Predictive energy management in large-scale grid connected PV-batteries system

PV systems can mitigate their production unpredictability with storage system employment. Installation of storage systems need to be designed with precise energy storage system (ESS) rating. ESS become saturate in case of lower capacity, which results into failure of achieving the targeted power commitment and cause penalties. To manage the integration of grid connected PV-Batteries system; a precise energy forecasts is required to improve the power quality, security and reliability. This paper investigates a predictive energy management (PEM) method which is as an integration of forecasting method and batteries storage system. This PEM to control the real-time power production of a large-scale grid connected PV-Batteries system and to ensure reliability of power production all the time during the day. The short-term forecasting method will support the batteries storage system to manage the power production from PV system for the next day. The model contains of short term forecasting artificial neural networks (ANN) method connected to a 1 MW PV system and 1 MW lithium-ion batteries connected to the grid. This full model will develop a smart injection system to control the energy flows and to predict PV power production ahead of time of twenty-four hours (24h). PV-batteries system is simulated by MATLAB-SIMULINK under different circumstances of various irradiance levels during the day. The output results are plotted and evaluated with respect to power of PV system, batteries and grid.