Modal Transient Vibration Response of a Cantilever Beam Subjected to Base Excitation
By Tom Irvine Email:
[email protected] January 29, 2013 _______________________________________________________________________________ Normal Modes Consider the cantilever beam in Figure 1. EI,
L
Figure 1.
The governing differential equation for the displacement y(x,t) is
4 y 2 y EI x 4 t 2
(1)
where E I L
is the modulus of elasticity is the area moment of inertia is the length is the mass density (mass/length)
Note that this equation neglects shear deformation and rotary inertia.
1
Separate the dependent variable.
y( x, t ) Y( x)T(t )
(2)
4 Y( x)T( t ) 2 Y( x)T( t ) EI x 4 t 2
(3)
d4 d2 EI T(t ) Y( x) Y( x) T(t ) 4 2 dx dt
(4)
d 4 4 Y( x ) EI dx Y( x )
d 2 2 T(t ) dt T(t )
d 4 Y( x ) EI dx 4 Y( x )
d 2 2 T(t ) dt c2 T(t )
(5)
Let c be a constant
(6)
Separate the time variable. d 2 2 T(t ) dt c2 T(t )
(7)
d2 T(t ) c 2 T(t ) 0 2 dt
2
(8)
Separate the spatial variable.
d 4 4 Y( x ) EI dx c2 Y( x )
(9)
d4 Y( x) c 2 Y( x) 0 EI dx 4
(10)
A solution for equation (10) is
Y( x) a1 sinhx a 2 coshx a 3 sinx a 4 cosx
(11)
dY( x) a1 coshx a 2 sinh x a 3 cos x a 4 sinx dx
(12)
d 2 Y( x) a1 2 sinh x a 2 2 cosh x a 3 2 sin x a 4 2 cos x 2 dx
(13)
d 3 Y( x) a1 3 cosh x a 2 3 sinh x a 3 3 cos x a 4 3 sin x 3 dx
d 4 Y( x) a1 4 sinh x a 2 4 cosh x a 3 4 sin x a 4 4 cos x 4 dx
(14)
(15)
Substitute (15) and (11) into (10).
a14 sinhx a 24 coshx a 34 sinx a 44 cosx c 2 a1 sinh x a 2 cosh x a 3 sin x a 4 cos x 0 EI
(16)
3
4 a1 sinh x a 2 cosh x a 3 sin x a 4 cos x c 2 a1 sinh x a 2 cosh x a 3 sin x a 4 cos x 0 EI
(17) The equation is satisfied if 4 c2 EI
(18)
1/4 c 2 EI
(19)
The boundary conditions at the fixed end x = 0 are
Y(0) = 0
(zero displacement)
dY 0 dx x 0
(zero slope)
(20) (21)
The boundary conditions at the free end x = L are d 2Y dx 2 d 3Y dx 3
0
(zero bending moment)
(22)
0
(zero shear force)
(23)
x L
x L
Apply equation (20) to (11). a2 a4 0
(24)
a4 a2
(25)
4
Apply equation (21) to (12). a1 a 3 0
(26)
a 3 a1
(27)
Apply equation (22) to (13).
a1 sinhL a 2 coshL a 3 sinL a 4 cosL 0
(28)
Apply equation (23) to (14).
a1 coshL a 2 sinhL a 3 cosL a 4 sinL 0
(29)
Apply (25) and (27) to (28).
a1 sinhL a 2 coshL a1 sinL a 2 cosL 0
a1 sinL sinhL a 2 cosL coshL 0
(30) (31)
Apply (25) and (27) to (29).
a1 coshL a 2 sinhL a1 cosL a 2 sinL 0
a1 cosL coshL a 2 sinL sinhL 0
(32) (33)
Form (31) and (33) into a matrix format. sin L sinh L cos L cosh L
cos L cosh L a1 0 sin L sinh L a 2 0
(34)
By inspection, equation (34) can only be satisfied if a1 = 0 and a2 = 0. Set the determinant to zero in order to obtain a nontrivial solution.
