Professional Documents
Culture Documents
Solid mechanics covers the analysis techniques to determine how structures transmit load
(force) and how they deform under load. There is a close link to Materials (SESG1003),
which covers aspects of elastic deformation, failure and factors that control life (fatigue,
creep).
Throughout SESG 1001 we will be invoking three basic principles (sometimes called the
“three great principles”):
1) Equilibrium of forces
2) Compatibility of displacements
3) Constitutive behavior of materials
We will see how these apply to several classes of structure during the module.
Finally, we will be constructing models for how structures behave. These are
idealizations (simplifications) of the real structure that enable us to analyze it. A key skill
for an engineer is to be able to develop models for physical situations that allow him/her
to design and analyze engineering systems. For the moment we will rely on the
experience and wisdom of those who have gone before us to examine existing models. In
due course we will start developing our own models.
1.1 Trusses
A truss is a structural configuration consisting of bars connected at joints. Generally these
are 3-D, but we will only consider 2-D cases in SESG1001. This will start us thinking
about models for structures and how we can analyze internal forces in structures.
EXAMPLE 1: A simple truss (3 bar truss). Note idealizations at support points: pins
and/or pins on rollers:
+
! F1 = 0 ""# $ HA + HB = 0 (1)
! M3 = 0 A+ ⇒ 10 • 200 ! 5 HB = 0 (3)
HB = 400N
From (1)
H A = !400 N
Is this statically determinate? – so far yes, we have been able to evaluate all reactions
using equilibrium equations.
Three reactions apply to the three degrees of freedom (translation in x1 and x2 and
rotation about x3) give three equilibrium equations that we can write down.
Page 4
Redrawing FBD
Draw FBD for joint C, replace bars by equivalent forces (FCB and FCA)
200 N
5
! = tan "1 = 26.6!
FCB 10
5 1 10 2
sin ! = = , cos! = =
θ 125 5 125 5
FCA
Apply equilibrium
+
!F1 ""# = 0 $ %FCB cos& % FCA = 0
!F2 ' + = 0 $ FCB sin& + 200 = 0
FCB = %200 ( 5 = %447N
NOTE: there is no opportunity to apply equilibrium of moments, all forces meet at C.
Page 5
Substitute in !F1
2
! "("200 5) " FCA = 0
5
⇒ FCA = +400N ⇐
FCB is positive i.e. tensile, pulls out of the joint, FCB is negative, i.e. compression, pushes
into the joint. Redraw joint:
200√5 N
200 N
400 N
Now consider joint, B:
200 N
400 N θ
FBC
FBA
Apply equilibrium
FBA = 0
So what is the purpose of FBA ?
EXAMPLE 2
Problem: to find force in bar ED
E D
L
45°
A C
2L 2L
2P
First draw FBD
E D
x2 45°
A C HC
x1 ~ 2L
~
2L
2P
VA
VC
E F ED D
Closed Must
surface include
45° F AB reactions
A crossing
B surface
F EB
2L
We want to find FED - try to find one equation with only one unknown
!
! F2 = 0 " + # FEBSin45 + P = 0
!= 1 #
" 2$
FEP = + 2P
Suppose we had wanted FAB
Is this correct?
- Think about cutting members ED - shortens AB – lengthens
NOTE:
In both methods of joints, method of sections, we solve for the internal forces by isolating
part of the structure.
Two Tips:
• Reduce computation by intelligent choice of method and section to analyze
1 equation, 1 unknown
• Check, check and double check as you go!
FL
!= + "#TL
AE
A= Cross-section area
[ L2 ] m2
α= Coefficient of thermal
expansion
[ L L.! ] M
MK
ΔT = Temperature difference [! ] K
FL " AE $
!= and rearranging gives # !=F
AE L %
AC +400 10 571 6
AE = 7 ! 10 N
BC -447 125 -714
Page 12
Note; the bar extensions are very small compared to bar lengths
Consider what this implies about deformations of 3 bar truss.
Definition Stress is the measure of the intensity (per area) of force acting at a point.
Use coordinate system x, y, z ! x1, x2 , x 3
Example:
!A = A nm
The stress is the intensity, force per unit area, so let m and n go to infinity.
Force, F, is a
vector quantity
F = F1 i1 + F2 i 2 + F3 i 3
Then take the limit as the force on the face is carried by a smaller set of areas.
lim !F
!A " 0 =0 (but stress vector has other components)
!A
! = ! 11 i1 + ! 12 i 2 + ! 13 i 3
Where ! 1 is the stress vector and !1n are the components acting on the x1 normal face, in the x1,
x2 and x3 directions.
! 2 = ! 21 i 2 + ! 22 i 2 + ! 23 i 3 on i2 face
! 3 = ! 31 i1 + ! 32 i 2 + ! 33 i 3 on i3 face
! m = ! mn i n
Subscripts or indices
By Convention
Latin subscripts m, n, i, j, p, q, etc.
Take values 1, 2, 3, (3-D)
2. A subscript occurring once in a term is called a free index, can take on the range 1, 2, (3) but
is not summed. It represents separate equations.
x1 = F11Y1 + F12Y2 + F13Y 3
x2 = F21Y1 + F22Y2 + F23Y 3
x 3 = F31Y1 + F32Y2 + F33Y 3
3. No index can appear in a term more than twice.
! mn
Stress acts on
Acts in xn direction
face with
normal in the
xm-direction
Page 4
Types of Stress
We can identify two distinct types of stress:
#1 & #1 &
2! 23"x 3 "x 1 % "x 2 ( ) 2! 32"x 2 "x 1 % "x 3 ( = 0
$2 ' $2 '
simplifies to: ! 23 = ! 32
Thus, in general ! mn = ! nm
Stress tensor is symmetric. Six independent components of the stress tensor.
