Assignment 6.

Statics and dynamics of plates

Problem 1.
Consider a circular plate with Young’s modulus E, Poisson’s ratio ν, thickness h and radius R, resting
on elastic foundation of stiffness kd and subject to two loads:

1. a point load P at the center of the plate

2. a line load q at the edge of the plate; the latter is uniformly distributed along the edge

First, we consider that the point load is replaced by a disk/surface load p, which is related to the the
amplitude of the point load as follows:

P
p= (1)
πRd2
where Rd denotes the radius of the circular area the disk load occupies.
We proceed to consider two regions in the circular plate:

1. the region 1 below the disk load (inner circle)

2. the (unloaded) region 2, being the circular plate excluding the part below the disk load.

For region 1, the equation of motion of the plate reads:

D∇4 w− = p (2)
w− (Eh3 )/(12 ν2


where is the displacement of the plate (only for region 1) and D = 1− is the plate
bending stiffness. The solution to Eq. (2), i.e. a circular plate on elastic foundation and subject to a
surface load p, reads:

− −
w− (r) = whom (r) + wparticular (r) (3)
where the homogeneous and particular solutions are expressed as:

 1 !  1 !
− 1 + i kd 4 1 + i kd 4
whom (r) = C1 J0 √ r + C 2 Y0 √ r
2 D 2 D
 1 !  1 !
1 − i kd 4 1 − i kd 4
+C3 J0 √ r + C4 Y0 √ r (4)
2 D 2 D

− p
wparticular (r) = (5)
kd
For region 1, we do not have any boundary conditions, but there is the condition for bounded response
(i.e. not infinite) at r = 0, which leads to:

C2 = 0, C4 = 0 (6)
Therefore for region 1, the full solution reads:

1
Assignment 6. Statics and dynamics of plates

 1 !  1 !
1+i kd 4 1−i kd 4 p
w− (r) = C1 J0 √ r + C3 J0 √ r + (7)
2 D 2 D kd
Similarly, for region 2 the equation of motion of the plate reads:

D∇4 w+ = 0 (8)
where w+ is the displacement of the plate (only for region 2). The solution to Eq. (8), i.e. a circular
plate on elastic foundation without a surface load, reads:

 1 !  1 !
1 + i kd 4 1 + i kd 4
w+ (r) = C5 J0 √ r + C 6 Y0 √ r
2 D 2 D
 1 !  1 !
1 − i kd 4 1 − i kd 4
+C7 J0 √ r + C8 Y0 √ r (9)
2 D 2 D

It is important to remark here that:

1. the solution is expressed in terms of Bessel functions and not Hankel functions, because the plate
is of finite and not infinite extent.

2. none of the coefficients can be eliminated directly as this region does not include neither r = 0, nor
r → ∞, which are usually associated with such conditions.

The boundary conditions at r = R are the following:

 2 +
1 ∂w+ 1 ∂ 2 w+
 
+ ∂ w
Mr (r = R) = −D +v + 2 =0 (10)
∂r2 r ∂r r ∂θ2 r=R

∂ ∂ 2 w+ 1 ∂w+ 1 ∂ 2 w+
  
+
Vr (r = R) = −D + + 2 =q (11)
∂r ∂r2 r ∂r r ∂θ2 r=R
Finally, the interface conditions that hold for the two regions at r = Rd are the following:

w+ (r = Rd ) = w− (r = Rd ) (12)

∂w+ ∂w−
= (13)
∂r r=Rd ∂r r=Rd

Vr+ (r = Rd ) = Vr− (r = Rd ) (14)

Mr+ (r = Rd ) = Mr− (r = Rd ) (15)

As can be seen, we have 6 unknown coefficients C1 , C3 , C5 , C6 , C7 andC8 and 6 equations, i.e.
Eqs. (10) to (15), which are all dependent solely on the unknown coefficients (and the plate and load
parameters which are considered known). Therefore, a system of 6 algebraic equations can be formed
and solved.
It is noted that the solution up to this point is valid for a disk load of radius Rd and amplitude
p = P/(πRd2 ). The final solution wP for the case of a point load P at the center of the plate is obtained
as:

w+ (r)


wP (r) = lim (16)
Rd →0
Assignment 6. Statics and dynamics of plates

Problem 2.
Consider a rectangular plate with Young’s modulus E, mass density ρ, Poisson’s ratio ν, lengths of Lx
and Ly along the x and y coordinates, respectively, resting on elastic foundation of stiffness kd . The two
opposite edges of the plate are simply supported (y = 0 and y = Ly ), whereas the other two edges are
free.
The equation of motion of the plate (in expanded form) reads:

