Optimizing Residential Energy Consumption Through Predictive Demand Side Management Using IoT and Edge Computing
Abstract
Smart cities use Information and Communication Technologies (ICT) and Internet of Things (IoT) to manage energy sustainably with smart grids and demand-responsive strategies. Unfortunately, the existing residential energy management systems have incremental challenges including poor electrical grid infrastructure, persistent outages in many developing countries, and cloud dependent data infrastructure challenges faced due to latency, privacy, security, and bandwidth constraints to efficiently address real-time demands in a timely manner. The current research proposes an Edge-IoT Based Predictive Demand Side Management Framework to address these challenges by integrating IoT devices with edge computing to achieve efficient residential energy consumption. The Framework does not transmit data to the cloud. It processes data at the network edge, with the energy demand preparation computing done using a Long Short-Term Memory (LSTM) neural network to predict new energy demand from historical consumption information and environmental conditions. The proposed method shows much better performance than existing baseline methods with Mean Absolute Percentage Error (MAPE) of 4.6%, Mean Absolute Error (MAE) of 1.1, Root Mean Square Error (RMSE) of 1.6 with peak load post intervention cases of over 26.4% which significantly led to grid stability and more optimized energy demand reduction.