! 11 ! 12 " = ! 21 %
! 22 ! 23 $$ = ! 32 ''
! 33 ! 31 # = ! 13 &
What happens if stress varies with position?
x3 !11
#x2
"! 31
! 31 + #x3
"x 3
"! 21
!21 ! 21 + #x2
"x2
#x3 x2
#x1
!31
x1
"! 11 f1
! 11 + #x
" x1 1
Force equilibrium ! F1 = 0 +
#
% ! 11 + "! 11 dx1 &( ( dx2 dx3 ) ) ! 11( dx2 dx3 )
$ "x1 '
# "! &
+ % ! 21 + 21 dx2 ( (dx1dx2 ) ) ! 21 (dx1dx 3 )
$ dx 2 '
# "! &
+ % ! 31 + 31 dx 3( (dx1dx2 ) ) (! 31 )( dx1dx 2 )
$ "x3 '
+ f1dx1dx2 dx3 = 0
Canceling out terms and dividing through by the common: dx1, dx 2 , dx3 .
Gives:
!"11 !" 21 !" 31
+ + + f1 = 0 x1 direction
!x1 !x 2 !x3
Similarly,
!" 12 !" 22 !" 32
+ + + f2 = 0 x2 direction
! x1 ! x2 ! x3
Page 6
!"13 !" 23 !" 33
+ + + f3 = 0 x3 direction
!x1 !x2 !x3
!" mn
+ fn + 0
!xm
Stress Transformations
Just as we need to resolve forces with respect to structural axes - need to resolve stress. Identify
directions which "see" maximum stress
!˜11dx3dx˜ 2 " !11dx3 [#x˜ 2 cos$ ]cos $ " ! 22 dx3[#x˜2 sin $ ] sin$ " !12 dx3[#x˜2 sin $ ] cos$ " ! 21dx3[#x˜2 cos$ ] sin$ = 0
Note: two angles: one for area; one for resulting force
Tensor form
Transformation requires direction cosines (see below)
Stresses are second order tensors (2 subscripts) and require two direction cosines for
transformation.
Thus:
!˜ mn = ! m˜ p! n˜ q! pq
• 2nd order tensor requires 2 direction cosines
• transforms stress from xn system to x˜m system
one angle for area, one for component of force
Direction Cosines
Page 8
Define direction cosine as:
Angle between
! m˜n = cos[ x m x˜ n ]
such that
xm = ! m˜n x n˜
Tensor equations
or x˜n = !n˜ m x m
where we mean
x1 = !1 ˜1x˜1˜ + !12˜ x˜ ˜2 + !13˜ x˜3˜
x2 = ! 2˜n x˜ n˜
x3 = ! 3˜n x˜n˜
2-D Example
y1 y2 y3
One further way to transform stress is to use "Mohr's Circle" graphical method for rotating 2nd
order tensors in 2-D coordinate system
G C F Ext
2!p
A
H'
SESG1001 Lectures 7 & 8: Strain
We have examined stress, the continuum generalization of forces, now let's look at the
continuum generalization of deformations:
Definition of Strain
Strain is the deformation of the continuum at a point. Or, the relative deformation of an
infinitessimal element.
Ldeformed ! Lundeformed
tensile strain =
Lundeformed
Shear
This produces an angle change in the body (with no rotation for pure shear)
Consider infinitesimal element.
Undeformed:
Deformed :
Good for range of use of most engineering materials, most structural applications.
Allows us to neglect higher order terms and leads to:
Strain - Displacement Relations (for small strains)
Conceptually, want to separate out rigid body translations from
deformations
Consider infinitesimal element, side length dx1 undergoing displacements and deformations
in the x1 direction defined by
u1 is a field variable = u1( x1,x 2 , x3 ) , i.e. displacements vary with position
let u1 be the displacement of the left-hand side of the infinitesimal element.
