The Interval Estimation of MTBF Based on Markov Chain Monte Carlo Method

The distribution of time between failures of numerical control (NC) system follows the Weibull distribution, thus it’s estimation of Mean Time Between Failures (MTBF) in reliability engineering is of significance. But there are great difficulties in interval estimation of MTBF using traditional method for Weibull distribution. After the introduction of the approximate estimation, the Markov chain Monte Carlo (MCMC) method is proposed. Combined with the specific characteristics of two-parameter Weibull distribution, Markov chain is established to calculate the interval estimation of MTBF, which solves the problems effectively. And MCMC is more accurate than that of engineering approximation. By analyzing various results in different conditions of MCMC transition kernel, the paper proves that MCMC is a good method for solving interval estimation of Weibull distribution parameters, which has systematic solution process and good adaptability. It greatly enhanced the robustness, effectiveness and accuracy of the calculation.