Math Methods · Learn
Some problems are only hard because of the shape you're trying to solve them on. Find a transformation to a friendlier shape, solve the easy version, and map the answer back — angles, and the physics built on them, survive the trip unchanged.
A complex function $w=f(z)$ is conformal at a point if it preserves angles there — two curves crossing at some angle in the $z$-plane cross at that same angle in the $w$-plane, even though lengths and areas can stretch wildly. Any function that's complex-differentiable with nonzero derivative is conformal, because locally it acts like multiplication by the complex number $f'(z)$ — a rotation and a uniform scaling, and rotations don't change angles between curves.
This matters for fluid flow because the equations governing 2D potential flow — Laplace's equation for the velocity potential — only care about angles and local structure, not which particular shape the boundary happens to be. So a conformal map carries a valid flow solution to another valid flow solution, just around a different-looking boundary.
Pick a circle in the $z$-plane, centered at $z_0=-\varepsilon+i\delta$ (a small offset controls thickness and camber), with radius $R$ chosen so the circle passes through exactly one special point, $z=b$. Map every point on that circle through $w=z+b^2/z$, and the image is a closed curve shaped unmistakably like an airfoil — rounded on one end, pointed on the other.
The derivative is $dw/dz = 1-b^2/z^2$, which vanishes at $z=\pm b$. A conformal map fails to preserve angles only where its derivative is zero — there, angles get doubled. The circle is built to pass through $z=b$, so that point becomes a sharp corner (the trailing edge); $z=-b$ is placed safely inside the circle, where its angle-doubling never touches the boundary. One singularity exposed, one hidden — that asymmetry is the entire reason the shape has a rounded nose and a sharp tail instead of two of either.
Flow around a plain circle of radius $R$, centered at $z_0$, with free-stream speed $U_\infty$ at angle of attack $\alpha$, plus circulation $\Gamma$, has a textbook closed-form complex potential. Writing $\zeta=z-z_0$:
The complex velocity is $dF/d\zeta = u - iv$. Everything here — uniform flow, the image-doublet term that keeps the circle a streamline, the circulation term — is standard, easy, and was solved long before airfoils were the point.
Without circulation, the flow has two stagnation points symmetric about the free-stream direction — and mapped through the Joukowski transform, one of them sits right at the airfoil's sharp trailing edge, where real flow can't physically turn that fast. The Kutta condition says: nature picks exactly the circulation that moves the rear stagnation point to coincide with the trailing edge, so the flow leaves smoothly instead of wrapping around a corner.
Setting $dF/d\zeta=0$ at the one point on the circle, $\zeta_{TE}=Re^{i\theta_{TE}}$, that maps to the trailing edge gives a clean closed form:
where $\theta_{TE}$ is just the angle of the point $b-z_0$ — known the moment the circle's geometry is fixed.
Verified before shipping: plugging this $\Gamma$ back into $dF/d\zeta$ at $\zeta_{TE}$ gives a value within machine precision of exactly zero, for every angle of attack — the Kutta condition isn't assumed, it's checked. Open the lab →
The Kutta-Joukowski theorem gives lift per unit span directly from the circulation: $L'=\rho U_\infty\Gamma$. In the thin, flat-plate limit ($\varepsilon,\delta\to0$), the circle degenerates to radius $b$ centered at the origin, the chord becomes exactly $4b$, and the lift coefficient simplifies to:
which is exactly the classical thin-airfoil-theory result — confirming the whole construction is self-consistent, not just a fit to known numbers after the fact.
Conformal mapping is one member of a much larger family of problem-solving moves: transform into a domain where the problem is easy, solve it there, transform the answer back.
| Transform | Hard problem | Easy domain |
|---|---|---|
| Logarithm | Multiplication and division | Addition and subtraction |
| Laplace transform | Linear ODEs/PDEs with derivatives | Algebraic equations |
| Electrical-circuit analogy | An unfamiliar diffusion or vibration system | A familiar resistor/capacitor network |
| Conformal mapping | Potential flow around a hard shape | Potential flow around a circle |
The specific mechanics never repeat — but the shape of the move always does: find the transform, solve where it's easy, and trust that the transform carries the answer back faithfully. Recognizing that move itself, independent of which field you're in, is often more useful than any single technique on this list.
EngineeringCandy · Learn · the theory behind the playground