CHAPTER ONE INTRODUCTION - To Restore The Dignity … · i To derive a set of differential...
Transcript of CHAPTER ONE INTRODUCTION - To Restore The Dignity … · i To derive a set of differential...
1
CHAPTER ONE INTRODUCTION
1.1 PREAMBLE
A thin-walled structure is one which is m1ade from thin plates joined along their
edges. The plate thickness however is small compared to other cross sectional
dimensions which are in turn often small compared with the overall length of the
member or structure. Quantitatively, thin walled structures are structures in which the
ratio of the thickness t, to the two other linear dimensions (length l, and width w)
ranges within the limits ( t/l or t/w = 1/50 to 1/10), (Rekach 1979). Thus a thin walled
structure has two dimensions of the structural element much larger than the third
one, i.e., the thickness. The elements of a thin walled structure are plates and shells,
the later being plates having a curvature in one or two dimensions. Numerous
structural configurations require partial or even complete enclosure that can easily be
accomplished by plates. Thin walled structures have gained special importance and
notably increased application in recent years. The wide use of these thin walled
structures is due to their great carrying capability and reliability and to the economic
advantage they have over solid (column and beam) structures, (Heins 1975). Thin
walled structures which are designed with the nature of work in view, are lighter than
solid structures. Consequently their mechanical theory is of great importance in
engineering practice.
When two or more plates are joined together to form an open or closed structure
strength and rigidity are increased. For example, tanks, boilers, etc, are cylindrical
shell structures with increased strength and rigidity. Conical shell structures are also
common features in construction, mechanical engineering and aeronautical design.
Thin-walled structures are used extensively in steel and concrete bridges, ships, air
crafts, mining head frames and gantry frames. These are seen in the form of box
girders, plate girders, box columns and purlins (z and channel sections).
Generally, a thin walled structure can be of open cross section e.g. channel and
prismatic sections, or a closed cross section such as circular and rectangular
sections. Thin walled structures of closed sections are generally referred to as ‘box
structures’. Thus a girder structure cellular in section can be called box girder
structure. These find their uses in different fields of modern engineering such as box
culverts, and box girders in highway and bridge engineering.
2
Thin walled structures of open cross section are subjected only to bending
stresses, no axial load or torsion, Heins (1975). Because of their thin wall
thicknesses, the shearing resistances are constant across the thickness of the plate.
On the other hand thin walled box structures may also be subjected to bending,
torsional and distortional stresses. Distortion alters the geometry of the cross section
and generates some additional stresses thereby reducing the bearing capability of
the box structural component.
Basically, a curved structural element has two interacting forces: bending and
torsion. Knowledge of this interaction leads to a successful design of the element.
The study of curved elements offers only one instance where torsional and bending
forces simultaneously occur. In other instances, e.g., straight girders (in either
bridges or buildings) may develop such forces and thus require appropriate analysis
and design. This situation arises where the load is eccentric with respect to the girder
axis or shear centre. The engineer is then forced to determine the shear centre,
resolve the forces into appropriate bending and torsional forces, Fig.1.1, and then
determine the stresses and deformations.
The torsional response of structural elements can be classified into two
categories: pure torsion and warping torsion or bi-moment. Most structural engineers
are familiar with the concept of pure torsion as this type of torsion is studied in
strength of materials courses. The second type of torsion, warping torsion and
distortion are probably new terms and phenomena which need to be fully
investigated in order to avoid their undesired effects particularly on box girder
bridges.
(a) Beam under eccentric load (b) Load transferred to shear
centre (c) Restoring moment at shear centre
= Sc
P
+ x
y
MZ
• x
y
• • sc c
x
z
P
y
ex xo
y
-
• • sc c c
Fig.1.1.1 Decomposition of eccentrically loaded beam
3
The general theories described by Heins (1975) relative to straight and curved
girders assume thin walled sections which may be open or closed (box section).
Development of general equations was based on prismatic girders (solid
beams).Their solution then permits a proper engineering examination of box girder
elements when subjected to vertical torsional loads.
One major topic of interest presently challenging the engineer is box girder
bridges, straight and curved. Initially, design of box girder bridges was related to the
design of plate girder bridges. However, such design knowledge does not contain
important primary conditions of cross sectional deformations. Initially, box girders
could be examined as open sections to allow the application of general plate
theories. Then the section was modified to consider continuity and warping. All of
these investigations assumed no distortion of the cross section. The application of
cross sectional deformation equations formulated by Vlasov (1958) and Dabrowski
(1968), has opened a new way to analyze the torsional and distortional effects of
loads on such girders. This is the main thrust of this work.
1.2 AIMS AND OBJECTIVES OF THE STUDY
A complicated state of forces develops in eccentrically loaded straight girders
when they carry generalized loads. The forces that develop include bending
moments, shear forces, pure torsion, warping torsional moments and bi-moments.
Plane sections under torsion will generally warp as shown in Fig.1.2.1, i.e., will not
remain plane and Bernoulli’s beam theory is therefore violated. If restrained, these
out of plane deformations create additional normal and shear stresses which when
integrated over the cross section yield the bi-moment and warping torsional moment;
the warping stresses. Warping stresses can be as large as or even larger than the
bending stresses, and they can not be ignored. It usually has its maximum value at a
free edge of the cross section and plays a significant role in initiating local buckling
because free edges are the weak parts of a cross section.
The aims and objectives of this study are as follows.
i To derive a set of differential equations governing the behaviour of thin- walled box
girder bridge structures of arbitrary cross section on the basis of Vlasov’s theory.
ii To apply the obtained differential equations in the analysis of single cell and multi-
cell thin walled structures with mono–symmetric and doubly symmetric cross
sections.
4
iii To apply generalized strain fields (modes) in computing the coefficients of the
differential equations in order to reduce the number of unknown displacement
functions.
iv To compare the results of analysis obtained using the derived equations with the
results obtained using other methods of analysis.
1.3 SIGNIFICANCE OF THE STUDY.
Structural components such as thin walled box girders are economical and
effective in resisting torsional and bending loads but due to thinness of the box walls,
loads applied to this type of structures give rise to warping and distortion of the cross
section. Consequently the geometry of the cross section is altered and some
additional stresses are generated thereby reducing the structural capability of the box
section.
Fortunately structural designers in steel are careful enough not to ignore the
effects of torsion on a structural member. Unfortunately however, the effects of
warping and distortion on a structural component are either poorly evaluated or
ignored in the analysis simply because of the rigorous mathematics involved in their
evaluation. There is therefore the need to develop a simple analytical model to
enable designers put into consideration the primary condition of cross sectional
deformations in the analysis and design of box girder structures. Thus a better and
more elaborate assessment of all the effects of loads on a thin walled box girder
bridge structure can be achieved by a consideration of the phenomena of warping
torsion and distortion in addition to bending and shear
1.4 SCOPE OF WORK
In this study, cellular box girders, with and without side cantilevers, are examined.
These include (a) single cell doubly symmetric section (b) multi cell doubly symmetric
Fig.1.2.1 Warping of closed thin-walled cross section
5
section (c) single cell mono symmetric section and double cell mono symmetric
section.
The girders are assumed to be subjected to purely torsional / distortional loading,
therefore the bending action of the girders are neglected as the structure will undergo
torsional stresses, distortional stresses and distortional displacements only. However,
the lateral displacements of the box girders due to torsional loads are considered.
Since the basic equations of equilibrium form the fundamental principle in the
development of stress/strain relations, dynamic loads which are characteristic of
bridge loads are taken care of by static loads stipulated by appropriate codes of
practice.
Although the requirements for road way alignment may give rise to curved bridge
structure, the structural behaviour of such curved thin walled box girder bridges is
outside the scope of this work. Elastic analysis (not elasto-plastic analysis) is the
mode adopted for the work.
Various methods of analysis of thin walled box girders are outlined in the
literature review but the method of Vlasov, was adopted in this work because it
captures all peculiarities of cross sectional deformations
Finally it is noted that the behaviour of thin-walled structures is usually sensitive to
the nature and magnitude of initial imperfections which arise inevitably during
fabrication. The effects of such initial imperfections fall outside the scope of this work.
1.5 RESEARCH METHODOLOGY
The existence of distortional and warping stresses in a thin walled structure
whose transverse cross section is arbitrary and which contains several closed
contours was first substantiated by V.Z.Vlasov (1949), who also developed a theory
for their analysis. Research has shown that using Vlasov’s original displacement
fields for the analysis of thin walled closed structure leads to a large number of
displacement functions to be determined. Verbernov (1958) has shown that by using
generalize strain fields on Vlasov’s formulation, the number of unknown displacement
functions can be reduced to seven irrespective of the number of independent
displacements possessed by the structure.
In this work, Vlasov’s theory as modified by Verbernov was adopted. The
potential energy of the box girder bridge structure under the action of a distortional
load was obtained using the principle of minimum potential energy. By minimizing the
6
energy functional using Euler-Lagrange technique the governing differential
equations of distortional equilibrium were obtained.
7
CHAPTER TWO
LITERATURE REVIEW
2.1 REVIEW OF PAST WORK
The curvilinear nature of box girder bridges along with their complex deformation
patterns and stress fields have led designers to adopt approximate and conservative
methods for their analysis and design, Khaled et al (2002). Recent literatures, Hsu et
al (1995); Fan and Helwig (2002); Tetsuya et al (1995); Sennah and Kennedy (2003),
on straight and curved box girder bridges deal with analytical formulations to better
understand the behaviour of these complex structural systems. Few authors, Okil and
El-tawil (2004); Sennah and Kennedy (2003); have undertaken experimental studies
to investigate the accuracy of existing methods. These are discussed in this review of
past work.
Before the advent of Vlasov’s ‘theory of thin-walled beams’ the conventional
method of predicting warping and distortional stresses is by beam on elastic
foundation (BEF) analogy. This analogy ignores the effect of shear deformations and
takes no account of the cross sectional deformations which are likely to occur in a
thin walled box girder structure.
A modification of BEF analogy was developed by Hsu et al (1995) as a practical
approach to the distortional analysis of steel box girders. The equivalent beam on
elastic foundation (BEF) method as it is called is an enhancement of the BEF
method. It is adoptable to the analysis of closed (or quasi-closed) box girders and
provides a simplified procedure to account for deformation of the cross section, the
effect of rigid or flexible interior diaphragms and continuity over the supports
Fan and Helwig (2002) proposed a model of steel box girder bridge where
distortional stresses are controlled by internal cross frames that restrain the cross
section from distorting. They presented an analytical study on the distortional
behaviour of box girders with a trapezoidal cross sectional shape and obtained
distortional components from different torsional loads and used them to derive
expressions for the brace forces in the internal cross frames for quasi closed box
girders.
In a paper titled ‘Strength of thin walled box girders curved in plan’, (Yakubu et al
1995) used numerical methods to study the influence of local buckling and
distortional phenomenon on the ultimate strength of thin walled, welded steel box
8
girder curved in plan. They proposed a modified stress strain curve to allow for local
buckling of plate components. In their non-linear flexure analysis, the effect of
distortion was considered through incorporating additional strain in the reference
stage of the incremental process.
Okeil and El-Tawil (2004) carried out detailed investigation on warping –related
stresses on 18 composite steel concrete box girder bridges. The differences between
stresses obtained taking warping into consideration and those calculated by ignoring
warping were used to evaluate the effect of warping. They concluded that warping
has little effect on both shear and normal stresses in all bridges.
Sennah et al (2003) used finite element model to carry out parametric study on
the shear distribution characteristics under dead and live loads in horizontally curved
composite multiple steel box girder bridges and used the results from tests on five
box girder bridge models to verify the finite element model. Based on the results from
parametric study simple empirical formulas for maximum shear (reactions) were
developed for the design office.
Osadebe and Mbajiogu (2006) employed the variational principles of cross
sectional deformation on the assumption of Vlasov’s theory and developed a fourth
order differential equation of distortional equilibrium for thin walled box girder
structures. Their formulation took into considerations shear deformations which were
reflected in the equation of equilibrium by second derivative term. Numerical analysis
of a single cell box girder subjected to distortional loading enabled them to evaluate
values of distortional displacement, distortional warping stresses and distortional
shear which they compared with BEF analogy results and concluded that the effect of
shear deformations can be substantial and should not be disregarded under
distortional loading.
Several investigators; Bazant and El-Nimeiri (1974), Zhang and Lyons (1984),
Boswell and Zhang (1984), Usuki (1986,1987),Waldron (1988), Paavola (1990),
Razaqpur and Lui (1991), Williams et al (1992), Fu and Hsu (1995), Tesar (1996),
have combined thin-walled beam theory of Vlasov and the finite element technique to
develop a thin walled box element for elastic analysis of straight and curved cellular
bridges.
Well over sixty literatures were reviewed. Various theories were postulated by
different authors examining methods of analysis, both classical and numerical. A few
others however carried out tests on prototype models to verify the authenticity of the
9
theories. At the end of it all, it was concluded that Vlasov’s theory captures all
peculiarities of cross sectional deformation such as warping torsion distortion etc, and
that generalized strain fields can be used to reduce the number of differential
equations which poses difficulties in the application of Vlasov’s theory.
2.2 BOX GIRDER BRIDGE TYPES
The cross section of box girder bridges may take the form of single cell, multi cell
(cellular) or multi spine (multi box) profiles shown in Fig.2.2.1. Multi cell box girder
type provides greater torsional stiffness than multi spine due to the high efficiency of
the contiguous cells in resisting eccentric loadings.
Goodall (1971), Grant (1993) and Dean (1994), derived solutions to the torsional
stiffness and shear flow of regular multi cellular sections that are frequently used in
modern curved high way bridges. Fu and Yang (1996) extended the solution to the
torsional design of multi- box and multi-cellular reinforced concrete bridges. A
summary of research work done on load distribution and design of various box girder
cross sections is presented in the following sections.
2.2.1 Single cell box girder bridges
Trapezoidal steel box girder bridges are often used in curved bridges due to the
large torsional stiffness that result from the closed cross section. However the use of
single cell box girders in bridge construction is a common feature in curved bridges,
highway interchanges and ramps.
Fan and Helwig (2002), carried out an analytical study on the distortional
behaviour of single cell box girder bridge with a trapezoidal cross sectional shape.
Typical torsional loads on curved box girders were presented and distortional
components of the applied torsional loads were identified and used to derive
expressions for the brace forces in the internal cross frames for quasi-closed box
girders. Turkstra and Fan (1978), investigated the effect of warping on the
longitudinal normal stresses as well as on the transverse normal stresses in single
cell curved bridges.
10
2.2.2 Multi-cell box girder bridges
In 1959, California design engineers who appreciated the large torsional rigidity of
closed cellular sections suggested to the American Association of State Highway and
Transportation Officials (AASHTO), a change in the load distribution factor (s / 5) then
used for bending moment in straight reinforced concrete box girder bridges, (Davis et
all (982). They recommended (s / 8) for one lane traffic and (s / 7) for two or more
traffic lanes, where s represents the cell width in feet. The National Standards of
Canada for design of highway bridges (CSA 1988) adopted the above specified
moment distribution factors for straight multicell bridge cross section even though
they do not give designers much information on the behaviour of the bridge or the
parameters influencing its response.
