Hw5 Sol
Hw5 Sol
Hw5 Sol
PROBLEM 1(P. 18 on page 374) An optics experiment is set up on a vibration isolation table, which may be
modeled as a single degree of freedom system for vertical oscillations. The system has very little damping.
• The floor has considerable vertical vibration at 30Hz. What must be the natural frequency of the table
be in order to reduce the motion of the table and optics experiment by a factor of four, when compared
to the motion of the floor?
• Each time you work on the experiment, you bump the table. This results in motion that takes a long
time to die out. You decide to add 10% damping to the system. Will this increase or reduce the
vibration of the table in response to floor motions at 30 Hz?
SOLUTION
The ground motion is harmonic, and can be described by the following function: ug ( t ) = u%g ei ω t .
The equation of motion in terms of absolute displacements, is m u&& + c u& + k u = c u& g + k u g , where
u ( t ) = u% ei ω t .
u
k c
ug
The transfer function for absolute displacement of the system, due to harmonic ground motion is:
1+ 2i ξ r
H (r ) = , where r = ω / ω n = tuning ratio.
(1 − r 2 ) + 2 i ξ r
u ( t ) = u% g A ( r) ei (ω t −φ )
1 + 4ξ 2 r 2 2 ξ r3
A (r ) = φ = a tan
(1 − r 2 )2 + 4 ξ 2 r 2 1 − (1 − 4ξ 2 )r 2
For ω = 2π ⋅ 30 = 60 π [rad/sec] and for damping ξ = 0, we want to reduce the motion of the table by a
factor of four, i.e.:
1 1 1
A (r ) = = =
(1 − r 2 )2 1− r 2 4
ω ω
∴ r= = 5 ⇒ ωn = = 84.3 [rad/sec]
ωn 5
If we add damping to the system, ξ = 0.10, the amplification function will be:
1 + 4ξ 2 r 2
A (r ) =
(1 − r 2 )2 + 4 ξ 2 r 2
( 5) = 1 + 4 ⋅ 0.1 ⋅ 5
2
A = 0.272 > 0.25
(1 − 5)2 + 4 ⋅ 0.12 ⋅ 5
Therefore, the value of the amplification function for the tuning ratio r 2 = 5 > 2, increased with the addition
of damping to the system.
ξ=0.00
ξ = 0.0 0
6
ξ=0.10
ξ = 0.1 0
4
ξξ=0.20
= 0 .20
1/2
rr=(5)
= √5
0
0 1 2 3 4 5
Tuning Ratio ω / ωn
However, increasing the damping ratio, the transient response due to the bumping of the table, mainly
controlled by the table frequency (ωd ) will decay as follows:
1
N50% ; = 1 cycle
0.10 ⋅ 10
Therefore, the amplitude will decrease 50% every cycle, and for the frequency of the table being 13Hz, i.e.
13 cycles per second, after 1 sec, the transient response due to the table bumping will be (1/2)13 =1.22×10-4.
PROBLEM 2(P. 19 on page 374) A one-story structure that can be modeled as a single degree of freedom
system has a natural frequency of 10Hz. This structure is subjected to an earthquake whose response
spectrum can be idealized by the function
ug 0
S d (ω ) =
( ) ( )
2 2
ω + 4b 2 ω
2
1 − ω0 ω0
in which:
Determine:
SOLUTION
ω 2 ug 0
S a (ω ) = ω 2 S d (ω ) = {1}
2
ω 2 2
2 ω
1 − + 4 b
ω 0 ω0
The ground acceleration will coincide with the absolute acceleration of the structure for an infinitely rigid
structure, i.e.
2
ω
ω 2 ug 0 ω0
u&&% g = lim u&&% (ω ) = lim = u g 0 ω02 lim
ω →∞ ω →∞ 2 ω →∞ 2
ω 2 2
ω 2 2
2 ω 2 ω
1 − + 4b 1 − + 4b
ω 0 ω0 ω 0 ω0
u&&% gmax = u g 0 ω02 = 0.47 g
(20 π ) 2 0.03
S a (20 π ) = (20π )2 S d (20π ) = = 4.908 m /sec2 ; 0.5 g
2
20 π 2 2
2 20 π
1 − + 4 ⋅ 0.25
4 π 4π
log Sv = log Sv
Sv [m/sec]
Sd [m] 0.1
Sv [ m / sec ]
Sa [g]
0.01
0 = 0.03
ugmax u&&gmax
0 = 0.47
0.001
0.01 0.1 1 10 100
Tuning Ratio [ ω / ω0 ]
PROBLEM 3(P. 20 on page 374) Consider the following periodic forcing function:
p (t)
p0
t d
t p
• Sketch the amplitude of the product H p% j , assuming fn = 0.4 Hz, ξ = 0.05 and m = 1kg.
