Accelerated Lab Test Methodology for Steering Gearbox Bracket Using Fatigue Damage and Reliability Correlation.
Steering system controls the direction of an automotive vehicle; thus, its components should be designed for higher durability and reliability. During product development stage, steering gearbox bracket designs were validated in our customized proving ground. Even though after design validation, few failures of steering gearbox bracket were reported at the customer site (Refer Figure-1&2). Hence, it was decided to revalidate the steering gearbox bracket but at a shorter span of time, culling for an accelerated test .
The forces acting on the steering gearbox are directly transferred to the bracket which is used to integrate it with vehicle chassis.
FIELD FAILURE ANALYSIS
Field Data Analysis
Field failure data of steering gearbox bracket was collated and reliability analysis was performed to infer the failure trend [2, 3]. Weibull plot made the best fit with a correlation coefficient of 0.988 as shown in Figure-3.
From the reliability analysis (Table-1&2), it was found that 10% of failures (B10 Life) occur at 205,245 km, whereas the intended vehicle life is 500,000 km. The shape factor([beta]) and scale factor([eta]) were 4.30 and 346,234 respectively.
Note: Percentile represents the estimated field failure kilometer. This result will be referred in the later sections to compare with lab testing.
Failure Cause Analysis
There are many probable reasons for failure of a component. Knowing the potential causes of failures is the fundamental to prevent them . Figure-4 illustrates fishbone diagram to categorize the potential causes for failure of steering gearbox bracket [4,5].
Some variables that could be the causes of the failure were eliminated after detailed analysis of each case, and were not considered to be origin of failures. They are: Inadequate quality control, assembly error, corrosion/thermal environment. Thus, we focused on analyzing the major causes i.e. use of wrong material, manufacturing defects, design error and overloading.
Chemical composition, microstructure and brinell hardness tests were conducted on three of the field samples to evaluate the failure causes. Additionally, tension testing was conducted on three sample specimens to evaluate mechanical strength of the material. Tested samples met the material specifications.
From the results of metallurgical, hardness and tension testing (Refer Figure-5 and Table-3), it was found that the material meets the specification [6,7]. Also, there were no manufacturing defects present in the component . Hence the quality and mechanical strength of the component were removed from the probable causes of failure.
Since the failure was neither due to improper material nor manufacturing defect; the focus was mainly on analyzing the forces and moments acting on the steering gearbox bracket.
Forces on Steering Gearbox Bracket
The road reactions are transmitted to the steering gearbox through draglink. Hence it induces forces and moments on the steering gearbox bracket. Additionally, the weight of the steering gearbox and its corresponding moment plays a significant role (Refer Figure-6).
The steering gearbox design specifications are tabulated in Table-4. Calculated draglink force at 100% rated torque condition is 27.5 kN.
To comprehend the forces acting on the draglink in field conditions, draglink was instrumented to acquire the strain corresponding to the forces in the field. For this, draglink was calibrated by applying a known force using hydraulic actuator (Refer Figure-7&8).
Draglink force (kN) = [0.0798*microstrain] + 0.2676 (1)
Equation-1 was used to convert the measured strain in terms of draglink force. Prior to field data acquisition, proving ground data was acquired to analyze the reason for not simulating the failure during its initial design validation process.
Proving Ground Data Acquisition and Analysis
The calibrated draglink was assembled in a 25T vehicle (field failure model) and the draglink force data was acquired at 6 of our PG tracks with laden and overload conditions (Refer Figure-9).
From the draglink force time history in Figure-10, the maximum tension and compression force on the draglink was 28.5 kN and -33.34 kN respectively .
From the analysis of draglink force distribution in Figure-11, 1% of the draglink force was greater than 27.5 kN (100% rated).
Field Data Acquisition and Analysis
Draglink force data was acquired at customer site with typical customer duty cycle in five routes. From the draglink force time history in Figure-12, the maximum tension and compression force on the draglink was 37.7 kN and -39.8 kN respectively.
From the analysis of draglink force distribution in Figure-13, 6.7% of the draglink force was greater than 27.5kN.
