Predictive Modeling and Analysis of Smart Building Energy Management: A Multi-model Approach

Abstract— In the context of sustainability and smart infrastructure, this research focuses on Smart Building Energy Management (SBEM) using Machine Learning and Time-Series Analysis. By combining real-world energy meter (EM) data with EnergyPlus (EP) Simulations, the study aims to enhance energy efficiency, cost-effectiveness, and environmental sustainability in modern buildings. Various Machine Learning (ML) models, including Artificial Neural Networks (ANN), Long Short-Term Memory (LSTMs), K-Nearest Neighbors (KNNs), Auto Regressive Integrated Moving Average (ARIMA), and Decision Trees, are examined to decipher energy consumption patterns across diverse smart building scenarios. Through systematic experiments and case studies, this research underscores the importance of precise data acquisition, preprocessing, and model selection. It emphasizes the need for adaptive energy management strategies and offers a Multi-Model approach to improve sustainability and energy conservation in smart buildings. Keyword—EnergyPlus, Energy consumption, Energy efficiency, Energy conservation, Machine learning, Smart building energy management (SBEM)