" !u1 %
And $ u1 + dx1 'is the displacement of right-hand side
# !x1 &
Recall:
! deformed # ! undeformed
!11 " !1 =
! undeformed
! undeformed
% $u (
' 1 1
dx + u + dx1 # u1* # dx1
& $x1 )
!11 =
dx1
Rigid body
$u translation
= 1
$x1
"u 3
and ! 33 =
"u 3
=
1
2
( undeformed - deformed )
1$! $! ''
= & " & " #12 ))
2% 2 % 2 ((
e.g:
NOTE: The strain tensor is defined such that there seem to be two parts in each angle
change. But the stress tensor is symmetric and we would like the strain tensor also to be
symmetric. Thus:
1
!12 = ! 21 = ( angle change in x1 " x 2 plane)
2
1
!13 = ! 31 = (angle change in x1 " x3 plane)
2
1
! 23 = ! 32 = ( angle change in x 2 " x3 plane)
2
Formal Definitions
"u1
u1 + dx2 # u1
"x2
!1 =
Hence: dx2
"u
u2 + 2 dx1 # u2
"x1
!2 =
dx1
!12
(We have dealt with elongation and shear strains but remember, we were concerned to
eliminate translation and rotational displacements)
"u 2
u2 + dx # u
"x1 1 2 "u2
!1 = = as before
dx1 "x1
$ #u '
" %u1 + 1 dx2 " u1(
& #x2 ) "#u1
or !2 = =
dx2 #x 2
But may also have shear deformation of the cube !1 " ! 2
Define average rotation (of diagonal)
1 $"u2 "u1 '
! = % # ( note (-) sign
2 & "x1 "x 2 )
Transformation of Strain
Identical to transformation of stress
!˜mn = ! m˜ p ! n˜ q! pq
Mohr's Circle in 2-D
Principal strains/directions
-no associated shear strains
Alternative notations for stress and strain
!11 " !x
! 22 " !y
Matrix Notation
"! 1 1 ! 1 2 ! 1 3 %
" !11 !12 !13 % $ '
! mn = $ ! 21 ! 22 ! 23 ' ! mn = $! 2 1 ! 2 2 ! 2 3'
$ '
$# ! ' $#! 3 1 ! 3 2 ! 3 3'&
31 ! 32 ! 33 &
- principal values are eigenvalues
- principal directions are eigenvectors
Next we will link stress and strain by constitutive behavior and to do this we will need to
look at material properties.
SESG1001 Lectures 9 & 10: Elastic Stress-Strain Relations
These lectures will complete our quick journey through continuum mechanics, to provide
you with a continuum version of a constitutive law - at least for linear elastic materials
! pq = E ? " mn
Elasticity
Where does it
come from?
Increasingly, materials are designed along with the structure, you need insight into what
contributes to material properties. These lectures will link to topics you will learn about
in materials lectures and form a foundation for material selection in design.
transverse strain !T . The axial strain will be tensile for a tensile applied stress, and the
transverse strain will usually be compressive. We can measure the strains using
resistance strain gauges.
For many materials, over some range of applied stress, the applied stress and the resulting
strains will follow a linear relationship. This observation is the basis for the definition of
the engineering elastic constants. The Young’s modulus, E, is defined as the constant of
proportionality between a uniaxial applied stress and the resulting axial strain, i.e:
! L = E" L
Note. This only applies for a uniaxial applied stress, and the component of strain in the
direction of the applied stress.
We can also define the Poisson’s ratio, ! , as the ratio of the transverse strain to the axial
strain. Since for the vast majority of materials the transverse strain is compressive for a
tensile applied stress, the Poisson’s ratio is defined as the negative of this ratio, to give a
positive quantity. I.e:
#
! =" T
#L
An applied shear stress leads to an applied shear strain. The shear strain, ! , is defined in
engineering notation, and therefore equals the total change in angle: ! = ".
Consistent with the definition of the Young’s modulus, the Shear modulus. G, is defined
as:
! = G"
Again, note, that this relationship only holds if a pure shear is applied to a specimen.
For reasons that will become apparent shortly, The Young’s modulus, Shear Modulus and
Poisson’s ratio are linked. For most materials Poission’s ratio’s are approximately 0.33,
3
and for these materials G ! E . However, values of the Young’s modulus can vary
8
widely.
From: Material Selection in Mechanical Design, M.F Ashby, Pergamon Press, Oxford,
1992
Note, how the different classes of material tend to cluster: Metals, Polymer, Glasses and
ceramics have similar ranges of density and modulus. There are also some materials that
have quite wide ranges of moduli (and densities), while others are relatively narrowly
banded. Finally note how wide an overall range of moduli is represented, from 0.01 GPA
for foams to 1000 GPA for diamond. The range of densities is somewhat less, but still
spans more than two orders of magnitude. See SESG1003 for more details.
Now we want to add a mathematical formalism to this physical basis, i.e. our 3rd great
principle, that of constitutive behavior.
A couple of problems we would like to be able to solve: a) what are the strains due to
stresses at an arbitrary angle to the loading direction:
b) We would also like to be able to deal with any state of multiaxial stress and convert to
the resulting strains, or vice versa. To do this we need to revisit tensor stress and strain.
i.e. we want the elastic property that links the stress tensor to the strain tensor:
σmn=Emnpqεpq
Where Emnpq is the 4th order (i.e. 4 subscripts) ELASTICITY (or STIFFNESS) tensor.
e.g:
!11 = E1111"11 + E1112"12 + E1113"13 ( p = 1,sum on q)
"! 11 & ) E1111 E1122 E1133 2E1123 2E1113 2E1112 , " /11 &
$! $ + E
22 1122 E2222 E2233 2E2223 2E2213 2E2212 . $/ 22 $
$ $ + .$ $
$! 33 $ + E1133 E2233 E3333 2E3323 2E3313 2E3312 . $/ 33 $
# '= + # '
$ ! 23 $ + E1123 E2223 E3323 2E2323 2E1323 2E1223 . $/ 23 $
.
$! 13 $ + E1113 E2213 E3313 2E1323 2E1313 2E1213 . $ /13 $
$ $ + .$ $
%! 12 ( * E1112 E2212 E3312 2E1223 2E1213 2E1212 - %/12 (
Since it links strains to stresses, Emnpq is also termed the "stiffness tensor". By inverting
the matrix we could also obtain the tensor linking stresses to strains, i.e. εmn=Smnpqσpq
Where Smnpq is the "Compliance" tensor (Why S is for compliance and E is used for
stiffness is unclear to me!)
