B Flexibility Method

In this text, we have focused our attention on the matrix stiffness method of struc-
tural analysis, which is the most commonly used method in professional practice
today, and which forms the basis for most of the currently available commercial
software for structural analysis. However, as stated in Section 1.3, another type
of matrix method, called the flexibility method, can also be used for structural
analysis. While the stiffness method can be applied to both statically determinate
and indeterminate structures, the flexibility method is applicable only to indeter-
minate structures. The flexibility method is essentially a generalization in matrix
form of the classical method of consistent deformations, and is generally consid-
ered convenient for analyzing small structures with a few redundants.
In this appendix, we present the basic concept of the flexibility method,
and illustrate its application to plane trusses. A more detailed treatment of this
method can be found in [3] and [52].
Essentially, the flexibility method of analysis involves removing enough
restraints from the indeterminate structure to render it statically determinate.
This determinate structure, which must be statically stable, is called the pri-
mary structure; and the reactions or internal forces associated with the excess
restraints removed from the given indeterminate structure to convert it into the
determinate (primary) structure, are termed redundants. The redundants are
then treated as unknown loads on the primary structure, and their values are
determined by solving the compatibility equations based on the condition that
the deformations of the primary structure due to the combined effect of the re-
dundants and the given external loading must be the same as the deformations
of the original indeterminate structure.
Consider, for example, a plane truss supported by five reaction compo-
nents, as shown in Fig. B.1(a) on the next page. The truss is internally deter-
minate, but externally indeterminate with two degrees of indeterminacy. This
indicates that the truss has two more, or redundant, reactions than necessary
for static stability. Thus, if we can determine two of the five reactions by using
compatibility equations based on the geometry of the deformation of the truss,
then the remaining three reactions and the member forces can be obtained from
equilibrium considerations.
To analyze the truss by the flexibility method, we must select two of the
unknown reactions and member forces to be the redundants. Suppose that we
select the horizontal and vertical reactions, R1 and R2, at the hinged support
at joint 5 to be the redundants. The hinged support at joint 5 is then removed
from the given indeterminate truss to obtain the statically determinate and stable
613

Copyright 2022 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 2 4

3m 4 5
6 7

1 2 5
¯
R 1
1 608 3

300 kN R¯2

2 panels at 4 m = 8 m
EA = constant
(a) Indeterminate Truss

5
259.81

346.41 2 4

346.41
259.81

259.81

2
3.0
43 0

496.41 5
150 0

1 150 kN 3

259.81 kN
(b) Primary Truss Subjected to External Loading—QaO Vector

1
2 4

0 0 0 0

1 1
1 1 kN 3 R¯1
1 3 5

Fig. B.1 ¯ —First Column of b Matrix

(c) Primary Truss Subjected to Unit Value of Redundant R1
614

Copyright 2022 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 Appendix B   Flexibility Method  615

1
1

2.6667 2 4

2.6667

7
66 1.6
1 1.6 0 66
7

2.6667 5
1.3333 1.3333
3 R¯2
1 3

1 kN

¯ —Second Column of b Matrix

(d) Primary Truss Subjected to Unit Value of Redundant R2

123.324

17.548 2 4
17.548
123.324

4 22
259.81
5.5 7.4
20 75

5
239.43 75 75 106.98

1 608 3

300 kN
136.485

(e) Support Reactions and Member Forces for Indeterminate Truss

Fig. B.1 (continued)

primary truss, as shown in Fig. B.1(b). The two redundants R1 and R2 are now
treated as unknown loads on the primary truss, and their magnitudes can be
determined from the compatibility conditions that the horizontal and vertical
deflections at joint 5 of the primary truss due to the combined effect of the
known 100 k load and the unknown redundants R1 and R2 must be equal to 0.
This is because the deflections in the horizontal and vertical directions of the
given indeterminate truss at the hinged support at joint 5 are 0.
The compatibility equations can be conveniently established by superim-
posing the deflections due to the external loading and the redundants, R1 and R2,

