D

Cylindrical coordinates
Cartesian coordinates x; y; z and cylindrical coordinates1 r; ; z are related by
x D r cos ; y D r sin ; zDz (D.1)
with the range of variation 0  r < 1, 0   < 2, and 1 < z < 1. The inverse
transformation is
p y
r D x2 C y2;  D arctan ; zDz (D.2)
x
The two first equations in both transformations simply define polar coordinates in the xy-
plane, whereas the last, z D z, is included to emphasize that this is a transformation in
three-dimensional space.
z
eOz
D.1 Local basis vectors 6
6
O
The cylindrical basis vectors follow from the geometry (see the margin figure), * e
u


x 
@x @R eOr
r
@
eOr D D .cos ; sin ; 0/; (D.3a) -y
@r @ r
1 @x ......
..........
 .....
..............................
@
eO D D . sin ; cos ; 0/; (D.3b) @r
r @
x
@x Cylindrical coordinates and basis
eOz D D .0; 0; 1/: (D.3c)
@z vectors.
They are clearly orthonormal and satisfy eOr  eO
eOz D 1.
A Cartesian vector field U may be resolved in this basis
U D eOr Ur C eO U C eOz Uz ; (D.4)
where
Ur D U
eOr ; U D U
eO ; Uz D U
eOz : (D.5)
A tensor field T may similarly be resolved in dyadic products of the local basis vectors.
1 Some texts use ‚ instead of  as the conventional name for the polar angle in the plane. Similarly, the radial

variable r is sometimes denoted s to distinguish it from the spherical radial distance.

Copyright
c 1998–2010 Benny Lautrup
622 PHYSICS OF CONTINUOUS MATTER

D.2 Line, surface and volume elements

The differentials along the local coordinate axes,

dr x D eOr dr; d x D eO rd; dz x D eOz dz; (D.6)

allow us to resolve the Cartesian line, surface and volume elements in the local basis,

d `  dr x C d x C dz x D eOr dr C eO rd C eOz dz; (D.7)

d S  d x  dz x C dz x  dr x C dr x  d x
D eOr r ddz C eO drdz C eOz r ddr; (D.8)

d V  dr x  d x
dz x D r ddrdz (D.9)

Using these infinitesimals, all integrals can be converted to cylindrical coordinates.

D.3 Resolution of the gradient

The derivatives with respect to the cylindrical coordinates are obtained by differentiation
through the Cartesian coordinates,

@ @x @
D
D eOr
r D rr ;
@r @r @x
@ @x @
D
D r eO
r D rr :
@ @ @x

Nabla may now be resolved on the local basis

@ 1 @ @
r D eOr rr C eO r C eOz rz D eOr C eO C eOz : (D.10)
@r r @ @z

Finally, we note that the only non-vanishing derivatives of the basis vectors are

@eOr @eO
D eO ; D eOr ; (D.11)
@ @

These are the fundamental tools necessary to convert differential equations from Cartesian to
cylindrical coordinates.

Copyright
c 1998–2010 Benny Lautrup
D. CYLINDRICAL COORDINATES 623

D.4 First order expressions

Here follows a list of various combinations of a single nabla and various fields. In writing
out the results we refrain from using the nabla projections, rr etc, but express everything in
conventional partial derivatives, @=@r etc.
The three basic first order expressions are the gradient, divergence and curl,

@S 1 @S @S
r S D eOr C eO C eOz ; (D.12)
@r r @ @z
@Ur 1 @U @Uz Ur
r
U D C C C ; (D.13)
@r r @ @z r
     
1 @Uz @U @Ur @Uz @U 1 @Ur U
r  U D eOr C eO C eOz C :
r @ @z @z @r @r r @ r
(D.14)

The tensor gradient is used in solid and fluid mechanics to calculate the stress tensor. In dyadic
notation (see appendix B), we have

@Ur @U @Uz

r U D eOr eOr C eOr eO C eOr eOz
@r @r @r
   
1 @Ur U 1 @U Ur 1 @Uz
C eO eOr C eO eO C C eO eOz
r @ r r @ r r @
@Ur @U @Uz
C eOz eOr C eOz eO C eOz eOz (D.15)
@z @z @z

The dot product with a vector V from the left becomes,

 
@Ur V @Ur @Ur V U
.V
r /U D eOr Vr C C Vz
@r r @ @z r
 
@U V @U @U V Ur
C eO Vr C C Vz C
@r r @ @z r
 
@Uz V @Uz @Uz
C eOz Vr C C Vz (D.16)
@r r @ @z

It is used for calculating the advective terms in equations of motion.

Finally, the divergence of a full-fledged tensor field becomes (also in dyadic notation)
 
@Trr 1 @Tr @Tzr Trr T
r
T D eOr C C C
@r r @ @z r r
 
@Tr 1 @T @Tz Tr Tr
C eO C C C C
@r r @ @z r r
 
@Trz 1 @Tz @Tzz Trz
C eOz C C C (D.17)
@r r @ @z r

This may be used in formulating the equations of motion for continuum physics, although it
is normally not necessary.

Copyright
c 1998–2010 Benny Lautrup
624 PHYSICS OF CONTINUOUS MATTER

D.5 Second order expressions

Expressions involving two nabla factors also turn up everywhere in continuum physics. They
can of course be derived by combinations of first order expressions, but it is nevertheless
useful to list them separately.
The Laplacian of a scalar field is calculated from the divergence of the gradient, r 2 S D
r
.r S /, and becomes after the dust has settled,

@2 S 1 @2 S @2 S 1 @S
r 2S D C C C : (D.18)
@r 2 r 2 @ 2 @z 2 r @r
The Laplacian can also be applied to a vector field, and may be obtained from the divergence
of the gradient of the vector field, r 2 S D r
.r U /. It is somewhat more complicated,
 2
1 @2 Ur @2 Ur

@ Ur 1 @Ur 2 @U Ur
r 2 U D eOr C C C
@r 2 r 2 @ 2 @z 2 r @r r 2 @ r2
 2
1 @2 U  @2 U

@ U 1 @U 2 @Ur U
C eO C C C C
@r 2 r 2 @ 2 @z 2 r @r r 2 @ r2
 2
1 @2 Uz @2 Uz

@ Uz 1 @Uz
C eOz C C C (D.19)
@r 2 r 2 @ 2 @z 2 r @r

Another second order expression is the gradient of a divergence,

 2
1 @2 U @2 Uz

@ Ur 1 @Ur 1 @U Ur
r .r
U / D eOr 2
C C C
@r r @r@ @r@z r @r r 2 @ r2
 2
1 @2 U 1 @2 Uz

1 @ Ur 1 @Ur
C eO C 2 C C 2
r @@r r @ 2 r @@z r @
 2 2 2 
@ Ur 1 @ U @ Uz 1 @Ur
C eOz C C C (D.20)
@z@r r @z@ @z 2 r @z

In practice we do not need more complex expressions than these.

Copyright
c 1998–2010 Benny Lautrup