Bending Theory.

Why ?
1 In practice shell will not behave as membrane
2 Behavior (not free from bending)

3 Stresses & displacements are different.

Nx

Example

N

Actual boundary condition can be realized by applying some correction

 •Apply load to counter act N & Nx

• This gives rise to bending & twisting

• which is Away from membrane

•Bending theory is essential

 Bending analysis in three steps

1. A membrane analysis for the given load.

2. Bending analysis of unloaded shell (edge

disturbances )

3. Super positioning of 1 and 2 to get actual boundary

condition (Assumption edge- straight and ).
3. Pre- requirment

Uniaxial state of stress

x dx

T T

A B
Before T, length AB = dx

After application of load

Let A has changed by u & B has changed by u  u dx

x

u
change in length  dx
x
u
strain 
x
u
 
x x
 State of Plane Stress
consider an element ABCD taken out of body acted by plane stresses
u u
u dx  dy
x y
C’
u
u dy v v
y D’ v dx  dy
v x y
v dy
y D C

B’
v
v dx
A’ x
v
B
A
u u
u dx
x

u,v are components of displacement of A to A’

The transformation of ABCD to A’B’C’D’can be considered in three steps.
 D
1 deviation in x direction

B
A
2 deviation in y direction u u
u  dx
x

v
v dy
y

3 angular distortion (r1 + r2) as we have seen earlier

u
strain in x direction  x 
x
v
strain in y direction  y 
y
v u
shear strain  xy  
x y
 Strains In circular Cylindrical Shell
use coordinate x and a
u,v,w are the displacement in respective direction.
1. u displacement in x direction
2. v in the direction of the tangent at point. x

3. w in the normal inward direction.

u

v
w

a
Now we have already seen for plane element 

u v  v u 
 x  ,  y  ,  xy    
x y  x y 
but if the element is curved then in x and a  system
 u
1 x  will remain same
x
2  xy will remain same only y will be replaced by a 
 v 1 u 
 x   
 x a  
The strain in the  direction will consist of two parts
Let the point be @ A has gone to A1 &
come down to A’ because of w.
v 1 v
1 Strain corresponding to plane state of stress elemments converted into a
y a 

A1 2 Circumferntial strain caused by w 

 a  w d  ad
w ad
A
A’
w
 -( )
a
a
dΦ  1 v w 
Totalstrai n      .  
 a  a 
Change in Rotation and circumferential
Curvature
 x

Finally we have
A
B
u x
x 
x  d

 1 v w 
   .  
 a  a  a

v+v/a.a
 1 u v  
 x   .  
 a  x  v
A1 B

A B’
w
Now rotation of tangent w+w/a.
A2
a

Rotation of tangent @ A is v/a

due to A to A & A1to A2

d
 Rotation of A to A1 is change of v
Therefore, Angle change = arch length / radius = r/q
 1 w 
Tan  =  . a   1 w
a  
  .
a 
a
total rotation @ A
v 1 w
  .
a a  w

1 w  w
  v   /a 
.a
a   

change in circumferential arc ds  a.d

1 
 curvature
a s
Now change in curvature is contributed by two
1) Change in arc length
2) Change in angle substended

Now we have seen that rotation @ A is given by

Similarly there will be rotation @ B and changes to New form

 2w 
The differential rotation @ B and A =   2 d
 

Hence central angle now become

 Hence central angle now become

 2w 
d +  2 d
  
and arch length becomes (a-w) d

now we have seen that

1   2 /  2
change in circumference =  1/ a
aw
 1  2
Change in curvature   2 ( w  )
a  2

Let (a-w) a = a.a as w is small

1  2
  2 ( w  2 )
a 
 Stress resultants

The forces and moments acting on a shell per unit

length are use known as stress resultants.

In a cylindrical shell under bending there are,

in general, 10 stress resultants to be
determined , the stress resultants may be
related to the corresponding stress as follows:
 d / 2az
Nx    x dz - - - - - (a)
d / 2 a

d / 2 az
Mx   d / 2 xz dz..........(b)
a

d / 2 az
Mx    x z dz..........(c)
d / 2 a

d / 2 az
Qx    xz dz - - - - - (d)
d / 2 a
 d / 2 az
Nx  d / 2 x z dz - - - - - (e)
a
d /2
N   d / 2 dz - - - -(f)
d /2
M   d / 2
dz - - - -(g)
d /2
Mx  d / 2  x zdz..........(h)
d / 2
Q  d / 2  z dz - - - (i)
d /2
N  d / 2  z dz - - - (j)
The factor (a+z)/a = (1+z/a) appearing in some of these expressions takes care of
trapezoidal shape of the element of the shell cross section . if the shell is thin, z/a < 1 and
can be ignored. If is done, it is verified that M x = M  x and Nx = Nx, because x =
=x easily
Stress Strain Relation

y

x
x = (x/E- y/E) y= (y/E-x/E)

