitz, K. G., Heuler, P., Kotte, K. L.,Naundorf, H., Schuetz, W., Sonsino, C. M., Wimmer

A., and Zenner, H., Betriebsfestigkeit in Germany An Overview, International Journal of Fatigue ,Vol. 24, 2002, pp. 603-625.

2004SAEZoroufiFatemi2004-01-0628

Documents

Transcript of 2004SAEZoroufiFatemi2004-01-0628