Professional Documents
Culture Documents
In each of the following three problems nd a particular solution to the differential equa
tion. Use complex exponentials where possible.
Problem 1: y(3) + y y + 2y = 2 cos x
Problem 2: y 2y + 4y = e x cos x
Problem 3: y 6y + 9y = e3x
Problem 4: Find the real general solution to the DE
d3 x
x = e2t
dt3
Problem 5: Find a particular solution to the differential equation
y 4y =
1 2x
e + e2x
2
zp =
2eix
2eix
2(1 + 2i )eix
= 3
=
p (i )
i + i2 i + 2
(1 2i )(1 + 2i )
Thus
zp =
2 + 4i
(cos x + i sin x )
5
, and
y p = Re(z p ) =
2
4
cos x sin x
5
5
Problem 2: y 2y + 4y = e x cos x
Solution: complex replacement: z 2z + 4z = e(1+i)x . p(s) = s2 2s + 4; p(1 + i ) =
(1+ i ) x
(1 + i )2 2(1 + i ) + 4 = 2, so z p = e 2 , and thus
y p = Re(z p ) =
1 x
e cos x
2
Problem 3: y 6y + 9y = e3x
Solution: p(s) = s2 6s + 9 = (s 3)2 so y p = cx2 e3x . p(3) = p (3) = 0, p (3) = 2 = 0.
This the generalized ERF gives
1
y p = x2 e3x
2
Problem 4: Find the real general solution to the DE
d3 x
x = e2t
dt3
OCW 18.03SC
r = 1, e2i/3 , e4i/3 = 1, 21 + i 23 , 12 23
So x1 = et , x2 = et/2 cos 23 t , x3 = et/2 sin 23 t
Real solutions: xh = c1 x1 + c2 x2 + c3 x3 so the general solution x = x p + xh is
1 2t
3
3
x = e + c1 et + c2 et/2 cos
t + c3 et/2 sin
t
7
2
2
1 2x
e + e2x
2
Solution: y p = y p,1 + y p,2 where p( D )y p,1 = 12 e2x and p( D )y p,2 = 12 e2x by superposition
principle. p(r ) = r2 4 so p(2) = p(2) = 0. Using the generalized ERF p (r ) = 2r. Thus,
y p,1 =
y p,2 =
1
2
p (2)
1
2
p (2)
xe2x =
1
xe2x = xe2x
8
y p = y p,1 + y p,2 =
1 2x
xe
8
x 2x
e e2x
8
MIT OpenCourseWare
http://ocw.mit.edu
Fall 2011
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
Part II Problems
Problem 1: [Superposition; exponential response formula] Find the general real solution
of:
(a) x + 2x = e3t cos(4t). (You could do this using an integrating factor, but its easier to
write the equation as the real part of a complex-valued equation with exponential right
hand side, and use the ERF.)
..
(b) x + x + 2x = cos t. (Express the sinusoidal solution of this equation in two ways: as
a cos t + b sin t, and as A cos(t ).)
Now open the Mathlet Forced Damped Vibrations. As in Damped Vibrations, the initial
conditions are set by the box at left. There is now a forcing term, A cos(t), and the values
of A and are adjustable by sliders, as are the values of the mass, damping constant, and
spring constant. The force in this system is applied directly to the mass (rather than being
mediated by the spring or the dashpot).
Set m = 1.00, b = 1.00, and k = 2.00.
(c) Start by setting A = 0.00, so you are looking at the homogeneous case. Set the initial
.
condition to x (0) = 0 and x (0) = 1, select the Solution curve only, and measure the pseu
doperiod. (Notice that a slider above the graphing window lets you adjust the vertical
scale.) How well does your measurement agree with what you found in (b)?
(d) Now set A = 1.00 and = 1.00. Select the green Steady State curve. Measure its am
plitude and the value at t = 0. You computed this solution in (b). Compare the computed
.
amplitude and value at t = 0 with what you measured. Also compute x p (0).
