Neural Modelling and Control of a Synchronous Generator

Abstract

Neural network techniques have been applied to excitation control of a synchronous generator. A radial basis function network was constructed using a hybrid training algorithm, combining linear optimization of the output layer weights with singular value decomposition, and non-linear optimization of the basis function centres and widths using second-order gradient descent BFGS. The network Jacobian was subsequently calculated to provide instantaneous linear models of the plant. Using these models, internal model control, Kalman and generalized minimum variance schemes were implemented on an industry-standard VME platform. The performance of the neural models and neural control schemes was investigated on a laboratory micromachine, with comparisons being made with a self-tuning regulator, employing a generalized minimum variance strategy.