Research on Energy-saving Regulation Model of Climate Compensation for Central Heating Station Based on Artificial Neural Network

Heating plants are usually based on heating specifications, local climatic conditions, building structure types, actual conditions of central heating systems, meteorological data and heating operation experience in the past years. The central heating system is a complex multi-variable nonlinear strong coupling system. So far, there is no precise model to get each parameter. The empirical parameters have proven to be effective and reliable in practice, but they are not necessarily optimal and comprehensive, and they cannot be directly used by computer control systems. In recent years, many new algorithms for fitting nonlinear systems have emerged, among which artificial neural network algorithm is one of the most successful algorithms. One of the essences of neural network algorithms is machine learning-the network is trained through continuous and extensive use of historical observation data and even empirical data to make the model optimally approximate the real model. Therefore, this paper analyzes the principles and objectives of heat exchange station regulation, clarifies the control parameters, and focuses on the control model based on neural network to identify the heat output of the heat exchange station and the outdoor atmospheric temperature changes. The advantages and disadvantages of the BP neural network (BPNN), general regression neural network (GRNN) and radial basis neural network (RBF) models were compared. Finally, the correctness of the model was verified with the actual meteorological data and heating data of Karamay in province Xinjiang using Matlab.