Robust Design Methodology of Topologically Optimized Components Under the Effect of Uncertainties

Abstract

In the present work, a robust design methodology is presented for topologically optimized components. An integrated methodology combining design of experiment and reliability based topology optimization is proposed to capture the performance of optimized components in realistic environment including various uncertainties. In the present work, Mechanical Advantage, output displacement and the maximum von-Mises stress values are considered as performance functions. Volume fraction, force and aspect ratio are set as design-factors. The uncertainties of design factors are incorporated in the design using reliability method. The uncertainties of non-controllable factors are simulated by creating random field of material properties. Considering uncertainties, the performance of the topology optimization problem is simulated for a space of design factors. The simulated results are analyzed using statistical tools such as, analysis of mean. This technique helps to identify statistical significance and the effect on the performance variations. The proposed methodology is illustrated on a Force inverter. This analysis provides a design methodology in a realistic environment that helps in achieving targeted performance and robust design.