Section 16.2 Line Integrals
Section 16.2
Line Integrals
Goals:
Compute line integrals of multi variable functions.
Compute line integrals of vector functions.
Interpret the physical implications of a line integral.
Multivariable Calculus 1 / 130
Section 16.2 Line Integrals
Types of Integrals
We have integrated a function over
The real number line∫ ba f (x)dx
The plane∫∫D f (x , y)dA
Three space∫∫∫R f (x , y , z)dV
Multivariable Calculus 2 / 130
Section 16.2 Line Integrals
Line Integrals
In this chapter we will integrate a function over a curve (in either two orthree dimensions, though more are possible).
A two-variable function f (x , y) over a plane curve r(t).
Multivariable Calculus 3 / 130
Section 16.2 Line Integrals
Parameterizations and the Line Integral
The naive approach to integrating a function f (x , y) over a curver(t) = x(t)i + y(t)j would be to plug in x(t) and y(t). Now we areintegrating a function of t.
But this is dependent on our choice of parameterizations. x(2t)i + y(2t)jdefines the same curve but moves twice as fast. Integrating thiscomposition would give half the ∆t and half the area:
Multivariable Calculus 4 / 130
Section 16.2 Line Integrals
Integrating Independent of Parameterization
Instead, our variable of integration should measure somethinggeometrically inherent to the curve:
Multivariable Calculus 5 / 130
Section 16.2 Line Integrals
Arc Length
A more geometrically relevant measure is to integrate with respect todistance traveled. The rectangles approximate the (signed) area betweenC and the graph z = f (x , y).
Multivariable Calculus 6 / 130
Section 16.2 Line Integrals
Integrating with Respect to Arc Length
Area ≈∑
f (x∗i , y∗i )√
∆x2 + ∆y2
=∑
f (x(t∗i ), y(t∗i ))
√(∆x
∆t
)2
+
(∆y
∆t
)2
∆t
∆t→0−−−−→∫
f (x(t), y(t))
√(dx
dt
)2
+
(dy
dt
)2
dt
Alternately:
∫f (r(t))|r′(t)|dt
Multivariable Calculus 7 / 130
Section 16.2 Line Integrals
Integrating with Respect to x or y
We can also integrate with respect to change in just x or just y .∑f (x∗i , y
∗i )∆x =
∑f (x(t∗i ), y(t∗i ))
∆x
∆t∆t
∆t→0−−−−→∫
f (x(t), y(t))x ′(t)dt
Multivariable Calculus 8 / 130
Section 16.2 Line Integrals
Comparison to Integrating a Vector Function
Remark
A plane curve C is defined by a vector function r(t), but we should notconfuse integrating r(t) itself with integrating a function f (x , y) over C .These two types of integrals are almost completely unrelated.
We visualize
∫ b
ar(t)dt as the total displacement achieved by
traveling with velocity r(t) (a vector).
We visualize
∫Cf (x , y)ds as the area between the graph of f and the
curve C (a number).
Multivariable Calculus 9 / 130
Section 16.2 Line Integrals
Applications
We defined∫C fds as an area. It can also be useful for integrating any
function that is a rate with respect to distance:
Example
Over varied terrain, if p(x , y) gives the price per mile to build railroadtracks at point (x , y), then
∫C p(x , y)ds gives the total cost to construct a
railroad following C .
Example
Over varied terrain, if f (x , y) gives the fuel consumption per mile traveledat the point (x , y), then
∫C f (x , y)ds gives the total fuel consumption to
travel along C .
Multivariable Calculus 10 / 130
Section 16.2 Line Integrals
Example 1
Let C be the line segment from (0, 0) to (3, 4). Letf (x , y) = x2 + cos(πy). Compute the line integral∫
Cf (x , y)ds.
Multivariable Calculus 11 / 130
Section 16.2 Line Integrals
Parameterizations to Know
In Chapter 13 we learned parameterizations for the following plane curves:
A line segment from A to B
A circle of radius a
The graph of an explicit function y = f (x)
Multivariable Calculus 12 / 130
Section 16.2 Line Integrals
Example 2
What does integrating∫C 1ds compute?
Formula∫C
1ds = arc length× height
= arc length
Multivariable Calculus 13 / 130
Section 16.2 Line Integrals
Example 3
Calculate the arc length of r(t) = (t2 − t)i + 23 (2t)3/2j on the interval
0 ≤ t ≤ 4.
Multivariable Calculus 14 / 130
Section 16.2 Line Integrals
Summary Questions
Why do we convert to a different differential when setting up a lineintegral?
What does ds mean? What is its differential in terms of dt?
How do we compute arc length?
Multivariable Calculus 15 / 130
Section 16.1 Vector Fields
Section 16.1
Vector Fields
Goals:
Recognize real world phenomena that are modeled by vector fields.
Determine the geometric behavior of a vector field from its equation.
Compute line integrals of a vector field over a curve.
Multivariable Calculus 16 / 130
Section 16.1 Vector Fields
The Definition of a Vector Field
Definition
A vector field in R2 is a function that assigns a two-dimensional vectorF(x , y) to each point in R2.
A vector field in R3 is a function that assigns a three-dimensional vectorF(x , y , z) to each point in R3.
Multivariable Calculus 17 / 130
Section 16.1 Vector Fields
Drawing a Vector Field
We draw a vector field by attaching the vectors F(x , y) to the points(x , y) by the tail. For obvious reasons, we only draw these vectors from afinite set of points.
Multivariable Calculus 18 / 130
Section 16.1 Vector Fields
Notation for a Vector Field
A vector field is defined by component functions P and Q (and R inthree dimensions)
F(x , y) = P(x , y)i + Q(x , y)j
or F(x , y) = 〈P(x , y),Q(x , y)〉 .
F(x , y , z) = (0.2x + 0.04y)i + (0.03z − 0.1)j + 0.2 sin(xz)k
Multivariable Calculus 19 / 130
Section 16.1 Vector Fields
Examples of Vector Fields
Wind speed at each point on the ground.
