Research On Intelligent Fault Diagnosis System Of Wind Farm PMSM Based On Ontology And Knowledge Map

Wind power generation,as a typical representative of new energy,has prospered and developed recently.China’s installed wind power capacity ranks among the top in the world and is still growing continuously.However,wind power PMSM(Permanent Magnet Synchronous Machine)units,as a typical large complex system,have high failure incidence and difficult fault diagnosis.The fault diagnosis method based on ontology and knowledge map is proposed for the strong correlation of PMSM system,the complex heterogeneous information,and the real causes of the traditional fault diagnosis method.With the open fault of PMSM drive system,the main work is as follows:(1)Firstly,for IGBT open circuit faults of converters with obvious fault characteristics and the highest incidence,the improved normalized error current data analysis method is used to identify the fault switch quickly and accurately.This method can get rid of the strong dependence of current signal on load mutation and the algorithm is simple and feasible.However,this method cannot diagnose and deduce the source fault inducements leading to open IGBT,and the multi-dimensional heterogeneous fault causes leading to open IGBT cannot be integrated by a single algorithm.Therefore,based on data analysis and fault domain knowledge,ontology model construction,knowledge classification and reasoning are carried out to locate the fault sources.(2)Secondly,the PMSM fault diagnosis domain ontology model is formally defined and standardized to integrate the dense heterogeneous information in the PMSM system.Based on the improved normalized error current data analysis method to accurately and quickly locate the fault switch tube,the fault characteristics and recognition results obtained from quantitative analysis are mapped to the fault domain ontology model,which expands the fault diagnosis scope.SWRL rule base)and rule-based reasoning,and integrate ontology technology to trace back and diagnose information such as fault inducement and fault type.(3)The knowledge map based on ontology and fuzzy reasoning was imported into Neo4 j for storage,and a variety of PMSM abnormal fault diagnosis cases were analyzed and verified by experiments.This method improved the utilization rate of information among subsystems and could reasonably trace the fault causes.It provides a shareable and easily updated standardized method for fault diagnosis of large and complex systems.(4)Finally,based on the fault knowledge map,the PMSM intelligent fault diagnosis system was developed based on the Neo4 j graph database and C #programming.In addition to the design of the fault diagnosis function,the automatic storage,update and query function was added according to the internship requirements,and the effectiveness of the system was tested through the same fault experiment in the ontology model and Neo4j database.