Even with the simplifications, 21 independent terms seems rather too many to have to
deal with!. Let's go back to the engineering elastic constants and see if we can see how to
simplify this list further.
We know that there are several different classes of material. Most metals and ceramics
are isotropic, that is they have the same properties in any direction that you measure. By
contrast, fiber-reinforced composites may have different properties in different directions,
i.e. they are anisotropic, generalized elasticity provides the framework to analyze these.
We also know that for small strains, and elastic materials the contributions of the separate
components of stress will superimpose. So let's consider the case of only σx applied and
all the other components of strain are zero:
$
And εy , εz are given by the Poisson contractions, so: ! y = ! z = "#! x = "# x
E
This allows us to fill in the first line of the matrix, and also by noticing that we could
have equally well applied σy or σz, and obtained similar relationships, we can fill in all of
the top left hand quadrant of the matrix:
# !x & # 1 )* )*
_ _ _& # + x &
% % E E E (% (
! y ( % )* 1 )*
_ _ _( % + y (
% ( E E E
% z ( % )*
! )* (%+
_( z (
1
=% E _ _
% % (
" zy ( %
E E
,
% ( _ _ _ _ _ _ ( % zy (
% " zx ( % ( % , zx (
_ _ _ _ _ _
% ( %% (( % (
$ " xy ' $ _ _ _ _ _
,
_ ' $ xy '
If instead of applying an extensional stress we applied a shear stress, we know that the
shear stress and shear strain are linked by the shear modulus, so:
# 1 )* )*
_ _ _ &#+
# !x & % E E E ( x&
%
!y ( % )* 1 )*
_ _ _ (%+y (
% ( % )E*
E E
(% (
)* 1
% !z ( % _ _ _ ( % +z (
% =% E E E
% (
" zy ( _ _ _
1
0 0 ( % , zy
% ( % G ( (
% " zx ( % _ _ _ 0
1
0 ( % , zx (
% ( G % (
$ " xy ' % 1 (( $ , xy '
% _
_ _ 0 0
$ G'
Finally we note that for isotropic materials the application of an extensional stress does
not result in a shear strain or vice versa, so the top right and bottom left quadrants are
populated by zeros:
# 1 )* )*
0 0 0& +
# !x & % E E E (# x &
% )* )*
0 (%+y (
1
!y ( % E 0 0
% ( % )* E E
(% (
)* 1
% !z ( % 0 0 0 ( % +z (
% ( =% E E E
% (
0 ( % , zy
" 1
% zy ( % 0 0 0
G
0
( (
% " zx ( % 0 0 0 0
1 % ,
0 ( zx (
% ( G % (
$ " xy ' %% 1 (( $ , xy '
$ 0 0 0 0 0
G'
In terms of the Mohr's circles of stress and strain these appear as (note the factor of two
for the representation of shear strain on the Mohr's circle):
For the biaxial tensile and compressive stress the resulting (principal) strain is given by:
1
but for the case of pure shear: ! I = " (From Mohr's Circle -remember the
2
factor of two between tensor and engineering shear strain)
and:
! = " II = #" I
'1 &
$% = 2( + )* ! + G =
E
E E 2(1 + & )
So we actually only have two independent elastic constants for an isotropic material.
Note that this only applies for isotropic materials
If we want to go in the reverse direction (i.e. have known strains and want to calculate
stresses) we need to invert the matrix of elastic constants. Note, this situation may arise
because we can experimentally measure strains using strain gauges. The inverse matrix
is usually expressed in terms of groupings of the elastic constant, known as Lamé's
constants, µ & ! , where:
!
µ= =G
2(1+ " )
"E
#=
(1 + " )(1 $ 2" )
# ! x & # ) + 2µ ) ) 0 0 0& # * x &
% ( % % (
! ) ) + 2µ ) 0 0 0( * y
% y( % (% (
% !z( % ) ) ) + 2µ 0 0 0 ( % *z (
Thus: % ( = % % (
" 0 0 0 µ 0 0 ( + xy
% xy ( % (% (
% " xz ( % 0 0 0 0 µ 0 ( % + xz (
% ( % (% (
$ " yz ' $ 0 0 0 0 0 µ' $ + yz '
The key feature of all these structural members is that one dimension is longer than the
others (i.e. they are one dimensional).
There is a basic logical set of steps that we will follow for each in turn:
1) We will make general modeling assumptions for the particular class of structural
member
In general these will be on:
a) Geometry
b) Loading/Stress State
c) Deformation/Strain State
First start with a working definition - from which we will derive our modeling
assumptions:
"A rod (or bar) is a structural member which is long and slender and is capable of
carrying load along its axis via elongation"
Modeling assumptions
a.) Geometry
(think about the implications of this - what does it imply about the magnitudes of stresses
and strains)
! 31 = 0 = !13
! 32 = 0 = ! 23
! 33 = 0
and
" ! 11dA = P
A
P P
# ! 11 = =
bh A
c.) deformation
Rod cross-section deforms uniformly (is this assumption justified? - yes, there are no
shear stresses)
So much for modeling assumptions, Now let's apply governing equations and solve.
