Research On Intelligent Modeling And Thermal Insulation Performance Prediction Of Unsteady Heat Transfer Process Of Vacuum Glass

New energy-saving materials-vacuum glass once come out,it has been the world's attention,especially in the field of thermal insulation,it is more excellent to the same type of building materials.At present,the vacuum glass manufacturing technology has made a major breakthrough,began a large number of industrial productions,and had a lot of economic benefits.But its important thermo physical parameters-heat transfer coefficient,how to measure it online is very difficult,and with the accumulation of vacuum glass using time,the heat transfer coefficient is also increasing,resulting in decreased thermal insulation properties.Therefore,how to quickly and accurately measure the heat transfer coefficient of the vacuum glass in use becomes a difficult subject.The heat transfer process of the vacuum glass is very complicated.Especially the heat transfer process of the Invalid vacuum glass is the coupling of heat conduction,convection and radiation.It is impossible to find the exact solution in mathematics.Based on advanced numerical simulation technology,this paper uses fluent software to research the heat transfer process in vacuum glass.Using the simulation results to guide the unsteady test method and analyze the unsteady test method,it is concluded that the radius of the heating plate and the thickness of the vacuum glass should be greater than 3 when heating the vacuum glass side with a round heating plate in order to ensure that the center of the heating plate for one-dimensional heat transfer,non-heating surface center temperature measurement has meaning.Temperature is an important measurement parameter.In this paper,MSP430 series of microcontrollers as the core,the use of DS18B20 temperature sensor for the formation of temperature measurement equipment,and the factors that affect the accuracy of heat transfer coefficient measurement device is analyzed,and the following conclusions:1,After the heating plate work for 10 minutes,then measure the temperature;2,the contact surface of the temperature sensor and the vacuum glass has the contact resistance,by its influence,the thermometer temperature rise has a certain lag,which will cause the calculated heat transfer coefficient low compared with the real value.Based on MATLAB software,a neural network intelligence model was established and the traditional BP neural network was optimized.First the genetic algorithm was used to reduce the dimension of the independent variable,and then the Mind Evolutionary Computation algorithm was used to optimize the initial weight and threshold.Use the optimized BP neural network intelligence model to predict the heat transfer coefficient of the vacuum glass insulation performance,the average absolute error = 0.076,with high accuracy...