Pure Bending and Cylindrical Bending of Rectangular Plates

Plate is subjected to bending moment M1 and M2 per unit length

a  2w 2w 
x   , M x   D 2  v 2   M1 , M xy  0
2  x y 

b  2w  2w 
x   , M y   D 2  v 2   M 2 , M xy  0
2  y x 
 Case 1
If M1, M2 ARE POSITIVE AND EQUAL , the deformed
surface WILL HAVE SAME CURVATURE

Case 2
If M1, M2 ARE POSITIVE , the deformed surface is in synclastic form

Case 3
When M1 IS POSITIVE AND , M2 IS NEGATIVE, the will form anti-clastic
form
Case 4
Cylindrical bending of the plate

Four case we have studied in depth

a x

b
x y
y q  q0 sin .sin
a b

q0 is intensity at center
Consider 4th order differential equation

 4 w 2 4 w  4 w q
 2 2 4 
x 4
x y y D

Boundary condition
w = 0, Mx = 0 for x = 0 & a
w = 0, My = 0 for y = 0 & b
It is observed that if the solution for w is considered as

x y
w  c sin sin
a b
Boundary condition will be satisfied

w x y 
 [c cos sin ]
x a b a

2w x  y 2
 [c sin sin ] 2
x 2
a b a

3w 4w  4 x y
& similarly 4  4 sin sin
x 3
x a a b

 4
Similarly operate to 4 And put the values in equation
y y
  4 w 2 4 w  4 w q0 sin  x sin  y
 2 2 4 
x 4
x y y D a b

2
4  1 1 q0
  2  2 c
a b  D

2
q0  1 
C  4 
 D 1/ a 2  1/ b 2 
Replace value of c

2
q0  1  x x
w 4  .sin sin
 D 1/ a 2  1/ b 2  a b
Once you know
w  Mx, My, Mxy, Qx, Qy

 2w 2w 
M x   Dx  2  2 
 x y 

q0 m x m y
w sin sin
2 2 a a
m n 
2
 D 2  2 
4

a b 
Navier Solution ( assumed Kirchhoff Theory ) solution
technique The solution of fourth order differential
equation for all four edges simply supported plate
can be obtained by use of double infinite series.
Navier assumed solution in following form similar
to double trignometric series
 
w( x, y )   Amn f m ( x) g n ( y )
m 1 n 1  
w   A mn sin  m x sin  n y
m 1 n 1
m n
where  m  , n  and A mn is cons tan t
a b 2
w
along x  0, a w  2  0
x
2w
along y  0, b w  2  0
y
Navier assumed the loading in the following form to
match w assumedearlier

q ( x, y )   qmn sin  m x sin  n y
m 1 n 1

q ( x, y )...is...loading m n
where  m  , n 
qmn ...is..arithmatic..cons tan t a b
multiply..both.....sides...by
a b

   n ydxdy 
' '
q ( x , y ) sin m x sin
0 0
a b  

   mn m     
' '
( q sin x sin n y ) sin m x sin n ydxdy
0 0 m 1 n 1
a

 m    
' '
sin x sin m xdx {0 for m m
0

{a / 2 for m  m '
 b

 n    
' '
sin y sin n ydy {0 for n n
0

{b / 2 for n  n '
a b
4
q mn    q ( x , y) sin  m x sin  n ydxdy
ab 0 0
Now....consider.....Kirchhof ...Eqn  

q w   A mn sin  m x sin  n y
  w
2 2
m 1 n 1
D
put..Value ...of ..w  

and ..value....of .. q ( x, y )   qmn sin  m x sin  n y

   
1
2 2  Amn sin  m x sin  n y   qmn sin  m x sin  n y
. m 1 n 1 D m 1 n 1
Differentiate the earlier and simplify

A mn   m   n   D  sin  m x sin  n y  0
 
 q mn 

m 1 n 1 
2 2 2

Can not
be zero
hence


A 
 mn m
2
  2 2
n q mn 
D   0

 q mn
A mn 

D  
2
m n
2 2

a4   qmn
4 
w sin  m x sin  n y
D m 1 n 1  2 n a 
2 2 2

m  2 
 b 
(a)Sinusoidal load
Total reaction=

Total reaction=
(b)uniformly distributed load
(c) Patch load
Fourier coefficient for different types of loads

a b
4  qx 
q      sin  m x sin  n ydxdy
ab 0 0  a 

16q m = 1, 3 ,5 ………
q mn 
mn 2 n = 1, 3 ,5 ………
 c b
4
q mn    q sin  m x sin  n ydxdy
ab 0 0

q mn  
8q
(cos  m c  1)
m = 1, 2 ,3 ………
mn 2

n = 1, 3 ,5 ………
 c

2 b
4 p
q mn 
ab c 0  c  sin  m x sin n ydxdy

2

Strip load

16q  c
q mn  sin  m  sin m
mn 2
2

Line load (c  0)

8p m = 1, 2 ,3 ………
q mn  sin  m 
an
n = 1, 3 ,5 ………
It is clear that Navier solution provides considerable
mathematical advantage as the solution is reduced to a
simple algebraic equation and can handle complex load
distributions.
However, the solution is limited to only all-round simply
supported plate.
In some of the examples given above, it has been
shown that the convergence of the series for deflection is
rapid for the case of distributed loads while it becomes
slow for highly discontinuous loads including
concentrated load.
Further, the convergence of the series for bending
moments is not as fast as that for deflection. Hence in
most of the cases it becomes necessary to sum up a
large number of terms, thus making the computation,
particularly manual computation, unwieldy.
This disadvantage can be nullified to a large
extent if the summation of the series is carried
out on a digital computer in which case,
consideration of a large number of terms does
not pose any difficulty.