Multidisciplinary Design Optimization Supported by Knowledge Based Engineering

10Uncertainty-Based Multidisciplinary Design Optimization*

10.1 Introduction

In practical product engineering, there exist a vast range of uncertainties arising from lack of knowledge in the product system itself and its operational environment. As an example, consider a structural design where uncertainties may arise from prediction errors induced by model assumption and simplification (Morris, 2008), structure performance variation resulting from the manufacturing tolerances, deviations in the material properties, and the uncertainties in the operational loads. These may cause the system performance to change or fluctuate or even result in severe function fault and mission failure. So it is clearly essential to take uncertainties into account from the beginning of the system design when optimal solutions are being sought. In a traditional optimal design process to account for uncertainties, the constraints imposed on the design are often reformulated with empirical or other predefined safety factors based on the marginal design philosophy to maintain the feasibility of the system in presence of uncertainties. However, this is based mainly on past experience for defining the margins which may be inappropriate or obsolete for new products. So we need to develop more advanced and accurate analytical approaches based on mathematical uncertainty theory to handle uncertainties systematically and rationally during the design process. These new approaches are termed uncertainty-based ...