Regression Analysis Using Neutral Networks for Nondestructive Control of the Thermal Characteristics of Polymers

Abstract

An intelligent information-measuring system for controlling the thermal characteristics of materials is a relevant topic related to the issue of finding the middle of a thermogram working section using a neural network as an advanced and accurate way of processing experimental results. The object of study is an information-measuring system for nondestructive testing of structural transitions in polymers. The aim of this study was to derive a new regression equation for the middle of the working section of a thermogram, depending on the thermal activity of a material and the specific thermal power of a flat heater, using a neural network. The obtained experimental dependences of the temperatures in the middle of the working section of a thermograms can be used by technologists who deal with the development of new polymers and the use of existing ones. New results have been obtained using a neural network and reliably described by the derived regression equation. Using the results, the regression equation has been refined to determine the middle of the working section in the method for nondestructive testing of structural transitions in polymers. Using the regression equation, one can predict the temperatures in the middle of a thermogram working section, depending on the power of the heater and the coefficient of thermal activity of the material under study.