Research On Energy-saving Control Of Secondary Pipe Network Of Central Heating System Based On BP Neural Network

Building energy consumption is one of the three major energy consumption in China,among which the energy consumption of central heating system accounts for a huge proportion in building energy consumption.For a long time,China's traditional central heating system,especially the secondary pipe network system,has suffered from hydraulic imbalance due to its extensive management and low level of intelligence.As a result,the indoor temperature of heating users is uneven,resulting in a large amount of heat loss.Aiming at this problem,a central heating district in China's severe cold region(Binzhou,Shandong Province)is taked as the research object in this article.First of all,the influence of hydraulic imbalance on the secondary pipe network in heating is analyzed,and the control strategies for the heating pipe network in recent years are summarized,and the intelligent balancing valve to control the secondary pipe network in front of the building is proposed in this paper.Then,the kindergarten,a representative building on the heating time and the heating end,was selected as the control object,and the equipment and test instruments were designed and installed.The experimental data obtained were used to obtain the transfer function of the heating system of the kindergarten,which laid a foundation for the subsequent control optimization.Considering the different time-lag effects of different controllers,the traditional PID controller,improved single neuron PID controller and BP neural network PID controller were used to simulate and analyze the transfer function of the kindergarten heating system.Simulation results show that compared with the conventional PID control simulation curve,the improved single neuron PID reduces the overshoot by 75%,the adjustment time is reduced by 0.5s,and the steady state time is shortened by 1s.Compared with the improved single neuron PID control simulation curve,the BP neural network PID control reduced the overshoot by 4%,adjusted the time by 0.6s,and shortened the system stabilization time by 3.5s.The results show that the BP neural network PID control algorithm has the best time-sharing control effect on the opening degree of the intelligent balance valve.In order to verify the industrialization feasibility of BP neural network control algorithm,the selection of circuit system components,circuit design,programming and simulation experiments are involved in this paper.The results show that the intelligent control of heating system can be realized by using the BP neural network control algorithm through the single-chip controller.In terms of energy saving research,the optimal BP neural network thermal load prediction and valve opening degree prediction model control method was applied to the district heating system after the transformation to control the energy saving of the heating system.By comparing and analyzing the operating data of the same period before and after the system renovation,it is concluded that on the basis of meeting the indoor temperature requirements of users,using BP neural network prediction method to control the system,the end user's heat consumption can be reduced by 7.4%,and the total system energy consumption can be reduced by 9.5%.The experimental data show that the energy consumption of district heating system can be effectively reduced by BP neural network predictive control method.The experience in energy saving transformation and engineering application of secondary pipe network of central heating system and similar system engineering can be taken as the corresponding basis and reference through the research in this paper.