Research On Fault Diagnosis Expert System For Key Components Of Pure Electric Truck

The energy consumption and pollution of traditional truck is high,which does not meet the low carbon economic development trend of the 21 st century.The emergence of pure electric truck has solved the problems above well.However,the pure electric truck is available for a short time,some of the key components are more prone to fault,coupled with the lack of relevant maintenance experience and technical staff,often lead to the fact that the fault of trucks can not be timely solved.Therefore there is an urgent need for the development of a fault diagnosis expert system for key components of the pure electric truck.The thesis researches on the fault diagnosis expert system for the key components of pure electric truck relying on the technology support program of the Department of Science and Technology of Sichuan Province-"The research and demonstration on integrated key technology of the pure electric truck(five high-end)"(project number: 2016GZ0020).First of all,the thesis studied on the key components of the pure electric truck,focusing on the battery system and the permanent magnet synchronous motor,and studied on their structure and working principle.Then we collected and sorted out the common faults of these two components,and sorted out the hierarchical relationship of each fault event by establishing the fault tree,which reduced the redundancy when designing the knowledge base.Because of the obvious causal relationship between fault and fault symptom of the battery system and the permanent magnet synchronous motor,the knowledge base of the fault diagnosis expert system is established by the production representation method,and because some faults may show more than one fault symptom which occurs at a certain probability when the fault actually happens,with strong uncertainty,therefore the thesis contrasted and analyzed several commonly used fuzzy reasoning methods,by comparing the advantages and disadvantages of these methods and combing with the actual situation of the pure electric truck,decided to use Bayesian network to establish the reasoning machine of the expert system,use the causal relationship questionnaire to determine the causal relationship between fault and fault symptom,and then established a Bayesian network,and determined the occurrence of a fault in the case corresponding to the conditional probability of a fault symptom as the confidence of the reasoning result.Because there is a strong subjectivity when determining the conditional probability by experts,the error caused by this may affect the accuracy of the diagnosis results.Therefore,the thesis presents a method to support the decision by diagnosis results through mining the historical diagnosis records when designing the reasoning machine,so that the expert system can gradually improve the diagnostic accuracy with the increase in the number of usages.Finally,the thesis programmed to achieve a set of B/S structure of the fault diagnosis expert system using the Java language.The expert system only need to use the browser to login,which is of great convenience.The thesis makes full use of the fuzzy reasoning theory and puts forward a new idea for the fault diagnosis of the key components of the pure electric truck through the development of the 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,has solved the difficulty acquiring knowledge of the expert system.The expert system also has a strong versatility,which can be applied to other areas of causal relationship uncertainty reasoning.