Академический Документы
Профессиональный Документы
Культура Документы
1-39)
Dynamic Response of
Second Order Mechanical
Systems with Viscous Dissipation
forces
d2 X dX
M + D + K X = Fext (t )
dt 2
dt
dV
FX = M dt
= Fext ( t ) FD FK
dX
FD = DV = D : Viscous Damping Force
dt
Fk = K X : Elastic restoring Force
d2 X
FI = M a = M : Inertia Force
dt 2
d2 X dX
Since V = d X write the equation of motion as: M + D + K X = Fext (t )
d t2 dt
dt
d
Torques = I dt
= Text ( t ) T D T K
d2 d
I + D + K = Text ( t )
dt 2
dt
+ Initial Conditions in angular velocity and displacement at t=0:
(0) = o and (0) = o
Let the external force be null (Fext=0) and consider the system to have an
initial displacement Xo and initial velocity Vo. The equation of motion for a
2nd order system with viscous dissipation is:
d2 X dX
M + D + K X =0 (1)
dt 2
dt
with initial conditions V (0) = Vo and X (0) = X o
d2 X dX
+ 2 n + n2 X = 0 (2)
dt 2
dt
X (t ) = A e s t (3)
Where A is a constant yet to be found from the initial conditions.
(s 2
+ 2 n s + n2 ) = 0 (5)
s1,2 = n n ( 2 1 )
1/ 2
(6)
The nature of the roots (eigenvalues) clearly depends on the value of the
damping ratio . Since there are two roots, the solution to the differential
equation of motion is now rewritten as:
X (t ) = A1 e s1 t + A2 e s2 t (7)
s1 = i n ; s2 = i n (8)
where i = 1 is the imaginary unit.
iat
Using the complex identity e = cos(at) + i sin(at), renders the
undamped response of the conservative system as:
V02 V0
Where XM = X + 2
and tan ( ) =
0
n2 n X 0
XM is the maximum amplitude response.
Notes:
In a purely conservative system, the motion never dies out, it is harmonic
and periodic.
Motion always oscillates about the equilibrium position X = 0
For an underdamped system, 0 < < 1, the roots are complex conjugate
(real and imaginary parts), i.e.
s1,2 = n i n (1 2 )
1/ 2
(10)
Using the complex identity eiat = cos(at) + i sin(at), write the solution for
underdamped response of the system as:
d = n (1 2 )
1/ 2
where is the system damped natural frequency.
V0 + n X 0
Then C1 = X 0 and C2 = (11.b)
d
Equation (11) representing the system response can also be written as:
X (t ) = e n t X M cos (d t ) (11.c)
C2
where X M = C12 + C22 and tan ( ) =
C1
Note that as t , X(t) 0, i.e. the equilibrium position only if > 0;
0.5
X(t)
0
0 10 20 30 40
-0.5
-1
-1.5
time (sec)
0.5
X(t)
0
0 10 20 30 40
-0.5
-1
-1.5
time (sec)
s1 = n + ( 2 1 ) ; s2 = n ( 2 1 )
1/ 2 1/ 2
(12)
V0 + n X 0
Then C1 = X 0 and C2 = (14)
*
st
The solution form X(t) = A e is no longer valid. For repeated roots, the
theory of ODEs dictates that the family of solutions satisfying the
differential equation is
X (t ) = e n t ( C1 + t C2 ) (16)
0
-0.2 0 5 10 15 20 25 30 35 40
-0.4
-0.6
-0.8
-1
time (sec)
0.6
0.4
0.2
X(t)
0
0 5 10 15 20 25 30 35 40
-0.2
-0.4
-0.6
-0.8
time (sec)
0.5
X(t)
0
0 5 10 15 20 25 30 35 40
-0.5
-1
-1.5
time (sec)
mV _ = mV+ + M xo (1)
xo
(2) & (3) lead to the undamped free response: x(t ) = sin (nt ) = sin (nt )
n
given = 0.012 m as the largest deflection of the spring-mass system.
