In principle this is simple (all you need is a heat source and heat sink and two temperature sensors). In practice, however, it is very difficult to obtain reliable results due to the difficulty in controlling the heat flow paths and heat losses. There are standards describing measurement methods (both steady-state and transient), however, in order to obtain acceptable accuracies (typically 2-5%) the requirements of the equipment setup and calibration becomes quite involved. If your intention is to measure conductivities (and not develop an instrument for academic or commercial use), I think it would be best to use an existing (certified) setup; you may risk spending an awful lot of time and frustrations in developing your own apparatus.

(As a hint: If you want to develop your own rig, perform thorough CFD/FEM-calculations of the setup in order to ascertain that the heat flow paths and heat losses are under control. You may discover that this is extremly difficult to achieve.)

Dear, Erling

Thanks for the speedy reply. I understand that its a tough initiative to construct ones own rig but I was looking for a rough setup. Inaccuracies of heat losses are acceptable given that its an academic project. I just needed to know a simple step by step method of making your own setup. If you could provide me some internet resource or link or anything I ld appreciate it. 

regards

Please refer the following site.

http://www.ccfltd.co.uk/Trade-Support/Calculating_U_Values 

Hello Sami

if you need a very, very, very rough estimate, then go to fundamental Fourier law for conduction heat flux. Perfectly insulate all but two opposite side walls and apply heat source (heating electric foil with metered power), well insulated from the ambience and in the same time measure temperature under the foil, i.e. with TC embedded just below the surface of your heated side. Of course you can expect a quite non uniform temp. distribution across the wall, but depends on how many TC you use... And measure temperature on the opposite side. Then, very roughly you get thermal conductivity by applying the Fourier law. To avoid radiation losses from the "cold" side, cover the wall with a low emissivity very thin Aluminium or copper foil, or just calculate radiation heat transfer (you find in any heat transfer texbook). And to eliminate convective heat transfer from the cold side, I would place all the equipment in a box, maybe with transparent walls to see what´s happening inside. But this is something you should play with.

regards

You can use the Thermal conductivity analyzer supplied by KD2 Pro for the purpose.  You can measure the thermal conductivity of solids and liquid samples as well as the specific heat capacity using this instrument with standard transient hot wire technique.

Thanks Miroslav!! I am beginning to get the idea of it. sounds like some work but fun too.... great advice. (Y)

Yes we can measure your samples and analyse heat transfer mechanisms in any temperature range

[email protected] 

If you ca bore a hole up to the center of your sample, you may also introduce in it a thermocouple and a small resistor and heat it. A thermistor may do the two jobs, since its resistance depends on its temperature. All the power given to this resistor will go out through your sample and the electric wires. The relation between this power and teh temperature difference between the center and the surfaces of your cube can be found by soving the Posson equation. Comparing the power and temperature difference you will get at seady state with this relation will give you the thermal conductivity.

There are also dynamic instruments, using a hot wire or a hot disk posed on nthe suface of the sample (as mentionned by Kinari Pradekh, but these require a very good tehrmla contact between the wire or disk and the sample, hence a flat surface.

Build an highly-insultated frame (to emulate one-dimensional heat flow), generate to different microclimates on both sides, and measure one surface temperature.  Following fourier-law (or if you want it easy, just check out ISO 6946) you can easily calculate a rough estimate of an u-value. However... i ask myself, if the material you are talking about is not easier put to the classical labs with thermochambers (they can guarantee a smaller deviation frame around your measurements)...

Sorry, I forgot to mention two standards related to measurements using local probes adapted to small samples. These are ISO 8894 (hot wire) and ISO CD 22007-2 (hot disc). I attach the drafts of these.