Development of a Hybrid Fuzzy Logic and Linear Programming Algorithm for Optimal Load Shedding in Islanded Microgrid

Abstract

The islanded microgrid (IMG) entirely depends on Distributed Generations (DGs) like Micro Hydro Power (MHP), solar photovoltaic (PV), wind, and fuel cells among other sources of energy. The stochastic nature of solar PV, wind and local loads creates an imbalance between generation and the loads. These disturbances can plunge the IMG into an emergency power crisis which can lead to a cascaded blackout if no remedy strategy is brought on board to restore the power to a balance. To avert the power crisis in the IMG load shedding (LS) is done as a last resort after all control mechanisms have been exhausted. The conventional methods of LS used in grids perform poorly when applied to the IMG because of low convergence and settling time. Recent researchers have found that adaptive methods for LS specifically the hybrid method perform optimal LS to curb the power system from collapsing in times of contingencies. The hybrid method of LS using a Fuzzy Logic Controller (FLC) and Linear Programming (LP) was used to optimize the amount of LS in the IMG. In this method, the objective function was formulated and solved by the Fuzzy Linear Programming (FLP) algorithm. The inputs to the controllers are power generated and power demand of the IMG. The loads were classified according to priorities using fuzzy membership functions while optimization of loads shed was achieved by the LP algorithm. The simulations consisted of generation contigencies, power demand in which 10 overload contingencies were simulated, voltage profiles and power losses. The results depict FLP algorithm finds the best steady-state operating point with a minimal amount of load curtailment. The scheme minimizes loading at the buses until total load demand matches generation to restore power balance within a single LS step. In comparison to GA 77.04%, ABC-ANN 84.03%, PSO-ABC 85.50% the proposed FLP algorithm was able to shed optimal amount of load quantities resulting in 86.10% voltage profile recovery. The developed algorithm was tested by performing simulations on IEEE 14 bus systems on a Matlab Simulink platform.