Research On Abnormal Identification And Intelligent Diagnosis Of Wind Turbine Main Bearings

As a typical clean energy source,wind power has been vigorously developed in China in recent years.Due to the long-term operation of wind turbines in a complex and changeable natural environment,the harsh environment leads to frequent wind power failures.Frequent failures bring high operation and maintenance costs.As the core component of the wind turbine transmission system,the main bearing of the wind turbine can directly affect the safety and stability of the entire unit.The main bearings are extremely prone to failure due to their constant exposure to alternating shocks and loads.Once the main bearing fails,its maintenance cost is high,and the downtime for operation and maintenance is long,resulting in large economic losses.Monitoring the main bearing,discovering its abnormality in time,and diagnosing the cause of the abnormality can not only assist the operation and maintenance engineer in operation and maintenance,but also prolong the service life of the main bearing.The specific work based on this paper is as follows:1.Starting from the structure of the wind turbine.Firstly,the structure and power generation principle of wind turbine are briefly analyzed.Secondly,the structure of the main bearing and its logical relationship with other components are described in detail.Then,the common faults of bearings are listed and the causes of the faults are analyzed.Finally,the reasons for the abnormality of the main bearing are analyzed and summarized,which lays a foundation for the establishment of the abnormal identification model of the main bearing and the intelligent diagnosis model of the abnormal cause.2.For the problem that the accuracy of wind turbine main bearing anomaly identification is greatly affected by wind speed fluctuation,a BPNN-NCT based wind turbine main bearing anomaly identification method is proposed.Firstly,the principles of back propagation neural network(BPNN)and non-central t distribution(NCT)are briefly analyzed.Next,the parameters related to the main bearing state are determined by using the correlation coefficient analysis method,and a state parameter prediction model with temperature as the state indicator is constructed based on the BPNN.Then,based on the NCT distribution,the distribution characteristics of the residuals of the main bearing condition parameters under different wind speed fluctuation intervals are depicted,and the wind speed fluctuation rate under different intervals were related to the distribution characteristics of the main bearing condition parameters,and the quantified index of the main bearing abnormality of wind turbine taking into account the effect of wind speed fluctuation was proposed.Finally,the effectiveness and accuracy of the proposed method are verified by taking a direct-drive wind turbine of a wind farm as an example.3.For the diagnosis of the causes of main bearing abnormalities in wind turbines,an ontology-based intelligent diagnosis method of main bearing abnormalities in wind turbines is proposed.Firstly,the principle and construction method of ontology and fault tree are explained.Secondly,the wind turbine main bearing ontology knowledge base and abnormal fault tree model are constructed,and the probability of triggering main bearing abnormal events and their probability importance are calculated,and this is used as a class basis for reasoning.Then,in order to clearly represent and discover the implicit factual relationships among main bearing abnormalities,the fault tree and semantic web rule language(SWRL)library rules are fused,and the main bearing abnormality inference ontology model is built by using Protégé tool,and the diagnosis of main bearing abnormality causes is realized by Pellet inference engine.Finally,the effectiveness of the proposed method is verified by examples.