Optimization Coupling Approach for/with Non-Static Point Based CAT-models

Abstract

This paper proposes a concept to optimize tolerance synthesis process. Tolerance synthesis is getting more and more important to limit inaccuracies of the product during the manufacturing process. Today tolerance synthesis is used to assign the “right tolerance range” to the called out geometrical and dimensional tolerances. This can be achieved using tolerance analysis tools. However to reach the defined target value (e.g. maximum allowed deviation of a gap-/ flushness- measurement) there are multiple input parameters available towards the tolerance simulation model, which can be modified to a certain range. To address these issues, the approach in this paper is showing a way to couple an optimization tool to a CAT-simulation. This approach shall consider all the input parameters in a tolerance simulation model to stand out from existing tolerance optimization approaches. Also a unique feature of the approach is using a non-static simulation deck for the coupled tolerance analysis. This means all the tolerance simulation models are generated subjected to the deviating input parameters defined by the optimization tool. The optimization tool then uses the results of the CAT-simulations to calculate the response surface model. Using an optimization algorithm on this response surface model (RSM) helps to find an optimal set of parameters to reach the required quality features. Thus the tolerance planning engineer can give a well-founded feedback to the departments involved.