noc18-me58 Lec 67-Shear Bucking in Rectangular Composite Plate (Part-I)

IIT Kanpur July 201
20 followers
Dec 3, 2023
Hello, welcome to advanced composites, today is the start of the last week of this course
that is the twelfth week of this particular course and what we plan to do over the span
of this course is finish our discussion on buckling which we have had in the specifically
in the last week and then in the remaining part of the week, we will introduce a new
topic related to short fiber composites because these types of composites are very widely
used in the industry and we will discuss how to analyze these types of composites also.
So, the first part we will discuss a little bit more about buckling phenomena and then
we will move into the area of short fiber composite and with this we will have a closure
to this particular course on advanced composites. So, earlier we have discussed buckling of
composite plates when they are subjected to compressive loads and the direction of the
compressive loads is normal to the plane of the, to the edge. So, if this is the edge
then the compressive force is normal to this edge, it may also be normal to the other edge.
Today, we are going to discuss a slightly different type of buckling known as shear
buckling when in this the plate is subjected to in-plane shear loads and because of those
shear loads also the plate may buckle if the shear loads exceed a particular threshold.
So, this is a type of buckling which we will like to understand better especially in context
of composite plates and let us see how we analyze these types of problems.
So, our theme is shear buckling. So, to give you an idea how shear buckling works is suppose
this is a composite plate. So, if I subject it to normal loads it buckles like this or
if I subject it to the normal load in this direction then it will buckle like this, but
it may also buckle in shear because if I subject it to shear forces. So, then it tends
to do this thing. So, in this case the deflection pattern is something like this, but when it
buckles in shear then its deflection pattern is a little more complicated.
So, this is the type of buckling we will like to discuss. So, schematically if I have
a rectangular composite plate and I want to I will be subjecting it to shear and so this
is being subjected to shear stresses or shear forces on all of its edges. Now of course,
if I subject a plate to a shear it will also tend to rotate. So, I do not want that the
plate should rotate otherwise it will be free to rotate and just. So, the way what
I do is that I pin these points, corners of the plate. So, what are my boundary conditions?
So, for specifying boundary conditions first I need to know what is the coordinate system.
So, my coordinate system is this is my x axis and this is the y axis. So, at x is equal
to 0. So, this is the corner x is equal to 0 and at x is equal to a, w is equal to 0
and mx is equal to 0 which means that the plate is simply supported on these two edges
and similarly I am also going to simply support the plate on the other two edges as well.
So, if that is the case then on y is equal to 0 and y is equal to b. So, what is a? a
is the length of the plate and b is the width of the plate. So, on the other two edges also
the out of plane deflection w0 is 0 and my is 0 because it is simply supported.
But the other extra condition I am going to specify. So, that buckling actually happens
is that I am going to pin these points. So, let us call this point 1 and let us call this
point 2. So, at points 1 and at point 2 u is equal to v is equal to w is equal to 0.
So, these are all mid plane displacements and this is what the boundary conditions look
like and what is the shear force which is being applied. So, let us call that. So, I
am applying a shear force resultant and it is S on all four sides. So, it is S Newton's
per meter. So, that is the shear force resultant. Now, so the question is at what value of S
will the plate. So, once again to solve this problem, we have to solve the differential
equation associated with the z direction and before we do that we assume that the plate
is symmetric. So, lamination is symmetric and if it is symmetric, then I can say that
the B matrix is a null matrix. The other thing is that the plate is especially orthotropic
and if that is the case, then I can say that A16 is equal to A26 is equal to D16 is equal
to D26 and this is all 0. So, because of these reasons, because of symmetry, the out of plane
equilibrium equation and the in plane equilibrium equation, they are decoupled.
So, because B is 0, out of plane equilibrium equation is decoupled with in plane equilibrium
equation. So, what that means is that I can now use this only the third equation and if
I solve the third equation, then I should be able to predict at what value of S will
the system buckle. So, the partial differential equation which I am going to address is minus
N x del 2 W over del x square minus N y del 2 W over del y square minus N x y del 2 W
over del x del y plus D11 del 4 W over del x square plus 2 D12 plus 2 D66 del 4 W over
del x square del y square plus D22 del 4 W over del y 4. So, this del should be del
x 4 and all this should equal q of x y. Now, in our equation, q is equal to 0 and
because the plate is nice rectangular symmetric lamination, then because of the geometry
and all that N x inside the plate is also 0, N y is also 0 and N x y is same as S, because
the plate is nice rectangular and it is uniformly loaded on the external edges by uniformly
applied N x y whose value is S. So, at every point, we will assume that it is the same
thing. So, if that is the case, then my simplified PDE becomes minus S del 2 W over del x del
y plus D11 del 4 W over del x 4 plus 2 D12 plus 2 D66 del 4 W over del x square del y
square plus D22 del 4 W over del y 4 equals 0.
