Intelligent Energy Management and Prediction of Micro Grid Operation Based On Machine Learning Algorithms and Genetic Algorithm

Micro grid energy management has become critically important due to inefficient power use in the residential sector. High energy consumption necessitates developing a strategy to manage the power flow efficiently. For this purpose, this work has been divided into two phases: The first is the "ON/OFF" operation, which has been executed using a genetic algorithm for the hybrid system, including diesel generator, solar photovoltaic (PV), wind turbine, and battery. Then, in the second phase, the output results were used as input in three algorithms to predict load and supply dispatch one month ahead. This study has two objectives; the first is to decide which energy source should meet the load one month ahead. The second is to compare the outcomes of machine-learning techniques, namely Random Forest (RF), Decision Tree (DT), and K-Nearest Neighbours (KNN), to determine the one that performs the best. The results indicated that the DT technique has the best performance in the application of classification with an accuracy of 100%. The findings also show that the RF approach gives acceptable results with an accuracy of up to 98%, and the KNN algorithm was poor in terms of accuracy with a value of 28%.