Development Of Wind Turbine Remote Monitoring And Health Prediction System

With the gradual scarcity of non-renewable energy and the increasing environmental pollution,renewable and clean energy such as wind energy has been favored and strongly supported by governments of various countries and regions around the world.Wind turbines have many complex structures,and they work continuously on land and sea in harsh natural conditions for a long time,so they are prone to failure.If problems can't be found in time,it is easy to make minor faults develop into major accidents,resulting in irreparable losses.Therefore,how to detect the potential faults of wind turbines in time by means of intelligent detection and timely warning through real-time remote monitoring system is an important topic being studied in the wind power industry.In this paper,the common faults of wind turbines are analyzed,two fault diagnosis methods are proposed from different directions,and the remote monitoring and health prediction system of wind turbines based on B/S mode is designed and implemented.The main work of this paper includes the following aspects:(1)According to the structural characteristics of wind turbines,the failure causes of important components are analyzed.(2)According to the regularity of vibration signals in key parts of wind turbines under different health conditions,a fault diagnosis method based on vibration signal processing is proposed.Vibration signals of key parts are decomposed into a series of IMF components by CEEMDAN method.Useful IMF components selected by hausdorff distance and cross-correlation coefficient are used to form a matrix.The matrix is decomposed by singular value,several larger singular values are selected and normalized as a set of state eigenvalues.The eigenvalues of various states are trained and predicted by using the extreme learning machine algorithm.The accuracy and rapidity of the proposed method are verified by the comparison of support vector machine(SVM)algorithms.(3)Considering that the failure rate of components will also play an important role in equipment fault diagnosis,a method of equipment fault diagnosis based on the failure rate of components and fault symptoms is proposed.Based on Weibull Distribution Model,an algorithm for calculating the failure rate of components is established,and a fault diagnosis method based on failure is constructed by using the method of fuzzy comprehensiveevaluation.The failure rate of components is introduced into the fault diagnosis model,and the applicability of the proposed method is verified by an example.(4)Applying B/S architecture and Apache+MySQL+PHP combination,the remote monitoring system of wind turbines is developed.The functions of user login,management,operation status monitoring of wind turbines,report drawing and so on are realized.The fault diagnosis and prediction function based on vibration signal proposed in this paper is realized with Python language.