σ x = Nx/(1 * d) d = thickness
σ y = N / (1 * d)
u
x = = (x/E- y/E)
x
= Nx/(E.d) -  N/Ed
u
= x = Nx/(E.d) -  N/Ed
x
e = (y/E-x/E)
1 v
e  (  w) = N/Ed -  Nx/Ed
a 
G = E/2(1+)
1 u v
Nx/Gd = x = 
a  x

1 u v
Nx/Gd = Nx 2(1+) / Ed = 
a  x

M    D  Moment – curvature relations

 M  ... flatten...out M/I = E/R
..shell .. & .. M = EI /R
.reduces ...curvature M/D = 1/R

Ed 3 1 2w
M  ( 2 (w  ))
12(1   ) a
2
 2
 Membrane displacements

stress – strain relationship provides to get u,v,w

consider following equation

1 u v
Nx/Gd = Nx 2(1+) / Ed = 
a  x

1 u v
Nx 2(1+) / Ed = 
a  x
 put  = 0
1 u v
2Nx / Ed = 
a  x
Differentiate with respect to x

2 N x 1  2u  2v
( ) 
Ed x a x x 2
our interest is to get
1  u 2
 2v
or
a x x 2
consider equation (1)

u
= x = Nx/(E.d) -  N/Ed
x
=0
u = Nx/(E.d)
differentiat with respect to d
x
  2u 1 Nx
 ( )
x Ed 

2 N x 1 Nx  2v
( )  
Ed x Eda  x 2

2 N x 1 Nx  2v
( ) 
Ed x Eda  x 2

By integrating we will get the values of v similarly u and w

In Membrane analysis we know
the values of in plane forces so
we can calculate membrane
displacements u v w
 Classical theory of Bending
Finsterwalder (1932)

Assumptions
1. material is homogeneous, isotropic & Obey Hook’s law.
2.Stresses normal to shell surface are neglected
3. A rectilinear elements normal to middle surface of shell remains
rectilinear and normal even after deformation
4. All displacements of the shell surface are small.
5. Mx, Qx & Mx = 0
 In bending analysis we are to find

Nx, N, Nx, Nx

Mx, M, Mx, Mx, Nx, N, Nx,

Mx, M, Mx,
Qx, Q, Qx, Q,
( In all 10) Remain 8 Because of
Noho
Nx, N, Nx,
As per Finsterwalder Mx, M, Mx,
Mx =0, Mx =0, Qx=
Qx, Q,
0 Remain 5 Because of
Finstarwalder
Aim to know
Nx, N, Nx, M, Q,
 Equation of equilibrium for unloaded shell

By summing all the forces In x directions

Nx 1 Nx
 ( )0 Eq.no1
x a 
By Summing the forces in  direction

N Nx
 a( )  Q  0 Eq.no2
 x
By Summing the forces in Normal inward direction

Q
N  ( )0 Eq.no3

By adding sum of moments about directrix AD
1 M
 Q  0 Eq.no4
a 
 Finsterwalder differential equation
introduced f() a function such that M = - f()cos (nx)/l
nx
M   f ( ) cos
l
put
na n n
n   hence
l a l
n
M   f ( ) cos( x)
a
Consider Eq.no 4

1 M
 Q  0
a 
 1 M
Q 
a 
Put the value of Mø

1  n x
Q  (  f ( ) cos )
a  a

Let ( f ( ))  f 0

1 n x
Q   f 0
cos( )
a a

Now consider Eq. no3

 
N  ( Q )  0



N   ( Q ) Eq.no x1


Substitute the value of

1 n x
Q   f 0
cos( )
a a
In Eq.no x1

1 n x Eq.no x2
N   f 00
cos( )
a a
 Consider Eq.no2

N 
 a( Nx )  Q  0
 x

Nx 1 
 ( N )  Q  0
x a 
Substitute the values of

1 n x
Q   f 0
cos( ) and
a a
1 n x
N   f 00
cos( ) Get
a a
 1 n x
Nx   2 { f 000
 f } cos(
0
)
x a a
Now integrate it
n x
{f 000
 f 0
} sin( )
1 a
Nx   2
a n
( )
a

1 n x
Nx   {f 000
 f } sin(
0
)
a n a
Consider Eq. no 1
Nx 1 
 ( Nx )  0 Re-arranging
x a 
Nx 1 
  ( Nx )
x a 
Put
1 n x
Nx   {f 000
 f 0 } sin( )
an a

