Enhancing the efficiency of HVAC and heat exchanger systems using artificial intelligence and predictive control techniques

In this study, a numerical and Artificial Intelligence (AI) supported investigation of heat dissipation and energy optimization problem of HVAC and heat exchanger systems under steady state conditions is presented. A serpentine water–air heat exchanger with a tube diameter of 6 mm and 20 vertical passes was modeled using ANSYS Fluent within a computational domain of \(300× 300\) mm. The effects of water mass flow rate, inlet temperature, and flow configuration on heat transfer performance, outlet air temperature, and pressure drop were systematically analyzed. Results demonstrated a significant increase in heat transfer with increasing mass flow rate from 0.5 to 4 kg/s, while the pressure drop increased from approximately 8 to 110 kPa. The maximum heat transfer rate reached nearly 115 kW under AI-optimized operating conditions. AI-driven optimization further improved the thermal-hydraulic performance by reducing pressure losses by approximately 25–35% compared with baseline operation. In parallel, machine learning and neural-network-based predictive control models were developed to classify HVAC efficiency states, forecast indoor temperature behavior, and optimize energy consumption under comfort constraints. The proposed AI-based predictive control framework achieved nearly perfect HVAC efficiency classification accuracy and reduced cumulative HVAC energy consumption by approximately 65.3% while maintaining indoor thermal comfort within the 22–25 \(∘\) C range. Moreover, the coefficient of performance (COP) proxy exhibited substantial improvement under AI-assisted operation, confirming enhanced thermodynamic efficiency. The combined CFD and AI framework demonstrated strong capability for improving heat exchanger effectiveness, reducing energy consumption, and enabling predictive energy-aware HVAC management for smart building applications.