论文部分内容阅读
基于维护时收集的数据,以生存分析方法构建了冷水主机的生存曲线,识别了故障事件间隔时间的概率模型,并估计了模型参数的极大似然估计量。利用统计假设检验法比较了2台冷水主机生存率的差异显著性,藉以推断这2台冷水主机维护质量的差异。利用随机过程更新理论确立了设备故障事件发生的计数过程及累积维修成本随机过程的统计特性,并建立了期望累积维修成本与使用时间的关系。结果显示,故障事件间隔时间的概率模型近似指数分布,故障事件发生次数随时间增加的计数过程可以泊松过程模拟。
Based on the data collected during maintenance, survival curves were constructed by survival analysis. The probability model of fault event interval time was identified and the maximum likelihood estimator of model parameters was estimated. The statistical significance hypothesis tests were used to compare the significant differences in the survival rates of two chiller hosts in order to infer the differences in the maintenance quality of the two chiller hosts. Using the theory of stochastic process updating, the counting process of equipment failure events and the statistical characteristics of stochastic process of cumulative maintenance costs are established, and the relationship between the expected cumulative maintenance costs and the operating time is established. The results show that the approximate exponential distribution of the probabilistic model of the fault event interval and the counting process of the number of fault events increasing with time can be simulated by Poisson process.