511 06 PDF

Section 6: The Flexibility Method - Beams
Washkewicz College of Engineering
Preliminaries: Beam Deflections – Virtual Work

There are several methods available to calculate deformations (displacements and rotations)
in beams. They include:
• Formulating moment equations and then integrating to find rotations and
displacements
• Moment area theorems for either rotations and/or displacements
• Virtual work methods
Since structural analysis based on finite element methods is usually based on a potential
energy method, we will tend to use virtual work methods to compute beam deflections.
The theory that supports calculating deflections using virtual work will be reviewed and
several examples are presented.
1
Consider the following arbitrarily loaded beam
Identify
M  M(x)
 M oment at any section in the beam due to external loads
m  m(x)
 M oment at any section in the beam due to a unit action
my
~ 
I
 Stress acting on dA due to a unit action
2
The force acting on the differential area dA due to a unit action is

~
f  ~ dA
m y
   dA
 I 
The stress due to external loads is
M y
 
I
The displacement of a differential segment dA by dx along the length of the beam is
   dx
 
   dx
E
M y
   dx
 E I 
3
The work done by the force acting on the differential area dA due to a unit action as the
differential segment of the beam (dA by dx) displaces along the length of the beam by an
amount  is ~
dW  f 
m y M y
   dA   dx
 I   E I 
 M m y2 
  2
 dA dx
 E I 
The work done within a differential segment (now A by dx) due to a unit action applied to
the beam is the integration of the expression above with respect to dA, i.e.,

  M m y2 
cT


 dW    

2
 dA  dx

A  cB  E I  
 cT
  Mm  2 

Wdiffernetial segment    2   y dA  dx
  EI c B
 

  Mm    Mm 
   2  I  dx    dx
  EI    EI  4
The internal work done along the entire length of the beam due to a unit action applied to
the beam is the integration of the last expression with respect to x, i.e.,
 M x  mx  
L
WInternal  0  EI  dx
The external work done along the entire length of the beam due to a unit action applied to
the beam is
WExternal  1D
With
WExternal  WInternal
 M  x  m x  
L
1 D   0  EI  dx
 M  x  m x  
L
D   0  EI  dx
or the deformation (D) of the a beam at the point of application of a unit action (force or
moment) is given by the integral on the right. 5
Example 6.1
6
Example 6.2
Flexibility Coefficients by virtual work
7
Perspectives on the Flexibility Method
In 1864 James Clerk Maxwell published the first consistent treatment of the flexibility
method for indeterminate structures. His method was based on considering deflections, but
the presentation was rather brief and attracted little attention. Ten years later Otto Mohr
independently extended Maxwell’s theory to the present day treatment. The flexibility
method will sometimes be referred to in the literature as Maxwell-Mohr method.
With the flexibility method equations of compatibility involving displacements at each of

the redundant forces in the structure are introduced to provide the additional equations
needed for solution. This method is somewhat useful in analyzing beams, frames and
trusses that are statically indeterminate to the first or second degree. For structures with a
high degree of static indeterminacy such as multi-story buildings and large complex trusses
stiffness methods are more appropriate. Nevertheless flexibility methods provide an
understanding of the behavior of statically indeterminate structures.
8
The fundamental concepts that underpin the flexibility method will be illustrated by the
study of a two span beam. The procedure is as follows
1. Pick a sufficient number of redundants corresponding to the degree of

indeterminacy
2. Remove the redundants
3. Determine displacements at the redundants on released structure due to external or

imposed actions
4. Determine displacements due to unit loads at the redundants on the released

structure
5. Employ equation of compatibility, e.g., if a pin reaction is removed as a redundant

the compatibility equation could be the summation of vertical displacements in the
released structure must add to zero.
9
Example 6.3
The beam to the left is statically
indeterminate to the first degree.
The reaction at the middle support
RB is chosen as the redundant.
The released beam is also shown.
Under the external loads the
released beam deflects an amount
DB.
A second beam is considered
where the released redundant is
treated as an external load and the
corresponding deflection at the
redundant is set equal to DB.
5
RB   wL
8
10
A more general approach consists in finding the displacement at B caused by a unit load in
the direction of RB. Then this displacement can be multiplied by RB to determine the total
displacement
Also in a more general approach a consistent sign convention for actions and displacements
must be adopted. The displacements in the released structure at B are positive when they are
in the direction of the action released, i.e., upwards is positive here.
The displacement at B caused by the unit action is

