The Heating Load Forecasting For Heating Pipe Network Control System

With the development of the society and the constant improvement of China’s economic and social system, heating mode in our country has made a lot of improvement.The way to charge by heating area replaced the way to charge by the heat based on wireless remote system. It becomes more and more attention by the department of heating problem that how to realize the heat data collection and management, heating prediction and reasonable control of heating. Due to the large and the complexity of the central heating pipelines network system, it is very important for environmental protection and energy saving that how to improve the accuracy of the forecast in heating load. Therefore, the predictive control of the heating load has become a hotspot research of current heating industry.In this paper, through the study of artificial neural networks, compared the features between feedforward neural networks and recurrent neural network, and analysis the feed-forward neural network-BP neural network and generalized regression network detailly. Finally, selected the BP neural network to predict the heating load.In this paper, we analyze and summarize the characteristic of load forecasting in heating pipe network and the factors of heating load, selecting date、outdoor temperature、water temperature、return water temperature、water pressure、the heating load that the same time of the previous day’s as the inputs, the heating load is selected as output. Using the BP neural network that has been selected as the tool to accomplish the heating load forecasting, to wit the inputs selected are used as the BP neural network’s inputs. We programed and adjusted the weights and thresholds in order to obtain a more perfect neural network model in the effect of prediction.After selecting the input and output, using MATLABr2012 a to make simulation about the heat load forecasting, resulting the map of the error between forecasting and expected. By adjusting the input volume, weights, thresholds and other parameters to make the effect of the heat load forecasting accurate. And make the error allowable.Finally, after completion of the heating load forecasting, using the development platform in MATLAB—GUI to design the interface of the heating pipe network system. Completed the login screen, basic information interface, data query interface, statistical analysis interface design.