Adaptive static load shedding schemes for Nigerian grid system based on computational intelligence techniques

Abstract

This paper has proposed fast convergence particle swarm optimization (FCPSO) based technique for solving the optimal load shedding problem leading to adaptive static load shedding (ASLSS) schemes. The candidate load buses to be shed are determined based on cost functions that considered distributed revenue profiles and social cost associated with critical and essential loads in the Nigerian power system. In order to develop the ASLSS, several optimal power flow (OPF) based load shedding scenarios via modified Matpower program were carried out. The optimal load shedding data generated there from were used for the training of radial basis function neural network (RBFNN) architecture. This enabled an ASLS framework to be developed specifically for the Nigerian grid system to achieve generation load balance requirements. The ASLSS has been exhaustively tested on the Nigerian power system. Several far reaching results obtained are presented and discussed extensively. In particular, the RBFNN forecast track the actual load sheds with maximum absolute predictive error of 16.3%.