Application of fuzzy numerical techniques for product performance analysis in the conceptual and preliminary design stage

Uncertainty and variability modelling tools greatly enhance the value of virtual prototypes at the different design stages of a CAE process. The fuzzy analysis technique is suited to deal with models containing subjective non-deterministic parameters. This technique is finding its way to different disciplines of mechanical engineering. The objective of this paper is to increase the value of this technique in early stages of mechanical design procedures. For this purpose, new numerical procedures are proposed. First, the degree of influence is introduced. This new concept measures the relative effect of highly uncertain design properties on the performance of a design. Next, this paper proposes a new reduced optimisation scheme in order to improve the computational efficiency of the interval analysis, which is at the core of the implementation of the fuzzy technique. The practical applicability of the newly developed procedures is demonstrated on two numerical applications from the automotive industry. The analysed models represent the design at the conceptual stage, and contain parameters with a high and subjective level of uncertainty. The parametrised models are used to demonstrate the value and efficiency of the developed numerical procedures: significant parameters are identified using the degree of influence analysis, the optimal configuration is identified through an interval analysis based on the reduced optimisation scheme, and finally the fuzzy technique is applied as design space exploration tool.