• What is the response at t=0 ? (Consider a finite number N of terms in the Fourier series. How many?)
• Using MATLAB, sketck the response for N = 16, 32, 64 terms.
SOLUTION
t
p ( t ) = p 0 sin 2π = p 0 sin (α t ) {1}
td
2π 2π
α= =
td 5
{2}
2π 2π α αj
∆ω = = = ω j = j ∆ω =
Tp 2 td 2 2
The Fourier coefficients are evaluated by integration over 1 period, i.e. over Tp :
Tp td Tp
p% j = ∫0 p (t ) e dt = ∫ p (t ) e dt + ∫ p (t ) e
−i ω j t −i ω jt −i ω j t
dt =
0 td
td td
1
= ∫ sin (α t ) e ∫0 eiα t − e−iα t e
−i ω j t −i ω j t
dt = dt =
0 2i
1
{3}
i ( α −ω ) t − i (α + ω ) t
1 e −1
td
1 i ( α −ω ) t − i (α + ω ) t −1 e j d j d
2 i ∫0
= e j
−e j
dt = + =
2 i i (α −ω j ) i (α + ω j )
1 ei (α −α / 2⋅ j )t − 1
d
e− i (α +α / 2 ⋅ j ) t − 1 d
= +
2i i (α − α / 2 ⋅ j ) i (α +α / 2 ⋅ j )
2π
± i (α mα / 2 ⋅ j )
We now substitute in {3} for td : e± i (α mα / 2 ⋅ j ) t = e
d α = e± i ( 2 π mπ j)
{4}
e 2 π i = 1 for j = even
Therefore, from equation {4} we have: e± i ( 2π m π j ) = π i {5}
e = − 1 for j = odd
1 1−1 1− 1
p% j j = even = − + =0
j≠2 2 (α − α / 2 ⋅ j ) (α + α / 2 ⋅ j )
1 −1 − 1 −1 − 1 2 α (1 +1 / 2⋅ j ) + (1 −1/2 ⋅ j )
p% j
{6}
= − + = =
j = odd 2 (α −α / 2 ⋅ j ) (α +α / 2⋅ j ) 2 (α −α / 2⋅ j ) ⋅(α + α / 2 ⋅ j )
2 2 5
= = =
α (1 − j / 4 ) π π (1 − j 2 / 4 )
2
(1− j 2 / 4 )
2
1
td
e− i 2α t
td td
1 0 − i 2ω t 1
p% 2 = ∫ e −e dt = j
∫0
e 0
− e − i 2α t
dt = t + =
2i 0 i 2α
d
2i 2 i 0
{7}
1 e −1 1
− 4π
1−1 td 2.5
= td + = td + = =
2i i 2α 2 i i 2α 2 i i
In Figure 1, the Fourier coefficients are plotted as a function of circular frequency ω j , i.e. the Fourier
amplitude spectrum:
1
Note.- For j = 2, we encounter a zero denominator, therefore the case j = 2 will be solved separately.
2.5
1.5
0.5
0
-20 -15 -10 -5 0 5 10 15 20
Frequency [rad/sec]
2.5
Fourier Coefficients |Fj|
1.5
0.5
0
0 1 2 3 4 5 6 7 8 9 10
Frequency [rad/sec]
Figure 1. Fourier Amplitude Spectrum (in bottom Figure, the Fourier coefficients between 0-10 rad/sec are
shown enlarged): Note that FJ = p% j in the plots.
H (ω j ) =
1
{8}
ω j 2 ω j
k 1 − + 2 i ξ
ω n ωn
0.5
0
-10 -8 -6 -4 -2 0 2 4 6 8 10
Frequency [rad/sec]
0.6
0.5
0.4
|Fj*H|
0.3
0.2
0.1
0
-10 -8 -6 -4 -2 0 2 4 6 8 10
Frequency [rad/sec]
Figure 2. Transfer Function | H(ωj ) | and Response Spectrum H (ω j )⋅ p% j : Note that FJ = p% j in the
plots.