Copyright 2022 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
616  Appendix B   Flexibility Method

acting individually on the primary truss, as shown in Figs. B.1(b), (c), and (d),
respectively. Thus,
ΔO1 1 f11R1 1 f12R2 5 0 (B.1a)
ΔO2 1 f21R1 1 f22R2 5 0 (B.1b)
in which ΔOi (i 5 1, 2) represents the deflection at joint 5 of the primary truss in
the direction of the redundant Ri , due to the external loading; and the flexibility
coefficient fij (i 5 1, 2 and j 5 1, 2) denotes the deflection of the primary truss
at the location and in the direction of a redundant Ri due to a unit value of a
redundant Rj . Equations (B.1) can be expressed in matrix form as

3ΔΔ 41 3 f 4 3 RR 4 5 3004
f11 f12
O1 1
(B.2)
O2 21
f22 2

From the foregoing discussion for the example truss with two degrees of
indeterminacy, we realize that the compatibility equations for a general indetermi-
nate structure with ni degrees of indeterminacy can be symbolically expressed as

DO 1 fR 5 0  (B.3)

in which the ni 3 1 vectors R and DO denote, respectively, the unknown re-

dundants, and the deflections of the primary structure at the locations and in
the directions of the redundants due to external loads; and the ni 3 ni matrix
f is called the structure flexibility matrix. The reader may recall from a pre-
vious course in mechanics of materials or structural analysis [18], that Max-
well’s law of reciprocal deflections states that for a linearly elastic structure,
the deflection at a point i due to a unit load applied at a point j is equal to the
deflection at j due to a unit load at i. As the flexibility coefficient fij denotes
the deflection of the primary structure at the location of the redundant Ri due
to a unit value of the redundant Rj , and the flexibility coefficient fji denotes the
deflection corresponding to Rj due to a unit value of Ri , according to Maxwell’s
law fij must be equal to fji (i.e., fij 5 fji). We can thus deduce that for linearly
elastic structures, the flexibility matrices are symmetric.
From Eqs. (B.1) through (B.3), we can see that the elements of the vector
DO and the flexibility matrix f represent deflections of the primary (statically
determinate) structure. Once these deflections have been evaluated, the com-
patibility equations (Eqs. (B.3)) can be solved for the unknown redundants.
With the redundants known, the other response characteristics of the structure
can be evaluated, either by equilibrium or superposition.
The deflections (and the flexibility coefficients) of a primary structure can
be conveniently expressed in terms of the forces and properties of its members,
using the virtual work method. Recall from a previous course in mechanics
of materials or structural analysis [18], that the expression of the virtual work
method for truss deflections is given by
NM Qar Qay L
Δ5 o
i51 EA
 (B.4)

Copyright 2022 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 Appendix B   Flexibility Method  617

in which NM denotes the number of members of the truss; Qar represents the
axial forces in truss members due to the real loading that causes the deflection
Δ, and Qay represents the axial forces in the truss members due to a virtual
unit load acting at the location and in the direction of the desired deflection Δ.
Equation (B.4) can be expressed in matrix form as
Δ 5 QTayfMQar (B.5)
in which Qay and Qar denote the NM 3 1 vectors containing member axial forces
due to virtual (unit) and real (actual) loads, respectively; and fM is a NM 3 NM
diagonal matrix containing the member flexibilities (L /EA) on its main diagonal
(i.e., fMij 5 Li /Ei Ai for i 5 j, and fMij 5 0 for i Þ j). The diagonal matrix fM is some-
times called the unassembled flexibility matrix. In order to develop the expressions
for DO and f in terms of the member forces and properties, let us define a NM 3 1
vector QaO which contains the axial forces in the members of the primary truss due
to the given external loading, and a NM 3 ni matrix b, the jth column of which con-
tains member axial forces due to a unit value of the jth redundant (i.e., Rj 5 1). The
matrix b is commonly referred to as an equilibrium matrix. In both QaO and b
the member axial forces are stored in sequential order of member numbers; that
is, the axial forces in the ith member are stored in the ith rows of QaO and b, and so
on. The member forces QaO for the example truss are shown in Fig. B.1(b). Note
that since the primary truss is statically determinate, the forces in its members due
to the given external loading can be conveniently evaluated by applying the method

FG
of joints. By using the member forces shown in Fig. B.1(b) and the member num-
bers given in Fig. B.1(a), we form the QaO vector for the truss as

