Concurrent optimisation of parameter and tolerance design via computer simulation and statistical method

摘要

This study optimises component parameters and component tolerances simultaneously via computer simulation and response surface methodology (RSM). The approach first generates a set of experimental data through computer simulation, then the data are converted to a total cost as a response value before applying RSM for statistical analysis and mathematical optimisation. The response value (total cost) includes quality and related costs which reflect the combined effect of the parameter and tolerance values being assigned. The results provide designers with the optimal component design values, the critical components, and the response function of a product or process design, which are very important to know during design activities as they give designers information about repeated applications, accurate feedback and appropriate suggestions, particularly under uncertain design conditions. Three examples are provided. They are mechanical assembly design, machining process planning, and electronic circuit design.