Search Results

Abstract: The increasing variation in residential electricity consumption has created a need for accurate electrical load forecasting to support efficient energy management and reliable operation of smart-home power systems. This paper proposes a deep learning-based technique combining Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM) for residential electrical load forecasting. The proposed approach utilizes historical household electricity consumption data to learn temporal patterns and nonlinear variations in residential load demand. The RNN component captures sequential relationships in the electrical load profile, while the LSTM component effectively learns long-term dependencies and reduces the limitations associated with conventional recurrent networks. The proposed model is evaluated for short-term residential load forecasting using suitable performance measures such as Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Mean Absolute Percentage Error (MAPE). The forecasting results demonstrate the capability of the RNN-LSTM approach to accurately track variations in household electricity demand and improve prediction reliability. Accurate load forecasting can support peak-load management, demand response, appliance scheduling, battery energy storage operation, rooftop photovoltaic integration, and smart-home energy management. The proposed technique provides an effective electrical load forecasting framework for improving energy utilization and supporting reliable operation of future smart residential distribution systems.