A PRACTICAL SIMULATION METHODOLOGY TO IMPROVE FATIGUE LIFE PREDICTION OF ENGINE OIL COOLER UNDERGOING PRESSURE CYCLE TESTING

Abstract

<p>Computer simulation is widely used to predict the fatigue life of engine oil coolers that fail under pressure cycles. The objective of this study is to develop a practical simulation methodology to accurately predict the fatigue life of an engine oil cooler undergoing pressure cycle testing. The study focuses on two key areas of the simulation process. First, it investigates the effect of using linear and nonlinear FEA to provide stress or strain results for subsequent fatigue analysis. Second, due to lack of fatigue material properties for the aluminum coreplate material, approximate material models derived from tensile properties are used in fatigue life calculation. The study has attempted to find out the material model that gives the best correlation in life prediction. The life prediction correlation based on the Seeger, the Modified Universal Slopes and the Modified Mitchell models, together with the Modified Universal Slopes-Al model, are evaluated.</p> <p>It is concluded that the Modified Universal Slopes-Al model, which is a re-assessment of the Modified Universal Slopes model based on the fatigue data of 16 wrought aluminum alloys, gives the best life prediction for simulations using either linear or nonlinear approaches. Life prediction using nonlinear finite element results together with this approximate material model is recommended to be the best approach. On the other hand, a simple and quick linear analysis, followed by fatigue life calculation using this material model still gives life estimates with an acceptable level of confidence.</p> <p>In the last part of the study, the life prediction performance using different strain-life criteria, together with either Morrow or Smith-Watson-Topper (SWT) mean stress correction, are evaluated. It is found that SWT mean stress correction method is worse than that of Morrow in EOC fatigue life prediction in both linear and nonlinear approaches. Using the principal strain criterion with SWT mean stress correction gives conservative life prediction in both approaches. On the other hand, there are no significant differences in life prediction correlations using the principal strain, the Brown-Miller combined strain and the maximum shear strain strain-life criteria, with Morrow mean stress correction. As such, the Brown-Miller combined strain criterion with Morrow mean stress correction is the recommended strain-life model used in fatigue life calculation.</p>