jurnal analisa kegagalan pada rangka truk

13
Failure analysis of frame crack on a wide-body mining dump truck Sen Zheng a , Kai Cheng a , Jixin Wang a,, Qingde Liao b , Xiaoguang Liu c , Weiwei Liu a a School of Mechanical Science and Engineering, Jilin University, Changchun, China b Xiamen XGMA Heavy Industry Co., Ltd., Xiamen, China c Aviation University of Air Force, Changchun, China article info Article history: Received 24 March 2014 Received in revised form 15 November 2014 Accepted 17 November 2014 Available online 27 November 2014 Keywords: Wide-body mining dump truck Frame Crack failure Dynamic test Stress concentrations abstract The wide-body mining dump truck is a type of heavy-duty, off-highway truck that is mainly used for transporting rock and ore in open-pit mines. Because of various potholes, obstacles, slopes and curves on the bumpy road, the frame of the truck is impacted by the multiform large loads from ground. After five to six months in service, cracks tend to appear in the frame of the truck, near the rear seating of the front leaf springs. To identify the cause of these failures and propose an approach for improving the design, a practical method combined with finite element analysis (FEA), as well as static and dynamic testing, was applied. FEA was used to analyze the cause of the cracking, after which the design of the frame was improved. Static and dynamic tests were conducted to verify the FEA results of the improved frame. Analysis results indicated that the stresses are concentrated in the frame near the rear seating of the front leaf springs, which results in the premature appear- ance of fatigue cracks. A solution for preventing the appearance of these cracks was pro- posed. The improved frame has been in service for more than twelve months in the mine and no cracks have appeared to date. Ó 2014 Elsevier Ltd. All rights reserved. 1. Introduction The wide-body mining dump truck (WMDT) is used in many small-scale mines in China. Considering the bad mine roads having potholes, obstacles, slopes and curves, and the influence of manufacturing cost and service cycle, with a short design- life of 2 years, the WMDT uses leaf spring for its suspension instead of hydro-pneumatic suspension. Hence it is a transition vehicle between traditional mining dump truck and the highway heavy dump truck. One type of WMDT is shown in Fig. 1. The model that was the subject of this study has an unladen weight of 24 t, a maximum load capacity of 72 t, and a normal speed of 10 km/h when carrying a full load or 50 km/h when empty. However, after five to six months in service, cracks tend to appear in the frame near the rear seating of front leaf spring (RSFLS), which results in significant downtime. Either finite element analysis (FEA) or testing alone could not fully analyze the causes of these frame failures. Rather, any analysis of the cracking would require a method that combined both FEA and testing [1–4]. Mi et al. presented a method for predicting the fatigue life of the frame of a 220-t mining dump truck through multibody dynamic analysis and the applica- tion of the finite element method [5,6]. Feng et al. analyzed the static, modal, and response spectra of the FEA model and confirmed its feasibility as a means of verifying the failure of a dump truck’s push rod [7]. Shao et al. presented an analysis http://dx.doi.org/10.1016/j.engfailanal.2014.11.013 1350-6307/Ó 2014 Elsevier Ltd. All rights reserved. Corresponding author. Tel.: +86 130 6920 1036. E-mail address: [email protected] (J. Wang). Engineering Failure Analysis 48 (2015) 153–165 Contents lists available at ScienceDirect Engineering Failure Analysis journal homepage: www.elsevier.com/locate/engfailanal

Upload: baskara-surya-widagdo

Post on 04-Oct-2015

33 views

Category:

Documents


6 download

DESCRIPTION

Analisa kegagalan pada rangka truk penambang

TRANSCRIPT

  • Article history:Received 24 March 2014Received in revised form 15 November 2014Accepted 17 November 2014Available online 27 November 2014

    2 t, and a normalrvice, cracntime.ilures. Rath

    analysis of the cracking would require a method that combined both FEA and testing [14]. Mi et al. presented a metpredicting the fatigue life of the frame of a 220-t mining dump truck through multibody dynamic analysis and the ation of the nite element method [5,6]. Feng et al. analyzed the static, modal, and response spectra of the FEA modconrmed its feasibility as a means of verifying the failure of a dump trucks push rod [7]. Shao et al. presented an analysis

    http://dx.doi.org/10.1016/j.engfailanal.2014.11.0131350-6307/ 2014 Elsevier Ltd. All rights reserved.