From the road load data analysis, 1% of the draglink force was greater than rated load in PG whereas in customer site it was 6.7%. Hence, root cause for the failure was overloading due to typical customer duty cycle.
However, this failure is not acceptable in terms of reliability; hence the bracket needs to be redesigned to eliminate failures in useful life period.
However after redesigning, the improved design samples need to be validated. This can be done by modifying the proving ground cycle or by accelerated lab testing.
Accelerated lab testing is preferred, since it is less time consuming and the methodology can be correlated after simulating the field failure.
ACCELERATED LAB TESTING
Existing Design Validation
Lab test facility was developed by reproducing similar boundary conditions as in vehicle. A Servo hydraulic actuator was used to apply force on the draglink. Steering gearbox with its mounting bracket was assembled in a rigid fixture to ensure that entire force acts on the steering gearbox bracket.
The hotspot (high strain location) identified from CAE analysis was found to be as same as field failure location. Hence, a rosette strain gauge was bonded at the hotspot to measure strain in three directions.
Cyclic force in steps of 2.5kN till 40 kN (maximum force in field data acquisition) was applied using a servo-hydraulic actuator for 10 cycles. For each step, observed peak strain value in three directions were tabulated and principal strain was calculated using a data analysis software [10,11].
Using the calculated principal strain and fatigue properties of the bracket material, fatigue damage was calculated using strain-life approach (with no correction method). Refer Figure-16 & Table-5.
From Table-5, at rated load (27.5kN); the calculated principal strain (1553 [micro][??]) is less than the strain corresponding to 0.2 % proof strength of the material (1975 [micro][??]). Hence design is safe for the rated load applications. However, failures were observed in the field due to the typical customer duty cycle with overloading.
Draglink force cycle was counted in steps of [+ or -]2.5kN for the customer duty cycle data and extrapolated to the vehicle life of 500,000 km. Cumulative fatigue damage was calculated and tabulated in Table-6.
From the cumulative fatigue damage calculation results in Table-6; it is observed that within [less than or equal to]12.5kN of draglink force, the damage induced on the steering gearbox bracket is negligible. Also, 80% of the cumulative damage is accumulated when draglink force is greater than the rated load. i.e. 20%, 53%, 27% of the cumulative fatigue damage cycles were from [less than or equal to] 100%, 100-125%, 125-150% of rated load respectively.
To develop the lab test sequence, 100%, 125%& 150% of steering gearbox design torque was considered and number of cycles required to attain the equivalent cumulative fatigue damage was calculated.
To simulate field conditions, target cycles in Table-7 was arranged in a block cycle sequence as in Table-8.
For 90% Reliability and 50% Confidence level, six samples of existing design were tested in lab for the test conditions derived in Table-8 . Field failure was simulated in all the samples.
Refer Table-9 and Figure-18 for lab test results of existing design samples.
All samples failed to meet the target specified in Table-8.
From the probability table of existing design (Table-10), it was found that 10% of failures (B10 Life) occur at 24578 cycles. The shape factor([beta]) and scale factor([eta]) were 9.92 and 30836 respectively.
Reliable life (B5, B10 and B50) of existing design samples in both field kilometer and lab cycles are tabulated in Table-11 (Refer Table-2 and Table-10).
Since, our requirement is 90% reliability for 500,000 km in field, B10 life was taken for comparison.
Equating the B10 life of existing design in both lab and field, lab cycles equivalent to the field km was derived [13, 14].
24578 cycles = 205,245 km in field Hence, 1 cycle = 8.36 km in field (2)
From Equation-2, for the intended vehicle life of 500,000 km in the typical customer duty cycle; the steering gearbox bracket has to survive 59875 cycles (120 repeats) in the block cycle sequence mentioned in Table-8.
Target lab cycles arrived in Table-11 based on reliability correlation was found to be severe than lab cycles derived in Table-8 based on cumulative fatigue damage. Considering higher level of severity in reliability correlation method, the block cycle test sequence mentioned in Table-8 needs to be extended for 120 repeats.
