The application of the artificial neural network and hot probe method in thermal parameters determination of heat insulation materials Part 2 - application of the neural network

Abstract

In part 1 of the paper, the discrete model of a nonstationary heat flow process in the sample of material with a hot probe and an auxiliary thermometer based on a two-dimensional heat-conduction model was presented. To create the model the finite element method (FEM) implemented in the Matlab environment was used. The part two of the paper is concentrated on possibility of using a neural network for the thermal parameters determination. The artificial neural network (ANN) is used to estimate the coefficients of the inverse heat conduction problem for solid. The network determines the value of the effective thermal conductivity and the effective thermal diffusivity on the basis of temperature responses of the hot probe and the auxiliary thermometer. During selection of optimal ANN architecture several configurations were evaluated. The influence of measurands uncertainty on identified values of the thermal parameters was also analyzed. Training process and simulation analysis were conducted in the Matlab environment.