    Corresponding author. Tel.: +86 130 6920 1036.E-mail address: [email protected] (J. Wang).

    Engineering Failure Analysis 48 (2015) 153165

    Contents lists available at ScienceDirect

    Engineering Failure AnalysisThe model that was the subject of this study has an unladen weight of 24 t, a maximum load capacity of 7speed of 10 km/h when carrying a full load or 50 km/h when empty. However, after ve to six months in seto appear in the frame near the rear seating of front leaf spring (RSFLS), which results in signicant dow

    Either nite element analysis (FEA) or testing alone could not fully analyze the causes of these frame faks tend

    er, anyhod forpplica-el and1. Introduction

    The wide-body mining dump truck (WMDT) is used in many small-scale mines in China. Considering the bad mine roadshaving potholes, obstacles, slopes and curves, and the inuence of manufacturing cost and service cycle, with a short design-life of 2 years, the WMDT uses leaf spring for its suspension instead of hydro-pneumatic suspension. Hence it is a transitionvehicle between traditional mining dump truck and the highway heavy dump truck. One type of WMDT is shown in Fig. 1.Keywords:Wide-body mining dump truckFrameCrack failureDynamic testStress concentrationsThe wide-body mining dump truck is a type of heavy-duty, off-highway truck that ismainly used for transporting rock and ore in open-pit mines. Because of various potholes,obstacles, slopes and curves on the bumpy road, the frame of the truck is impacted by themultiform large loads from ground. After ve to six months in service, cracks tend toappear in the frame of the truck, near the rear seating of the front leaf springs. To identifythe cause of these failures and propose an approach for improving the design, a practicalmethod combined with nite element analysis (FEA), as well as static and dynamic testing,was applied. FEA was used to analyze the cause of the cracking, after which the design ofthe frame was improved. Static and dynamic tests were conducted to verify the FEA resultsof the improved frame. Analysis results indicated that the stresses are concentrated in theframe near the rear seating of the front leaf springs, which results in the premature appear-ance of fatigue cracks. A solution for preventing the appearance of these cracks was pro-posed. The improved frame has been in service for more than twelve months in themine and no cracks have appeared to date.

    2014 Elsevier Ltd. All rights reserved.Failure analysis of frame crack on a wide-body mining dumptruck

    Sen Zheng a, Kai Cheng a, Jixin Wang a,, Qingde Liao b, Xiaoguang Liu c, Weiwei Liu aa School of Mechanical Science and Engineering, Jilin University, Changchun, ChinabXiamen XGMA Heavy Industry Co., Ltd., Xiamen, ChinacAviation University of Air Force, Changchun, China

    a r t i c l e i n f o a b s t r a c t

    journal homepage: www.elsevier .com/locate /engfai lanal

  • method based on dynamic strain measurements of actual road surface conditions combined with FEA, which was applied tothe analysis of the failure of the drive axle housing of a mining dump truck, using ANSYS software [8].

    These previous efforts provided important guidelines for this study, given that the application of the nite element model,the loads to be applied, and the boundary conditions can be referenced in these papers [914]. The interaction between theframe and the leaf springs also plays an important role in the appearance of cracks in the frame. The design of the frame mustnot only provide sufcient mechanical strength, but also satisfy the technical process requirements. With a goal of solvingthese practical problems, we set out to devise an effective method of improving the design of the frame. To identify the rea-sons for the frame cracking and improve the mechanical strength, FEA and static/dynamic tests are carried out. The framewas analyzed using FEA, after which measures were implemented to improve the stress distribution. A static test was con-ducted to verify the improvement in the frame design by comparing the measured results with those obtained with FEA. Adynamic test was performed to further conrm the improvement, after which the results were compared with those of thestatic test to determine a matching coefcient. Based on the FEA and static test results, the dynamic test data for the originalframe could be acquired by multiplying the FEA results by the matching coefcient. The research conducted during thisstudy is shown in Fig. 2.