Trikha(1972) developed a set of design coefficients for two lane twin cell curved
girder reinforced concrete bridges as an aid to practical design of such bridges. The
effect of intermediate diaphragms was not considered in his study which he used
finite strip method to accomplish. Ho et all (1989) also used the finite strip method to
analyze straight simply supported, two-cell box girder and rectangular voided slab
bridges without intermediate diaphragms. Empirical expressions and design curves
were deduced for the ratio of the maximum longitudinal bending moment to the
equivalent beam moment.
2.2.3 Multiple spine box-girder bridges
Johanston and Mattock (1967) and Fountain and Mattock (1968) used a computer
program for the analysis of folded plate structures to study the lateral distribution of
loads in simple span composite multi-spine box girder bridges without intermediate
diaphragms. To verify their analysis and the authenticity of the computer program
they built and tested two-box and three-box bridge models under AASHTO truck
loading. The results were used to develop an expression for the live load bending
moment distribution factor for each box girder as a function of roadway width and
number of boxes. Their findings formed the bases for lateral distribution of loads for
(a) Single Cell (b) Multiple Spine (Multi-box) (c) Multi-cell (Cellular)
Fig.2.2.1 Box girder cross section types
11
bending moment evaluation currently used by AASHTO (1998), OMTC (1983) and
CSA (1988), for the multi-spine box-girder bridges. However, the principle does not
account for the beneficial effect of cross bracing inside and between the boxes. More
over its application is limited to bridges with a number of spines equal to the number
of lanes. Therefore, AASHTO (1998) provides another expression for load distribution
for the ultimate limit state to obtain the live load-bending moment and shear force in
each box of the cross section.
Tung and Fountain (1970) presented an approximate method to determine the
torsional moments and angle of twist in curved box girders of single and multiple
spans. The box girders were assumed to be adequately stiffened by internal
transverse diaphragms so that both warping and distortional stresses were
considered negligible. Heins (1978) developed two expressions for the live load
longitudinal bending moment and torsional moment in a straight box girder, with two
modification factors for curvature. The expressions were developed from data provide
by Ho (1972) resulting from a thorough system analysis of 90 curved multi-spine box-
girder bridges. The work also included approximate formulas, for the box section
geometry, calculation of internal forces, and top chord and cross bracing
requirements.
Bakht and Jaeger (1985,1992) presented a method for analysis of a particular
case of multi-spine bridges having at least three spines, zero transverse bending
stiffness, with the load transfer between the various spines through transverse shear.
Based on these simplifications they proposed load distribution factor for bending
moment and shear. These formed the basis for live load distribution [OMTC 1972] for
multi-spine bridges. Salaheldin and Schmidt (1991) reported a detailed study, using
the finite element method, of the flange transverse membrane stresses of a simply
supported box girder reinforced with a rigid end diaphragm under symmetric
concentrated loads acting at the mid span. Empirical expressions were formulated to
predict the distribution and magnitude of the transverse flange stresses. Normandin
and Massicotte (1994) presented the result of a refined finite element analysis to
determine the distribution of live load patterns in multi-spine box girder bridges with
different characteristics and geometry. Results indicated that in some cases both the
AASHTO (1996) and OMTC (1992) distribution factors underestimate the live load
effects by a significant margin. Two empirical equations for longitudinal bending
12
moment and shear in each box were developed for simply supported bridges when
the width of the steel section is of the same order of magnitude as the lane with.
Using the finite strip method, Cheung and Foo (1995) proposed expressions for
the ratio of the moment distribution factor of the curved composite multiple box-girder
bridge to that of the straight one. The disadvantage of this study is that it did not take
into account the beneficial effect of the diaphragms inside the box or cross bracings
between boxes. Furthermore, the effect of the number of boxes and dead load
distribution were not included. Foinquinos et all (1977), using the finite element
method assessed the influence of intermediate diaphragms on the live load
distribution of a straight three-spine box girder bridge, within the context of AASHTO
(1998).
2.3 BRIDGE LOADINGS AND RESULTANT STRESSES
Dead loads on a bridge structure are estimated based on the dimensions and
materials given in blue prints provided by the appropriate codes of practice. In USA,
live loads are considered according to AASHTO LRFD (1998) provisions. Following
the HL93-loading a uniform lane load of 9.3N/mm distributed over a 3.0 m width, is
considered in addition to a tandem load of two 110kN axles. Transversely, the loads
are positioned at the outermost possible location to generate the maximum torsional
effects, Fig. 2.3.1. Centrifugal forces (CE) are taken into account for curved segment
of the bridge. The tandem load is considered to act along the longitudinal direction of
the bridge axis, while the uniform lane load is positioned on each span separately.
(b) Tandem Load (one axle
shown)
600 b 1800
110KN
CE
1800
(a) Lane Load
3000
9.3 N/mm
Fig.2.3.1 Positioning of live loads for single-lane bridge
13
Analysis of these load cases is used to generate envelopes for the live load force
distributions. Each set includes the bending moments (Mx, My) and bi-moment (Mw)
which causes warping, the shear forces (Vx and Vy). The torsional moment (T) acting
on the cross section has two components, pure torsion (Ts) and warping torsion (Tw).
Normal stresses (including the effect of warping) are calculated as follows, (Nakai
and Yoo (1988).
y ω ωxexact approx
x y ω ω
M M MMσ = y + x + ω =σ + ω
I I I I (2.3.1)
where,
Ix and Iy = the moments of inertia about x and y axes,
x and y = distances from the centroid of the cross section,
ωM = the bi-moment
ωI = the warping constant and
ω = warping function.
The sum of the first two terms in Eqn. 2.3.1 is from classical beam theory, which does
not account for warping and is hereby referred to as approxσ .
t approx , is the approximate shear calculated from classical beam theory by ignoring
the effects of warping, using the following equation, ( Okeil and El-Tawil 2004).
τ y xapprox x y
x y o
V V T= S (s)+ S (s)+
tI tI 2tA (2.3.2)
where
Vy and Vx = shear forces,
T = torsional moment acting on the cross section.
The other terms in the expression represent the geometric properties of the closed
cross section namely; box thickness (t), moments of area [ Sx and Sy], and enclosed
cross sectional area (Ao). The third term assumes that the entire torsional moment is
Saint Venant torsion.
If warping is considered the torsional moment s w
T τ τ= +
Where,
sτ = pure torsion
wτ = warping torsion
The shear stresses can be calculated as follows;
14
τ ττ y s ωxexact x y ω
x y o ω
V V= S (s)+ S (s)+ + S (s)
tI tI 2tA tI (2.3.3)
wS = sectorial moment of area.
2.3.1 Bi-moment Effects
A bi-moment consists of two equal and opposite moments (couples) acting about the
same axis and separated from one another. Its value is the product of the moment
and the distance of separation.
The longitudinal stress at a point whose coordinates are (x, y) is given by;
y ωx
y x ω
M MMPσ = + y + x + ω
A I I I (2.4.1)
In Fig.2.3.2 the single load P can be treated as a combination of four sets of loads,
one of which represents axial loading, two of which represent bending about the axis
of symmetry of the cross section, and the last is a bi-moment. The first three sets of
loads will result in the deformation pattern familiar to engineers.
The bi-moment will result in distortions of the cross section parallel to the longitudinal
axis of the column and in a twisting of the column about its longitudinal axis [Fig.2.3.3
(b)]. In Fig 2.3.3(c) the web and flanges are separated from one another and a pair of
moments Pa/4, applied at the top of each flange. The flanges each bend as shown,
while the web remains straight. The integrity of the cross section can be restored by
twisting the web and each of the flanges through an angle q and since these
individual plates are very flexible the forces required to do this are likely to be small.
This pure twisting of each plate results in another set of stresses which are shear
stresses. They are called Saint Venant stressesst
τ , and the total torque required to
twist all of the plates in this way is the Saint Venant torque Mst. Thus it is seen that a
bi-moment can result in the twisting of a column and longitudinal stresses in the
flanges, and for the design engineer this can be very important consideration.
15
Distortion of box girders occurs when the torsional loads are not distributed to the
plates of the cross section in proportion to a uniform Saint Venant shear flow (Fan
and Helwig 2002)
2.3.2 Pure tortional and distortional loads
Depending on the source of torsion, external torsional load on a box girder can be
modeled by either two opposing vertical or two opposing horizontal forces, as shown
(a) (c)
(d) (d)
(b)
Fig.2.3.2 Components of single load P resulting in warping of the profile.
(Saint Venants principle is not valid and Bernoulli’s hypothesis is not applicable for part (e))
(a ) I-column carrying (b) Distortional component of (c) Bimoment effects:
eccentric load P the eccentric load P (distortion and twisting)
Fig.2.3.3 A bi-moment applied to a column results in twisting and longitudinal stress
16
in Fig 2.3.4 (a and b) respectively. Fig. 2.3.5 shows that torsional loads consisting of
opposing vertical forces usually result from gravity loads that are eccentric to the
centre line of the girder. The eccentric load can be modeled as a bending
component, Fig 2.3.5(b), superimposed on the torsional component, Fig 2.3.5(c).
The distortional behaviour of a box girder depends on the manner in which the
external torque is applied to the girder. A torsional load either a vertical couple or a
horizontal couple, can be modeled as a uniform torsional component superimposed
on a distortional component as shown in Figs. 2.3.6 and 2.3.7. The rectangular thin-
(a) Two opposing vertical force model
(c) Torsional Component
mT/2h mT/2h Tm
b
Tm
b
+ h mT/2b mT/2b
=
mT/2b
b mT/2b
mT/2h mT/2h (c) Distortional component (b) Pure torsional
component
(a) Torsional load
Fig.2.3.6 Torsional and distortional loads on rectangular box girder
due to vertical forces
b
(b) Bending Component
q
=
q
+
h
(a) Eccentric load
q
Concrete Slab
q
b
(b) Two opposing horizontal force model
e
Fig. 2.3.4 External torsional loads on box girder
w Girder
central
line
Fig.2.3.5 Decomposition of eccentric concrete load on box girder
2
w
wew
b ew
b
17
walled box has a respective dept and width of h and b. Each couple can be modeled
by the uniform torsional components given in Fig.2.3.6(b) or Fig.2.3.7(b). Although
the boxes in Figs.2.3.6(a) or 2.3.7(a) are subjected to the same magnitude of torsion
(MT), the resulting distortional stresses are opposite in direction since the distortional
loads shown in Figs 2.3.6(c) and 2.3.7(c) are opposite. The pure torsional
component, Figs.2.3.6(b) and 2.3.7(b) generate a uniform Saint Venant shear flow
along the circumference of the box girder cross section and warping stresses due to
this torsional component are usually negligible. However, significant distortional
warping stresses may be induced due to the distortional components if the box is not
properly braced.
Vlasov (1961) was the first to study distortion of box girders while investigating the
torsional behaviour of thin-walled beams with closed cross section. Dabrowski (1968)
established a more rigorous theory when he developed the governing equations for
box girder distortion and provided solutions for several simple cases.
A fundamental issue for the distortional analysis of a box girder is the
determination of the distortional loads which are generally not explicitly applied to box
girders, but rather are included in the applied torsional loads as shown in Figs 2.3.6
and 2.3.7 for rectangular boxes. The pure torsional components of the applied
torsional loads [Figs 2.3.6(b) and 2.3.7(b)] do not cause the cross section to distort. A
distortional load consists of four force components on the plates of the box [Figs
2.3.6(c) and 2.3.7(c)]. These four components cause the cross section to distort,
however they do not induce torsion on the girder.
The distortional behaviour of box girders can be studied by separating the
distortional component from the applied torsional loads. The distortional components
are generally resisted by both in-plane and out of plane shears in the girder plates.
Fig. 2.3.7 Torsional and distortional loads on rectangular box girder due to horizontal forces
=
h
(a) Torsional
load
(b) Pure torsional
component (c) Distortional
component
mT/h
mT/h
mT/2h
mT/2h
mT/2b mT/2b mT/2b mT/2b
mT/2h
mT/2h
b
+
18
These two shear components result in different distortional stresses. Fig.2.3.8(a)
shows the typical distortional shape of the box girder that result in out-of-plane
bending of the plate components. Fig.2.3.8(b) shows the shears that develop in the
trough thickness direction as a result of the distortion. The distortional loads on the
flanges and webs are partially resisted by these through thickness shears that
develop in the plates. Out of plane bending stresses are induced with corresponding
moments shown in Fig.2.3.8(c).
Distortional loads are also partially resisted by the in-plane shears that develop on
the cross section of the individual plates, Fig.2.3.9. The large arrows represent the in-
plane shears that resist the distortional loads represented by the small arrows on the
girder plates. The individual plates will experience in-plane bending from these
shears, and longitudinal bending stresses may be induced on the cross section. The
longitudinal bending stresses are known as the distortional warping stresses.
Fig.2.3.9(b) shows a typical distribution of warping stresses in a trapezoidal box
girder.
(a) Typical distortional shape of box gird (b) Shears in trough thickness direction as a result of distortion (c) Corresponding moments due to out of plane bending
Fig.2.3.8 Out of plane distortional stresses in box girders
Fig. 2.3.9 In-plane distortional warping stresses in box girder
(a) In-Plane shears (b) Longitudinal bending stress
19
A pure torsional load consists of a shear component proportional to the Saint Venant
shear flow which intensity is constant along the perimeter of the box cross section,
Fan and Helwig (2000). Equilibrium and mechanics of thin-walled members can be
used to determine the intensity of the uniform shear flow f, as shown in Fig 2.3.10. A
torsional moment of T = f (2Ao) results from this uniform shear flow, where Ao is the
enclosed area of the box. Therefore, for an applied torsional moment T, the
equivalent pure torsional component is represented by uniform shear flow shown in
Fig 2.3.10, in which f = T/ 2Ao. If this distributed load along the perimeter of the cross
section is replaced with the shear loads on individual plates, the shear load on the ith
plate of the box is equal to biT / (2Ao), where bi is the width of the plate independent
of plate overhangs.
Figs 2.3.11 and 2.3.12 demonstrate that the applied torsional load is a function of the
pure torsional components and distortional components. Therefore if the equivalent
pure torsional components are known [Fig 2.3.11(b) or Fig 2.3.12(b)], they can be
subtracted from the applied torque [Fig.2.3.11 (a) or Fig.2.3.12(a)] to determine the
distortional components [Fig.2.3.11(c) or Fig. 2.3.12(c). For example, if the applied
torque consists of two opposing vertical forces P (kN/m), as shown in Fig.2.3.11 (a),
the resulting distributed torsional load is Pb (kNm/m). For a trapezoidal cross-section,
the vertical forces may be resolved into components in the web and in the flanges,
Fig.2.3.11(b). Thus,
2 tant
P P θ= (a)
/ cosw
P P θ= (b) (2.5.1)
0b
P = (c)
in which the subscript t, w, and b refer to the top plate, web plates and the bottom
plate respectively.
ƒ
Fig .2.3.10 Saint Venant shear flow in box girder
20
The intensity of the ‘shear flow’ of the equivalent pure torsional load is P x b / (2Ao).