 1 1 0000 00 n x
Nx   [  {f  f }sin( )]
x a an a
By integrating

1 n x
Nx   {f 0000
 f 00
} cos( )]
an
2
a
Now expressions for u,v,w can be had by stress strain relationship

Nx u

Ed x
u
Ed  Nx Put the value of Nx
x
1 n x
 {f 0000
 f 00
} cos( )]
an
2
a

Integrate it

1 0000 00 n x
Edu   {f  f }sin( )]
 3
n a
 Consider

2(1   ) 1 u v
Nx  
Ed a  x

Put poisson’s ratio =0

2 Nx 1 u v
 
Ed a  x

Re-arranging
v 2 Nx 1 u
 
x Ed a 
Multiply by Ed

v Ed u
Ed  2 Nx 
x a 
Put

1 n x
Nx   { f  f } sin(
000 0
)
a n a
And

1 0000 00 n x
Edu   {f  f }sin( )]
 3
n a
 2 0 000 1 000 00000 n x
Edv  [ 2 { f  f }  4 { f  f }] cos( )
n n a
 Now consider
N Nx 1  dv 
      w
Ed Ed a  d 
N 1  dv 
   w
Ed a  d 
aN  dv 
or   w
Ed  d 
 dv 
Edw   aN  Ed 
 d  
dv
put value of aN and Ed
d
 2 1 n x
Edw  [ f 00  2
{ f 00
 f 00000
}  4
{ f 0000
 f 000000
}] cos( )
n n a

Now we know v & w we can go rotation of tangent

1 w 
  v  
a  
After simplification we get
2
Ed  [ 2 { f  2 f  f
0 000 00000
}
an
1 000 00000 0000000 1 000 n x
 4
{f  2 f f )  f }] cos( )
an a a
 1  2w 
  2  w  2 
a   

1 2
   [ f 00
 f 0000
 ( f 00
 2 f 0000
 f 000000
)
n
2 2
Eda
1 n x
 4 (f 0000
2f 000000
f 00000000
)] cos( ).......eq..no..nn
n a
 We know that initially we have started with
n
M   f ( ) cos( x).
a
n
M   f cos( x).....eq.no...mm
a
Also M   D
Ed 3
M   
12(1   )
2

Ed 3
M   
12
substitute...the...value..of ..
 .. from..eq.no..nn......and
M ... from...no..eq..mm
 After putting the value of Mø and χø in above expression
We get Finsterwalder 8 th order Deff. Equation

00 0000 2 00 0000 000000

[f  f  2
(f 2f  f )
n
2
1 0000 000000 a
 4
(f 2f  f 00000000 )]  12 2 f  0
n d
Or

( f 00000000  2 f 000000  f 0000 )  22 n ( f 00  2 f 0000  f 000000 )

2
a
 4 n ( f 00  f 0000 )  124 n 2 f  0
d
Mathematical Technique
finsterwalder differential equation..can be satisfied..if
- f( )  He m
Now substituti ng and rearrangin g..the solution in following equation

( f 00000000  2 f 000000  f 0000 )  22 n ( f 00  2 f 0000  f 000000 )

2
a
 4 n ( f 00  f 0000 )  124 n 2 f  0
d
m  2m (1  n )  m (1 - 4n  n ) 
8 6 2 4 2 4

m (n  2n )  12n a / d  0

2 4 2 4 2 2

The above is the characteristic eq. of Fansterwalder

Methods of calculating the roots of above equation is as under

Steinman Rule: if the polynomial is as follows then

mx  ax
n n -1
 bx n -2
 cx n -3
 dx n -4
 ....
If xo is the approximate value of one of its
real roots then the exact value is given as
under
2b 3c 4d
a  2  3  ....
xo xo . xo
x1 
b 2c 3d
m  2  3  4  ....
xo xo . xo
 How to apply for particular example In
Fansterwalder Eq.