Hence, xo = n = 0.715 m/s
(c) Ball velocity after impact: from Eq. (1)) (b) Graph of motion
M m m
V+ = V xo = ( 6.26 7.15 ) = 0.892
m s s
(upwards)
V+2
h+ = 2g = 41 mm
1.5
damping ratio=0.1
1 X1
X2
X1+n
0.5
X(t)
0 t1 t1+n
t
0 10 20 30 40
-0.5
Td
-1
time (sec)
and
Or, since d Td = 2 .
X1
=
{ X o e-n t1 cos (d t1 )}
=
e-n t1
= en ( nTd )
X 1+ n {
X o e-n ( t1 + nTd ) cos (d t1 + 2 ) } e -n ( t1 + nTd )
(3)
1 X1 2
= ln = (5)
n X 1+n (1- ) 2 1/ 2
= 1/2
(6),
( 2 ) + 2
2
and for small damping ratios, ~ .
2
The logarithmic decrement method to identify viscous damping ratios
should only be used if:
a) the time decay response shows an oscillatory behavior (i.e. vibration)
with a clear exponential envelope, i.e. damping of viscous type,
b) the system is linear, 2nd order and underdamped,
c) the dynamic response is very clean, i.e. without any spurious signals
such as noise or with multiple frequency components,
d) the dynamic response X(t) 0 as t. Sometimes measurements are
taken with some DC offset. This must be removed from your signal
before processing the data.
e) Use more than just two peak amplitudes separated n periods. In
practice, it is more accurate to plot the magnitude of several peaks in
a semi-log paper and obtain the log-decrement () as the best linear fit
to the following relationship [see below Eq. (7)].
i.e., plot the natural log of the peak magnitudes versus the period
numbers (n=1,2,) and obtain the logarithmic decrement from a straight
line curve fit. In this way you will have used more than just two peaks
for your identification of damping.
Always provide the correlation number (goodness of fit = r2) for the
linear regression curve (y=ax+b), where y=ln(X) and x=n as variables
x 1
ln o = ln
1 1.0
= = 0.2231
n x2 2 0.64
2
= ; = ~ = 0.035
1 2 2 + 4 2 2
underdamped system with 3.5% of critical damping.
2 rad
Damped natural frequency, d = = 10.053
Td sec
d rad
Natural frequency, n = = 10.059
1 2 sec
d) Equivalent mass of system:
from n = K M
K 200 lb/in
M= = 2
= 1.976 lb/ sec /in
n 2 10.0592 1sec2
EXAMPLE
x 4.8
(a) logarithmic decrement = n = n = 1.1631
o
x
1 1.5
Note that the physical damping coefficient (D) does not change whether car is
loaded or not, but does change.
D
Damping ratio when car is empty e =
2 K M empty
The ratio
M empty M full 1/ 2
35,000
= e = = 0.182
e M full M empty 8,000
e 0.381
AE lb 8000 lb lb sec 2
The cable stiffness is just K = = 20, 000 and M = = 20.70 , and
L in 386 in/sec 2 in
rad
n = 31.08
sec 2
Xo
The motion [soln. of (1)] is: X (t ) = sin (n t). (2)
n
With
Xo
maximum Dynamic Displacement is: X d = (3)
n
W Mg 8000 lb
XS = = = 0.4 in
K L 20000 lb/in
The figure below depicts the motion. The system dynamics is governed
by different EOMs if motion is to the left (X decreasing) or to the right (X
increasing) since the friction force changes sign. It is of importance to know
the amplitude decay () every period of motion and also the time elapsed
until the system stops.
