Research Of Heating Network System Simulation And The Primary Network Optimization Control

With the rapid development of modern city, the central heating system becomes the important system in northern city of our country. Because the heating system has characteristics of nonlinear, multi variable, large lag, and the heat load has characteristics of time varying and uncertainty, so the heating system is not an easy control system. How to save energy, reduce loss and control the heating system, is an important problem of heating enterprises.Firstly, the heating network model is researched. In the past, the heating network model is established according to the network mechanism, in which the transfer function and reduce the amount of calculation of heat supply network is simplified, and some network parameters are ignored. So some gaps are existed between the model and the real network system. In this paper, a one-dimensional fluid simulation software Flowmaster is used to build model of the central heating system. The simulation model of central heating system is build which has heat source, actual thermal power station, heat users, and in which the number of stations is fourteen, is a relatively large central heating system. The model can be simulated in some simulation environments, which need to be set many different parameter. the model is close to the real system.AS the central heating system heat load and temperature of atmosphere is uncertainty, the system need to be established short-term forecasting model, which used in a constraint condition of heat distribution. Each thermal power station use the neural network prediction model predict heat load, that using dynamic K mean clustering algorithm and recursive least squares (ROLS) modified RBF neural network, and thermal heat station use the gradually refresh history data to realize the short-term load forecasting. A primary optimal control problems is designed objective function of a heat source to produce the fewest calories. The particle swarm optimization (Particle Swarm Optimization, PSO) algorithm optimiza the objective function,were found in each time and follow the trend of total amount of heating heat load prediction model of each substation to the distribution of heat.The PSO optimization can make the heating system provides the least heat and satisfy most users,which achieve the purpose of saving energy.The simulation results show that the total heat distribution of RBF neural network. Useing centralized heating system simulation model of PSO algorithm and Flowmaster to build a joint simulation.Through the model data curve simulation can be found after the control algorithm of the model can be used in accordance with the thermal heating trend of user’s energy demand, the system energy consumption has certain guiding significance.