1
$\begingroup$

The generally quoted formula foe resistance is

\begin{equation} R = \rho \ell/A \end{equation}

some special cases are easy to solve. For example the case where the current flowing along the z-axis and $\rho$ or $A$ only depend on $z$ can be easily solved and is commonly mentioned in introductory textbooks.

Usually, the resistance is calculated by dividing the voltage (V) by the current (I) according to Ohm's law (R = V/I). However, consider the case when neither the voltage nor the current are known but the resistivity $\rho(r, z)$ is known.

How to compute the resistance when $\rho$ depends on both $r$ and $z$ i.e. $\rho(r,z)$ is an arbitrary but continuous function?

$\endgroup$
4
  • $\begingroup$ Turn the object into a variety of 'resistors' in series and parallel. $\endgroup$
    – Jon Custer
    Commented Feb 12 at 21:25
  • $\begingroup$ I thought that would have been done and published $\endgroup$
    – wander95
    Commented Feb 12 at 21:27
  • $\begingroup$ FYI, this is not a homework. I have removed that tag. $\endgroup$
    – wander95
    Commented Feb 12 at 21:56
  • $\begingroup$ Note also this tip about working with conductances in parallel, for radial variation in resistivity. But if the resistivity changes strongly with both $r$ and $z$, these approaches don't work because the current is no longer moving in the $z$-direction. Consider reviewing Resistivity of an object with arbitrary shape, which involves the same complicating issue. Summary: A numerical solution is probably required. $\endgroup$ Commented Feb 12 at 22:55

2 Answers 2

1
$\begingroup$

Resistance is a property of a component with linear relationship between current density $\mathbf{J}$ and electric field $\mathbf{E}$. In such cases one can show that there will be linear relationship between the voltage drop ($V$) between two points on that component and the current ($I$) flowing between them.

The general equations that govern this situation are (isotropic case):

$$ \begin{align} \mathbf{J}=\sigma\mathbf{E}&=-\sigma\boldsymbol{\nabla}\phi \\ \nabla^2\phi &= 0\\ \end{align} $$

Where $\phi$ is the electrostatic potential. You would solve them and then get relationship between voltage and current. Below is an example of a simple solution. In general solution may not be that simple, but it will give you what you are after - linear coefficient that relates voltage and current.

Simple solution for a cylindrical segment of a conductor

The resistance equation you have quoted is a simple solution of the equation I gave above.

In your case you could arrive at it by choosing a cylinder with axis of symmetry along z-axis, cross-section area $A=\pi R^2$, height $l$ and its centre at the origin. We choose the boundary conditions to be $\phi=0$ at the bottom face of the cyllinder $-\phi=V$ at the top face of the cylinder and $\partial_r\phi=0$ at the sides of the cylinder (no current electric field perpendicular to cyllinder walls).

One then finds, in cyllindrical coordinates with symmetry $$ \nabla^2\phi=\frac{1}{r}\partial_{r}\left(r\partial_r\phi\right)+\partial_{zz}\phi=0 $$

Assuming a more general $\partial_r\phi=0$ which means no current is flowing side-ways, only along the length of the cyllinder, we get the solution $\phi=-\frac{V}{l}\cdot \left(z-l/2\right)$.

Current ($I$) is then the current density, $\mathbf{J}=-\sigma\boldsymbol{\nabla}\phi=\mathbf{\hat{z}}\sigma\cdot V/l$, flowing through the cross-section of the cyllinder, so $I=\pi R^2 \mathbf{\hat{z}}.\mathbf{J}=A\sigma \cdot V/l$. Finally, introducing $\rho=1/\sigma$:

$$ \frac{V}{I}=l\rho/A $$

$\endgroup$
1
$\begingroup$

Since $\sigma(=1/\rho)$ is not constant in the cylinder, that is, \begin{equation} \nabla\cdot\mathbf{J}=-\nabla\cdot\left(\sigma(r,z)\nabla\phi\right) =-\nabla\sigma(r,z)\cdot\nabla\phi-\sigma(r,z)\Delta\phi=0 \end{equation} is the equation to solve. This equaition differs from the pure laplace equation. Numerical solver, for example finite element method, is preferable.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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