Co-ordinated static and dynamic monitoring and optimisation of power systems using a parallel architecture and pattern recognition techniques

The pattern recognition approach developed by the author has been implemented in a parallel architecture for the monitoring and optimisation of power system static and dynamic security. Initial results from this approach have indicated that it is feasible for large-scale applications where systems have to be operated and optimised in a multiobjective environment within various security limits. This architecture has many of the basic characteristics required for parallel computations, namely low memory requirements, high evaluation accuracies and computational savings, and large flexibility of formulation/expansion. The development and results of four suboptimisation algorithms using linear programming to enhance system security is also described. Designed for the proposed parallel environment, these algorithms perform recursive and separate suboptimisations on critical clearance time, line-flows, generator excitation and shunt reactive compensation