Multiple Weibull Statistical Model of Random Censored Data of NC Machine Tools and Optimal Estimation of Parameters

Abstract

According to the truncation feature of NC machine tools, this paper adopts Johnson rank adjustment method to deal with censored data, finding the distribution model of time between failures. At the same time, in order to increase the accuracy of parameter estimation of reliability data distribution model of NC machine tools, and to avoid the shortcoming that conventional optimization algorithms is difficult to get global optimal solution due to the influence of iteration initial value, this paper uses particle swarm optimization to solve the parameter of weibull mixture model. The result shows that particle swarm optimization can balance solution efficiency and convergence performance, it is not only feasible to estimate the parameter of mixture weibull distribution, but also to get more accurate results.