Software Design And Implementation Of Fuzzy Reliability Analysis

This dissertation originates from a project of National Natural Science Foundation of China. The aim of the dissertation is to develop analytical software for fuzzy reliability based on the newest methods of fuzzy reliability analysis. The software can be used not only in fuzzy reliability analysis, but also in random reliability analysis. The research contents are of vital theoretical significance and practical application.As the development of science and technology, the research and application of reliability obtains more and more attention. In fuzzy reliability design of machine products, the failure probability are often needed to be computed. But in fuzzy reliability analysis, the software of fuzzy reliability analysis is short because of the complexity of fuzzy reliability analysis. Therefore, when the research and application of fuzzy reliability are studied, computing procedures must be programmed. In this way, it not only wastes a lot of time and manpower, but also increases the possibility of error.In order to change this situation, based on the fuzzy reliability model and algorithm of the transformation from fuzzy variables to random variables, the fuzzy reliability analysis software is developed by using .NET platform and the core language C# development tools in Microsoft .NET strategy. The software can be used to analyze not only fuzzy reliability but also random reliability. Firstly, in this paper some basic mathematic knowledge which is needed for the fuzzy reliability theory is described, including fuzzy reliability and random reliability. Secondly, the development of the system requirements analysis, architecture, function and specific design model is discussed. In the process of developing modules, including random reliability module and fuzzy reliability module, the implementation process of details is discussed. Some important algorithms which are useful in the process of the procedure realization are analyzed mainly, such as the "dichotomy", the transformation of fuzzy variables to equivalent random variables, fuzzy design checking point method, equivalent normal distribution analysis method, and so on. Some key problems which should be solved in the process of software development are discussed, including "formula identification", "Numerical I.", "equation". Finally, some examples are given and the computed results are analyzed, which show that it is of universality and practicality for the developed software.