Free Vibration Analysis of A Uniform Can
Free Vibration Analysis of A Uniform Can
J.-S. WV
Department of’ Naval Architecture and Marine Engineering, National Cheng-Kung University. Turnan.
Taiwan 70101, Republic of China
AND
T.-L. LIN
Department of Industrial Safety and H.vgiene, Chia-Nan Junior College of Pharmacy, Tainan Hsien,
Republic of China
The natural frequencies and mode shapes of a uniform cantilever beam carrying any
number of concentrated masses were determined by using an analytical-and-numerical-
combined method (ANC method). The eigenvalue equation was first derived analytically
by using the expansion theorem and then the eigenvalues and eigenvectors were calculated
numerically. In comparison with the general finite element method (FEM), the ANC
method has the advantage that there is no necessity to derive the property matrices of
each beam element and then to develop the overall ones of the entire beam. In comparison
with the transfer matrix method (TMM) the ANC method requires no tedious matrix
multiplication so that some computer time may be saved. The ANC method is also better
than the pure analytical (closed form) method, since, instead of a few special cases, a
uniform beam carrying any number of point masses of various magnitudes or distributions
along the beam length can be easily solved. Besides, by using the ANC method, there is
no difficulty in taking more modes in the mode superposition equation and hence more
higher mode natural frequencies of better accuracy may be obtained.
1. INTRODUCTION
Before the advent of computers, seeking the analytical (closed form) solution was one
of the most popular ways to solve an engineering problem. In comparison with numerical
methods, the closed form solutions have the advantage of elegance and are time-saving.
However, for most practical engineering problems no closed form solutions can be found,
unless they are significantly simplified. Because of this drawback of the analytical (closed
form solution) method, it has gradually been replaced by numerical methods after the
invention of computer, particularly in recent years. One of the greatest shortcomings of
the numerical approach is that it is time-consuming, but because of its powerful ability
for solving the practical problems this drawback seems allowable.
For various (supporting) boundary conditions the closed form solutions of uniform
beams without concentrated mass loads may be found in many books [l-3]. Most of the
closed form solutions have been obtained under the assumption of negligible shear and
rotatory effects. As for the frequently encountered engineering problem of a uniform
cantilever beam carrying a concentrated mass at the free end, the closed form solution
was presented by Laura and his colleagues in 1974 for a rigidly supported cantilever
201
0022-460X/90/020201 + 13 %03.00/O @ 1990 Academic Press Limited
202 J.-S. WU AND T.-L. LIN
beam [4] and in 1975 for an elastically supported one [5]. In 1984 Giirgoze presented a
theory for obtaining the closed form solutions of natural frequencies and mode shapes
of a uniform beam or rod carrying any numbers of point masses and springs [6]. In recent
years, various approximate methods have been presented to solve similar problems [7-91.
Because of the complexity of a beam with more than two point masses or springs, only
the closed form solutions of that carrying a single point mass or spring were presented
explicitly in reference [6]. In other words, it is impractical to obtain the closed form
solution for the natural frequencies and mode shapes of a uniform beam with more than
two point masses or springs by using the theory of reference [6], particularly when the
mode number is greater than 2. These difficulties in applying the theory of reference [6]
may be overcome by using the ANC method introduced here.
For any kind of boundary conditions, the natural frequencies and mode shapes of a
uniform beam with or without point masses or springs can be determined by using
numerical approaches such as the transfer matrix method (TMM) or finite element method
(FEM). Since any continuous system must be replaced by an equivalent discrete systems,
and then such numerical methods can be applied, discrete concentrated masses or springs
can be easily accommodated. In the TMM [lo] a concentrated mass is considered as a
part of the mass of the station closest to the concentrated mass, and in the FEM a
concentrated mass is distributed to the diagonal elements of the finite element mass matrix
[ill.
In this paper the analytical process is first used to formulate the problem of a uniform
cantilever beam carrying any number of concentrated masses, and then a purely numerical
method is employed to solve the eigenvalue problem. Because there is no necessity to
apply either TMM or FEM for deriving the eigenvalue equation some computer time
may be saved. Also, since the eigenvalue problem is solved numerically, tedious analytical
derivations for arriving at explicit mathematical expressions such as those given in
reference [6] are avoided.