    2. FEA models

    2.1. Frame and leaf spring seating

    The frame, which connects the engine, cab, dump body, and other major parts, is a supportive and connective componentof a mining dump truck. The frame supports its own weight and other complex loads, such as impact loads from the suspen-

    154 S. Zheng et al. / Engineering Failure Analysis 48 (2015) 153165sion and gravitational loads imposed by other components. Therefore, the reliability of the frame directly affects the servicelife of the WMDT and the technology that it employs. As shown in Fig. 3, the frame consists of two side rails and sevenwelded crossbeams of different thicknesses. The side rails are designed as D-type box beams, and the second and third cross-beams are connected to the side rails through the leaf spring seatings. The other crossbeams, however, are welded to theinner plates. The frame is fabricated from high-strength, low-alloy quenched and tempered Q460CFD steel, the specicationsof which are listed in Table 1.

    The leaf springs transfer forces and torques from the ground over which the WMDT is traveling. The leaf springs are rig-idly xed to the front axle by two U-bolts, and their ends are connected to the frame through the leaf spring seatings, whichare bolted to the bottom surface and outside plates of the frame. Fig. 3 shows the openings in the frame near the RSFLS thatare required in order to access and tighten the xing bolts. When the truck travels over rough ground, the frame is subject toalternating dynamic loads that cause bending and twisting. The leaf springs are connected to each RSFLS with pin rolls,which bear more force than the slide of the front seating of the leaf spring, as shown in Fig. 3. Based on the results obtainedfor the combined effects of TX, TY, TZ and RX, RY, RZ, it is clear that each RSFLS plays a pivotal role in transferring loads.When the WMDT is turning or traversing a tilting surface, there is an X-direction force in the spring seating, which actstoward the outer plate of the frame, while the main loads in the Y direction are the weight of the WMDTs dump body,its payload, and its powertrain. When the WMDT starts to move or brakes, an inertia force is generated in the forwarddirection, that is, the Z direction. Given the tight interfaces between the bolts, seating and frame, a torque is applied thatis a function of the distances, in the X, Y, and Z directions, between the bolts in this area. As a result, forces are generatedbetween the leaf spring and the side rail in all three directions [15,16].

    Fig. 1. Crack failure of frame.

  • S. Zheng et al. / Engineering Failure Analysis 48 (2015) 153165 1552.2. Finite element model

    To create a nite element model, those frame components of different thicknesses were analyzed separately usingSHELL181. The leaf springs were simulated using the COMBIN14 axial spring damper, which is a one-dimensional tensileor compression unit. Regardless of bending or twisting, each node has three degrees of freedom (DOF), namely, the X, Y,and Z axial movement. BEAM188 and LINK180 elements were used to model the connection between the frame and axles,and the nodes were coupled to limit the DOF of the trunnion shaft. The nite element model had a total of 226,018 nodes and234,097 elements, as shown in Fig. 4.

    2.3. Loads and boundaries

    In this analysis, 72 t and 8.3 t are selected as the quality as the payload and dump body respectively, giving a total of 80.3t, uniformly distributed along the upper surface of the frame side rails and in the same position as in the actual WMDT. Theweight of the cab, engine, and dump body are simplied to a single downward force acting on a corresponding node on theupper surface. Furthermore, the X, Y, and Z directions of the translational DOF of the tires are constrained, while the X, Y, and

    Fig. 2. Paper analysis ow chart.

    Fig. 3. RSFLS on the frame.

    Table 1Physical properties of Q460CFD.