∂2w
 4
∂4w ∂4w

∂ w
ρh 2 + D +2 2 2 + + kd w = 0 (17)
∂t ∂x4 ∂x ∂y ∂y 4
To study the free vibrations of the rectangular plate considered, we assume the following solution:
 
nπy
w(x, y, t) = sin(ωt)Wmn (x) sin (18)
Ly
where Wm,n (x) is the mode shape dependency along x and the index m denotes the m-th mode of the
modal set corresponding to n. It is remarked that Eq. (18) corresponds to a single free vibration mode.
By substituting Eq. (18) into Eq. (17), the following equation is obtained:

" 2 4 !#
d4 Wmn (x) d2 Wmn (x)
  
nπy nπ nπ
sin(ωt) sin (kd − ω 2 ρh)Wmn (x) + D −2 + Wmn (x) =0
Ly dx4 Ly dx2 Ly
(19)
For the above expression, the time-dependent term is discarded as we are interested in non-trivial
solutions, so we have:

" 2 4 !#
d4 Wmn (x) d2 Wmn (x)
  
nπy nπ nπ
sin (kd − ω 2 ρh)Wmn (x) + D −2 + Wmn (x) =0
Ly dx4 Ly dx2 Ly
(20)
At this point we take advantage of the orthogonality condition:
Z Ly    
jπy nπy Ly
sin sin dy = δn,j (21)
0 Ly Ly 2
 
thus we multiply by sin mπy
Ly and integrate along y from 0 to Ly , leading to (only for n = j):

" 2 4 !#
d4 Wmn (x) d2 Wmn (x)
 
Ly nπ nπ
(kd − ω 2 ρh)Wmn (x) + D −2 + Wmn (x) =0 (22)
2 dx4 Ly dx2 Ly

Finally, the following ODE can be obtained:

2 4 !
d4 Wmn (x) d2 Wmn (x) kd − ω 2 ρh
 
nπ nπ
−2 + + Wmn (x) = 0 (23)
dx4 Ly dx2 Ly D
By substitution of Wmn (x) = exp(kx) into Eq. (27), we obtain the following:
 2  4 !
nπ nπ k − ω 2 ρh
d
k4 − 2 k2 + + =0 (24)
Ly Ly D
The above equation has 4 roots (except for certain values of ω, when roots with multiplicity may
appear) and can be trivially solved by setting λ = k 2 , leading to:

k = α, k = −α, k = β, k = −β (25)
and the values of α and β can be found as follows:
Assignment 6. Statics and dynamics of plates

r  q  r  q 
D Dn2 π 2 − DL4y hω 2 ρ − DL4y kd D Dn2 π 2 + DL4y hω 2 ρ − DL4y kd
α= , β= (26)
DLy DLy

Therefore, the full solution to Eq. (27) may be written as:

Wmn (x) = C1 exp(−αx) + C2 exp(αx) + C3 exp(−βx) + C4 exp(βx) (27)

At this point we employ the boundary conditions along the free edges (x = 0 and x = Lx ), which are
the following:

Mx (x = 0) = 0 (28)

Vx (x = 0) = 0 (29)

Mx (x = Lx ) = 0 (30)

Vx (x = Lx ) = 0 (31)
As can be seen, we have 4 unknown coefficients C1 , C2 , C3 , C4 and 4 equations (i.e. Eqs. (28) to (31));
however, the natural frequencies ω still remain unknown and the system cannot be solved. The process
to obtain the natural frequencies and modal shapes is as follows;

• form a system of 4 algebraic equations and 4 unknown coefficients and require the determinant of
the coefficient matrix to be equal to zero. This determinant will be a function of frequency ω and
needs to be evaluated numerically.

• assuming that the roots of that determinant (natural frequencies) have been obtained, we substitute
each frequency in the coefficient matrix and set one of the coefficients (C1 , C2 , C3 , C4 ) equal to a
value, e.g. C4 = 1. Then we proceed to obtain the remaining 3 coefficients (C1 , C2 , C3 ) by solving
3 of the algebraic equations (i.e. boundary conditions). In that manner for each natural frequency
ωmn , we can obtain the associated mode shape.