Tolerance allocation in electric machine based on stochastic response surface methodology

In electric machine design, tolerance analysis and allocation must be adopted to predict the effect of mechanical tolerances or assembly inaccuracies as regards the performance of the machine designed by optimization technique. Accordingly, this paper proposes stochastic response surface method combined with genetic algorithm to achieve more accurate tolerance design. The proposed method is applied for the allowance on torque characteristic of a switched reluctance motor with barrier.