0
$\begingroup$

I'm attempting to write a FORTRAN program that calculates the magnetic field, B, at any point outside of a bar magnet.

I'm going to use a first order euler scheme, where each side of the bar magnet is split into small cells, each with centres at (xi,yi,zi). I know I can ignore all of the sides that have any z values, and just focus on the top and bottom sides that are orientated in the x-y plane. So the method says this:

$\int f(x,y,z)dS = \Delta S \cdot \sum f(x_{i},y_{i}, z_{i})$

where the integral is over the surface S, and the summation is over i.

Delta S, each area, is given by $\hat{n}\cdot d\vec{S}$ , so if the cell is oriented in the x-y plane it's just $\Delta x\cdot \Delta y$ .

Here is a screenshot of the specific method instructions with a figure that demonstrates it

The function for the magnetic field is this: $\vec{B}(\vec{r}) = \frac{\mu _{0}}{4\pi }\cdot \int\frac{(r-{r}')\cdot M(r)\cdot \hat{n}}{|(r-{r}')|^{3}}$

Where the integral is over the surface S

I'm struggling to understand this method. I've tried to construct a flow chart, but can't get very far so I figured the problem is with the mathematics. Any help to understand it would be appreciated, and also any help with the flow chart would be fantastic. Here is my flowchart

$\endgroup$

migrated from physics.stackexchange.com Feb 12 '18 at 6:28

This question came from our site for active researchers, academics and students of physics.

  • $\begingroup$ Cross-posting is discouraged on the StackExchange network, so people don't waste their time with an answer you already received on the other site. The usual procedure is to wait a few days, and then either raise a flag and ask the moderators for migration (if there are some answers already) or delete the old and ask a new question. Croos-post math.stackexchange $\endgroup$ – Mauro Vanzetto Feb 12 '18 at 14:27
1
$\begingroup$

Let us simplify the problem to understand better what is going on.

Let us begin with the following 1D integral: $$\int_{x_0}^{x_N}{f(x)\,dx}$$ We know that since the integral is a sum, the sum of the cut integral is equal to the previous one, i.e. $$\int_{x_0}^{x_N}{f(x)\,dx}=\int_{x_1}^{x'}{f(x)\,dx}+\int_{x'}^{x_N}{f(x)\,dx}$$ Where $x'$ is a point such as $x'\in[x_0,x_N]$

Imagine, we cut the integral in $N$ intervals... if this number is big enough we are "seeing" that the function $f(x)$ along a cut is almost a straight line (e.g the Earth is a sphere but from our point of view is a plane). If we do further assumptions, we can say even that the function $f$ in this tiny interval has an almost constant value. Let us name the corresponding value of $f(x)$ in the $i$-th cut $f_i$, thus the cut integral is simplified to: $$\int_{x_0}^{x_N}{f(x)\,dx}=\sum_{i=0}^{N-1}\int_{x_i}^{x_{i+1}}{f(x)\,dx}\approx\sum_{i=0}^{N-1}\int_{x_i}^{x_{i+1}}{f_i\,dx}=\sum_{i=0}^{N-1}f_i\int_{x_i}^{x_{i+1}}{\,dx}=\sum_{i=0}^{N-1}{f_i\Delta x_i}$$ The last step is due the fact that $f_i$ is a constant.

There are some type of functions that decay really quickly. Imagine that $f(x)=\frac{1}{x^2}$ (like in your problem), the integrand will be then: $$\int_{x_0}^{x_N}{\frac{dx}{x^2}}$$ The contributions of the integral are negligible once $x$ is large enough, i.e. the error in the integration is proportional to $1/x$. Therefore you can stop once $x$ is large enough, i.e. it is not necessary to sum the $N$ intervals you have.

$\endgroup$

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

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