2
$\begingroup$

How does one find the effective thermal conductivity of a cylindrical body made of multiple layers? E.g. Copper core (k1), insulator 1 (k2) and insulator 2 (k3). I need the effective properties for performing lump calculations.

An example can be seen below. I appreciate any kind of help! enter image description here

$\endgroup$
  • $\begingroup$ Which axis is the cylinder conducting through: radial, or axial? $\endgroup$ – CR Drost Jul 10 '15 at 16:42
  • $\begingroup$ It would be nice for have the formula for both. $\endgroup$ – Physther Jul 10 '15 at 17:17
3
$\begingroup$

Axial conduction, along the cylinder

Let's first talk about electricity. We know that effective resistance of a parallel portion of a circuit found by the relation: $$\frac{1}{R} = \sum \frac{1}{R_i}$$If you just figure out the relation between conductance and resistance, $$G = \frac{1}{R}$$So $G = \sum G_i $. In fact, conductance in parallel is additive, just add each one up and you'll get the total conductance.

Now about electrical conductivity. $$G = \sigma \frac{A}{l}$$ If you have several conductor materials of the same length, then the total conductance would be $$G = \sum G_i = \sum \sigma_i\frac{A_i}{l} = \frac{1}{l} \sum \sigma_iA_i$$ For the effective "total" conductivity, $$ \sigma = G \frac{l}{A} = \frac{l}{A} \cdot\frac{1}{l}\sum \sigma_iA_i = \frac{\sum \sigma_iA_i}{A}$$ Since we know thermal conduction and electrical conduction is analogous, this yields: $$ k = \frac{\sum k_iA_i}{A_{total}}$$ or in words:
The effective conductivity of a multi-layer composite material is the weighted mean of each component layer's conductivity, where the weight is the cross-sectional area of each layer.

This formula should work for any composite material, be it radially, vertically or block-layered, as long as each cross-section is identically structured. As a demonstration with your question's examples, the effective conductivity would be: $$ k = \frac{k_1A_1 + k_2A_2 + k_3A_3}{A_{total}}$$

Note: there is a possibility that the formula even work if each cross-section is not identically structured, as long as the cross-sectional area of each component remains the same, though I didn't bother to proof that.


Radial conduction

This one would involve calculus. Starting from the resistance of each layer, $$R_i = \int^{r_i}_{r_{i-1}} \frac{1}{k_iA}dr = \int^{r_i}_{r_{i-1}} \frac{1}{2k_i\pi rl}dr = [\frac{\ln r}{2k_i\pi l}]^{r_i}_{r_{i-1}} = \frac{\ln \frac{r_i}{r_{i-1}}}{2k_i\pi l}$$ Evidently this formula doesn't allow a point source of heat, so you'll need to assume that temperature is homogeneous in any cross-section of the copper component, and then find the resistance of the outer two resistor layers.

$\endgroup$
0
$\begingroup$

In order to solve this, we must first find the total thermal resistance ($R_{total}$). The thermal resistors (copper, insulator 1, and insulator 2) are acting in parallel. Thus the total resistance is give by the following equation:

${\frac{1}{R_{total}}}={\frac{1}{R_1}}+{\frac{1}{R_2}}+{\frac{1}{R_3}}+...+{\frac{1}{R_n}}$ where ${R_n}={\frac{L}{k_nA_n}}$ (this is very similar to parallel resistors in an electric circuit)

${R_n}$ is the thermal resistance of each layer, $L$ is the legnth of the cylinder, ${k_n}$ is the thermal conductivity of each layer, and ${A_n}$ is the cross section area of each layer. When we plug in $k_1,k_2,k_3,A_1,etc.$ we find that ${R_{total}}=\frac{L}{k_1A_1+k_2A_2+k_3A_3+...+k_nA_n}$

If we want to find the effective thermal conductivity (${k_{eff}}$), we know that $R_{total}=\frac{L}{k_{eff}A_{total}}$

Thus $k_{eff} = \frac{L}{R_{total}A_{total}}=\frac{L(k_1A_1+k_2A_2+...+k_nA_n)}{L(A_1+A_2+...+A_n)}=\frac{(k_1A_1+k_2A_2+...+k_nA_n)}{(A_1+A_2+...+A_n)}$

Edit: Made a math error

$\endgroup$
  • $\begingroup$ Thanks for the answer. Is L the length of the circle (2piR) or the thickness of layer? Also, shouldn't it have a ln function because of cylindrical coordinates? Or cylindrical coordinates are not necessary? $\endgroup$ – Physther Jul 10 '15 at 17:09
  • $\begingroup$ L is the length of the cylinder (not the circumference). There would be a ln(R_out/R_in) term if the heat conduction was from the center of the cylinder outward. The answer I provided applies for heat conduction along the length of the cylinder (in cylindrical coordinates this would be along the Z axis). $\endgroup$ – Joe Jul 11 '15 at 0:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.