L3
B 
48EI
The displacement at B caused by RB is δB RB. The displacement caused by the uniform load
w acting on the released structure is
5 w L4
DB  
384 EI
Thus by the compatibility equation
DB 5
DB   B RB  0 RB     w L
B 8 11
Example 6.4
If a structure is statically indeterminate to more

than one degree, the approach used in the
preceeding example must be further organized
and more generalized notation is introduced.
Consider the beam to the left. The beam is

statically indeterminate to the second degree. A
statically determinate structure can be obtained
by releasing two redundant reactions. Four
possible released structures are shown.
12
The redundants chosen are at B and C. The

redundant reactions are designated Q1 and Q2.
The released structure is shown at the left

with all external and internal redundants
shown.
DQL1 is the displacement corresponding to Q1

and caused by only external actions on the
released structure
DQL2 is the displacement corresponding to Q2

caused by only external actions on the
released structure.
Both displacements are shown in their

assumed positive direction.
13
We can now write the compatibility equations for this structure. The displacements
corresponding to Q1 and Q2 will be zero. These are labeled DQ1 and DQ2 respectively
DQ1  DQL1  F11Q1  F12Q2  0
DQ 2  DQL 2  F21Q1  F22Q2  0
In some cases DQ1 and DQ2 would be nonzero then we would write
DQ1  DQL1  F11Q1  F12Q2
DQ 2  DQL 2  F21Q1  F22Q2
14
The equations from the previous page can be written in matrix format as
D 
Q  D  QL  F  Q
where:
{DQ } - vector of actual displacements corresponding to the redundant

{DQL } - vector of displacements in the released structure corresponding to the
redundant action [Q] and due to the loads
[F] - flexibility matrix for the released structure corresponding to the redundant
actions [Q]
{Q} - vector of redundants
 DQ1   DQL1  Q1 

D 
Q    D QL    Q   
 DQ 2   DQL 2  Q2 
F F12 
  11
F22 
F
 F21
15
The vector {Q} of redundants can be found by solving for them from the matrix equation
on the previous overhead.
F  Q  D 
Q  D 
QL
Q  F 1  DQ    D 
QL
To see how this works consider the previous beam with a constant flexural rigidity EI. If
we identify actions on the beam as
P1   2P M  PL P2   P P3  P
Since there are no displacements imposed on the structure corresponding to Q1 and Q2,
then
0
DQ   
0
16
The vector [DQL] represents the displacements in the released structure corresponding to
the redundant loads. These displacements are
13PL3 97 PL3
DQL1  DQL 2 
24 EI 48EI
The positive signs indicate that both displacements are upward. In a matrix format
26
D QL 
PL3
 
48EI 97 
17
The flexibility matrix [F ] is obtained by subjecting the beam to unit load corresponding
to Q1 and computing the following displacements
L3 5L3
F11  F21 
3EI 6 EI
Similarly subjecting the beam to unit load corresponding to Q2 and computing the
following displacements
5L3 8L3
F12  F22 
6 EI 3EI
18
The flexibility matrix is

L3  2 5 
F  
6 EI  5 16
The inverse of the flexibility matrix is

 16  5 
F 1 
6 EI
 5 2 
7 L3  
As a final step the redundants [Q] can be found as follows
Q 
Q   1  F 1   DQ   D  
QL
Q2 
  16  5   0
 6 EI  PL3  26
  3        
 7L   5 2   0  48 EI  97 
 P  69 
   
