Load Forecasting Algorithm Of Building Air Conditioning Systems

Building energy conservation in the speedy development of city construction today is increasingly becoming one of the topics of social concern.Air conditioning load is one of the main sources of building energy consumption,and its energy saving effect determines the effect of energy saving in the whole building.The load forecasting is to estimate the macroscopic load,is based on the principle of similarity,approximation in the same demand characteristics under load,which is based on the historical data to estimate the load at a certain moment in the future the same or similar conditions,so as to adjust the equipment input,system optimization control,to achieve the purpose of saving energy.In view of this,the accurate air conditioning load is the key to the energy saving operation of the air conditioning system.In the design phase,the energy consumption simulation software is used to analyze the building air conditioning load forecasting,so as to rationally design the air conditioning system and select the air conditioning equipment,and the hourly load forecasting is needed in the operation stage,and the precision of the prediction is higher.This paper studies the load forecasting algorithm for the air conditioning system of the Wangfujing shopping mall in Leshan,Sichuan,and collects the historical data of the air conditioning system load,and pretreats the data collected on the spot.On the basis of that,the principle and structure of the SVR(support vector regression)and BP(back propagation)network are studied,such as the single prediction algorithm,the fusion prediction algorithm and the combined prediction algorithm,and the factors affecting the air conditioning load.The structure differences of the support vector regression machine and the BP(back propagation)network are studied.The load forecasting model based on neural network and support vector machine is built,and the parameters of the model are optimized with the genetic algorithm GA(genetic algorithm).The hourly prediction of the air conditioning load is realized.The construction of the load forecasting model is realized in the MATLAB and the prediction accuracy of several forecasting methods is also made.The calculation and analysis of the running time are carried out.The results show that the support vector machine(SVM)algorithm based on the genetic algorithm has better prediction accuracy and better prediction accuracy under the condition of less sample size.