I'm having trouble understanding how to map the streamlines from one plane to another using the Joukowski transform.

In the $\zeta$ plane, I'm considering flow around a cylinder, with the complex potential given by $$w(\zeta)=U(\zeta e^{-i \alpha} + \frac {a^2}{\zeta}e^{i\alpha})$$

where $\zeta = re^{it}$.

From here I understand it is more useful to consider the streamlines first in the $\zeta$ plane and then use the Joukowski map $z=\zeta+\frac{b^2}{\zeta}$ to transform them to the $z$ plane. It's not a problem finding the streamlines in the $\zeta$ plane, as I can use the stream function given by $$\psi(r,t)=Usin(t-\alpha)(r-\frac{a^2}{r})$$ and equate it to constant values. From here though, the notes I'm looking at say it's easy to then use the transform to get the streamlines in the $z$ plane, but I'm not very sure how. I'm aware it's supposed to look like flow around an ellipse, but my ideas on Maple all draw a blank.



I would find the parametric equation for the streamlines in the $\zeta$ plane; then the transformation to the $z$-plane is indeed easy. As a parameter, take $s=t-\alpha$. The streamline equation is $U\sin s (r-a^2/r) = C$. Solving this for $r$ yields $$r=r(s)=\frac12\left(\frac{C}{U\sin s}+\sqrt{\left(\frac{C}{U\sin s}\right)^2+4a^2}\right)$$ So, the parametric equation of streamline is $\zeta=r(s)e^{i(s+\alpha)}$. Applying the Жуковский transformation (I hate that transliteration) yields $$z= r(s)e^{i(s+\alpha)} +\frac{b^2}{r(s)} e^{-i(s+\alpha)}$$ which is something that Maple's complexplot can handle.

  • $\begingroup$ I understand the parameterisation, but why is the parametric equation $r(s)$ multiplied by $e^{i(s+\alpha)}$? I think I'm missing something simple there. Thank you for the answer, it's no doubt clear to most people! $\endgroup$ – Mike Miller Nov 27 '13 at 2:00
  • $\begingroup$ @BritMiller I understand that $\zeta=re^{it}$ in your notation. After introducing $s=t-\alpha$, this relation becomes $\zeta=re^{i(s+\alpha)}$. This gets plugged into the formula for $z$... By the way, I should have said that the range of $s$ can be taken $-\pi<s<\pi$. $\endgroup$ – Post No Bulls Nov 27 '13 at 2:14
  • $\begingroup$ Ah, of course! Yes I can see why it's a lot easier doing it this way now. So I should be able to plot the streamlines as a function of $s$; something like complexplot(z(s),s=-Pi..Pi), for certain values of C? $\endgroup$ – Mike Miller Nov 27 '13 at 2:37
  • 1
    $\begingroup$ @BritMiller Yes. Perhaps with a slightly smaller interval: going to $\pm$Pi amounts to going to infinity in both directions. Or maybe Maple will cut it automatically; I don't have. $\endgroup$ – Post No Bulls Nov 27 '13 at 3:17
  • $\begingroup$ Hiya. I've given this a go in Maple. The graphic I get after doing the transform is not what I expected - the streamlines cross (which I don't think should happen!) Though the basic shape I expected is there. Is there something simple I am missing do you think? I'd show a graphic if I could, but not sure how in the comments. $\endgroup$ – Mike Miller Nov 27 '13 at 16:35

There are two additional issues here. First, your choice of $a=2$ and $b=3$ on the plot. The transformation $z=\zeta+b^2/\zeta$ is not conformal at $\zeta=\pm b$, where the derivative vanishes. This point should not be in the domain where you consider the flow; otherwise you don't have a conformal transformation (it is not even 1-1). The domain is the exterior of the disk of radius $a$. So, the allowable parameters are $0<b\le a$. For example, $a=3$ and $b=2$.

Second, the range of $s$ should not be $(-\pi , \pi)$. Indeed, $s$ is introduced as a polar angle (with some offset) on a streamline that stays in one halfplane; therefore its range is an interval of length $\pi$. It will be $(0,\pi)$ if $C>0$ and $(-\pi,0)$ if $C<0$. If this is unclear, consider the polar parametrization of horizontal line $y=C$. When $C>0$, it is $r=C/\sin t$, $0<t<\pi$; when $C<0$, it is $r=C/\sin t$ with $-\pi<t<0$.

With the above corrections, the plot comes out right.


Maple code:

a:=3; b:=2; 
display(complexplot([seq(subs(C=i,z),i=0..20)],s=0.01..Pi-0.01,-15..15,-10.5..10.5,thickness=3), complexplot([seq(subs(C=i ,z),i=-20..-0)],s=-Pi+0.01..-0.01,-15..15,-10.5..10.5),thickness=3);
  • $\begingroup$ Oops, the $b>a$ was a mix up - I did understand that part. I'm not so great with parameterisations, so thank you very much for the example $y=C$; it's perfectly understood now! :) $\endgroup$ – Mike Miller Nov 27 '13 at 18:39

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.