Method for predicting heat transfer property of building exterior wall at severe-cold region based on neural network

Abstract

The invention discloses a method for predicting heat transfer property of a building exterior wall at a severe-cold region based on neural network. The method aims at utilizing a neural network model to analyze influence of outdoor air-flow speed in a severe-cold region in winter on heat transfer property of a building exterior wall based on actually-measured data. The method comprises: measuring a multi-story building peripheral outdoor wind speed and wind direction, indoor and outdoor average temperature and humidity, and density value of a heat flow passing through the exterior wall at corresponding moment on site in a coldest month in Haerbin; taking environment parameters and the heat flow density of the building exterior wall as input / output parameters to establish a neutral network prediction model, and using actually-measured data groups to train the network model; and employing a control variate method and utilizing the neutral network prediction model to generate a data curve of the heat flow density passing through the building exterior wall along with the variation of outdoor air-flow speed.

Description

technical field [0001] The invention relates to a method for testing the heat transfer density of building exterior walls. It specifically relates to a method for testing the heat transfer density of building exterior walls under the influence of different outdoor airflow velocities in winter in severe cold regions based on measured data. Background technique [0002] The heating energy consumption of buildings in severe cold areas accounts for 50%-60% of the total energy consumption of buildings, and most of them come from the heat transfer loss of building exterior walls. Heat loss through building exterior walls is mainly carried out in three ways: convective heat transfer, solid heat transfer and radiation heat transfer. The outdoor air velocity significantly affects the convective heat transfer process. Existing data show that the change of convective heat transfer caused by wind disturbance accounts for 20% of the total heat transfer change, and there is uncertainty ...

AI Technical Summary

Problems solved by technology

However, wind tunnel experiments mostly use scale miniature models, and there are deviations between the experimental boundary conditions and the actual environment under natural conditions, which affects the research accuracy.

Images