Developing Metaheuristic Optimization Techniques for Optimal Scheduling of Distributed Energy Resources on Smart Grid Environment

Abstract

In the MicroGrid (MG) environment, the high penetration of uncertain energy sources such as solar Photovoltaics (PVs), Energy Storage Systems (ESSs), Demand Response (DR) programs, Vehicles to Grid (V2G or G2V) and Electricity Markets make the Energy Resource Management (ERM) problem highly complex. All such complexities should be addressed while Virtual Power Player (VPP) or aggregator looks for the maximization of profit or minimization of operating cost by optimal scheduling of all types of available energy resources from the MG system. Moreover, it may use its own assets, e.g., energy storage systems (ESS) to supply the load demand. In addition, a V2G feature that allows the use of energy in the battery of electric vehicles (EV) is also possible. The VPP establishes bilateral energy contracts with those who seek electricity supply, e.g., residential and industry customers. In this case, it is assumed that the VPP does not make profits from the supply of energy to ESSs and EVs charging. The main idea is that the optimization algorithm can perform the energy resource scheduling of the dedicated resources in the day-ahead context for the 24 hours of the following day. newlineSince the VPP performs the scheduling of resources for the day-ahead (i.e., the next 24 hours), it relays in the forecast of weather conditions (to predict renewable generation), load demand, EV trips, and market prices. However, the assumption of perfect or highly accurate forecast might bring catastrophic consequences into the operation of the grid when the realizations do not follow the expected predictions. Due to this situation, it is desired that the VPP determines solutions that are robust to the uncertainty inherent in some parameters and the environment. Four aspects newlineVI newlineof uncertainty that affects the performance of a solution are considered in this work, namely: a) Weather conditions, b) Load forecast, c) Planned EVs trips, and d) Market prices. Therefore, the VPP should find solutions that provide not only an optimal (or near-optima