0% found this document useful (0 votes)
23 views

Curvilinear Co-Ordinates: Peter Symonds

This document discusses different coordinate systems used in multivariable calculus problems, including orthogonal curvilinear coordinates, plane polar coordinates, cylindrical coordinates, and spherical coordinates. It provides the definitions and relationships between Cartesian and each curvilinear coordinate system. It also defines the gradient, divergence, curl, and Laplacian operators in each coordinate system and provides examples of their use.

Uploaded by

Hansraj Rahul
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
23 views

Curvilinear Co-Ordinates: Peter Symonds

This document discusses different coordinate systems used in multivariable calculus problems, including orthogonal curvilinear coordinates, plane polar coordinates, cylindrical coordinates, and spherical coordinates. It provides the definitions and relationships between Cartesian and each curvilinear coordinate system. It also defines the gradient, divergence, curl, and Laplacian operators in each coordinate system and provides examples of their use.

Uploaded by

Hansraj Rahul
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 30

Curvilinear

co-ordinates
Peter
Symonds

Curvilinear co-ordinates
Peter Symonds
University of Manchester

CHEN20041, 2016

Orthogonal curvilinear co-ordinates


Curvilinear
co-ordinates
Peter
Symonds

Certain problems may be more suited to other co-ordinate


systems. For example., a problem with circular symmetry is
likely to be easier to state and to solve in terms of polar
co-ordinates.
When working in 2 dimensions, we may use plane polar
co-ordinates. When working in 3 dimensions, two commonly
used co-ordinates systems are cylindrical co-ordinates and
spherical co-ordinates.

Plane polar co-ordinates


Curvilinear
co-ordinates
Peter
Symonds

As we have seen previously, plane polar co-ordinates take the


form (r , ), where r is the distance of a given point from the
origin, and is the angle between the position vector of the
point and a fixed direction (usually the positive x-axis).
There are the following relations between Cartesian and plane
polar co-ordinates.
p
x = r cos
r = x2 + y2
y = r sin
tan = y /x

Cylindrical co-ordinates
Curvilinear
co-ordinates
Peter
Symonds

In the cylindrical co-ordinate system we will label our


co-ordinates as (, , z). and can be chosen to lie in the
xy -plane, and can be thought of as the usual plane polar
co-ordinates, while z gives the position along the axis
perpendicular to this plane.

Cylindrical co-ordinates
Curvilinear
co-ordinates
Peter
Symonds

There are the following relations between Cartesian and


cylindrical co-ordinates.
p
x = cos
= x2 + y2
y = sin
tan = y /x
z =z
z =z

Spherical co-ordinates
Curvilinear
co-ordinates
Peter
Symonds

Spherical co-ordinates are labelled (r , , ). For a given point,


r is the distance from the point to the origin, is the angle
between the position vector of the point and a fixed direction
(sometimes called the zenith), and is the angle between the
position vector of the point and a fixed direction in a plane
orthogonal to the zenith direction.
For example, positions on the surface of the Earth are often
given in terms of latitude and longitude - these correspond to
and , respectively.

Spherical co-ordinates
Curvilinear
co-ordinates

Peter
Symonds

(r, , )
r

Spherical co-ordinates
Curvilinear
co-ordinates
Peter
Symonds

There are the following relations between Cartesian and


spherical co-ordinates.
p
x = r sin cos
r = x2 + y2 + z2
y = r sin sin
cos = z/r
z = r cos
tan = y /x

Example
Curvilinear
co-ordinates
Peter
Symonds

Example

Let p = (1, 1, 2) be a point in 3-dimensions, given by


Cartesian co-ordinates.
Write p as a point in spherical co-ordinates.

Example
Curvilinear
co-ordinates
Peter
Symonds

Example (cont)
Using the formulae from the previous slide, we have
q

r = (1)2 + (1)2 + ( 2)2 = 4 = 2.

Also, = cos1 ( 2/2) = 4 .


We have to be a little careful when working out . We have
tan1 ((1)/(1)) = /4. However, (1, 1) lies in the third
quadrant, so we require < < /2. Therefore we must
take
= /4 = 3/4.
Hence, in spherical co-ordinates, p = (2, /4, 3/4).