(e) Finally, select initial condition x (0) = 0, x (0) = 0: so-called rest initial conditions. (Click
ing on a hashmark gives the exact value.) If you select Steady State, Transient, and Solu
tion, you can see all three: x = x p + xh . Please compute xh .
Problem 2: [Operators, resonance] Use the Resonant ERF to nd the general real solution
of
.
x + x = et
(a) x + 2x = e3t cos(4t). (You could do this using an integrating factor, but its easier to
write the equation as the real part of a complex-valued equation with exponential right
hand side, and use the ERF.)
..
(b) x + x + 2x = cos t. (Express the sinusoidal solution of this equation in two ways: as
a cos t + b sin t, and as A cos(t ).)
Now open the Mathlet Forced Damped Vibrations. As in Damped Vibrations, the initial
conditions are set by the box at left. There is now a forcing term, A cos(t), and the values
of A and are adjustable by sliders, as are the values of the mass, damping constant, and
spring constant. The force in this system is applied directly to the mass (rather than being
mediated by the spring or the dashpot).
Set m = 1.00, b = 1.00, and k = 2.00.
(c) Start by setting A = 0.00, so you are looking at the homogeneous case. Set the initial
.
condition to x (0) = 0 and x (0) = 1, select the Solution curve only, and measure the pseu
doperiod. (Notice that a slider above the graphing window lets you adjust the vertical
scale.) How well does your measurement agree with what you found in (b)?
(d) Now set A = 1.00 and = 1.00. Select the green Steady State curve. Measure its am
plitude and the value at t = 0. You computed this solution in (b). Compare the computed
.
amplitude and value at t = 0 with what you measured. Also compute x p (0).
(e) Finally, select initial condition x (0) = 0, x (0) = 0: so-called rest initial conditions.
(Clicking on a hashmark gives the exact value.) If you select Steady State, Transient, and
Solution, you can see all three: x = x p + xh . Please compute xh .
e(3+4i)t
e ( 3 + 4i ) t
1
=
= e3t ei(4t) ,
p ( 3 + 4i )
5 + 4i
41
..
OCW 18.03SC
1 it
(b) z + z + 2z = eit has solution z p = eit /p(i ); p(i ) = 1 + i + 2 = 1 + i so z p = 1+
ie =
1 i
1
1
2ei/4 ,
2 (cos t + i sin t ) and x p = 2 cos t + 2 sin t. We can also write p (i ) = 1 + i =
so z p =
1 ei/4 eit
2
2 i (t/4)
2 e
= s2 + s +
2 =
2
2
cos(t
4 ). The characteris
7
(s + 12 )2 + 74 withroots 12
, so
the homo
2
7
7
t/2
xh = e
a cos 2 t + b sin 2 t
or xh =
and x p = Re (z p ) =
(c) I measure the rst six zeros to be 2.36, 4.76, 7.12, 9.48, 11.89, 14.19. The successive differ
ences are 2.40, 2.36, 2.36, 2.41, 2.30, with an average of 2.366. My measuredpseudoperiod
is twice this, 4.73. The damped circular frequency of this system is d = 27 so the pseu
4
doperiod is 2
4.7496. Not bad agreement.
d =
7
(d) I measure the amplitude as 0.71. It looks like x p (0) 0.50. Computed amplitude is
2
2
0.707.
Computed
value
is
x
(
0
)
=
p
2
2
.
22 sin(t 4 ) so x p (0) = 22 sin( 4 ) = 12 .
cos( 4 ) =
1
2.
(e) We need to select xh so that xh (0) = x p (0) = 12 and x h (0) = x p (0) = 12 . Its
1
more convenient
to use the
expression
rectangular
for
xh for this. xh (0) = a so a = 2 .
7
7
3
3
t/2 1 cos
2 x h (0) + 1 a = 2 1 1 =
.
x
=
e
t
sin
.
h
2
2
4
2
2
2 t
7
2 7
2 7
Problem 2: [Operators, resonance] Use the Resonant ERF to nd the general real solution
of
.
x + x = et
Solution: p(s) = s + 1 and p(1) = 0, so we are in resonance. p (s) = 1 so the
ERF/Resonant gives
x p = tet .