The force exerted by gravity (or magnetism or charge) at each pointin space.
The gradient of a differentiable function.
Multivariable Calculus 20 / 130
Section 16.1 Vector Fields
Example 1
Sketch the two dimensional vector field F(x , y) = y2 i− 1
2 j.
Multivariable Calculus 21 / 130
Section 16.1 Vector Fields
Example 2
How can we visualize the vector field
F(x , y) =x i + y j√x2 + y2
?
Multivariable Calculus 22 / 130
Section 16.1 Vector Fields
Example 3
How can we visualize the vector field
F(x , y) = −y i + x j?
Multivariable Calculus 23 / 130
Section 16.1 Vector Fields
Exercise
Consider the vector field
F =−x i− y j
(x2 + y2)3/2.
1 What direction do the vectors of this field point?
2 How do the lengths depend on the location of (x , y)?
3 Can you sketch the field?
Multivariable Calculus 24 / 130
Section 16.2 Line Integrals
Dot Products and Work
The dot product measures the angle of two vectors, as well as theirmagnitude.
F · s = |F||s| cos θ
This models the work done by a force F on a displacement s, since onlyFproj contributes to work.
W = Fproj · s = F · s
Multivariable Calculus 25 / 130
Section 16.2 Line Integrals
Line Integrals of Vector Fields
The formula W = F · s assumes that F is constant, and the displacement sis along a straight line. A vector field introduces the possibility that F isdifferent at different points. To compute the work done by a vector field,we use an integral.
Definition
The line integral of the vector field F(x , y) over the vector function r(t) isdefined: ∫
CF · dr =
∫F(r(t)) · r′(t)dt
This is sometimes called a work integral.
We’ll see later that line integral of a vector field has applications beyondphysics.
Multivariable Calculus 26 / 130
Section 16.2 Line Integrals
Alternate Notation
If F(x , y) = P(x , y)i + Q(x , y)j we can rewrite this integral without avector operation:
∫C
F(x , y) · dr =
∫C〈P,Q〉 ·
⟨x ′(t), y ′(t)
⟩dt
=
∫CPx ′(t)dt + Qy ′(t)dt
=
∫CPdx + Qdy
Notation
The line integral of a vector field can also be written:∫CP(x , y)dx +
∫CQ(x , y)dy .
Multivariable Calculus 27 / 130
Section 16.2 Line Integrals
Example 4
Suppose an object travels once anticlockwise around the unit circle and isacted on by a force field F(x , y) = y
2 i− 12 j. Does F do positive or negative
work on the object? Calculate the total work done.
Multivariable Calculus 28 / 130
Section 16.2 Line Integrals
Independence of Parameterization
Theorem
If C1 and C2 are two parameterizations of the same curve then∫C1
F(x , y)dr =
∫C2
F(x , y)dr
To prove this, let C1 be given by r1(t). Then there is some function u(t)such that C2 is given by r2(t) = r1(u(t)). The proof is a u-substitution.
This makes physical sense, because work does not care about speed, onlydisplacement. Thus traveling more quickly or slowly along a curve will notchange the total work done.
Multivariable Calculus 29 / 130
Section 16.2 Line Integrals
Summary of Line Integrals
NotationGeometricInterpretation
Calculation∫Cf (x , y) ds Area below the graph
∫ b
a
f (x(t), y(t))√
(x ′(t))2 + (y ′(t))2dt∫Cf (x , y) dx
Area when projectedonto xz plane
∫ b
af (x(t), y(t))x ′(t)dt∫
CF(x , y) · dr Work done by F
∫ b
aF(x(t), y(t)) ·
⟨x ′(t), y ′(t)
⟩dt
or
∫ b
aP(x(t), y(t))x′(t) + Q(x(t), y(t))y′(t) dt∫ b
ar(t) dt
Not a line integral,computes change inposition from velocity
∫ b
ax(t)i + y(t)j dt
Multivariable Calculus 30 / 130
Section 16.2 Line Integrals
Summary Questions
How do we represent a vector field, graphically?
What does a line integral of a vector field measure?
How can we see whether a vector field is doing positive or negativework on a path?
What does dr mean? What is its differential in terms of dt?
Multivariable Calculus 31 / 130
Section 16.3 The Fundamental Theorem for Line Integrals
Section 16.3
The Fundamental Theorem for Line Integrals
Goals:
Use the fundamental theorem to evaluate line integrals ofconservative vector fields.
Determine when a vector field is conservative.
Multivariable Calculus 32 / 130
Section 16.3 The Fundamental Theorem for Line Integrals
Choosing a Path
We asserted previously that two parameterizations of the same curve orvector function yield equal line integrals. However, changing the course ofthe curve will usually change the value of the integral, even if the startingand ending points are left the same.
Multivariable Calculus 33 / 130
Section 16.3 The Fundamental Theorem for Line Integrals
Example 1/Exercise
Consider the vector field F(x , y) = ∇f (x , y), where f (x , y) =x2 + y2
4. If
A = (−4, 0) and B = (4, 0), what is the work done by F traveling from Ato B along:
1 A line segment?
2 A semicircle of radius 4?
Multivariable Calculus 34 / 130
Section 16.3 The Fundamental Theorem for Line Integrals
Line Integrals of Gradient Fields
Gradient fields have the following property.
Theorem (The Fundamental Theorem of Line Integrals)
If F = ∇f and C travels from point A to point B, then∫C
F · dr = f (B)− f (A)
Multivariable Calculus 35 / 130
Section 16.3 The Fundamental Theorem for Line Integrals
Proof of the Fundamental Theorem
F = ∇f =∂f
∂xi +
∂f
∂yj
dr = r′(t)dt =
(dx
dti +
dy
dtj
)dt
∫C
F · dr =
∫ b
a
∂f
∂x
dx
dt+∂f
∂y
dy
dtdt
=
∫ b
a
d
dtf (r(t))dt
= f (r(b))− f (r(a))
= f (B)− f (A)
Multivariable Calculus 36 / 130
Section 16.3 The Fundamental Theorem for Line Integrals
Visualizing the Fundamental Theorem
By our previous calculation, F · r′(t) computes the rate of change off (r(t)) with respect to t. This can be realized as the change in height ofthe graph z = f (x , y).
