ZPEM2309 Engineering Mathematics 2A MVC Tutorial Problem Set 2 (Sections 3.23.

7) ANSWERS page 1 of 12

CHAPTER 3

Vector Differentiai'ion

1. SLS 3.133.16 Answer: See SLS (below):

3.13. Let A = (2x 2y - x4 )i + (crY

+ (x 2 cos y)k. Find:

y sinx)j

(a)

~~,

(b) : .

Solution

(a) aA = ~(2x2y _ x4 )i + ~(e'Y _ y sinx)j + ~(x2 cosy)k

ax ax
ax
ax

= (4xy -

4x3 )i + (yeTY

ycosx)j + 2xcosyk

(b) aA = ~(2x2y _ x4 )i + ~V)' - ysinx)j

ay ay
ay

+~(~ cosy)k
ay

= 22i + (xe'Y - sinx)j - ~ sinyk

3.14. Let A be the vector in Problem 3.13. Find:

Solution

2
a 4
3.
a 'Y
).
a
ax2 = a/ xy - 4x )1 + ax (ye - ycosx j + ax (2x cos y)k

(a) a A

(b) a2A
ay2

= (4y -12~)i + (y2e'Y

+ ysinx)j + 2cosyk

= ~(2~)i +~(xe'Y _

sinx)j -

ay

ay

~(~ siny)k

ay
2
2
= 0 + x e'Yj - ~ cosyk = x e'Yj - x2 cosyk

3.15. Let A be the vector in Problem 3.13. Find:

a2A

(a) axay'

a2A

(b) ayax'

Solution
(a) a2

A a (aA) =-(2x
a
a
)I+-(xe
2.

-=- -

axay

ax ay

ax

= 4xi + (xyeT)'

(b) -a2

X).

ax

+ e\)' -

a (x-smy)k
2 .
-smx)J
-ax

cos x)j - 2x sin yk

A= -a (aA)
a
a . - ycosx)J. + -(2xcosy)k
a
- = -(4xy
- 4x3)1. + -(ye\)

ayax

ay ax

ay

= 4xi + (xyCl'

Note that a2Ajayax

ay

+ e'Y -

ay

cosx)j - 2xsinyk

= a2Ajaxay, that is, the order of differentiation is immaterial. This is true in general if A has

continuous partial derivatives of the second order at least.

3.16. Suppose cfJ(x, y, z) = xy2 Z and A = xzi - xy2j

+ yz2k. Find ax

az

(cfJA) at the point (2, -1, 1).

Solution

A = (,,-y2 z)(xzi - xlj

+ yik) =

~lZ2i - ~lti

~(A) = ~(x2lii - ~lti + xy3 Z3 k) =

az

az

2lzi - x2lj

2x

~(A) = ~(2X2lzi - ~lj + 3xlz2k) = 4Ay2zi axaz

+ xy3Z3k

ax

2xy4j + 3lz2k

2
a::az (A) = :x (4xy2Zi - 2xy4j + 3lz k) = 4lzi - 2lj
If x

= 2, y = -1, and z = 1, this becomes 4( -1)2(1)i -

2( _1)4j

= 4i -

+ 3xy3Z2k

2j.

ZPEM2309 Engineering Mathematics 2A MVC Tutorial Problem Set 2 (Sections 3.23.7) ANSWERS page 2 of 12

u and 2 u for the following vector fields u = (u, v, w), and also match them to
one of the figures (e.g. (i) is shown in figure b; note that the vector lengths may have been scaled):
u v
u , u , v ,
x , y ,
y x

2. Calculate
(i)

u = (1, 0, 0)

(ii)

u = (1, 2, 0)

(iii)

u = ( x, 0, 0)

(iv)

u = (2 y, 2, 0)

(v)

u = (3, 2x, 0)

(vi)

u = (y, x, 0)

y0

y0

y0

a
4

b
4

(vii)

u = ( x, y, 0)

(viii)

u = (y2 , 0, 0)

u = ( x, y, 0)

y0

y0

y0

(ix)

d
4

e
4

y0

y0

y0

g
4

h
4

Answer:

Q
(i)
(ii)
(iii)
(iv)
(v)
(vi)
(vii)
(viii)
(ix)

