An adaptive radial basis function neural network (RBFNN) control of energy storage system for output tracking of a permanent magnet wind generator

Abstract

The converters of a permanent magnet synchronous generator have to be properly controlled to achieve maximum transfer of energy from wind. To achieve this goal, this article employs an energy storage device consisting of an energy capacitor interfaced through a voltage source converter which is operated through a smart adaptive radial basis function neural network (RBFNN) controller. The proposed adaptive strategy employs online neural network training as opposed to conventional procedure requiring offline training of a large data-set. The RBFNN controller was tested for various contingencies in the wind generator system. The adaptive online controller is observed to provide excellent damping profile following low grid voltage conditions as well as for other large disturbances. The controlled converter DC capacitor voltage helps maintain a smooth flow of real and reactive power in the system.