Multivariable Calculus 37 / 130
Section 16.3 The Fundamental Theorem for Line Integrals
Conservative Vector Fields
Definition
A vector field is conservative if line integrals depend only the endpointsof the curve.
Theorem
For a vector field F = P i + Qj on a simply connected (no holes) domain,the following are equivalent (if one is true, the others are true).
1 F is conservative.
2 F = ∇f for some function f .
3∫C F · dr = 0 for all closed curves (start and end at same pt).
4 Py = Qx
The function f is sometimes called a potential function for F.
Multivariable Calculus 38 / 130
Section 16.3 The Fundamental Theorem for Line Integrals
Proving the Equivalence
Here’s an outline of how we’d prove the theorem.
F conservative F = ∇f
Closed curves integrate to 0 Py = Qx
define f (x , y) =
∫ (x,y)
(0,0)
F · dr
FTLI
compare toconstant curve
combine C1 anda backward C2
fxy = fyxhard
Multivariable Calculus 39 / 130
Section 16.3 The Fundamental Theorem for Line Integrals
Examples of Conservative Vector Fields
There are a couple cases of conservative fields that are easy to recognize:
A constant field F(x , y) = ai + bj.
A sum of the form F(x , y) = P(x)i + Q(y)j.
What is the potential function for each of these?
Multivariable Calculus 40 / 130
Section 16.3 The Fundamental Theorem for Line Integrals
Example 2
1 Is F = (3x2 − 2xy)i + (cos y − x2)j conservative?
2 What is its potential function?
3 If C is a path from (1, 0) to (3, 0), what is∫C F · dr?
Multivariable Calculus 41 / 130
Section 16.3 The Fundamental Theorem for Line Integrals
Exercise
For each vector field, determine whether it is conservative. If it is, find apotential function.
1 F1 = (√xy − y)i + (
√xy − x)j (for x , y ≥ 0)
2 F2 = (ey + 2x)i + (xey − 4y3)j
Multivariable Calculus 42 / 130
Section 16.3 The Fundamental Theorem for Line Integrals
Summary Questions
What does is mean for a vector field to be conservative?
What is the relationship between the gradient and a conservativevector field?
How do we test that a vector field is conservative?
What does the Fundamental Theorem of Line Integrals say?
Multivariable Calculus 43 / 130
Section 16.4 Green’s Theorem
Section 16.4
Green’s Theorem
Goals:
Use Green’s Theorem to replace a line integral with a double integralor vice versa.
Multivariable Calculus 44 / 130
Section 16.4 Green’s Theorem
Conclusions from Line Integrals of Closed Curves
Multivariable Calculus 45 / 130
Section 16.4 Green’s Theorem
Comparison to the Fundamental Theorem of Calculus
The fundamental theorem of calculus says that integrating (adding up insmall pieces) a rate of change on the interval [a, b] gives the total changebetween the boundary points a and b.
∫ b
af ′(x)dx = f (b)− f (a)
We will attempt to find a similar correspondence for two-dimensionaldomains.
Multivariable Calculus 46 / 130
Section 16.4 Green’s Theorem
Green’s Theorem
Theorem
Suppose D is a simply connected (no holes), bounded region and r(t)defines C , a piecewise smooth curve that traces the boundary of Dcounterclockwise. If F = P i + Qj is a vector field, then∫
CF · dr =
∫∫D
(∂Q
∂x− ∂P
∂y
)dA
Multivariable Calculus 47 / 130
Section 16.4 Green’s Theorem
Alternate Line Integral Notation
To avoid mentioning vectors, your textbook uses the notation∫CPdx + Qdy
Thus we can also write Green’s Theorem
Theorem ∫CPdx + Qdy =
∫∫D
(∂Q
∂x− ∂P
∂y
)dA
Multivariable Calculus 48 / 130
Section 16.4 Green’s Theorem
Significance of ∂Q∂x −
∂P∂y
If ∂Q∂x > 0, then the upward work on
the right side of C outweighs theupward work on the left side of C .
If ∂P∂y < 0, then the rightward work on
the bottom of C outweighs therightward work on the top of C .
Multivariable Calculus 49 / 130
Section 16.4 Green’s Theorem
Proof of Green’s Theorem
We approximate the line integral about C by summing line integralsaround ∆x by ∆y rectangles. Notice that the interior edges cancel eachother out. The remaining outer edges approximate C .
∫C
F·dr = lim∆x ,∆y→0
n∑i=1
∫Ci
F·dri
Multivariable Calculus 50 / 130
Section 16.4 Green’s Theorem
Proof of Green’s Theorem
In order to approximate the line integral around a ∆x by ∆y rectangle, welinearize F. We’ll use differential notation. Unless noted otherwise, allfunctions evaluated at (x , y).
L(x + dx , y + dy) =F + dF
=F + Fxdx + Fydy
=[P + Pxdx + Pydy
]i
+[Q + Qxdx + Qydy
]j
Multivariable Calculus 51 / 130
Section 16.4 Green’s Theorem
Proof of Green’s Theorem
Let’s parameterize the edges of a ∆x by ∆y rectangle. For each segment,0 ≤ t ≤ 1.
(x + t∆x)i + y j
(x + t∆x)i + (y + ∆y)j
x i + (y + t∆y)j (x + ∆x)i + (y + t∆y)j
(x , y) (x + ∆x , y)
(x + ∆x , y + ∆y)(x , y + ∆y)
Multivariable Calculus 52 / 130
Section 16.4 Green’s Theorem
Proof of Green’s Theorem
When we replace F with L and compute the line integral, we getconvenient cancellation. Here are the top and bottom segments.
