design and failure modes of automotive suspension springs

20
Design and failure modes of automotive suspension springs Y. Prawoto a, * , M. Ikeda a , S.K. Manville a , A. Nishikawa a,b a NHK International Corporation, 50706 Varsity Court, Wixom, MI 48393, United States b NHK Spring Corporation, 3-10 Fukuura Kanazawa-ku, Yokohama 236-0004, Japan Received 23 July 2007; accepted 19 November 2007 Available online 21 February 2008 Abstract This paper is a discussion about automotive suspension coil springs, their fundamental stress distribution, materials characteristic, manufacturing and common failures. An in depth discussion on the parameters influencing the quality of coil springs is also presented. Following the trend of the auto industry to continuously achieve weight reduction, coil springs are not exempt. A con- sequence of the weight reduction effort is the need to employ spring materials with significantly larger stresses compared to similar designs decades ago. Utilizing a higher strength of steel possesses both advantages and disadvantages. The advan- tages include the freedom to design coil springs at higher levels of stress and more complex stresses. Disadvantages of employing materials with higher levels of stress come from the stresses themselves. A coil’s failure to perform its function properly can be more catastrophic than if the coil springs are used in lower stress. As the stress level is increased, material and manufacturing quality becomes more critical. Material cleanliness that was not a major issue decades ago now becomes significant. Decarburization that was not a major issue in the past now becomes essential. To assure that a coil spring serves its design, failure analysis of broken coil springs is valuable both for the short and long term agenda of car manufacturer and parts suppliers. This paper discusses several case studies of suspension spring failures. The failures presented range from the very basic including insufficient load carrying capacity, raw material defects such as excessive inclusion levels, and manufacturing defects such as delayed quench cracking, to failures due to complex stress usage and chemically induced failure. FEA of stress distributions around typical failure initiation sites are also presented. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Metallurgical failure analysis; Decarburization; Inclusion; Delayed quench crack; Coil spring 1. Introduction A mechanical spring is defined as an elastic body which has the primary function to deflect or distort under load, and to return to its original shape when the load is removed. The long-established compression spring design theory involves over simplification of the stress distribution inside the wire. One of the simplest 1350-6307/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.engfailanal.2007.11.003 * Corresponding author. Tel.: +1 248 563 0169; fax: +1 248 926 2022. E-mail address: [email protected] (Y. Prawoto). Available online at www.sciencedirect.com Engineering Failure Analysis 15 (2008) 1155–1174 www.elsevier.com/locate/engfailanal

Upload: heliselyay

Post on 18-Nov-2014

192 views

Category:

Documents


15 download

TRANSCRIPT

Page 1: Design and Failure Modes of Automotive Suspension Springs

Available online at www.sciencedirect.com

Engineering Failure Analysis 15 (2008) 1155–1174

www.elsevier.com/locate/engfailanal

Design and failure modes of automotive suspension springs

Y. Prawoto a,*, M. Ikeda a, S.K. Manville a, A. Nishikawa a,b

a NHK International Corporation, 50706 Varsity Court, Wixom, MI 48393, United Statesb NHK Spring Corporation, 3-10 Fukuura Kanazawa-ku, Yokohama 236-0004, Japan

Received 23 July 2007; accepted 19 November 2007Available online 21 February 2008

Abstract

This paper is a discussion about automotive suspension coil springs, their fundamental stress distribution, materialscharacteristic, manufacturing and common failures. An in depth discussion on the parameters influencing the quality ofcoil springs is also presented.

Following the trend of the auto industry to continuously achieve weight reduction, coil springs are not exempt. A con-sequence of the weight reduction effort is the need to employ spring materials with significantly larger stresses compared tosimilar designs decades ago. Utilizing a higher strength of steel possesses both advantages and disadvantages. The advan-tages include the freedom to design coil springs at higher levels of stress and more complex stresses. Disadvantages ofemploying materials with higher levels of stress come from the stresses themselves. A coil’s failure to perform its functionproperly can be more catastrophic than if the coil springs are used in lower stress. As the stress level is increased, materialand manufacturing quality becomes more critical. Material cleanliness that was not a major issue decades ago nowbecomes significant. Decarburization that was not a major issue in the past now becomes essential.