In order to confirm the reliability of the theory, most of the results obtained with the
ANC method are checked against exact solutions or results obtained from TMM and
good agreement is achieved. In addition, the influence of various factors on the natural
frequencies of the system is also studied.
2. THEORETICAL ANALYSIS
For a uniform beam carrying p concentrated masses (Figure l), the equation of motion
with shear and rotatory effects ignored is [6]
a4w(x,t) + ~ a2wb,t) +i a2w(x,t)
EI mj 6(x-x,)=0, (1)
iJX4 at2 j=l at”
where E is the Young’s modulus, I is the moment of inertia of the cross-sectional area
of the beam, r3 is the beam mass per unit length, mj is the magnitude of the jth concentrated
mass, w(x, I) is the vertical deflection of the beam at location x and time r, and F( x - x, )
is the Dirac delta function.
Equation (1) can be rewritten in the form
in which the primes (‘) and the overdots (‘) denote the derivatives of M’with respect to
the co-ordinate x and time t, respectively.
Equation (2) is equivalent to the equation of motion of a forced vibrating system with
the inertia1 forces of the p concentrated masses as the excitation forces. According to the
expansion theorem [2] or the mode superposition methodology [3] one may assume that
where W,(x) are the normal mode shapes of the uniform beam without any concentrated
masses and n,(t) are the generalized coordinates.
Substituting equation (3) into equation (2), multiplying the resulting equation by M’,(x ),
and then integrating over the beam length, one obtains
Note that with respect to the concentrated masses mj (j = 1,. . , p) the normal mode
shapes are not orthogonal to each other.
For the mode shapes normalized with respect to the beam mass per unit length, ti,
one has M,, = 1 .O, and equation (6) is reduced to
w, =JK7; (9)
is the rth natural frequency of a uniform beam “without” concentrated masses.
204 J.-S. WU AND T.-L. LIN
where W denotes the natural frequency of the uniform beam “with” concentrated masses.
Substituting equation (10) into equation (7~) and then placing the result into equation
(6) gives
[B’l=$, mj{*(xj)Hw(xj)IT
(14-16)
or
where
[B]=[Z]+[B’]. (19)
Equation (18) is a standard eigenvalue problem, from which one obtains the frequency
equation
The symbol { } in equations (12), (13), (16), (17) and (18) denotes a column matrix.
The determination of the natural frequencies w, (r = 1,. . . , n) of a uniform beam
“without” concentrated masses is given in the Appendix. Based on this, the natural
frequencies f& (s = 1, . . . , n) of the uniform beam “with” concentrated masses may be
obtained from equation (20). The substitution of the values of W, into equation (18) leads
to the determination of the generalized co-ordinate vector { ?j}(“) and, referring to equation
(3), one obtains the approximate mode shapes of the uniform beam “with” concentrated
masses as
or
In order to check the reliability of the theory, the uniform cantilever beam carrying a
single concentrated mass at its free end (see Figure 2) was first investigated. The magnitude
of the concentrated mass was assumed to be WI, = fil = 35564 lbm.
In Table 1 are shown the five lowest natural frequencies W, (i = 1,. . . ,5) obtained by
three different methods, of which those obtained from reference [4] are the exact solutions.
From the table one sees that the values of 0; (i = 1, . . . ,5) obtained by the ANC method
are very close to the exact solutions, particularly those of W, and LQ. Of course, the errors
become larger when the mode number (i) increases. For the three lowest modes, the
values of Gi (i = 1, . . . ,3) obtained from the transfer matrix method (TMM) are also in
close agreement with the exact solutions.
iii,d,E
5
Figure 2. A uniform cantilever beam with a single concentrated
fi = 0.8891 Ibm/in, d = 2 in, E = 30 x 1Ohpsi, I = 40 in, m, = 35.564
mass rn< at the free end. Numerical
Ibm.
values:
TABLE 1
The jive low est natural frequencies W i (i = 1, . . . , 5) of a uniform cantilever beam carry ing
a concentrated mass (m,) at the free end
In Table 2 are shown the three mode shapes, Wi([) (i = 1,3,5), corresponding to the
three natural frequencies &ii(i = 1,3,5) of Table 1; each mode shape was normalized so
that the maximum modal displacement is equal to 1*O.The numbers in column A are the
exact mode shape values obtained from reference [4], and those in column B are the
approximate ones obtained by the present method. From Table 2 one sees that the results
of the present paper are also close to the exact ones, particularly the first and the third
mode shapes.
