大修汽车可靠性分析中故障分布规律的研究

从分布函数拟合的基本原理入手，把回归分析方法和优化技术有机结合起来，提出了一种由观测数据拟合其分布函数的优化方法，并以威布尔分布为例，说明其具体过程，通过试验数据应用表明，该方法正确可行。

FAILURE DISTRIBUTION IN RELIABILITY ANALYSIS OF MAJOR OVERHAULED AUTOMOBILES

In the reliability analysis of major overhauled automobiles, it is very important to fit the statistical distribution function to the test data that can reflect the characteristics of their life. But in practical analysis, it may happen to that a group of data can pass through several different hypotheses of distribution functions. In this case, it can not differentiate the merits and demerits of each distribution qualitatively by using the general statistical test. Therefore, authors took the ordering probability calculated from the sample data as the test data and the accumulative failure probability calculated from the fitted distribution function as the theoretical value. Then, an accumulative value of the one order absolute errors between above two values was taken as a criterion of the curve fitness. As an example, the Weibull distribution was taken to judge the test data by above procedure. The result shows that it is reasonable and straightforward for determining the fitness quality.