Critical Performance Evaluation of a Few Intelligent Controllers for Effective Power Control in Variable Speed Wind Generators

Abstract

Active and reactive power controls are the two major grid code issues for effective grid integration of wind farms. In this chapter, a novel inverse artificial neural network (ANN) controller has been proposed for effective control of grid side power in a variable speed wind generator. Its performance was tested on a full capacity grid-connected squirrel-cage wind generator. A two-level ac-dc-ac converter was used as buffer between the grid and the wind generator. The inverse ANN controller was designed using neural network system identification (NNSYSID) approach. The performance of inverse ANN controller was compared with adaptive neuro-fuzzy inference system (ANFIS) and a conventional proportional-integral (PI) controller. ANFIS and inverse ANN controllers can cognize the nonlinear dynamics of a plant for performing the control action. A vast amount of input and output data generated from a plant can be used for training while designing these nonlinear controllers. Hence, their design is relatively easier as compared to any other nonlinear controller. A PI controller is designed by linear approximation of a nonlinear plant near an operating point. However, if the operating point of the plant shifts beyond the range of design, the performance of PI controller deteriorates. The use of intelligent ANN and ANFIS controllers can overcome this problem. This study shows the better performance of proposed inverse ANN controller than the ANFIS and PI controllers.