Fault Detection in the Wind Farm Turbine Using Machine Learning Based on SVM Algorithm

Abstract

In this paper, for failure identification and insulation in a horizontal axis variable speed turbine made up of three sheets and one total converter, the vector support machines (SVM) is used. Data is based on the SVM method and so know-how is robust. It is also focused on the reduction of systemic risk this increases generalization and encourages method non-linearity accounting for the use of modular kernels. A radial function as a kernel has been used in this work. Various parts of the process, including actuators, sensors and process failures, have been investigated.