Behavior-aware predictive energy control for sustainable operation of urban metro transport systems

Passenger-induced cooling loads in metro stations vary rapidly with passenger demand, whereas conventional building automation system (BAS) control responds only after changes in indoor conditions are detected, often resulting in insufficient or excessive cooling. This study proposes a passenger-flow-driven feed-forward control framework that integrates passenger-flow forecasting, behavior-induced dwell-time correction, cooling-load prediction, and proactive chilled-water flow regulation. Short-term passenger-flow forecasts are combined with a calibrated delay-probability model to estimate the additional passenger residence time associated with operational disturbances. The resulting behavior-corrected cooling load is then predicted 15 min ahead and used to generate the feed-forward control command. The framework was evaluated at residential-area, scenic-area, and terminal stations characterized by different passenger-flow patterns. After incorporating the dwell-time correction, the cooling-load prediction MAPE ranged from 2.0% to 3.3%, with a maximum error reduction of 81.5%. During a representative morning peak, the deviation between the supplied and actual cooling loads decreased from 27.2 kW under conventional BAS control to 1.5 kW under the proposed strategy. Over a continuous one-week evaluation period, the overall HVAC energy-saving rates ranged from 16.3% to 21.6% across the three station types. The corresponding cooling-season energy-saving rates ranged from 15.7% to 20.2%, indicating that the energy-saving performance was maintained under extended seasonal operation. These findings show that incorporating passenger dwell-time behavior improves the representation of passenger-induced thermal disturbances and enables a timelier response to rapid cooling-load variations. The proposed framework therefore provides a practical approach to improving load tracking and reducing HVAC energy consumption using routinely available metro operational data.

Building automation · cooling load prediction · feedforward control · HVAC energy savings · passenger behavior · passenger demand forecasting · Building Energy and Comfort Optimization · Integrated Energy Systems Optimization · Railway Systems and Energy Efficiency

• Uncertainty method and sensitivity analysis for assessment of energy consumption of underground metro station

  Open Access•Gangqiang Kong, Shuaijun Hu et al.•Sustainable Cities and Society•2023

  Add to list

• Energy consumption of subway stations in China

  Open Access•Bowen Guan, Liu Xiao-hua et al.•Sustainable Cities and Society•2018

  Add to list