Therefore, utilizing the expression, qi = bi T / 2Ao for the shear flow in the ith plate,
the component of the pure torsional load on the top plate is, from Fig. 2.3.11(b),
2
t,pureT
o o
Pb bq = b = P
2A 2A (2.5.2)
The distortional component on the girder plates can be derived by subtracting the
pure torsional components from the applied torsional load. For example, the
distortional component on the top flange of the box is given by;
2 2
t,dist t t,pureT
o o
Pb aq =P - q = 2Ptanφ - = - P
2A 2A (2.5.3)
in which the relation, 2Ao = h(a+b) and φtan = (b-a) / 2h , have been applied. Other
torsional and distortional components on the web and bottom plates of the box can
be similarly derived. The results are given in Table 2.1 where the expressions
represent the magnitude of the force components. The directions of pure torsional
and distortional components are shown in Fig. 2.3.11.
b
qt,dist
p
qw,dist
h
qw,Pure T
Pt
=
Pw
qt,pure T
=
p
p
h
a
p
θ
qw,Pure T
c
qb,Pure T
Pw
qw,dist
=
qb,dist
+
,t pureTq
Fig. 2.3.12 Pure torsional and distortional components in horizontal couple
,b pureTq
(a)
,wpureT
q
(b)
,wpureT
q
(c)
,t distq
,w distq
,w distq
,b distq
(a) (b) (c) (d)
Fig. 2.3.11 Pure torsional and distortional components in a vertical couple
+
21
Table 2.1 Pure torsional and distortional components of applied torsional
loads (Fan and Helwig 2002)
If the torsional load consists of two opposing horizontal forces acting on the top and
bottom plates of the box girder [Fig. 2.3.12(a)], the applied torque is Ph. The applied
torsional load consist of Pt = Pb= P and Pw = 0. The pure torsional and distortional
load components shown in Fig.2.3.12 (b and c) can be derived in a similar way
outlined above. The results are also presented in Table 2.1. The directions of the
force components are indicated in Fig. 2.3.12(b and c).
2.4 METHODS OF ANALYSIS OF THIN-WALLED BOX GIRDERS
During the time when ignorance superseded knowledge of the behaviour of box
girder beams, elementary beam theory (Bernoulli beam theory) was applied. The
usual assumption of beam bending theory is that plane sections remain plane over
the entire cross section of say a box girder, after bending. However it is easily
demonstrated that when certain thin-walled structures are twisted there is a so-called
warping of the cross section and Bernoulli hypothesis is violated. In other words,
warping brings about total violation of Bernoulli’s theory. Thinness of wall also leads
to distortion of box cross-section thereby also violating Bernoulli’s beam theory. All
these suggest that application of ordinary beam theory is not suitable for analysis of
Type / Magnitude
of load
Components in plate
members
Pure torsional
component
Distortional
component
Torsional moment
by vertical forces
(Pb)
Top plate, Pt =2Ptan θ
Bottom plate, Pb = 0
Web plate, Pw=P /cos θ
Pb2 / 2Ao
Pab / 2Ao
Pbc / 2Ao
Pa2 / Ao
Pab / 2Ao
Pac / 2Ao
Torsional moment
by horizontal
forces (Ph)
Top plate, Pt = P
Bottom plate, Pb = P
Web plate, Pw = 0
Pb / (a+b)
Pa / (a+ b)
Pc / (a+b)
Pa / (a+b)
Pb / (a+b)
Pc / (a+b)
22
box girder structures. This necessitated for a search for more accurate method.
Vlasov’s ‘theory of thin-walled beams’ (1951), translated into English language by
American National Science Foundation (USA NSF) and called ‘Thin-walled elastic
foundation (1961)’ was most accepted and adopted.
Some codes of practices, AASHTO (1996), CHBDC (2000), have recommended
several methods of analysis for only straight box girder bridges. These methods
include; (i) orthotropic plate theory, (ii) finite difference technique, (iii) grillage
analogy, (iv) folded plate, (v) finite strip and (vi) finite element techniques. Several
authors have applied these methods along with the thin-walled beam theory to
analyze straight and curved box girder bridges. These different approaches to box
girder bridge idealizations are further discussed in the following sections.
2.4.1 Orthotropic plate theory
In this method, the stiffness of the flanges and webs are lumped into an
orthotropic plate of equivalent stiffness and the stiffness of diaphragm is distributed
over the girder length. This method is suggested mainly for multi spine straight and
curved bridges. Bakht et al (1981) presented the various methods of calculating the
equivalent plate parameters which are necessary for analysis of straight cellular and
voided slab bridges. Cheung et al (1982) used the orthotropic plate method to
calculate the longitudinal moments and transverse shear in multi-spine box girder
bridges. The results were compared to those obtained from 3D analysis using the
finite-strip method to establish the limits of validity of the orthotropic plate method. It
was concluded that the orthotropic plate method gives accurate results provided that
the number of spines is not less than three.
2.4.2 Grillage analogy method
Canadian Highway Bridge Design Code (CHBDC 2000) limits the application of
grillage analogy method to voided slab and box-girder bridges in which the number of
cells or boxes is greater than two. In this method, the multi cellular super structure
was idealized as a grid assembly by Hambly and Pennels (1975). Similar idealization
was applied to curved multi-spine box girder bridges by Kissane and Beal (1975).
One difficulty in the grillage-analogy method lies in the representation of torsional
stiffness of closed cells. Satisfactory but approximate representation can be achieved
in modeling the torsional stiffness of a single closed cell by an equivalent I-beam
torsional stiffness,(Evans and Shanmugam 1984).
23
2.4.3 Folded plate method
The folded plate method utilizes the plane stress elasticity theory and the classical
two-way plate bending theory to determine the membrane stresses and slab moment
in each folded plate membrane The folded plate system consists of an assemblage of
longitudinal annular plate elements interconnected at joints along their longitudinal
edges and simply supported at the ends. No intermediate diaphragms are assumed.
Solution of simply supported straight and curved girders is obtained for any arbitrary
longitudinal load function by using direct stiffness harmonic method. The method has
been applied to cellular structures by Meyer and Scordelis (1977), Al-Rifaie and
Evans (1979) and Evans (1984). It was evident that the method was complicated and
time consuming.
2.4.4 Finite strip method
The finite strip method may be regarded as a special form of the displacement
formulation of the finite-element method. In principle, it employs the minimum total
potential energy theorem to develop the relationship between unknown nodal
displacement parameters and the applied load. In this method, the box girders and
plates are discretized into annular finite strips running from one end support to the
other and connected transversely along their edges by longitudinal nodal lines. The
displacement functions of the finite strips are assumed as a combination of
harmonics varying longitudinally and polynomials varying in the transverse direction.
2.4.5 Beam on elastic foundation (BEF) analogy
The response of a box cell to a loading that causes deformation of the cross
section can be represented by a differential equation identical in form to that for
beams on elastic foundations.
A direct analogy exists between the physical properties of the box cell and the beam
on elastic foundation. To develop an approximate theory of deformation of the cross
section one basic assumption regarding this in-plane motion is borrowed from the
theory of torsion of thin-walled beams of open section. It is assumed that the
distortions are accompanied by sufficient warping to annul the average shear strains
in the plates which form the cross section. The measure of distortion ω leads to
warping displacements proportional to ω' . Warping stresses depend on ω'' and
shear on ω''' . The forces per unit length which have their origin in warping and resist
24
deformation are then proportional to ω''' . These considerations suggest a governing
differential equation for ω in the form:
iv
ωωEI ω +kω =P (2.6.1)
in which
ωωI = warping constant
K = measure of deformation stiffness of a unit length of the box, and
P = the applied generalized distortional load per unit length
For distributed load P and beam deflection ω , the differential equation for the BEF
is
iv 4ω + 4β ω =P (2.6.2)
In which 4
b
Kβ =
4EI (2.6.3)
2.4.6 Finite element method
The finite element method which has been used extensively in structural
mechanics has also been used in the analysis of box girder bridge decks. Many
computer programs have been developed using the finite element procedure.
Modern computerized systems such as NASTRAN and BERSAFE for buckling and
stress analysis incorporating different types of finite elements have been developed
for elastic and in-elastic structures. Several finite element programs have also been
developed specifically for the analysis of box-girder bridges by Scordelis (1971).
There are some simplified approaches for box analysis such as those developed by
Sawko and Cope (1969) and Crisfield (1971) which retain the finite element method
of solution with reduced computer time.
2.5 SUMMARY OF LITERATURE SURVEY
Based on the numerous literatures consulted during the literature survey the
following observations and comments can be made.
1. Research work done on thin-walled box girder structures covers essentially three
types of cross section. (a) Single cell steel box girder structure, straight and curved.
(b) Twin cell reinforced concrete box girder, straight and curved. (c) Two-box and
three-box multiple spine box girder, steel and reinforced concrete
2. Literature on cellular steel box girders with three or more cells appears to be
scarce. Those available for two-celled box girder are in RC area. Thus, there appears
25
to be a dirt of information on the torsional-distortional behaviour of thin-walled box
girder bridge structure with three or more cells in cross section.
3. Specific areas covered by the available literature include:
(a) torsional stiffness and shear flow
(b) distortional design of multi-box and multi-cellular sections.
(c) distortional behaviour and brace forces
(d) effect of warping on longitudinal and transverse normal stresses
(e) design coefficients for loads and aids to practical design
(f) analysis of deformable sections / design curves
(g) load distribution
(h) approximate method to determine torsional moments- non deformable
sections
(i) bending moment expressions, shear and torsional moments for live loads
(j) bracing requirements
4. Available literature on torsional-distortional elastic behaviour of thin-walled
deformable box girder structure is surprisingly few. This could be attributed to the fact
that most authors assume that the use of intermediate stiffeners and diaphragms on
thin-walled closed and quasi-closed structures is an effective way of handling non-
uniform torsion and its attendant problems of warping and distortion.
5. It is believed that the use of triangular cells on trapezoidal cross section will
improve the torsional-distortional qualities of trapezoidal box girders and at the same
time make the use of diaphragms and intermediate stiffeners irrelevant. This is a
novel area for research.
2.6 THEORY OF THIN-WALLED BOX GIRDERS
2.6.1 Warping stresses and shear centre When a thin-walled girder is restrained from warping additional stresses arise in
the longitudinal and transverse directions. These stresses do not arise in the case of
uniform (Saint Venant ) torsion. Batch (1909) first noted these stresses by
considering a channel for which the centre of gravity S and the shear centre M do not
coincide. He observed that a transverse load P acting at the end of a cantilever
through S produced vertical deflection and rotation of the cross section about a
longitudinal axis. He also noticed that plane sections do not remain plane and that
warping out of the plane had occurred. By systematically changing the position of P
26
he found the position M (the shear center) and noted that when P acted through M,
Benoulli’s hypothesis was satisfied and the additional stresses also disappeared
Thus, in a thin-walled beam which is twisted only, there can be two sets of shear
stresses (and shear strains), that is, the Saint Venant shear stresses τ s and the
warping shear stress τω
.These are quite different from shear stresses associated
with bending.
There are no longitudinal stresses associated with Saint Venant (uniform ) torsion,
but there are with warping (non-uniform) torsion and they, together with the
associated warping shear stressesτω
, are called the warping stresses.
Every cross section has a shear centre M which is the point through which the
shear force must be applied if there is to be no twisting. When a cross section has
two planes of symmetry (doubly symmetric), the shear centre M and the centre of
gravity S coincide. If there is one plane of symmetry ( e.g. as there is for a channel)
M and S are located on the same plane and it is a relatively simple matter to
establish their coordinates. However when the cross section is less regular a
systematic method for locating M is required. The analysis in section 2.6.2 treats
general case of both open and closed profiles (mono symmetric and doubly
symmetric sections).
2.6.2 Saint Venant torsion of thin-walled cross section
The purpose of this section is to present back up theory for the analysis of beams
of various cross sections, i.e., open profiles, closed profiles and mixed profiles in
which part is open and the other part is closed.
Open Profile
The simplest example of open profile is a solid shaft of circular cross section.
Consider the case of a cylindrical shaft of radius R, length L, and applied torque M.
From the study of strength of materials the following are the expressions for the
maximum shear stress τmax , the maximum shear strain γ max ( i.e. at the outer
surface), and the angle of rotation φ of one end relative to the other end
τmax
p
MR=
J (2.6.1)
27
maxmax
G
τγ = (2.6.2)
p
MLφ =
GJ (2.6.3)
where, 2
pJ =πR /2 (2.6.4)
is the polar moment of inertia (for solid shaft) and GJp is the torsional stiffness of the shaft. Closed Profile
In the case of closed profile, Fig.2.6.1, it is seen that for longitudinal equilibrium of
the element A, τ τ1 1 2 2t dz = t dz i.e. T1 = T2 where T1 and T2 are shear flow, tτ=T
Thus the shear flow around a closed profile is constant. The moment of a force on an
element at C about an arbitrary point B is
BdM = Tρ ds (2.6.5)
By integrating around the cross section noting that
BdA =1/2ρ ds and ∫dA =1/2∫ Βρ ds = A , it is seen that;
M = 2TA (2.6.6) where A is the area enclosed by the profile.
The angle of twist is found by equating the internal and external energies.
ττ∫ ∫� �
21 L L tMφ = tγds = ds
2 2 2 G
Inserting eqn. (2.8.6), we obtain that
Bρ
B A
a
b C
•
1t
2t
2τ
1τ
dz
A
Fig.2.6.1 Closed Profile
28
∫
∫�
� 2
ds/tφ dφ 1 T M= = ds =
L dz 2GA t G 4A (2.6.7)
Comparing with eqn. (2.6.4) we see that for a closed profile with one cell the
equivalent property of the cross section to the polar moment of inertia is
∫�
2
p
4AJ =
ds/t (2.6.8)
This property is called the Saint Venant torsion constant or simply the torsion
constant.
Thin rectangular profile
Equation (2.6.5) can be used to derive the torsional property of a thin rectangle, (Fig
2.6.2), the element being treated as an open profile (Murray 1984).
Thus, the torsion constant of an isolated plate is
3
p
1J = at
3 (2.6.9)
τ 2
max
1M = at
3 (2.6.10)
For open cross section comprised of several thin rectangles joined together the
torsion constant is ∑ 3
p n n
1J =K a t
3 (2.6.11)
where K is a factor which makes allowance for small fillet. K is 1 for thin sections bent
from a sheet and 1.2 to 1.3 for rolled sections.
Closed Profile and one or more Open Profiles (Mixed Cross Section)
For mixed cross section the results obtained in closed profile and open coss
section of several thin rectangles joined together can be added. The total torque M is
the sum of the torques carried by the open profiles, ∑ openM and that carried by the
closed profile closedM . Hence,
maxτ a
t
Fig. 2.6.2 Rectangular profile
29
( ) ( ) ( ) ∑ ∑closed p popen open closed
dφM = M +M = G J + J
dz (2.6.12)
where Jp is the torsion constant of the entire cross section.
The torque carried by any single plate or the closed profile is obtained by simple
proportion.
( )∑ p open
open
p
JM =M
J ;
( )
p
p closedclosed
JM =M
J (2.6.13)
Having obtained the torque carried by, say, a plate of the cross section, the stress
can be found by applying the relevant formula.
