Research on machine learning-based room temperature prediction algorithm and fault warning system for central air conditioning of water system

Focusing on the challenges posed by the significant lag in room temperature response within the complex central water system, this paper introduces an innovative room temperature prediction algorithm and fault warning system, leveraging machine learning. This system is designed to cater to varying ambient and water supply temperature conditions. Leveraging historical data, the system employs machine learning techniques to automatically learn and fine-tune the room temperature prediction model. Three distinct machine learning algorithms are employed to construct the room temperature prediction model. Through rigorous comparison, it is discovered that the artificial neural network (ANN) model exhibits the highest degree of accuracy. This algorithm boasts remarkable prediction precision, enabling real-time monitoring of room temperature variations, thereby ensuring a comfortable indoor environment for users. Furthermore, the system incorporates a fault detection and alarm functionality. This feature allows for the timely identification and forecasting of potential faults within the room temperature control system, ensuring seamless system operation. Extensive testing and verification have been conducted in real-world projects, and the experimental results are promising. The system effectively enhances the accuracy and reliability of room temperature control, holding immense potential for widespread application.