Fig
field
du/dx dv/dy div(u) du/dy dv/dx curl(u) lap(u)
b
(1,0,0)
0
0
0
0
0
(0,0,0) (0,0,0)
d (81,2,0)
0
0
0
0
0
(0,0,0) (0,0,0)
a
(x,0,0)
1
0
1
0
0
(0,0,0) (0,0,0)
f (28y,2,0)
0
0
0
81
0
(0,0,1) (0,0,0)
i
(3,2x,0)
0
0
0
0
2
(0,0,2) (0,0,0)
e
(y,8x,0)
0
0
0
1
81 (0,0,82) (0,0,0)
h
(x,y,0)
1
1
2
0
0
(0,0,0) (0,0,0)
c (y^2,0,0)
0
0
0
2y
0 (0,0,82y) (2,0,0)
g
(x,8y,0)
1
81
0
0
0
(0,0,0) (0,0,0)

3. If A = ( x2 z2 , 2y2 z2 , xy2 z), find

i.

ii.

iii.

( A)

iv.

( A)

v.

( A)

vi.

2 A

(NB: worked solution in SLS is wrong)

(NB: working in SLS is wrong)

ZPEM2309 Engineering Mathematics 2A MVC Tutorial Problem Set 2 (Sections 3.23.7) ANSWERS page 3 of 12
Answer:
i.

A = 2xz2 4yz2 + xy2

ii.

A = (2xyz + 4y2 z, 2x2 z y2 z, 0) (see SLS 4.23 for working but note the corrections below!)

iii.

( A) = 2yz 2yz = 0 (as it must be, by identity (x))

iv.

( A) = (2xz2 4yz2 + xy2 ) = (2z2 + y2 , 2xy 4z2 , 4xz 8yz) (using result from (i))

v.

( A) = (y2 2x2 , 2xy + 4y2 , 2xz 8yz) (using result from (ii); see SLS 4.24 with corrections below!)

vi.

2 A = ( A) ( A) = (2z2 + y2 , 2xy 4z2 , 4xz 8yz) (y2 2x2 , 2xy + 4y2 , 2xz 8yz) = (2z2 +

(see SLS problem 4.15)

y2 y2 + 2x2 , 2xy 4z2 2xy 4y2 , 4xz 8yz 2xz + 8yz) = (2z2 + 2x2 , 4z2 4y2 , 4xz 2xz)
from (iv) and (v))
2 f
x2

4. Show that 2 f =
2 f
x2

2 f
y2

2 f
.
z2

2 f
y2

2 f
z2

f f f

in Cartesian coordinates. Answer: 2 f = ( f ) = ( x , y , z ) =

5. Prove the following vector identities by writing them out in Cartesian components.
(i)

( + ) = +

(ii)

( A + B) = A + B

(iii)

( A + B) = A + B

(i.e. curl is a linear operator)

(iv)

(A) = () A + ( A)

(the product rule for div)

(v)

(A) = () A + ( A)

(ix)

() = 0

(x)

( A) = 0

(i.e. grad is a linear operator)

(i.e. div is a linear operator)

(using results

(the product rule for curl)

(the curl of a gradient field is zero, i.e. is irrotational)

(the divergence of a curl field is zero, i.e. A is solenoidal)

ZPEM2309 Engineering Mathematics 2A MVC Tutorial Problem Set 2 (Sections 3.23.7) ANSWERS page 4 of 12
Answer: Worked solutions are provided in Spiegel, Lipschutz & Spellman:
(i)

See SLS problem 4.2a:

(ii)

See SLS problem 4.18a:

(iii)

See SLS problem 4.25a:

(iv)

See SLS problem 4.18b:

ZPEM2309 Engineering Mathematics 2A MVC Tutorial Problem Set 2 (Sections 3.23.7) ANSWERS page 5 of 12
(v)

See SLS problem 4.25b:

(ix)

See SLS problem 4.27a:

(x)

See SLS problem 4.27b:

Do the following problems from Spiegel, Lipschutz & Spellman 2nd Edition (SLS) without looking at the worked
solutions, and then cross-check with the worked solutions where available. (Those with a * are more advanced.)
p
6. SLS 4.74 (where r = x2 + y2 + z2 is the distance from the origin) Answer:

ZPEM2309 Engineering Mathematics 2A MVC Tutorial Problem Set 2 (Sections 3.23.7) ANSWERS page 6 of 12

7. SLS 4.76 (where r = ( x, y, z) is the position vector; hint: check the table of vector identities and see section 3.3 in

the notes) Answer: Using identity (vi), v = ( r ) = ( r ) = 0 since is constant and r is curl-free.