Consider the following data

Span 83.25 feet, radius 25, thickness d-
0.25. ft

a
n   0.9434212174
l

2
a
Let..  12n 2  95061.311784
8 4

d
 Fansterwalder equation

m  2m (1  n )  m (1 - 4n  n ) 
8 6 2 4 2 4

m (n  2n )  12n a / d  0

2 4 2 4 2 2

Substitute m  m/ 

8 2(1  n 2 ) 6 (1 - 4n 2  n 4 ) 4 (n 4  2n 2 ) 2

m  m  m  m 1  0
 2
 4
 6

a
Substitute n   0.9434212174 a2
values l
8
  12 2  95061.311784
4

d
 8 6 4 2
m  0.0125241882m  0.0057342889m  0.0001829796m  1  0
2
y  m ....substitute...so..eq...becomes...
biquadratic

y 4  0.0125241882y 3  0.0057342889y 2  0.0001829796y  1  0

Eq...is...of ... form
y 4  py 3  qy 2  ry  s  0
To ...e lim inate.. y 3 ... put Eq...is...of ... form
y  ( x  p / 4) x 4  ax 2  bx  c  0
Now...eq...becomes
x 4  0.00579310996 x 2  0.0001463254 x  1.0000005149  0
so..a  0.00579310996
b  0.0001463254....
c  1.0000005149
Now...eq...becomes
x  0.00579310996 x  0.0001463254x 
4 2

1.0000005149  0
Eq...is..in.. form
x  ax  bx  c  0
4 2

so..a  0.00579310996
b  0.0001463254....
c  1.0000005149
Now..resolven..cubical...eqn..will ..be
z 3  2az 2  (aa  4c) z  bb  0
so..new...eqn..becomes
z 3  0011586219 2 z 2  3.999684995 z  0.0000000214  0
To ..apply...Steinman..rule..eqn..rearranged
z 3  0011586219 2 z 2  3.999684995 z  0.0000000214
Steinman Rule: if the polynomial is as follows then

2b 3c 4d
a  2  3  ....
xo xo . xo
x1 
b 2c 3d
m  2  3  4  ....
xo xo . xo
The first root is assumed as zo=+2

2b 3c 4d
a  2  3  ....
xo xo . xo
x1 
b 2c 3d
m  2  3  4  ....
xo xo . xo
then we get z1 =2.005785260

Then assuming this value we get z1=

2.005793267
Third cycle gives the same value
shows that there is no further scope for
improvement
The second root is assumed as zo=-2 as
new trial then we get z2 =--1.994198990

Then assuming this value we get z2= -

1.994207403
Third cycle gives the value equals to zero
Hence roots of the equation are
z1= 2.005793267
z2= - 1.994207403
Z3=0
And roots of eq can be written as
roots...of ....eq
x  0.00579310996 x  0.0001463254x 
4 2

1.0000005149  0
can....as..written...as... follows ...given..by..
DESCAAARTES
 1
x1  { z1  i ( z 2  z3 )}
2
1
x2  { z1  i ( z 2  z3 )}
2
1
x3  { z1  i ( z 2  z3 )}
2
1
x1  { z1  i ( z 2  z3 )}
2

Now
y  ( x  p / 4)
hence.. y1..to.. y 4..canbe..calculated
We ...have..substituted
2
ym
m y
can..be.calculated
We ..know..
m  m/ m1, 2,3, 4   (3.869508850  i1.602028646)

m5, 6, 7 ,8   (1.599522594  i3.87557140)

Check ..if ..summation..of ..sq..of .roots ..

(3.869508850)  (1.602028646) 
2 2

(1.599522594) 2  (3.87557140) 2  95061.35927

must , , be..eq..to... 
8

then..the..values..are...correct
Mathematical Technique
finsterwalder differential equation..can be satisfied..if
- f( )  He m
Now substituti ng and rearrangin g..the solution in following equation
m  2m (1  n )  m (1 - 4n  n ) 
8 6 2 4 2 4

m (n  2n )  12n a / d  0

2 4 2 4 2 2

m1  1  i1   m5

m2  1  i1   m6
m3   2  i 2  m7

m4   2  i 2  m8
 Expression forM nx
M   f ( ) cos
..is..assumed l

by ..assuming - f( )  He m ...

solution...can..be...had

m  2m (1  n )  m (1 - 4n  n ) 
8 6 2 4 2 4

m (n  2n )  12n a / d  0

2 4 2 4 2 2

M  {H 1e (1  i1 )  H 2 e (1 i1 )

( 1  i1 ) ( 1  i1 )
 H 3e  H 4e
( 2  i  2 )  (  2  i  2 )
 H 5e  H 6e
(   2  i 2 ) (   2  i  2 ) nx
 H 7e  H 8e } cos
l
M  {H 1e (1 )(cos 1  i sin 1 )
(1 )(cos 1  i sin 1 )
 H 2e
 (1 )(cos 1  i sin 1 )
 H 3e 
H 4 e  (1 )(cos 1 i sin 1 )
( 2 )(cos 1  i sin 1 )
 H 5e
 H 6 e ( 2 )(cos 1 i sin 1 )
 ( 2 )(cos 1  i sin 1 )
 H 7e 
 ( 2 )(cos 1  i sin 1 ) nx
H 8e } cos
l
 Observe this eq. more terms are multiplied by negative
and four are by +index

