Heating Analysis and Prediction Based on GOOSE-BP Model Mathematical Modeling and Neural Networks

Abstract

: The wide application of air source technology has imposed considerable pressure on the power grid system. By accurately predicting indoor temperatures and reducing unnecessary energy consumption, this will help alleviate the burden on the power system and bring about certain social and economic benefits. This article first provides the relevant background information and collects relevant data from two locations. Different identification parameters are set and combined with the thermodynamic differential equation model to solve the parameters for different locations. Secondly, GOOSE-BP and BP neural networks are used to predict different locations respectively, and their effects are also evaluated. Finally, based on the above two results, a constant temperature control strategy is established, and a dynamic programming model is established based on the electricity prices during peak and off-peak periods and constraints to calculate the optimized electricity costs for different locations. This article proposes different strategies for the two different locations to reduce the electricity burden of enterprises. Meanwhile, a feasible neural network model is proposed to optimize power distribution and reduce the burden on the power grid.