Artificial intelligence based fault diagnosis mechanism in wind turbines

The increasing energy needs of the global energy sector, fossil resource reserves and the need to reshape within the framework of green environmental factors have progressed towards renewable energy sources with the development of technology. In terms of the continuity and accessibility of wind energy, wind energy has had the largest share among these resources. The size of the energy requirement also increases the turbine dimensions. The increase in turbine sizes prevents the easy accessibility of turbines. Increasing turbine sizes and the difficulty of interfering with the system have made it necessary to have a comprehensive control structure in the systems. Growing turbine sizes, increasing electrical power, safety and efficiency factors require fault prediction, detection or diagnostic systems as well as control in the system. In this study, an artificial intelligence-based fault diagnosis study was performed on a 4.8MW wind turbine with three wings, horizontal axis, pitching angle control. Linear classification methods cannot provide effective control because wind turbine systems have healthy operating conditions that vary under variable conditions. Various data received through the system were trained using the Neural Network Toolbox. It is aimed to make decisions about the system status by being processed by the decision structure. The input values measured or obtained by various calculations were used in the fault prediction with the structure of the neural network (YSA). The 'A Wind Turbine Benchmark Model for a Fault Detection and Isolation Competition, Silvio Simani' wind turbine competition benchmark model used in the international IFAC competition was used to implement and develop the fault diagnosis system. It is aimed to test the performance of the mechanism by performing eight different failure scenarios in MATLAB/SIMULINK environment.