A proactive 2-stage indoor CO2-based demand-controlled ventilation method considering control performance and energy efficiency

Abstract

This paper presents a novel method, named proactive 2-stage demand-controlled ventilation (P2S-DCV) method, to maintain indoor air quality (IAQ) and reduce the energy consumption of multi-zone ventilation systems. The proposed P2S-DCV method applies a proactive control scheme, which predicts future indoor CO2 concentration and supplies proper ventilation to each zone. The method includes two stages. In Stage I, a DNN prediction model is established to predict the future CO2 concentration to calculate the corresponding demand airflow. In Stage II, a reinforcement learning method is designed to achieve rapid and accurate control, and further reduce the energy consumption by optimizing the fan pressure and damper positions. A 5-zone ventilation system is established to validate the proposed P2S-DCV method. The experiment verifies that: a) it can maintain comfortable IAQ via predicting the change of future indoor CO2 and applying effective ventilation control in advance; b) it can improve the control performance, the accuracy is maintained within 8 % (satisfied the ASHRAE Standards), and the control time is maintained within minutes. It can reduce the regulating time by 83.62 % compared with ASHRAE Ratio method, and up to 51.68 % compared with PID method; c) it can reduce the fan energy consumption by 16.4 % compared with ASHRAE Ratio method, and up to 21.8 % compared with PID method; d) it has good generalization ability for various IAQ requirements and ventilation systems with different topologies.