M Is disturbance and should be expected to be o as we move

away from edge or we increase angle phi

Is disturbance and This is possible with - exponetial

hence + terms are neglected to realise the boundary
condition and eq takes the following form
  (1 )(cos 1  i sin 1 )
M  H 3e 
 (1 )(cos 1  i sin 1 )
H 4e
 ( 2 )(cos 1  i sin 1 )
 H 7e 
 ( 2 )(cos 1  i sin 1 ) nx
H 8e } cos
l

Or this can be written as follows

 Observe this eq. more terms are multiplied by negative
M  { and four are by +index

 (1 )
e ( H 3  H 4 ) cos 1
 (1 )
e ( H 3  H 4 )i sin 1 
 (1 )
e ( H 7  H 8 ) cos  2 
 (1 ) nx
e ( H 7  H 8 )i sin  2 } cos
l

H3 , H4, H7 and H8 are complex number but M

is real number so it follows that
(H3+H4) , i((H3-H4) ,( H7+H8 ) and i(H7-
H8) have to be real. And this to satisfy…..
numbers should be conjugate pair
To further simplify

( H 3  H 4 )  An
i ( H 3  H 4 )  Bn
( H 7  H 8 )  Cn
i ( H 7  H 8 )  Dn

Hence eq. will be

M 
[e  (1 ) ( An cos 1  Bn sin 1 ) 
 ( 2 ) nx
e (C n cos  2  Dn sin  2 )] cos
l
M ....is... inf inite..series
n..can..have.1...3..5....up..to. inf inity..values
An ...Bn ..C n ...Dn ..and cos ...can..have..these
values..of ..n..1...3..5...have..inity..only
nowone...has... four..values..of ... ..and
so.. An ...Bn ..C n ...Dn ..canbe..evaluated
 Re curance... formula

by...differntta iting ..
M


 (1 )
{e {(1 An  1 Bn ) cos 1 
(  1 An  1 Bn ) sin 1
 e  ( 2 ) {( 2C n   2 Dn ) cos  2 
(   2C n   2 Dn ) sin  2 )}
nx
cos
l
observ...the..Expresssion
M
if ...is....operated ...then..it .

will ..not...alter..the.. form..
An , Bn , C n , Dn ...will ..change
we..denote...the..values..as..under

A(1)  ( 1 An  1 Bn )
B (1)  (  1 An  1 Bn )
C (1)  (  2C n   2 Dn )
D (1)  (  1C n   2 Dn )
So.. forr ...epeated ..diffi
( k i ) ( k 1)
A( k )  ( 1 An  1 Bn )
( k i ) ( k i )
B (k )
 (  1 An  1 Bn )
( k i ) ( k i )
C (k )
 (  2C n   2 Dn )
( k i ) ( k i )
D (k )
 (  1C n   2 Dn )

If ...these...are...known...then..all ..the..
stress ..resul tan t ..can...be.. found ...out
.
 DKJ Bending Theory

D(Donnel,1938-34) K(Karman,1941) J(Jenkin,1947)

Published 1947 as known DKJ.

Simplified assumptions of Finsterwalder

Mx=0 , Mx=0. Qx = 0 DKJ disagreed
Structural Action under Bending combination of
disk, plate and membrane

Disk :- Disk action corresponds to shell loaded in it’s own plane

Plate :- Plate action corresponds to shell loaded normal to it’s plane
Membrane :- action corresponds to shell as a flexible membrane

D-K-J modified the equation of equilibrium as follows

 Nx 1 Nx
X  0   0 - - - - - -(1)
x a 
N Nx
 0  Q  a. 0
 x
Q is dropped as it does not occure in equation
of equilibrium of disk, plate on membrane
modified equation in  direction N  a. Nx  0 - - - - - -(2)
 x

 in normal inward direction Qx Q

a   N  0 - - - - - - - (3)
x 
4 th Equation by equating sum of all moment about AD
Mx M
a   aQ  0 - - - - - - - -(4)
x 
5th Equation  moments @ AB
Mx Mx
a  aQx  0 - - - - - - - (5)
 x
 6 th Equation  moments @ normal to element

(Nx - Nx).a  Mx  0 - - - -(6)

