The Idea of the Measurement System for Quick Test of Thermal Parameters of Heat-Insulating Materials

Abstract

For the sake of climate and atmosphere conservation the emission of gasses must be bounded. A significant reduction of emission can be obtained by rational heat energy consumption, which is a substantial percent of the world’s consumed energy. One of the ways is using in the building engineering and industry suitable insulating materials: foamed polystyrene, mineral wool, glass fiber, polyurethane foam, synthetic clothes, foam glass or cellular concrete. The existing methods of determination of material’s thermal parameters are based mainly on stationary heat transfer conditions (Bayazitoglu & Ozisik, 1988; Bejan, 1993; Janna, 2000; Minkina & Chudzik 2004; Platunov, 1986). These methods allow determining in the experiment only a single thermophysical parameter of the tested material. They require the use of big and heavy measuring systems and a long period of time to conduct the measurement. Author do not know a commercial solution of portable measuring system which in relatively short time could assess fulfilling the requirements of insulating materials delivered to building site or leaving the factory from the point of view of thermal conductivity. Therefore, it seems to be crucial to work on design of such a measuring system. The research in this field concentrates, among other things, on possibility of application of artificial neural networks to solve the coefficient inverse problem of diffusion process ( Alifanov et al., 1995; Beck, 1985). To determine the usability of network an analysis of its response for known values of thermal parameters is needed. It is necessary to generate input data for network training process using mathematical model of the tested sample of heat insulation material. The discrete model of a nonstationary heat flow process in a sample of material with hot a probe and an auxiliary thermometer based on a two-dimensional heat-conduction model was presented. The minimal acceptable dimensions of the material sample, the probe and the auxiliary thermometer were determined. Furthermore, the presence of the probe handle was considered in the heat transfer model. The next stage of the research is solving the inverse problem in which the thermal parameters will be estimated on the basis of recorded temperatures. Methods employing the classical algorithm of the mean square error minimization in the inverse problem of the heat conduction equation have an advantage of making it possible to take into consideration the arbitrary, varying boundary conditions that occur during the 18