    Material Density (kg/m3) Elastic modulus (GPa) Poissons ratio Yield strength (MPa) Ultimate strength (MPa)

    Q460CFD 7850 200 0.3 460 510

  • 156 S. Zheng et al. / Engineering Failure Analysis 48 (2015) 153165Fig. 4. FEA model of frame.Z directions of the translational DOF at the trunnion shaft are coupled. In addition, the translational DOF limits the leafsprings in the X direction at the corresponding positions. If no special instruction is given, the previously mentioned loadsand constraints are applied to the following operating conditions, as shown in Fig. 5.

    3. Static analysis

    3.1. FEA of the original frame

    Due to the versatility of the truck and the diversity of the conditions in which it is operated, several typical cases werecarried out in detailed analyses.

    3.1.1. Static case with full loadThe frame experiences pure bending when the WMDT is static on at ground. The deformation and stress in the frame

    were analyzed for this case. It was observed that the maximum stress at that point in the side rail where a crack appearsis 165 MPa, as shown in Fig. 6(a). The overall strength of the model is expressed by the von Mises stress.

    3.1.2. Hoisting conditionWhen the dump body is being hoisted up, the stress in the frame varies as the angle of the dump body increases. The

    frame experiences the maximum bending load when the dump body is lifted up at the hoisting moment, which correspondsto a typical pure bending working condition. From Fig. 6(b), the frame experiences large amounts of stress, peaking at197 MPa.

    Fig. 5. Loads and boundaries.

  • S. Zheng et al. / Engineering Failure Analysis 48 (2015) 153165 157A.88MPaA.140MPa 3.1.3. Obstacle surmounting conditionThe frame twists as the truck travels over uneven ground. The stress of the frame is worse under these conditions because

    of the combined effects of bending and torsion. In this condition, the left front tire was lifted 200 mm above the ground tosimulate the surmounting of an obstacle. As shown in Fig. 6(c), the maximum stress was 235 MPa near the crack.

    3.1.4. Sinking conditionSinking is experienced as frequently as the surmounting of obstacles during actual driving and has a signicant effect on

    the frame at the RSFLS. To simulate this condition, the left front tire was allowed to sink 200 mm below the ground surface.As shown in Fig. 6(d), the maximum stress was 200 MPa at the bottom surface of the frame side rail near the RSFLS.

    (a) Static case with full load (b) Hoisting condition

    (c) Obstacle surmounting condition (d) Sinking condition

    (e) Braking condition

    B.200MPa

    A.139MPa

    B.235MPa

    A.166MPa

    A.132MPa

    B.143MPa

    B.197MPaB.165MPa

    Fig. 6. FEA results of the original frame at RSFLS.

  • 3.1.5. Braking conditionCracks most frequently occurred at that part of the frame where it connects to the suspension. This phenomenon is clo-

    sely related to emergency braking. Therefore, it was necessary to analyze the stress intensity in the frame under the inu-ence of emergency braking. The RSFLS experiences greater impact forces than the front seating of front leaf spring becausethe rear one is connected to the leaf spring by the pin roll. As a result, the reaction force acting on the bottom surface of theframe is even larger. As shown in Fig. 6(e), the maximum stress was 143 MPa when the truck was in static balance.

    3.1.6. Summary of FEA results for original frameThe FEA results show that the stresses at the crack position were larger and changed more abruptly than in other parts of

    the frame. The maximum stress at the bottom surface of the frame at the crack position was 235 MPa when the WMDT wassurmounting an obstacle and 200 MPa when the WMDT was sinking. It is a load-transfer point between the bottom surfaceof the frame and the RSFLS, and the stress increases as the RSFLS becomes narrower. The U-shaped hole forms a discontin-uous section in the inner plate. This gives rise to the discontinuous and high-level stress. Modifying the shape of the hole

    158 S. Zheng et al. / Engineering Failure Analysis 48 (2015) 153165could decrease the level of the stress at the point where the cracks occur.

