Brake Analysis and NVH Optimization Using MSC . NASTRAN

Brake Analysis and NVH (Noise, Vibration and Harshness) Optimization have become critically important areas of application in the Automotive Industry. Brake Noise and Vibration costs approximately $1Billion/year in warranty work in Detroit alone. NVH optimization is now increasingly being used to predict the vehicle tactile and acoustic responses in relation to the established targets for design considerations. Structural optimization coupled with frequency response analysis is instrumental in driving the design process so that the design targets are met in a timely fashion. Usual design targets include minimization of vehicle weight, the adjustment of fundamental eigenmodes and the minimization of acoustic pressure or vibration at selected vehicle locations. Both, Brake Analysis and NVH Optimization are computationally expensive analyses involving eigenvalue calculations. From a computational sense and the viewpoint of MSC.Nastran, brake analysis exercises the CEAD (Complex Eigenvalue Analysis Dmap) module, while NVH optimization invokes the DSADJ (Design Sensitivity using ADJoint method DMAP) module. In this paper, two automotive applications are presented to demonstrate the performance improvements of the CEAD and DSADJ modules on NEC vector-parallel supercomputers. Dramatic improvements in the DSADJ module resulting in approx. 8-9 fold performance improvement as compared to MSC.Nastran V70 were observed for NVH optimization. Also, brake simulations and experiences at General Motors will be presented. This analysis method has been successfully applied to 4 different programs at GM and the simulation results were consistent with laboratory experiments on test vehicles.