Physics 430: Lecture 4

Quadratic Air Resistance

Dale E. Gary
NJIT Physics Department
Linear Air Resistance Recap-

1
When a projectile moves through the air (or other medium—such as gas
or liquid), it experiences a drag force, which depends on velocity and
acts in the direction opposite the motion (i.e. it always acts to slow the
projectile).
f  f (v) vˆ
 Quite generally, we can write this force as , where the function
f(v) can in general be any function of velocity.
 At relatively slow speeds, it is often a good approximation to write
f (v) bv  cv 2  f lin  f quad
where flin and fquad stand for the linear and quadratic terms, respectively:
f lin bv and f quad cv 2
 For linear air resistance, the equation of motion ismr mg  bv or in terms
of velocity, it is a first-order differential equationmv mg  bv , which has
component equations: mv x  bv x
mv y mg  bv y
 v x b v y b
Equations of this form can be written:  ; 
where v  mg is the terminal velocity. vx m v y  vter m
ter
b
September 10, 2009
Linear Air Resistance Recap-

2
Such equations are said to be in separable form (terms involving v on
one side, and no dependence on v on the other side). Solutions of this
particular form, e.g. dv b dv y b
x
 dt ;  dt
vx m v y  vter m
have exponential solutions:
v x v xo e  t / v y v yo e  t /  vter (1  e  t / )  m / b 
which we then integrate to get x and y positions:

  y (t ) v t  v  v  (1  e )
x(t )  x 1  e  t / ter yo ter
 t /
 We can then combine these equations by eliminating t, to get a single
equation for the trajectory:
v yo  vter
 x 
y x  vter ln 1  
 v xo R, i.e. the value
Finally, we solved this for the range   x for which y = 0,
 v xoof
valid for low air resistance:
 4 v yo 
R Rvac  1  
 Linear air resistance applies only to tiny
3 vprojectiles
ter  or viscous fluids.

September 10, 2009

2.4 Quadratic Air Resistance
 For more normal size projectiles (baseball, cannon ball), it is the
quadratic drag force that applies.
 We are now going to follow exactly the same procedure, but starting
with the quadratic form of the drag
f quad force:
cv 2
 The equation of motion (in terms of v) then becomes: mv mg  cv 2 vˆ
with component equations:mv x  cv 2  c v x2  v y2 v x
mv y mg  cv 2 mg  c v x2  v y2 v y
 As we noted last time, these two equations are coupled, and are
generally not solvable analytically (in terms of equations), although
they can be solved numerically.
 However, we can solve these equations for special cases of either
solely horizontal motion (vy = 0), or solely vertical motion (vx = 0), in
which case the equations
mv become
cv 2 [horizontal motion]
mv mg  cv 2 [vertical motion]
 Let’s look at these one at a time.

September 10, 2009

 Horizontal Motion with Quadratic

Drag-1
As before, we write the equation in separable form (move the terms
dv it isc trivial:
involving v to one side). For the horizontal equation,  dt
v2 m
 This equation is called a non-linear differential equation because
one of the derivatives (the zeroth one, in this case) has a non-linear
dependence. Such equations are significantly harder to solve, in
general. In this case, however, the separable form allows v dvus to
c t
integrate both sides directly vo v2  m 0 dt 
 1 1 ct v
    v(t )  o
to get  vo v  m or 1  t /  , where I have introduced the
m

characteristic time, , in terms of constants: cvo .