 
56  64  19
The redundants have been obtained. The other unknown reactions can be found from
the released structure. Displacements can be computed from the known reactions on
the released structure and imposing the compatibility equations.
Discuss the following sign conventions and how they relate to one another:
1. Shear and bending moment diagrams
2. Global coordinate axes
3. Sign conventions for actions
- Translations are positive if the follow the direction of the applied force
- Rotations are positive if they follow the direction of the applied moment
20
Example 6.5
A three span beam shown at the left is
acted upon by a uniform load w and
concentrated loads P as shown. The
beam has a constant flexural rigidity EI.
Treat the supports at B and C as
redundants and compute these
redundants.
In this problem the bending moments at B
and C are chosen as redundants to
indicate how unit rotations are applied to
released structures.
Each redundant consists of two moments,
one acting in each adjoining span.
21
The displacements corresponding to the two redundants consist of two rotations – one for
each adjoining span. The displacement DQL1 and DQL2 corresponding to Q1 and Q2.
These displacements will be caused by the loads acting on the released structure.
The displacement DQL1 is composed of two parts, the rotation of end B of member AB
and the rotation of end B of member BC
wL3 PL2
DQL1  
24 EI 16 EI
Similarly,
PL2 PL2 PL2
DQL 2   
16 EI 16 EI 8EI
such that
L2 2wL  3P 
DQL   
48EI  6P 
22
The flexibility coefficients are determined next. The flexibility coefficient F11 is the sum
of two rotations at joint B. One in span AB and the other in span BC (not shown below)
L L 2L
F11   
3EI 3EI 3EI
Similarly the coefficient F21 is equal to the sum of rotations at joint C. However, the
rotation in span CD is zero from a unit rotation at joint B. Thus
L
F21 
6 EI 23
Similarly
L L 2L
F22   
3EI 3EI 3EI
L
F12 
6 EI
The flexibility matrix is
L 4 1 
F  1 4
6EI  
The inverse of the flexibility matrix is
1 2 EI  4  1
F   1 4 
5L  
24
As a final step the redundants [Q] can be found as follows
Q 
Q   1  F 1  DQ   D  
QL
Q2 
2 EI 4  1 0 L2 2wL  3P 
  1 4  0  48EI  
5L      6 P 
L  8wL  6 P  
   
120  2wL  21P 
and
wL2 PL wL2 7 PL
Q1    Q2  
15 20 60 40
25
Example 6.6
26
Joint Displacements, Member End Actions And Reactions

Previously we focused on finding redundants using flexibility (force) methods. Typically
redundants (Q1, Q2, … , Qn) specified by the structural engineer are unknown reactions.
Redundants are determined by imposing displacement continuity at the point in the structure
where redundants are applied, i.e., we imposed
D 
Q  D 
QL  F  Q
If the redundants specified are unknown reactions then after these redundants are found other
actions in the released structure could be found using equations of equilibrium.
When all actions in a structure have been determined it is possible to compute displacements
by isolating the individual subcomponents of a structure. Displacements in these
subcomponents can be calculated using concepts learned in Strength of Materials. These
concepts allow us to determine displacements anywhere in the structure but usually the
unknown displacements at the joints are of primary interest if they are non-zero.
.
27
Instead of following the procedure just outlined we will now introduce a systematic
procedure for calculating non-zero joint displacements, reaction, and member end actions
directly using flexibility methods.
Consider the two span beam below where the redundants Q1 and Q2 have been computed
previously in Example 6.4. The non-zero joint displacements DJ1 and DJ2, both rotations, as
well as reactions AR1 and AR2. can be computed. We will focus on the joint displacements
DJ1 and DJ2 first. Keep in mind that when using flexibility methods translations are
associated with forces, and rotations are associated with moments.
Reactions other than redundants will be denoted {AR} and these quantities can be
determined as well. The objective here is the extension of the flexibility (force) method so
that it is more generally applied. 28
The principle of superposition is used to obtain the joint displacement vector {DJ}, which is
a vector of displacements that occur in the actual structure. For the structure depicted on
the previous page the rotations in the actual structure at joints B ( = DJ1) and C ( = DJ2) are
required. When the redundants Q1 and Q2 were found superposition was imposed on the
released structure requiring the displacement associated with the unknown redundants to be
equal to zero. In finding joint displacements in the actual structure superposition is used
again and displacements in the released structure are equated to the displacement in the
actual structure. Focusing on joint B, superposition requires
DJ 1  DJL1  DJQ11Q1  DJQ12Q2