From these results, it is evident that for the several lowest modes the accuracy of the
method introduced in this paper is reasonable.
In theory, equation (3) is satisfied only when the mode number (n) approaches infinity
(i.e., n + w). However, in practice equation (3) will give satisfactory results with a small
value of n. In order to realize the effect of the mode number (n) on the accuracy of the
natural frequencies Wi, four cases, each with n = 2,3,4, 5, respectively, were studied. The
numerical analysis results are shown in Table 3; the numbers in the parentheses denote
the percentage errors of the approximate value of Gi comparing with the exact solutions.
From Table 3 one sees that the larger the number of modes (n) the smaller the errors
of the approximate natural frequencies c;ii, which agrees with the theoretical prediction.
It is also seen that the error of the ith natural frequency 6i will be small enough so long
as the number of modes is greater than i + 1. Therefore, it is not necessary to take a very
large number of modes in practice.
TABLE 2
The jirst, third and jifth mode shapes corresponding to 6, , W3 and W 5 of Table 1
Mode shapes
TABLE 3
TABLE 4
Influence of mass ratio (m,/ fil) on the naturalfrequencies Gi (i = 1, . . ,5) w ith concentrated
mass (m,) at free end
1 3 5 7 9
Mass ratio, m,/riIl
Figure 3. The influence of mass ratio m,/riil on the natural frequencies W, (i = 1,. ,5).
160
70 -
40 -
zi,
IO-
I I I I
01
2 4 6 8 to
Station numbor al wtbich point moss m, ia loc0Md
Figure 4. The influence of the location of the concentrated mass m, on the natural frequencies Gi (i = 1,2,3).
BEAM W ITH POINT M ASSES 209
TABLE 5
from 1 to 9, the variations of Wj (i = 1,. . . , 5) are shown in Table 4 and Figure 3. Although
thevaluesofC&(i=2,..., 5) decrease slightly in the range of m,/+il= 1-3, the variations
of 6; (i = 2,. . . ,5) are, however, negligible in the range of m,/riil= 3-9.
The numbers in the parentheses of Table 4 show the percentage errors between the
natural frequencies (sj (i = 1, . . . , 5) obtained by the present method (listed in rows A)
and those obtained from reference [4]. It is seen that the mass ratio (m,/tiiZ) does not
affect the accuracy of the present method.
In Table 5 and Figure 4 is shown the influence of the location of a single concentrated
mass (m,) on the three lowest natural frequencies (si (i = 1,2,3). It is seen that the
fundamental natural frequency Oi decreases gradually as the concentrated mass m, is
moved from station No. 2 (near the fixed end) to station No. 10 (at the free end), but
this trend does not apply to the second and third natural frequencies (sz and 6,. The
reason for this is as follows.
For the mode shape corresponding to the first natural frequency (5,, the spring constant
of the cantilever beam is given by k = 3EZ/13; thus the larger the effective length 1 the
smaller the spring constant k and the first natural frequency 6,. The effective length 1
increases when the concentrated mass m, is moved from the fixed end to the free end of
the cantilever beam. For the mode shapes corresponding to the second and third natural
frequencies O2 and G3, however, there are one and two nodes, respectively, between the
fixed end and the free end of the cantilever beam (see Figure Al in the Appendix). Thus
210 J.-S. WU AND T.-L. LIN
the relationship k = 3 EI/ l3 is no longer valid. Of course, no definite trend can be obtained
from the two curves for (3* and W3 as shown in Figure 4.