To assure that a coil spring serves its design, failure analysis of broken coil springs is valuable both for the short andlong term agenda of car manufacturer and parts suppliers. This paper discusses several case studies of suspension springfailures. The failures presented range from the very basic including insufficient load carrying capacity, raw material defectssuch as excessive inclusion levels, and manufacturing defects such as delayed quench cracking, to failures due to complexstress usage and chemically induced failure. FEA of stress distributions around typical failure initiation sites are alsopresented.� 2007 Elsevier Ltd. All rights reserved.

Keywords: Metallurgical failure analysis; Decarburization; Inclusion; Delayed quench crack; Coil spring

1. Introduction

A mechanical spring is defined as an elastic body which has the primary function to deflect or distort underload, and to return to its original shape when the load is removed. The long-established compression springdesign theory involves over simplification of the stress distribution inside the wire. One of the simplest

1350-6307/$ - see front matter � 2007 Elsevier Ltd. All rights reserved.

doi:10.1016/j.engfailanal.2007.11.003

* Corresponding author. Tel.: +1 248 563 0169; fax: +1 248 926 2022.E-mail address: [email protected] (Y. Prawoto).

Page 2: Design and Failure Modes of Automotive Suspension Springs

1156 Y. Prawoto et al. / Engineering Failure Analysis 15 (2008) 1155–1174

approaches available is referenced here [1]. The so called un-wound spring as shown in Fig. 1 is commonlyused. It is based on the assumption that an element of an axially loaded helical spring behaves essentiallyas a straight bar in pure torsion. The following notations are typically used: P: Applied load, a: Pitch angle,s: Shear stress, R: Coil radius, and d: Wire diameter. The torsion is then calculated as PR cos a, the bendingmoment as PR sin a. the shear force as P cos a, and the compression force as P sin a. Traditionally, when thepitch angle is less than 10�, both the bending stresses and the compression stresses are neglected.

Assuming that the shear stress distribution is linear across the wire cross section, and PR cos a = PR, thefollowing should be valid:

s ¼ 16PR

p � d3: ð1Þ

The shear stress here is usually called uncorrected shear stress. The total length l is 2pRn, where n is the num-ber of active coils. Using the fact that c = s/G, it can be rewritten as 16PR/(p � d3G), and the total angulartorsion u becomes:

u ¼Z 2pRn

0

2cd

dx ¼ 32PR

pd4Gdx ¼ 64PR2n

Gd4; ð2Þ

where G is the modulus of rigidity. The total deflection caused by the angular torsion is:

d ¼ Ru ¼ 64PR3n

Gd4¼ 8PD3n

Gd4: ð3Þ

The spring rate therefore becomes:

k ¼ Pd¼ Gd4

8nD3: ð4Þ

Eq. (4) is still commonly used to estimate the spring rate by suspension designers. As opposed to the uncor-rected shear stress in Eq. (1), Wahl [2] proposed corrected shear stress. The uncorrected shear stress neglects agreat many factors which modify the stress distribution in actual helical springs. The corrected shear stress, sa,is obtained by multiplying the uncorrected stress with a correction factor K, which depends upon the springindex D/d. Fig. 2 shows the typical corrected shear stress distribution.

Furthermore, by taking x as the distance from the cross point where the shear stress is zero, Wahl provedthat the following equation holds:

sa ¼32xPR2

p � d4ðR� d2=16R� xÞð5Þ

With the introduction of the spring index c = D/d, the maximum shear stress at the inner side of the coil, wherex = d/2 -d2/16R, becomes:

Fig. 1. Wound and un-wound coil springs.

Page 3: Design and Failure Modes of Automotive Suspension Springs

Fig. 2. Uncorrected shear stress vs. corrected shear stress distribution.

Y. Prawoto et al. / Engineering Failure Analysis 15 (2008) 1155–1174 1157

sa1¼ 16PR

p � d3

4c� 1

4c� 4ð6Þ

Additional shear stress caused by the neutral surface of a cantilever of circular cross section loaded by force P,the term 4.92P/pd2 should be added to obtain maximum shear stress:

smax ¼16PR

p � d3

4c� 1

4c� 4þ 0:615

c

� �ð7Þ

Fig. 3. Materials used for coil springs.