Several Closed Profiles
When a cross section consists of several closed profiles joined together, Fig. 2.6.3,
equilibrium considerations at a joint such as A, show that the shear flow in the web is
;
web i i-1T = T - T
For the ith cell we have from eqn (2.6.6)
ii iM = 2A T
Hence, for the whole cross section
∑ i iM = 2A T (2.6.15)
The angle of twist is obtained from eqn (2.6.7), thus
∫ ∫ ∫ ∫B D
i-1 i i+1
ii i A C
dφ 1 Tds 1 ds ds ds= = -T +T - T
dz 2GA t 2GA t t t (2.6.16)
Since dφ
dz is the same for all cells, it is seen that a set of simultaneous equations in
the unknowns iTψ =
Gdφ/dz (2.6.17)
can be established and solved.
Substituting ψ value in eqns. (2.6.15) and (2.6.16) respectively the following
relationships are obtained.
∑ i i
dφM = G ψA
dz (2.6.18)
∫ ∫ ∫B D
i i-1 i i+1
A C
2A = -ψ ds/t +ψ ds/t -ψ ds/t (2.6.19)
i = 1
2
3
a
b
c
A B
D C
Fig. 2.6.3 Several closed profiles
2 3webT T T= −
2T
3T
30
Hence the torsion constant of a multi-celled profile is ∑p i iJ = 2 ψA .
When the profile consists of a single cell only then
∫�T 2A
ψ = =Gφ' ds/t
(2.6.20)
2.6.3 The out - of – plane displacement of profiles
Consider an element from a beam with open profile, Fig 2.6.4. The shear strain at
the middle surface is assumed to be zero. Hence,
( ) ( )γ∂ ∂
∂ ∂zs
u v= z,s + z,s = 0
s z (2.6.21)
It is also assumed that the shear strain normal to the surface, γ zn , is zero. By
integrating eqn.(2.6.21) between arbitrary starting point V and point i on the profile
distance s from V, Murray (1984) obtained the expression for the displacement in the
longitudinal, z direction as;
( ) ( ) ( ) ( ) ( ) ( ) ( ) ( )∫ ∫ ∫s
s s' ' '
B B B0 0
0
u z,s = ξ z - ε z cosα s ds -η z sinα s ds -φ z ρ s ds (2.6.22)
where, ( )zξ is a function of z only and arises from the integration, and the primes
indicate differentiation with respect to z.
Noting that ( )dx = dsCosα s and ( )dy = dsSinα s (2.6.23)
The following function is defined.
( ) ( )∫s
B B0
ω s = ρ s ds (2.6.24)
This term is called the warping function. It represents a sectorial area, the
magnitude depends on the location of the pole B and the point V in the profile from
which the integration is started (sectorial origin).
z , u
s , v
n , w dz ds
( )st
Fig.2.6.4 Element from a beam with 0pen profile
31
Substituting eqns.(2.6.23) and (2.6.24) into eqn. (2.6.22) the following expression
for u(z, s) is obtained.
' ' '
B B Bu(z,s) = ζ(z) - ε (z)x(s) -η (z)y(s) -φ (z)ω (s) (2.6.25)
where ζ(z) ≡ longitudinal extension, uniform over the cross section
≡'
Bε (z)x(s) rotation about the y axis
≡'
Bη (z)x(s) rotation about the x axis
≡'
Bφ (z)ω (s) the distortion or the warping of the cross section out of its plane
due to twisting.
Murray (1984) also considered the normal and tangential components of
displacement of a thin walled section rotating as a rigid body through an angle φ .He
obtained expressions for the components of the displacements as follows;
B B Bv(z,s) = ε cosα+η sinα+ρ φ (2.6.26)
B B Bω(z,s) = -ε sinα+η cosα+q φ (2.6.27)
where, the terms Βρ and Bq are the components of a radius vector Ai. He also
proved that Βρ is the perpendicular from the pole B to the tangent to the middle line
of the profile at the point i.
2.6.4 Von Karman and Christensen’s theory for closed profile
Consider an element from the ith cell of a box girder with closed profile .Von
Karman and Christensen , in their approximate theory for closed profile, observed
that the shear strain is not zero in this case but is given by the stress-strain
relationship expressed in terms of Saint Venant shear flow Ti in the ith cell.
Thus,
γ∂ ∂
∂ ∂zs
u(z,s) v(z,s) Ti(z)= + =
s z Gt(s) (2.6.28)
From eqns. (2.6.15) and (2.6.18) we have
ii '
T(z)ψ (z) =
Gφ (z) [for multi cell]
and
∫i
T(z) 2Aψ (z) = =
Gφ'(z) ds/t [for single cell] (2.6.29)
32
Hence,γ 'izs
ψ (z)= φ (z)
t(s) (2.6.30)
From eqns. (2.6.28), (2.6.30) and (2.6.26) we obtain
∫ ∫ ∫s s s
' ' ' iB B B
0 0 0
ψ (z)u(z,s) = ς(z) - ε (z) cosα(s)ds -η (z) sinα(s)ds -φ (z) ρ (s) - ds
t(s) (2.6.31)
This equation has the same form as eqn (2.6.22). Thus eqn. (2.6.22) is valid for both
open and closed profiles provided ( )sΒρ is correctly evaluated from eqn.(2.6.24) or
eqn.(2.6.26) and eqn.(2.6.27).
2.6.5 Warping function
From Fig 2.6.5 and eqn (2.6.24) it is seen that the warping function is an area whose
magnitude depends on the location of the pole, B and of point V in the profile from
which the integration is started (starting point).
Suppose that the beam is simply twisted about the axis through the pole B without
bending in either x or y directions and without longitudinal extension.
Under these circumstances,
' '
B Bζ(z) = ε (z) =η (z) = 0
Furthermore, if the angle of twist '( )zφ could be made equal to -1 radian per unit
length, eqn. (2.6.25) becomes
Bu(z,s) =ω (s) .
Thus it is seen that the warping function is the out of plane displacement of the cross
section when the beam is twisted one radian per unit length about its axis through the
pole without bending in either the x or y directions and without longitudinal extension.
B
A x
y ds
Middle line of
profile
i ½dωB(s)
ρB(s)
y
x
ωi(s) s B
Fig 2.6.5 The physical meaning of warping function
33
Warping function represents a scaled value of the warping or out of plane distortion
of the cross section. When they are plotted on the cross section in the direction of z
axis they give a pictorial representation of the warping.
34
CHAPTER THREE
ENERGY FORMULATIONS AND GOVERNING EQUATIONS OF EQUILIBRIUM
3.1 BACKGROUNG PRINCIPLES
The elastic strains and stresses as well as their corresponding strain relations for
a thin walled box girder can be obtained after the formulation of Vlasov’s
displacement functions. In this study generalized strain fields (modes) are employed
in order to reduce the number of unknown displacement functions in the equations of
equilibrium. Torsional- distortional analysis is accomplished by using energy
principles on the basis of minimum potential energy to obtain the energy functional of
thin walled box girder structures with arbitrary deformable and or non -deformable
cross sections. Using the principles of variational calculus the energy functional is
minimized with respect to its functional variables which are displacement functional
and their first and second order derivatives, to obtain the equations of equilibrium in
the form of a system of linear differential equations in displacement quantities.
The coefficients of these equations are determined using Morh’s integral for
displacement computations. By solving these linear differential equations the
displacement functions (longitudinal and transverse displacements) are obtained and
hence, normal and shear stresses are computed and the distortional bending
moments are also evaluated. These require the evaluation of cross sectional
parameters such as shear centre, torsion constant, warping constant and warping
function.
3.2 PRINCIPLE OF MINIMUM POTENTIAL ENERGY
Let us consider a structure in equilibrium whose deformed configuration is
characterized by the displacement field ui. Recall the principle of virtual work which
states that if a structure is in equilibrium and remains in equilibrium while it is subject
to a virtual distortion the external virtual work ( EδW ) done by the external forces
acting on the structure is equal to the internal virtual work (δU) done by the internal
stresses. If we consider a class of arbitrary displacements iu which are consistent
with all constraints imposed on the body, the iu will in general, differ from the actual
displacements by some amount, say iδu . That is
i i iu = u +δu (3.2.1)
35
where the variation ( iδu ) must vanish at those points in the structure where
constraints (supports) exist. Quite clearly the variation iδu is equivalent to the virtual
displacement ( iδu ). Likewise, the internal virtual work (δu ) may be interpreted as the
variation in the strain energy resulting from the variation in the displacement field,
(Touchert 1974). Thus we see that the internal virtual work may be regarded as the
first variation of the strain energy due to variation in strain components ije .
Similarly, the external virtual work ( EδW ) may be regarded as the work done by
the surface and body forces during a variation iδu in the displacement field. For most
types of external forces it is possible to define potential functions which, when
differentiated with respect to the displacement components, yield the corresponding
force components. For example, if i
G(u ) and ig(u )represent the potentials of the
surface forces iT and body forces if , respectively, then
∂
∂i
i
GT = -
u and
∂
∂i
i
gf = -
u (3.2.2)
When such potential functions exist, the external virtual work becomes
∂ ∂E EW = V (3.2.3)
where the potential of the external forces EV is given by
∫ ∫E
s v
V = Gds+ gdv (3.2.4)
Suppose for instance that the surface and body forces are functions of positions only;
i.e., they are independent of the deformation of the structure. Such external forces
are said to be conservative, and from eqn (3.2.2) it follows that
i iG = -Tu and i ig = -fu (3.2.5)
In this case the potential of the external force becomes
∫ ∫E i i i i
s v
V = - Tu ds - fu dv (3.2.6)
Thus, for a structure which possess a strain energy u and an external potential VE
the principles of virtual work equation may be written as
E Eδu -δW = δ(u+ V ) = 0 (3.2.7)
or δπ = 0 (3.2.8)
in which Eπ = u+ V , (3.2.9)
36
is the total potential energy of the structure.
Equation (3.2.8) is a mathematical statement of the principle of minimum potential
energy (pmpe) of a system. It may be stated as follows:
Of all displacement fields which satisfy the constraint conditions, the correct state is
that which makes the total potential energy of the structure a minimum.
It is important to note that while the principle of virtual work (pvw), is valid for both
elastic and inelastic structures subject to arbitrary loads, the pmpe is applicable only
to elastic structures (linear and non linear) acted upon by forces which are derivable
from potential functions.
3.3 PRINCIPLE OF VARIATIONAL CALCULUS
Many problems in engineering and physics are most naturally formulated in terms
of extremum principles. Ordinary extremum problems of the differential calculus
involve finding the extreme values (maxima or minima) of a function of one or more
independent variables. Recall that a function of a single variable u(x) will posses an
extremum at a certain point only if the first derivative of the function vanishes at that
point ( du
= 0dx
). Likewise a necessary condition for the existence of an extreme value
of a function of n variables u(xi . . . .,xn ), is that all its partial derivatives of first order
are zero (∂ ∂
∂ ∂i n
u u=..... = 0
x x ).
Let us now find, among a set of admissible functions, that function which maximizes
or minimizes a certain functional (i.e., a function of functions). Consider the problem
of determining the function u(x) which makes the following integral a minimum.
∫b
aI(u) = F(u,u',x)dx (3.3.1)
And which in addition, satisfies the prescribed end conditions, Fig 3.3.1.
u(a) = ua ; u(b) = ub (3.3.2)
The integrand F(u,u,x) is presumed to be a known function of 'u, u = du/dx, and .... x;
I is the functional to be minimized. The branch of mathematics concerned with
problems of this nature is known as the calculus of variation.
37
Let us hypothesize that the curve u(x) shown in Fig.3. 3.1 is the actual minimizing
function. In the calculus of variation we are interested in the behaviour of I(u) when
the curve u(x) is replaced by slightly different curve , say u(x) . A function u(x) in
the neighbourhood of u(x) can be represented in the form
u(x) = u(x)+ εη(x) (3.3.3)
where ε is a small parameter. The difference between u(x) and u(x) is called the
variation in u(x) and is denoted by
δu = u(x) -u(x) = εη(x) (3.3.4)
In other words the variation δu is a small arbitrary change in u from the value of u
which minimizes the integral Iu
Similarly, the difference between the slope of the minimizing curve u(x) and the
slope of the varied curve u(x) is called the variation in the slope and is denoted by
δu' : that is, δu' = u'(x) -u'(x) (3.3.5)
Comparing eqn.(3.3.5) with the expression obtained by differentiating eqn (3.3.4),
namely, (δu)' = u'(x) -u'(x) = εη'(x) (3.3.6)
it is seen that δu' = (δu)' = εη' (3.3.7)
Hence the process of variation and differentiation are permutable.
The difference I∆ between the minimum value of the integral I in eqn.(3.3.1) and
the value of I evaluated from the varied curve u(x) is given as ( Tauchert 1974),
∫ ∫ ∫b b b
a a a∆I = F(u,u',x)dx - F(u,u',x)dx = ∆Fdx (3.3.8)
For higher order variations ∆F is given by
21∆F = δF+ δ F+...
2! (3.3.9)
or ∫b
2
a
1∆I = (δF+ δ F+...)dx
2! (3.3.10)
The first and second order variations of the integral I are defined, respectively, as
a b
u
du
u(x)
_
( )u x
au
bu _
_
u
x
Figure 3.3.1 Boundary conditions of an integral
38
∫b
aδI = δFdx (3.3.11)
And ∫b
2 2
aδ I = δ Fdx (3.3.12)
Similarly the nth variation of I is defined by
∫b
n n
aδ I = δ Fdx (3.3.13)
Introducing these definitions into eqn. (3.3.10) gives
21∆I = δI+ δ I+...
2! (3.3.14)
A necessary condition for the integral I to have an extremum is that the first
variation Iδ vanishes. If the second variation 2δ I is positive definite, the extremum is a
maximum; if 2δ I is negative definite, it is a maximum.
3.3.1 Euler-Lagrange Equations
As demonstrated above, a necessary condition that the definite integral eqn.