1. Equilibrium
!" mn
+ fn = 0
!x n
!" 11
+ f1 = 0 f1 = body force =0
!x1
!" 11
= 0 # " 11 = constant = P
!x1 A
Constitutive Laws
stress - strain equations:
"1%
! 1 1 = $ ' (1 1
# E&
" )* %
! 2 2 = $ ' (1 1
# E&
" )* %
! 3 3 = $ '( 1 1
# E&
Compliance Form
Hence:
! 11 "u1
=
E "x1
"u1
# P AE =
"x1
integration gives:
Px1
u1 = + g(x2 , x3 )
AE
Apply B. C. u1 = 0 @ x1 = 0 ! g(x2 , x 3 ) = 0
Px1
! u1 =
AE
!"P
similarly u2 = x2
AE
!"P
u3 = x3
AE
Further assumptions
u1 = 01 , u2 = 01, u3 = 0
"Remote from the boundary conditions internal stresses and deformations will be
insensitive to the exact form of the boundary condition."
This is the importance of the "long slender" wording of the rod definition.
This should have been all fairly obvious. Next time we will start an equivalent process
for beams - which require a little more thought.
SESG1001 Lectures 12 and 13
Beam Theory: Bending Moment and Shear Force
Distributions
Reading: Benham, Crawford and Armstrong, 1.10, Ch 6 and Ch 7
A beam is a structural member which is long and slender and is capable of carrying
bending loads. i.e loads applied transverse to its long axis.
Beams are very common structural members and are to be found across all areas of
engineering. Wings, masts, gas turbine blades even whole ships, aircraft or spacecraft
can be treated as beams to varying degrees.
c) Deformation
We will talk about this later
2) Boundary conditions
As for rods, trusses
Pinned, simply supported, point load Cantilever, built in support, distributed load
Analysis approach: Draw free body diagram, apply equilibrium to determine reactions.
Before moving through this process we need to understand how beams transmit load.
Opposite
directions on two
faces - equilibrium
where M = Bending movement Beam
S = Shear force Bending "beam bar"
F = Axial force bar, rod
Examples of calculating shear force and bending moment distribution along a beam.
Equilibrium: H A = 0, VA = P, M A = !PL
Take cut at point X ,distance x from left hand end (root). 0< x < L. Replace
the effect of the (discarded) right hand side of beam by an equivalent set of forces and
moments (M, S, F) which vary as a function of position, x.
Apply equilibrium
+
! Fx = 0 " F(x) = 0
! Fz = 0 # + P $ S(x) = 0 S( x) = P
! MX = 0 + PL + M( x) $ Px = 0 M( x) = $ P(L $ x)
Draw "sketches" - bending moment, shear force, loading diagrams
These representations are useful because they provide us with a visual indication of
where the internal forces on the beam are highest, which will play a role in determining
where failure might occur and how we should design the internal structure of the beam
(put more material where the forces are higher).
Example 2: A simply supported beam
P P
! " S(x) = 0 # S(x) =
2 2 0<x<L
P Px
M x = 0 + # " x + M (x) = 0 # M (x) =
2 2
L
take cut at <x<L
2
P P
" ! F+ = 0 : # P # S(x) = 0 $ S(x) = #
2 2
P % L(
"M =0: # x + P' x # * + M (x) = 0
2 & 2)
P
M (x) = ! (x ! L)
2
P
= (L ! x)
2
Draw Loading, Shear Force and Bending Moment Diagrams
Observations
Distributed loads
e.g. gravity, pressure, inertial loading. Can be uniform or varying with position.
"
q(x) = q0 # 1! $% &= q0 @ x = 0, = 0@ x = L
x
q(x) = q 0
L
[ q ! ] = [force/length]
We deal with distributed loads in essentially the same way as for point loads.
Example: Uniform distributed load, q (per unit length), applied to a simply supported
beam.
Apply equilibrium:
x
qL
"! + # Sx # $ q dx = 0
2 0
qL
% # qx = S(x)
2
qLx x
! Mx + " + Mx + # qxdx = 0
2 0
qL qx2 qLx qx2
" + Mx = = 0 $ Mx = "
2 2 2 2
Plot:
Observations
• Shear load varies linearly due to uniform distributed load.
• Moment varies quadratically (parabolically) over region where distributed load
applied
• This suggests a relationship between M & S & P
!M
M+ dx
!x
!F
F+ dx
!S !x
S + dx
!x
Now use equilibrium, replace partial derivatives by regular derivatives (F, S, M varying
only in x).
+ dF dF
! Fx = 0" : # F + F + dx
dx = 0
dx
=0
dS
! Fz = 0 " + S # S # dx + q(x)dx = 0
dx
dS
= q(x)
dx
dM dx dS dx
+ ! M0 / +M
"M / + dx " S " (S + dx) = 0
dx 2 dx 2
note: q(x) has no net moment about O.
dM 1 dS
dx ! Sdx + (dx) 2 = 0
dx 2 dx
Now we will move on to analyze how these internal forces and moments correspond to
the internal stress distribution and deformation of the beam.
Engineering Beam Theory
Reading: Benham, Crawford and Armstrong Ch 6 and Ch 7
We have looked at the statics of a beam and we have seen that loads are transmitted by
internal forces: shear forces and bending moments.
Now look at how these forces imply stresses, strains, and deflections.
