Active fault-tolerant control based on DDQN architecture applied to HVAC system

This paper presents the development and application of an active fault-tolerant control for temperature regulation in an intelligent building. The architecture uses a combination of fault detection and diagnosis techniques and optimal control based on deep reinforcement learning to ensure correct system operation in the presence of faults with good performance. Promising results have been obtained with this approach. It provides an effective solution for maintaining heating, ventilation, and air-conditioning system performance and reliability, which is essential for ensuring user comfort and safety.