 |
International Society of Science and Applied Technologies |
| Robust Design Tools for Dynamic Systems Using Least-Squares Meta-Models | ||||
| Author |
Gordon J. Savage
|
|||
| Co-Author(s) |
Young Kap Son
|
|||
| Abstract | Robust design encompasses the theories and methodologies that make performance measures (responses, critical times, energy, etc.) invariant to uncertainties in the design variables (environmental conditions, manufacturing processes, material dimensions and material properties, etc.). In the following work, we develop three of the main tools needed for robust design based on Least-Squares meta-models. (Probabilistic methods are excluded herein.) The tools are applicable to both static and dynamic systems. We start with worst-case analysis where the lows and highs for all parameters are examined to assess the low and high points of a response. Then we present sensitivity functions wherein we find changes in a single response for a change in a single parameter. As a reasonable check of accuracy, a simple difference formula is used along with the mechanistic model. Finally, we investigate the important second-moment method wherein the means and variances of the responses are found from the means and variances of the multiple parameters. It is important to note that the necessary derivatives are obtained from the fitting polynomial used in the original meta-model. The results are compared to true values and found to be sufficiently accurate to provide guidance in robust design activities.
|
|||
| Keywords | Least-Squares Meta-model, Robust Design, Sensitivity function, Dynamic systems | |||
| Article #: RQD2026-110 | ||||
Proceedings of 31st ISSAT International Conference on Reliability & Quality in Design |