 To find the position, we t vagain integrate the velocity equation to get

x(t )  xo   0
dt  vo ln(1  t /  ) [if xo 0]
0 1  t/ 

September 10, 2009

 Horizontal Motion with Quadratic

Drag-2
The final solutions for v(t) and x(t) are:
v
v(t )  o  m / cvo [for quadratic drag]
1 t /
x(t ) vo ln(1  t /  )
 Graphs of these functions are:
vx x
vo

t t
 

They may look similar at first to the linear case, but now the velocityt as
approaches zero much more slowly, like 1/t, so the position does not approach
some limiting value xlike in the linear case, but rather continues to increase
forever. If this sounds impossible, you are right. What really happens is that
as the speed drops, quadratic drag gets swamped by linear drag.
September 10, 2009
Vertical Motion with Quadratic Drag
 We now consider motion solely in the vertical direction, governed by
the equation of motion:mv mg  cv 2 [vertical motion]
 Before we write the vertical equation in separated form, however, we
notice as before that the gravity force mg is balanced by the drag force
cvy2 at terminal velocity mg
vter 
c
after whichv y 0 , i.e. the velocity becomes constant. In terms of vter,
the separated form for the vertical equation is:
dv
 g dt
1  v 2 / vter
2

 In this separated form, we can integrate both sides directly (assuming

vo = 0). v dv  t
0 1  v2 / vter2  g 0 dt 
dx
1  x 2 arctanh x
 Looking at the inside front cover of the book we find
which is what we have if we write x = v/vter. What the heck is arctanh?
September 10, 2009
 Hyperbolic Functions—Problem

2.33(a)
Statement of the Problem:
 The hyperbolic functions cosh z and sinh z are defined as follows:
e z  e z e z  e z
cosh z  and sinh z 
2 2
for any z, real or complex. (a) Sketch the behavior of both functions over
a suitable range of real values of z.

cosh z sinh z
ez ez
e z 2 2
1
2
1 1
2 2

z z
 1
2 ez

2

September 10, 2009

 Hyperbolic Functions—Problem

2.33(b)
Statement of the Problem, cont’d:
 (b) Show that cosh z = cos(iz). What is the corresponding relation for sinh z?
 Solution:
 To do this part, you have to know the relations:
e ix  e  ix e ix  e  ix
cos x  and sin x 
2 2i
 Then the solution is very easy:
e i ( iz )  e  i ( iz ) e  z  e z
cos iz   cosh z
2 2
e i ( iz )  e  i ( iz ) e  z  e z e z  e z
sin iz   i i sinh z
2i 2i 2
 So
sin iz
sinh z   i sin iz
i

September 10, 2009

 Hyperbolic Functions—Problem

2.33(c)
Statement of the Problem, cont’d:
 (c) What are the derivatives of cosh z and sinh z? What about their
integrals?
 Solution:
 The derivatives
d coshare:
z d z
 12(e  e  z )  12 (e z  e  z ) sinh z
dz dz
d sinh z 1 d z
 2 (e  e  z )  12 (e z  e  z ) cosh z
dz dz

 The integrals are equally straightforward:

cosh z dz sinh z
sinh z dz cosh z

September 10, 2009

 Hyperbolic Functions—Problem

2.33(d)
Statement of the Problem, cont’d:
 (d) Show that cosh2 z – sinh2 z = 1.
 Solution: 2
 e z  e z 
 Since cosh z 
2
  14 (e 2 z  2e z e  z  e  2 z )  14 (e 2 z  2  e  2 z )
 2 
2
 e z  e z 
and 2
sinh z    14 (e 2 z  2e z e  z  e  2 z )  14 (e 2 z  2  e  2 z )
 2 

the difference is cosh 2 z  sinh 2 z  12  12 1

September 10, 2009

 Hyperbolic Functions—Problem

2.33(e)
Statement of the Problem, cont’d:
dx
 (e) Show that
the
 1 x 2
arcsinh x . [Hint: One way to do this is to make

substitution x = sinh z.]