sin2 L sinh2 L
2 0
cos L cosh L
5
(35)
sin2 L sinh2 L cos2 L 2 cosL coshL cosh2 L 0 (36)
sin 2 L sinh 2 L cos2 L 2 cosL coshL cosh 2 L 0 (37)
2 2 cosL cosh L 0
(38)
1 cosL coshL 0
(39)
cosL coshL 1
(40)
There are multiple roots which satisfy equation (40). Thus, a subscript should be added as shown in equation (41). (41) cos n L cosh n L 1
The subscript is an integer index. The roots can be determined through a combination of graphing and numerical methods. The Newton-Raphson method is an example of an appropriate numerical method. The roots of equation (41) are summarized in Table 1, as taken from Reference 1.
Table 1. Roots Index
n L
n=1
1.87510
n=2
4.69409
n=3
7.85476
n=4
10.99554
n>5
(2n-1)/2
Note: the root value formula for n > 5 is approximate.
6
Rearrange equation (19) as follows EI c2 n 4
(42)
Substitute (42) into (8).
T( t ) n 4 dt 2
EI T( t ) 0
(43)
EI EI t b 2 cos n 2 t 0 T( t ) b1 sin n 2
(44)
d2
Equation (43) is satisfied by
The natural frequency term n is thus n n2
EI
(45)
Substitute the value for the fundamental frequency from Table 1.
. 187510 2 EI 1 L
(46)
1 3.5156 EI 2 L2
(47)
Y( x) a1 sinhx a 2 coshx a 3 sinx a 4 cosx
(48)
f1
Find the eigenvectors.
a 4 a 2
(49)
a 3 a1
(50)
Y(x) a1 sinhx a 2 coshx a1 sinx a 2 cosx
(51)
Y(x) a1sinhx sinx a 2 coshx cosx
(52)
7
a1sinhL sinL a 2 coshL cosL 0
a 2 a 1
a1 a 2
sinhL sinL coshL cosL
(53)
(54)
coshL cosL sinhL sinL
(55)
coshL cosL sinhx sinx Y( x ) a 2 coshx cosx sinhL sinL
(56)
Eigenvectors with arbitrary scale. Y( x ) cosh n L cos n L sinh x sin x aˆ n cosh n x cos n x n n sinh n L sin n L
(57) Again, the n terms are given in Table 1. The eigenvectors are normalized such that L
0 Yn
2 ( x ) dx 1
(58)
Normalize with respect to mass. The leading coefficient is 1 for each mode. eigenvectors normalized with respect to mass are
Thus the
1 Y1 ( x ) cosh1 x cos1 x 0.73410 sinh 1 x sin 1 x L
(59)
1 Y2 ( x ) cosh 2 x cos 2 x 1.01847 sinh 2 x sin 2 x L
(60)
8
1 Y3 ( x ) cosh3 x cos3 x 0.99922 sinh 3 x sin 3 x L
(61)
1 Y4 ( x ) cosh 4 x cos 4 x 1.00003 sinh 4 x sin 4 x L
(62)
ρ and L are numerical values only. Y is non-dimensional. The units must be consistent, however. Again, the n values for equations (59) through (62) are given in Table 1.
CANTILEVER BEAM MODE SHAPES 3 n=3 n=2 n=1
DISPLACEMENT * SQRT ( L )
2
1
0
-1
-2
-3
0
0.1
0.2
0.3
0.4
0.5
0.6
DISTANCE (X/L)
Figure 2. The first three mode shapes are plotted in Figure 2.
9
0.7
0.8
0.9
1.0
Participation Factors The participation factors for constant mass density are L 0
n Yn ( x ) dx
(63)
The participation factors from a numerical calculation are
1 0.7830 L
(64)
2 0.4339 L
(65)
3 0.2544 L
(66)
4 0.1818 L
(67)
The participation factors are non-dimensional. Effective Modal Mass The effective modal mass is
2 L m ( x ) Y ( x ) dx n 0 m eff , n L 2 0 m(x) Yn (x) dx
(68)
The eigenvectors are already normalized such that
L
0
m( x ) Yn ( x )2 dx 1
(69)
Thus,
L m eff , n n 2 m( x ) Yn ( x )dx 0
2 (70)
The effective modal mass values are obtained numerically.
m eff , 1 0.6131 L
10
(71)
m eff , 2 0.1883 L
(72)
m eff , 3 0.06474 L
(73)
m eff , 4 0.03306 L
(74)
The effective modal mass terms have units of mass.
