摘要：Aimed at the reparable characteristic of the mechanical product, a method of the reliability model recognition of mean time between failures based on the BP neural network was developed, and a method of parameter estimation of reliability model based on adaptive linear network was proposed by means of theory of artificial neural network with MATLAB and reliability engineering theory. By network test and numerical simulation, reliability model recognition system and reliability parameters estimation system are verified. The results obtained from the simulation is better than those from the reliability paper for the common reliability model in engineering reliability and indicate the method is feasible. According to the method, the distribution model and function of reliability for mean time between failures of mechanical product were gained by this means.
关键词：mean time between failures; model recognition; parameter estimation; artificial neural network
