Short-term Heating Load Prediction And Development Of Heat Network Dispatching System

Central heating is an important livelihood project in northern China.However,the traditional heating enterprises are generally lack of information and intelligence,and there are various problems to be solved in the actual scheduling production process.Based on the field investigation of Hongshan thermal power company of Shaanxi Yulin Energy Group,it is found that there are three main problems:(1)in the process of central heating,only relying on industrial configuration software for equipment control,the degree of intelligence is low,the interface is relatively simple,and the traditional dispatching mode needs to be upgraded;(2)the heating mode is extensive,and lack of precise control of heat users;(3)dispatchers mainly use subjective experience in heat load forecasting,lack of theoretical model support,it is difficult to achieve on-demand heating,resulting in a waste of resources.In order to effectively solve the practical demands of energy saving,on-demand heating and information upgrading of heating enterprises,this paper focuses on the research of short-term heat load forecasting and the development of heat supply network dispatching system:(1)Establish the heat load forecasting model.Based on the analysis of the influencing factors and numerical variation characteristics of heat load in the process of central heating,combined with machine learning technology and particle swarm optimization algorithm,a multi-dimensional heat load forecasting model based on pso-xgboost is realized.The MAPE value is 3.81%,and the overall forecasting accuracy is 96.19%(2)Combined with pso-xgboost model,Internet of things technology,springboot framework,websocket communication,multithreading,redis cache,ecarts data visualization and other mainstream frameworks and technologies,the heat supply network scheduling system is developed.The system consists of four layers: user interface layer,business logic layer,data access layer and data perception layer.The heat source,primary network,heat exchange station,secondary network and heat users are connected into a complete heating network through Internet of things technology.Combined with Modbus protocol and websockt technology,the remote control of heating field equipment and key data such as temperature,pressure and flow of supply and return water are realized In order to improve the performance of the system,we need to use multi thread technology to reduce the waiting time of data storage,reduce the access pressure of My SQL database with redis cache technology,use NB IOT technology to achieve accurate control of hot house temperature,integrate PSO xgboost multi-dimensional forecasting model to assist the dispatcher in load forecasting,and use ecarts data visualization technology to clear all kinds of data Clearly and intuitively show to the system page,so that the dispatcher can more clearly and intuitively control the overall situation,scientific and reasonable production dispatch.The heat supply network dispatch system developed in this paper integrates remote control,information archiving,load forecasting,abnormal alarm,data visualization and other functions.It has been applied to the heat supply company for heat supply production dispatch,and the operation effect is good.It has guiding significance for the actual heat supply production of the heat supply company,and has been affirmed by the heat supply company.