Response to Base Excitation
y(x, t) w(t)
Figure 3.
The forced response equation for a beam with base motion is taken from Reference 1, page 345. y(x,t) is the relative displacement.
EI
4y x 4
2y t 2
2w
(75)
t 2
where w is the base displacement. The term on the right-hand-side is the inertial force per unit length.
11
y( x , t )
m
Yn (x)T n (t )
(76)
n 1
m 2w Yn ( x )T n ( t ) 2 t n 1
(77)
m m 4 2 2w EI T n ( t ) Yn ( x ) Yn ( x ) T n ( t ) x 4 t 2 t 2 n 1 n 1
(78)
m m d4 d2 d2w EI T n ( t ) Yn ( x ) Yn ( x ) T n ( t ) dx 4 dt 2 dt 2 n 1 n 1
(79)
EI
4 m 2 Y ( x ) T ( t ) n n x 4 n 1 t 2
d4 dx
4
Yn ( x ) n 4 Yn ( x )
m m d2 d2w EI n 4 T n ( t )Yn ( x ) Yn ( x ) T n ( t ) dt 2 dt 2 n 1 n 1
m m d2 d2w EI n 4 T n ( t )Yn ( x ) Yn ( x ) T n ( t ) dt 2 dt 2 n 1 n 1
(80)
(81)
(82)
Multiply each term by Yp ( x ) .
m m d2 d2w EI n 4 T n ( t )Yn ( x )Yp ( x ) Yn ( x )Yp ( x ) T n ( t ) Y (x) 2 2 p dt dt n 1 n 1
12
(83)
Integrate with respect to length. L
m m d2 4 EI T ( t ) Y ( x ) Y ( x ) Y ( x ) Y ( x ) T ( t ) 0 n n n p n p dt 2 n dx n 1 n 1 L d2w Y ( x )dx 0 dt 2 p
(84)
L m L m d2 EI n 4 T n ( t )Yn ( x )Yp ( x )dx Yn ( x )Yp ( x ) T n ( t )dx 0 0 dt 2 n 1 n 1 L d2w Y ( x )dx 0 dt 2 p
(85) m m 2 L L d EI n 4 T n ( t ) Yn ( x )Yp ( x ) dx T n ( t ) Yn ( x )Yp ( x )dx 0 0 n 1 dt 2 n 1
d2w L
dt 2 0
Yp ( x )dx (86)
m 2 L L EI n 4 m d T ( t ) Y ( x ) Y ( x ) dx n 0 2 T n ( t ) 0 Yn ( x )Yp ( x )dx n p n 1 n 1 dt
d2w L
dt 2 0
Yp ( x )dx (87)
13
The eigenvectors are orthogonal such that L
0 Yn (x)Yp (x) dx 0 L
0 Yn (x)Yp (x) dx 1 d2 dt 2
T n (t)
for n ≠ p
(88)
for n = p
(89)
EI n 4 d2w L T n (t) Yp (x)dx dt 2 0
EI
(91)
EI n 4
(92)
n n 2
n 2
d2 dt
(90)
T ( t ) n 2 T n ( t ) 2 n
d2w L
dt 2 0
d2
d2w
dt
dt 2
T ( t ) n 2 T n ( t ) n 2 n
Yp ( x )dx
(93)
(94)
Add a damping term.
d2
d2w 2 T n ( t ) 2 n n T n ( t ) n T n ( t ) n dt 2 dt 2
(95)
(t ) 2 T (t ) 2 T (t ) w T n n n n n n ( t )
(96)
The damped natural frequency d is d n 1 - 2
(97)
14
The time variable can be determined from a ramp invariant digital recursive filtering relationship from Reference 2. Note that Δt is the time step.
