Neural network approach to identification and control of sub-synchronous resonance in series compensated systems

The development of a precise mathematical model for SSR is a complicated process, especially, in power systems involving multiple generators and series compensated transmission lines. Further, the analysis of the SSR phenomenon for the entire operating range is computationally intensive. In view of these situations, it is envisaged that neural networks can be effectively employed to design SSR countermeasures. This paper presents such a contribution. Examples on systems with thyristor controlled shunt reactor are used to illustrate the benefit gained from this approach.