International Journal of Engineering Sciences, 2(6) June 2013, Pages: 250-258

TI Journals
ISSN
2306-6474

International Journal of Engineering Sciences

Vibrational Modes of Thin Flat Circular Plates Calculated

AB STR AC T

Keywords:

This paper gives the results of a study aimed at making an analytic definition of the resonance
frequencies and the modal forms of vibration of a thin flat circular plate fixed at the central hole of
radius r1 and free on the outside circumference of radius r2 . The calculation considers a state of bidimensional tension and assumes the vertical movement of each point of the plate to be identical to
that of the upper or lower point on the median surface. The theory of vibrating thin plates provides
the Dynamic bi-Laplacian Equation which is the analytic representation of the model. The problem
was approached using cylindrical geometry to obtain an explicit formulation of the bi-Laplacian
operator. The Dynamic bi-Laplacian Equation was integrated by the variable separation method.
The paper shows the principal analytic functions obtained by integrating the Dynamic bi-Laplacian
Equation in bi-dimensional cylindrical geometry. These functions can solve many kinds of
structural problems that involve any different boundary conditions of thin flat circular plates. It is a
possibility that they could also allow to analyze any bi-dimensional atomic and nuclear models
based on Electromagnetic Atomic Theory (EAT) [Bellotti, Giuseppe (2012). The Ideas Behind the
Electromagnetic Atomic Theory. Advances in Natural Science, 5(4), 711, DOI: 10.3968/j.ans.
1715787020120504.2014].

Dynamic bi-Laplacian Equation

2013 Int. j. eng. sci. All rights reserved for TI Journals.

1.

Introduction

The resonance frequencies of a thin flat circular plate fixed at the central hole of radius r 1 and free on the outside circumference of radius r2
, were calculated by considering a state of bi-dimensional tension. The theory of vibrating thin plates provides the following differential
equation to the partial derivatives as an analytic representation of the model (Reference 1):
0
D 4 w hw
(1)
where:

Eh 3
12 1 2

E = Youngs module
h = thickness of thin plate
= Poissons coefficient

4 2 2 = is the Laplace operator consecutively applied twice

* Corresponding author.
Email address: [email protected]

Vibrational Modes of Thin Flat Circular Plates Calculated Using the Dynamic bi-Laplacian Equation

251

Internat ional Jour nal of Engineeri ng Science s, 2(6) June 2013

R (r1 ) 0
dR(r1 )
0
dr
r2 M r (r2 ) 0
r2 Q (r2 ) 0
h
2

Mr

where:

rr

dz

is the flexing moment which tends to inflect the disc in the radial direction

h
2

h
2

rz

dz

is the shearing stress.

h
2

This integration makes it possible to define the eigenvalues, and hence, the resonance frequencies of the disc. Since the conditions are
homogeneous, the eigenvalues are obtained by imposing a zero value on the determinant of the linear system of four equations in four
unknowns generated by the boundary conditions. The values of the determinant can be obtained numerically against the frequency. The
intersections with the frequency axis of this function correspond to the resonance frequencies. Finally, it is possible to obtain the
eigenvectors of disc deformations corresponding to each eigenvalue.
Rewriting equation (1), we obtain:
0
4 w w
(3)
where:

D
h

and

w(r , , t ) R(r )( )T (t )

This equation, written using cylindrical coordinates, becomes:

r 2 2
r
t 2
r r

1 1 2 1 I
1
II
II
II

r 2 2 R T R T 2 R T RT 0
r

r
r
r

1 1 I
1 II
2
1 I II
III
II

r 2 R T r R T R T 3 R T 2 R T
r
r
r r r

II

RT 0
1 1 R I II T R II II T 1 R IV T

r 2 r

r2

From this, by dividing the equation by w(r , , t ) R(r )( )T (t ) , we obtain:

1 1

r 2
r r r r

IV
2 R III 1 R II 1 R I
2 II R II 1 R I 1 IV
II T II
R

2
3
2

4
0

4
r R r R r R r R r R r
T
R

(4)

(5)

(6)

(7)

Since the term in curly brackets does not depend on the time and T(t) does not depend on and r, we could put:

T II
2
T

(8)

T (t ) S cos t T sin t

(9)

where - is a constant.
The solution of equation (8) is:
with S, T numeric constants, and this represents the harmonic oscillation, at resonance angular frequency , of the modal form associated
with the angular frequency .
If we change to + 2 , the function
( ) b1 sin b2 cos
(10)
which represents a possible solution of
() ,
and hence the deflection at a generic point of the elastic surface
w(r , , t ) R(r )( )T (t ) , need not change. This condition is translated by putting m N . The case = - m may be ignored
since it implies only a change of sign which can be absorbed within the arbitrary constant. Supposing, then:
( ) b1 sin m b2 cos m
(11)
we obtain:

II m 2

IV

(12)
(13)

Bellotti Giuseppe

252

Int ernational Journal of Engi ne ering Sc iences, 2(6) June 2013

Substituting in equation (7) and bearing in mind equation (8), we thus obtain:

d 4 R 2 d 3 R 1 2 m2 d 2 R 1 2m 2 dR m

dr
dr 4 r dr 3
r2
dr 2
r3
where:

m
r

R R

(14)

With m fixed there are infinite values of which satisfy equation (14) for defined boundary conditions. In order to integrate equation (14)
we use the series integration method with generalized series of the type:

R ( r ) an r s n

(15)

n 0

Deriving equation (15) four times, we obtain:

R I (r ) an s n r s n 1

(16)

n0

R II (r ) an s n s n 1 r s n 2

(17)

n0

R III (r ) an s n s n 1 s n 2 r s n 3

(18)

n0

R IV (r ) an s n s n 1 s n 2 s n 3 r s n 4

(19)

n0

We substitute these generalized series in equation (14) and make equal the terms of equal power. If we assume n = 0 , the left-hand side of
equation (14) starts from the power rs-4 and the right-hand side starts from rs. In the same way, supposing n = 1, n = 2 and n = 3, terms are
generated on the left-hand side with the powers rs-3, r s-2 and rs-1. These terms must be equal to zero, which happens if the coefficients a3, a2,
a1 and a0 of rs-1, rs-2, rs-3 and rs-4 are zero or for particular values of s obtainable by known mathematical procedures, i.e. for n = 0 the
coefficient of rs-4 is equal to zero only if:

a0 s s 1 s 2 s 3 2s s 1 s 2 2m 2 1 s s 1 2m 2 1 s m 2 m 2 4 0

(20)

And this happens if, and only if, a0 = 0 or if the expression in curly brackets is equal to zero, i.e. when:
s=m; s=2-m; s=-m; s=2+m
(21)
If we assume s = m , two arbitrary constants can be introduced: a0 and a2 ; while a1 and a3 must be presumed to be equal to zero. The
generalized series (15) would then only be formed by powers of r where its coefficients are linked to a0 and a2, whereas all the coefficients
of terms of the type rm + (2t+1) for t N are zero. Expressing all the coefficients of powers of the series against a0 and a2 [making equal,
term for term, the coefficients of the series to the left-hand and right-hand sides of equation (14)], we discover that the series which solves
equation (14) are:

t m 1 ! r 4 t
t m ! r 4t
Rm (r ) r m a0 t
a2 r 2 t
a0 S m (r ) a2 Pm (r )
t 0 32 t ! 2t m 1 ! 2t 1 !!

t 0 32 t ! 2t m ! 2t 1 !!

(22)

having considered (-1)!!=1. With the ratio criterion it is possible to check that the above series are convergent and that their convergence
radius is infinite. It is possible to confirm that equation (22) is a solution to equation (14). Let us consider, for example, the series:

S m (r )
t 0

t m ! r 4t m

f m, , t r 4t m

32t t ! 2t m ! 2t 1 !! t 0

(23)

It is easy to obtain S mI , S mII , S mIII and S mIV and substitute these in equation (14). Each term of the resulting series on the left-hand side must
be exactly equal to each term of equal power on the right-hand side. Thus, the following identity is confirmed:
4t m 4t m 1 4t m 2 4t m 3 f m, , t r 4t m 4 2 4t m 4t m 1 4t m 2 f m, , t r 4 t m 4
(24)
2m 2 1 4t m 4t m 1 f m, , t r 4 t m 4 2m 2 1 4t m f m, , t r 4 t m 4 m2 m2 4 f m, , t r 4t m 4

4t m 4

f m, , t 1 r
When t = 1 on the right-hand side, we can obtain the initial term of the series (in rm) which exactly corresponds to the left-hand side where
we have an equal term (in rm). On the left-hand side a residual term (in rm-4) would remain, but this is zero. If we presume s = - m we obtain
solutions which are linearly dependent on equation (22). If we define to be a number which is neither whole nor partially whole and we
assume s = and s = - in the generalized series, we can satisfy the following differential equation:

d 4 R 2 d 3 R 2 2 1 d 2 R 2 2 1 dR

dr
dr 4 r dr 3
r2
dr 2
r3

R R

(25)

four linearly independent series are obtained:

S (r ) r
t 0

t 1 r 4 t
32t t ! 2t 1 !! 2t 1

(26)

Vibrational Modes of Thin Flat Circular Plates Calculated Using the Dynamic bi-Laplacian Equation

253

Internat ional Jour nal of Engineeri ng Science s, 2(6) June 2013

P (r ) r

t 2 r 4t

32 t ! 2t 1 !! 2t 2

(27)

t 0

t 1 r 4t

S (r ) r
t 0

P ( r ) r

(28)

32 t ! 2t 1 !! 2t 1
t

t 2 r 4t

32 t ! 2t 1 !! 2t 2

(29)

t 0

in which is Eulers Gamma function.

R (r ) C1 S (r ) C2 S (r ) C3 P (r ) C4 P (r )

(30)

with C 1 , C2 , C 3 and C4 numeric constants.

2.

Solutions for various values of m

2.1 Case m = 0
In order to obtain two solutions which are linearly independent of S0( r ) and P0( r ) we must solve the following two limits:

S (r ) S (r )

H 0 (r ) lim

(31)

sin

P (r ) P (r )

K 0 (r ) lim

(32)

sin

In order to solve these limits we must remember that:

x 1 x x

(33)

n n 1 !

(34)

lim r lim1 log r

(35)

lim 1 lim e c lim1 c

(36)

lim 1 lim ec lim1 c

(37)

with c = Eulers constant

x x1 x x2 x x3 x xn x n xl x n 1 (1)n 1 xl
l 1

l 1

l 1

n
1

n
x (1) xl
xl
l 1

(38)

When the limits (31) and (32) are carried out we obtain:

H 0 (r )

2t
2
2
t r 4t
1
log(r ) log(r ) t

t 1
l 1 l 32 t !(2t )!(2t 1)!!

(39)

K 0 (r )

2 2
2r 2
r log(r )

(40)

2t 1

l 1

t r 4t

1 log(r ) l 32 t !(2t 1)!(2t 1)!!

With the ratio criterion it is possible to check that these two above series are convergent with infinite convergence radius and that they
solve the differential equation:

d 4 R 2 d 3 R 1 d 2 R 1 dR

R
dr 4 r dr 3 r 2 dr 2 r 3 dr

(41)

obtained from equation (14), supposing m = 0.

R0 r AS 0 r BP0 r CH 0 r FK 0 r

(42)

with A, B, C and F numeric constants to be decided by the boundary conditions. For a thin flat circular plate fixed at the central circular
hole, the boundary conditions are [Reference 1 , equations (6.130) (6.134)]:

Bellotti Giuseppe

254

Int ernational Journal of Engi ne ering Sc iences, 2(6) June 2013

R r1 0
R I r1 0

rM r r r r r D I
2
r

r r2

I
rQ r r r r D II

0
2
r r 2

r r2
since the vibrating plate is not loaded and its deformations are small, then [in accordance with equation (6.134) of reference 1 where Q = 0
] we can assume:

RI r
then the boundary conditions are:

d2R
dR
rM r r r r r 2
0

2
dr r r
dr

2
d 3 R d 2 R 1 dR
rQ r r r r 3 2
0

2
r dr r r
dr
dr
2
Making successive derivatives of equation (42) and substituting in the four boundary conditions, we obtain a homogeneous linear system of
four equations in the four unknowns A, B, C and F. The solution of this system is significant only if the determinant of the coefficients is
equal to zero. Since R0( r ) contains the coefficient =(), we may proceed to the calculation of the determinant against the frequency.
When the frequency f has been chosen, we can calculate:

48 2 f 2 1 2
Eh 2

and thus calculate the numeric coefficient of the unknowns A, B, C and F allowing to calculate the determinant of the coefficients and by
making f vary, we can define the frequency at which it becomes zero. In such way the resonance frequencies are defined. Once A, B, C and
F, have been found (unless there is a constant of proportionality) it is possible to substitute in equation (42) and find the profile of

R0 j r , f j (where j N ) along the radius. In order to obtain the elastic vibrant surface it is necessary to multiply by () and T(t) the

function R0 j r , f j

where f is a resonance frequency.

2.2 Case m =1
Equation (14) becomes:

d 4 R 2 d 3 R 3 d 2 R 3 dR 3

R R
dr 4 r dr 3 r 2 dr 2 r 3 dr r 4

(43)

We already know two solutions:

S1 (r ) r
t 0

t r 4t
32 t ! 2t 1 ! 2t 1 !!
t

t r 4t
32t t ! 2 t 1 ! 2t 1 !!