Page 4: Design and Failure Modes of Automotive Suspension Springs

1158 Y. Prawoto et al. / Engineering Failure Analysis 15 (2008) 1155–1174

And minimum shear stress:

smin ¼16PR

p � d3

4cþ 1

4cþ 4� 0:615

c

� �ð8Þ

Eqs. (7) and (8) are usually used by the design engineer for coil springs when neglecting the curvature. Furthertheory can be found in Ref. [2]. Also, since the equations were derived by over simplification, the larger thepitch angle, the more error that will result. In reality, coil spring makers today use equations that are generallyconfidential, and therefore will not be discussed here. The equations require the design engineer to input thecoil diameter, design height, design load, spring rate, etc. The equation will calculate the optimum possibleshape and dimension of the coil. After this step, for more accurate stress distribution, it is usually too cum-bersome not to use FEA to design.

1.1. Automotive coil spring materials

Although the history of the gasoline powered automobile can be traced back to 1870 when the first car wasmade in Austria, the mass production of cars did not start until about the early 1900s both in Germany and inthe US. The first automotive coil spring was on the model-T (Ford) in 1910, where the suspension combinedthe leaf spring and the coil spring. The earliest coil spring material used had approximately a 500 MPa designstress level. Coil spring materials have developed to the point where today it is common to have a coil springwith a design stress of around 1200 MPa. The approximate time line is shown in Fig. 3.

2. Designing coil spring using FEA

A major reason to use the FEA in coil design is the ability to reduce error caused by the simplification ofequations, mainly by the pitch angle. An FEA based design begins with the selection of the element type, howthe model should be constructed, how accurate the results should be, and how fast the model should be run.The most accurate FEA results can be obtained by creating 3D parts of a coil spring and its seats, followed bymeshing the parts with a 3D solid element. Finer meshing with higher order elements in general will producethe most accurate results.

Fig. 4. Finite element model.

Page 5: Design and Failure Modes of Automotive Suspension Springs

Fig. 5. Example of compressed coil spring.

Y. Prawoto et al. / Engineering Failure Analysis 15 (2008) 1155–1174 1159

However, because of the higher number of elements and a non-linearity due to the contact between a coilspring and seat, or the coil itself, each analysis could take hours. While the accuracy of the result is important,the computational time must be reasonable to incorporate FEA into the coil spring design. To resolve lengthy

Fig. 6. Simple flowchart of coil spring design.

Page 6: Design and Failure Modes of Automotive Suspension Springs

1160 Y. Prawoto et al. / Engineering Failure Analysis 15 (2008) 1155–1174

computational time in a solid model, a 3D beam element is usually selected to model a coil spring and seat [3].Since the deformation of a seat under compression is very minimal and can be ignored, the material propertiesof seats are set very high to act as rigid. Contact between a coil and seat, or the coil itself, is detected by gapelements. A typical FEA model is shown in Fig. 4.

In-house software was developed to model coil/seats (pre-processor) and to display the analysis results(post-processor). The simulation is performed by commercial FEM software, MARC or ABAQUS. In gen-eral, a new FEA model is created each time a coil/seat profile is modified, and the simulation will be repeateduntil all requirements are satisfied. Sometimes it may require numerous simulations to achieve a desired coilspring design, therefore a faster simulation is necessary.

This FEA model may not generate accurate analysis results due to model simplification and assumptions,but will generate the approximated results and significantly reduce the amount of time in designing. Because ofthis approximation, the determination of whether or not the coil design using FEA is good enough will bedone by comparing past analysis results or actual experimental results if available. FEA is also performed afterfinishing the initial design, such as evaluation of the actual product, or redesigning in case of testing failureFig. 5 shows model of coils at free, normal, and compressed coils. Fig. 6 shows the steps of coil design.

3. Major imperfections in coil manufacturing

Raw material selection is always the most important decision in obtaining the best quality of any product,including coil springs. The selection of the raw material usually includes the enforcement of cleanliness, micro-structure, and decarburization inspection. Fig. 7 shows a typical raw material defect in the form of an inclu-sion; also shown is a microstructure matrix defect and decarburization.

Other sources of defects include improper heating patterns prior to coiling. The control of the prior aus-tenite grain size is an important step in coil manufacturing. Fig. 8 shows the difference between a small grainsize and a large grain size. This example was taken from identical materials processed with different param-eters. Although not reflected by other mechanical properties, except by metallography when interpreted byan expert, larger prior austenite grain size is proven to have less advantage in fatigue life than that of smallsize. Some argue that this is due to the fewer number of the grain boundaries passed during crack propagation[4].