Here
DJ1 = non-zero displacement (a rotation) at joint B in the actual structure, at

the joint associated with Q1
DJL1 = the displacement (a rotation) at joint B associated with DJ1 caused by
the external loads in the released structure.
DJQ11 = the rotation at joint B associated with DJ1 caused by a unit force at
joint B corresponding to the redundant Q1 in the released structure
DJQ12 = the rotation at joint B associated with DJ1 caused by a unit force at joint
C corresponding to the redundant Q2 in the released structure 29
Thus displacements in the released structure must be further evaluated for information
beyond that required to find the redundants Q1 and Q2 . In the released structure the
displacements associated with the applied loads are designated {DJL} and are depicted
below. The displacements associated with the redundants are designated [DJQ ] and are
similarly depicted.
In the figure to the right unit

loads are shown applied at the
redundants. These unit loads
released structure
were used earlier to find
flexibility coefficients [Fij ].
These coefficients were then
used to determine Q1 and Q2 .
Now the unit loads are used to
find the components of [DJQ ].
30
A similar expression can be derived for the rotation at C ( = DJ2), i.e.,
DJ 2  DJL 2  DJQ 21 Q1  DJQ 22 Q2
Here
DJ2 = non-zero displacement (a rotation) at joint C in the actual structure, at

the joint associated with Q2
DJL2 = the displacement (a rotation) at joint C associated with DJ2 caused by the
external loads in the released structure.
DJQ21 = the rotation at joint C associated with DJ2 caused by a unit force at joint B
corresponding to the redundant Q1 in the released structure
DJQ22 = the rotation at joint C associated with DJ2 caused by a unit force at joint C
corresponding to the redundant Q2 in the released structure
31
The expressions DJ1 and DJ2 can be expressed in a matrix format as follows
DJ   DJL   D Q

JQ
where
D 
DJ    J1 
 DJ 2 
D 
DJL    JL1 
DJL 2 
 
D 
JQ  
DJQ11
DJQ 21
DJQ12
DJQ 22 
 
and
Q 
Q   1
Q2 
which were determined previously 32
In a similar manner we can find reactions via superposition
AR1  ARL1  ARQ11Q1  ARQ12Q2
AR 2  ARL 2  ARQ 21Q1  ARQ 22Q2

For the first expression
AR1 = the reaction in the actual beam at A

AR2 = the reaction in the actual beam at A
ARL1 = the reaction in the released structure due to the external loads
ARL2 = the reaction in the released structure due to the external loads
ARQ11 = the reaction at A in the released structure due to the unit action
corresponding to the redundant Q1
33
The expressions on the previous slide can be expressed in a matrix format as
AR   ARL   A Q

RQ
where
A   
A 
ARL    RL1  A   
ARQ11 ARQ12

AR 
RQ
  R1   ARL 2   ARQ 21 ARQ 22 
 AR 2 
Q 
Q   1
Q2 
34
In a similar manner we can find member end actions via superposition

AM 1  AML1  AMQ11Q1  AMQ12Q2

For the first expression
AM1 = is the shear force at B on member AB

AML1 = is the shear force at B on member AB caused by the external loads on
the released structure
AMQ11 = is the shear force at B on member AB caused by a unit load
AMQ12 = is the shear force at B on member AB caused by a unit load
The other expressions follow in a similar manner. 35
The expressions on the previous slide can be expressed in a matrix format as follows
AM   AML   A Q

MQ
where
 AM 1   AML1   AMQ11 AMQ12 

A  A   
AM  
 M2
  AML  
 ML 2 
  A  MQ 
 AM Q 21 AMQ 22 
 
A
 M3 A
 ML3   AMQ31 AMQ32 

 AM 4 
 
 AML 4 
 A A 
 MQ 41 MQ 42 
Q 
Q   1
Q2 
The sign convention for member end actions is as follows:

+ when up for translations and forces
+ when counterclockwise for rotation and couples 36
Example 6.7
Consider the two span beam to the left
where it is assumed that the objective is
to calculate the various joint
displacements DJ , member end actions
AM , and end reactions AR. The beam has
a constant flexural rigidity EI and is acted
upon by the following loads
P1  2P
M  PL
P2  P
P3  P
37
Consider the released

structure and the attending
moment area diagrams.
The (M/EI) diagram was
drawn by parts. Each
action and its attending
diagram is presented one at
a time in the figure starting
with actions on the far
right.
38
From first moment area theorem

