Application of deep reinforcement learning to intelligent distributed humidity control system

The indoor environment of buildings is complex and changeable, and it is difficult to ensure that the indoor humidity is uniform and stable while employing a centralized humidity control system. To address this challenge, this paper proposes an intelligent distributed humidity control system based on model-free deep reinforcement learning. The proposed system consists of three parts: an intelligent controller, distributed facilities, and distributed sensors. The distributed sensors are used to monitor the environmental parameters. This study developed a reinforcement learning algorithm called RH-rainbow and deployed it in distributed facilities. In RH-rainbow, the reward consists of the mean absolute difference of humidity and the energy consumption of distributed facilities. The action is the humidity setpoints and fan settings of the constant humidity machines. The performance of RH-rainbow was evaluated and compared to that of other algorithms in two scenarios with different air outlet settings under different sensor numbers, reporting time intervals, and external interference modes. It was found that RH-rainbow is superior to manual strategies, the traditional analog control strategy, DQN, and PID in terms of uniformity, anti-interference ability, and energy consumption.