Once the raw material is heated properly, the coil is usually formed. Physical defects due to coiling some-times cause the coil to fail early. Following coil formation, a heat treatment process is performed by means ofquenching, followed by tempering. Heat treatment related defects are another major cause of the coil failingearly. These defects include, but are not limited to, quench cracking, insufficient tempering, and over-tempering.

After tempering, the coil spring is shot peened. The shot peening process is beneficial for two reasons: itcleans the surface of defects and scale caused by quenching, and introduces compressive residual stresses atthe surface. Fig. 9 shows the typical residual stress distribution formed by shot peening. When a load isapplied to the coil spring, the net stress is the superposition of the beneficial residual stress from shot peeningand the applied stress.

Fig. 7. Typical defects inherited by raw materials: inclusion, microstructure different from intended one, and decarburization.

Page 7: Design and Failure Modes of Automotive Suspension Springs

Fig. 8. Identical raw materials heated with different heating patterns. on the left, the prior austenite grain size is clearly larger than that ofthe right.

Fig. 9. Typical residual stress distribution caused by shot peening. on the left, the sample is heavily decarburized, while on the right, thesample has a normal surface condition.

Fig. 10. Pre-treatment coverage.

Y. Prawoto et al. / Engineering Failure Analysis 15 (2008) 1155–1174 1161

Setting usually does not have a detrimental effect on the coil. After setting, coating is typically the last stepof coil making. The process of coating consists of two major steps: pre-treatment and coating application. The

Page 8: Design and Failure Modes of Automotive Suspension Springs

1162 Y. Prawoto et al. / Engineering Failure Analysis 15 (2008) 1155–1174

main ingredient in the pre-treatment is usually zinc. The zinc works as a sacrificing anode to protect the steel.Fig. 10 shows the appearance of pre-treatment coverage. Insufficient pre-treatment coverage not only causesthe coil to have poor coating quality, but also worsens the corrosion resistance.

4. Failure analysis case studies

4.1. Raw materials defect

A typical raw material defect is the existence of a foreign material inside the steel, such as non-metallicinclusions. Figs. 11a and 11b show the fracture surface and SEM fractograph, as well as the EDS spectrumof an inclusion located �1 mm below the surface. This particular coil failed early despite all other parametersbeing normal.

In general, there are two types of foreign materials that can become trapped inside the steel solution: largeimperfections such as spinells, and smaller imperfections such as inclusions that are caused by alloying ele-ments. ASTM differentiates inclusion types by thin and heavy, in addition to composition and shape. TypeA is sulfide-type with a boundary of thin and heavy classification of 4 lm. Type B is aluminate-type with aboundary of 9 lm. Type C is Silicate-type with a boundary of 5 lm, and Type D is globular oxide with aboundary of 8 lm. It is also worth noting that thin inclusion rarely cause a coil spring to fail early.

Fig. 12 shows a raw material defect that is usually very difficult to find after a coil is formed. This type ofdefect is easy to detect during the cold drawing process of coil manufacturing preparation. An ideal raw mate-rial has the form of ferrite pearlite. However, a raw material can also have local bainite inside the ferrite pearl-ite matrix. Due to a hardness difference, such raw materials may exhibit internal cracking.

4.2. Surface imperfections

Surface imperfections can occur as small hardening cracks, tool marks, scale embedded to the base materialduring cold drawing, or surface flaws inherited by the raw material. Fig. 13 shows two different surface flawsdeep enough to cause a coil spring to fail early. On the left side, the surface imperfection is inherited from theraw material. This type of defect can occur when the surface flaw detector does not function normally. It isusually easy to determine if such a flaw was inherited from the raw material and not due to coil manufacturing.A pre-existing defect usually has surrounding decarburization after the raw material is heated during coil man-ufacturing, while a surface defect caused by coil manufacturing is often not accompanied by decarburization;see the right side of the figure.

Poorly shot peened surfaces can also be classified as surface imperfections. Fig. 14 shows a comparisonbetween two different coils that failed at similar locations, but possessed completely different fatigue lives.On the left side, the surface was shot peened poorly and therefore exhibited a shorter life. On the right side,the surface was shot peened sufficiently and therefore had a longer life.

Fig. 11a. Fracture surface of a coil that failed early due to an inclusion and Its SEM appearance.

