Papers by Keyword: Censored Data

Abstract: The calculation of mean time between failures (MTBF) is of significance in reliability engineering. The precondition before calculation of MTBF is to make sure what failure model the numerical control (NC) system follows. For different distributions, the corresponding calculation methods of MTBF are always different. Undertaken a 2 year period of timing censored test with replacement of NC lathe, authors apply a type I censored likelihood function to make the distribution fitting of time between failures of NC system. The tests of goodness-of-fit applying Hollander’s method demonstrate that the time between failures of NC system follows the Weibull distribution. The conclusion not only deeply analyzes the NC system failure law, but also establishes the basis of calculation for the mean time between failures of NC system based on censored data with replacement.

Abstract: Integral inference method for multi-state type-Ⅱcensored data is presented, which is used to evaluate the rupture life of composite material. The best unbiased estimators for the model parameters, the standard deviation and their covariance are also given. The confidence limits of the acceleration equation and the confidence intervals of percentiles for Weibull distribution are discussed in detail. According to the presented method, the experimental information of accelerated rupture life is fully exploited, and the cross information among test data in different stresses are analyzed as an integration.So its information quantity is much bigger than traditional group test method. The method provides a feasible and effective approach to rupture life evaluation and reliability estimation of composite material.

Abstract: An algorithm is presented for maximum likelihood estimation (MLE) of three-parameter Weibull distribution in this paper, which is not only suitable for uncensored data but also appropriate for various censored data. Firstly, a unified likelihood equation under censored data and uncensored data was built. Here, the second-order convergent Newton-Raphson iteration method was applied to solve the MLE of two-parameter Weibull distribution, which had only scale parameter and shape parameter with a given positional parameter. Then, the steady and rapid Brent search method was applied to solve the optimal solution of the likelihood function just with a single variable of positional parameter. Finally, several examples were given to demonstrate the stability and efficiency of this algorithm.

Abstract: It is difficult to identify distribution types and to estimate parameters of the distribution for small sample censored data when you deal with mechanical equipment reliability analysis. Here, an intelligent distribution identification model was established based on statistical learning theory and the algorithm of multi-element classifier of Support Vector Machine (SVM), and also applied to parameter estimation of small sample censored data, in order to improve the precision of traditional method. Firstly, the algorithm of training based on SVM and the RBF kernel function was selected; secondly, the parameters of the distributions characteristics were drawn; on the basis of these conditions, the distributions identification model and the parameter estimation model were finally constructed. And the model was verified with Monte Carlo simulation method. The results indicate that the new algorithm has more preferable performance in distribution type identification and parameter estimation than the traditional methods.