Smart Residential Energy Management System (REMS) Using Machine Learning

Electricity consumption is increasing day by day in every corner of the world which leads to imbalance of supply and demand. In the current scenario, everybody searches for cheaper and environmentally friendly approaches in accessing electricity. In order to mitigate falling into a huge energy crisis, both the utility and consumer could involve in energy management which is a convenient and trending approach. We believe it should be started from the ground level, which is the consumer scope. This paper proposes a method for a smart Residential Energy Management System (REMS) using Machine Learning. Specifically, the proposed system (REMS) effectively switches pre-prioritized possible loads without limiting consumption, between the grid and renewably energized local storage with rooftop solar at the residential premises, using Machine Learning algorithms. Reduction of the electricity bill with a reliable power supply as much as possible in residential premises is also concerned, with the use of by load shifting algorithm. Available average solar power prediction using Artificial Neural Network and Optimum utilization of available solar power generation and the energy storage using Reinforcement Learning features are also included in the system. Ultimately, the grid dependency is reduced at the Residential premises.