Development of an Energy Management System for a Microgrid Using Neural Networks. Case Study: San Cristobal Island, Galapagos Archipelago

Abstract

In this paper, an energy management system EMS hourly energy management system for a renewable energy system (HRES) is presented. The proposed HRES is composed of hybrid wind turbine (WT), solar photovoltaic (PV) panels, a diesel generator (DG) and a Distributed Collector System (DCS), as primary energy sources. In turn, an energy storage system (ESS), which is a battery sub-system. The wind turbine, PV panels and DCS system are made to work at peak power, while the battery acts as storage. The EMS uses intelligent rule-based controllers and optimizers to meet the energy demanded by the load and maintain the state of charge (SOC) of the battery between certain target margins, while trying to optimize the utilization cost and lifetime of the BESS. Simulation results show that optimization-based control meets the objectives set for the HRES EMS and achieves a total cost savings of 23.5% over other simpler control state-based EMSs.