 Solution:
 Making that substitution,
cosh z dz we have dx
cosh z dzz dz, so:
= cosh
 1  sinh 2
z

2 2
cosh z  sinh z  sinh z 2
dz  z

but sinh z  x  z arcsinh x

so we have shown that dx

 1 x 2
arcsinh x

September 10, 2009

Vertical Motion with Quadratic Drag

sinh z
tanh z 
Likewise, you can do Problem 2.34, which gives the definition:
cosh z
dx
and leads you through the steps needed to show 1  x 2 arctanh x .

v dv  t
0 1  v2 / vter2
 Now back to our equation:  g 0 dt 

v dv v / vter dx
 The left side is v ter 
0 1  v2 / v 2 0 1 x2
ter

v / vter  v 
vter arctanh x 0 vter arctanh  
 vter 
 gt 
v(t ) vter tanh  
while the right side is just gt, so solving for v, we get  vter 
 To get the position, integrate (see Prob. 2.34) to get
2
vter   gt 
y (t )  ln  cosh   
g   vter 
September 10, 2009
 Example 2.5
 A Baseball Dropped from a High Tower
 Find the terminal speed of a baseball (diameter D = 7 cm , mass m = 0.15
kg). Make plots of its velocity and position for the first six seconds after
it is dropped from a tall tower.
 Solution
 Recall that the constant c can be written c = D2., where = 0.25 Ns2/m2.
So mg (0.15 kg)(9.8 m/s 2 )
vter   2 2 2
35 m/s
c (0.25 Nm /s )(0.07 m)

which is nearly 80 mph.

 You can sketch the velocity and position, or you can calculate it in
Matlab. Here are the plots. As expected, the velocity increases more
slowly than it would in a vacuum under gravity (dashed line), and
approaches vter = 35 m/s (dotted line). As a consequence, the position is
less than the parabolic dependence in vacuum.

September 10, 2009

 Quadratic Drag with Horizontal
and Vertical Motion
 As we said before, the general problem of combined horizontal
and vertical motion yields a set of coupled equations
mv x  c v x2  v y2 v x
mv y  mg  c v x2  v y2 v y
where now we take y positive upward.
 The projectile does not follow the same x and y equations we just
derived, because the drag in the x direction slows the projectile
and changes the drag in the y direction, and vice versa. In fact,
these equations cannot be solved analytically at all! The best
we can do is a numerical solution, but that requires specifying
initial conditions. That means we cannot find the general
solution—we have to solve them numerically on a case-by-case
basis.
 Let’s take a look at one such numerical solution.

September 10, 2009

 Example 2.6
 Trajectory of a baseball
 The baseball of the previous example is thrown with velocity 30 m/s
(about 70 mph) at 50o above the horizontal from a high cliff. Find its
trajectory for the first 8 s of flight and compare with the trajectory in a
vacuum. If the same baseball were thrown on the same trajectory on
horizontal ground, how far would it travel (i.e. what is its horizontal
range)?
 Solution
 First, what are the initial conditions for the position and velocity? For
the position, we are free to choose our coordinate system, so we
certainly would choose xo = 0 and yo = 0 at t = 0. For the velocity, the
statement of the problem gives the initial conditions vxo = vocos  = 19.3
m/s, vyo = vosin  = 23.0 m/s. mUsing
v x  cthese
v x2 values,
v y2 v x we need to write a
program in Matlab that performs a numerical solution to the equations
mv y  mg  c v x2  v y2 v y

for the time range 0 < t < 8 s. We will use the routine ode45 (ode stands
for ‘ordinary differential equation’).

September 10, 2009

 Example 2.6, cont’d
 Solution, cont’d
 First we have to write a function that will be called by ode45. The
heart of that routine is quite simple, just write expressions for the two
equations: Vdot_x = -(c/m)*sqrt(v(1)^2+v(2)^2)*v(1);
Vdot_y = -g-(c/m)*sqrt(v(1)^2+v(2)^2)*v(2);
= [block 2]
 Here, v is the velocity vector, so v(1) is the horizontal velocity vx and
v(2) is the vertical velocity vy. Before these equations will work, we
have to define the constants,
m = 0.15; in kgm. Recall that c = D2.
g, c, and
% Mass of baseball,
g = 9.8; % Acceleration of gravity, in m/s
diam = 0.07; % Diameter of baseball, in m
gamma = 0.25; % Coefficient of drag in air at STP, in Ns^2/m^2
c = gamma*diam^2; = [block 1]