∫ 1
0L(x + t∆x , y) · (∆x i)dt −
∫ 1
0L(x + t∆x , y + ∆y) · (∆x i)dt
= ∆x
[∫ 1
0P + Px(t∆x)dt −
∫ 1
0P + Px(t∆x) + Py (∆y)dt
]= ∆x
∫ 1
0−Py∆ydt
= −Py∆y∆x
Similarly the left and right segments sum to Qx∆x∆y .
Multivariable Calculus 53 / 130
Section 16.4 Green’s Theorem
Proof of Green’s Theorem
Finally we return to our original limit approximation. Note ∆y∆x is thearea of a ∆x by ∆y rectangle, so our expression conforms to the limitdefinition of a double integral.
∫C
F · dr = lim∆x ,∆y→0
n∑i=1
∫Ci
F · dri
= lim∆x ,∆y→0
n∑i=1
(Qx − Py )∆y∆x
= lim∆x ,∆y→0
n∑i=1
(Qx − Py )∆A
=
∫∫D
(Qx − Py )dA
Multivariable Calculus 54 / 130
Section 16.4 Green’s Theorem
Example 1
Let F(x , y) = (y2 − 3x)i + xy j. Let C be the path that travels along theline segment from (2, 4) to (−1, 1) and then back to (2, 4) along theparabola y = x2.
Compute∫C F · dr.
Multivariable Calculus 55 / 130
Section 16.4 Green’s Theorem
Exercise
Let F(x , y) = (3y − ex)i + (2x − sin y)j. Let C be a circle of radius 3traveling counterclockwise once around the origin.
1 Set up the line integral∫C F · dr as a single-variable integral of t.
2 If F were conservative what would the value of this integral be? Is Fconservative?
3 How would you apply Green’s theorem to the integral? What is itsvalue?
4 What would∫C F · dr be if C traveled clockwise instead?
Multivariable Calculus 56 / 130
Section 16.4 Green’s Theorem
Example 2
Let C be a semicircle from (2, 0) to (−2, 0) above the x-axis. Compute∫C
(x2 − y3)dx + (x3 + ey2)dy
Multivariable Calculus 57 / 130
Section 16.4 Green’s Theorem
Steps for Line Integrals
First decide whether you’re taking the line integral of a function: f (x , y)or a vector field: F(x , y).
1 Function f
a Parameterize C and set up the integral, replacing dx , dy , ds with theappropriate differential.
b Evaluate the integral.
2 Vector Field Fa Is F conservative? Use FTLI.b Can you draw the curve C? Is it closed? Use Green’s.c If neither works, parameterize C and set up the integral, replacing dr
with r′(t)dt, and evaluate.
Multivariable Calculus 58 / 130
Section 16.4 Green’s Theorem
Summary
What kind of integrals can we evaluate with Green’s theorem?
Why would we ever want to replace a single integral with a doubleintegral?
Multivariable Calculus 59 / 130
Section 16.5 Curl and Divergence
Section 16.5
Curl and Divergence
Goals:
Compute the curl and divergence of a vector field.
Interpret curl and divergence geometrically.
Multivariable Calculus 60 / 130
Section 16.5 Curl and Divergence
Derivatives of Vector Fields
If we compare the fundamental theorem of calculus to the fundamentaltheorem of line integrals,
f (b)− f (a) =
∫ b
af ′(x)dx
f (B)− f (A) =
∫C∇f · dr
we see that ∇f fills the role of a derivative of the multivariable function fin this context. In this section we try to define some derivative-likeoperations of vector fields.
Multivariable Calculus 61 / 130
Section 16.5 Curl and Divergence
The ∇ Operator
Notation
We define the gradient operator ∇ (“del”), which behaves in some wayslike a vector. Depending on our choice of dimension we can have
∇ =
⟨∂
∂x,∂
∂y
⟩∇ =
⟨∂
∂x,∂
∂y,∂
∂z
⟩
Multivariable Calculus 62 / 130
Section 16.5 Curl and Divergence
The Gradient of a Function
Given a function f (x , y), we can reexamine ∇f in terms of the gradientoperator:
∇f =
⟨∂
∂x,∂
∂y
⟩f
=
⟨∂
∂xf ,
∂
∂yf
⟩=
⟨∂f
∂x,∂f
∂y
⟩
Multivariable Calculus 63 / 130
Section 16.5 Curl and Divergence
Divergence of a Vector Field
The ∇ operator is more exciting when we apply it to vector fields.
Definition
The divergence of a vector field F = P i + Qj at a point (x0, y0) measuresthe extent to which F points away from (x0, y0). The divergence functionis
div F = ∇ · F =∂
∂xP +
∂
∂yQ
Divergence is defined analogously for 3-dimensional vector fields.
Notice that ∇ · F(x0, y0) is a number, not a vector.
Multivariable Calculus 64 / 130
Section 16.5 Curl and Divergence
Recognizing Divergence Visually
It is easiest to recognize divergence when F(x0, y0) is the zero vector.
∇ · F > 0 ∇ · F < 0 ∇ · F ≈ 0
Multivariable Calculus 65 / 130
Section 16.5 Curl and Divergence
Divergence when F(x , y) 6= 0
When F(x0, y0) is not zero, it can be harder to tell whether F is expandingaway from (x0, y0) or accelerating toward it. If F(x0, y0) = v we cansubtract the constant vector v from F .
∇ · F = ∇ · (F− v)
Multivariable Calculus 66 / 130
Section 16.5 Curl and Divergence
Exercise
If F = xz i + xyz j− y2k, compute ∇ · F at (−2, 2, 1). What does it mean?
Multivariable Calculus 67 / 130
Section 16.5 Curl and Divergence
Green’s Theorem
Recall that ∂Q∂x −
∂P∂y measured the amount that a vector field curled around
a point. Green’s theorem related this to the line integral of a curve aroundthat point. This is a special case of our other “derivative,” the curl of F.
