Vanderbilt University
Engineering Capability Brief

Reliability and Optimization of Automotive Systems

S. Mahadevan, T. Zou, A. Sopory
Civil and Environmental Engineering, Vanderbilt University
VU Station B 351831, Nashville, TN 37235; (615)-322-3040; fax (615)-322-3365
E-mail: sankaran.mahadevan@Vanderbilt.edu

Overview: In engineering quality management, the assessment of a product design's compliance to customer-focused specification needs to include design and process variation and uncertainty. Traditional test-based methods to assess product reliability are too expensive for complicated systems. Therefore, model-based simulation and design methods have gained popularity in recent years.

Vanderbilt University is combining system computational models with probability and optimization methods to predict the performance reliability of vehicle systems and subsystems, funded by General Motors Corporation. Current reliability estimation methods are based on expensive full-scale tests. At Vanderbilt University, analytical and simulation-based methods are combined to perform accurate and efficient reliability analysis. For nonlinear problems with implicit limit state functions, an adaptive response surface method is combined with a nonlinear optimization technique to identify the most probable failure condition and provide sensitivity and first-order estimates for both component and system-level reliability. An adaptive Monte Carlo sampling method, which provides an accurate estimate of the reliability with a small number of simulations, then improves this first-order estimate. The reliability estimates and sensitivity information are then used to develop a simulation-based reliability-based design optimization (RBDO) and robust design optimization framework that enables decision-making under uncertainty.

Example Applications: The developed probabilistic methods are applied to the body-door subsystem as a demonstration problem. Two quality issues are considered - wind noise and door closing energy, with the limit state values calculated by the finite element software. The random variables include spatial variations, geometric properties and mechanical properties. The two quality issues are found to conflict for certain random design variables, which makes this example a bi-objective RBDO problem. Various multi-objective optimization approaches are applied, and a series of trade-off solutions are provided for further decision making.

Body FEM Model
Door FEM Model

Potential Applications: Most practical system computational models are nonlinear and do not have explicit solutions. The application of model-based simulation to reliability assessment and design optimization for such problems can be computationally expensive. The developed reliability analysis and optimization methods achieve remarkable efficiency and accuracy, and are applicable to a wide variety of RBDO problems with competing objectives and reliability constraints at both component and system levels, or robust design problems in which optimizing the mean and minimizing the variance of the objective are of concern simultaneously.

Publications:
Zou, T., Mahadevan, S., Mourelatos, Z., and Meernik, P., "Reliability Analysis of Automotive Body-Door Subsystem," Reliability Engineering and System Safety, 78(3), 315-324, 2002.
Zou, T., Mahadevan, S. and Mourelatos, Z., "Reliability-Based Evaluation of Automotive Wind Noise Quality," Reliability Engineering and System Safety, 82(2), 217-24, 2003.

ACKNOWLEDGEMENTS
This study is supported by funds from the National Science Foundation through the Vanderbilt University IGERT program on Risk and Reliability Engineering and is supported, in part, by General Motors Research and Development.

 

©