    3.2. Improvement of the frame

    To overcome the problem of cracking, measures should be implemented to improve the strength of the frame. Increasingthe thickness of the frame side rails would increase the strength but the stress concentration would remain as is. Using theFEA method, while referring to design experience and the technical process requirements, the following methods wereadopted to improve the frame, as shown in Fig. 7.

    The U-shaped hole was changed to a circular hole, moved forward 135 mm, and elevated to the middle of the inner plate,corresponding to the stress neutral layer. The number of holes and their total area are kept to a minimumwhile conform-ing to the technological requirements and the stress demands.

    By widening the RSFLS by 100 mm and increasing the distance between the two bolts on the bottom surface, the bendingmoment acting on the frame is reduced.

    The section of the frame is changed from a D-type box beam to a form congured by welding two C-bend plates together.This enhances the bending strength of the cover plate.

    As shown in Fig. 8(a)(e), the stress of the improved frame is reduced at the RSFLS, and structures are implemented easily.Overall, the stresses of frame are reduced signicantly. Comparison of the FEA results between the original and improvedframes is shown in Table 2. As the WMDT is surmounting an obstacle, the stress at the point where the cracks occur fallsfrom 235 MPa to 43 MPa, a reduction by 81.7%. The maximum stress of the improved frame is lower than of the originalframe. These ndings show that the maximum stresses of the dangerous areas are decreased. Furthermore, by modifyingthe geometrical shape, the fatigue life of the frame could be greatly extended. In short, the use of FEA provides a reasonablemeans for nding the position at which fatigue cracks occur. To further verify the rationality and practicality of the improvedframe, however, static and dynamic tests are necessary.

    4. Static and dynamic tests

    4.1. Static test on the test site

    To check the validity of the FEA results, a static test of the improved frame design was carried out. The 32 measuringpoints with 60 response channels consist of sensors and temperature compensators, which access to KYOWA strength test-ing system, as shown in Fig. 9. Based on the FEA results, the cases where the WMDT surmounts an obstacle and sinks into theground are shown in Fig. 10. For the complex stress state under such operating situations, the frame was measured by 45

    (a) Original frame

    (b) Improved frametwo C-bend type

    D-type

    Fig. 7. Original and improved frames.

  • S. Zheng et al. / Engineering Failure Analysis 48 (2015) 153165 159A.13MPastrain gauge rosettes to observe the stress distributions in the X, Y, and Z directions. The single strain gauge was used to mea-sure the unidirectional bending strength of the frame. Two crucial locations near the crack were studied, with the aim ofobtaining the stresses labeled 6 and 8. Some of the measuring points are shown in Fig. 11.

    As shown in Fig. 12(a)(d), the static full loads and hoisting loads were symmetrical. The values obtained show that thestresses increased with the load. In the static test under the full-load condition, the maximum stress at measuring point 6was 12 MPa, while that at measuring point 8 was 22 MPa. Similarly, for the hoisting condition, the maximum stress atmeasuring point 6 was 19 MPa while that at measuring point 8 was 22 MPa. The data were reset to zero at 0 t under both

    (a) Static case with full load

    (b) Hoisting condition

    A.100MPa

    B.24MPa

    B.23MPa

    (c) Obstacle surmounting condition

    A.47MPa

    B.43MPa

    Fig. 8. FEA results of the improved frame at RSFLS.

  • 160 S. Zheng et al. / Engineering Failure Analysis 48 (2015) 153165B.53MPaA.45MPaworking conditions. As shown in Fig. 12(e)(h), under the obstacle surmounting and sinking conditions, the data obtained atmeasuring points 6 and 8 showed that the stresses on the frame were smaller.

    To check the validity of the results of FEA, the FEA stress value was compared with the peak stress obtained in the statictest, as shown in Table 3. This comparison revealed that the maximum stress value obtained with the FEA most closely cor-responded to the static test results, indicating that the nite element model of the frame was reliable.