(3.3.1) have an extremum under the boundary conditions eqn. (3.3.2) is
∂ ∂
∂ ∂∫b
'
'a
F FδI = ( δu+ δu )dx = 0
u u (3.3.15)
Noting that d
δu' = (δu)dx
, the second term in the integrand of eqn.(3.3.15) can be
integrated by parts giving:
∂ ∂ ∂
∂ ∂ ∂ ∫ ∫
bb b
I
I I Ia aa
F F d Fδudx = δu - δudx
u u dx u (3.3.16)
Therefore ∂ ∂ ∂
∂ ∂ ∂ ∫
bb
aa
F d F FδI = - δudx + δu = 0
u dx u' u' (3.3.17)
Since the variation δu = u(x) -u(x) vanishes at the end points x = a and x = b, the
integrated term in eqn (3.3.17) vanishes also. Furthermore since δu is arbitrary in the
range a < x < b, the bracketed term inside the integral must also vanish
independently; thus
∂ ∂
∂ ∂
F d F- = 0
u dx u' (3.3.18)
39
This differential equation is called the Euler-Lagrange equation. It represents a
necessary but not sufficient condition which the function u(x) must satisfy if it is to
yield an extremum for I(u). The Euler-Lagrange equation can be generalized to the
case in which the integrand F contains higher order derivatives of u. For example the
Euler-Lagrange equation for the integral
( )∫b
aI = F u,u',u'',x dx , is given by
∂ ∂ ∂
∂ ∂ ∂
2
2
F d F d F- + = 0
u dx u' dx u'' (3.3.19)
3.4 V.Z.VLASOV’S THEORY
3.4.1 Introduction
The theory of thin-walled structures whose transverse cross sections are arbitrary
and which contains several closed contours was developed by Vlasov (1958). A
prismatic frame is considered to consist of an infinite number of narrow strips having
the form of planer frames connected by absolutely rigid bars; the bars transmit
normal (Ns) and shearing (S) forces from frame to frame. This means that normal
and tangential stresses σ and τ , in a cross section (x = constant) are distributed
uniformly across the thickness δt of a strip frame. Normal (Ns), shearing (S), and
transverse (Q), forces and also bending moments (Ms) develop on areas of
longitudinal section (s = constant), Fig. 3.4.1. The position of a point on the middle
surface of a prismatic frame is determined by the coordinates x and s. Since the
plates are in plane stress state, the strain in them can be completely described if the
displacements, u(x,s) and v(x.s), of point c in the x-direction and along the tangent to
the contour of the transverse section, produced by the load P(x,s) and q(x,s) are
known.
o
o o o
o o
o
o o
o o o
o o
o o
o o o
o o
s
τ σ s
Q
40
3.4.2 Energy formulation of the equilibrium equations
The longitudinal warping and transverse (distortional) displacements given by Vlasov
(1958) are
ϕu(x,s) =U(x) (s) (3.4.1)
v(x,s) = V(x)ψ(s) (3.4.2)
where U(x) and V(x) are unknown functions governing the displacements in the
longitudinal and transverse directions respectively, and ϕ and ψ are generalized
warping and distortional strain modes respectively. These strain modes are known
functions of the profile coordinates, and are chosen in advance for any type of cross
section. The displacements may be represented in series form as;
ϕ∑m
i ii=1
u(x,s) = U (x) (s) (3.4.3)
∑n
k kk=1
v(x,s) = V (x)ψ (s) (3.4.4)
where, Ui(x) and Vk(x) are unknown functions which express the laws governing the
variation of the displacements along the length of the space frame.
ϕi(s) and kψ (s) are elementary displacements of the strip frame, respectively out of
the plane (m displacements) and in the plane (n displacements).
These displacements are chosen among all displacements possible, and are called
the generalised strain coordinates of a strip frame.
3.4.2.1 Vlasov’s stress – strain relations
Fig. 3.4.1 Vlasov’s model for stress and strain in a strip frame
41
From the theory of elasticity the strains in the longitudinal and transverse
directions are given by;
ϕ∂
∂∑m
i ii=1
u(x,s)= U'(x) (s)
x and
∂
∂∑
n
k kk=1
v(x,s)= V '(x)ψ (s)
x (3.4.5)
The expression for shear strain is γ∂ ∂
∂ ∂
u v(x,s) = +
s x
or ( ) ϕγ ∑ ∑m n
i i k ki=1 k=1
'(s)U (x)+ ψ (s)V '(x)x,s = (3.4.6)
Using the above displacement fields and basic stress-strain relationships of the
theory of elasticity the expressions for normal and shear stresses become:
ϕ∂
∂∑m
i ii=1
u(x,s)σ(x,s) = E = E (s)U'(x)
x (3.4.7)
γ ϕτ ∑ ∑m n
i i k ki=1 k=1
(x,s) = G (x,s) = G '(s)U (x)+ ψ (s)V '(x) (3.4.8)
The (m + n) functions sought for, iu (x) and kv (x) , are determined from (m + n)
equations for the strip frame, obtained by equating to zero the work done by external
and internal forces in (m + n) independent virtual displacements, Fig. 3.4.2.
Every virtual displacement is as a result of an infinitesimal variation experienced
by one of the generalized strain coordinates which determine the position of all joints
and bars of the frame. This application of the principle of virtual displacements is
called the method of variations.
Transverse bending moment generated in the box structure due to distortion is given
by; ( ) ∑n
k kk=1
M x,s = M (s)V (x) (3.4.9)
where Mk(s) = bending moment generated in the cross sectional frame of unit width
due to a unit distortion, V(x) = 1
42
3.4.2.2 Potential energy functional
The potential energy of a box structure under the action of a distortional load of
intensity q is given by:
Eπ =U+ W (3.4.10)
Where,
π = the total potential energy of the box structure,
U = Strain energy,
EW = External potential or work done by the external loads.
From strength of material, the strain energy of a structure is given by
τ
∫ ∫2 2 2
(s)L S
1 σ (x,s) (x,s) M (x,s)U = + t(s)+ dxds
2 E G EI (a)
and work done by external load is given by;
EW = qv(x,s)dxds= ϕ∑∫∫ h h
s x
q V (x) (s)dsdx = ∑∫ h h
x
q V dx (b)
Substituting expressions (a) and (b) into Eqn.( 3.4.10) we obtain that,
τ
∫ ∫
2 2 2
L S
1 σ (x,s) (x,s) M (x,s)π = + t(s)+ - 2qv(x,s) dxds
2 E G EI(s) (3.4.11)
where,
o 1
m = 6
n = 5
• • •
• • •
1 2 3
4 5 6
o o
o 2
3 o
4 5
(b) Double cell
(c) Triple triangular cell
•
• •
• • 2 1 3
4 5 o
o o
o 1
2
3 4
m = 5
n = 4
• •
• •
1 2
3 4 o
o o
o 1
2
3 4
(a) Single cell
m = 4
n = 4
Fig.3.4.2 Independent (m+n) virtual displacements
43
σ(x,s)= Normal stress
τ (x,s) = Shear stress
M(x,s) = Transverse distortional bending moment
q = Line load per unit area applied in the plane of the plate
3
(s) 2
t (s)I =
12(1- ν )= Moment of inertia
E = Modulus of elasticity
G = Shear modulus
ν = poisson ratio
t = thickness of plate
Substituting the expressions for σ(x,s) eqn (3.4.7), τ (x,s) eqn.(3.4.8), M(x,s)
eqn.(3.4.9) and v(x,s) eqn.(3.4.2) into eqn.(3.4.11) we obtain that:
ϕ ϕ∑ ∑i i j jπ =E (s)U'(x) * (s)U ''(x) * t(s)dsdx +
+ i j
ϕ ϕ ∑ ∑ ∑ ∑i k k j h hG '(s)U (x)+ ψ (s)V '(x) * '(s)U (x)+ ψ (s)V '(x) * t(s)dsdx +
+ ∑ ∑
n n
k k h hk=1 h=1
1M (s)V (x) * M (s)V (x) dsdx
EI - ∑∫ hh
x
q V dx
Simplifying further noting that t(s)ds = dA we obtain;
ϕ ϕ∑ ∑m m
i j i ji=1 j=1
1π = E U'(x)U '(x) * (s) (s) * dAdx
2+
+ ϕ ϕ∑ ∑m
j=1
m1 ' 'G U (x)U (x) (s) (s)dAdxi j i j2 i =1
+
+j
ϕ∑ ∑n n
j k kj=1 k=1
1G U (x)V '(x) * '(s)ψ (s) * dAdx
2 +
+ ϕ∑ ∑m m
i h i hh=1 i=1
1G U (x)U '(x) * '(s)ψ (s) * dAdx
2 +
44
+ ∑ ∑n n
k h k hh=1 k=1
1G V '(x)V '(x) * ψ (s)ψ (s) * dAdx
2 +
+ ∑ ∑n n
k hk h
k=1 k=1(s)
M (x)M (x)1* V (x)V (x) * dxds
2 EI
- ∑ h hq V dx (3.4.12)
Let,
ϕ ϕ∫ij ji i ja = a = (s) (s)dA (a)
ϕ ϕ∫ij ji i jb = b = '(s) '(s)dA (b)
ϕ∫kj jk k jc = c = '(s)ψ (s)dA (c)
ϕ∫ih hi i kc = c = '(s)ψ (s)dA (d)
∫kh hk k hr = r = ψ (s)ψ (s)dA; (e)
∫ k hkh hk
(s)
M (s)M (s)1s = s = ds
E EI (f)
∫h hq = qψ ds (g) (3.4.13)
Substituting eqns. (3.4.13) into eqn. (3.4.12) gives the potential energy functional:
∑ ij i j
1π = E a U'(x)U '(x)dx
2
+ ∑ ∑ij i j kj k j
1G b U (x)U (x)+ c U (x)V '(x)
2
+ ∑ ∑ih i h kh k h
1G c U (x)V '(x)+ r V '(x)V '(x) dx
2
+ ∑ hk k h
1E s V (x)V (x)dx
2 ∑- q V dx
h h (3.4.14)
3.4.3 Governing equations of equilibrium
The governing equations of distortional equilibrium are obtained by minimizing the
above functional eqn.(3.4.14), with respect to its functional variables u(x) and v(x)
using Euler Lagrange technique, eqns. (3.3.19)and (3.4.15).
∂ ∂ ∂ ∂
'
j j
π d π- = 0
U dx U (a)
45
'
∂ ∂
∂ ∂ h h
d- = 0
V dx V
π π (b) (3.4.15)
Carrying out the partial differentiation of eqn.(3.4.14) with respect to Uj and jU ' gives
∂ ∂∑ ∑ '
ij i kj k
j
= G b U (x)+ c V (x)U
π ,
∂
∂∑ '
ij i'
j
=E a U (x)U
π,
∂ ∂
''
ij i'
j
d=Ea U (x)
dx U
π
Therefore, ∂ ∂ ∂ ∂
'
j j
d- = 0
U dx U
π π
⇒ ∑ ∑ ∑' ''
ij i kj k ij iG b U (x)+ c V (x) -E a U (x) = 0
or ∑ ∑ ∑'' '
ij i ij i kj kE a U (x) -G b U (x) -G c V (x) = 0
Dividing through by G, and re-arranging we obtain;
∑ ∑ ∑m m n
k a U''(x)- b U (x)- c V '(x)=0ij i ij i kj k
i=1 i=1 k=1 (3.4.16a)
where υE
κ = = 2(1+ )G
Performing similar operations on eqn.(3.4.14) with respect to hV and hV ' we obtain
the second equation as follows.
∂
∂∑ ∑hk K h
h
=E s V (x) - qV
π
'
∂ ∂∑ ∑ih i kh k
h
= G c U (x)+ r V '(x)V
π
'
∂ ∂ ∑ ∑ih i kh k
h
d= G c U'(x)+ r V '(x)
dx V
π
'
∂ ∂∴
∂ ∂ h h
d- =
V dx V
π π - ∑ ∑ih i kh kG c U'(x)+ r V '(x) + ∑ ∑hk k hE s V (x) - q = 0
∑ ∑ih i kh kc U'(x)+ r V ''(x) - ∑ ∑hk k h
1κ s V (x)+ q = 0
G (3.4.16b)
Equations (3.4.16a) and (3.4.16b) are Vlasov’s differential equations of distortional
equilibrium for a box girder. The evaluation of the coefficients of these equations is
presented in the next chapter.
46
Equations (3.4.16a) and (3.4.16b) can be written in matrix notation. Thus,
''
''
''
11 12 13 1
21 22 23 2
31 32 33 3
a a a U
κ a a a U
a a a U
-
11 12 13 1
21 22 23 2
31 32 33 3
b b b U
b b b U
b b b U
-
'
'
'
'
1
11 12 13 14
2
21 22 23 24
3
31 32 33 34
4
Vc c c c
Vc c c c
Vc c c c
V
=0
(3.4.17a)
'
'
'
11 12 13
1
21 22 23
2
31 32 33
3
41 42 43
c c cU
c c cU
c c cU
c c c
-
111 12 13 14
221 22 23 24
331 32 33 34
441 42 43 44
Vs s s s
Vs s s sκ
Vs s s s
Vs s s s
+
1 ''
''
''
''
11 12 13 14
21 22 23 24 2
31 32 33 34 3
41 42 43 44 4
r r r r V
r r r r V
r r r r V
r r r r V
+
1
2
3
4
1
G= 0
q
q
q
q
(3.4.17b)
Expanding the matrix notation eqns. (3.4.17 (a) and (b)) we obtain:
11 1 12 2 13 3 11 1 12 2 13 3 11 1 12 2 13 3 14 4
'' '' '' ' ' ' 'κa U + κa U + κa U - b U - b U - b U - c V - c V - c V - c V = 0
21 1 22 2 23 3 21 1 22 2 23 3 21 1 22 2 23 3 24 4
'' '' '' ' ' ' 'κa U + κa U + κa U - b U - b U - b U - c V - c V - c V - c V = 0
1 32 2 33 3 31 1 32 2 33 3 31 1 32 2 33 3 34 4
'' '' '' ' ' ' 'κa U + κa U + κa U - b U - b U - b U - c V - c V - c V - c V = 0
31 (3.4.18a)
11 1 12 2 13 3 11 1 12 2 13 3 14 4 11 1 12 2 13 3 14 4
' ' ' '' '' '' ''c U + c U + c U - κs V - κs V - κs V - κs V + r V + r V + r V + r V = -q/G
21 1 22 2 23 3 21 1 22 2 23 3 24 4 21 1 22 2 23 3 24 4
' ' ' '' '' '' ''c U + c U + c U - κs V - κs V - κs V - κs V + r V + r V + r V + r V = -q/G
31 1 32 2 33 3 31 1 32 2 33 3 34 4 31 1 32 2 33 3 34 4
' ' ' '' '' '' ''c U + c U + c U - κs V - κs V - κs V - κs V + r V + r V + r V + r V = -q/G
41 1 42 2 43 3 41 1 42 2 43 3 44 4 41 1 42 2 43 3 44 4
' ' ' '' '' '' ''c U + c U + c U - κs V - κs V - κs V - κs V + r V + r V + r V + r V = -q/G
47
CHAPTER FOUR
GENERATION OF THE STRAIN MODES AND EVALUATION OF THE COEFFICIENTS OF THE GOVERNING EQUATIONS OF EQUILIBRIUM
4.1 PREAMBLE
The governing differential equations of distortional equilibrium are given by eqns.
3.4.16(a) and 3.4.16(b).For ease of reference they are stated here below.
∑ ∑ ∑m m n
ji i ji i jk ki=1 i=1 k=1
k a U''(x) - b U (x) - c V '(x) = 0 4.1.1(a)
∑ ∑ ∑ ∑hi i kh k hk k h
1c U'(x)+ r V ''(x) - k s V (x)+ q = 0
G 4.1.1(b)
where, ji ij ji ij kj jk hi ih hk kh hk kha = a , b = b , c = c , c = c , r = r , s = s are called Vlasov’s coefficients
given by;
ϕ ϕ∫ij ji i ja = a = (s) (s)dA (a)
ϕ ϕ∫ij ji i jb = b = '(s) '(s)dA (b)
ϕ∫kj jk k jc = c = '(s)ψ (s)dA (c)
ϕ∫ih hi i kc = c = '(s)ψ (s)dA (d)
∫kh hk k hr = r = ψ (s)ψ (s)dA; (e)
∫ k hkh hk
(s)
M (s)M (s)1s = s = ds
E EI (f)
Before the solution of these differential equations can be obtained, the following
steps need to be taken.