The general solution is therefore
x = tet + cet .
MIT OpenCourseWare
http://ocw.mit.edu
Fall 2011
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
Part I Problems
Problem 1: a) Find the periodic solution to
x + x = 2 cos t
in amplitude-phase form.
b) What is the gain and the phase lag ?
MIT OpenCourseWare
http://ocw.mit.edu
Fall 2011
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
1
1 i 4
=
1
i =
2 e
2
z p = ei 4
x p =
2 cos(t +
4 )
2
b) Gain = |
1
p (i )
Phase lag
=
1
2
|=
MIT OpenCourseWare
http://ocw.mit.edu
Fall 2011
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
Part I Problems
Problem 1:
Find the solution satisfying the initial conditions:
y y = x2 ,
y(0) = 0,
y (0) = 1
y(0) = 0,
y (0) = 1
Solution: yh = c1 e x + c2 e x .
Try:
y p = a1 x 2 + a2 x + a3
yp = 2a1
x2 = a1 x2 a2 x + 2a1 a3
Thus a1 = 1, a2 = 0, a3 = 2. So
y = c1 e x + c2 e x x 2 2
y(0) = 0 gives us c1 + c2 2 = 0; y (0) = 1 gives us c1 c2 = 1, so c1 = 12 , c2 = 32 .
Thus,
1
3
y = e x + e x x2 2
2
2
OCW 18.03SC
1 2
3x 2x
8
3
8
and B = 14 . Thus,
MIT OpenCourseWare
http://ocw.mit.edu
Fall 2011
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
Part II Problems
Problem 1: Find the general real solution of
d3 x dx
= t2 + 1
dt3
dt
.
= t2 + 1
dt3
dt
.
Solution: characteristic polynomial p(s) = s3 s, p(0) = 0 and so we cant apply
.
..
undetermined coefcients directly. Let u = x, so u u = t2 + 1. Try u = at2 + bt + c, so
..
..
u = 2a and t2 + 1 = u u = at2 bt + (2a c) implies a = 1, b = 0, 2a c = 1 or
c = 3: so u p = t2 3. Then x p is the integral of u p :
1
x p = t3 3t.
3
To solve the homogeneous equation, factor p(s) = s(s 1)(s + 1) so
xh = c1 + c2 et + c3 et . General solution: x = x p + xh .
MIT OpenCourseWare
http://ocw.mit.edu
Fall 2011
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
Part I Problems
Problem 1: Find a particular solution to the DE
..
x + x = t2 + cos(2t 1)
..
x + x = t2 + cos(2t 1)
Solution: x p = x p,1 + x p,2 where p( D ) x p,1 = t2 and p( D ) x p,2 = cos(2t 1), by superposi
tion. Here p( D ) = D2 + 1.
x : try a solution of form x p,1 = At2 + Bt + C. p( D ) x p,1 =
x p,1 + x p,1 = 2A + ( At2 + Bt + C ) = t2 A = 1, B = 0, C = 2. Thus, x p,1 = t2 2.
.. p,1
..
x p,2 : try solution of the form x p,2 = A cos(2t 1) + B sin(2t 1). Then p( D ) x p,2 = x p,2 +
x p,2 = cos(2t 1)
cos(2t 1) A = 13 , B = 0
3
3
yh = e x/2 c1 cos
x + c2 sin
x
2
2
To nd the particular solution, try
y p = a1 xe x + a2 e x
yp = a1 e x ( x + 1) + a2 e x
yp = a1 e x ( x + 2) + a2 e x
2xe x = 3a1 xe x + (3a1 + 3a2 )e x
So a1 = 23 , a2 = 23 .Thus the general solution is
1 3
2
so the
y = e x/2
OCW 18.03SC
3
3
2
c1 cos
x + c2 sin
x + e x ( x 1)
2
2
3
x2 e x
( x 3)
24
MIT OpenCourseWare
http://ocw.mit.edu
Fall 2011
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
Part I Problems
Problem 1: Find a particular solution to the DE
x + x = 2 cos t
MIT OpenCourseWare
http://ocw.mit.edu
Fall 2011
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
x p = Re(z p ) = t cos t.
