Study On The Interturn Short Circuit Fault Of The Rotor Windings Of Doubly-fed Asynchronous Wind Turbine

The wind turbine plant is generally established in the remote area of no man,and the unit is installed on the higher tower.When the inter-turn short circuit fault occurs,it is not easy to find,which causes huge losses to the power plant.In addition,the wind turbine environment is relatively bad,difficult maintenance,so it is very important to accurately judge the degree of inter-turn short circuit fault,so that the maintenance staff better grasp the maintenance time,improve the economic model of wind turbine.In this paper,the following research is carried out on the turn-to-turn short circuit of doubly-fed wind turbine:(1)The diagnosis results based on D-S evidence theory are compared with the diagnosis results of single information,and the method proposed in this paper is more suitable for fault identification of DFIG rotor windings.Ansys software is used to build the fault model of doubly-fed wind turbine.When the rotor has inter-turn short circuit fault,it is not easy to be detected in the early stage.Then the theory is used to accurately judge whether the rotor winding has inter-turn short circuit fault.According to the change of relevant parameters before and after the failure of doubly-fed wind turbine,the appropriate short-circuit fault characteristic quantity is analyzed and selected as the evidence body.Multiple sets of fault characteristic evidence body in generator are fused according to evidence theory,and the conclusion of fault determination with high confidence is obtained.The effectiveness of multi-source information fusion in turn fault identification of generator rotor windings is verified by comparing with single feature.(2)Combining the improved entropy weight theory with the grey correlation theory,a fault identification framework is constructed to accurately identify the turn-to-turn short circuit trouble level of doubly-fed wind turbine.The improved entropy weight theory and grey correlation theory are combined to get a better diagnosis method.Based on the fundamental frequency amplitude of A phase current in generator stator winding,A phase current(1-2s)f harmonic amplitude of stator winding,U phase voltage amplitude of rotor winding,DC component of electromagnetic torque,2sf harmonic amplitude of electromagnetic torque,fault domain of 0,4%,8%,12%,16%,trouble identification framework is built to diagnose the short-circuit level of generator rotor winding,and it is verified by Ansys Maxwell software simulation.(3)An improved Elman neural network is used to diagnose the turn-to-turn short circuit trouble phase of DFIG.Elman neural network is designed,and then the Elman neural network is improved to improve the approximation ability and dynamic characteristics of the network.Through Ansys Maxwell simulation,the experimental data are obtained,and the characteristic values of the rotor winding are extracted from the experimental data when the rotor winding is normal,the U phase is short,the V phase is short,and the W phase is short.The extracted data are trained in the network,and then the neural network is trained by two methods,respectively,to compare the diagnostic effect of the Elman neural network and the improved Elman neural network,and to verify that the improved Elman neural network has better approximation ability and dynamic characteristics.