So, this is the differential equation I want to solve and I have to also meet all the boundary
conditions and if I can do that, then for some value of S, I can calculate when the
value of W will become a real number. So, we have to solve this equation and the way
we are going to solve it is that we are going to use special Galerkin method. So, if that
is the case, then what is the first step? The first step is that we have to choose a
displacement function for W. So, we have to guess a displacement function for W which
meets all the kinematic boundary conditions. So, select appropriate W. So, once again these
are all mid plane displacements. So, we assume that W 0 x y is a series and this is a m n
sin m pi x over a sin n pi y over p and this is being integrated from 1 to M capital M
and 1 to capital N. So, this W x y, it is equal to 0 on all the four edges. So, it is
0 at x is equal to 0 edge, x is equal to a edge, y is equal to 0 edge and y is equal
to b edge. So, what does that mean? All kinematic BCs satisfy. So, this is number 1. So, once
we have selected such a function, then what is the next step? We have to calculate the
residue, we have to plug this back into the equation, find out the residue and integrate
that residue, its weighted integral over the domain of the plate.
So, the second step is I integrate this thing from 0 to a, 0 to b, the residue and I just
do not integrate the residue, I integrate it after applying some appropriate weight
function and I am integrating it over the domain and I set it to 0. Now, in special
Galerkin method, this weight function is what? It is same as assumed displacement function.
What will this be? So, it will be, so what is our assumed displacement function? It is
sin m pi x over a times sin n pi y over b. So, this is same as sin m pi x over a times
sin n pi y over b. So, we actually plug this in and it will be times W m n 1, some constant.
So, what I do is, I plug it all this in. So, what is the residue? First we will develop
an expression for a residue. So, my residue is minus S times the second derivative of
W with respect to x and y. So, it is m n pi square over a b cosine pi x m over a cosine
pi y over b plus and of course, I have to multiply it by a m n. So, I will take the
a m n outside the bracket, that is not a problem and then I get all these terms related to
d's. So, d 1 1 m pi over a to the power of 4 plus 2 d 1 2 plus 2 d 6 6 m n pi square
over a b, the whole thing squared plus d 2 2 n pi over b whole thing to the power of
4 and this entire thing in parenthesis curly brackets, this is being multiplied by sin
pi x m over a sin pi y n over b and I forgot an n here. So, I am going to put it back there
and this entire thing is multiplied by a m n and it is being added up, summed from 0
to M capital M and 0 to or 1 to capital N. So, this is my residue. So, this is this long
expression of residue I put it here and weight function also I put m pi x m pi y.
So, one thing I will make a distinction that here I will not put indices m and n, but rather
I will put i and j and then I will just iterate. So, total number of unknowns. So, if I have
to, so what is the unknown, what is it that we are trying to calculate? We are trying
to calculate, we are trying to calculate W 0 x y, we already know these functions, the
thing which we do not know is a m n. So, if m and n these are the values then total
unknowns, total number of unknowns is what m times n. So, what we are interested in finding
out is m times n unknowns which are a m n. So, a 11, a 12 and so on and so forth till
a capital M and a capital N. So, what do I do? The way we do it is that from this we
have to generate m times n equations. So, the way we generate it is that first I multiply
the residue by sin pi x over a. So, I set the value of i and j as 1, I get 1 equation.
Then in the second equation I put i is equal to 1 and j is equal to 2 and like this I keep
on increasing j till it becomes capital N, then I change the value of i to 2 and then
I again. So, these are trial functions or weight functions and we can generate these
m times n equations by selecting different values of i and j and I can change this index
i from 1 to capital M and j from 1 to capital N and in that way I can generate m times n
equations. If I do all the math, finally what we get is this is the equation we get in a
general way, pi to the power of 4 times d 11 m to the power of 4 plus 2 d 12 plus 2
d 66 m square n square r square plus d 22 n 4 r 4 a m n minus 32 m n r cube b square
s summation of m i j a i j equals 0. So, this is a long equation and this is not just
one equation. So, here i goes from 1 to m and j goes from 1 to. So, this single relation
it represents actually m times n relations, it is not one equation. So, these are different
m times n equations and there is a typo here. So, here it should be r to the power of 4
and what is r? r is the aspect ratio a over b and then you will say what is s, s we already
know it is the external shear resultant and then what is m i j. So, m i j if we do the
integral it is equal to i times j divided by m square minus i square times n square
minus j square. So, this is the value of m i j if m plus or minus i is odd and n plus
minus j is odd. So, both of the conditions m plus minus i should be odd and also n plus
minus j should also be odd, if that is not the case then. So, this is the case else.
So, if either one of these m plus minus i or n plus minus j is even then this m i j
is equal to 0. So, in this way we get m plus n equations and then we solve for these equations,
we solve for these equations. So, this is what we will actually discuss further tomorrow.
So, what we have done till so far is that using the equilibrium equation for the third
direction we have developed this longish expression and this one single expression represents
m times n equations and using this m times n equations we can calculate the critical
value of s at which the plate will start to buckle and how is it that we are going to
calculate that is something we will see in tomorrow's lecture.
Thank you and I look forward to seeing you tomorrow.
