Analysis On The Influence Factors Of Indoor Temperature And Study On Thermal Load Forecasting

Data statistics display that in recent years, the national energy consumption rose sharply. So energy saving and emission reduction as a basic national policy of the country is imperative. In order to achieve energy-saving emission reduction of heating industry, central heating of urban buildings has become the main application form of winter heating in North China. However, the central heating system is controlled by multi variables, and the operation is complicated, so the disadvantages of the internal correlation, hysteresis and nonlinearity are serious. Therefore, in the heating system, thermal load for short term prediction of building is very important. Thermal load forecasting as the preparatory work of the heating system control, which is a reasonable and efficient guidance of heating operation, improve the thermal efficiency of the heating system, to achieve energy saving, is significant to energy-saving emission reduction and environmental protection significance.In this thesis, we analyze the factors that affect the indoor temperature. Collected heating period of relevant data, including outdoor temperature、windspeed、sunshine、the time and the temperature of the room, to analyze and research the data using radial basis function neural network (RBF), and ultimately determine the outdoor temperature^ wind speed、sunshine、the time is the influence factors of the indoor temperature and the size of the extent of their influence on the indoor temperature.After analyzing the characteristics and current situation of centralized heat load forecasting, Using gray prediction、BP neural network and grey neural network to model the historical data collected during heating period in North China. Through the training of a lot of historical data, the best model parameters are determined in order to get the best thermal load forecast model for the actual heating situation. Next, MATLAB is used to predict the thermal load, and use the forecast results to guide the heating. The results show that under the same or similar conditions, using the forecast results of gray neural network model to guide the building heating, the building interior temperature is more close to setting the comfort temperature.