Deep Learning-Based Energy Consumption Forecasting for Smart Buildings Using Internet of Things Sensors

Abstract

Energy-efficient buildings play a vital role in achieving sustainable urban development and reducing greenhouse gas emissions. Accurate forecasting of building energy consumption enables intelligent energy management and operational cost reduction. This paper proposes a hybrid deep learning framework that combines Temporal Convolutional Networks (TCNs), Gated Recurrent Units (GRUs), and attention mechanisms to forecast short-term and long-term energy consumption using data collected from IoT-enabled smart buildings. Environmental conditions, occupancy behavior, and historical energy usage patterns are incorporated into the prediction model. Experimental results demonstrate improved forecasting accuracy and lower prediction error compared with conventional machine learning and statistical forecasting methods. The proposed framework supports intelligent building automation systems by optimizing energy utilization while maintaining occupant comfort.