M X + K X = F = W (1)
for X < 0
with I.C., X ( 0) = X o , X ( 0) = 0
M X + K X = F = - W (2)
for X > 0
with I.C.,
X = ( X o X ), X = 0
2 2
F F
X (t ) = ( X o ) cos (nt ) + (2)
K K
after period, at , the position of the mass-spring
t= =
2 n
(XL) is:
The sketch below shows the response X(t) for the first period of
motion
M X + K X = - N (4)
for X > 0
with I.C.,
X = ( X o X ), X = 0
2 2
)
F F
X (t ') = ( X L + ) cos (n t ') - (7)
K K
Now, at t =Tn (1 full period), i.e. t = ; i.e. the mass-spring system is
at its rightmost position (XR), as shown in the graph below
F F
X = X R = ( X L + ) cos ( )
2 K K
(8)
2F
XR = XL = X L X
K
XR = X o 2X X 0
2 X=
and note that X ( ) = 0 XR
/2
t
0 /2 t
XL
X
-X 0
Xo to XR = Xo -2 X = Xo -
4F W
where = 2X = =4
K K
d2 X dX
M + D + K X = Fo (21)
dt 2
dt
with initial conditions V (0) = Vo and X (0) = X o
d2 X dX Fo Fo 2
+ 2 n + 2
n X = = n (22)
dt 2
dt M K
The solution of the Non-homogeneous Second Order Ordinary Differential Equation with
Constant Coefficients is of the form (homogenous + particular):
Fo
X (t ) = X H + X P = Ae s t + (23)
K
Where A is a constant found from the initial conditions Note: Xss=Fo / K is
and XP=Fo/K is the particular solution for the step load. equivalent to the static
displacement if the
force is applied very
slowly.
MEEN 617 Notes: Handout 2a Luis San Andrs (2008) 2-30
Substitution of (23) into (22) leads to the CHARACTERISTIC EQUATION
of the system:
(s 2
+ 2 n s + n2 ) = 0 (25)
s1,2 = n n ( 2 1 )
1/ 2
(26)
The nature of the roots (eigenvalues) clearly depends on the value of the
damping ratio . Since there are two roots, the solution to the differential
equation of motion is now rewritten as:
X (t ) = A1 e s1 t + A2 e s2 t + Fo K (27)
For an underdamped system, 0 < < 1, the roots are complex conjugate
( real and imaginary parts), i.e.
s1,2 = n i n (1 2 )
1/ 2
(28)
d = n (1 2 )
1/ 2
where is the system damped natural frequency.
and X ss = Fo K
V0 + n C1
Then C1 = ( X 0 X ss ) and C2 = (30)
d
The larger the viscous damping ratio , the fastest the motions will damp
out to reach the static position Xss.
1.5
X(t)
0.5
0
0 10 20 30 40
time (sec)
For an overdamped system, > 1, the roots of the characteristic eqn. are
real and negative, i.e.
s1 = n + ( 2 1 ) ; s2 = n ( 2 1 )
1/ 2 1/ 2
(34)
* = n ( 2 1 )
1/ 2
where . Do not confuse this term with a frequency
since motion is NOT oscillatory.
V0 + n C1
Then C1 = ( X 0 X ss ) and C2 = (36)
*
Note that as t , X(t) Xss = Fo/K for > 1, i.e. the steady-state
(static) equilibrium position.
The larger the damping ratio , the longer time it takes the
system to reach its final equilibrium position Xss.
For a critically damped system, = 1, the roots are real negative and
identical, i.e.
s1 = s2 = n (37)
X (t ) = e n t ( C1 + t C2 ) + X ss (38)
Note that as t , X(t) Xss = Fo/K for > 1, i.e. the steady-state
(static) equilibrium position.
A critically damped system does not oscillate and it is the fastest to reach
the steady-state value Xss.
1.6
1.4
1.2
X(t)
1
0.8
0.6
0.4
0.2
0
0 5 10 15 20 25 30 35 40
time (sec)
(a) Using free body diagrams: Note that = X/R is a kinematic constraint.
The EOM's are:
M X = F K X D X FG (1)
I = FG R (2)
X
Then from (2) FG = I =I (3) ;
R R2
(3) into (1) gives
I
M + 2 X + D X + K X = F (4)
R
or Using the Mechanical Energy Method:
1 1 1 I
(system kinetic energy): T = M X 2 + I 2 = M + 2 X 2 (5)
2 2 2 R
1
(system potential energy): V = K X2
2
I
(b) define M eq = M + = 3 Kg , K = 27 N/m, D = 7.2 N.s/m
R2
and calculate the system natural frequency and viscous damping ratio:
1/ 2
K rad D
n = = 3 ; = = 0.4 , underdamped system
M
eq sec 2 KM
rad 2
d = n 1- 2 = 2.75 , and Td = = 2.28 sec is the damped period
sec d
of motion
(c) The step response of an underdamped system with I.C.'s X (0) = X (0) = 0 is: