CAE Prediction and Test Correlation for Body Sheet Metal 2006-01-0828
Finite element based stress analysis and fatigue predictions are practiced routinely in automotive body structural design and development. The accuracy of these simulation results is not fully understood or at least not well documented. Automotive body structures have many kinds of notches, metal thinning due to stamping and cold working etc. Modern fatigue assessment tools do take into account many of these complexities by Neuber corrections, mean-stress correction, critical plane selection, etc. Other challenges exist in the sensitivity to element quality, including warpage, size, element type, interpretation of results, etc. This case study is based on static loading and accelerated fatigue test conducted on a front-end body buck. The stress and fatigue correlations are designed to build confidence in the model and load inputs. The fatigue results are intended to reproduce durability issues that developed during a proving ground test and were then used to verify potential fixes. Two front end structure designs are analyzed in this study, design “A” the original design and design “B” the new and improved design. The overview covers both experimental testing and predictions from the finite element analysis. The first part consists of stress correlation for both design “A” and “B”, where stresses from the NASTRAN analysis are correlated to measured strain gage data from the laboratory. With the good correlated stress analysis, a sheet metal fatigue analysis for design “B” was conducted utilizing an in house fatigue software FLAP and is correlated to 5 test samples. Better approaches for making approximations of fatigue life, problems that may be encountered for both over and under predictions and general recommendations are included.