P1 (r ) 2r 3
t 0

(44)

(45)

We needs to find two more solutions linearly independent of these. Supposing = 1 + with 0 , we obtain from equations (26)
(29):

t 2 r 4 t 1

S (r ) r
t 0

32t t ! 2t 1 !! 2t 2

t 3 r 4t 3

P (r ) r
t 0

32 t ! 2t 1 !! 2t 3
t

S ( r ) r
t 0

P ( r ) r
t 0

Now we find two solutions of the type:

t r 4t 1
32t t ! 2t 1 !! 2t

t 1 r 4t 1
32t t ! 2t 1 !! 2t 1

(46)

(47)

(48)

(49)

Vibrational Modes of Thin Flat Circular Plates Calculated Using the Dynamic bi-Laplacian Equation

255

Internat ional Jour nal of Engineeri ng Science s, 2(6) June 2013

H1 (r ) lim
0

S (r ) P (r )

(50)

sin

32
S ( r )

sin

P (r )
K1 (r ) lim
0

(51)

The solutions obtained by developing the two limits are:

H1 (r )
K1 ( r )

2
2
r log(r )

2t

t r 4t 1

log(r ) 2t 1 l 32 t !(2t 1)!(2t 1)!!

l 1

2 t 1
32 2
3t 2
t r 4t 3
1
2 log(r )
2 t
r t 0
2t 2
l 1 l 32 t !(2t 2)!(2t 1)!!

(52)
(53)

2.3 Case m = 2
Supposing = 2 + we may solve the limits:

32
S (r )

sin

S (r )
H 2 (r ) lim
0

96
P ( r )

sin

(54)

P (r )
K 2 (r ) lim
0

(55)

and find the functions which confirm equation (14) for m = 2. In Section IV we will calculate K2 ( r ).
2.4 Case m = 3
Supposing = 3 + we may solve the limits:

96
P (r )

sin

S (r )
H 3 ( r ) lim
0

322
S (r )
2
sin

(56)

P ( r ) 12
K 3 (r ) lim
0

(57)

2.5 Case m = 4
Supposing = 4 + we may solve the limits:

32 2
S (r )
2
sin

S (r ) 36
H 4 (r ) lim
0

322
P (r )
2
sin

(58)

P (r ) 60
K 4 (r ) lim
0

And so on.

(59)

Bellotti Giuseppe

256

Int ernational Journal of Engi ne ering Sc iences, 2(6) June 2013

3.

Typical modal forms

4.

Calculation of the function K2 (r)

We have to calculate the following limit:

K 2 (r ) lim

P (r ) P (r )

sin

(60)

Since = 2 + , from equation (27) and equation (29) we have:

K 2 r lim
0

t 1 1 r 4t 4

r
1

t
t 1
t 0 32 t ! 2t 1 !! 2t 4
t 0 32 t 1 ! 2t 3 !! 2t 2
sin

r
K 2 r lim
0

t 4 r 4t 4

t r 4t

r
1

t
t
t 0 32 t ! 2t 1 !! 2t 4
t 1 32 t ! 2t 1 !! 2t
sin

(61)

t 4 r 4t 4

(62)

We substitute in equation (62) the following equation (63):

r
K 2 r lim
0

t 0

t r 4t 4
32t t ! 2t 1 !!

G 1

r 1

2t 4
t 0 32 t 1 2t 3 2t 2
sin

(63)

G 4

If we remember equations (33) (37) , we can write:

(64)

Vibrational Modes of Thin Flat Circular Plates Calculated Using the Dynamic bi-Laplacian Equation

257

Internat ional Jour nal of Engineeri ng Science s, 2(6) June 2013

1 log r

G 3 2 1 1
2t 4

t0

K 2 r lim

sin

G 1
1 log r 1

t 0 32 t 1 2t 3 2t 2

sin

(65)

G 3 2 1 1 c
G 1 c

1 log r
1 log r

t 0 2t 3 2t 2 2t 2
t 0 32 t 1 2t 3 2t 2
1 log r
K 2 r lim

0
sin
sin
sin

(66)
If we define:
from equation (66) we obtain:

K 2 r lim
0

1 log r

sin

16 log r 3 2 1 1 c
2 log r 1 c
(67)

lim G
G

0
sin

16

2
t

2
t

2
t

sin

16

2
t

2
2
t

2
t

t 0

16 3 2 1 1 c
2 1 c

lim G
G

0
sin 16 2t 2 2t 3 2t 2
t 0
sin 16 2t 3 2t 2 2t 2

K2 r

96log r log r

16

2
t

3
2
t

2
2
t

1
!

t 0

2

16 3 2 1 1 c 2t 2 2t 3 2t 3 1 c 2t 3 2t 2 2t 2
lim G

0
sin 16 2t 2 2t 3 2t 2 2t 3 2t 2 2t 2
t 0

(68)
If we remember equation (63) and define:

96

(69)

We can remember equation (33), and write:

K2 r

96 12
t r 4t 4

log r t

t0
32 t ! 2t 3 ! 2t 1 !!

