PANDEMIC: Occupancy driven predictive ventilation control to minimize energy consumption and infection risk.

Area of Science:

• Building energy efficiency
• Infectious disease transmission modeling
• HVAC control systems

Background:

• Pandemic-driven increases in ventilation rates have raised building energy consumption.
• Energy-efficient strategies are needed to balance infection control and energy use.
• Accurate occupancy prediction is key to optimizing ventilation demand.

Purpose of the Study:

• To develop and evaluate an occupant-number-based model predictive control (OBMPC) algorithm for building ventilation.
• To reduce building energy consumption while maintaining low infection risk during the COVID-19 pandemic.
• To assess the impact of occupancy prediction accuracy on ventilation control.

Main Methods:

• Collected occupancy and HVAC data (March-July 2021).
• Employed four predictive models (ARMA_MLP, RNN, LSTM, NH_Markov) for occupancy forecasting (15 min to 24 h ahead).
• Calculated ventilation demand using the Wells-Riley equation based on predicted occupancy.

Main Results:

• Occupancy prediction achieved 85% accuracy (1-person offset), reaching 95% for 15-min ahead predictions.
• The OBMPC model maintained infection risk below 2% for 93% of the day.
• Reduced coil load by 52.44% and shifted peak load by 3 hours compared to a 24/7 full outdoor air system.

Conclusions:

• The OBMPC algorithm offers an energy-efficient ventilation strategy for buildings.
• Accurate occupancy prediction is crucial for effective ventilation control and energy savings.
• The developed model demonstrates significant potential for reducing HVAC energy consumption and peak loads.