Metaheuristic Optimization Algorithm for Day-Ahead Energy Resource Management (ERM) in Microgrid Environment of Power System

Abstract

The day-ahead Energy Resource Management (ERM) problem with the aim to backing the functioning decisions of Virtual Power Player (VPP) in the microgrid environment. The aim of the VPP is to manage the available distributed energy resources as practically as possible with the objective of minimizing the operational cost and maximizing profits by reducing the need to buy energy from the external supplier or electricity market at high prices. The day-ahead ERM is executed the day before the energy trades are due. Typically, the considered trades periods are one-hour corresponding to 24 scheduling periods. A vital input to the ERM is each hour forecasting demand, which can be done using correct forecasting methods. VPP can aggregate the all types of energy resources like, DGs, PV, electric vehicles, energy storage, demand response and electricity market. The use of Vehicle to Grid (or G2V), PV, and energy storage technology can help to increase the penetration of non dispatchable uncertain renewable based DGs. The drawback of large DERs penetration is that the optimal scheduling problem turns into a complex optimization problem and becomes hard to be addressed by deterministic techniques, because these techniques can take a large execution time for obtaining the optimal solution. On the other hand, the VPP has its own optimal scheduling related time constraints. For these reasons, metaheuristic techniques are very useful to support the VPP in the computation of a good solution with a low execution time. This paper proposed the new metaheuristic algorithm called Cross-Entropy Variable Neighborhood Differential Evolutionary Particle Swarm Optimization (CE-VNDEPSO) for addressing the Energy Resource Management (ERM) problem of 25-bus microgrid systems. The effectiveness of CE-VNDEPSO algorithm is finding out by comparing its performance with the well-known optimization algorithms like, Variable Neighborhood Search (VNS), Differential Evolutionary Particle Swarm Optimization (DEEPSO), Particle Swarm Optimization (PSO) and Differential Evolution (DE).