Intelligent HVAC System Control Method Based on Digital Twin

With the development of digitalization and intelligence, the heating, ventilation and air conditioning (HVAC) system is facing the challenges of changing environments, stricter quality standards and increased energy efficiency requirements. In this paper, a digital twin generation method for environmental temperature and humidity control based on time series prediction (TSP) is proposed. The system is comprised of a forecasting model attention-based temporal convolutional network (ATCN) and a control model proximal policy optimization (PPO). Through the analysis of temperature sensor data, we propose a regulatory method underpinned by PPO. The effectiveness of the forecasting model has been verified on five datasets, and the control model can provide strategies that are superior to manual methods in the real environment.