MIT OpenCourseWare
http://ocw.mit.edu
Fall 2011
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
Part I Problems
Problem 1: A driven spring-mass-dashpot system is modeled by the DE
..
mx + cx + kx = F0 cos t
with m = 1, c = 6, and k = 45. F0 = 50. Find the amplitude A( ) of the response as a
function of the input frequency and nd the frequency which gives the largest system
response. Is this a system for which practical resonance occurs?
..
mx + cx + kx = F0 cos t
with m = 1, c = 6, and k = 45. F0 = 50. Find the amplitude A( ) of the response as a
function of the input frequency and nd the frequency which gives the largest system
response. Is this a system for which practical resonance occurs?
Solution: Using the formulas derived in this session, we have
k
m
1
2
c 2 12
A( ) = F0
A( ) = 50
+ c2 2
+ 36 2
k m 2
45 2
12
12
max =
m
= the frequency which gives practical resonance if
MIT OpenCourseWare
http://ocw.mit.edu
Fall 2011
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
Part II Problems
Problem 1: [Models and complex gain] This problem employs the Mathlet Amplitude
and Phase: Second Order I, which illustrates the steady state sinusoidal system response
of a spring/mass/dashpot system driven through the spring, as discussed in the session
on Gain and Phase Lag. The amplitude of the input signal is 1, so the amplitude of the
system response is the gain. Select b = 0.5, k = 4.00, and = 2.00. Animate the system.
(a) Verify the displayed values of time lag and gain.
(b) The input signal in this system is the position y(t) = cos(t) of the top of the spring
(the cyan box). The system response is the position of the mass (the yellow box), and
we are looking just at the steady state solution x (t). Determine the complex gain for this
system. From it determine the gain of the system, as a function of . Finally determine
tan where is the phase lag.
Problem 2:
[Frequency response] This problem will use the applet Amplitude and
Phase: Second Order I again (as in a previous problem). Set k = 4.00, b = 0.50. These
settings will be in force for parts (a) through (c).
(a) In a previous problem you studied the response of this system when = 2.00. The gain
is pretty large with that setting. Lets investigate the gain for other values of . You can
see a graph of the gain as a function of the input circular frequency by invoking [Bode
Plots]. The top window shows the gain as a function of , and the bottom window graphs
as a function of . Move the slider and verify that these readings correspond to the
graph of the system response at left. You can see a readout of the value of the gain and the
phase lag for the selected value of by rolling the cursor over the relevant window.
From your experimentation with the applet, do you believe that the gain maximal for
= 2.00, or is the practical resonance peak at a different value of ?
In the previous problem mentioned you wrote down a formula for the gain as a function of
(for these values of k and b), g( ). Now nd the value = r which maximizes g( ).
(Hint: itll be easier to minimize the square of the denominator.) Is it at = 2? Finally,
what is the maximal gain?
(b) Experiment to nd the value of giving phase lag as close to 45 as you can. In
previous problem mentioned you also gave a formula for tan . Determine the positive
value of for which the phase lag equals 45 . Compare.
(c) Now invoke the [Nyquist Plot]. This shows the trajectory of the complex gain H ( )
as runs from 0 to . The value of H ( ) corresponding to the selected value of is
shown as a yellow diamond. This means that the length of the yellow strut equals the
Part II Problems
OCW 18.03SC
gain, and the size of the green arc equals the phase lag. Again grab the slider and move
it slowly from 0 to 4. Please submit a sketch of the Nyquist plot with such that ( ) is
as close to 4 as you can get it.
(d) Finally, set = 2 and leave k = 4.00, but adjust the value of b by grabbing the b
slider. What do you observe about the position of the yellow strut in the Nyquist plot?
Try setting k to a different value, and adjust so that the phase lag is close to 2 . Now
vary b and comment on what happens to the phase lag. Please explain this observation
as follows. Write down a formula for the complex gain H ( ) for general values of b, k,
and . What does = 2 say about the complex gain? Finally, what relationship does this
imply about b, k, and ? Does this relationship explain your observation?
lag is t0 = = 4 0.7854.