1 PL 1 PL
DJL1   2  1 L  1.5  0.5  L
2 EI 2 EI
PL 1 PL L
 L 
EI 2 EI 2
5 PL2

4 EI
1  2 PL  1  3PL  3L
DJL 2    2 L   
2  EI  2  2 EI  2
 PL  1 PL L
   L 
 EI  2 EI 2
13PL2

8 EI
PL2 10
DJL   13
8EI   39
Consider the released beam with a unit load at point B
1 L
DJQ11 
1 L
L DJQ 21  L
2 EI 2 EI
L2 L2
 
2 EI 2 EI
40
Consider the released beam with a unit load at point C
2L
DJQ12 
1 L
2  1 L DJQ 22 
1 2L
2L
2 EI 2 EI
3L2 2 L2
 
2 EI EI
leading to
L2 1 3

 DJQ   1
 2 EI  4
 41
Previously in Example 6.4

P  69 
Q   
56  64 
with
DJ   DJL   D Q

JQ
then the displacements DJ1 and DJ2 are
PL2 10  L2 1 3  P  69 
 DJ   13  1  
8 EI   2 EI  4 56  64 
PL2 17 
  
112 EI 5
42
Using the following free body diagram of the released structure
Then from the equations of equilibrium
M A  0
L 3L
 ARL 2  2P  PL  P  P2L
2 2
PL
ARL 2  
2
F Y  0
 ARL1  2P  P  P
ARL1  2P
43
Using a free body diagram from segment AB of the entire beam, i.e.,
then once again from the equations of equilibrium
F Y  0
 AML1  2P  2P
AML1  0
M B  0
L PL
 AML 2  2P   2 PL
2 2
3PL
AML 2  44
2
Using a free body diagram from segment BC of the entire beam, i.e.,
then once again from the equations of equilibrium
F Y  0
 AML3  P  P
AML3  0
M  0
B
PL
 A   PL
ML 4 2
PL
A   45
ML 4 2
Thus the vectors AML and ARL are as follows:
 0 
 
 
 3 PL   2P 
 2  
 

   
AML    ARL 
 0   PL 
  
 2  
 
 PL 
 
2 
Reactions in the released structure.
 
Member end actions in the released structure.
46
Finally with
AR   ARL   A Q

RQ
then knowing [ARL], [ARQ] and [Q] leads to
 2P 
  P  1 1   69 
 AR    PL    L   64 
  56  2 L   
 2 

P 107 
  
56 31L 
47
In a similar fashion, applying a unit load associated with Q1 and Q2 in the previous
cantilever beam, we obtain the following matrices
1 1 
0 L 
A 
MQ 




1 
0

0  L

 1 1 

 ARQ    L

 2 L 

48
Similarly, with
AM   AML   A Q

MQ
and knowing [AML], [AMQ] and [Q] leads to
 0 
 3PL  1 1 
    69
 2  P 0 L   
 AM        
 0  56 0 1   64 
 PL  
0 L
 
 2 
 5 
 
P  20 L 
  
56  64 

36 L 
49
Summary Of Flexibility Method
The analysis of a structure by the flexibility method may be described by the following steps:
1. Problem statement
2. Selection of released structure
3. Analysis of released structure under loads
4. Analysis of released structure for other causes
5. Analysis of released structure for unit values of redundant
6. Determination of redundants through the superposition equations, i.e.,
D 
Q  D 
QS  F  Q
D 
QS  D  QL  D 
QT  D 
QP  D 
QR
Q  F 1  DQ   D  

QS 50
7. Determine the other displacements and actions. The following are the four flexibility
matrix equations for calculating redundants member end actions, reactions and joint
displacements
DJ   DJS   D Q
JQ
AM   AML   A Q

MQ
AR   ARL   A Q

RQ
where for the released structure
DJS   DJL   DJT   DJP   DJR 

All matrices used in the flexibility method are summarized in the following tables
51
MATRIX ORDER DEFINITION
Q qx1 Unknown redundant actions (q = Number of redundant)
Displacements in the actual structure Corresponding to the