As Nx  Nx being shell thin Validity of 6 equation is only when Mx  0
(in built drawback of equation of DKJ)
stress  strain relation hold' s good
u Nx N
x    - - - -(A)
x Ed Ed
1  v  N Nx
    w   - - - -(B)
a    Ed Ed
Nx 2(1   ) Nx 1 u v
x    - - - -(C)
Gd Ed a  x
Moment curvature relationship
bending moment Mx and twisting moment Mx
ignored by Finsterwalder was tanken care by D-KJ/ together
with M (transverse moment)
the relationship similar to plate action was considered byD_KJ
  2w D 2w
Mx  -D 2 - - - (A1) M  - 2 - - - (B1)
x a  2

D 2w
Mx  -    (C1)
a x

D  2w
M  - 2 2 is different from M  - D o considered by finsterwalder
a 
(w term is dropped from the equation)
1  2
 w  2
1  w D 2w
   w     M   D  - 2 2
a 2   2 
 a 2  2 a 

Consider stress - strain & moment curvature relationsh ip

& substitut e in equation of equilibriu m
to get the three simultaneous Equations
equation involving u, v, w by Flu gge's Equation
Consider I st equation Nx 1 Nx
 0
x a 
Now   0
u Nx  2 u Nx
 (differentiate with dx) Ed  - - - -(1)
x Ed x 2
x

2(1   ) Nx 1 u v 1  1 u v 
and    Nx  Ed   
Ed a  x 2  a  x 

Nx 1  1  2u  2 v 
 Ed   
 2  a  2
x 
Nx Nx
replacing &
x 
 2 u 1  1  1  2u  2v  
Ed 2   Ed     0
x a  2  a  2
x  
 a2
multiply by
Ed
 2
u 1   2
u  2
v 
a 2
  2 a 0 - - - - - F1
x 2
2   x 

Now consider N  a Nx  0

 x
N

Nx 1  v
 

 w   0
Now Ed Ed a   
  N  Ed  w 
2
Ed  v  v
N    w  
  
a   a   2
 

 

Ed  1 u v    Ed  1  2
u  2
v
and Nx     
Nx  
  

 2
2  a  x  x 2  a x x 
Now substitute in equation as

Ed   2 v w  Ed  1  2u  2v 
 2    a   2   0
a     2  a x x 
a
multiply by
Ed
  2v w  1   2u  2
v
     a a 2
  0 - - - - - F2
2 
 
2
  2  x x 

Consider equation 5

Mx Mx 1  Mx Mx 

a  aQx  0  Qx   a 
 x a   x 
From moment curvature relation

2w
Mx   D
x 2

D  2w
M   2
a  2

D 2w
Mx  
a x

D 1  3w a 3 w
Qx   [  ]
a a x 2
x 3
Consider.equation.4
Mx M
a   aQ  0 - - - - - - - -(4)
x 
Mx M 1
[a  ]  Q
x  a
D  3w 3w
Q   [ 2  2 3]
a x  a 
 Qx Q
a   N  0 - - - - - - - (3)
x 

D 1  3w a 3 w
Qx   [  ]
a a x 2
x 3

D 3w  3w
Q   [  2 ]
a x 2 a  3

1  v  N Nx
    w    - - - -(B)
a    Ed Ed

Ed  v 
  w   N
a   
 we......will......get..........F 3

 v  d2  4
w  4
w  4
w

   w 
4 2
  12a 2 [ a x 4  2a x 2  2   4 ]  0
 
 D - K - J - equation
f f
 f o
&  f ' convention
 x
  2v w  1   2u 2  v
2
consider F2      a a 0
2 
 
2
  2  x x 

oo o1 o 2 2
(v - w )  (au ' a v' ' )  0 Differenti ate with respect to 2
2 x
1
(v ' '-w ' ' )  (au o ' ' ' a 2v' ' ' ' )  0 - - - - M
oo o
2
consider.. ... F1 2  2
u 1   2
u  2
v 
a   a 0
x 2 2   2 x 
1 oo 2
a u ' ' (u  av o ' )  0
2
operate (a )  ao '
2 x
1
a 3u o ' ' ' (au ooo '  a 2v oo ' ' )  0 - -M1
2
 Now consider F2
2 1
operate 2 (v oo
 w )  ( au o ' a 2 v' ' )  0
o

 2
oooo 1ooo ooo 2 oo
(v  w )  ( au ' a v '')  0 - - - N
2
1
a u ' ' ' (au ooo '  a 2v oo ' ' )  0 - -M1
3 o
2
By comparing M1 & N a 3u o ' ' '  v oooo  wooo
Now consider equation M
1
(v ' '-w ' ' )  (au o ' ' ' a 2 v' ' ' ' )  0
oo o

