Research on nonparametric method of reliability evaluation in context of mixed distribution

With regard to the small sample and randomly right censored problem of field data in reliability evaluation of aviation equipment, this paper compares the nonparametric method of reliability evaluation. We design a simulation method to compare the advantages and disadvantages for reliability evaluation under mixed distribution with Monte Carlo method. The method takes double mixed Weibull distribution for an example, comparing four nonparametric evaluation methods through producing field failure data of aviation equipment with features of randomly censored. The simulation result shows the optimal and worse order is successively sorted by Kaplan-Meier method and Mean Rank Order method, then Herd-Johnson method and last Nelson-Aalen method under double mixed Weibull distribution.