A math-based method is described for including design and process variation and uncertainty into the assessment of a product's performance. The computational efficiency and accuracy of several probabilistic methods are demonstrated using an automotive body-door system. The performance of the body-door system is assessed with respect to freedom from wind noise and water leaks and acceptable door closing effort using Monte Carlo simulations, advanced mean value plus, and adaptive importance sampling techniques. Physics-based models for calculating body-door seal gap and door closing energy are used to determine a design's performance in terms of freedom from wind noise and water leaks and acceptable door closing effort. To reduce the computation time, response surface approximations are created as an alternative to the more complex physics-based models. The accuracy of the approximate probabilistic techniques is evaluated by comparing their results to probabilities obtained from Monte Carlo simulations.
[1]
T. Cruse,et al.
Advanced probabilistic structural analysis method for implicit performance functions
,
1990
.
[2]
Y.-T. Wu,et al.
COMPUTATIONAL METHODS FOR EFFICIENT STRUCTURAL RELIABILITY AND RELIABILITY SENSITIVITY ANALYSIS
,
1993
.
[3]
R. S. Thomas,et al.
An Alternative Approach to Robust Design: A Vehicle Door Sealing System Example
,
1997
.
[4]
R. V. Lust,et al.
Probabilistic structural analysis—An introduction
,
1998
.
[5]
Donald R. Jones,et al.
Efficient Global Optimization of Expensive Black-Box Functions
,
1998,
J. Glob. Optim..
[6]
Donald R. Jones,et al.
A Taxonomy of Global Optimization Methods Based on Response Surfaces
,
2001,
J. Glob. Optim..
[7]
Thomas J. Santner,et al.
Design and analysis of computer experiments
,
1998
.