Abstract：

Based on practically collected data, constructs the survival curves of the water chillers by survival analysis method, identifies the probability model of interval time between two sequential breakdown events for the water chillers and estimates the maximum likelihood estimators (MLE) of the parameters of the model. By means of statistic hypothesis testing, compares discrepancy significance of survival rates in order to judge the discrepancy of maintenance quality between two chillers. Based on the theory of random renewal processes, detects the statistical characteristics of the counting process of breakdown events and the cumulative repair cost process. Constructs the relation between the expected cumulative cost and time. The results show that the interval time follows approximately exponential distribution, and the counting process of breakdown events can be simulated as a Poisson process.