The Fault Diagnosis Of Wind Turbine Gearbox Based On Improved SVM Of Parameter Optimization

The problem of resource shortage and environmental pollution is getting worse, and it is a direct threat to human survival and development. The development and utilization of new energy is imminent. Wind energy is a clean, non-polluting, renewable green energy, and it is under increasing scrutiny. With the large scale wind generators are put into operation, fault diagnosis problem is much accounted of people. In order to reduce the losses of unit fault, a perfect wind turbine fault diagnosis system is required. Gear box, an important part of the wind turbine, is part of the high incidence of failure. The damage caused by the gear box failure is very serious. This paper focuses on the identification and classification of gearbox fault.SVM theory is a type of machine learning theory for small sample and the common fault types of wind turbine gearbox are random and small sample study. In this paper, the support vector machine is used for building model of fault diagnosis. The accuracy of fault diagnosis classification is affected by the selection of SVM kernel function. Typically, single kernel function is global or localized, so SVM with a single kernel function has poor classification accuracy. From the perspective of the kernel function, composite kernel, a new kernel function is proposed in this paper and the new function has two single-core characteristics. Compared with a single kernel function, support vector machines with composite kernel have better classification accuracy. The parameters of SVM are optimized by SAPSO and the improved SVM model is used for simulation experiments in UCI.Finally, gearbox fault diagnosis signal is collected and extracted on the small fault diagnosis system in laboratory. Based on decision tree and voting rules, multiclass SVM classifier is established and used for the gear box failure data classification and identification. Experimental results show that the composite kernel SVM model based on SAPSO algorithm has good classification accuracy and improves gearbox fault diagnosis accuracy. The small wind generator fault diagnosis system with this method achieves good diagnosis results in the experiment. It provides important technical support for the normal operation of the wind turbine and is of great practical significance.