Fault Diagnosis Of Wind Turbine Based On Minimum Entropy Deconvolution

Wind power generation is becoming more and more important in the electric power industry. However, due to the poor conditions of the environment, the failure rate of the wind turbine and the maintenance cost are relatively high. So it is of great significance to the condition monitoring and fault diagnosis of wind turbine generator. One of the major challenges in the fault diagnosis of wind turbine generator is how to extract the fault features of rolling bearing and gear in the strong background noise.The minimum entropy deconvolution method is a convenient and effective fault feature extraction and intelligent diagnosis method. It is more advantageous to identify the weak fault of equipment and improve the accuracy and reliability of fault prediction by the minimum entropy deconvolution method.In this paper, we mainly study the method of bearing fault feature enhancement based on minimum entropy deconvolution and its application in field measurement. The main research work:(1) To explore the theoretical basis of minimum entropy deconvolution method and the validity of the method is verified by the simulation signal and the actual signal of the wind turbine.(2) It is verified that the minimum entropy deconvolution method can still be stable and reliable when the speed of the object signal is changed. And by changing the parameters of the filter, the influence of the different factors on the minimum entropy deconvolution results is presented, and the reasonable parameters are given.(3) The time domain characteristic value and frequency domain characteristics of the signal after the minimum entropy deconvolution is studied. Introduces the self-organizing feature map (SOM) neural network theory, and train vibration signal input neural network with minimum entropy deconvolution method in processed, which can be used to recognize the normal and fault signal and applied to the fault diagnosis of wind turbine effectively.