CONTRIBUTED ARTICLE Application of Neural Networks to Load-Frequency Control in Power Systems

Abstract--This paper describes an apphcatton of layered neural networks to nonhnear power systems control. A single generator unit feeds a power hne to various users whose power demand can vary over trine. As a consequence of load variatwns, the frequency of the generator changes over ttme A feedforward neural network is trained to control the steam admtsston valve of the turbme that drtves the generator, thereby restoring the frequency to its nominal value. Frequency transients are minimized and zero steady-state error is obtained The same techmque is then applied to control a system composed of two single umts tted together through a power hne Electrw load variations can happen independently m both units. Both neural controllers are trained wtth the back propagatwn-through-time algorithm Use of a neural network to model the dynamtc system ts avoided by introducing the Jacoblan matrices of the system in the back propagation chain used in controller training Keywords--Power system, Load-frequency ontrol, Feedforward neural network, Back propagation-through-time. 1