Research On The Management And Application Of Multi-source Heterogeneous Knowledge For Wind Turbines Based On Ontology

Equipme nt Maintenance and fault diagnosis play critical roles in keeping windturbines working normally to reduce operational costs. The improvement of the skillsof maintenance and fault diagnosis for wind turbines will not only help enterprises toincrease their revenues, but also promote the harmonious development of environmentand society. With the development of modern science and technology, the techno logiesand ideas about the equipment maintenance and fault diagnosis are constantly updating,and the utilization of existing equipment knowledge has become an important way torealize the equipment maintenance optimization and inte lligent fault diagnosis.Knowledge-based equipment ma intenance can reduce the complexity of equip mentmaintenance and maximize the maintenance effectiveness by ma intenance knowledgeintegration and maintenance process optimization. Knowledge-based fault diagnosiscan inte lligently generate diagnosis strategies for maintenance personne l from theintegration of diagnosis knowledge and the utilization of knowledge reasoning.Knowledge management is the foundation of knowledge-based equip mentmaintenance and fault diagnos is, while the wind turbine knowledge is difficult to fuseefficiently with traditional methods, for this scattered knowledge existing in differententerprises and departments. In order to improve the knowledge manage ment aboutwind turbines, ontologies were introduced into the knowledge representation andretrieva l to provide supports for equipment ma intenance optimization and intelligentfault diagnosis. Sponsored by the State High-Tech Development Plan of China (No.2009AA04Z414), the present dissertation did a profound and systematic research onthe management and application of multi-source heterogeneous knowledge for windturbines.The main research work and innovative achievements of this dissertation aresummarized as follows(1) Considering the demands of the knowledge management for wind turbine s, amulti-source heterogeneous knowledge manage ment model based on ontology for windturbines was introduced in the present dissertation. With this model, the knowledge invarious domains was described in the form of ontology, and then the global knowledgefus ion was achieved by virtue of the mappings between global and local ontologies.Maintenance optimization and inte lligent fault diagnosis for wind turbines were realized on the basis of knowledge fus ion, which can be used to achieve the effic ientutilization of equipment knowledge.(2) In the traditional FMECA (Failure Mode, Effects and Criticality Analysis)method, the attributes of the risk are assigned vague ly, and the influence of the weightof risk was ignored. To overcome these shortcomings, an ontology was employed toexpress the fuzzy knowledge and a method of fuzzy multi-criteria decis ion wasproposed to realize the quantitative evaluation. For the issue that the risk attributes ofdifferent faults have the same RPN (Risk Priority Number) value, a method based onthe data envelopment analysis was used to improve the model of assessment. Thecritical ability of FMECA can be enhanced and the optimization of equip mentmaintenance can be achieved by the proposed method.(3) To meet the demands of the maintenance plan optimization for wind turbines,a FTF (Fault Tree Failure) method, which is the combination of FMECA ontology andFTA (Fault Tree Analysis) ontology, was proposed in this dissertation. Based on theproposed method, all the knowledge of relevant doma ins was expressed by ontologies,the ma intenance plan optimization was imple mented according to the RPN values ofthe minimal cut sets. This method can solve the problem that the FMECA can tinvestigate multiple fa ilure modes, and hence the maintenance effic iency for windturbines is improved.(4) Aimed at the proble m that the fault diagnosis methods for wind turbines aretoo complicated to grasp by the maintenance personnel, an intelligent fault diagnosismethod for wind turbines was developed based on knowledge retrieva l. This methoddescribed the doma in knowledge by ontology, and built the rule sets for fault diagnosisreasoning. Thus, it can assist the maintenance personnel to select the appropriate faultdiagnosis methods according to knowledge reasoning.(5) An inte lligent fault diagnosis method based on FMECA ontology wasproposed. This method taken the FMECA ontology as the diagnosis knowledge base,the doma in rules about FMECA were also established. Reasoning on this knowledgebase by virtue of the JESS (Java Expert System Shell) rule engine, the maintenancepersonnel can find the causes of faults of a wind turbine quickly, and choose the propersolutions. The problem solving capacity of fault diagnosis is improved by thereasoning ability of the ontology model. And the reasoning results can providesupports for diagnostic decision making.(6) A knowledge manage ment prototype system for wind turbines was developed.The R&D demands and overall framework of the prototype system were illustrated. The development processes (inc lud ing the design of ontology knowledge base andknowledge reasoning, the module of maintenance optimization and fault diagnosis)were also described. The example validation showed that the prototype system iseffective.