Page 9: Design and Failure Modes of Automotive Suspension Springs

Fig. 11b. Eds spectrum mapping of the inclusion.

Fig. 12. Appearance of different microstructures extracted from the same bar. on the left side, the microstructure is normal ferrite pearlite.on the right side, the same material has bainite structure inside the ferrite pearlite matrix.

Y. Prawoto et al. / Engineering Failure Analysis 15 (2008) 1155–1174 1163

An example of a small quench crack that can be classified as a surface imperfection is shown in Fig. 15. Inthis case the heating process and the heat treatment itself were not wrong, however, the quench oil was con-taminated with water, causing an extremely high cooling rate locally.

Page 10: Design and Failure Modes of Automotive Suspension Springs

Fig. 13. Inherited from raw material (left) and surface imperfection due to manufacturing (right).

Fig. 14. Surface imperfections due to poor shot peening condition.

1164 Y. Prawoto et al. / Engineering Failure Analysis 15 (2008) 1155–1174

Similar to a quench crack, a delayed crack can also sometimes occur. This could be caused by either insuf-ficient tempering time or temperature, or by prolonged time between quenching and tempering. Fig. 16 showsa typical fractograph of such a failure.

4.3. Improper heat treatment

Improper heat treatment can be easily overlooked since a temperature difference in heating does not relatedirectly to the hardness of the material. Extensive evaluations are usually needed to identify this problem.Fig. 8 shows a typical example of an improper heat treatment. Prolonged heating can cause the prior austenitegrain size to grow significantly.

Improper heat treatment can also result in the microstructure becoming pearlite instead of the requiredmartensite. This type of defect is easier to identify due to the clear difference in hardness. Fig. 17 showstwo different coils of the same product with varying microstructure. This defect usually occurs when the heat-ing system does not operate normally. Again, referring to the figure, the left hand side coil has a much lowerlifetime than that of the right side.

Bainitic formation is another form of improper heat treatment. Unlike martensite, bainitic ferrite usuallycontains only slight excess of carbon in ferrite solution. Most of the carbon in a transformed sample of bainiteis in the form of cementite particles, which in turn tend to be coarser than those associated with tempered mar-tensite. The effects of tempering are therefore always milder than is the case when the microstructure is mar-tensite. Furthermore, bainitic structures are usually accompanied by a greater percentage of retained austenite

Page 11: Design and Failure Modes of Automotive Suspension Springs

Fig. 15. Appearance of surface defect caused by small quench crack. the intergranular surface is covered by an oxide due to trappedquench oil being heated in tempering furnace.

Fig. 16. Appearance of intergranular fracture surface caused by delayed quench crack.

Y. Prawoto et al. / Engineering Failure Analysis 15 (2008) 1155–1174 1165

than martensitic structures [5]. Tempering induces the decomposition of the retained austenite into mixture offerrite and carbides. Fig. 18 shows the microstructure of bainite steel.

4.4. Corrosion

Corrosion is a more common cause of spring breakage than is usually understood by users; however, recentcoating technology has reached a point where it is able to protect the metal from even the hardest cold stone

Page 12: Design and Failure Modes of Automotive Suspension Springs

Fig. 17. Improperly heat treated sample (left) vs. properly heat treated sample (right).

1166 Y. Prawoto et al. / Engineering Failure Analysis 15 (2008) 1155–1174

chipping. Fig. 19 shows the appearance of a coil which failed due to corrosion. The right side of the figuredepicts the cross section of the corresponding coil, with the 1 line illustrating its approximate originaldimension.

4.5. Decarburization

Decarburization may be considered as the least severe offender in the entire list of defects. Partial decarbu-rization is usually permissible in spring wire, at least to a slight extent. When a complete, full ferrite ringaround the circumference is found, the wire is always subject to rejection. Fig. 20a shows a case where a coilspring failed due to an excessive decarburized layer. The fractography of the sample also shows that the firstouter layer consisted of softer layer where almost no striations were visible.

Further evaluation of the longitudinal and cross section revealed that the sample had significant decarbu-rization. Subsequently, when the micro hardness near the surface was profiled, one can see that apparentdecarburization was also found, see Fig. 20b.

1 For interpretation of color in Figs. 19 and 23, the reader is referred to the web version of this article.

Page 13: Design and Failure Modes of Automotive Suspension Springs

Fig. 19. Coil spring which broke due to corrosion.

