基于ESCN的8K型电力机车故障诊断系统的应用研究

【作者】 陈平；

【导师】 彭军；

【作者基本信息】 中南大学 ， 系统工程， 2007， 硕士

【摘要】 铁路运输安全关系着国家的经济命脉，而机车故障诊断系统是保障列车安全运行的重要组成部分。论文以8K电力机车为应用背景，为了实现机车电气故障诊断、机车运行状态监控等功能，针对8K型电力机车电气逻辑关系的复杂性、故障源的多样性、机车的重联特性，设计实现了基于专家系统与通信网络(ESCN，Expert System and Communication Network)的在线故障诊断系统。为了清楚的表达机车电气故障的传播途径，采用了基于布尔型Petri网的知识库设计方法。论文利用微机控制系统提供的电气节点状态，对机车电气线路进行了知识建模，解决了专家系统知识获取的“瓶颈”问题。基于产生式系统的推理机制确保能够快速、实时的查找到机车电气故障的故障源。产生式系统总数据库存储着机车电气节点的实时状态，论文采用节点遍历和反向推理的推理策略，利用产生式规则快速定位故障源，提高了故障查找的效率。另外，论文采用的基于预制文本法的解释器设计将故障信息和故障解决方法以用户易于理解的形式提供给用户；总线型的网络拓扑结构、令牌传递总线方式的运用，确保了系统数据采集网络通信顺畅。最后，依据系统总体结构，通过对数据采集系统各单元的硬件设计和对监控诊断软件的模块化设计逐步实现整个诊断系统，实现了机车电气故障诊断功能与机车运行状态监控等功能。论文设计的基于ESCN的8K型电力机车故障诊断系统已成功应用到北京铁路局8K型电力机车上，实际应用证明，该系统在保障列车安全运行方面是完全可行的、有效的。更多还原

【Abstract】 The safety of railway transportation is relative to national economic vitals. Fault diagnosis system is an important part to ensure the safety running of the locomotive. Aiming at the complexity of 8K Locomotive logical relation , the diversity of fault source and the trait of multi-locomotives, the dissertation designs an on-line fault diagnosis system based on ESCN (Expert System and Communication Network) to realize the electric fault diagnosis and working state detection for 8K Locomotive.In the dissertation, the repository design based on Boolean logical Petri Nets is adopted to distinctly clarify the transmission approach of electric fault for locomotive. The knowledge modeling is achieved by using the states of electric nodes provided by Microcomputer Control System, which can solve the "bottleneck" of knowledge acquisition.A reasoning mechanism based on production system is proposed to ensure the fast and real-time looking-up of electric fault source for 8K Locomotive. In the mechanism, all the states of the electric nodes are stored in the database of production system. By employing the node traversal and backward reasoning strategies in the production system, the fast localization of the fault sources can be achieved. Consequently the efficiency of fault looking-up can be well improved.Furthermore, the dissertation presents a design method of interpreter based on prefabricated text method, which can provide the fault information and solving solution to users with a comprehensible form. The bus network topology structure and token accessing control are used to solve the capability problem of the network communication in data collection.Finally, according to the general system structure, the dissertation realizes the function of electric fault diagnosis as well as working state detection through the hardware design of data collection system and the modular design of diagnostic system.The proposed fault diagnosis system based on ESCN has already been applied in the 8K Locomotive of Beijing Railway Administration. The experiment results show that the system is feasible and effective.更多还原