(i) Appropriate choice of girder cross sections
(ii) Generation of strain modes for the girder sections
(iii) Evaluation of the coefficients of the differential equations for each girder cross
section
(iv) Development of the equations for various strain modes interactions
Steps (i) to (iii) are systematically considered and presented in this chapter while step
(iv) is considered in the subsequent chapter.
48
4.2 CHOICE OF BOX GIRDER CROSS SECTIONS
As stated in the aims and objectives of this study, thin walled bridge girders, both
mono-symmetric and doubly symmetric sections are examined in this report. From
literature survey it is observed that doubly symmetric sections are not common
features in current bridge design and construction, but they are prominently used in
box culverts and similar bridge structures of less importance. Consequently, the
following box girder sections, Figs. 4.2.1 and 4.2.2, are adopted in this work for the
purpose of torsional-distortional analysis. (a) Single cell mono symmetric section,
(b) double cell mono symmetric section, (c) single cell doubly symmetric section and
(d) multi cell doubly symmetric section,
A non symmetric section, Fig.4.2.3, is also chosen to highlight the effect of full
interaction of strain modes on the formulation of governing differential equations and
the analysis of thin walled box girders with such sections.
t
b
t
t
h
x
y
(a) Single Cell doubly symmetric
section
• C & S
Fig. 4.2.2 Doubly-symmetric box girder sections
b b b
h
t
t t t
x
y
C & S •
(b) Multi-Cell doubly symmetric Section
3050
2745 2745 1830 1830
915 915 3660 3660
203
203 203
203
(b) Double cell box girder section
Fig. 4.2.1 Mono-symmetric box girder sections
915 7320
2745 2745 3660
(a) Single cell box girder section
3050
915
t = 203
5490
100
49
4.3 STRAIN MODES AND VLASOV COEFFICIENTS
4.3.1 Introduction
From the energy formulation of the equilibrium equation in section 3.4.2 it was
noted that ϕ and ψ represent generalized warping and distortional strain modes
respectively and from eqns.( 3.4.3 and 3.4.4) ( )ϕi s and kψ (s) are elementary
displacements of the strip frame out of the plane (m displacements) and in the plane (
n displacements) respectively. It was also noted that these displacements are chosen
among all displacements possible and are called the generalized strain coordinates
of a strip frame. Thus, Vlasov’s coefficients of differential equations of equilibrium,
eqn.(3.4.13), which involve a combination of these elementary displacements and
their derivatives may be obtained by consideration of the box girder bridge cross
section as a strip frame and then applying unit displacement one after the other at
the nodal points of the frame in longitudinal direction, to determine the corresponding
out of plane displacement of every joint (regarded as fixed) on the frame. By applying
another set of unit displacements at the joints in n possible transverse directions, the
corresponding transverse (in-plane) displacements can also be obtained. The first
order derivatives of these displacement functions may be obtained by numerical
differentiation and used for computation of the coefficients with the aid of Morh’s
integral for displacement computations (appendix A).
4.3.2 Nodal displacement strain modes
Consider the single cell mono-symmetric section shown in Fig.4.3.1(a). The frame
has four degrees of freedom in the longitudinal direction and four in the transverse
direction. By applying unit displacement at the nodes in each of these directions the
50
longitudinal strain modes ( )ϕi s and transverse strain modes kψ (s) are obtained and
hence their first derivatives, Figs.4.3.1b and 4.3.2.
x
ψ1
x
ψ2
φ1(s) 1
1.0
1.0
x
y
φ2(s)
2
1.0
1.0
x
y
φ4(s)
4
1.0
x x
y
(a) Longitudinal Nodal Strain Modes
Fig.4.3.1 Longitudinal strain modes and their first derivatives
0.444
0.294
x
y
φI1(s)
0.294
0.444
x
y
φI2(s)
0.437
0.444
x
y
φI3(s) 0.444
0.437
x
y
φI4(s)
(b) Derivatives of Nodal Strain Modes
φ3(s)
3
1.0 y
x
51
Expanding the coefficients using eqn 3.4.13 we have;
(a) ϕ ϕ ϕ ϕ∑∫ji ij j i j ia = a = dA = dA
(b) ϕ ϕ ϕ ϕ∑∫ji ij i j i jb = b = ' 'dA = ' 'dA
(c) ϕ ϕ∑∫jk kj j k j kc = c = 'ψ dA = 'ψ dA (4.3.1)
(d) ϕ ϕ∑∫hi ih h i h ic = c = ψ 'dA = ψ 'dA
(e) ∑∫hk kh h k h kr = r = ψ ψ dA = ψ ψ dA
The coefficients aij, bij, ckj, cih, and rkh are determined using Morh’s integral chart.
4.3.3 Generalized strain modes
A consideration of the single cell mono-symmetric strip frame in Fig.4.2.1(a)
shows that it has four degrees of freedom in the longitudinal direction and four in the
transverse direction. From eqns.(3.4.3 and 3.4.4), where in this case m = 4 and n = 4,
it follows that we have sixteen displacement quantities to compute and hence,
sixteen differential equations of distortional equilibrium will be required.
For multi-celled profiles the number of degrees of freedom will increase and
hence the number of independent displacement quantities (m+n) will require 2(m x n)
differential equations to solve for the displacement quantities and this can be quite
cumbersome.
1.032
y 1.032
y
1.0
x
y
ψ3
1.0 y
ψ4
Fig. 4.3.2 Transverse strain modes
x
52
The application of Vlasovs generalized strain modes as modified by Varbanov
(1976) reduces the number of displacement quantities and hence the differential
equations of equilibrium required to solve for them to seven, irrespective of the
number of degrees of freedom possessed by the structure.
In the generalized strain modes, there are three strain fields in the longitudinal
direction, 1 2,ϕ ϕ and 3ϕ .Thus, from eqn. (3.4.3) we have,
ϕ ϕ ϕ1 1 2 2 3 3u(x,s) =U (x) (s)+U (x) (s)+U (x) (s)
or ϕ∑3
i ii=1
u(x,s) = U (x) (s) (4.3.2a)
In the transverse direction four strain modes are also recognized; 1 2 3ψ ,ψ ,ψ and 4ψ .
From eqn (3.4.4) we have;
1 1 2 2 3 3 4 4v(x,s) = V (x)ψ (s)+ V (x)ψ (s)+ V (x)ψ (s)+ V (x)ψ (s)
or ∑4
k kk=1
v(x,s) = V (x) ψ (s) (4.3.2b)
where 1ϕ = out of plane displacement parameter when the load is acting (vertically)
normal to the top flange of the girder, i.e., bending is about horizontal
axis.
2ϕ = Out of plane displacement parameter when the load is acting tangential
to the plane of the flanges, i.e., bending is about vertical axis.
3ϕ = Out of plane displacement parameter due to distortion of the cross
section, i.e., the warping function .
1ψ = In-plane displacement parameter due to the load giving rise to 1ϕ
2ψ = In-plane displacement parameter due to the load giving rise to 2ϕ
3ψ = In-plane displacement parameter due to the distortion of the cross
section, i.e., non uniform torsion.
4ψ = In-plane displacement function due to pure rotation or Saint Venant
torsion of the cross section.
4.3.4 Plotting of warping functions
In the preceding section, it was stated that 3ϕ , the out of plane displacement
parameter due to distortion of the cross section is obtained by plotting the warping
53
function of the cross section. The summary of the procedure for plotting the warping
function of any given cross section is presented below, followed by numerical
example calculations.
(i) Chose a coordinate axis (preferably along the axis of symmetry) and plot (a)
variation of x-ordinates along x-axis and (b) variation of y-ordinates along y-axis.
These give x diagram and y diagram as shown in Fig.4.3.3 (b) and (c) respectively
(ii) Choose a pole B and a starting point V. Evaluate and draw the warping function
with respect to pole B using eqns (2.6.24 and 2.6.29)
(iii) Evaluate the following section properties;
(a) F = Cross sectional area of profile
(b) Fx = first moment of area of section about y-axis
(c) Fy = first moment of area of section about x-axis
(d) ωF = First moment of sectorial area about pole B
(e) xxF = Second moment of area of profile about y-axis
(f) yyF = Second moment of area of profile about x-axis
(g) xyF = Product moment of area of profile, i.e., Fx.Fy.
(h) B
xωF = Sectorial product of area, Fx.FwB
(i) ByωF = Sectorial product of area, Fy.FwB
(j) B Bω ωF =Warping constant with respect to pole B
(iv) Locate the centroid s of the profile using the relations
__x
s
Fx =
F and
__y
s
Fy =
F (4.3.3)
The warping function with respect to the centroidal axis becomes
C B 0ω =ω -ω , where, 0 ωBω =F /F (4.3.4)
(v) Using the new (centroidal) axis plot x and y diagrams as in i(a) and (b) above.
(vi) Calculate the product integrals, Fxx, Fyy, Fxy, Fwx, Fwy, Fww, (based on the
centroidal axis), using the following transformation formulae:
54
__ _
__ _
__ _ _
__ _ _
__ _ _
__ _
2
xxxx x
2
yyyy y
xyxy x y
xωxω x ω
yωyω y ω
2
ωωωω ω
F =F -F /F
F =F -F /F
F =F -F F /F
F =F -F F /F
F =F -F F /F
F =F -F /F
(4.3.5)
(vii) Obtain the coordinates of the shear centre M from the relation
yω xx xω xy
m B 2
xx yy xy
F .F -F .Fx - x =
F .F -F (4.3.6)
and yω xy xω yy
M B 2
xx yy xy
F .F -F .Fy - y =
F .F -F (4.3.7)
(viii) The warping function Mω , with pole at shear centre M, is obtained from the
following transformation formulae (Murray 1984);
M C M B M Bω (s) =ω +(y - y )x - (x - x )y (4.3.8)
4.3.4.1 Warping functions for mono-symmetric box girder sections
(a) Single cell mono-symmetric box girder section
The warping functions ( Bω ) are given as follows:
∫B Bω (s) = ρ (s)ds for open sectional profile
∫B B
ψω (s) = ρ - ds
t for closed sectional profile
x
6.774
2.195
6.774
6.774
6.774
_
_ _
+
+
+
1
2
3
5 6
ωB
1
+ +
2
3 4
5
6 x
915 7320
2745 2745 3660
(a) Single cell box girder section
3050
915
t = 203
y
(b) x Diagram
y
+
1830
3660
4575
+
+ _
_
_ 4575
3660
1830
B,V 1 2
3
4
5 6 x
55
The computations of the warping functions are shown in Appendix One.
4.3.4.2 Warping function for multi cell doubly symmetric section
h=b
t t t
t
x C & S •
A3 A2 A1
Fig.4.3.3 : Single cell mono symmetric section and warping function
Fig.4.3.4 Double cell mono-symmetric section and warping function
(a) Double cell box girder
3050
2745 2745 1830 1830
915 915 3660
203
203
203
1889.4
1160.58
3660
y
x
(b) x Diagram
y
+
1830
3660
4575
+
+
_
_
_ 4575
3660
1830
B,V 1 2
3
4
5 6 x
(c) y Diagram
3050 3050 +
1
+ +
2
3 4
5
6 x
+
3050
y
B, V x
6.774
y 2.195
2.195
6.774
6.774
6.774
_
_ _
_
+
+
+
1
2
3
4
5 6
(d) Warping function ωB
ωB
56
For the doubly symmetric section in Fig.4.3.5 (a) the centroid C and the shear centre
S, coincide. Cross sectional area F = 10bt
Enclosed area A1 = A2 = A3 = b2
Fx = 0, symmetry about x-x axis
Fy = 0 , symmetry about y-y axis
∫ i
M M
ψω (s) = ρ - ds
t
For closed sections we have
∫ ∫ ∫B D
i i-1 i i+1
A C
ds ds ds2A = -ψ +ψ -ψ
t t t
For Cell 1 we obtain; 1 24ψ -ψ = 2bt (1)
Cell 2 ⇒ 1 2 3-ψ + 4ψ -ψ = 2bt (2)
Cell 3 ⇒ 2 3ψ + 4ψ = 2bt (3)
Solving eqns.(1, 2, & 3) above gives;
1 3
2
ψ =ψ = 5bt/7
ψ = 6bt/7
Using ( )∫M M iω = ρ -ψ /t ds we obtain,
+ +
+
+ +
+
_ _ _
_ _
_
2 3 8
4 7
tb
+
2 3 8
4 7
tb
+
2 3 8
4 7
tb
+
2 3 8
4 7
tb
+
2 1 3
4 7
tb
+
2 1 3
4 7
tb
+
(b) Warping function diagram
Fig.4.3.5 Doubly symmetric section and warping function diagram
x
y
57
For cell 1 and cell 3, ( )∫M Mω = ρ - 5bt/7 ds
Cell 2; ( )∫M Mω = ρ - 6bt/7 ds
Integrating from point V to point 1 we obtain;
M
b b 6btρ = - , ds = , ψ =
2 2 7
∴
2
1
1 3tω = -b +
4 7
From point 1 to point 2 we have;
M
b 5btρ = - , ds = b, ψ =
2 7
∴
2
2
3 8tω = -b +
4 7
From point 0 to point 4 we have;
M
b b 6btρ = ,ds = ,ψ = -
2 2 7
∴
2
4
1 3tω = b +
4 7
From point 4 to point 3 we have;
M
b 5btρ = , ds = b, ψ = -
2 7 ∴
2
3
3 8tω = b +
4 7
By reason of symmetry we observe that;
2
5 1
1 3tω = -ω = b +
4 7
2
6 2
3 8tω = -ω = b +
4 7
2
7 3
3 8tω = -ω = -b +
4 7
2
8 4
1 3tω = -ω = -b +
4 7.
The diagram of the warping function with the pole located at shear centre S is on
Fig.4.3.5 (b).
4.3.4.3 Warping function for non-symmetric section
3050
1.475
8.245
8.245
6.88
0 1
2 3
. c 1348
1702
x
200
5490
200
200
x
58
The warping function of the single cell non-symmetric profile (Fig.4.3.6a), based on
the centroidal axes is shown on Fig. 4.3.6b while that with respect to the shear centre
is on Fig. 4.3.13f
The computations of the warping functions are shown in Appendix One.
4.3.5 Strain mode diagrams
Consider a simply supported girder loaded as shown in Fig 4.3.7(a). If we assume
the normal beam theory, i.e., neutral axis remaining neutral before and after bending,
then the distortion of the cross section will be as shown in Fig.4,3.7(b) where, θ is
the distortion angle (rotation of the vertical axis). The displacement φ1 at any distance
R, from the centroid is given by 1ϕ = Rθ . If we assume a unit rotation of the vertical
(z) axis then 1ϕ = R , at any point on the cross section. Note that 1ϕ can be positive
or negative depending on the value of R, in the tension or compression zone of the
girder. Thus, 1ϕ is a property of the cross section obtained by plotting the
displacement of the members of the cross section when the vertical (z-z) axis is
rotated through a unit radian.
Similarly, if the load is acting in a horizontal (y- y) direction, normal to the x-z plane
in Fig.4.3.7(a), then the bending is in x-z plane and y axis is rotated through angle θ2
giving rise to 2ϕ , displacement out of plane. The values of 2ϕ are obtained for the
members of the cross section by plotting the displacement of the cross section when
y-axis is rotated through a unit radian.
