AI-enhanced BIM Systems for Real-time Energy Performance Simulation

Abstract

Amid urgent climate goals, the building sector’s high energy use demands smarter design and operation strategies. This paper critically explores recent peer-reviewed literature on AI-enhanced Building Information Modeling (BIM) systems for real-time energy performance simulation, highlighting how artificial intelligence (AI) techniques are integrated with BIM to improve building energy efficiency, reviewing global studies that combine BIM-based energy modeling with machine learning, optimization algorithms and digital twin frameworks. Key findings indicate that AI can dramatically accelerate energy simulations and enable real-time predictive analysis in both design and operational phases. Surrogate models (neural networks and gradient-boosted trees) trained on BIM-generated data achieve prediction accuracies above 90%, providing instant feedback on design alternatives. In operation, AI-driven digital twins linking BIM with IoT sensor data allow continuous monitoring and predictive control of building systems. These approaches have led to significant energy savings (often >10%) and support net-zero energy goals. However, challenges persist in data interoperability, model generalization and industry adoption. This paper contributes an integrated perspective on current methods, empirical outcomes and emerging themes (explainable AI and uncertainty analysis), outlining future research directions to fully realize real-time energy simulation in smart sustainable buildings.