Study of Sensorless Controller for Direct Drive Wind Electric Conversion Systems

Abstract

The direct drive machine used in wind turbine systems (WECS) is efficient due to absence of losses rotor losses and lower no load current. The goal of this study is to propose and develop a neuron-based controller to optimize wind turbine output power by considering two objectives; maximize power extraction by wind turbine and predict and analyze pitch angle to gain maximum power. The proposed controller is designed to take the following inputs: wind velocity, pitch angle and wind turbine electrical output. The controller then finds the optimal pitch to predict the maximum power that can be extracted. The controller has a neural network model that is a feed forward, three-layered perceptron. This model explains the integration of controller providing wind speed estimation and robust control of maximum power extraction with no rapid drift on the power coefficient curve.