New Design Validation
The steering gearbox bracket was redesigned to eliminate failures in the useful life period. Six samples of new design were tested in lab with the same test conditions used for existing design (Table-8, 11). Refer Table-12 and Figure-19 for lab test results of new design.
All samples exceeded the specified target and test was continued till the failure of samples.
From the probability table of new design (Table-13), it was found that 10% of failures (B10 Life) occur at 77414 cycles. The shape factor([beta]) and scale factor([eta]) were 12.98 and 92060 respectively.
From the lab test results, it was confirmed that new design samples have 300% more life (based on B10 Life comparison) than the existing design samples and it is approximately equivalent to 646,465 km in field which is surpassing the vehicle intended life of 500,000km. Hence, it was implemented in our vehicle.
SUMMARY AND CONCLUSION
* Potential causes for failure of existing steering gearbox bracket in field conditions were analyzed using material and road load data analysis. It was found that overloading due to typical customer duty cycle was the root cause for failure.
* Accelerated lab test facility (maintaining vehicle boundary conditions) was developed and block cycle sequence for lab testing was derived based on cumulative fatigue damage analysis.
* Field failure was simulated with existing design and reliability analysis was performed to find the reliable life (B10 Life) in both lab and customer duty cycle.
* Lab cycle target equivalent to the intended vehicle life was derived by comparing the damage severity in cumulative fatigue damage and reliability correlation methods. New design of steering gearbox bracket was validated and it surpassed the target requirements.
* Similar methodology can be used to derive block cycle target for accelerated testing of any automotive components.
[1.] Senniappan, M., More, R., Bhide, S., and Gowda, S., "Optimization of Commercial Vehicle's Steering Tie Rod Arm Design Based on Strain Life Approach," SAE Technical Paper 2016-01-0381, 2016, doi:10.4271/2016-01-0381.
[2.] Patrick D.T.O'Connor, "Practical Reliability Engineering", Fifth edition 2012, John Wiley & Sons Ltd.
[3.] "Minitab 16 Software reference", Release 16.1.0 , Year 2010
[4.] Guimaraes a A.V., Brasileiro P.C., "Failure analysis of a half-shaft of a formula SAE racing car", Elsevier - Case Studies in Engineering Failure Analysis 7 (2016) 17-23,2016
[5.] Rao Vinay N., Eischen Jeffrey W., "Failure analysis of mixed mode crack growth in heavy duty truck frame rail", Elsevier-Case Studies in Engineering Failure Analysis, 5-6 (2016)
[6.] Bureau of Indian standards, "Iron castings with spheroidal or nodular graphite - Specification", Third revision, IS1865 : 1991.
[7.] ISO International Standard, "Cast Irons - Part 1: Materials and properties for design", ISO/PDTR 10809-1,2008
[8.] More A.P. and Baxi R.N. Dr., "Review of Casting Defect Analysis to Initiate the Improvement Process", Int J Engg Techsci Vol 2(4) 2011, 292-295,2011
[9.] "RPC Pro Tools Software Reference", MTS Systems Corporation , Manual Part no.100-114-918 K,2010
[10.] Yung-Li Lee and Jwo Pan, "Fatigue Testing and Analysis (Theory and Practice)", Elsevier's Science & Technology, First edition 2005, pp.16-25.
[11.] "GlyphWorks Fatigue Theory Guide", 2012 HBM UK
[12.] Dodds Colin J, "Structural testing of complete vehicles, aggregates and components in the laboratory - The Test Engineer's Handbook", Second edition, 2015.
[13.] Schuh, F., Corso, L., and Hoss, L., "Methodology for Fatigue Life Durability Prediction Applied to Commercial Vehicles," SAE Technical Paper 2014-36-0038, 2014, doi:10.4271/2014-36-0038.
[14.] Subramanian, D., Sridhar, N., Karthikeyan, N., and Chandra, V., "Reliability Testing: Predictor Effect Analysis on Engine Mounts," SAE Technical Paper 2015-01-2757, 2015, doi:10.4271/2015-01-2757.