Fig. 18. Bainitic structure – same material as in Fig. 16.

Y. Prawoto et al. / Engineering Failure Analysis 15 (2008) 1155–1174 1167

5. Analysis

5.1. Procedure

A finite element analysis was performed to check the local stress distribution around a given defectusing a typical coil spring (Fig. 21a). First, the overall stress distribution was checked without any defectin the material, and then at the location where the highest stress was found, each defect was added. Sincethe size of the defect is significantly smaller than the whole model, a submodeling technique [6] was used.This technique is used to study a local part of a model with refined meshing based on the FEA result of aglobal model with coarse meshing. Boundary conditions for the submodel will be automatically interpo-lated from the global model solution. As shown in Fig. 21, the submodeling technique was used twicefor this study.

Submodel 2 was modified to apply various defects. For meshing, either the quadratic brick element(C3D20) or quadratic tetrahedron (C3D10) was used. For material specifications, typical spring steel proper-ties, E = 210 GPa and m = 0.3, were used except for the decarburized layer. The commercially available FEAsoftware, ABAQUS, was used here to study each stress distribution.

Page 14: Design and Failure Modes of Automotive Suspension Springs

Fig. 20a. Fracture surface and SEM Fractograph of a coil which failed due to excessive decarburization.

Fig. 20b. Cross-sectional metallograph of the broken coil (left) and micro hardness profile near the surface (right).

Fig. 21. FEA model of a coil spring and its submodels.

1168 Y. Prawoto et al. / Engineering Failure Analysis 15 (2008) 1155–1174

Page 15: Design and Failure Modes of Automotive Suspension Springs

Fig. 22. Von Mises stress result of no-defect model.

Y. Prawoto et al. / Engineering Failure Analysis 15 (2008) 1155–1174 1169

5.2. FEA result of model without defect

For comparison with the defect model, Submodel 2 was analyzed first without any defects. Boundary con-ditions were interpolated from the result of Submodel 1 to the inner and side surfaces of Submodel 2. Fig. 22shows the Von Mises stress distribution. The highest stress was found at the outer surface and the lowest atinner side of wire. The highest Von Mises stress was about 1715 MPa, which matches the stress level of theglobal model. The gray area around the outer edge shows a stress concentration, however this is ignored sinceit is where the boundary condition was applied.

5.3. Defect FEA models and results

5.3.1. Inclusion

A cubic hole was placed about 1 mm below the outer surface; its size is 50 lm (Fig. 23, red dot is the inclu-sion). Instead of using a foreign material for the cubic area, it was left as a hole for simplification. Since ahigher stress concentration was expected around the inclusion area, a finer mesh was used at the center andcoarser mesh was used at the outer area (Fig. 23b).

The stress distribution is shown in Fig. 24. As expected, a local stress concentration is observed at the inclu-sion area, and the highest Von Mises stress reached 2000 MPa, which is higher than the outer surface stresslevel. Stress on other areas, such as outer surface, was at the same level as the no-defect model.

5.3.2. ImperfectionA model was created based on Fig. 13 (left side); the surface imperfection is inherited from the raw material.

A crack (50 lm width, 500 lm depth) alongside of the centerline of the wire was applied to the Submodel 2 asshown in Fig. 25.

The stress distribution is shown in Fig. 26. A high stress concentration is observed at the crack location,and the Von Mises stress exceeded 4000 MPa, which is much higher than the outer surface stress level. There-fore, the product would likely fail from this point. A stress concentration is also observed at the vertical edge,however this concentration occurred due to the boundary condition and should be ignored.

Page 16: Design and Failure Modes of Automotive Suspension Springs

Fig. 25. Part model with imperfection. (Left) and its FEA model (right).

Fig. 23. Part model with inclusion. (Left), FEA model with inclusion (display model is cut in half to show inside) (right).

Fig. 24. Von Mises stress result of inclusion model.

1170 Y. Prawoto et al. / Engineering Failure Analysis 15 (2008) 1155–1174

Page 17: Design and Failure Modes of Automotive Suspension Springs

Fig. 26. Von Mises Stress result of imperfection model (display model was cut at crack location).

Fig. 27. Part model with corrosion (left) and its FEA model (right).