2150 1
0 2
346.41 3
QaO 5 259.81 4 kN (B.6)
259.81 5
2433.02 6
0 7
in which the tensile member axial forces are considered to be positive. The
first column of b is obtained by subjecting the primary truss to a unit value of
the redundant R1, as shown in Fig. B.1(c), and by computing the correspond-
ing member forces by applying the method of joints. The second column of b
is generated similarly by subjecting the primary truss to a unit value of the

F G
redundant R2, and by computing the corresponding member axial forces (see
Fig. B.1(d)). The equilibrium matrix b thus obtained is
1 1.3333
1 1.3333
0 22.6667
b5 0 21 kN/kN (B.7)
0 0
0 1.6667
0 21.6667
Copyright 2022 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
618  Appendix B   Flexibility Method

The first element, ΔO1 , of the DO vector represents deflection of the pri-
mary truss at the location and in the direction of R1 due to the given exter-
nal loading. Therefore, for the purpose of calculating ΔO1 via the virtual work
method, the real system consists of the given external loading as shown in
Fig. B.1(b), and the virtual system consists of a unit load applied at the location
and in the direction of the redundant R1, which is the same as the system shown
in Fig. B.1(c) (without the multiplier R1). Thus, the virtual work expression for
ΔO1 can be obtained by substituting QaO for Qar and the first column of b for
Qay in Eq. (B.5); that is,

ΔO1 5 bT1 fMQaO (B.8)

in which b1 denotes the first column of b. The expression for the second ele-
ment, ΔO2 , of DO , in terms of member axial forces, can be obtained in a similar
manner, and is given by

DO2 5 bT2 fMQaO (B.9)

with b2 denoting the second column of b. By combining Eqs. (B.8) and (B.9),
we obtain

DO 5 bT fMQaO  (B.10)

The expressions for the elements of the flexibility matrix f, in terms

of member forces and properties, can be obtained in a similar manner. For
example, the virtual and real systems for the evaluation of f12 are shown in
Figs. B.1(c) and (d), respectively, with the corresponding member forces stored
in the first and second columns of b. Therefore,

f12 5 bT1 fMb2

Thus, the entire flexibility matrix f can be expressed in terms of the member
forces and properties as

f 5 bT fMb  (B.11)

Finally, by substituting Eqs. (B.10) and (B.11) into Eq. (B.3), we obtain the
structure’s compatibility equations in terms of its member forces and proper-
ties, as

bT fMQaO 1 (bTfMb) R 5 0  (B.12)

To illustrate the application of the flexibility method to the analysis

of plane trusses, let us reconsider the truss of Fig. B.1. The QaO vector and
the b matrix for this truss were determined previously, and are given in

Copyright 2022 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 Appendix B   Flexibility Method  619

F G
Eqs. (B.6) and (B.7), respectively. The unassembled flexibility matrix for
the structure is
4 0 0 0 0 0 0
0 4 0 0 0 0 0
0 0 4 0 0 0 0
1 0 0 0 3 0 0 0
fM 5  (B.13)
EA 0 0 0 0 3 0 0
0 0 0 0 0 5 0
0 0 0 0 0 0 5

Substituting Eqs. (B.6), (B.7), and (B.13) into Eq. (B.10), and performing the
required matrix multiplications, we obtain

3 4
1 2600
Do 5 bT fMQaO 5  (B.14)
EA 28,883
Substitution of Eqs. (B.7) and (B.13) into Eq. (B.11) yields the flexibility
matrix:

3 4
1 8 10.667
f 5 bT fMb 5  (B.15)
EA 10.667 73.444
Next, we substitute Eqs. (B.14) and (B.15) into the compatibility equations
(Eqs. (B.12)), and solve the resulting system of simultaneous equations for the
unknown redundants. This yields

R5 32106.98
136.4854
kN (B.16)

With the redundants known, the member axial forces in the actual indeter-
minate structure, Qa, can be conveniently evaluated by applying the superposi-
tion relationship (see Figs. B.1(a) through (d)):

Qa 5 QaO 1 bR  (B.17)

F G
Substituting Eqs. (B.6), (B.7), and (B.16) into Eq. (B.17), we determine the
axial forces in the members of the indeterminate truss to be

275
75
217.548
Qa 5 123.324 kN
259.81
2205.54
2227.475

These member axial forces are shown in Fig. B.1(e).

Copyright 2022 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s).
Editorial review has deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.