Tn , i 2 exp n t cosd t T n , i 1 exp 2 n t T n , i 2 n 2 i 2exp n t cos d t 1 exp n t 2 1 sin d t n t n w 3 m n t d
1
n 2 i 1 2 n t exp n t cos d t 21 exp 2 n t 2 2 1 exp n t sin d t n w 3 m n t d
1
n 2 i 2 2 1 sin d t 2 cos d t n w 2 n t exp 2 n t exp n t 3 d m n t 1
(98)
15
T n , i 2 exp n t cosd t T n , i 1 exp 2 n t T n , i 2 n i exp( n t ) cos(d t ) sin d t 1 n w 2 d m n t 1
n i 1 exp 2 n t 2 exp( n t ) sin d t 1 n w 2 m n t d 1
n i 2 exp( t ) cos( t ) sin t exp 2 t n d d n n w 2 d m n t 1
(99)
, 2 exp t cos t T , T n i n d n i 1 exp 2 n t T n , i 2
exp n t sin d t md t
n w i 2 w i 1 w i 2 (100)
16
The beam relative displacement is y( x , t )
m
Yn (x)T n (t )
(101)
n 1
The beam relative velocity is
y ( x, t )
m
Y n (x)T n (t )
(102)
n 1
The beam relative acceleration is
y( x, t )
m
Y n (x)T n (t )
(103)
n 1
The beam absolute acceleration A is ( y) y(x, t ) A(x, t ) w
(104)
References 1. W. Thomson, Theory of Vibration with Applications, Second Edition, Prentice-Hall, New Jersey, 1981. 2. T. Irvine, Modal Transient Analysis of a System Subjected to an Applied Force via a Ramp Invariant Digital Recursive Filtering Relationship, Revision K, Vibrationdata, 2012. 3. T. Irvine, Steady-State Vibration Response of a Cantilever Beam Subjected to Base Excitation, Revision C, Vibrationdata, 2012.
17
APPENDIX A
Example Consider a beam with the following properties: Cross-Section
Circular
Boundary Conditions
Fixed-Free
Material
Aluminum
Diameter
D
=
0.5 inch
Cross-Section Area
A
=
0.1963 in^2
Length
L
=
24 inch
Area Moment of Inertia
I
=
0.003068 in^4
Elastic Modulus
E
=
1.0e+07 lbf/in^2
Stiffness
EI
=
30680 lbf in^2
Mass per Volume
v
=
0.1 lbm / in^3 ( 0.000259 lbf sec^2/in^4 )
Mass per Length
=
0.01963 lbm/in (5.08e-05 lbf sec^2/in^2)
L
=
0.471 lbm (1.22E-03 lbf sec^2/in)
=
0.05 for all modes
Mass Viscous Damping Ratio
The analysis is performed using Matlab script: continuous_beam_base_accel.m. The normal modes results are given in Table A-1. Again, both the mode shape and participation factor are considered as dimensionless, but they must be consistent with respect to one another. Table A-1. Natural Frequency Results, Fixed-Free Beam Effective Modal Mass ( lbf sec^2/in )
Effective Modal Mass (lbm)
Mode
fn (Hz)
Participation Factor
1
23.86
0.02736
0.00074837
0.289
2
149.52
0.01516
0.00022982
0.089
3
418.67
0.00889
7.9028e-05
0.031
4
820.42
0.00635
4.0361e-05
0.016
18
Now apply a 1 G, 24 Hz sinusoidal base acceleration input to the beam. This case could be solved more properly using Laplace transforms. But it provides a good test of the ramp invariant digital recursive filtering relationship. The relative displacement and absolute acceleration at the free end are shown in Figures A-1 and A-2 respectively.
Figure A-1.
The peak relative displacement is 0.2656 in. This is consistent with the steady-state response in Reference 3.
19
Figure A-2.
The peak acceleration is 15.53 G. Again, this is consistent with the steady-state response in Reference 3.
20