N. Obuli Karthikeyan, R. Dinesh Kumar, and V. Srinivasa Chandra
Ashok Leyland Ltd.
Obuli Karthikeyan N
Deputy Manager, Component Test Laboratory
Ashok Leyland Technical Centre, Chennai, India
The authors would like to express their gratitude to
Dr.N.Saravanan, Senior Vice President - Product Development
Mr.S.Ravishankar, GM - Product Development
Mr.V.Vijaykumar, AGM- Product Development
Mr.Shiju Subramaniom, Divisional Manager - Product Development
Mr.R.Senthil Kumar, Senior Manager - Product Development
Ashok Leyland Ltd. for providing guidance and permission to publish this work.
OEM - Original Equipment Manufacturer
PG - Proving Ground
CAE - Computer Aided Engineering
Table 1. Characteristics of Distribution Standard 50.0% Normal CI Estimate Error Lower Upper Hean (MTTF) 315147 25843.2 298190 333069 Standard Deviation 82745.3 27343.7 66213.1 103405 Hedian 317967 27312.7 300069 336933 Table 2. Probability table of field failure data Standard 50.0% Normal CI Percent Percentile Error Lower Upper 1 118S89 49957.6 89547.5 157845 2 139S31 50251.0 109732 178185 3 153828 49812.8 123646 191378 4 164660 49168.8 134622 201399 5 173632 48452.8 143843 209591 6 181367 47713.4 151878 216581 7 188212 46971.7 159053 222717 8 194385 46237.2 165571 228212 9 200028 45514.8 171569 233209 10 205245 44806.8 177143 237805 20 244344 33546.9 219680 271777 30 272478 33568.9 250752 296088 40 296198 29759.0 276790 316965 50 317967 27312.7 300069 336933 60 339271 26681.2 321744 357753 70 361495 28484.6 342784 381227 80 386717 33552.5 364736 410023 90 420277 44061.7 391585 451073 Table 3. Result of Material analysis Details Specification Observation Sample-1 Sample-2 Sample-3 Material Grade SG 500/7 SG 500/7 Hardness 160-240BHN 212 207 223 Nodulargraphite Nodularity: Graphite Nodular graphite Micro structure Form- V & VI Nodularity: Graphite Form - VI Matrix-Ferrite Ferrite (50%) Pearlite (50%) Pearlite Table 4. Steering gearbox specification Steering gearbox rated torque T 6780 Nm Drop arm length L 0.246 m Draglink force (at 100% rated torque) [F.sub.R] 27.5 kN Table 5. Principal strain and Fatigue damage calculation Minor Draglink Major Principal Fatigue Force Principal damage for Strain ([+ or -]kN) Strain ([micro][??]) a Cycle [micro][??] 2.5 128 -122 0.0000 5 250 -243 0.0000 7.5 383 -358 0.0000 10 516 -482 0.0000 12.5 633 -624 0.0000 15 782 -750 4.208 E-08 ... ... ... ... 27.5 1553 -1519 1.207E-05 ... ... ... ... 35 2163 -2116 6.333E-05 ... ... ... ... 40 2655 -2708 14.803E-05 No. of Cycles Draglink for failure Force (1 / Fatigue ([+ or -]kN) damage for a cycle) 2.5 Infinite 5 Infinite 7.5 Infinite 10 Infinite 12.5 Infinite 15 23.76E06 ... ... 27.5 82.84E03 ... ... 35 15.79E03 ... ... 40 6.75E03 Table 6. Cumulative fatigue damage calculation Extrapolated to 500,000 km in field Draglink Fatigue No. of cycles Force damage for ([+ or -]kN) a Cycle 2.5 0.0000 ... ... 0.0000 ... 15 4.208E-08 41174 ... ... ... 27.5 1.207E-05 17156 ... ... ... 