Research On Dynamic Prediction And Supervision System Of Energy Consumption In An Office Building

According to the Statistics of China Building Energy Consumption Report in 2020,the energy consumption of China's buildings has reached 37% of the total energy consumption of the society,and shows a large trend of growth.As a kind of extensive buildings,the area of the office building accounts for nearly 30% of the completed area of all kinds of buildings.Therefore,the paper conducts a detailed study on the energy consumption prediction model and supervision system of the office building,with the specific contents as follows:(1)A survey on a number of office buildings was conducted,to understand the background and basis of the study on energy consumption of office buildings,and the problems existing in the current energy consumption supervision system were summarized,such as improper supervision methods,unclear energy consumption indicators,and insufficient data mining.Therefore,the standard data set of energy consumption of office buildings was established according to the energy consumption data of office buildings obtained from the survey,and the energy consumption of five typical office buildings was compared.Then,taking an office building in Xi'an with less ideal energy saving effect as an example,the sub-measurement method of its energy consumption monitoring system is designed,and based on this method,the original energy consumption data of the office building from January to December of 2017 is analyzed,and the corresponding energy saving strategy is proposed.(2)The energy consumption checking and prediction model was established based on the operating energy consumption of an office building in Xi'an.Firstly,Bayes and Expected Maximum Algorithm are used to identify and repair the energy consumption data of an office building.Then determining the transfer function of the BP neural network,the network layers and each layer of the number of neurons.Considering thedata noise interference and the BP neural network easy to fall into local optimum,the research adopted the Fuzzy C-means for grouping the data of energy consumption,and the BP neural network weights and threshold were optimized by the Wolf Algorithm,so as to set up a combined forecasting model of the office building energy consumption.Simulation results show that the accuracy of the optimized BP neural network prediction model is improved significantly.(3)The development of energy consumption monitoring and management system for an office building were completed.The office building energy consumption supervision software database,system and function were designed firstly.Then,the corresponding software program were developed to realize the building energy consumption analysis,checking and prediction in a certain office building,and further realize several functions such as the air conditioning terminal equipment optimization control.The simulation results show that the system can realize the comprehensive management of energy consumption in office buildings and has potential application value in practical projects.Compared with the current office building energy consumption supervision system,this paper solves the problems of unclear energy consumption index,lack of energy efficiency analysis function,and errors in energy consumption data of office building,and emphatically discussed the dynamic prediction of office building energy consumption.This study provides a scientific basis for the efficient supervision of energy consumption in office buildings,and its results can be applied and popularized in practical projects,which is of great significance for the supervision of energy consumption and energy conservation and consumption reduction in office buildings of China.