Recall model assumptions: slenderness General, symmetric, cross section
Geometry
L >> h,b
Loads
Leads to assumptions on stresses
! yy = ! xy = ! yz = 0
Load in x - z plane →
Also L >> h,b implies ! xx ,! xz >> ! zz
Assumptions on Deformations:
The key to simple beam theory is the Bernoulli - Euler hypotheses (1750)
"Plane sections remain plane and perpendicular to the mid-plane after deformation."
It turns out that this is not really an assumption at all but a geometric necessity, at least
for the case of pure bending.
To see what the implications are of this, consider a beam element, which undergoes
transverse (bending)
deformation.
Note, negative sign here due to use of consistent definitions of positive directions for w, x
and dw/dx.
" xx (4)
! xx =
E
"
! yy = #$ xx
E
"
! zz = #$ xx
E
& xy
% xy = Note inconsistency -
G ! xz = 0, " xz # 0 (shear
& yz forces are non zero)
% yz =
G
&
% xz = xz
G
The inconsistency on the shear stress/strain arises from the plane/sections remain plane
assumption. Does not strictly apply when there is varying bending moment (and hence
non-zero shear force). However, displacements due to ! xz are very small compared to
!" mn
+ fm = 0
!x m
!" xx !# xy !# zx !" xx !# zx
+ + = 0$ + =0 (5)
!x !y !z !x !z
!" xy !# yy !" zy
+ + =0 $0= 0
!x !y !z
!" xz !" yz !# zz !"
+ + = 0 $ xz = 0
!x !y !z !z
Summarizing:
dw
u( x, y,z ) = !z (1)
dx
w( x, y,z ) = w( x) (2)
d 2w
! xx = "z (3)
dx 2
" xx
! xx =
E (4)
!" xx !# zx
+ =0
!x !z (5)
d 2w
! xx = E" xx = #Ez 2
dx
Now substitute this in (7)
h2
h
d 2w 2 " d 2 w z2 %
F = !E 2 ( zbdz = $!E 2 b' =0
dx !h $# dx 2 &'
2 !h 2
Which is correct, since there is no axial force in the case of pure beam bending.
Similarly in (8)
+h
2
d w 2
2
M= E 2 " z bdz
dx !h
2 .
We can write this more succinctly as:
d 2w
M = EI 2 this is the moment curvature relation - positive moment results in
dx positive curvature (upward). The quantity “EI” defines the
curvature stiffness of beam – sometimes called “flexural rigidity” – note
that it depends on both the shape and the material property, E.
h
2 2
Where I = " z bdz
!h
2
is the second moment of area (similar to moment of inertia in dynamics)
d 2w M
Recall that ! xx = "Ez 2 = "Ez
dx EI
This is the moment - stress relationship
Mz
! xx = " Implies a linear variation of stress,
I
Maximum stress at edges (top and bottom) of beam
In bending centroid is unstrained – also known as the “neutral axis” or “neutral plane”
Can also use parallel axis theorem to calculate second moments of area
I total = ! I i + ! Ai d2i
is equivalent to:
1
bh 3 b3
12
I total = I y1 y1 + A1d12 + I y 2 y 2 + I y 3 y 3 + A3 d 32
Changing area distribution by moving material from the web to the flanges increases
Returning to the derivations of simple beam theory, the one issue remaining is to
calculate the shear stresses in the beam. We would like to obtain an expression for:
! zx (z ) .
Recall:
Shear stresses linked to axial (bending) stresses via:
!" xx !# zx !# !"
+ = 0 $ zx = % xx
!x !z !z !x (5)
Also, shear force (for rectangular cross-section, width b) linked to shear stress via:
h
2
S = ! # " xzbdz
!h
2 (7)
Multiply both sides of (5) by b and integrate from z to h/2 (we want to know ! zx (z )
" h%
and we know that ! zx $ ± ' = 0 )
# 2&
h h
2 !" zx !% xx 2
"Mz
#b dz = $ # bdz ! xx =
z !z z !x Iyy
h
# "dM & zb
h 2
[ xz ]
b! (z) = " ) %$ 2
z
dx
(
' I
dz
z
dM
recall =S
dx
# zb &
h
( ( ) )
2
% (
2 " ! xz (z ) = )z S % I ( dz
b ! xz h
$ yy '
h
Note that at z = (top surface) ! zx = 0
2
Also define
h
2
Q = ! zbdz first moment of area above the center
z
Hence:
SQ
! xz (z) = "
I yy b
shear stress -force relation
For a rectangular section:
h h2
2 !z2 $ b !h 2 $
2
Q = ' zbdz = # b& = # (z &
z #" 2 &% 2 #" 4 &%
z
But need to beware cases in which we have thin walled open or closed sections where
or
We will defer discussion of these cases until 2nd year courses. Next: Look at examples of
applying Engineering Beam Theory to examples of beams.
Lectures 17 and 18 Examples of Application of Beam
Theory
Reading: Benham, Crawford and Armstrong Chs 6 and 7
Example: Calculate deflected shape and distribution of stresses in beam. Refer back to
example problem in lecture 13 & 14.