The warping function 3ϕ ,of the beam cross section is obtained as detailed in
section 4.3.4. It has been explained that the warping function is the out of plane
displacement of the cross section when the beam is twisted about its axis through the
pole, one radian per unit length without bending in either x or y direction and without
longitudinal extension.
59
ψ1 and ψ2 are in-plane displacement of the cross section in x-z and x-y planes
respectively while ψ3 is the distortion of the cross section. They can be obtained by
numerical differentiation of 1ϕ , 2ϕ and 3ϕ diagrams respectively.
ψ4 is the displacement diagram of the beam cross section when the section is
rotated one radian in say, a clockwise direction, about its centroidal axis. Thus, ψ4 is
directly proportional to the perpendicular distance ( radius of rotation) from the
centroidal axis to the members of the cross section. It is assumed to be positive if the
member moves in the positive directions of the coordinate axis and negative
otherwise.
The generalized strain modes for (a) single cell mono-symmetric section and (b)
double cell mono-symmetric section are shown in Figs 4.3.8 and 4.3.9 respectively.
1
+
2
3
5
6 + 0.857
(d) Transverse strain mode in z-
direction
+ 0.857
+ y
z
4
'1 1
ϕ =ψψψψ
z
1.945 +
1
+ +
2
3 4
5 6
1.945
-1.105
1.945
1.105
-1.105
(c) Longitudinal Strain Mode Diagram
(Bending about o-y axis)
. y ϕ1111
1
+
2
5 6
4.575
3.66 3.66
+
- -
4.575
ϕ2222 1
+
0.514
2 5 6
0.514
1.0 _
-
'2 2
ϕ =ψψψψ
x
y
(a) Simply supported girder section
z +
-
θ
(b) Cross section distortion
y
Fig. 4.3.7 Simply Supported Girder and Cross Section Distortion
915 7320
2745 2745 3660
(a): Single cell box girder section
3050
915
t = 203
_
(b) Pure Rotation Diagram
. +
+
_
2.57
0
1.105
1ρ
2ρ
3ρ
4ρ
1.945 2.57
0
4
ψψψψ
y
60
(c)Longitudinal Strain Mode Diagram
(Bending About y-y Axis)
1.889 1.889
1.16 1.16 1.16
c
+
+ + +
- - -
-
1 2
3 4
5 6
y
z
c 0.857
0.857
+0.857
(d) Transverse stain mode in y-y direction
+ +
+
+0.857 +1.00
y
y
'1 1
ϕ =ψψψψ
Fig.4.3.8 Generalized strain modes for single cell mono-symmetric box girder frame
+
1 +
2
3 4
5
6
-0.159
(h) Distortion Diagram '
3 3( ) ( )s sϕ ψ=
+ -
-
-
y
z
-0.492
+1.162
-0.417
0.123
+1.162
+0.417
'3 3
ϕ =ψψψψ 0.966
+
1
2
4
5 6
+
z
+
-
- -
-
+ + + -
y
3
-0.966 -0.582
+0.582
+0.582
+0.901 -0.901
(g) Warping Function Diagram
+0.901
3Mω ϕ=
_ 1.00
+4.575 -3.66 +
- +3.66
3050
2745 2745 1830 1830
915 915 3660
203
203
203
(a) Double cell concrete box girder
1889.4
1160.58
3660
y
z
. +
+
_
_
1.0
0.514 0.514
0.514
1ρ
2ρ
3ρ
4ρ
(b) Pure rotation Diagram
4ψψψψ
y
61
Others are Fig.4.3.10 for single cell doubly symmetric section, Fig. 4.3.11 for multi-
cell doubly symmetric section and Fig.4.3.12 for single cell non- symmetric section.
Fig. 4.3.9: Generalized strain modes for double cell mono-symmetric box girder section
0.901 +
-
1 2
3
4
5 6
(g) Warping function, 3( )M
sω ω= diagram
0.901
+
+ +
-
-
-
-
0.901
0.966 0.582
0.582
0.582
0.966
3ϕ 0.417
0.492
+
+
1 2
4
5 6
(h) Distortion diagram
_
_ 0.417
0.159
3
'3 3
ϕ =ψψψψ
z
y
3.60
(b) Rotation diagram
⊕
⊕
-
-
4ψ
z
3m
7.2m
y
(a) Single cell doubly-symmetric section
200mm
200mm
3.60 1.50
1.50
y ⊕ ⊕
ϕ1 1
' =ψ ϕ1(s) y
- -
- 1.5
62
3.0m 3.0m 3.0m
y
3.0m 200
200
200
200
z
• 0,M
(a) Multi cell doubly symmetric section
1.50
4.50 4.50
(b) Pure rotation diagram
1.50
1.50
+ +
+ + -
- -
-
4ψ
y
1.0 1.0 1.0 1.0
+ +
+
y
1.50 - -
- - -
1.50 1.50 1.50 1.50
ϕ '1 1=ψ ϕ
1
(f) Transverse strain mode
in z-direction
z
y
⊕
⊕
⊕
(e) Longitudinal strain mode diagram
( bending about z-z axis)
2ϕ
y
z
ϕ2 2
' =ψ
1.0
1.0 3.60
3.60
-
-
-
3.60
3.60
y
⊕
z
− ⊕
ϕ3 M=ω
(g) Warping Function (h) Distortion diagram
z
y
Fig.4.3.10 Generalized strain modes for single cell doubly symmetric section
⊕
⊕
−
2.223
2.223 2.223
2.223
1.482
1.482
0.618
0.618
63
0 1
2 3
. C y
1.0 0.857
ϕ '1 1=ψ
0 1
2 3
. C y
1.348 1.348
+
+
-
-
-
1ϕ
(b) Rotation
0 1
2 3
. C y
z
+
+
_
_
1.944
1.348 2.414
1.702
4ψ
0 1
2 3
. C
2414 1246
1348
1702
(a) Cross section dimensions
5490
t = 100mm
(e) Longitudinal strain mode Diagram.
Bending about z-z axis
1 2 3 4
5 6 7 8
+
+
+ +
1.50 4.50 4.50
4.50 4.50
-
-
-
-
(f) Transverse strain mode in y-y direction
6 5 7
1 2 3 4
8
-
-
1.0
1.0
ϕ '2 2=ψ
ϕ2
8.808
8.808
_ _ _ +
+ +
+ + +
_ _
_
8.808 8.808
8.808 3.021 3.021
3.021 3.021
(g) Warping function 3 3ω ϕ=
Fig.4.3.11: Generalized strain modes for multi-cell doubly symmetric section
1.929
2.014 5.871
5.871 1.929
(h) Distortion '
3 3ϕ ψ= Diagram
2.014
ϕ '3 3=ψ
ϕ3
64
4.3.6 Tests for accuracy of strain modes
We define a function 0ϕ which is orthogonal to the strain modes 1 2,ϕ ϕ and 3ϕ
such that ϕ∫2
o dt = area of cross section. Since 0ϕ is orthogonal to 1 2,ϕ ϕ and 3ϕ it
follows that; ϕ ϕ∫ 0 idt = 0 , (for i =1,2,3) (4.3.10)
Equation (4.3.10) can therefore be used to verify the accuracy of the strain mode
diagrams. Additionally, the strain mode 1ϕ can be obtained from the relation,
1 1 0ϕ ϕ αϕ= + (4.3.11)
where,
1ϕ = auxiliary shape function
α = a factor
Substituting eqn (4.3.11) into eqn (4.3.10) gives, after simplification;
0 1
2 3
. C
0.608
(g) Warping function
0.421 0.430
0.633
3Mω ϕ= 028 y
z z
y
Fig.4.3.12 Generalized strain modes for single cell non symmetric section
0 1
2 3
. C y
z +
+
_
_
0.339
0.289
0.155
0.349
(h) Distortion Diagram
'3 3
ϕ ψ=
0 1
2 3
. C y
z
(e) Longitudinal strain mode:
bending in z-z direction
2ϕ
1.246
3.076
3.076 2.414
2.414
+
+
+
_
_
_
y
0 1
2 3
. C
z
(f) Transverse strain mode in z-z
direction
ϕ '2 2=ψ
1.00
0.514
1.0
_
_
+
65
ϕ ϕ
ϕ∫∫
0 1
2
0
dtα = -
dt (4.3.12)
With values of α obtained from eqn (4.3.12), 1ϕ can be computed from eqn (4.3.11)
and compared with 1ϕ diagram obtained using Morh’s integral.
The single cell mono-symmetric section is used to demonstrate the accuracy of
the strain mode diagrams. The cross sectional dimensions and the generalised
strain modes ( 1ϕ , 2ϕ and 3ϕ ) diagrams for this profile are shown in Fig.4.3.8. The
auxiliary shape function 1ϕ diagram, Fig.4.3.14(a), is obtained by arbitrary choice of
coordinate axis (different from the centroidal axis) and by drawing the displacement
diagram of the cross section when rotated through a unit radian about the x-x axis.
The 0ϕ diagram, Fig.4.3.14(b), is obtained by giving a unit displacement to each
member of the cross section in a positive direction.
Using Morh’s integral chart (appendix A), we have
ϕ∫2
0 dt =1.992 , 0 1ϕ ϕ∫ dt =1.839
From eqn(4.3.10) α = -0.923
From eqn (4.3.11) 1ϕ = 1ϕ -0.923 0ϕ (4.3.13)
A plot of eqn (4.3.13) gives exactly Fig.4.3.8 (b) thus, confirming that 1ϕ diagram is
correct.
From eqn. (4.3.10) we obtain
0 1ϕ ϕ∫ dt = (1.0)(-0.332)(9.15)(0.097)+½(1.0)(2.715-0.332)(3.557)(0.016)x2 +
+ (1.0)(2.718)(3.66)(0.016) = 0
By inspection it can also be seen that
1
+
2
3 4
5 6
+
+
+
1.0
1.0
1.0
1.0
(b) Orthogonal shape function 0ϕ
(Unit displacement diagram)
Fig.4.3.13 Auxiliary shape function and orthogonal shape function
1
+
2
3 4
5 6
+
+
3.05 3.05
(a) Auxiliary shape function 1ϕ
3.05
z
y
66
ϕ ϕ∫ 0 2dt = 0 and ϕ ϕ∫ 0 3dt = 0 are satisfied. Thus the accuracy of the strain mode
diagrams is confirmed.
4.4 COMPUTATION OF VLASOV’S COEFFICIENTS
As explained earlier, the coefficients ij ij kj iha , b , c , c and khr , of the differential
equations of equilibrium are computed with the aid of Morh’s integral chart.(Appendix
Three). Expanding eqn (3.4.13), the coefficients are obtained as shown in Table 4.1
The detailed calculations are shown in Appendix Two
Table 4.1a: Summary of Vlasov’s aij coefficients for various sections
Coefficient Elements
Mono-symmetric sections Doubly-symmetric Sections
Non-Symmetric Section
(a) Single Cell
(b) Double Cell
(c) Single Cell
(d) Multi- Cell
(e) Single Cell
11a 6.453
7.023
10.930
9.900
5.264
12 21a a=
0
0
0
0
1.536
13 31a a=
0
0
0
0
-0.001
22a 25.05 25.073 27.994 27.000 11.018
23 32a a=
-0.270
0.425
0
0
-0.038
33a 0.757 0.750 6.721 127.752
0.293
Table 4.1b: Summary of Vlasov’s bij coefficients for various sections
Coefficient Elements
Mono-symmetric Sections Doubly-symmetric Sections
Non Symmetric Section
(a) Single Cell
(b) Double Cell
(c) Single Cell
(d) Multi-Cell
(e) Single Cell
11b 1.060 1.680 1.200 2.400 1.132
12 21b b=
0 0 0 0 0.314
0 0 0 0 -0.037
67
Table 4.1c: Summary of Vlasov’s Ckj coefficients for various
sections
Table 4.1d: Summary of Vlasov’s Cih coefficients for various
sections
13 31b b=
22b
2.982 2.982 2.880 3.600 2.018
23 32b b=
-1.066 -0.449 0 0 0.028
33b 1.407 1.533 3.736 55.727 0.244
Coefficient Elements
Mono-symmetric Sections
Doubly-symmetric Sections
Non Symmetric Section
(a) Single Cell
(b) Double Cell
(c) Single Cell
(d) Multi-Cell
(e) Single Cell
11c 1.060
1.680
1.200
2.400
1.132
12 21c c=
0 0 0 0 0.314
13 31c c=
0 0 0 0 -0.037
22c 2.982 2.982 2.880 3.600 2.018
23 32c c=
-1.066 -0.449 0 0 0.028
33c 1.407 1.533 3.736 55.727 0.244
41c 0 0 0 0 0.148
42c -1.561 1.112 0 0 0.476
43c 1.265 1.295 3.732 24.759 0.262
Coefficient Mono-symmetric Doubly- Non Symmetric
68
Elements Sections symmetric Sections
Section
(a) Single Cell
(b) Double Cell
(c) Single Cell
(d) Multi-cell
(e) Single Cell
11c
1.060
1.680
1.200 2.400 1.132
21 12c c=
0 0 0 0 0.314
31 13c c=
0 0 0 0 -0.037
41 14c c=
0 0 0 0.148
22c
2.982 2.982 2.880 3.600 2.018
32 23c c=
-1.066 -0.449 0 0 0.028
33c
1.407 1.533 3.736 55.727 0.244
24c
-1.561 3.4041.112 0 0 0.476
34c
1.265 1.295 3.732 24.759 0.262
Coefficient Elements
Mono-symmetric Sections Doubly-symmetric Sections
Non Symmetric Section
(a) Single Cell
(b) Double Cell
(d) Single Cell
(e) Multi-Cell
(f) Single Cell
11r
1.060
1.680
1.200
2.400 1.132
12 21r r= 0 0 0 0 0.314
13 31r r= 0 0 0 0 -0.037
14 41r r= 0 0 0 0 0.148
Table 4.1e: Summary of Vlasov’s rkh and hks coefficients for various
sections
69
4.4.1 The hks Coefficients
The coefficients hks given by eqn.(3.4.13(f)) depend on the bending deformation
of the strip characterized by kM (for k = 1, 2, 3, 4). To compute the coefficients we
need to construct the diagram of the bending moments due to the strain modes,
1 2 3 4ψ , ψ , ψ and ψ ,. Incidentally, 1ψ , 2ψ and 4ψ strain modes do not generate
distortional bending moments as they involve pure bending and pure rotation. Only
3ψ strain mode generates distortional bending moment which can be evaluated
using the distortion diagram for the relevant cross section. Consequently the relevant
expression for the coefficient becomes
∫ 3 3hk kh
ss
M (s)M (s)1s = s =
E EI, (4.4.1)
where 3M (s) is the distortional bending moment of the relevant cross section.
22r 2.982 2.982 2.880 3.600 2.018
23 32r r=
-1.066 -0.449 0 0 0.028
24 42r r=
-1.561 1.112 0 0 0.476
33r
1.407 1.538 3.736 57.727 0.244
34 43r r=
1.265 1.295 3.732 24.759 0.262
44r
14.616 14.485 22.032 35.100 10.358
=Shk
0.261x Is 0.723x Is 1.593x Is
65.065x Is
-
70
The hks coefficients for mono symmetric and doubly symmetric sections are shown in
Table 4.1e.
