Optimising the design of mechanical components for reliability and cost

This paper discusses a design optimisation methodology that uses probabilistic methods. The methodology addresses product performance, reliability and cost that allow the engineer to evaluate the impact of design changes on all relevant measures of the design simultaneously. By considering variation in the design parameters, the influence of uncertainty can be determined and the design can be optimised to prevent early failures while minimising overall cost. The sensitivities of the design variables to the statistical distribution parameters are determined. The sensitivities are then used in an optimisation model to determine the best combination of design parameters including the manufacturing tolerances and acceptance testing routines. A demonstration problem that illustrates the application of the method is provided.