Development Of A Predictive Control System For Managing Hybrid Renewable Energy Sources Using Machine Learning Algorithms

Abstract

In this article, three algorithms are proposed to improve a predictive control system to manage a hybrid energy system. This system consists of a solar cells station, wind turbines station, and a fossil fuel energy station that represents the energy of the public electrical grid to feed a dynamic load represents the load of  Lattakia city.

 The priority is to invest in renewable energy resources and then meet the rest of the load requirements by fossil fuel energy, thus reducing the quantities of fossil fuel and their economic return. These algorithms depend on the predicted values ​​of three basic variables: solar radiation intensity, wind speed, and dynamic electrical loads. These values ​​were predicted by machine learning and deep learning algorithms in a previous research. The importance of the proposed algorithms comes from their ability to predict the appropriate quantity of fossil fuel energy during the hours of the day under changing conditions of weather parameters and electrical loads. Matlab/Simulink software is used to achieve the simulations, where three simulations are achieved, each with a duration of ten days, representing different seasons of the year, in order to verify the efficiency of the proposed algorithms and compare their performance. The simulation results show that the proposed Slope Algorithm (SA) is the best in terms of system efficiency on the one hand and savings in fossil fuel on the other hand.