Example: Uniform distributed load, q (per unit length), applied to simply supported
beam.
qL
S(x) = ! qx :
2
qLx qx 2
M (x) = ! :
2 2
Let beam have rectangular cross-section, h thick, b wide. Material, Young’s modulus E:
d 2w d 2w M (x)
EI =M! =
dx 2 dx 2 EI
Integrate to obtain deflected shape w(x)
M (x)
w( x ) = ! ! dx dx
EI
and for EI constant over length of the beam:
1
! w( x ) = " " M (x)
EI
For our case:
!qx
M (x) = (x ! L)
2
dw 1 qx
= ! " (x ! L)dx
dx EI 2
!q )# x 3 Lx 2 &,
= ++% ! (. + C1
EI *%$ 6 4 ('.-
1 q# x 3 Lx 2 &
So w(x) = "( % ! ( + C1 )dx
EI 2%$ 6 4 ('
!q " x 4 Lx 3 %
w(x) = $ ! + C1 x + C2 '
2EI $# 24 12 '&
C1 and C2 are calculated from values of w(x) and w’(x) at boundary conditions.
Hence:
w(x) = !
q
48EI
[
x 4 ! 2Lx 3 + L3 x ]
dw(x)
Maximum deflection occurs where: =0
dx
dw L
! 4 x 3 " 6 L x2 + L3 = 0 satisified at x =
dx 2
5 qL4
w=!
384 EI yy
Find Stresses:
Mz 1 3
! xx = " , I= bh , substitute for M(x)
I yy 12
L qL2
Max tensile (bending) stress at x = , M max = , z = ±h2
2 8
3qL2
! xx max =
4bh 2
bh 2 qL 12 1 3 qL
! xzmax = =
8 2 bh 3 b 4 bh
Note that this is much smaller than ! xx if L/h is large (i.e slender beam).
max
Example
Bending moments:
"Mz
Note in passing M(x) and hence ! xx = = constant for L # x # 2L - useful for
I
mechanical testing of materials – “four point bending”
Could solve for deflections by writing separate equations for: M(x) for 0 < x < L, L <
x < 2L, 2L < x < 3L and integrating each equation separately and matching slope and
displacements at interfaces between sections. This is a tedious process.
Or we can use:.
Macaulay’s Method (BCA 7.3)
Relies on the knowledge that the integral of the moment (i.e. the slope, and deflection)
are continuous, smooth functions. Beams do not develop “kinks”. This is a particular
example of a class of mathematical functions known as “singularity functions” which are
discontinuous functions, defined by their integrals being continuously defined.
Write down equilibrium equation for section, leaving quantities such as {x-L} grouped
within curly brackets { }
{x-a} is treated as a single variable, only takes a value for x > a otherwise { } = 0
Proceed as before - integrate moment-curvature relationship twice, applying boundary
conditions
d2w
M = Px ! P(x ! 2L) ! P(x ! L) = EI
dx2
dw Px 2 P 2 P 2
EI = ! {x ! 2L} ! {x ! L} + A
dx 2 2 2
Px 3 P 3 p 3
EIw = ! {x ! 2L} ! {x ! L} + Ax + B
6 6 6
27PL3 PL3 P
x = 3L !
6
"
6
( )
" 8L3 + A x= 0
6
for
! A = "PL2
1 " Px3 P P %
$ ! { x ! 2 L} ! {x ! L} ! PL x ' (
3 3 2
w( x) =
EI $# 6 6 6 '&
Provides complete description of deflected shape of beam in one equation. Can also use
for distributed loads.
2.) Superposition
Decompose problem into several statically determinate structures.
Replace constraints by applied loads
Solve for loads required to achieve geometric constraints.
Example: cantilever beam under uniform distributed load, with pin at right hand end.
RB = ?
Equivalent to:
(1)
+
(2)
From “standard solutions” or calculation obtain expressions for deflections in case (1)
and case (2):
q! 4
!1 = " ,
8EI
R L3
!2 = + B
3EI
!1 + !2 = 0
qL4 RB L3 3qL
" = " RB =
8EI 3EI 8
Note. For superposition to work we must be superimposing the loads and boundary
conditions on the same beam. Only change one variable in each loading case.
L19 Shafts: Torsion of Circular Shafts
Reading: Benham, Crawford and Armstrong, Ch 4
A shaft is a structural member which is long and slender and subject to a torque
(moment) acting about its long axis. We will only consider circular cross-section shafts
in SESG1001. These have direct relevance to circular cross-section shafts such as drive
shafts for engines of all types. However, the basic principles are more general and will
provide you with a basis for understanding how structures with arbitrary cross-sections
carry torsional moments. Torsional stiffness, and the shear stresses that arise from
torsional loading are important for the design of aerodynamic surfaces such as wings, and
also play a key role in the design of ship hulls, automotive chassis and even hip and knee
replacement prosthetics.
Modelling assumptions
(a) Geometry (as for beam). Long slender, L >> r (b,h)
Note: For the time being we will work in tensor notation since this is all about shear
stresses and tensor notation will make the analysis more straightforward. Remember
we can choose the system of notation, coordinates to make life easy for ourselves!
(b) Loading
Torque about x1 axis, T (units of Force x length). We may also want to consider
the possibility of distributed torques (Force x length/unit length) (distributed aerodynamic
moment along a wing, torques due to individual stages of a gas turbine)
No axial loads (forces) applied to boundaries (on curved surfaces with radial normal, or
on x1 face)
!11 = ! 22 = ! 33 = 0
(c) Deformation
-Cross sections rotate as rigid bodies through twist angle φ, varies with x1 (cf. beams
– plane sections remain plane and perpendicular)
- No bending or extensional deformations in x1 direction
Cross-section:
where r = x22 + x 32
x x
sin $ = 3 cos $ = 2
r r
!
