| Zugriffsnummer | 16133 |
| Dokumenttyp | Konferenzartikel |
| Sprache | Englisch |
| Titel | Thermal transport properties of layered composites by transient measurement : identification by a new numerical algorithm |
| Autor(in); Institution |
Model, Regine; 8.12, Biomedizinische Optik und NMR-Messtechnik, PTB-Berlin
|
| Quelle/Jahr | Fifteenth Symposium on Thermophysical Properties, Boulder, CO, USA, June 22 - 27, 2003 [preprint volume]:(2003) |
| Availability | [CD-ROM] ; file name: 356.pdf |
| Herausgeber(in) |
Committee on Thermophysical Properties
|
| Verlag | Boulder, Colo.: NIST |
| Konferenzangaben | Fifteenth symposium on thermophysical properties, Boulder, CO, 22-27, June, 2003, USA |
| Freie Schlagworte | thermal transport properties ; layered composites ; inverse problem ; analytic solution |
| Zusammenfassung | Transient methods are widely used to determine the thermal transport properties of dielectrics and other materials. In some situations they can be used in case of homogeneous media to measure several properties either simultaneously or separately. In this context analytical model are available and a well-posed inverse problem of parameter identification has to be solved. The examination of composite media is more complicated and only a few results are known. The algorithm proposed here allows to simultaneously determine the thermal conductivity and thermal diffusivity of layered dielectrics by transient measurements. It is based on a plane source which acts both as a resistive heater and temperature sensor. For the technique to be successful two essential aspects have to be considered: firstly, the mathematical modeling of the measured data (the forward problem) and secondly, the problem of ill-posedness of the inverse problem. For the proposed measurement configuration, a new fast data analysis algorithm based on an analytic solution for the forward problem is presented. In principle, a numerical solution such as FEM solution of the heat conduction equation can be used instead of the analytical one, but the computational effort is much greater. The inverse problem is formulated as an output-least squares problem, which leads to a transcendent algebraic system of equations. The method was successfully tested for different situations. |
