Research Combination Forecasting Based On BP Neural Network Inthe Energy Syetem

Each enterprise is very concerned about enterprise energy problem. But it isdifficult to make decision of the enterprise energy demand. How to forecast theenterprise energy consumption is related to the development of the enterprise. Thispaper uses a new forecasting method to forecast the enterprise energy consumption, inorder to make the enterprise energy managers master the enterprise energy trend, todirect the enterprise production, scheduling and balance, and to ensure the balancebetween the enterprise energy supply and demand. So the research contents of thispaper have some theory and application value.Combined with the characteristics of GM(1,1), pGM(1,1),Grey-Markov chainand BP neural network,this paper suggests a combination forecasting method based onBP neural network and is to the enterprise energy consumption of the XX steelenterprise model and analyze. The main research contents are as follows:1．Set up the single forecasting model of the enterprise energy managementsystem of energy consumption. Make use the energy consumption data of the XX steelenterprise to model and predict inspection, and analyses the characteristics andforecasting performance of the single forecasting model in order to lay the foundationof the combination forecasting.2. Construct two combination forecasting models based on BP neural network.The first combination forecasting mode GNN1:the predict results of the singleforecasting model is as the input of the BP neural network, then use the BP neuralnetwork combination; the second combination forecasting model GNN2: Form threesingle forecasting models select the best predict result of single forecasting model—grey-markov chain forecasting model is as one of the input of the BP neural network, at the same time the factors which have a major impact on the output is also as theinput of the BP neural network, then use the BP neural network combination.3. Set up the based on BP neural network combination forecasting mode of theenterprise energy management system of energy consumption. Make use the energyconsumption data of the XX steel enterprise to model and predict inspection. Theresults show that the based on BP neural network combination forecasting modelimproves the precision, even the t precision of the combination forecasting modelGNN2is more high and stability.