This is a special case of the results proved in SLS 4.88 and SLS 4.104. Note from the example in section 3.3 in the notes
that v is a solid-body rotation (with angular velocity vector ) so it makes intuitive sense that v has zero divergence.

(Weve just proven that v is solenoidal. It can therefore be written as the curl of a vector potential. So heres another question:

p
show that G = r2 /2 is a vector potential for v = r , where r = x2 + y2 + z2 . This is a neat result: this vector

potential is antiparallel to the rotation vector .)

8. SLS 4.93 (hint: check the table of vector identities) Answer:

Need to show that () = 0 (irrotational).

Method 1:
By the chain rule, = 12 2 (verify by components that this is true).

So by identity (ix), () = 21 (2 ) = 0 .

Method 2:
Let A = .

Thus we need to show that (A) = 0 .

By identity (v), (A) = () A + ( A).

But by identity (ix), A = () = 0 .

Similarly, loss in volume per unit time in y direction = aV2 fu.c,.y.c,.z

ay

ZPEM2309 Engineering
Problem
2 (Sections 3.23.7) ANSWERS page 7 of 12
loss Mathematics
in volume per 2A
unitMVC
time Tutorial
in z direction
= a;:Set
.c,.x.c,.y.c,.z.

Then,
in
volume
Therefore
(total
)loss
= (
) (per
). unit
Butvolume
this is per
theunit
zerotime
vector because it is the cross-product of two parallel

vectors.
So () = 0 , i.e. is irrotational.
3)
iNI iN2 iN

4 Gradient, Divergence, Curl

+ - +CHAPTER
(ax ay
az- .c,.x.c,.y.c,.,z .
Method 3:
=
=dlvv=Vv
.c,.x.c,.y.c,.z
The long way (by components) not recommended!
aH
aE
.
2
"a2 u
4.31. Suppose V E = 0, V H = 0, V x E = - at' V x H = at Show that E and H satIsfy V u = at2'
.is true
?nlycheck
in thethe
limit
as the.
shrinks
to P, i.e.
.c,.x, .c,.y,directly
and.c,.z from
approach
zero.(viii)
9. SLS 4.98 This
(where
v =exactly
k v k; hint:
table
of parallelepiped
vector identities)
Answer:
Thisasfollows
identity

If
no
of
anywhere, then V v 2= O. This is called the continuity equation for an incompressible
by setting
A = B = v and recalling that v v = k v k .
Solution
Smce
flUid
IS neither created nor
destroyed

flUid.
at any point, it is said to have no sources or sinks. A vector such
1E
, as v whose divergence is zero is sometimes aH)
called solenoidal.
v x (V x E) = V x ( Tt = - ata (V x H) = - a (aE)
Tt = - ift2
10. SLS 4.22 Answer: See SLS:

CH

at

4.22. Determine the constant a so that the following 2vector is solenoidal.

2
1E
By Problem 4.29, V x (V x E) = - V2E + V(V E) = -V E.

Then V E =

2
V = (-4x - 6y + 3z)i + (- 2x + y - 5z)j

Similarly, V x (V x H) = V x

Solution
"

aE) a
(Tt = at (V x E) =

'

ata (aH)
-Tt = -

ift2.

+ (5x + 6y + az)k
aH
at2 '

1H

But
V x (VVx isH)solenoidal
= - V2H +ifV(V
H) = - V H.
Then V H - iji2'
Avector
its divergence
is zero.
The given equations are related to MaxweU's equations of electromagnetic theory. The equation

vo V =

a2u(-2x+y
a2u 1u
a+ ua/5x+ 6y + az) = -4+ 1 + a = -3 + a.
ax(-4x - 6y+ 3z) + ay
--5z)
+
+
=
2
2
ax
ay2 az
at2

isThen
calledVoV=-3+a=Owhena=3.
the wave equation.

The
Curl4.32 Answer:
Miscellaneous
Problems
11.
* SLS
See SLS:
4.32. A vector V is called irrotational if curl V = O. (a) Find constants a, b, and c so that

4.23. Suppose A = x2z2 j

2y2z2j

+ xy2 zk. Find V x A (or curl A) at the point P = (1,

V = (-4x - 3y + az)i + (bx + 3y + 5z)j + (4x + cy + 3z)k

Solution

-1, 1).