Multivariable Calculus 68 / 130
Section 16.5 Curl and Divergence
Curl F
Definition
The curl is defined for a three-dimensional vector field F = 〈P,Q,R〉. It isdenoted ∇× F and computed as follows:
curl F = ∇× F =
∣∣∣∣∣∣i j k∂∂x
∂∂y
∂∂z
P Q R
∣∣∣∣∣∣=
(∂R
∂y− ∂Q
∂z
)i−(∂R
∂x− ∂P
∂z
)j +
(∂Q
∂x− ∂P
∂y
)k
Multivariable Calculus 69 / 130
Section 16.5 Curl and Divergence
Geometric Properties of ∇× F
Notice ∇× F(x0, y0, z0) is a vector.
∇× F(x0, y0, z0) is related to the line integrals of F around the point(x0, y0, z0).
The length indicates how large such integrals can be, as a multiple ofthe area they enclose.
The direction is perpendicular to the plane in which these lineintegrals are maximized, or around which F curls most strongly.
Multivariable Calculus 70 / 130
Section 16.5 Curl and Divergence
Exercise
If F = xz i + xyz j− y2k, compute ∇×F at (−2, 2, 1). What does it mean?
Multivariable Calculus 71 / 130
Section 16.5 Curl and Divergence
Vector Version of Green’s Theorem
Green’s theorem is two dimensional, so we assumeF(x , y , z) = P(x , y)i + Q(x , y)j + 0k. Most of the terms in the curl arezero. Specifically,
∇× F =
(∂Q
∂x− ∂P
∂y
)k.
Theorem (Green’s Theorem)
Suppose D is a simply connected, bounded region in the plane z = 0 andr(t) defines C , a piecewise smooth curve that traces the boundary of Dcounterclockwise. If F = 〈P,Q, 0〉 is a vector field, then∫
CF · dr =
∫∫D
(∇× F) · k dA
Multivariable Calculus 72 / 130
Section 16.5 Curl and Divergence
Composing ∇’s
We can also compose ∇ operators together. Here are some examples
Example
∇2f = ∇ · (∇f ) takes the divergence of the gradient vector field.
Example
∇ · (∇× F) computes the divergence of the curl of F.
Theorem
A vector field G on a simply connected 3-dimensional domain is equal to∇× F for some F, if and only if ∇ · G(x , y , z) = 0 for all (x , y , z).
Multivariable Calculus 73 / 130
Section 16.5 Curl and Divergence
Summary Questions
How do you compute divergence and curl?
How do you interpret divergence geometrically?
On what vector fields can you compute curl? Divergence?
If someone handed you two functions and tells you one is the curl of avector field and the other is the divergence, how could you tell whichis which?
Multivariable Calculus 74 / 130
Section 16.6 Parametric Surfaces and Their Areas
Section 16.6
Parametric Surfaces and Their Areas
Goals:
Parameterize a surface.
Compute tangent vectors to a parametric surface.
Multivariable Calculus 75 / 130
Section 16.6 Parametric Surfaces and Their Areas
Parametric Surfaces
A parametric surface is the set of points defined by two-variableparametric equations, usually in three-space.
r(u, v) = x(u, v)i + y(u, v)j + z(u, v)k
Where u and v are both parameters. Like with curves, we write this as avector function so we can do calculus, but we visualize it as a set of points.
Multivariable Calculus 76 / 130
Section 16.6 Parametric Surfaces and Their Areas
Parameterizing the Graph of an Explicit Function
The graph of an explicit function z = f (x , y) can be parameterized bysubstituting
x = u y = v z = f (u, v)
r(u, v) = ui + v j + f (u, v)k
Multivariable Calculus 77 / 130
Section 16.6 Parametric Surfaces and Their Areas
Parameterizing a Plane
Formula
If a plane p contains the point (x0, y0, z0) and the vectors a and b, then aparametrization of p is
r(u, v) = 〈x0, y0, z0〉+ ua + vb
Multivariable Calculus 78 / 130
Section 16.6 Parametric Surfaces and Their Areas
Parameterizations from Other Coordinate Systems
Another source of parameterizations comes from coordinate systems we’velearned. Constant multiples and constant terms stretch and shift thesurface.
r(u, v) = (a cos u sin v + x0)i + (b sin u sin v + y0)j + (c cos v + z0)k0 ≤ u ≤ 2π 0 ≤ v ≤ π
Multivariable Calculus 79 / 130
Section 16.6 Parametric Surfaces and Their Areas
Exercise
Describe the surfaces with the following parametric equations.
1 r(u, v) = 3 cos ui + 3 sin uj + vk 0 ≤ u ≤ 2π, 0 ≤ v ≤ 5
2 r(u, v) = (3− 3u − 3v)i + (6u + 2v)j + (2− 9v)k
3 r(u, v) = u cos π4 sin v i + u sin π4 sin v j + u cos vk
0 ≤ u ≤ 5, 0 ≤ v ≤ π
Multivariable Calculus 80 / 130
Section 16.6 Parametric Surfaces and Their Areas
Derivatives of a Parametric Surface
Definitions
The partial derivatives ru(u0, v0) and rv (u0, v0) are tangent vectors tothe surface S .
We can use them to produce a linearization of S at r(u0, v0).
L(u, v) = r(u0, v0) + ru(u0, v0)(u − u0) + rv (u0, v0)(v − v0)
Some algebra shows that this is the equation of a tangent plane.
Multivariable Calculus 81 / 130
Section 16.6 Parametric Surfaces and Their Areas
Summary Questions
How do we parameterize a plane?
How do we parametrize the graph z = f (x , y)?
How do we parametrize a sphere or a cylinder?
What is the relationship between the tangent vectors and the tangentplane of a surface?
Multivariable Calculus 82 / 130
Section 16.7 Surface Integrals
Section 16.7
Surface Integrals
Goals:
Understand the geometric significance of the different surfaceintegrals.
Set up and evaluate surface integrals.
