The Research On Building Energy Management System Based On The Internet Of Things

With the development of various fields of society and the improvement of people's quality of life,the use of energy has increased,resulting in a large amount of energy consumption,which is especially reflected in large buildings,such as commercial buildings and campus buildings.In order to carry out the idea of green campus construction in the 13 th Five-Year Plan of China,on the basis of summarizing the previous research work,the requirement analysis and research of building energy management system are carried out.Based on the current Internet of Things technology,a building energy management system with data query and energy consumption trend analysis is designed and developed,and a prediction model is established by using the actual data of a university in Guangdong Province to predict energy consumption.In order to carry better energy conservation planning.This study takes the energy consumption of office buildings in a university in Guangdong Province as the research object,designs the energy consumption data acquisition module with the CC2530 chip as the core,realizes the function of the ammeter data acquisition,and verifies the reliability of the data acquisition module through the experiment.Based on the historical energy consumption data of a university in Guangdong Province,the linear regression BP neural network prediction model and the exponential smooth BP neural network prediction model are established to predict the daily energy consumption and the monthly statistical energy consumption respectively.The basic principles are as follows:first,the linear regression method and exponential smoothing method are used to carry out the first prediction,and we use the BP neural network which have been trained to the secondary prediction.The simulation results show that it is reasonable to take the value of the influence factor between 0 and 1 in linear regression prediction,and the prediction result is closer to the actual value when the value of the influence factor is smaller.And in the exponential smoothing prediction,it is reasonable to take the value of the weighting factor between 0.1 and 0.5 according to the time series law of energy consumption data,and the prediction result is closer to the actual value when the value of the weighting factor is smaller.The prediction accuracy can be improved by using BP neural network to carry out the secondary prediction,which reduce the error between the prediction result and actual data by1.56%~2.1%.Compared with the first prediction result,the secondary prediction result is closer to the actual value,and the energy conservation planning can be better carried out,so asto achieve the purpose of energy saving.Finally,the software platform of building energy management system is designed and developed with WAMP as the development environment and ThinkPHP as the development framework.The platform combines multiple functions of system login,energy consumption data query,history energy consumption query,energy consumption trend analysis,energy-saving report,etc.into a whole.It also setups secure and extensible functions.The main characteristic and innovation of this paper is that the software platform of building energy management system is developed through PHP,which makes use of its advantages of free open source and less system resources when running and provides the platform support for the energy management of campus buildings.In view of the time series characteristics of daily energy consumption and monthly energy consumption,linear regression prediction and exponential smoothing prediction combined with BP neural network are applied to the prediction of energy consumption data.By setting reasonable parameters to improve the accuracy of data prediction,energy consumption is effectively reduced.