4.3

is irrotational. (b) Show that V can be expressed as the gradient of a scalar function.
i
j
k
Solution
a
a
a
VxA=
(a)

az

ax
ay
xZ Z2 -2y 2Z2

curl V = V x V

VxV=
) = [ ; (xy2 Zax

xy2 Z

a
2

(-2l z )}
ay -

-4x-3y+az bx+3y+5z
= (2xyz + 4yz2)i - (lz -a Wz)j

Thus V x A(P)

ay

= 2i + j.

b + 3y+5z

X2 Z2 )}
az 4x+cy+3z

ax
az
i-4x - 3y + az
4x+cy+ 3z

(-2lz2) + ; (x2l)

a
az
j
4x + cy + 3z

! ( [r

[r

ay

-4x - 3y + az

bx + 3y + 5z

= (c - 5)i - (4 - a)j + (b + 3)k.

This equals the zero vector when a = 4, b

= -3, and c =

5. So

V = (-4x - 3y + 4z)i + (-3x + 3y + 5z)j

(b)

Assume V

= V 4> =

a4>.
ax I

a4>.

a4>k Th

+ ay J + az'

+ (4x + 5y + 3z)k.

en

a4>
ax

= -4x -

3y + 4z

(1)

'

a4> = -3x+3y+5z

(2)

a4> = 4x + 5y + 3z
az

(3)

ay

Jk

4.3

CHAPTER 4

Gradient, Divergence, Curl

ZPEM2309 Engineering Mathematics 2A MVC Tutorial Problem Set 2 (Sections 3.23.7) ANSWERS page 8 of 12

CHAPTER 4

Gradient, Divergence, Curl

4> = -3xy +"2l + 5yz + g(x, z)

3
4> = 4xz + 5yz + "2Z2

(4)
(4)
(5)
(5)

+ h(x, y).

(6)

3
Comparison of (4), (5), and (6) shows
be a+common
4> =that
4xzthere
+ 5yzwill
+ "2Z2
h(x, y). value of 4> if we choose

(6)

Comparison of (4), (5), and (6) shows that there will be a common value of 4> if we choose
so that

3
4> = - 2.? + "2l

so that

+ "2 Z2 -

3xy + 4xz + 5yz

3 if 3V x V = 0, then we can find 4> so that V = V 4>.

Note that we can add any constant 4>to =0/.- In
2.?general,
+ "2l +
"2 Z2 - 3xy + 4xz + 5yz
A vector field V, which can be obtained from a scalar field 4>, so that V = V4> is called a conservative vector
and we
4> iscan
called
scalar
potential.
Note
conversely
V4>,
x V4> =so0that
(seeVProblem
field that
Note
addthe
any
constant
to 0/. In
general,
if V xthat,
V =ifY
0, =
then
wethen
canVfind
= V 4>. 4.27a).
A vector field V, which can be obtained from a scalar field 4>, so that V = V4> is called a conservative vector
4.33. field
Show
if cp(x,
z) ispotential.
any solution
of Laplace's
equation,
thenVVcp
vector
that 4.27a).
is both
and that
4> is called
they,scalar
Note conversely
that, ifY
= V4>, then
x Vis= a0 (see
Problem

12. SLS 4.33 Answer: See SLS:

'
solenoidal and irrotational.
4.33. Show that if cp(x, y, z) is any solution of Laplace's equation, then Vcp is a vector that is both
Solution and irrotational.
'
solenoidal

By hypothesis, 4> satisfies Laplace's equation V2 4> = 0, that is, V (V4 = O. Then V4> is solenoidal (see
Solution
Problems 4.21 and 4.22).
4.27a, VLaplace's
x (V4 =equation
0, so that
From Problem
By hypothesis,
4> satisfies
V2V4>
4> =is0,also
thatirrotational.
is, V (V4 = O. Then V4> is solenoidal (see
Problems 4.21 and 4.22).

4.34. Give
a Problem
possible4.27a,
definition
of grad
V x (V4
= 0,B.so that V4> is also irrotational.
From
Solution
4.34. Give
a possible definition of grad B.
13. SLS 4.84 Answer: It can be seen by inspection that A = 0, i.e. A is solenoidal.
grad
B as
Assume B = B1i + B2j + B3k. Formally, we can define
Solution
A = 4x + 8y2 z + 3x3 3x 8y2 z 2x3 = x + x3 6= 0, i.e. A is not solenoidal.
14. SLS Assume
4.85 Answer:
B = B1i + B2j + B3k. Formally, we can define grad B as

From identity (iv), (A) = () A + ( A). Here = xyz2 , so = (yz2 , xz2 , 2xyz)

so (A) = (yz2 , xz2 , 2xyz) (2x2 + 8xy2 z, 3x3 y 3xy, 4y2 z2 2x3 z) + xyz2 ( x + x3 )

= (2x2 yz2 + 8xy3 z3 ) + (3x4 yz2 3x2 yz2 ) (8xy3 z3 + 4x4 yz2 ) + ( x2 yz2 + x4 yz2 )
= 0, i.e. B = xyz2 A is solenoidal.

