Energy-Aware Scheduling for Sustainable Manufacturing: Integrating Production Systems and HVAC Control

Achieving sustainability in the manufacturing sector calls for systemic reductions in energy consumption and carbon emissions without compromising productivity. In the global energy consumption landscape, the manufacturing sector accounts for a significant proportion and is a major source of carbon emissions, with manufacturing systems and HVAC (Heating, Ventilation, and Air Conditioning) systems being the principal energy consumers. Existing research typically optimizes these two systems independently, neglecting their dynamic coupling; production scheduling determines equipment power and heat dissipation, which alters building thermal loads and consequently affects HVAC energy consumption. To address this problem and advance sustainable manufacturing practices, this study proposes an energy-aware scheduling framework integrating manufacturing and HVAC control. A WOA-XGBoost energy consumption prediction model is constructed, employing the Whale Optimization Algorithm to tune XGBoost hyperparameters, achieving a prediction accuracy of R2 = 0.937 on the Shanghai typical meteorological year dataset. The HVAC decision variables are defined as five operational control variables—supply air flow rate, fan total pressure, ERV sensible/latent heat recovery effectiveness, and ventilation air flow rate—ensuring the physical realizability of scheduling solutions. An integrated scheduling-and-control model incorporating production capacity constraints and electricity demand response is then formulated and solved using a hybrid Particle Swarm Optimization algorithm. Validation on a five-machine, four-buffer flow shop demonstrates that the proposed framework reduces total electricity cost by 8.85% and total energy consumption by 14.88% in summer compared with a physics-based coupling baseline, with all metrics exhibiting coefficients of variation below 4% across ten independent runs. These results demonstrate that the proposed data-driven framework provides a practical and scalable pathway toward sustainable manufacturing by jointly reducing energy use and associated carbon emissions while maintaining full production throughput.