Unit vectors
Curvilinear
co-ordinates
Peter
Symonds

When working with Cartesian co-ordinates, we use the unit


vectors ~i, ~j and ~k.
We need corresponding unit vectors for curvilinear co-ordinates.
Note: you do not need to memorise the following formulae for
unit vectors in non-Cartesian co-ordinates.

Unit vectors - polar


Curvilinear
co-ordinates
Peter
Symonds

For 2-dimensional polar co-ordinates, we denote our unit


vectors by ~r and
~ . We have the following relations.
~i = cos ~r sin ~

~r = cos ~i + sin ~j

~j = sin ~r + cos ~

~ = sin ~i + cos ~j

y
2.0

`
j

1.5

r`

1.0

r
0.5

0.0
0.0

x
0.5

1.0

1.5

Note that ~r and


~ are orthogonal.

2.0

2.5

Unit vectors - polar


Curvilinear
co-ordinates
Peter
Symonds

Alternatively, we can write the unit vector ~r and


~ in terms of
the Cartesian coordinates x and y as follows.
~

j
~r = x i+y
2
2
x +y

Since cos =
sin =

y
x 2 +y 2

x
x 2 +y 2

y i+x j

~=
2
2
x +y

(adjacent over hypotenuse) and

(opposite over hypotenuse), we see that these

expressions for ~r and


~ are equivalent to those on the previous
slide.

Unit vectors - cylindrical


Curvilinear
co-ordinates
Peter
Symonds

For cylindrical co-ordinates, we will denote our unit vectors by

~,
~ and ~k (sometimes we use ~z instead of ~k. We have the
following relations.
~i = cos ~
sin ~

~ = cos ~i + sin ~j

~j = sin ~
+ cos ~

~ = sin ~i + cos ~j

~k = ~k

~k = ~k

Unit vectors - spherical


Curvilinear
co-ordinates
Peter
Symonds

For spherical co-ordinates, we will denote our unit vectors by ~r ,


~ and

~ . To convert spherical unit vectors to Cartesian unit


vectors, we have the following relations.
~ sin ~
~i = sin cos ~r + cos cos

~ + cos ~
~j = sin sin ~r + cos sin

~k = cos ~r sin
~

Unit vectors - spherical


Curvilinear
co-ordinates
Peter
Symonds

While to convert Cartesian unit vector to spherical unit vectors,


we have the following relations.
~r = sin cos ~i + sin sin ~j + cos ~k
~ = cos cos ~i + cos sin ~j sin ~k

~ = sin ~i + cos ~j

grad, div, curl and 2


Curvilinear
co-ordinates
Peter
Symonds

We can now define grad, div, curl and the Laplacian with
respect to the curvilinear co-ordinate systems.

Plane polar co-ordinates


Curvilinear
co-ordinates
Peter
Symonds

~ a vector field.
Let f (r , ) be a scalar field and ~v = vr ~r + v
Then
1 f
f
~r +

~
r
r


1 (rvr ) v
div ~v =
+
r
r

grad f =

2 f =

2f
1 f
1 2f
+
+
r 2
r r
r 2 2

Plane polar co-ordinates


Curvilinear
co-ordinates
Peter
Symonds

We dont need to worry about how to derive the formulae on


the previous slide. However, the reason we get a factor of 1r
appearing in grad f , for example, is that an infinitesimal arc
length is given by r , so depends not just on the angle but
the radius r also. This is in contrast to the situation in
Cartesian co-ordinates, where an infinitesimal length in the
x-direction, for example, is just x.

Cylindrical co-ordinates
Curvilinear
co-ordinates
Peter
Symonds

Let f (, , z) be a scalar field and ~v = v


~ + v
~ + vz ~z a
vector field. Then
grad f =

f
1 f
f
~z

~+

~+

1 (v ) 1 v vz
+
+


z



~z
~
~

1
curl ~v = / / /z

v
v
vz

div ~v =

2 f =

2f
1 f
1 2f
2f
+
+
+
2 2 2 z 2

Spherical co-ordinates
Curvilinear
co-ordinates
Peter
Symonds

~ + v
~ a
Let f (r , , ) be a scalar field and ~v = vr ~r + v
vector field. Then
1 f ~
1 f
f
~r +
+