Allthevaluesof&(i=l,..., 5) shown in Table 1 are for the case of the concentrated
mass m, at the free end of the cantilever beam. For cases in which m, is at other locations
on the cantilever beam, exact values of Wi (i = 1, . . . , n) are not available. Since the values
of (3, (i = 1,2,3) obtained from the transfer matrix method (TMM) are very close to the
exact solutions of reference [4], as one may see from Table 1, the values of Oi (i = 1,2,3)
obtained by using the TMM was used as the basis of the comparisons shown in Table
5. The numbers shown in the parentheses of Table 5 represent the percentage errors
between the values of Wi (i = 1,2,3) obtained by the present method (listed in rows A)
and those obtained by the TMM (listed in rows C). As in the case of the effect of the
mass ratio (m,/fi,I), the locations of the concentrated mass m, do not affect the accuracy
of the present method.
With the magnitude of each point mass given by m,i = fil/ii, where ii denotes the total
number of point masses, and each point mass occupies one station, the influence of the
distribution of point masses on the natural frequencies of the cantilever beam is shown
in Table 6 and Figure 5. Note that A = 1 represents a single point mass located at station
TABLE 6
t:ee end) AS
cs 5.01108
5.01100 52.35136
52.20070 164.10220
163.45510 342.16890
-
-.-.
2’ ?? 1
-e-* -e-e-e
1 , 1 I L
1 2 3 4 5 6 7 8 9 1C
Total number of point mosses, 5
Figure 5. The influence of the distribution of point masses m,, (i = 1,. , ri) on the natural frequencies W,
(i = 1,2,3,4).
No. 10, fi = 2 represents two point masses distributed at station Nos. 10 and 9, fi = 3
represents three point masses at station Nos. 10, 9, and 8,. . . . As shown in Figure 2, the
free end and fixed end are designated as station No. 10 and No. 0, respectively.
In Table 6 the results obtained by the present method (A) are compared with those
obtained by the TMM (C), and it is seen that they are in good agreement. From Figure
5 one sees that when the single lumped mass at the free end is sequentially divided into
2,3,4,. .. point masses and distributed to station Nos. 10,9,8,7,. . . respectively, the
fundamental natural frequency W, of the cantilever beam increases gradually, but this is
not true for the other natural frequencies Ui (i = 2,3,4). This phenomenon may have
something to do with the mode shapes of a cantilever beam (see Figure A.l) as indicated
in section 3.4.
4. CONCLUSIONS
1. Usually, the five lowest natural frequencies wi (i = 1, . . . ,5) of a vibration system
are the most important ones. The analytical-and-numerical-combined method (the ANC
method) can determine the values of Wi (i = 1, . . . ,5) of a cantilever beam carrying any
number of concentrated masses with good accuracy.
2. The larger the number of modes (n) the more accurate is the ANC method. The
errors in the ith natural frequency (si will be small enough as long as the number of
modes used is greater than i + 1.
3. For a cantilever beam carrying a single concentrated mass m, at free end, the
variations of natural frequencies W, (i = 1, . . . ,5) are negligible in the mass ratio range
m,/ rid = 3-9.
212 J.-S. WU AND T.-L. LI?J
REFERENCES
where
From equation (Al) one obtains the natural frequencies OJ, and the corresponding
normal mode shapes #,,(&) as
CO,= (P,l)‘JEZ/ rit14, r= 1,2,3, . . . , (A3)
G,(5) = (l/m)[(cos I%&- cash pJ&) + D,(sinh &I&- sin &&)], (A4)
where
0, = (cos &l+ cash p,I)/(sin &l+ sinh &l), 5=x1& (A% A6)
BEAM WITH POINT MASSES 213
and the values of&l (called the frequency coefficients) are determined from the frequency
equation
Based on these expressions, the lowest five natural frequencies and the corresponding
normal mode shapes of a uniform cantilever beam may be determined. The results are
shown in Table Al and Figure Al, the latter being plotted from the data of Table Al.
Conversion factors for the English units used in this paper are as follows: 1 ft = 0.3048 m;
1 in = 25.4 mm; 1 lbm = 0.45 kg; 1 lbf = 4.45 N; 1 psi = 6894.76 N/m’.
TABLE Al
Coejicients for the low est jive natural frequencies and normal mode shapes sf‘a uniform
cantilever beam
0
/ I 1 I 1 1 I J
0.1 02 0.3 0.4 0.5 O-6 07 00 09 10
(=x/l
Figure Al. The five lowest mode shapes of a uniform cantilever beam (b = J’?d). Mode numbers: -. I;
---. 2; --( 3; ---, 4;. ,5.