Energy Management Design for Industrial Demand Considering PV Power Prediction and Battery SOC

Electricity is a very important need for daily life. It began to be developed and researched additional energy in the form of renewable energy sources, such as the utilization of solar energy using photovoltaic (PV). The utilization of solar energy in PV can be connected to the grid. For that, we need a stable performance that grid usage can be reduced. So we use Battery Energy Storage (BES). The role of the Energy Management System (EMS) is needed to determine the charge/discharge of battery and strategies to minimize grid power. The proposed research carried out an EMS on a system consisting of PV, BES, and load. First is predicting PV output power using Artificial Neural Network (ANN) by considering the parameters of irradiance, temperature, and time. Second, it is considering battery State-Of-Charge (SOC). The battery is kept at the minimum and maximum SOC limits to maintain the battery. These two parameters become the input of EMS. EMS using ANN to determine charge/discharge of the battery and minimize grid power usage by determining load usage. This research was simulated using MATLAB Simulink. EMS uses ANN then compared with results of EMS with the rule-base algorithm.