2 Conformal mapping

In Section 26.1 we saw that the real and imaginary parts of an analytic function each satisfies Laplace’s equation. We shall show now that the curves

u ( x , y ) = constant and v ( x , y ) = constant

intersect each other at right angles (i.e. are orthogonal ). To see this we note that along the curve u ( x , y ) = constant we have d u = 0 . Hence

d u = u x d x + u y d y = 0 .

Thus, on these curves the gradient at a general point is given by

d y d x = u x u y .

Similarly along the curve v ( x , y ) = constant, we have

d y d x = v x v y .

The product of these gradients is

( u x ) ( v x ) ( u y ) ( v y ) = ( u x ) ( u y ) ( u y ) ( u x ) = 1

where we have made use of the Cauchy-Riemann equations. We deduce that the curves are orthogonal.

As an example of the practical application of this work consider two-dimensional electrostatics. If u = constant gives the equipotential curves then the curves v = constant are the electric lines of force . Figure 2 shows some curves from each set in the case of oppositely-charged particles near to each other; the dashed curves are the lines of force and the solid curves are the equipotentials.

Figure 2

No alt text was set. Please request alt text from the person who provided you with this resource.

In ideal fluid flow the curves v = constant are the streamlines of the flow.

In these situations the function w = u + i v is the complex potential of the field.

2.1 Function as mapping

A function w = f ( z ) can be regarded as a mapping, which maps a point in the z -plane to a point in the w -plane. Curves in the z -plane will be mapped into curves in the w -plane.

Consider aerodynamics where we are interested in the fluid flow in a complicated geometry (say flow past an aerofoil). We first find the flow in a simple geometry that can be mapped to the aerofoil shape (the complex plane with a circular hole works here). Most of the calculations necessary to find physical characteristics such as lift and drag on the aerofoil can be performed in the simple geometry - the resulting integrals being much easier to evaluate than in the complicated geometry.

Consider the mapping

w = z 2 .

The point z = 2 + i maps to w = ( 2 + i ) 2 = 3 + 4 i . The point z = 2 + i lies on the intersection of the two lines x = 2 and y = 1 . To what curves do these map? To answer this question we note that a point on the line y = 1 can be written as z = x + i . Then

w = ( x + i ) 2 = x 2 1 + 2 x i

As usual, let w = u + i v , then

u = x 2 1 and v = 2 x

Eliminating x we obtain:

4 u = 4 x 2 4 = v 2 4 so v 2 = 4 + 4 u is the curve to which y = 1 maps.

Example 5

Onto what curve does the line x = 2 map?

Solution

A point on the line is z = 2 + y i . Then

w = ( 2 + y i ) 2 = 4 y 2 + 4 y i

Hence u = 4 y 2 and v = 4 y so that, eliminating y we obtain

16 u = 64 v 2 or v 2 = 64 16 u

In Figure 3(a) we sketch the lines x = 2 and y = 1 and in Figure 3(b) we sketch the curves into which they map. Note these curves intersect at the point ( 3 , 4 ) .

Figure 3

No alt text was set. Please request alt text from the person who provided you with this resource.

The angle between the original lines in (a) is clearly 9 0 0 ; what is the angle between the curves in (b) at the point of intersection?

The curve v 2 = 4 + 4 u has a gradient d v d u . Differentiating the equation implicitly we obtain

2 v d v d u = 4 or d v d u = 2 v

At the point ( 3 , 4 ) d v d u = 1 2 .

Task!

Find d v d u for the curve v 2 = 64 16 u and evaluate it at the point ( 3 , 4 ) .

2 v d v d u = 16 d v d u = 8 v . At v = 4 we obtain d v d u = 2. Note that the product of the gradients at ( 3 , 4 ) is 1 and therefore the angle between the curves at their point of intersection is also 9 0 0 . Since the angle between the lines and the angle between the curves is the same we say the angle is preserved .

In general, if two curves in the z -plane intersect at a point z 0 , and their image curves under the mapping w = f ( z ) intersect at w 0 = f ( z 0 ) and the angle between the two original curves at z 0 equals the angle between the image curves at w 0 we say that the mapping is conformal at z 0 .

An analytic function is conformal everywhere except where f ( z ) = 0 .

