Research And Implementation Of Wind Power Generator Condition Monitoring And Fault Diagnosis System

Wind power base on the wind to push the wind turbine, as only the place where have wind can build wind turbines, and the wind direction and the wind speed are changeable,so the wind turbines running in the long-term variable load conditions. the area where the wind strong, the running environment is bad.Under the disadvantage conditions, the wind turbine failure occurrence probability becomes large. Moreover, with the development of wind power in recent years, many wind power companies are increasingly concerned about the safe running of the wind turbine.So the condition monitoring and fault diagnosis system must have high performance. In addition, due to the inherent characteristics of the wind industry. The wind turbine is widespread in the wind farm, and it is unattended in the normal circumstances, so when the wind turbine break down we can not be repaired it in time. Therefore, research has practical significance for the wind turbine fault diagnosis method, which can not only improve the reliability and availability of the wind turbine, and the maintenance costs can also get the maximum decline, so this is has high engineering practical significance.In general the fault diagnosis technology is broadly divided into three categories: 1.a diagnostic method in the traditional sense.2. Mathematical diagnostic methods.3. Intelligent diagnostic methods. In this paper, a method of intelligent diagnostic has been adopted, Because the intelligent diagnosis method does not require precise mathematical model, so this gearbox vibration analysis that identify the signal is done by a computer program. The program can determine the fault type and fault location. The intelligent diagnosis method manly include expert system, fuzzy mathematics and neural network, intelligent algorithm, support vector machine (SVM). Each kind of intelligent algorithm has its own advantages and disadvantages. Therefore, this chapter proposes intelligent gearbox fault diagnosis method based on the use of the improved fruit fly algorithm to optimize BP neural network, and use the actual measurement data to test the fault diagnosis simulation and analyses.This article describes a comprehensive analysis of the wind industry development status and trends at home and abroad, and have a simple explanation introduction on the structure and working principle of wind turbines. Focus on presentation and analysis of wind turbine failure mechanisms occurring against, and detailed analysis the fault model of the gearbox. According to the inherent characteristics of the collected wind turbine vibration sensor gearbox vibration signals using the least squares method to eliminate vibration trend sample data items; five three-point method to eliminate vibration peak sample data makes data more smoothly; wavelet decomposition eliminate vibration frequency noise sample data. Finally, using the wavelet packet method to extract the vibration signal band energy, and as a feature vector of the signal.Through the analysis of feature vectors, and use the BP, the FOA-BP and the improved FOA-BP intelligent algorithm for fault conditions of wind turbine gearboxes to identify and build three models of the gearbox fault, use actual data to achieve a wind turbine gearbox fault identification and classification, analysis and comparison of the three models to conclude.