    4.2. Dynamic stress test

    A dynamic test was performed with the WMDT, carrying a 55-t load over a typical worksite driving route, as shown inFig. 13. In the same way as for the static test, 32 measuring points with 46 response channels were glued onto the framesurface [17]. The rst stage was the loading period, which began at 0 s and had a duration of 3185 s. During this period,the truck was fully loaded with 26 t of material extracted from the mining site. From 3186 s to 5500 s, the truck was driven

    (d) Sinking condition

    (e) Braking condition

    A.83MPa

    B.85MPa

    Fig. 8 (continued)

    Table 2Comparison of the FEA results between the original and improved frames (unit: MPa).

    Position Position A original(see Fig. 6)

    Position A improvement(see Fig. 8)

    Position B original(see Fig. 6)

    Position B improvement(see Fig. 8)

    Static case with full load 140 13 165 23Hoisting condition 88 100 197 24Obstacle surmounting condition 166 47 235 43Sinking condition 139 45 200 53Braking condition 132 83 143 85

  • Fig. 9. Testing photos of data acquisition system.

    (a) Obstacle surmounting condition (b) Sinking condition

    Fig. 10. Typical static testing condition at test site.

    Fig. 11. Measuring points.

    S. Zheng et al. / Engineering Failure Analysis 48 (2015) 153165 161

  • 162 S. Zheng et al. / Engineering Failure Analysis 48 (2015) 153165on a long route which included combinations of curves and slopes. During this period, the truck stopped to make way forother trucks for about 1000 s. Then, between 6501 s and 7000 s, the truck stopped to weigh and unload. The sampling fre-quency was 100 Hz, and the average speed was 10 km/h. The data collection system was DH5902, in which the strain signalsare amplied, ltered, and converted to digital signals. The conversion of the principal stress is done using the followingformula:

    (a) Point6 at static case with full load (b) Point 8 at static case with full load

    (c) Point 6 at hoisting condition (d) Point 8 at hoisting condition

    (e) Point 6 at obstacle surmounting condition

    (g) Point 6 at sinking condition

    (f) Point 8 at obstacle surmounting condition

    (h) Point 8 at sinking condition

    Fig. 12. The static test results of measuring point 6 and 8.

  • Onface orst stthe mdrivinincluding postress

    4.3. D

    Toresultstresscoefc

    Table 3Compar

    Posit

    StatiHoisObstSink

    S. Zheng et al. / Engineering Failure Analysis 48 (2015) 153165 163ison of the FEA and static test results at measuring point 6 and 8 (unit:MPa).

    ion P6.FEA (see Fig. 8) P6.TEST (see Fig. 12) Error (%) P8.FEA (see Fig. 8) P8.TEST (see Fig. 12) Error (%)

    c case with full load 13 12 8.3 23 22 4.5ting condition 100 19 373 24 22 9.1acle surmounting condition 47 46 2.2 43 40 7.5ing condition 45 34 32.3 53 50 6.0rmaxrmin

    Ee0 e90 21 l

    E21 lp

    e0 e45 2 e45 e90 2

    q1

    e strain gauge was afxed to the frame upper surface at measuring point 6, and another was afxed to the bottom sur-f the side rail at measuring point 8, both near the crack position. Fig. 14(a) and (b) show the three stages of the test. Theage was the loading period, during which the main impact was the payload being dropped into the dump truck, whenaximum stress was 55 MPa at measuring point 6 and 42 MPa at measuring point 8. The second stage was the full-loadg stage, during which the main impacts on the frame originated from the complex and irregular road conditionsing uphill and downhill, the tilting road surface, and the uneven ground. The maximum stress was 62 MPa at measur-int 6 and 54 MPa at measuring point 8. The third stage was the unloading of the truck, during which the maximumwas 77 MPa at measuring point 6 and 55 MPa at measuring point 8.

    etermining the matching coefcient for the static and dynamic tests

    determine the relationship between the static and dynamic tests, a matching coefcient is proposed for the twos. This is obtained from the improved frame by the application of the following equation. Furthermore, the dynamicof the original frame can be calculated by multiplying the FEA results by the matching coefcient. Matchingients of static and dynamic test are showed in Table 4.

