Research On Load Forecasting And Environment Optimization Building Atrium Based On BP Neural Network

Large atrium spaces are a prevalent feature in commercial buildings.However,the hollow structure of atriums can lead to a significant "chimney effect",resulting in severe temperature stratification during summer months.If the air conditioning system fails to adjust supply air parameters,such as airflow rate,temperature,and speed in response to the dynamic atrium load,it can cause deviations from the design temperature distribution,reduce thermal comfort,and waste energy.This thesis addresses the issues of excessive atrium top temperatures and stratification by optimizing the supply air scheme based on dynamic atrium load predictions,ultimately improving thermal comfort and energy efficiency.Focusing on an atrium in a commercial building in Shenyang,this study analyzes atrium load factors using meteorological parameters and historical load data through correlation analysis,subsequently establishing a BP neural network model.Outdoor temperature,humidity,and solar radiation were found to have strong correlations with atrium load and were used as input parameters for the BP neural network.The prediction error of the model is within10%,meeting the requirements for dynamic load prediction.Load condition clustering analysis reveals that nighttime load correlations are low and stable,while daytime correlations are higher,with a strong correlation between 7:00 and 14:00,consistent with partial correlation analysis results.As solar radiation affects both load prediction and Fluent simulations,this study calculates the solar radiation intensity at any horizontal position on the Earth’s surface based on the Earth’s orbit around the sun.By comparing measured and calculated results,the relative error is found to be less than 15% for a typical day,indicating that the solar radiation model is valid and can provide a theoretical basis for atrium thermal environment simulation analysis.Drawing on turbulence theory and the characteristics of fluid flow and heat transfer,this study employs Fluent simulation technology with k-ε control equations to simulate four ventilation methods under summer conditions with thermal pressure ventilation.airflow velocity and temperature fields are obtained under different conditions.The simulation results generally align with actual temperature and velocity distributions in engineering projects.The highest temperature typically occurs at the top of the atrium’s central area,and Large temperature gradient in the vertical direction of the atrium.This confirms the established model’s validity.Fluent is utilized to adjust air conditioning supply air speed,angle,roof exhaust volume,and shading measures to simulate the summer thermal environment in commercial building atriums,Analysis reveals that excessive or insufficient supply air speed,angle,and exhaust volume are detrimental to the indoor thermal environment.By selecting appropriate parameters and implementing internal shading,the vertical temperature gradient in atriums can be reduced.The results show that optimizing atrium environmental control strategies based on dynamic load prediction and numerical simulation can solve thermal comfort issues and achieve energy-efficient operation.