Multivariable Calculus 83 / 130
Section 16.7 Surface Integrals
Integrating on Parametric Surfaces
Just like with line integrals, we’d like to find ways of integrating a functionf (x , y , z) on a surface S that do not depend on the choice ofparameterization.
Definition
An integral dS is computed with respect to the geometric area on thesurface. We divide S into regions and let
∆Si be the area of the i th region.
(x∗i , y∗i , z∗i ) be a test point in the i th region.
D be the largest diameter of any of the regions (the longest distancebetween two points in the region).
We then define the surface integral:∫∫Sf (x , y , z)dS = lim
D→0
∑i
f (x∗i , y∗i , z∗i )∆Si
Multivariable Calculus 84 / 130
Section 16.7 Surface Integrals
Computing a Surface Integral
It’s easier to choose subregions defined by a change in u and a change inv . Still, we may not know the area of such a region, so we use our oldtrick and linearize r(u, v) at a point and use the area of the parallelogramgiven by ∆u and ∆v . The area is thus
dS = |ru × rv |dudv
Multivariable Calculus 85 / 130
Section 16.7 Surface Integrals
Example 1∫∫S
1dS computes the area of a surface. Compute the surface area of a
sphere of radius R.
Multivariable Calculus 86 / 130
Section 16.7 Surface Integrals
Exercise
Let L be the surface
r(u, v) = 3 cos ui + 3 sin uj + vk 0 ≤ u ≤ 2π, 0 ≤ v ≤ 5.
Set up (but do not evaluate) the surface integral∫∫Lx2zdS .
Multivariable Calculus 87 / 130
Section 16.7 Surface Integrals
Motivational Question
Suppose water, traveling at 3 meters per second is passing through acircular opening of radius 4m.
1 How much water flows through the opening per second?
2 What if the water is not flowing perpendicular to the circle?
Multivariable Calculus 88 / 130
Section 16.7 Surface Integrals
Motivational Question
If the water is flowing with velocity v and n is normal to the opening Swith length equal to the area of S , then the flow rate is v · n.
Multivariable Calculus 89 / 130
Section 16.7 Surface Integrals
Flux Integrals
3 What if the velocity of the water is not constant, but depends on thelocation where it is measured?
4 What if the opening isn’t a flat shape, but a surface in 3 dimensions?
Multivariable Calculus 90 / 130
Section 16.7 Surface Integrals
Flux Integrals
Definition
The flux integral of F through S is denoted∫∫S
F · dS.
For a parameterization r(u, v) we define dS = (ru × rv )dudv . F · dSmeasures the flow of F through the parallelogram dS .
Multivariable Calculus 91 / 130
Section 16.7 Surface Integrals
Example 2
Let F(x , y , z) = x i− y j be a flow of water and let N be a net given byN = {(x , y , z) : z = x2 − y2, x2 + y2 ≤ 1}.
1 Compute
∫∫N
F · dS.
2 What does the sign of your answer mean?
Multivariable Calculus 92 / 130
Section 16.7 Surface Integrals
Exercises
1 Let L be the surface
r(u, v) = 3 cos ui + 3 sin uj + vk 0 ≤ u ≤ 2π, 0 ≤ v ≤ 5.
Let F(x , y , z) = xy i + z j + exk. Set up, but do not evaluate the fluxintegral:
∫S F · dS.
2 An experimental solar car has solar panels in the shape of the graph
z =√x − y2
4 for z ≥ 0 and x ≤ 16. The direction and intensity of thesunlight is given by the constant vector field: F = c(2i + 3j− 2k). Setup a flux integral to compute the energy collected by the car.
Multivariable Calculus 93 / 130
Section 16.7 Surface Integrals
Orientation
Depending on our choice of parameterization, ru × rv could point in one oftwo normal directions. If a surface has two sides, then choosing a normaldirection defines an orientation of the surface.
Note that in general not all surfaces have two sides. Surfaces without twosides are non-orientable.
Multivariable Calculus 94 / 130
Section 16.7 Surface Integrals
The Unit Normal Vector
We can connect surface integrals to flux integrals via the followingdefinition
Definition
Given a surface S , the unit normal vector n to S at the point P is thenormal vector of length 1 at P whose direction agrees with the orientationof the surface.
Notice that at the point r(u, v), n = ru×rv|ru×rv | so
∫∫S
F·dS =
∫∫F(r)·(ru×rv )dudv =
∫∫F(r)·n|ru×rv |dudv =
∫∫S
F·ndS
This gives us an alternative notation for flux integrals.
Multivariable Calculus 95 / 130
Section 16.7 Surface Integrals
Summary Questions
What are the two kinds of surface integrals? What do they compute?
What expression do we substitute for the differentials dS and dS?
How many different orientations can a connected surface have? Doesthis change if the surface consists of two or more disconnected parts?
Multivariable Calculus 96 / 130
Section 16.8 Stokes’ Theorem
Section 16.8
Stokes’ Theorem
Use Stokes’ Theorem to evaluate integrals.
Multivariable Calculus 97 / 130
Section 16.8 Stokes’ Theorem
Line Integrals that Bound a Surface
Green’s Theorem related a line integral around C to a double integral of aregion bounded by C . In three dimensions, a curve C might bound asurface S . We can attempt to apply the method of Green’s to thissituation.
Multivariable Calculus 98 / 130
Section 16.8 Stokes’ Theorem
Subdividing and Linearizing a Surface
Like in two dimensions, given a subdivision of S into smaller pieces withperimeter curves ri , we have∫
CF · dr =
∑i
∫Ci
F · dri
We approximate the Ci with parallelograms from the linearization of S .This lets us write the line integrals in terms of the parameterization of S .