2
3 o
1 a u '''
(v oo ' '-w o ' ' )  ( 2  a 2 v' ' ' ' )  0
2 a
 1 1
 
(v ' '-w ' ' )   2 v oooo  wooo  a 2 v ' ' ' '  0
oo o

2 a 

ooo
1 1 w
(v oo ' '  2
v oooo
 a 2
v ' ' ' ' )  w' ' o

2a 2 2a 2
2
multiply it with 2a

(v oo  a 2v ' ' ) 2  2a 2 w' 'o  wooo

 (v oo  a 2 v' ' ) 2  2a 2 w' 'o  wooo
2
  2
a  2 2


   v  w000  2a 2 w''.
  2
 x 2


operate

2
   2
a  2 2

   v  w0000  2a 2 w0''
   2
 x 2

 w0000  2a 2 w'' 00  a 4 w''''  a 4 w''''

 w 00 2
 a w''  2
 a 4 w''''
2
  a  
2
4
2 2
 2  2 
w  a w''''
  x 
 2
 a  
2 2 2
4
 2  2 
w  a w' ' ' '
  x 
2
   0
 
2 2 2
a 
 2   v  w  a 4 w' ' ' '
   x 2

consider F3
 v  d 2  a 4 4 w  4
w  4
w
  w  
2
 2a 2
 4   0
   x  
4 2 2
 12a  x
 2
d 2
  2
a  2 2

 v  w 
    w  0
12a 2   2
 x 2

2
  a  2 2 2

on it operate it with   
  2
 x 2

2 4
  2
a 
2 2
 d 2
  2
a  2 2

    v  w     w  0
  2
 x 2
 12 a 2
  2
 x 2

put from
4
d 2
  2
a  2 2

 a w' ' ' '
4
   w  0
12a 2   2
 x 2

 12a 2
multiply with [ 2 ]
d
4
12a 6
  2
a 
2 2

w' ' ' ' 2   w  0
d 2
   x 2

d2
put 2

12a
4
a 
2
 
2 2
a4
 2  2  w w' ' ' '  0
 x   
Donell's equation
DKJ .........characteristic....Eq
nx m na 
w  He cos  n  
a  l 
Donnel eqatuion can be written

a w
8 ''''''''
 4a 6 w''''''..  6a 4 w....''''  4a 4 w......''  w........  
a4

w' ' ' '  0

Now assume solution is

 2 n 4 
 m  n  
2 4
0
  
DKJ characteristics roots of equation
1
 1
 2
1 1
m    n 2   n  

4
 
1  l 

 2
 
1
n 2
 1
 8

n 2
k

as Jakcobson parameter
 m1  1  i1  m5

m2  1  i1  m6

m3   2  i 2  m7

m4   2  i 2   m8
 1

1 
 
1  1  k 2  2 

 1  1  k 
2 
2
  
84
1

1 
 
1  1  k 2  2  
 1  1  k 
2 
2
  
84
1

2 
 
1  1  k 2  2 

 1  1  k 
2 
2
  
84
1

2 
 
1  1  k 2  2  
 1  1  k 
2 
2
  
84
SCHORER THEORY ()1936

•All The Assumptions Of Shell

•Like Finsterwalder Mx=0, Mx=0,Qx=0
•Contribution of Schorer
Tangential ....strain  0
1  v 
    w   0
a   
v
w .......... .......... .. Z 1

Shear strain assumed zero
 1 u v 
 x      shear strain
 a  x 
 1 u v 
  0
 a  x 
v 1 u
 .... Z2
x a 
Apart from all,
Schorer theory Applicable to Long shell
Regarding
Moment Curvature relation
D 2w
M   2
a  2

Assumed Similar to D - K - J (Plate theory)

Shorer' s equation of Equilirium (four and same Finsterwalder)
Fx  0
Nx 1 Nx
  0   1
x a 
F  0
N Nx
a  02
 x
F normal inward  0
Q
N   03

 Sum of moment @ AD  0
1 M
 Q  0    4
a 
Schorer Differential eqution
consider 4th equation
1 M
Q 
a 
2
D w D
M   2 2   2 w 
a  a
 D
M   2 w
 a
D
Q   3 w    A1
a
Consider second equation
N Nx
a 0
 x
Nx 1 N

x a 
 D  
 3 w 
Nx 1 a 

x a 
Nx D 
 4w     A3
x a
recall stress - strain relation
u Nx N
 
x Ed Ed
 0
Nx u 2
 operate 2
Ed x x
1  2 Nx  3u
2
 3
Ed x x
 2 Nx '''
2
 Edu    A4
x
From equation No 1
Nx 1 Nx
 0
x a 

operate
x
 2 Nx 1  2 Nx
 0
x 2
a x
 2 Nx 1   Nx 
     Re place from A3
x 2
a   x 
 2 Nx 1    Dw  


x 2
a   a 4 

 2 Nx  Dw  
    A5

x 2  a5 
 
From.....Schorer .....Z 2
v 1 u

x a 
3
1 
v '   u operate
a x 3
1
v ' ' ' '   u  i.e.
a
1   2 Nx 
v    2 
   Replace A5
aEd  x 
1  Dw
 

