Design And Implementation Of An Intelligent Diagnosis System For Operating Events Of Hydroelectric Generating Units

The Intelligent diagnostic system in a field of artificial intelligence research has been widely demanded in engineering equipment applications,but there are few cases satisfying user's demands,which mainly because the knowledge base construction is complex and difficult,the quality is not guaranteed and reasoning mechanism is not perfect.To solve these problems,this paper develops a set of Intelligent diagnostic system,and builds corresponding knowledge base for the operation of hydropower units in hydropower plant of Wuxi River based on the project to achieve automatic intelligent diagnosis.A set of B/S(Browser/Server)mode Intelligent diagnostic system is developed by JAVAEE technology to analyze the demand of Intelligent diagnostic system software according to project actual needs,and the SSM(Spring-SpringMVC-MyBatis frame technology is used to design the architecture,which involving data structure design,business logic design,design of control jump interface and so on.The real-time data is butt-joint with the power-plant monitoring system to adjust the corresponding reasoning mechanism on the basis of actual situation of the hydropower plant,which has mainly achieved following functions:(1)For current problems of knowledge acquisition and technology design,it develops the tool software in WEB form to construct knowledge base,achieving multi-user/multi-station knowledge editing and input,which significantly improves the efficiency of knowledge acquisition to facilitate collaborative development of knowledge base.(2)Knowledge model is input in webpage form that adopts Ali Druid connection technology to store in MySQL database as data sheet.Knowledge base is stored in database form to achieve knowledge base's portability and reusability.(3)The fault tree kernel expression of knowledge base is established through data-field information,which provides reasoning thought for fault diagnosis.As fault tree analysis has been proved to be a good method for engineering application,it not only can significantly improve the quality of knowledge acquisition,but also facilitates the engineers and team of experts to review and improve the knowledge base.(4)By research on diversity expression of fault naming,it has three ways of naming(cause,component,observable parameter),and the construction of fault tree can explain knowledge node information more vivid,adapt to the need of various fault diagnosis(fault cause,process,fault impact range,fault propagation direction),construct corresponding knowledge base for the actual situation of hydropower units,and build software through knowledge base to store in the data base.(5)Research and design of the data dictionary,as an important bridge of interaction between knowledge base and reasoning machine,provide domain experts team about the reference of knowledge node,for the construction of knowledge base can play a good supporting role.(6)Inference machine gives consideration to two databases of MySQL and SQLSERVER,the MySQL is used for knowledge base abutment,and SQLSERVER is used for factual evidence base abutment,which selects two reasoning modes of upwards matching and downwards searching(it adopts the reasoning mode of primarily downward searching with auxiliary upwards matching),which achieves the reasoning function of Intelligent diagnostic system.(7)The report page of fault diagnosis is designed by HTML and CSS technology that uses JAVASCRIPT scripting language to control the page timed refresh(the refresh mechanism is set to 90s based on the factual evidence base).The failure report page includes alarm test name,time,input knowledge base,related possibility data deduced by inference machine and relevant evidences provided by factual evidence base.(8)When generating the failure report,record current faults and list it in the failure log,including the failure frequency,cause of fault and trouble-shooting information to provide basis and convenience for the establishment of big data fault diagnosis.Through designing and achieving above functions,the team of experts can independently related knowledge nodes,build new knowledge base,review and complete previous knowledge base.Meanwhile,once the factual evidence base reads the abnormality or alarm,the data will automatically start the inference machine,transfer to corresponding knowledge base for fault diagnosis,input the diagnosis result in the report(failure cause,fault influence range),and record the occurred fault information to corresponding fault log.This Intelligent diagnostic system has strong generality,which provides strong support for the experts and engineers in this field to construct knowledge base,and reduces the time consuming of manual troubleshooting that conforming to the technology development of times.Under the premise that knowledge base's quality is guaranteed,it reduces man-made diagnostic mistakes and makes the contribution for artificial intelligence development.