Tolerance optimization of non-normal distribution characteristics for mechanical parts based on asymmetric cubic quality loss

Abstract

Purpose: The error distribution of product quality is different, not all follow the normal distribution, and sometimes there is a non-normal distribution of the product. Therefore, the triangular distributions and trapezoidal distributions are investigated based on asymmetric cubic quality loss. Design: Appropriate changes in the process mean value can effectively reduce the quality loss of the product. In order to study the optimization problem of the process mean with linear asymmetric quality loss, a mathematical model for calculating the expected quality loss considering the optimization of the process means is established. Finding: The proposed model is solved, and the analytical solution of the optimal process mean is obtained. Moreover, a numerical model of inherent reliability that conforms to the triangular distribution is introduced. Due to the existence of measurement fluctuations, it will bring fluctuation loss. The measurement loss is improved. Finally, the objective function calculates the optimal tolerance is calculated with the minimum total cost. An example is given to illustrate the effectiveness of the model.