Nonlinear mixed reliability model with non-constant shape parameter of aviation cables

Abstract

In this paper, a reliability model of aviation cables that using nonlinear mixed model and Bayesian estimation is proposed. Considering the small sample character of accelerated life test data, the sample is expanded by Bootstrap method, least square regression is used to get the initial Weibull distribution parameters, and non-constant shape parameter is got by maximum likelihood estimation. Then, considering the uncertainty of the scale parameter, a nonlinear mixed model is established, and Bayesian estimation based on the prior information is used to make the reliability model more accurately. Accelerated life test case study is given to illustrate the performance of the proposed method under temperature and vibration stresses. Simulation results indicate the effectiveness of proposed model compared with nonlinear mixed model or Bayesian estimation individually. The proposed model can be extended to the reliability modelling of other insulated cables.