ADAPTIVE AI scheduling of building HVAC to charge phase change thermal batteries with elevator regenerative braking heat

High-rise buildings consume significant energy due to intensive HVAC and elevator operations. This study proposes an adaptive energy management framework that captures regenerative braking heat from elevators to charge Phase-Change Thermal Batteries (PCTBs), which are subsequently used to assist HVAC operations. A hybrid co-simulation platform integrating TRNSYS, MATLAB/Simulink, and Python (TensorFlow) was developed to couple building thermal dynamics, elevator regenerative heat recovery, and AI-based HVAC scheduling. The proposed reinforcement learning (RL)–driven scheduler dynamically coordinates HVAC power modulation, PCTB charging/discharging cycles, and occupancy-based thermal demand in response to real-time weather and elevator activity. Simulation results demonstrate that the adaptive AI scheduler reduces HVAC energy consumption by up to 18 % and peak load by 12 %, while maintaining indoor comfort within ASHRAE-55 standards. Regenerative heat utilization efficiency exceeded 75 %, confirming effective capture and reuse of elevator braking energy. Extended seven-day simulations further validated system robustness and consistent performance under varying occupancy and climatic conditions. Economic assessment indicates a payback period of 5–7 years, emphasizing the financial viability of the proposed integration. By uniting regenerative energy recovery, latent thermal storage, and adaptive AI control, this research establishes a scalable framework for intelligent HVAC operation in smart buildings. The findings highlight substantial potential for achieving energy circularity, reduced carbon emissions, and enhanced sustainability in high-rise urban environments.