Y. Prawoto et al. / Engineering Failure Analysis 15 (2008) 1155–1174 1171

5.3.3. Corrosion

Instead of modeling the actual corrosion part, a simple oval shape was removed from the outer surface tosimplify the FEM model. Its size is approximately 300 lm in depth, 500 lm in height, and 1 mm in width.Finer meshing was used around the corrosion area since a higher stress concentration was expected there.The model is shown in Fig. 27.

The stress distribution is shown in Fig. 28. As expected, a local stress concentration is observed at the bot-tom edge of the corrosion area, and its Von Mises was about 3450 MPa, which is again much higher than theouter surface stress level. This high stress concentration will cause early spring breakage from this point.

5.3.4. Decarburization

The decarburization model is shown in Fig. 29. A softer material property from Table 1 was used for a0.15 mm layer on the outer surface. Inner side material is the same as original except yield stress was specifiedthis time for elastic–plastic analysis.

Analysis results are shown in Fig. 30. The stress level on the decarburized layer reached the yield stress andremained that value because the material was assumed to be perfectly plastic, and the plastic deformation

Page 18: Design and Failure Modes of Automotive Suspension Springs

Fig. 28. Von Mises stress result of corrosion model.

Fig. 29. Part model with decarburization (left) and its FEA model (right).

Table 1Material property of decarburization model

Original Decarburized

Young’s modulus, E [GPa] 210 124.7Yield stress, rYS [MPa] 1449 359Poisson’s ratio, m 0.3 0.3

1172 Y. Prawoto et al. / Engineering Failure Analysis 15 (2008) 1155–1174

occurred on the decarburized layer. The rest of the part never reached the yield stress; therefore, no plasticdeformation was observed inside of the decarburized layer.

5.4. Analysis result summary

Table 2 shows the summary of analysis results. As expected, a local stress concentration was observed in theinclusion, imperfection, and corrosion defect models at each defect area, and those stress values were much

Page 19: Design and Failure Modes of Automotive Suspension Springs

Fig. 30. Von Mises stress result (left) and plastic deformation result (right).

Table 2FEA summary

Defect Summary

None No stress concentration. The highest stress was found on the outer surface. Von Mises stress � 1715 MPa. Max.Principal stress � 1200 MPa. No plastic deformation occurred.

Inclusion Stress concentration is observed at the inclusion area. Von Mises stress = 2069 MPa. Maximum principalstress = 1922 MPa.

Imperfection Stress concentration is observed at the crack location. Von Mises stress = 4195 MPa. Maximum principalstress = 2670 MPa.

Corrosion Stress concentration is observed at the bottom edge of corrosion surface. Von Mises stress = 3453 MPa. Maximumprincipal stress = 3286 MPa.

Decarburization On decarburized layer, the stress reached the yield point, and a plastic deformation occurred.

Y. Prawoto et al. / Engineering Failure Analysis 15 (2008) 1155–1174 1173

Page 20: Design and Failure Modes of Automotive Suspension Springs

1174 Y. Prawoto et al. / Engineering Failure Analysis 15 (2008) 1155–1174

higher than the model without any defects. These high stress concentrations will cause an early failure; hencethe material needs to avoid these defects as much as possible.

6. Concluding remarks

Failure analyses of suspension coil springs were performed and summarized in this paper. Subsequently,finite element analyses of representative cases were also modeled. Integrating finite element modeling in met-allurgical failure analysis synergizes the power of failure analysis into convincing quantitative analysis. Thispresumably will be the trend in failure analysis.

Acknowledgements

All the samples used in this paper were provided by Nasco, Bowling Green KY. The fatigue tests were allperformed by the experimental team at NHK Wixom Lab.

References

[1] Sugano, Taihei, (Ed.), Design, manufacture, and testing methods of springs. Nikkan Kogyo Shimbunsha: Japan Society for SpringResearch; 2001 [in Japanese].

[2] Wahl AM. Mechanical springs. McGraw-Hill Book Company; 1984.[3] M. Shimozaki, FEM for springs. Nikkan Kogyo Shimbunsha, Japan Society of Spring Engineers; 1997 [in Japanese].[4] Prawoto Y. The effect of residual stress on fatigue crack propagation. J Pract Fail Anal, ASM Int 2002;2(5).[5] HKDH, Badhesia. Bainite in steels. The Institute of Materials; 1992.[6] ABAQUS user’s manual volume II analysis, ABAQUS Inc, 2006.