This same result could have been found directly, without using identity (iv).
15. SLS 4.88 (hint: check the table of vector identities) Answer:
Need to show
(
Vso
) =on,0 (solenoidal).
Thethat
quantities
ii,Uij,and
are called unit dyads. (Note that ij, for example, is not the same as ji.)

Define AA=quantity
U and
B =form
V.
of the

The quantities
ii, ij, and so on, are called unit dyads. (Note that ij, for example, is not the same as ji.)
Thus we need
to show that
( A B) = 0.
A quantity of the form

By identity (vi), ( A B) = B ( A) A ( B).

is called(ix),
a dyadic
a12,
...are
components.
But by identity
and
A =the
coefficients
(U ) =all,
0 and
B its
=
(V ) =An0 .array of these nine components in

the form

Therefore
(
V
) = 0, i.e. U all,
Va12,
is solenoidal
(nondivergent).
is called
aU
dyadic
... are its components.
An array of these nine components in

and the coefficients

al2
Note: this
is essentially the same as SLS 4.104, and SLS 4.76 is a special case of this result.
the question
form

al3]
al3]

a22 a23
3
)a32
x2 , (a33
1 a) xz2 ) = ( 0, 3z2 (1 a)z2 , 2( a 2) x ax ) =
16. SLS 4.92 Answer: A = ( axy z , ( a 2al2

a22
a23
( 0, ( a 4)z2 , ( a 4) x ) = 0 when a = 4.
a32 a33

17. SLS 4.104 (hint: check the table of vector identities) Answer:
Need to show that ( A B) = 0 (solenoidal).

By identity (vi), ( A B) = B ( A) A ( B).

But A = 0 and B = 0 (since we are told they are irrotational).

Therefore ( A B) = 0, i.e. A B is solenoidal (nondivergent) if A and B if are irrotational.

z constant, we obtain

-2.? - y3xy
+ f(Y, we
z) obtain
Integrating (1) partially with respect to4> x= keeping
and+z4xz
constant,
where f( y, z) is an arbitrary function of y and z. Similarly, we obtain from (2) and (3)
4> = -2.? - 3xy + 4xz + f(Y, z)
3
where f( y, z) is an arbitrary function 4>of=y -3xy
and z.+"2l
Similarly,
+ 5yzwe+ obtain
g(x, z) from (2) and (3)

and

Integrating (1) partially with respect to x keeping y and

and

l
l

ZPEM2309 Engineering Mathematics 2A MVC Tutorial Problem Set 2 (Sections 3.23.7) ANSWERS page 9 of 12
Note: this question is essentially the same as SLS 4.88, and SLS 4.76 is a special case of this result.
p
18. SLS 4.105 (where r = ( x, y, z) is the position vector and r = x2 + y2 + z2 ) Answer:

16.2

Q18

ZPEM2309 Engineering Mathematics 2A MVC Tutorial Problem Set 2 (Sections 3.23.7) ANSWERS page 10 of 12
19. * SLS 4.102 Answer:

3.7
XXX
3.7

ZPEM2309 Engineering Mathematics 2A MVC Tutorial Problem Set 2 (Sections 3.23.7) ANSWERS page 11 of 12
20. * SLS 4.103 (where r = ( x, y, z) is the position vector and r =

x2 + y2 + z2 is its magnitude, i.e. the distance from

the origin, and express as a function of r, so the last part of the question asks you to find such that ( a) = 0 at
some positive radius a) Answer:

16.2

Q20

21. * SLS 4.106 (In part (a) r = ( x, y, z) is the position vector. For part (b), look at section 3.3 and think about what a

constant curl represents; this will lead you to a result for V which differs from that given in SLS but you should be

able to check that it is correct and that the difference from the SLS answer is an irrotational field, i.e. they are both
valid vector potentials.) Answer:

ZPEM2309 Engineering Mathematics 2A MVC Tutorial Problem Set 2 (Sections 3.23.7) ANSWERS page 12 of 12