Multivariable Calculus 99 / 130
Section 16.8 Stokes’ Theorem
Line Integrals Around the Parallelograms
Let s(u, v) be a parameterization of S . We will parameterize the edges ofparallelogram that results from a change of ∆u and ∆v in the linearizationof s(u, v). The domains are all 0 ≤ t ≤ 1.
r′(t)︷ ︸︸ ︷
s + tsu∆u
s + sv∆v + tsu∆u
r(t) = s + tsv∆v s + su∆u + tsv∆v
F + Fut∆u
F + Fut∆u + Fv∆v
F + Fu∆u + Fv t∆vF(r(t)) = F + Fv t∆v
s s + su∆u
s + su∆u + sv∆vs + sv∆v
We approximate F(ri (t)) by L(u, v) = F + Fudu + Fvdv .Multivariable Calculus 100 / 130
Section 16.8 Stokes’ Theorem
Computations
∫Ci
F · dri
≈∫ 1
0(F + Fut∆u) · (su∆u)dt +
∫ 1
0(F + Fu∆u + Fv t∆v) · (sv∆v)dt
−∫ 1
0(F + Fv∆v + Fut∆u) · (su∆u)dt −
∫ 1
0(F + Fv t∆v) · (sv∆v)dt
=−∫ 1
0Fv∆v · (su∆u)dt +
∫ 1
0Fu∆u · (sv∆v)dt
=
∫ 1
0(Fu∆u) · (sv∆v)− (Fv∆v) · (su∆u)dt
=
∫ 1
0(Fu · sv − Fv · su)∆v∆udt
=(Fu · sv − Fv · su)∆v∆u
Multivariable Calculus 101 / 130
Section 16.8 Stokes’ Theorem
Computations II
(Fu · sv − Fv · su) =
[∂P
∂x
∂x
∂u+∂P
∂y
∂y
∂u+∂P
∂z
∂z
∂u
]∂x
∂v
+
[∂Q
∂x
∂x
∂u+∂Q
∂y
∂y
∂u+∂Q
∂z
∂z
∂u
]∂y
∂v
+
[∂R
∂x
∂x
∂u+∂R
∂y
∂y
∂u+∂R
∂z
∂z
∂u
]∂z
∂v
−[∂P
∂x
∂x
∂v+∂P
∂y
∂y
∂v+∂P
∂z
∂z
∂v
]∂x
∂u
−[∂Q
∂x
∂x
∂v+∂Q
∂y
∂y
∂v+∂Q
∂z
∂z
∂v
]∂y
∂u
−[∂R
∂x
∂x
∂v+∂R
∂y
∂y
∂v+∂R
∂z
∂z
∂v
]∂z
∂u
Multivariable Calculus 102 / 130
Section 16.8 Stokes’ Theorem
Computations III
(Fu · sv − Fv · su) =
(∂R
∂y− ∂Q
∂z
)(∂y
∂u
∂z
∂v− ∂z
∂u
∂y
∂v
)+
(∂P
∂z− ∂R
∂x
)(∂x
∂u
∂z
∂v− ∂z
∂u
∂x
∂v
)+
(∂Q
∂x− ∂P
∂y
)(∂x
∂u
∂y
∂v− ∂y
∂u
∂x
∂v
)= (∇× F) · (su × sv )
Multivariable Calculus 103 / 130
Section 16.8 Stokes’ Theorem
Taking a Limit
Using this computation we can see what happens as we let the size of oursubdivisions approach 0.∫
CF · dr =
∑i
∫Ci
F · dri
= lim∆u,∆v→0
∑i
(Fu · sv − Fv · su)∆v∆u
= lim∆u,∆v→0
∑i
(∇× F) · (su × sv )∆v∆u
=
∫∫S
(∇× F) · dS
Multivariable Calculus 104 / 130
Section 16.8 Stokes’ Theorem
Stokes’ Theorem
Theorem
If S is a smooth surface that is bounded by a simple closed boundarycurve C with positive orientation and F is a vector field with continuouspartial derivatives, then∫
CF · dr =
∫∫S
(∇× F) · dS
Positive orientation means the rotation of C and the direction of dS obeythe right hand rule. If the surface is negatively oriented, introduce a minussign.
Multivariable Calculus 105 / 130
Section 16.8 Stokes’ Theorem
Visualizing Stokes’ Theorem
Much like Green’s theorem, Stokes’ theorem is understood as adding upthe extent to which the vector field curls around each point to get thetotal work around the boundary. The dot product measures the extent towhich the overall curl of F takes place in the surface.
Multivariable Calculus 106 / 130
Section 16.8 Stokes’ Theorem
Example 1
Let C be the curve given by r(t) = cos(t)i + sin(t)j + (cos2(t)− sin2(t))k.Let F(x , y , z) = xyk. How does Stokes’ Theorem apply to∫
CF · dr?
Multivariable Calculus 107 / 130
Section 16.8 Stokes’ Theorem
Exercise
Suppose that F is a conservative vector field on R3 and C is a smoothcurve that bounds a positively-oriented surface S .
1 Can the Fundamental Theorem of Line Integrals compute∫C F · dr?
What is the value?
2 Use your characterization of F from 1 to compute ∇× F.
Multivariable Calculus 108 / 130
Section 16.8 Stokes’ Theorem
Example 2
Let F(x , y , z) be a smooth vector field on R3. Let S be a sphere. What is∫∫S(∇× F) · dS?
Multivariable Calculus 109 / 130
Section 16.8 Stokes’ Theorem
Physics Application of Stokes’ Theorem
Faraday’s law of induction says that the change in magnetic field througha surface S induces an electromotive force through a wire on its boundaryC . Physicists measure the induced voltage by integrating the change inmagnetic field dS.
Multivariable Calculus 110 / 130
Section 16.8 Stokes’ Theorem
Exercise
Suppose S is part of the paraboloid z = 16− x2 − y2 above the xy -plane,and C is its boundary. Is there an easier way to compute
∫∫S F · dS?
Multivariable Calculus 111 / 130
Section 16.8 Stokes’ Theorem
Summary Questions
What two types of integrals does Stokes’ Theorem equate?
What does positive orientation mean?
Multivariable Calculus 112 / 130
Section 16.9 The Divergence Theorem
Section 16.9
The Divergence Theorem
Use the Divergence Theorem to evaluate integrals.
