1
$\begingroup$

Given

$$ I[u(x,y)] = \frac12\int\int_G\left| \nabla u \right|^2$$

where G is defined by $x^2+y^2\lt R^2 $

Find u(x,y) that will provide the lowest possible value of the functional. Also given:

$$u(r=R,\theta) = ax^2 $$

So I used E-L relation, and found that I end up with Laplace's equation.

Using separation of variables, I get

$u(r,\theta) = R(r)\Theta (\theta): $

$$\frac{r}{R(r)}\frac{\partial}{\partial r}\left(r R(r)' \right) = -\frac{\Theta''}{\Theta} $$

I for $\Theta$ and got:

$$\Theta_n =\alpha_ncos(n\theta) + \beta sin(n\theta) $$

And for R i get two solutions depending on n:

If n = 0:

$$R_0 = Aln(r) + B$$

A has to be zero so the function doesn't diverge at r=0

When $n \ne 0$:

$$R_n = Cr^n + Dr^{-n} $$

Where d has to be zero, again so the function doesn't diverge at n=0.

I get the function:

$$u(r,\theta) = B_0 + \sum \left(\alpha_n cos(n\theta) + \beta_n sin(n\theta)\right)r^n$$

Using our BV, changing it to polar coordinates:

$$ u(r=R, \theta)= ar^2\left(\frac12 - \frac{cos2\theta}{2}\right) $$

$$u(r,\theta) = B_0 + \sum \left(\alpha_n cos(n\theta) + \beta_n sin(n\theta)\right)r^n = ar^2\left(\frac12 - \frac{cos2\theta}{2}\right) $$

Here we see that we only have terms when n=2:

$$ B_0 = \frac{aR^2}{2}$$

$$\beta_n = 0 $$

$$ \alpha_n = -a\frac{1}{2} $$

Therefore we get:

$$\frac{aR^2}{2} -a\frac{1}{2}r^2cos2\theta $$

Switching back to cartesian I get:

$$\frac{aR^2}{2} + \frac{a}{2}x^2 -\frac{a}{2}y^2$$

According to solutions, Im missing a term $by$.

I'd appreciate any type of help

  • 1
    The method and reasoning here looks fine, however you have a small typo: $\cos^2\theta = \frac{1\color{red}{+}\cos(2\theta)}{2}$ not $\color{red}{-}$. The last expression is the same as I get. What do you mean by a $by$-term? What is $b$? If $r = R$ then the last expression is correct on the boundary as written and adding an additional term would mean that the boundary condition is not satisfied.2017-02-09
  • 0
    It seems there's something wrong with the solution. Thanks for confirming that my answer is correct! I've spent way too long trying to get the solution given2017-02-09
  • 1
    Well we might be missing the same thing, however I can't see any mistakes here. The simplest way to convince yourself that the book answer is wrong is to evaluate it for $x^2+y^2 = R^2$ and compare to the given boundary condition.2017-02-09

0 Answers 0