# & x3
so u2 = "% x22 + x 32 () (x1 ) = ") (x1 )x 3
$ '
x22 + x 32
" % x2
u3 = $ x22 + x 32 '( (x1 ) = ( (x1 )x2
# &
! x22 + x 32
So
!
u1 = 0 (1)
u2 = "# (x1 )x 3 (2)
u3 = # (x1 )x2 (3)
Governing Equations
du !
!11 = 1 = 0
dx1
du2
! 22 = =0 Consistent with assumption of no axial,
dx2 or radial stresses, strains
du 3
! 33 = =0
dx 3
Net moment due to shear stresses must equal resultant torque on section:
"# mn
Apply equilibrium: + fn = 0
"x m
"# 11 "# 21 "# 31
i.e.: + + =0
"x1 "x2 "x 3
! "#
12 + "# 22 + "# 32 = 0
"x1 "x2 "x 3
! "# 13 "# 23 "# 33
+ + =0
"x1 "x2 "x 3
!
!
Retaining non-zero terms we obtain
"# 21 "# 31
+ =0 (9)
" x2 "x 3
"# 12
=0 (10)
"x1
"# 13
=0 (11)
"x1
Solution
Go back to stress-displacement relationships (4,5, 6 and 7)
!
% 1 d$ (*
!12 = 2G"12 = 2G' # x 3
& 2 dx1 )
d#
!12 = "Gx3
dx1
Similarly:
$1 d# ') d#
!13 = 2G13"13 = 2G& x2 = Gx2
%2 dx1 ( dx1
% d$ d$ (*
= ## ' x22 + x32 dx dx
& dx1 dx1 ) 2 3
d!
T =G
dx1
( 2
)
2
"" x2 + x3 dA
! 2 2
define J = %% " x2 + x 3 #$ dA
!R 4
= for circular cross-section
2
d!
Hence T = GJ
dx1
Torque-twist relation
d! !2 w
(note, we can compare T = GJ for shafts with M = EI for beams)
dx1 !x 2
2 2 T Tr
" res = # 12 + # 13 = x 32 + x22 =
J J
Tr $ Mz '
compare " = with &# xx = )
J % I (
!
Model works well for circular cross-sections and cylindrical tubes J = J1 ! J2
!
Does not work for open sections:
For a full treatment of torsion of slender members see 2nd year structures classes.
L20 Buckling of Simple Columns
Benham Crawford and Armstrong: Chapter 10
Elastic instabilities, of which buckling is the most important example, are a key limitation
on structural integrity. The key feature of an elastic instability is the transition from a
stable mode of deformation with increasing applied load to an unstable one, resulting in
collapse (loss of load carrying capability) and possibly failure of the structure. Examples
of elastic collapse are the buckling of bars in a truss under compressive load, the failure
of columns under compressive load, the failure of the webs of “I” beams in shear, the
failure of skin panels or hull plates in shear and many others. The only particular case we
will consider in SESG1001 is the failure of a bar or column loaded in axial compression,
however, as for the other slender members we have considered, the basic ideas will apply
to more complex structures.
A structure is in stable equilibrium if, for all possible (small)displacements/deformations,
a restoring forces arise.
Now apply this idea to a continuous structural member.
Model of a column
(a) Geometry – identical to a beam, long, straight, slender, symmetric cross-section etc.
L<< b , h
(b) Loading – axial compressive axial forces
(c) Deflection
At low loads - same as a rod, axial (x) stresses, axial deformation only
+
" Fx = 0 ! P + F(x) = 0 # F ( x) = $P
" F2 = 0 % + S(x) = 0
"MA =0 + M (x) $ F(x)w(x) = 0
M (x) + Pw(x) = 0
d 2w
EI + Pw = 0
2
dx
Tends to destabilize -
Stabilizing compressive load
(restoring) As w increases
w = e !x
Write as:
d 2w P
+ w=0
dx 2 EI
P
!2 e !x + e !x = 0
EI
"P P
!2 = # !± i i = "1
EI EI
P P
W = A sin x + B cos x + Cx + D
EI EI
For the simply supported case:
x=0 w=0
d 2w d 2w
M = EI =0! =0
dx 2 dx 2
x=L w=0
d 2w d 2w
M = EI =0! =0
dx 2 dx 2
d 2w P P P P
=! A sin x! B cos x
dx 2 EI EI EI EI
So using B.C.’s
w( x = 0) = 0 ! B + C = 0 B=0
C=0
d 2w
( x = 0) = 0 ! B = 0
dx 2
P
w( x = L ) = 0 ! + A sin L + DL = 0 A, B, C, D = 0
EI
d 2w P
( x = L ) = 0 ! + A sin L =0
dx 2 EI
! P $
So: A sin# L& = 0
" EI %
Satisfied if A=0, but that is not very interesting!
P P
So: sin L =0! L = n"
EI EI
integer
Thus buckling occurs if:
2 2
P = n ! 2 EI
L
deformed shape
n!x
w = A sin Note: A is not
L defined
n=1
n=3
! 2 EI
Pcrit =
L2
Other boundary conditions provide other values for A, B, C, & D, solution always of the
form:
P P
w = A sin x + B cos x + Cx + D
EI EI
and buckling load always has the form:
c!2 EI
Pcr =
L2
Simply Supported clamped-free clamped-clamped
c=1 c=1/4 c=4