Multivariable Calculus 113 / 130
Section 16.9 The Divergence Theorem
Divergence of a Vector Field and Flux Integrals
A vector field with positive divergence expands outward. Let S be aclosed surface. Visually, we expect the flux integral through that surfaceto be related to this divergence.
Multivariable Calculus 114 / 130
Section 16.9 The Divergence Theorem
The Divergence Theorem
Theorem
If E is a simple solid region, S is its boundary surface with outwardorientation, and F is a vector field with continuous partial derivatives.Then ∫∫
SF · dS =
∫∫∫E∇ · FdV
Outward orientation means that the normal vectors ru × rv point out of Erather than into E .
Multivariable Calculus 115 / 130
Section 16.9 The Divergence Theorem
Example 1
Let S be the surface of the cube with vertices (±1,±1,±1), orientedoutward. Let F(x , y , z) = x3i + yz j + exzk. Compute
∫∫S F · dS.
Multivariable Calculus 116 / 130
Section 16.9 The Divergence Theorem
Example 2
Let F = (2x − sin(y))i + (ez − 4y)j +−zk. Let L be given byr(u, v) = 3 cos(u)i + 3 sin uj + vk for 0 ≤ u ≤ 2π and 0 ≤ v ≤ 4. Compute∫∫
LF · dS.
Multivariable Calculus 117 / 130
Section 16.9 The Divergence Theorem
Exercise
Let S be a sphere of radius 5 with outward orientation. LetF(x , y , z) = (x + 3y)i + (x sin(z))j− zk. What does the divergencetheorem tell you about the value of∫∫
SF · dS?
Multivariable Calculus 118 / 130
Section 16.9 The Divergence Theorem
Application: Diffusion
This is a classical situation studied in a partial differential equations course.Suppose a function u(x , y , z , t) gives the density of smoke (or heat) at thepoint (x , y , z) at time t. Suppose F(x , y , z , t) gives the flow of the smoke.Then for any domain D with smooth boundary S we have
d
dt
∫∫∫DudV = −
∫∫S
F · dS
We can reverse the order of differentiation and integration on the left.
We can apply the Divergence Theorem on the right.∫∫∫D
∂u
∂tdV = −
∫∫∫D
(∇ · F) · dV
If this is true for all domains D, the integrands must be equal
∂u
∂t= −∇ · F
Multivariable Calculus 119 / 130
Section 16.9 The Divergence Theorem
Application: Integration by Parts
The Divergence Theorem allows us to derive a multivariable integration byparts which is also used to solve partial differential equations.
Product Rule
If s is a function and F is a vector field then
∇ · (sF) = (∇s) · F + s(∇ · F)
Theorem (Integration by Parts)
If R is a region with boundary ∂R then∫∫∫R
(∇s) · F +
∫∫∫Rs(∇ · F) =
∫∫∂R
sF
Multivariable Calculus 120 / 130
Section 16.9 The Divergence Theorem
The Divergence Theorem in One Dimension
Unlike Green’s and Stokes’ Theorems, the Divergence Theorem can begeneralized to other dimensions
The Fundamental Theorem of Calculus
For a vector field F(x) = P(x)i on a domain [a, b]. We have∇ · F = P ′(x). The outward normal vectors at a and b are −i and i. Sothe Divergence Theorem is∫
[a,b]P ′(x)dx = F(a) · (−i) + F(b) · i = −P(a) + P(b)
Multivariable Calculus 121 / 130
Section 16.9 The Divergence Theorem
The Divergence Theorem in Two Dimension
Two Dimensions
For a closed curve C = (x(t), y(t)) traveling counterclockwise around adomain D, we have the outward normal vector:
n(t) =
∣∣∣∣ i jx ′(t) y ′(t)
∣∣∣∣The Divergence Theorem is
∫∫D∇ · FdA =
∫ b
aF(x(t), y(t)) · n(t)dt.
Multivariable Calculus 122 / 130
Section 16.9 The Divergence Theorem
Summary Questions
What two types of integrals does the Divergence Theorem equate?
What needs to be true about the domain for the theorem to work?
Explain why the Divergence Theorem is more likely to be useful incomputation than Stokes’ Theorem.
Multivariable Calculus 123 / 130
Section 16.10 Summary of Integration Theorems
Section 16.10
Summary of Integration Theorems
Notice the parallels between the integration theorems we have learned.
Multivariable Calculus 124 / 130
Section 16.10 Summary of Integration Theorems
The Fundamental Theorem of Calculus
Given a function f and an interval [a, b], we have
∫ b
af ′(x)dx = f (b)− f (a)
Multivariable Calculus 125 / 130
Section 16.10 Summary of Integration Theorems
The Fundamental Theorem of Line Integrals
Given a function f and a vector function r(t) from A to B, we have
∫ b
a(∇f ) · dr = f (b)− f (a)
Multivariable Calculus 126 / 130
Section 16.10 Summary of Integration Theorems
Green’s Theorem
Given a vector field F and a domain D in R2 bounded by a curve C , we
∫∫D
(∂Q
∂x− ∂P
∂y
)dA =
∫C
F · dr
Multivariable Calculus 127 / 130
Section 16.10 Summary of Integration Theorems
Stokes’ Theorem
Given a vector field F and a smooth surface S in R3 that is bounded by asimple closed boundary curve C , we have
∫∫S
(∇× F) · dS =
∫C
F · dr
Multivariable Calculus 128 / 130
Section 16.10 Summary of Integration Theorems
The Divergence Theorem
Given a vector field F and a region E with (outward oriented) boundarysurface S we have
∫∫∫E∇ · FdV =
∫∫S
F · dS
Multivariable Calculus 129 / 130
Section 16.10 Summary of Integration Theorems
Summary Questions/Exercise
On each side of an integration theorem:
How are the functions related to each other?
How are the domains related to each other?
Multivariable Calculus 130 / 130