(b) If we call the complex gain H ( ), then we have H ( ) =
gain is | H ( )| =
(4 2 )2 + 14 2
4
4
=
. The
2
p(i )
4 + 12 i
Im( p(i ))
/2
.
=
Re( p(i ))
4 2
Problem 2:
[Frequency response] This problem will use the applet Amplitude and
Phase: Second Order I again (as in a previous problem). Set k = 4.00, b = 0.50. These
settings will be in force for parts (a) through (c).
(a) In a previous problem you studied the response of this system when = 2.00. The gain
is pretty large with that setting. Lets investigate the gain for other values of . You can
see a graph of the gain as a function of the input circular frequency by invoking [Bode
Plots]. The top window shows the gain as a function of , and the bottom window graphs
as a function of . Move the slider and verify that these readings correspond to the
graph of the system response at left. You can see a readout of the value of the gain and the
phase lag for the selected value of by rolling the cursor over the relevant window.
OCW 18.03SC
From your experimentation with the applet, do you believe that the gain maximal for
= 2.00, or is the practical resonance peak at a different value of ?
In the previous problem mentioned you wrote down a formula for the gain as a function of
(for these values of k and b), g( ). Now nd the value = r which maximizes g( ).
(Hint: itll be easier to minimize the square of the denominator.) Is it at = 2? Finally,
what is the maximal gain?
(b) Experiment to nd the value of giving phase lag as close to 45 as you can. In
previous problem mentioned you also gave a formula for tan . Determine the positive
value of for which the phase lag equals 45 . Compare.
(c) Now invoke the [Nyquist Plot]. This shows the trajectory of the complex gain H ( )
as runs from 0 to . The value of H ( ) corresponding to the selected value of is
shown as a yellow diamond. This means that the length of the yellow strut equals the
gain, and the size of the green arc equals the phase lag. Again grab the slider and move
it slowly from 0 to 4. Please submit a sketch of the Nyquist plot with such that ( ) is
as close to 4 as you can get it.
(d) Finally, set = 2 and leave k = 4.00, but adjust the value of b by grabbing the b
slider. What do you observe about the position of the yellow strut in the Nyquist plot?
Try setting k to a different value, and adjust so that the phase lag is close to 2 . Now
vary b and comment on what happens to the phase lag. Please explain this observation
as follows. Write down a formula for the complex gain H ( ) for general values of b, k,
and . What does = 2 say about the complex gain? Finally, what relationship does this
imply about b, k, and ? Does this relationship explain your observation?
Solution: (a) The square of the denominator is f ( ) = | p(i )|2 = (4 2 )2 + ( 12 )2 . Dif
ferentiate with respect to : f ( ) = 2(4 2 )(2 ) + 12 . Now f ( ) = 0 if either = 0
or = 4 18 . We are interested in the positive value, which is about 1.9685: just less
4
4
64
than 2. The corresonding gain is g(r ) =
=
=4
,
63
1
+ (1 1 )
(4 2 )2 + ( 1 )2
r
64
32
or about 4.0316.
(b) The closest we could get was = 1.70, giving a phase lag of 43 . The phase lag in
this system is the argument of p(i ) = (4 2 ) + 12 i. For 45 we should have real and
1
) so
imaginary parts equal and positive: 4 2 = 21 . 2 + 12 4 = ( + 14 )2 (4 + 16
651
1
1
the desired circular frequency is 4 + 4 + 16 = 4 1.7656.
(c)
OCW 18.03SC
(d) The observation is that the phase lag of 2 seems to occur at a circular frequency which
k
is independent of b. H ( ) =
. Phase lag 2 says that Arg( H ( )) = 2 .
2
(k ) + bi
This is the same as saying that the argument of the denominator is 2 , which says that the
denominator has zero real part and positive imaginary part. This says = k: indepen
dent of b, as observed [and equal to 2 when k = 4, as observed].
MIT OpenCourseWare
http://ocw.mit.edu
Fall 2011
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.