Dynamic Cloud and ANN based Home Energy Management System for End-Users with Smart-Plugs and PV Generation

Over the past decades, the importance of energy management has been raised due to increasing electricity demand and consumers' unawareness of their electricity consumption. The paper proposes a Home Energy Management System (HEMS) that implements an Artificial Neural Network (ANN) and reinforcement learning-based algorithm to schedule the home appliances as well as an optimized and efficient way of profiting from renewable energy source with the utilization of energy storage systems. The objective of the HEMS is to decrease energy cost, customer dissatisfaction, and grid overloading. Two types of appliances were considered: non-shiftable controllable, shiftable interruptible. A simulation of the case study where the forecasted values were fed to the HEMS algorithm demonstrated a total profit increase by 15% due to the renewable energy source, making the value of total profit 63.5 units in one day. The simulation was done for a single house loading profile and throughout the capacity change of the energy storage system, a maximum profit was derived. These results show the efficient function of HEMS with the utilization of the proposed ANN, reinforcement learning, and energy decision algorithm.