~
r
r
r sin


1
(r 2 sin vr ) (r sin v ) (rv )
div ~v = 2
+
+
r sin
r

grad f =


~r
1
curl ~v = 2
/r
r sin
vr
2 f =


~
r
r sin ~

/
/
rv r sin v

2f
2 f
1 2f
cos f
1
2f
+
+
+
+
.
r 2 r r r 2 2 r 2 sin r 2 sin2 2

Example 1
Curvilinear
co-ordinates
Peter
Symonds

Example
Suppose f (x, y ) = x + y is a scalar field given in 2-dimensional
Cartesian coordinates. Note that grad f = ~i + ~j.
(a) Convert f to polar coordinates (r , ).
(b) Calculate grad f using the formula for grad in polar
coordinates.
(c) Convert the answer to part (b) back to Cartesian
coordinates and check it agrees with the answer given
above.

Example 1
Curvilinear
co-ordinates
Peter
Symonds

We have x = r cos and y = r sin , so converting to polar


coordinates we have
f (r , ) = r cos + r sin = r (cos + sin ).
Recall that in polar coordinates
grad f =
Since
that

f
r

= cos + sin and

grad f

1 f
f
~r +

~.
r
r
f

= r ( sin + cos ) we find

1
= (cos + sin )~r + (r ( sin + cos ))~

r
= (cos + sin )~r + ( sin + cos )~
.

Example 1
Curvilinear
co-ordinates
Peter
Symonds

Now we need to convert back to Cartesian coordinates. We


need to convert the coordinate and the unit vectors ~r and
~.
The quickest way is to notice that cos = 2x 2 and
x +y

sin =

y ,
x 2 +y 2

and hence ~r =

~ ~
x i+y j
x 2 +y 2

y i+x j
and
~=
.
2
2
x +y

Example 1
Curvilinear
co-ordinates
Peter
Symonds

Substituting all this into our expression for grad f , we get


grad f =

x
p

x2

y2

+ p

y
x2

y2

(x~
i + y~
j)
+
p
x2 + y2

y
p

x2

y2

+ p

x
x2

y2

(y~
i + x~
j)
.
p
x2 + y2

Expanding this we find a lot of cancellation, and it becomes


grad f =
as required.

(x 2 + y 2 )(~i + ~j) ~ ~
= i + j,
x2 + y2

Example 2
Curvilinear
co-ordinates
Peter
Symonds

Example
An electric current I flowing in a thin wire along the z-axis in
the positive direction generates a magnetic field
~
~
~ = 0 I (y i + x j) .
B
2(x 2 + y 2 )
~ be the vector field
Let A
~ = 0 I log(x 2 + y 2 )~k.
A
4
~ to cylindrical co-ordinates. Then find curl A.
~
First convert A
~ by converting back to Cartesian
Show this is equal to B
co-ordinates.

Example 2
Curvilinear
co-ordinates
Peter
Symonds

Example (cont)
We have x = cos and y = sin , so in cylindrical
co-ordinates we have
~ =
A
=

0 I
log(2 cos2 + 2 sin2 )~k
4
0 I
log(2 )~k
4

(since cos2 + sin2 = 1).


Notice that we dont need to change ~k as it is already a
cylindrical unit vector.

Example 2
Curvilinear
co-ordinates
Peter
Symonds

Example (cont)
We now apply the cylindrical formula for curl.
~
curl A




~
~
~

z
1
/ / /z

A
A
Az




(A )
Az
~
Az
A ~



(A )
A ~
z
+


Az

(since A = A = 0 and neither or z appear in Az ).

Example 2
Curvilinear
co-ordinates
Peter
Symonds

Example (cont)
Using the chain rule, we have
Az

=
=
=

Therefore
~=
curl A



0 I
2
log( )

4
0 I 2
4 2
0 I
.
2


0 I
2

~=

0 I
~
.
2

Example 2
Curvilinear
co-ordinates
Peter
Symonds

Example (cont)
We now need to convert back to Cartesian co-ordinates. We
p
y ~i + x ~j
have
~=p
and = x 2 + y 2 , so
x2 + y2
~
~
~ = 0 I (y i + x j)
curl A
2
2
2(x + y )
~ as required.
which is equal to B,

You might also like