A Robust TCN-Based Energy Forecasting Framework for IoT-Controlled Smart Buildings in Smart Cities

In response to the critical need for energy conservation and carbon emission reduction globally, this paper introduces an innovative framework for hourly energy forecasting in smart buildings. Utilizing data collected from sensors, this framework forecasts future energy consumption, enabling efficient monitoring and control of electrical appliance usage through advancements in the Internet of Things (IoT). This reduces energy waste significantly. Diverging from traditional statistical models, which often falter with nonlinear power load sequences, our approach employs a Temporal Convolutional Network (TCN). This robust framework is designed to improve accuracy, robustness, and generalization in forecasting energy usage in buildings. It incorporates an adaptive method to enhance the practical applicability and dynamic updating of our proposed solutions in real-world scenarios. A comparative analysis with existing energy forecasting methods demonstrates the superior performance and adaptability of our framework, showing a reduction in error of more than 20%. This study aims to establish a new benchmark in energy forecasting for IoT-enabled smart buildings, offering a scalable and precise solution for energy consumption management aligned with global sustainability goals.