Developing a reliability model of CNC system based on weibull mixture distribution

The time-to-failure data is often used in reliability modeling, but inaccurate reliability models of the computer numerical control (CNC) system will be developed when ignoring the characteristics of the failure data itself-the failure mechanism. In order to solve the problem of insufficient utilization for failure information, a reliability model based on Weibull mixture distribution is proposed to deal with the failure information of CNC system. Considering the complexity of the model, the fuzzy clustering method is used to classify the time-to-failure data. Meanwhile, the parameters of the mixed Weibull distribution are obtained separately and the reliability model of the CNC system can be developed. Finally, the three kinds of evaluation indexes are used to compare the accuracy of fitting the data life grounded on the mixed Weibull model and the single Weibull model. And the validity of applying the Weibull mixture distribution to the reliability model of the CNC system is verified.