Task!

At which points is w = e z not conformal?

f ( z ) = e z . Since this is never zero the mapping is conformal everywhere.

2.2 Inversion

The mapping w = f ( z ) = 1 z is called an inversion . It maps the interior of the unit circle in the z -plane to the exterior of the unit circle in the w -plane, and vice-versa. Note that

w = u + i v = x x 2 + y 2 y x 2 + y 2 i and similarly z = x + i y = u u 2 + v 2 v u 2 + v 2 i

so that

u = x x 2 + y 2 and v = y x 2 + y 2 .

A line through the origin in the z -plane will be mapped into a line through the origin in the w -plane. To see this, consider the line y = m x , for m constant. Then

u = x x 2 + m 2 x 2 and v = m x x 2 + m 2 x 2

so that v = m u , which is a line through the origin in the w -plane.

Task!

Consider the line a x + b y + c = 0 where c 0 . This represents a line in the z -plane which does not pass through the origin. To what type of curve does it map in the w -plane?

The mapped curve is

a u u 2 + v 2 b v u 2 + v 2 + c = 0

Hence a u b v + c ( u 2 + v 2 ) = 0 . Dividing by c we obtain the equation:

u 2 + v 2 + a c u b c v = 0

which is the equation of a circle in the w -plane which passes through the origin.

Similarly, it can be shown that a circle in the z -plane passing through the origin maps to a line in the w -plane which does not pass through the origin. Also a circle in the z -plane which does not pass through the origin maps to a circle in the w -plane which does pass through the origin. The inversion mapping is an example of the bilinear transformation :

w = f ( z ) = a z + b c z + d where we demand that a d b c 0

(If a d b c = 0 the mapping reduces to f ( z ) = constant.)

Task!

Find the set of bilinear transformations w = f ( z ) = a z + b c z + d which map z = 2 to w = 1 .

1 = 2 a + b 2 c + d . Hence 2 a + b = 2 c + d .

Any values of a , b , c , d satisfying this equation will do provided a d b c 0 .

Task!

Find the bilinear transformations for which z = 1 is mapped to w = 3 .

3 = a + b c + d . Hence a + b = 3 c + 3 d .

Example 6

Find the bilinear transformation which maps

  1. z = 2 to w = 1 , and
  2. z = 1 to w = 3 , and
  3. z = 0 to w = 5
Solution

We have the answers to 1. and 2. from the previous two Tasks:

2 a + b = 2 c + d a + b = 3 c + 3 d

If z = 0 is mapped to w = 5 then 5 = b d so that b = 5 d . Substituting this last relation into the first two obtained we obtain

2 a 2 c 6 d = 0 a + 3 c 8 d = 0

Solving these two in terms of d we find 2 c = 11 d and 2 a = 17 d . Hence the transformation is:

w = 17 z 10 11 z + 2 (note that the d ’s cancel in the numerator and denominator).

Some other mappings are shown in Figure 4.

Figure 4

No alt text was set. Please request alt text from the person who provided you with this resource.

As an engineering application we consider the Joukowski transformation

w = z 2 z where is a constant.

It is used to map circles which contain z = 1 as an interior point and which pass through z = 1 into shapes resembling aerofoils. Figure 5 shows an example:

Figure 5

No alt text was set. Please request alt text from the person who provided you with this resource.

This creates a cusp at which the associated fluid velocity can be infinite. This can be avoided by adjusting the fluid flow in the z -plane. Eventually, this can be used to find the lift generated by such an aerofoil in terms of physical characteristics such as aerofoil shape and air density and speed.

Exercise

Find a bilinear transformation w = a z + b c z + d which maps

  1. z = 0 into  w = i
  2. z = 1  into  w = 0
  3. z = i  into  w = 1
  1. z = 0 , w = i gives  i = b d so that b = d i
  2. z = 1 , w = 0 gives 0 = a + b c + d so a + b = 0 so a = b .
  3. z = i , w = 1 gives 1 = a i + b c i + d so that  c i + d = a i + b = d + d i (using 1. and 2.)

We conclude from 3. that c = d . We also know that a = b = d i .

Hence w = d i z + d i d z + d = i z + i z + 1