纯电动载货汽车关键部件的故障诊断专家系统研究

发布时间：2018-07-01 08:26

本文选题：纯电动载货汽车 + 故障诊断　；参考：《电子科技大学》2017年硕士论文

【摘要】：传统载货汽车能耗高、污染大,不符合21世纪低碳绿色的经济发展趋势。纯电动载货汽车的出现很好地解决了上述问题。然而纯电动载货汽车面世时间短,一些关键部件比较容易发生故障,加上缺乏相关维修经验和技术人员,经常会导致车辆出现故障无法及时维修。因此为纯电动载货汽车的关键部件开发故障诊断专家系统具有迫切的需求。本文依托四川省科技厅科技支撑计划项目——“纯电动载货汽车集成关键技术研究及示范(五大高端)”(项目编号:2016GZ0020)的支持,开展针对纯电动载货汽车关键部件的故障诊断专家系统研究。本文首先重点分析了纯电动载货汽车两大关键部件——电池系统和永磁同步电机的结构和工作原理。接着搜集并整理出了这两个部件的常见故障,通过建立故障树梳理了各个故障事件的层次关系,减少了设计知识库时的冗余。由于电池系统和永磁同步电机的故障与故障征兆具有明显的因果关系,因此本文采用产生式规则表示法建立了知识库;通过对比分析几种模糊推理方法的优缺点并结合纯电动载货汽车的实际情况,应用贝叶斯网络构建推理机;通过因果关系调查表判定故障和故障征兆之间的因果关系,进而构造贝叶斯网络,并确定在某些故障发生与否的情况下对应于某一故障征兆的条件概率。由于专家在确定条件概率时具有很强的主观性,由此带来的误差可能影响诊断结果的准确度,因此本文在设计推理机时提出了一种通过对历史诊断记录进行数据挖掘实现诊断结果辅助决策的方法,使得专家系统能够随着使用次数的增多逐渐提高诊断的准确度。最后本文使用Java语言设计实现了一套B/S结构的故障诊断专家系统,该专家系统仅需使用浏览器登录便可使用,具有极大的便利性。本文充分利用贝叶斯推理理论,通过开发专家系统为纯电动载货汽车关键部件的故障诊断提出了新思路,具有较高的经济价值。此外,应用数据挖掘技术优化故障诊断专家系统,一定程度上解决了专家系统知识获取困难的问题。本专家系统还具有较强的通用性,可以适用于其他领域的因果关系不确定推理。
[Abstract]:Traditional truck has high energy consumption and high pollution, which does not accord with the economic development trend of low carbon green in the 21 ~ (st) century. The emergence of pure electric truck solves the above problem well. However, due to the short arrival time of pure electric truck, some key parts are easy to break down, and the lack of relevant maintenance experience and technical personnel will often lead to vehicle failure and failure in time. Therefore, it is urgent to develop fault diagnosis expert system for the key parts of pure electric truck. This paper relies on the support of the Science and Technology support Program of Sichuan Provincial Science and Technology Department-"Research and demonstration of key Technologies in the Integration of Pure Electric truck (five High end)" (Project No.: 2016GZ0020). The research of fault diagnosis expert system for the key parts of pure electric truck is carried out. In this paper, the structure and working principle of battery system and permanent magnet synchronous motor (PMSM) are analyzed. Then the common faults of these two components are collected and sorted out. The hierarchical relationship of each fault event is combed by establishing the fault tree, and the redundancy in the design of knowledge base is reduced. Because of the obvious causality between the fault and the fault symptom of the battery system and the permanent magnet synchronous motor, the knowledge base is established by using the production rule representation method in this paper. By comparing and analyzing the advantages and disadvantages of several fuzzy reasoning methods and combining with the actual situation of pure electric truck, the inference engine is constructed by using Bayesian network, and the causality relationship between fault and fault symptom is determined by causality questionnaire. Then the Bayesian network is constructed and the conditional probability corresponding to a fault symptom is determined when some faults occur or not. Because experts are highly subjective in determining conditional probability, the resulting errors may affect the accuracy of diagnostic results. So this paper puts forward a method to realize the decision aid of diagnosis result by data mining of the history diagnosis record, which makes the expert system improve the diagnostic accuracy with the increase of the number of times of use. Finally, a fault diagnosis expert system with B / S structure is designed and implemented in Java language. The expert system can be used only by using browser login, which has great convenience. This paper makes full use of Bayesian reasoning theory and puts forward a new idea for fault diagnosis of key parts of pure electric truck by developing expert system, which is of high economic value. In addition, the application of data mining technology to optimize the fault diagnosis expert system, to a certain extent, solve the problem of expert system knowledge acquisition difficulty. The expert system has strong generality and can be applied to causal uncertain reasoning in other fields.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U472.9;U469.72;TP182

【参考文献】