∞
H (ω j ) p% j e
1
u (t ) = ∑ − iω t
j
Tp −∞
∞
{9}
H (ω j ) p% j
1
u ( 0) =
Tp
∑
−∞
Observing Figure 1, we can see that the coefficients 0-5 are enough to evaluate an acceptable result for the
response at t = 0. Therefore we have:
H (ω j ) p% j
5
1
u ( 0) ≅
Tp
∑ {10}
−5
( )
% j are plotted. We can readily see that
In Figure 3, the real and imaginary parts of the product H ω j ⋅ p
the imaginary parts of the symmetric coefficients cancel out, and the real parts have identical value,
therefore:
{
∑ Re H (ω j ) p% j } = T1 ∑ 2 Re { H (ω ) p% } =
5 5
1
u ( 0) ≅ j j
Tp −5 p 0
1
= 2 {0.3581 − 0.035 − 0.4476 + 0.0813 } = − 8.64× 10− 3 m
Tp
0.6
0.4
Real Part |Fj*H|
0.2
-0.2
-0.4
-4 -3 -2 -1 0 1 2 3 4
Frequency [rad/sec]
0.6
Imaginary Part|Fj*H|
0.4
0.2
-0.2
-0.4
-4 -3 -2 -1 0 1 2 3 4
Frequency [rad/sec]
( )
% j : Note that FJ = p% j in the plots.
Figure 3. Real and Imaginary Part of the Response Spectrum H ω j ⋅ p
Figure 4 shows the displacement time history of the 1-degree of freedom system, when 16, 32 and 64 terms
of the Fourier series are considered.
It can be readily seen that the contribution of the higher coefficients is minor to the response of the system,
therefore, the result is essentially identical for the cases considered.
0.5
N=16
0
-0.5
-10 -5 0 5 10 15 20
Time [sec]
0.5
N=32
-0.5
-10 -5 0 5 10 15 20
Time [sec]
0.5
N=64
-0.5
-10 -5 0 5 10 15 20
Time [sec]
Figure 4. Response evaluated using N=16,32 and 64 terms of the Fourier Series
PROBLEM 4(P. 21 on page 375) A reciprocating machine that weighs 100,000N is known to develop
vertical harmonic forces having a maximum amplitude of 5000N at its operating speed of 30 Hz, when
attached to a rigid support. In order to limit the vibrations in the building in which the machine is to be
installed, it is to be supported by a spring at each corner of its base. What spring constant (stiffness) is
required to reduce the total harmonic force trasmitted to the building to 1000N?
k
k k
k
SOLUTION
mu&& + 4 k u + m e ( u&& − r Ω2 e i Ω t ) = 0 ⇒
( m + m e ) u&& + 4 k u = me r Ω 2 ei Ωt ⇒ {1}
( m + m e ) u&& + 4 k u = 5000 e iΩ t
We assume that the eccentric mass, m e, is very small compared to the total mass of the machine, therefore:
100000 N
m + me ≈ m = = 10000 kg
10 m /sec 2
− Ω 2 mu% + 4 k u% = me r Ω 2 ⇒
m Ω 2
2
me e r
r Ω2 m ω n
me r Ω2 {2}
u% = = m = H ( Ω) ⇒ A ( Ω) =
4 k − Ω2 m ω n2 − Ω2 Ω 2
2
1 −
ω n
We know that me r Ω 2 = 5000 N and Ω = 60 π rad /sec .
Fs ( t ) = 4 k u( t ) = 4 k u% (Ω ) e i Ω t
{3}
= 4 k u% ( Ω ) = 4 k u% (Ω ) = 1000 N (required)
max
Fs
Therefore, introducing {2} in equation {3}, the required stiffness is calculated as follows:
m Ω 2
2
e r
m ωn
= 4 k u% (Ω ) = 4 k 2 = 1000 N
max
Fs
Ω 2
1 −
ω n
2
me Ω me Ω 2
r r
m ωn m 4k /m me r Ω 2 5000
4k =4k = = = 1000 N
Ω
2
1−
Ω 2
1−
Ω 2
( 60 π )2
1− 4k /m 4k / m 1−
ωn 4 k /10000
( 60π ) ( 60π )
2 2
1− =5 ⇒ = 6 ⇒ k = 14804.41 kN
4 k /10000 4 k /10000