40 ... ... Cumulative Fatigue Damage Extrapolated to 500,000 km in field Draglink Fatigue Damage (Damage for Force a Cycle * No. of cycles) ([+ or -]kN) 2.5 0.0000 ... 0.0000 15 0.0017 ... ... 27.5 0.2071 ... ... 40 ... "X" Table 7. Target cycles To create Cumulative fatigue damage of "X" (for 500,000 km) % of Design Draglink No. of Cumulative Sl.No. Torque (%) force (kN) damage Cycles 100 [+ or -]27.5 20% 30000 1 125 [+ or -]35 53% 16000 2 150 [+ or -]40 27% 4000 3 Total number of cycles 50000 Table 8. Block cycle - Test Sequence Block Cycle for Fatigue Test (1 Repeat = 500 cycles) Design Draglink No. of No. of Sequence Torque (%) force (kN) Cycles Repeats 1 100 [+ or -]27.5 300 2 125 [+ or -]35 160 100 3 150 [+ or -]40 40 Table 9. Lab test results of existing design samples Existing Faile Sample No. of Cycles No. of Repeats 1 24516 49 2 31172 62 3 27586 55 4 29361 59 5 30856 62 6 33258 67 Table 10. Probability table of existing design Standard 50.01 Normal CI Percent Percentile Error Lower Upper 1 19394.6 4639.73 16504.6 22790.7 2 20S08.8 4253.00 18129.1 23884.5 3 2163S.0 3991.09 19156.5 24554.2 4 22337.8 3788.04 19923.5 25044.6 5 22S57.9 3620.19 20542.0 25434.8 6 23294.1 3476.08 21063.7 25760.8 7 23671.5 3349.19 21517.0 26041.8 8 2400S.2 3235.47 21919.2 26289.8 9 24305.1 3132.15 22281.7 26512.3 10 24578.1 3037.32 22612.5 26714.5 20 26509.4 2356.60 24966.6 28147.5 30 27792.7 1921.45 26526.5 29119.4 40 28817.5 1619.68 27745.5 29930.9 50 29717.9 1427.56 28770.4 30696.5 60 30565.8 1350.36 29668.4 31490.3 70 31418.8 1403.14 30486.5 32379.6 80 32351.6 1605.79 31286.4 33453.0 90 33540.9 2024.90 32202.5 34934.8 Table 11. Reliable life correlation for existing design Field Life kilometer Lab test cycles B5 173,632 22337 B10 205,245 24578 B50 315,147 29326 Table 12. Lab test results of new design samples New Failed at Sample No. of Cycles No. of Repeats 1 87654 175 2 67533 135 3 75421 151 4 70876 142 5 83654 167 6 69835 140 Table 13. Probability table of new design Standard 50.0% Normal CI Percent Percentile Error Lower Upper 1 64602.1 10541.5 57869.2 72118.2 2 68170.1 9553.63 62021.5 74928.3 3 70359.6 8914.39 64596.7 76636.6 4 71964.0 8431.93 66495.7 77882.0 5 73240.4 8040.60 68013.1 78869.6 5 74305.8 7709.50 69283.6 79692.1 7 75223.6 7421.44 703E0.9 80399.6 8 76032.3 7165.82 71349.4 81022.5 9 76756.8 6935.61 72218.5 81580.2 10 77414.3 6725.91 73008.2 82086.5 20 82019.1 5260.42 78546.7 85645.1 30 85035.2 4352.62 82149.5 88022.2 41 87419.9 3725.43 84942.9 89969.1 50 89498.6 3306.14 87296.2 91756.6
|Printer friendly Cite/link Email Feedback|
|Author:||Karthikeyan, N. Obuli; Kumar, R. Dinesh; Chandra, V. Srinivasa; Murali, Vela|
|Publication:||SAE International Journal of Commercial Vehicles|
|Article Type:||Technical report|
|Date:||May 1, 2017|
|Previous Article:||Design and Analysis of Lifting Pusher Drop Axle for Heavy Commercial Vehicle.|
|Next Article:||Design of a Thermoelectric Generator for Waste Heat Recovery Application on a Drivable Heavy Duty Vehicle.|