 The last step is to name the function and indicate the inputs and
outputs. ODE45 specifies that the function must have two inputs—the
limits of the independent variable (time
function vdot =in this case), and the array of
quad_drag(t,v)
…
initial conditions (start velocity in x and y in this case).
[block 1]
[block 2]
…
vdot = [vdot_x; vdot_y];

 After saving
[T,V]this function as quad_drag.m,
= ode45('quad_drag',[0 8],[19.3; 23.0]);we call ODE45 with

September 10, 2009


Example 2.6, cont’d
Solution, cont’d
 The arrays T and V that are returned are the times and x and y
velocities, but what we need is the trajectory, i.e. the x and y positions.
For those, we have to integrate the velocities. There is probably an
elegant way to do this in Matlab, but I wrote a simple (and rather
inaccurate) routine to do that, given the T and V arrays:
function y = int_yp(t,yp)
n = length(t);
y = yp;
y(1,:) = [0 0];
for i=1:n-1
dt = t(i+1)-t(i);
dy = yp(i,:)*dt;
y(i+1,:) = y(i,:)+dy;
end
y = y(1:n-1,:);
 Save this as int_yp.m, and then call it by
pos = int_yp(T,V);

which returns the position array [pos(1,:) is x, pos(2,:) is y].

 All that remains is to plot the trajectory ( i.e. pos(1,:) vs. pos(2,:) ).
plot(19.3*T,23.0*T-4.9*T.^2); % Plot vacuum case
hold on
plot(pos(:,1),pos(:,2),'color','red'); % Overplot quadratic drag case
hold off

September 10, 2009

 Example 2.6, cont’d
 Solution, cont’d
 Here is the resulting plot (somewhat improved by labels).
 Note that the range is about
60 m, much shorter than the
equivalent trajectory in a
vacuum.
 Note also that the baseball
does not reach quite as high
as in a vacuum, and reaches
its peak earlier.
 You will be given homework
problems in which I will ask
you to try your hand at such
numerical solutions and plotting.
I will help you learn these very
useful skills, or you can make
use of the Matlab helpers.

September 10, 2009

 2.5 Motion of a Charge in a
Uniform Magnetic Field
 You may recall from Physics 121 that the force on a charge
moving in a magnetic fieldFis
qv B
where q is the charge and B is the magnetic field strength. The
equation of motion then becomes
mv qv B
which is a first-order differential equation in v.
 In this type of problem, we are often free to choose our coordinate
system so that the magnetic field is along one axis, say the z-axis:
B (0,0, B )
and the velocity can in general have any directionv (v x , v y , v z ) .
Hence, v B (v y B,  v x B, 0)
mv x qv y B
and the three components of the equation of motionmare: v  qv B
y x

mv z 0

September 10, 2009

 Motion of a Charge in a
Uniform Magnetic Field-2
 This last equation simply says that the component of velocity
along B,
vz = const. Let’s now focus on the other two components, and
ignore the motion along B. We can then consider the velocity as a
two-dimensional vector (vx, vy) = transverse velocity.
v x  v y
 To simplify, we define the parameter  = qB/m:, so the equations of
motion become: v y   v x

 We will take the opportunity provided by these two coupled

equations to introduce a solution based on complex numbers.
 As you should know, a complex number is a number imaginary
part
like z = x + iy,
 =root
where i is the square vx + of
ivy 1. Let us define:
+ iv y
=v
x
 vy
and then plot the value of  as a vector in the real part
complex plane whose components are vx and vy. vx

September 10, 2009

 Motion of a Charge in a
Uniform Magnetic Field-3
 Next, we take the time derivative of :
 v x  iv y v y  iv x  i (v x  iv y )
or
  i
 So the equation in terms of this new relation has the same form
we saw in the previous lecture for linear air resistance, with the
familiar solution   Ae  it

 The only difference is that this time the argument of the

exponential is imaginary, but it turns out that this makes a huge
difference.
 Before we can discuss the solution in detail, however, we need to
introduce some properties of complex exponentials, which we will
do next time.

September 10, 2009