【作者】 郭伟伟；

【导师】 马捷中；

【作者基本信息】 西北工业大学 ， 计算机科学与技术， 2007， 硕士

【摘要】 随着新型装备在军事领域中的应用，为了保持武器的性能，投入在测试和故障诊断上的费用越来越高，基本贯穿了武器装备的整个生命周期。并且，随着第三代信息化武器装备的使用，传统的故障诊断已经不能满足现在的需要。远程故障诊断是目前的一种趋势。信号的采集、信号处理、数据的传输和故障诊断专家系统是远程故障诊断中需要解决的关键问题。故障诊断专家系统又是其中的重中之重。本文主要研究了基于故障树的远程故障诊断专家系统的构建和实现。 文章首先讲述了远程故障诊断专家系统的系统架构的构建和专家系统的知识，详细阐述了各个部分的组成和功能，分析了将故障树技术和专家系统融合的优点。然后，文章重点介绍了故障树分析法，对故障树分析法中的故障树的建立，故障树的预处理，故障树的定性分析和定量分析以及基于故障树定量分析的故障诊断法进行了详细的研究，并在利用二元决策图进行故障树定性分析时，提出了几条适用于决定底事件转换顺序的规则： (1)尽量对故障树进行化简并在化简后，考虑底事件出现的次数。出现次数少的优先考虑，若出现的次数一样，那么底事件所在项的项数最短的优先考虑。 (2)对于函数的子函数，则考虑所有子函数中每个变量出现的次数总合，如果一个函数中有一个函数中没有，那么应该优先考虑。 (3)在转换子函数的过程中，利用第三条决定了先分解的变量后，若其中某个子函数不含有即将分解的分量，则先不分解，等下一次。 然后，文章研究了故障树技术和专家系统的融合技术。对于专家系统知识的获取，本文采用了故障树技术来获取专家系统所需的知识。建立所研究对象的故障树的过程就是一个对系统的故障模式进行详细分析的过程，同时获得了专家系统所需的知识库。故障树的建立是近年来故障树技术的一个研究重点，本文采用了基于元件小故障树的计算机辅助建树法。对于专家系统知识的表示，本文将规则表示法和框架表示法相结合来表示知识库中的知识，同时将故障树的概念和知识的框架相融合，将最小割集和规则相融合，将最小割集中的底事件和规则中的条件融合起来。对于推理机制，这是专家系统的重点。故障诊断中，如果没有一个好的推理规则，对所有可能导致故障的事件一一进行测试，会导致效率低并且消耗也大。本文将故障树中基于最小割集重要度的思想运用到专家系统的推理机制中，这样可以大大提高一次命中的概率，减少盲目测试的次数。 最后，本文以某航空设备中某重要的电子设备为例，详细的讲述了基于故障树的专家系统的建立和应用过程，并对结果进行了分析，分析结果证明了将故障更多还原

【Abstract】 With the use of new equipment in the military field, in order to maintain the performance of them, the money invested in testing and fault diagnosis on them increased highly and throughout the entire life cycle. Moreover, with the use of weapons and equipment of the third generation of information, the traditional fault diagnosis has been unable to meet current demands. Remote Diagnosis is a trend. Signal acquisition, signal processing, Fault Diagnosis Expert System for data transmission and remote fault diagnosis are the key issues which need to be resolved. Fault Diagnosis Expert System is the top priority. This paper studies the design and implementation of fault diagnosis expert system. The expert system is related to knowledge acquisition, knowledge representation and reasoning mechanisms to achieve such issues.The article starts by describing Remote Diagnosis Expert System Architecture and Construction of the expert system knowledge, explained in detail the various parts of the composition and functions of a fault tree analysis and expert system integration advantages. Then, the article highlights the fault tree analysis, fault tree analysis to the establishment of the FTA, Fault Tree pretreatment Fault tree analysis of the quantitative and qualitative analysis and quantitative analysis based on the fault tree for a detailed study of fault diagnosis, and the use of binary decision diagram fault tree analysis, the number of incidents in the end decided to change the rules of order .In this process of transformation, the order of the incident should follow these rules:(1) After making the fault tree simple, taking into account the number of cases of late. The more frequency has less priority.(2) Functions of the function, taking into account all the variables for each functions aggregate number. If a function is a function not available, then priority should be given.(3) Functions in the conversion process, the decision to use the first third of the decomposition of variables, If one will not contain any Functions decomposition of the moment, I will not biodegradable, so the next.Then, we study the fault tree expert system technology and the integration of them. For knowledge acquisition of expert system, we use a fault tree expert system technology to acquire the necessary knowledge. Establish the study of the fault tree is a process of detailed analysis of the system failure modes in the process and gain the更多还原