6 11 1 c 2t 2 2t 3 2t 1 1 1
G

sin 2t 3 2t 2 2t 2 2t 2 2t 3 2t 2
lim

0
6 1 c 2t 3 2t 2 2t 1 1 1
t 0

G
sin 2t 3 2t 2 2t 2 2t 2 2t 3 2t 2

K2 r

(70)

96 12
t r 4t4

log r t

32
t
!
2
t

3
!
2
t

1
!!

t0

11

1 6 1 c 2t 2 2t 3 2t 1 1 1 c

G

6

2t 3! 2t 3!
lim

0 sin
t 0

1 c 2t 3 2t 2 2t 1 1 1 c

2t 3! 2t 3!

(71)

Bellotti Giuseppe

258

Int ernational Journal of Engi ne ering Sc iences, 2(6) June 2013

K2 r

96 12
t r 4t4

log r t

t0
32 t ! 2t 3 ! 2t 1!!

11

2 2
1 6 2t 2 2t 3 2t 1 1 1 c

G
6

2t 3! 2t 3 !
lim

0 sin
t 0

2t 3 2t 2 2t 1 1 1 c 2 2

2t 3! 2t 3!

(72)

If we remember Eq.(38), we can write:

K2 r

96 12
t r 4t 4

log r t

t0
32 t ! 2t 3 ! 2t 1 !!

2 t 1
11

1 2t 2 2t 3 2t 1! 2t 1!

l 1 l

G

2t 3 ! 2t 3!
6

lim

2 t 1
0 sin
1

t 0
2t 3 2t 2 4t 5 2 2t 1! 2t 1!

l 1 l

2t 3! 2t 3!

K2 r

96 12
t r 4t4

log r t

t0
32 t ! 2t 3 ! 2t 1!!

2 t 1
11
2 t 1

4t 5
1

6
l 1 l

l 1 l
G 2t 3 2t 2

lim
G

0 sin
2
t

3
!
2
t

3
!

t 0

K2 r

5.

(73)

2 t 1

4t 5
96 12 11
1
t r 4t 4
log r
t

t 0 12
2 2t 3 2t 2 l 1 l 32 t ! 2t 3 ! 2t 1 !!

(74)

(75)

Conclusions

The theory of vibrating thin circular plates provides the Dynamic bi-Laplacian Equation which is a useful analytic model. In order to
calculate the resonance frequencies and the modal forms of vibration, this paper considered a thin flat circular plate fixed at the central hole
of radius r1 and free on the outside circumference of radius r 2. By an explicit formulation of the bi-Laplacian operator in cylindrical
geometry, the Dynamic bi-Laplacian Equation was integrated using the variable separation method in order to solve the principal functions
of this dynamic problem. These functions can solve many other structural symmetrical problems that involve different boundary conditions.
A complete vibrational modal analysis of rotating discs might be achieved using the parameter m = . Since these modal forms show a
single nodal radial line, many asymmetrical shapes of vibration appear when the ratio between the angular frequencies of resonance linked
to m = and the angular frequency of rotation of the disc is a suitable whole. Moreover, since the vibrating thin circular plate can also
simulate a bi-dimensional e.m. standing wave in cylindrical coordinates, there is a possibility that the results obtained in the paper could
allow to analyze any bi-dimensional atomic and nuclear models based on the Electromagnetic Atomic Theory (reference 2 and reference 3).

Acknowledgement
Reviewed by Sally Pitt.

References
[1]
[2]
[3]

Dym, C.L., & Shames, I.H. (1973). Solid Mechanics. New York: MecGraw-Hill Book Company.
Bellotti, Giuseppe (2012). The Ideas Behind the Electromagnetic Atomic Theory. Advances in Natural Science, 5(4), 711. DOI: 10.3968/j.ans.
1715787020120504.2014
Bellotti, Giuseppe (2009). The dynamic bi-Laplacian equation in polar coordinates and the magic numbers of atomic nucleus. Physics Essays, 22(3),
268-287. DOI: 10.4006/1.3141024