    Cm rdmax rbasicrstc 2

    rstc Xn

    i1ri=n 3

    Fig. 13. Practical route on mine.

  • 164 S. Zheng et al. / Engineering Failure Analysis 48 (2015) 153165Cm: matching coefcient.rdmax: the maximum stress in dynamic test.rbasic: the basic stress in dynamic test.rstc: the average stress in static test.ri: the working stress in static test.

    (a) Time load history of measuring point 6

    (b) Time load history of measuring point 8

    Fig. 14. Stress data from measuring point 6 and 8.

    Table 4Matching coefcients of static and dynamic test (unit: MPa).

    Position Measuring point 6 (see Fig. 10) Measuring point 8 (see Fig. 10)

    Static test stress 26.5 29.3Dynamic test stress 77.0 55.0Basic stress 8.5 17.9Matching coefcient 3.2 2.5

    Table 5Dynamic stress of the original frame based on matching coefcients (unit:MPa).

    Position Measuring point 6 (see Fig. 10) Measuring point 8 (see Fig. 10)

    Static case with full load 451.8 410.8Hoisting condition 283.9 490.5Obstacle surmounting condition 535.7 585.1Sinking condition 448.5 497.9Braking condition 425.9 356.1

  • S. Zheng et al. / Engineering Failure Analysis 48 (2015) 153165 1654.4. Comparison between original and improved frames

    The results of the FEA and static testing for the improved design were found to be consistent. The maximum stress valueof the original frame, as observed in the dynamic test, could be determined by multiplying the static FEA result by the match-ing coefcient for the same given measuring point. For the modied design under the same loads, the stresses of theimproved frame could be smaller than the original frame. As shown in Table 5, some values of the stresses of the originalstructure at the measuring points near the crack were mostly close to or in excess of yield limit of 460 MPa. For this reason,cracks would appear within a relative short operating time.

    5. Conclusions

    Through analysis and the calculation of the matching coefcient for the static and dynamic tests, it was possible to cal-culate the dynamic data for the original frame. The resulting values were extremely large and the stress concentration wassevere. Furthermore, it was proven that the stresses in the frame near the RSFLS were a result of the unreasonable design.The two main reasons for the appearance of cracking in the original frame were the complex and large interaction loadsbetween the frame and the RSFLS, and another was the unreasonable design of the original frame at the RSFLS.

    The problemof the cracking could be resolved through the application of several improvements. Changing the frame sectionenhanced the bending strength of the frame.Modifying the openings in the frame ensured that the interface between the innerplate and bottom surfacewas continuous.Making the RSFLSwider reduced the bendingmoment between the bolts that affectsthe stresses. All the above measures were proposed to reduce the stress concentration, which leads to the cracking.

    The FEA method indicated that these improvements caused the stress in the frame to fall sharply from 235 MPa to 43 MPaat the position of the cracking. The required strength and technical requirements continued to be satised. The practicality ofthe improved frame was veried through static and dynamic tests. The improved frame has been in service for more thantwelve months without the appearance of any cracking.

    The matching coefcient used to calculate the dynamic testing results for the original frame involves a reference function,for which further research is necessary. In addition, a study should be conducted on a matching coefcient for the frame andrelated suspension system, and how it affects the dynamic response of the frame. This topic will be addressed in the nextphase of our research. However, the FEA, and static/dynamic tests conducted as part of this study cannot accurately analyzeor measure the residual stresses in the welds, although these should not substantially inuence the cracking problem.

    Acknowledgements

    This research was nancially supported by the National Natural Science Foundation of China under Contact No.51075179. The authors also would like to express their gratitude to Xiamen XGMA Heavy Industry Co. Ltd. for their supportof this work.

