Robust Design Optimization for Multiple Responses Using Response Surface Methodology and Taguchi Approach: Solid Rocket Motor Application

Abstract

Robust design methodology is concerned with the process and the product design while minimizing the effects of uncertainties in design and external noises. Designers are often required to explore the design space to predict the behavior of a multi-disciplinary system and to select optimum design(s). Selection of optimum design(s) may become very cumbersome (if not impossible) if the system has conflictive multi-objectives. In robust design, optimum system design is sought with special emphasis on the insensitiveness of objective functions due to uncertainties. This study presents the robust design of a solid rocket motor having multiple ballistic performance responses. Aim is to minimize the mean and the variation of the chamber pressure of the motor. Three approaches are compared. In the first approach, the means system objective and constraint functions along with the variation of the chamber pressure are approximated by response surface models. Control factors for the optimum design are sought by using Desirability Function approach. In the second approach, Signal to Noise (S/N) Ratios for the objective and constraint functions are approximated by response surface functions and maximized. Representative optimal solutions for the first and the second approaches are compared. In the third approach, S/N Ratio for the chamber pressure is maximized subject to probabilistic constraints. Probability of failure of the solid rocket motor is also calculated at the selected optimum states along with the confidence limits through Monte Carlo simulations.