The main goal of the work is to choose the theoretical distribution function most consistent with the empirical function of fault distribution based on the analysis of statistical information of previous replenishment periods about the failures of details of each type of par value. This information should be accumulated on daily information about the replacement of spare parts of failed parts in vehicles that arrived during the entire period of replenishment for maintenance at this service station. The choice of the best theoretical distribution function in this sense is made from a set of a finite number of competing parametric distributions (exponential, normal, log-normal, We bull, monotonic and no monotonic diffusion) by Kolmogorov-Smirnov's test. The advantage of this criterion in comparison with other consent criteria is that, along with an estimate of the accuracy of the approximation of the empirical failure distribution function. The mutual reversibility of the processes of distribution of the operating time to failure (to failure) and the number of failures is established, the relationship between the expressions for the distribution function of the operating time to a fixed number of failures and the function of the distribution of the number of failures for a fixed operating time to failure is obtained. This ratio allows you to choose the best distribution model based on the available fault statistics of parts (and replacing them with the corresponding spare parts) in the previous planning periods.
展开▼