    References

    [1] Koyuncu Atayl, Gkler Mustafa_Ilhan, Balkan Tuna. Development of a design verication methodology including strength and fatigue life prediction foragricultural tractors. Int J Adv Manuf Technol 2012;60:77785.

    [2] Lee Young-Joo, Song Junho. Finite-element-based system reliability analysis of fatigue-induced sequential failures. Reliab Eng Syst Safety2012;108:13141.

    [3] Grbovic Aleksandar, Rasuo Bosko. FEM based fatigue crack growth predictions for spar of light aircraft under variable amplitude loading. Eng Fail Anal2012;26:5064.

    [4] Araujo LS, de Almeida LH, Batista EM, Landesmann A. Failure of a bucket-wheel stacker reclaimer: metallographic and structural analyses. J Fail AnalPreven 2012;12:4027.

    [5] Mi Chengji, Zhengqi Gu, Yang Qingquan. Frame fatigue life assessment of a mining dump truck based on nite element method and multibodydynamic analysis. Eng Fail Anal 2012;23:1826.

    [6] Zhengqi Gu, Mi Chengji, Wang Yutao, Jiang Jinxing. A-type frame fatigue life estimation of a mining dump truck based on modal stress recoverymethod. Eng Fail Anal 2012;26:8999.

    [7] Feng Guosheng, Jia Sumei, Li Ning, Lin Hao. Push rod failure analysis of a mining heavy dump truck. Eng Fail Anal 2012;25:1939.[8] Shao Yimin, Liu Jing, Mechefske Chris K. Drive axle housing failure analysis of a mining dump truck based on the load spectrum. Eng Fail Anal

    2011;18:104957.[9] Cernescu Anghel, Dumitru I, Faur N, Branzei N, Bogdan R. The analysis of a damaged component from the connection system of the wagons. Eng Fail

    Anal 2013;29:93107.[10] Duncheva Galya V, Maximov Jordan T. A new approach to enhancement of fatigue life of rail-end-bolt holes. Eng Fail Anal 2013;29:16779.[11] Bonjak Srdan M, Petkovic Zoran D. Designing-in failures and redesign of bucket wheel excavator. Eng Fail Anal 2013;35:95103.[12] Bonjak Srdan M, Petkovic Zoran D. Failure investigation of the bucket wheel excavator crawler chain link. Eng Fail Anal 2013;35:4629.[13] Topaa MM, Ercan S, Kuralay NS. Fatigue life prediction of a heavy vehicle steel wheel under radial loads by using nite element analysis. Eng Fail Anal

    2012;20:6779.[14] Hans Prem. Off-highway mine haul truck dynamics simulation. Society of automotive engineers. SAE Paper No. 981982.[15] Langer Thomas H, Christensen Brian B, Mouritsen Ole , Hansen Michael R. Optimization of front axle suspension system of articulated dump truck. In:

    The 1st joint international conference on multibody system dynamics; May 2010.[16] Poh KW, Dayawansa PH, Dickerson AW, Thomas IR. Steel membrane oors for bodies of large rear-dump mining trucks. Finite Elem Anal Des

    1999;32:14161.[17] Trebuna F, imck F, Bocko J, Pstor M. Analysis of causes of casting pedestal failures and the measures for increasing its residual lifetime. Eng Fail Anal

    2013;29:2737.

    Failure analysis of frame crack on a wide-body mining dump truck1 Introduction2 FEA models2.1 Frame and leaf spring seating2.2 Finite element model2.3 Loads and boundaries

    3 Static analysis3.1 FEA of the original frame3.1.1 Static case with full load3.1.2 Hoisting condition3.1.3 Obstacle surmounting condition3.1.4 Sinking condition3.1.5 Braking condition3.1.6 Summary of FEA results for original frame

    3.2 Improvement of the frame

    4 Static and dynamic tests4.1 Static test on the test site4.2 Dynamic stress test4.3 Determining the matching coefficient for the static and dynamic tests4.4 Comparison between original and improved frames

    5 ConclusionsAcknowledgementsReferences