DQ qx1 redundant
Displacements in the released structure corresponding to the

DQL qx1 redundants and due to loads

DJQ qxq redundants and due to unit values of the redundants

DQT , DQP , DQR qx1 redundants and due to temperature, prestrain, and restraint
displacements (other than those in DQ)
DQS qx1 DQS  DQL  DQT  DQP  DQR

52
jx1 Joint displacement in the actual structure (j = number of joint

DJ
displacement)
jx1 Joint displacements in the released structure due to loads
DJL
jx1 Joint displacements in the released structure due to unit values

DQL of the redundants
Joint displacements in the released structure due to

DJT , DJP , DJR jx1 temperature, prestrain, and restraint displacements (other than
those in DQ)
DJS jx1 DJS  DJL  DJT  DJP  DJR
F qxq Matrix of flexibility coefficients

53

mx1 Member end actions in the actual structure
AM (m = Number of end-actions)
mx1 Member end actions in the released structure due to loads
AML
mxq Member end actions in the released structure due to unit
AMQ values of the redundants
AR rx1 Reactions in the actual structure (r = number of reactions)
ARL rx1 Reactions in the released structure due to loads
Reactions in the released structure due to unit values of the

ARQ rxq redundants
54

511 06 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

511 06 PDF

Uploaded by

Copyright:

Available Formats

Section 6: The Flexibility Method - Beams

Washkewicz College of Engineering

Preliminaries: Beam Deflections – Virtual Work

Consider the following arbitrarily loaded beam

The force acting on the differential area dA due to a unit action is

Flexibility Coefficients by virtual work

Perspectives on the Flexibility Method

With the flexibility method equations of compatibility involving displacements at each of

1. Pick a sufficient number of redundants corresponding to the degree of

2. Remove the redundants

3. Determine displacements at the redundants on released structure due to external or

4. Determine displacements due to unit loads at the redundants on the released

5. Employ equation of compatibility, e.g., if a pin reaction is removed as a redundant

The displacement at B caused by the unit action is

If a structure is statically indeterminate to more

Consider the beam to the left. The beam is

The redundants chosen are at B and C. The

The released structure is shown at the left

DQL1 is the displacement corresponding to Q1

DQL2 is the displacement corresponding to Q2

Both displacements are shown in their

DQ1  DQL1  F11Q1  F12Q2  0

DQ 2  DQL 2  F21Q1  F22Q2  0

DQ1  DQL1  F11Q1  F12Q2

DQ 2  DQL 2  F21Q1  F22Q2

{DQ } - vector of actual displacements corresponding to the redundant

 DQ1   DQL1  Q1 

The flexibility matrix is

The inverse of the flexibility matrix is

As a final step the redundants [Q] can be found as follows

The inverse of the flexibility matrix is

As a final step the redundants [Q] can be found as follows

Joint Displacements, Member End Actions And Reactions

DJ 1  DJL1  DJQ11Q1  DJQ12Q2

DJ1 = non-zero displacement (a rotation) at joint B in the actual structure, at

In the figure to the right unit

A similar expression can be derived for the rotation at C ( = DJ2), i.e.,

DJ 2  DJL 2  DJQ 21 Q1  DJQ 22 Q2

DJ2 = non-zero displacement (a rotation) at joint C in the actual structure, at

DJ   DJL   D Q

In a similar manner we can find reactions via superposition

AR1  ARL1  ARQ11Q1  ARQ12Q2

AR 2  ARL 2  ARQ 21Q1  ARQ 22Q2

AR1 = the reaction in the actual beam at A

The expressions on the previous slide can be expressed in a matrix format as

AR   ARL   A Q

In a similar manner we can find member end actions via superposition

AM 3  AML3  AMQ31Q1  AMQ32Q2

AM 4  AML4  AMQ41Q1  AMQ42Q2

AM1 = is the shear force at B on member AB

AM   AML   A Q

 AM 1   AML1   AMQ11 AMQ12 

The sign convention for member end actions is as follows:

Consider the released

From first moment area theorem

Consider the released beam with a unit load at point B

Consider the released beam with a unit load at point C

Previously in Example 6.4

DJ   DJL   D Q

then the displacements DJ1 and DJ2 are

Using the following free body diagram of the released structure

Then from the equations of equilibrium

then once again from the equations of equilibrium