 

v   5 
aEd  a 
 
D 
v   6     A6
w
a Ed
From Schorer Assupmtion Z1
v 
w  w   v replace A6

 
  D
  6
  a Ed

w 
  6
D
a Ed
w
Rearranging
6
  Ed w  0
w a
D
6
  Ed 12w  0
w a
Ed 3
  12 a 6
 12a 2 1 
w
2
w  0  2  
d  d k
4
a
  w  0    SDE
w
k
Schorer also assumed the solution in the from
 The schorer characteristic equation
m nx
w  He cos

 
1
a 1  2 1 2
1
84

 
1
1  1
2 1 2

m1  1  i1   m5 84
m2  1  i1   m6 
 
1
2  1
2 1 2

m3   2  i 2   m7 84

 
1

m4   2  i 2   m8 2  1
2 1 2

84
Comments on Schorer Tehory
Equation of schorer can be had from
equation of Finsterwalder
(Provided all the lower derivatives with  are
ignored in comparison with 8th order by this
The Ist and Last Termin Fin- walder Eqwhich
assumes the same form of SCH. EQN

Now in order to have the above

 2
 na 
 n    
2

 l 
na
n 
l
a
n 
l
n must be less than or equal to 1 (n  1)
or l/a  
Schorer theory applicable l/a   .
It is a special case of D - K - J
k0
Schorer theory applicable.....only ....
when l/a   .
Schorer ....theory...can..be
special case of D - K - J...
if
k0
 1
 
  
1  1  1  k 2  1   1  k 2 
2
 2
  
8 4

put k  0

 
1
1  1
2 1 2
84
which is nothing but characteristic roots of schorer
1
 
  
1  1  1  k 2  1   1  k 2 
2
 2
  
8 4

1
 
  
 2  1  1 k 2  1   1  k 2 
2
 2
  
8 4

1
 
  
 2  1  1 k 2  1   1  k 2 
2
 2
  
8 4
 ‘Long’ and Short Shell
• Long/Short (as no limits)
Several options
‘Lundgreen – Shell as beam curved in cross section. Hence Bending theory can apply

Limiting proportion for long to short was attempted (where distance of Longitudinal
stresses approaches to linear distance.)
For edge beam
l/a>= 3 & without edge beam
l/a >= 5
1/a exceeding above is called ‘Long’
In classical theory
Mx=0, Mx=0, Qx=0
l/a>=
Shell termed as long
 ‘Long’ and Short Shell
• Long/Short (as no limits) very
diffult

Several options

‘ Lundgreen – Shell as beam curved in cross section.

Hence Bending theory can apply. (Base)
Researchers tried to define Limiting proportions for
long to short (where distribution of Longitudinal
stresses approaches to linear distance. )
This was attended if
For edge beam ( Length/ radius ) l/a>= 3 without
edge beam l/a >= 5
l/a exceeding above is called ‘Long’
We know the classical Theories like
Finsterwalder , DKJ, Schorer
In classical theory
FINSTERWALDER Mx=0, Mx=0, Qx=0
DKJ, own assumption
, Schorer own assumption
Hence validty is a matter of concerned
Schorer and Finsterwalder Theories
are applied to long shell onlyfixed a
limit

Schorer fixed a limit l/a>=

This limit is frequently used to
classify shell in to long and
short
Classification was also attempted on the
basis of the of the extend to which the
distubancees emerging from the straight
edge penetrating in to body of shell

This was the base for ASCE

If l/a is less than 1.6 then short
The similar creteria was proposed by
Ribich

RABICH APPROACH

If B/(l2ad)1/4 < 3 shell is long

If B/(l2ad)1/4 < 5 shell is short

from 3 to 5 (Quasi shell)

Aas – Jakobsen was the first one to classify
the shell
The shell is classified as long, short or
intermediate on the basis of value of  and k.

For long shell

=4 to 7 and K=0.03 to 0.12

For short shell

= 10 to 20 and K = 0.15 to 0.3

For intermediate shell ( ρ = 7 to 10 ) And

intermediate shell are rarely used in practice
(This is matching with ASCE & RABICH)
Summary

Finally l/a>= Schorer

or Finsterwalder

DKJ for short say

l/a<=1.6