Intelligent Fresh Air System for Synergistic Optimization of Air Quality and Energy Consumption

As building energy consumption becomes increasingly prominent, optimizing the balance between energy use and indoor air quality in fresh air systems—key devices for regulating indoor air quality—has become a major research focus. Traditional static control strategies for fresh air systems often rely on fixed thresholds, leading to response delays and high energy consumption. This study proposes a two-stage cooperative optimization framework for intelligent fresh air systems based on “prediction-control.” It employs Random Forest and Transformer models to predict multivariate time series of environmental parameters such as CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> concentration, temperature, and humidity. Based on the predicted data, the PPO reinforcement learning algorithm is then used to optimize the fan control strategy. The results show that the intelligent agent can effectively reduce CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> concentration (with a set threshold of $\mathbf{1 0 0 0}$ ppm) and lower energy consumption, with cumulative rewards continuously improving.