Digital Twin-Driven Real-Time Monitoring and Control System for Building Energy Management in Smart Green Buildings

INTRODUCTION: Smart building energy management requires integrated monitoring, prediction, optimization, and control to improve energy efficiency while maintaining suitable indoor environmental conditions. OBJECTIVES: This study aims to develop a Digital Twin-based real-time monitoring and control system for smart building energy management and to optimize heating, ventilation, air conditioning, lighting, and ventilation settings. METHODS: The proposed Digital Twin-driven framework integrates Kalman Filter-based state refinement, Res-DL-PEM occupancy prediction, building-state transition modeling, energy-surrogate estimation, and DT-A-MOPSO optimization. The Kalman Filter is used to update the state of the Digital Twin; Res-DL-PEM to forecast occupancy. The state-transition model assesses the future condition of buildings and the energy surrogate predicts the energy consumption. DT-A-MOPSO derives optimal energy consumption, CO₂ concentration and thermal discomfort by choosing HVAC, light and ventilation settings. The Digital Twin is updated with the simulated control responses to make further predictions and optimizations. RESULTS: The occupancy prediction model achieved an accuracy of 0.9776 and an area under the receiver operating characteristic curve of 0.9927. The energy surrogate achieved an R² of 0.9822, while state-transition prediction achieved R² values ranging from 0.978 to 0.995. The optimization process reduced predicted energy consumption. CONCLUSION: The proposed Digital Twin-based framework integrates real-time monitoring, prediction, energy estimation, and multi-objective control, providing an effective approach for improving smart building energy management while considering carbon dioxide levels and thermal comfort.