4.5 DISTORTIONAL BENDING MOMENT
The procedure for evaluation of distortional bending moment is not straight
forward, but attempt is made here to outline the procedure, backed up with numerical
calculations.
Step 1: Obtain the distortion diagram from numerical differentiation of the warping
function and indicate the distortional movement directions of the members of the
cross section with arrows, Fig.4.5.1 (a)
Step 2: Draw the base system for the frame (Fig.4.5.1 (b)) assuming all the joints are
fixed and use vectors, R1, R2,…Rn , to represent the magnitudes and directions of the
forces required to give unit translations to the joints in the direction of the forces. The
directions of the forces are guided by the directional movements of the members of
the cross section at distortion.
Step 3: Apply unit rotation to each of the joints, one after the other, and designate
the forces required to accomplish this by X1, X2, …Xn, Fig.4.5.1(b). Obtain the
bending moment diagram for the frame and by considering the equilibrium of each
joint determine the un-balanced moment at each joint. These are then considered as
coefficients ijr in the equilibrium equation.
Note that _
iM denotes the bending moment diagram due to unit rotation of joint i. A
system with n number of joints will have n bending moments and ( n x n) coefficients.
Step 4: Apply unit transverse displacement in the directions of R1 ,R2…Rn, one at a
time and draw the bending moment diagrams generated on the frame.
Step 5: Establish the equilibrium equations based on equality of internal work ij ijr X ,
and external work i iR∆ , eqn.(4.5.3).
Step 6: For each bending moment diagram due to unit joint translation determine the
un-balanced reaction ijR (coefficient) using static equilibrium of the joints.
Step 7: Multiply the coefficients ijR with the actual values of the distortion i
∆ ,
obtained from the distortion diagram to obtain values of ij iR ∆ and sum up to obtain,
for each joint the value of ∑ ij iR ∆ , Table 4.2.
71
Step 8: Substitute the values of ijr and∑ ij iR ∆ into the equilibrium equation and solve
for the unknown reactions, 1 2 nX , X , ...X .
Step 9: The final bending moment M due to distortion of the cross section is obtained
from the relation;
∑ ∑_ _
i Rii iM(s) = M X + M ∆ where i 3i∆ =ψ values. (4.5.1)
4.5.1 Evaluation of distortional bending moment for single cell mono-symmetric section
For uniform plate thickness top-flange bot-flange webI = I = I = I
Top Flange; 2
4EI/L = 0.546EI
2EI/L = 0.273EI
6EI/L = 0.112EI
Bottom Flange; 2
4EI/L =1.093EI
2EI/L = 0.546EI
6EI/L = 0.448EI
Web 2
4EI/L =1.125EI
2EI/L = 0.562EI
6EI/L = 0.474EI
+
1 2
3 4
5 6
-0.159
(a) Distortion Diagram 3ψ (s)
+ -
-
y
z
-0.492
-0.417
0.123
+0.417
915 915 7320
(b) Base System
3050
1
2
3
4
5 6
3660
X2
X1
X3
X4
R4
R3
R1
R2
Fig. 4.5.1 Distortion diagram and base system for evaluation of hks
2
4
1
3
0.582
0.546
_
1M
1.125
X 1
21r
31r
41r
11r
0.546EI
0.273EI
1.125EI
0.562EI
72
32r
42r
12r
0.273EI
0.546EI
1.125EI
0.562EI
22r 1 2
4
3
0.546EI
+0.273EI
1.125EI
0.562EI
_
2M
X 2
(b) Bending moment due to unit rotation of joint 2 and joints equilibrium diagram
1 2
4
3
0.562EI
_
3M
X 3
1.125EI
1.093EI
0.546EI
23r
0.56EI
33r
43r
13r
1.125EI
0.546EI
1.093EI
(c) Bending moment due to unit rotation of joint 3 and joints equilibrium diagram
24r
34r
44r
14r
0.546EI
1.093EI
0.562EI
1.125EI
1 2
4
3
0.562EI
_
4M
X 4
1.125EI
0.546EI
1.093EI
(d) Bending moment due to unit rotation of joint 4 and joints equilibrium diagram
Fig.4.5.2 Bending moments due to unit rotation of joints and joints equilibrium diagrams for single cell mono-symmetric section
1 2
3 4
R 1
∆1=1 ∆’
∆’
β
Unit Displacement of Joint 1
R 21
R 31 R 41
R 11
0.406EI 0.406EI
Joint Equilibrium
1 2
3 4
0.406EI
0.406EI 0.406EI
Moment Diagram
1M
R
73
(b) Unit Displacement of Joint 1 in R2 Direction, Moment Diagram and Joint Equilibrium
1 2
3 4
∆2=1
R2
Unit Displacement of Joint 1
in R 2 Direction
R 22
R 32
R 42
R 12
0.384EI
0.096EI
Joint Equilibrium
4
1 2
3
0.096EI
R 22
R12
R42
R32
0.096EI
0.384EI 0.384EI
Moment Diagram
4
2M
R
(c) Unit Displacement of Joint 3 in R3 Direction, Moment Diagram and Joint Equilibrium
R 23
R 33 R 43
R 13
0.406EI 0.406EI
Joint Equilibrium
1 2
3
4
∆3=1
R3
Unit Displacement of Joint 3
in R 3 Direction
R 23 1 2
3 4
0.406EI
0.406EI 0.406EI R 13
R 43 R 33
Moment Diagram
3M
R
(d) Unit Displacement of Joint 3 in R4 Direction, Moment Diagram and Joint Equilibrium
1 2
3 4
∆4=1
Unit Displacement of Joint 3
R4
R 24
R 34
R 44
R 14 0.096EI
0.384EI Joint Equilibrium
Moment Diagram
R 24 1 2
3
0.096EI
R14
R44
R34
0.096EI
0.384EI
0.384EI 4
4M
R
74
The bending moment diagrams due to unit joint rotations and the static joints equilibrium are shown in Fig.4.5.2. By considering the joint equilibrium diagrams for
_ _ _
1 2 3M , M , M and _
4M we obtain the following coefficients as reactions.
r11 = 1.671EI; r21 = 0.273EI; r31 = 0; r41 = 0.562EI r12 = 0.273EI; r22 = 1.671EI; r32 = 0.562EI; r42 = 0 r13 = 0; r23 = 0.562EI; r33 = 2.218EI; r43 = 0.546EI r14 = 0.562EI; r24 = 0; r34 = 0.546EI; r44 = 2.218EI From the joint equilibrium diagrams in Fig.4.5.3 we obtain the following coefficients. R11 = 0.406EI R21 = 0.406EI; R31 = 0.406EI; R41 = 0.406EI
R12 = - 0.096EI R22 = 0.096EI; R32 = -0.384EI; R42 = -0.384EI
R13 = 0.406EI R23 = 0.406EI; R33 = 0.406EI; R43 = 0.406EI
R14 = -0.096EI R24 = -0.096EI; R34 = -0.384EI; R44 = 0.384EI
Since the moment diagrams in Fig. 4.5.3 are based on unit translation of the joints in
the direction of the members’ displacements we now multiply these unit diagrams by
the actual displacements (distortion) of the members of the cross section obtained
from the distortion diagram, i.e.
Member 1 – 2 displaced ∆1 = 0.159m, left
Member 2 – 3 displaced ∆2 = 0.417m, down
Member 3 – 4 displaced ∆3 = 0.492m, right
Member 4 – 1 displaced ∆4 = 0.417m, up
Table 4.2 shows the summation of ij i
R ∆ for the single cell mono-symmetric section.
The equations of equilibrium is given by eqn (4.5.3) Equilibrium equations
11 1 12 2 13 3 14 4 1 1r X +r X +r X +r X +R ∆ = 0
21 1 22 2 23 3 24 4 2 2r X +r X +r X +r X +R ∆ = 0
31 1 32 2 33 3 34 4 3 3r X +r X +r X +r X +R ∆ = 0
41 1 42 2 43 3 44 4 4 4r X +r X +r X +r X +R ∆ = 0 (4.5.2)
75
Table 4.2 Summation of work done on the joints due to distortion of single cell mono symmetric section
Substituting the coefficients ijr and ∑ ij 1R ∆ into the equilibrium eqn.(4.5.2) and
solving gives; 1 2X = X = -0.1068 and 3 4X = X = 0.042 .
The final bending moment Fig.4.5.4(a) is obtained by summation of the products
of 1 2 3 4M , M , M and M diagrams and 1 2 3X , X , X & 4X respectively plus the summation of
the products of R1 R2 R3M , M , M & R4M diagrams and 1 2 3∆ , ∆ , ∆ & 4∆ respectively as per
eqn. (4.5.2) The distortional bending moment diagrams for single cell and double cell
mono-symmetric sections are shown in Fig.4.5.4 (a) and (b) respectively. From eqn.
(4.5.1) we obtain the hks coefficients for the mono-symmetric sections using Morh’s
integral chart (Appendix Three).
Joints
Work done ij iR ∆
1 2 3 4 R1j ∆1 R2j ∆2 R2j ∆3 R4j ∆4
1 0.406EI 0.406EI 0.406EI 0.406EI 2 -0.096EI -0.096EI -0.384EI -0.384EI 3 0.406EI 0.406EI 0.406EI 0.406EI 4 -0.096EI -0.096EI -0.384EI -0.384EI
∑ ij iR ∆ = 0.184EI 0.184EI -0.056EI -0.056EI
(a) Single Cell Mono-symmetric Section
1 2
3
0.167EI
0.251EI
0.251EI 0.251EI
0.167EI 0.167EI
4
3333M(S)M(S)M(S)M(S)
Fig.4.5.4 Bending moments due to distortion of mono-symmetric Sections
(b) Double Cell Mono-symmetric Section
0.324
-0.243
+
+
- - - 1 2 3
4
5 6
-
+
+ +
+ +
-
-
-
+
-0.560
0.280
0.280
0.324
0.229
0.229 -0.458 0.243
3333M(S)M(S)M(S)M(S)
76
4.5.2 Evaluation of distortional bending moments for doubly symmetric
Sections
The distortion diagram and base system for evaluation of distortional bending
moment for multi-cell doubly symmetric section is shown in Fig.4.5.5. To determine
the distortional bending moment, we follow the same procedure outlined in section
4.5. For uniform plate thickness top-flange bot-flange webI = I = I = I
Flanges; ⇒4EI/h =1.333EI 2EI/h = 0.667EI , 26EI/h = 0.667EI
Webs ; ⇒4EI/b =1.333EI 2EI/b = 0.667EI , 26EI/b = 0.667EI
The bending moment diagrams due to unit joint rotations are shown in Fig.4.5.6.
By considering the joints equilibrium diagrams for 1 2 3M , M , M and 4M we obtain the
values of the coefficients as reactions. Thus,
From M1 diagram; 3111 21 81 41 51 61 71r = 2.667EI, r = 0.667EI, r = 0.667, r = r = r = r = r = 0
From M2 diagram; 12 22 32 72r = 0.667EI, r = 3.999EI, r = 0.667EI, r = 0.667, others are zero
From M3 diagram, 23 33 43 63r = 0.667EI, r = 3.99EI, r = 0.667EI, r = 0.667EI , others are
zero
From M4 diagram; 34 44 54r = 0.667EI, r = 2.667EI, r = 0.667EI,others are zero
From M5 diagram; 45 55 65r = 0.667EI, r = 2.667EI, r = 0.667EI, others are zero
Fig.4.5.5 Distortion diagram and base system for evaluation of hkS coefficient
(b) Base system
1.929
2.014 5.871
5.871 1.929
(a) Distortion Diagram
2.014 o
o o
o o
R 1 R 2 R3
R4
R5 R6
X1 X2 X3 X4
X5 X6 X7 X8
77
From M6 diagram; 36 56 66 76r = 0.667EI, r = 0.667EI, r = 3.999EI, r = 0.667EI, others are
zero.
From M7 diagram; 27 67 77 87r = 0.667EI, r = 0.667EI, r = 3.999EI, r = 0.667EI, others are
zero.
From M8 diagram; 18 78 88r = 0.667EI, r = 0.667EI, r = 2.667EI,others are zero.
(h) Bending moment due to unit rotation
7X
z
2 3 4
5 6 7 8
8888MMMM
z
1 2 3 4
5 6 7 8
1
0.667
0.667
0.667
7777MMMM
8X
1.333 1.333
0.667
0.667
z
6666MMMM
z
5555MMMM
0.667
0.667 0.667
0.667
0.667
1.333 1.333 5 5 6 6 7 7 8 8
(e) Bending moment due to unit rotation
of joint 5 (f) Bending moment due to unit rotation
of joint 6
y
z y
z 5X
6X
z z
1M 2222MMMM
(a) Bending moment due to unit rotation
of joint 1 (b) Bending moment due to unit rotation
of joint 2
1.333 0.667
0.667
1.333 0.667
1.333 0.667
0.667
1 2
y
z
y
z
1X 2X
z
3333MMMM
z
4444MMMM
(d) Bending moment due to unit rotation
of joint 4 (c) Bending moment due to unit rotation
of joint 3
1.333 1.333
0.667
0.667
0.667
0.667
0.667 3 4
y
z
y
z
4X 3X
78
1 2 3 4
5 6 7 8
1 2 3 4
5 6 7 8 4 o
0.667EI 0.667EI
R14 R24 R34
R44 RRRR4444MMMM
z
1 2 3 4
5 6 7 8
z
1 2 3 4
5 6 7 8
o
3
(e) Unit transverse displacement of joint
2
( Strain mode 3 )
0.667EI
0.667EI 0.667EI
0.667EI 0.667EI
(f) Bending moment diagram due to unit
displacement of joint 2 (Strain Mode 3)
R33
R23
R13
R83
R73
R63
0.667EI
R3R3R3R3MMMM
z
1 2 3 4
5 6 7 8
z
1 2 3 4
5 6 7 8
o
2
(c) Unit transverse displacement of joint
1
(Strain mode 2)
(d) Bending moment diagram due to unit
displacement of joint 1 (Strain Mode 2)
0.667EI
0.667EI
0.667EI
0.667EI
R12
R22
R72 R82
R2R2R2R2MMMM
z
1 2 3 4
5 6 7 8
z
1 2 3 4
5 6 7 8
1
0.667EI
0.667EI (a) Unit transverse displacement of joint
1 (Strain mode 1 ) (b) Bending moment due to unit
displacement of joint 1 (Strain mode 1)
o
0.667EI
0.667EI
R 11 R 21 R 31 R 41
R 51 R 61 R 71
R 81
R1R1R1R1MMMM
79
Similarly, a consideration of the equilibrium of joints for R1, R2, R3, R4, R5, and R6
moment diagrams gives;
From MR1 diagram; 11 21 31 41 51 61 71 81R =R =R =R =R =R =R =R = 0.667EI
From MR2 diagram; 12 22 72 82R =R =R =R = -0.667EI , 32 42 52 62R =R =R =R = 0
From MR3 diagram; 13 83R =R = 0.667EI , 33 63R =R = -0.667EI , 23 43 53 73R =R =R =R = 0
From MR4 diagram; 14 24 34 44 54 64 74 84R =R =R =R =R =R =R =R = 0.667EI
From MR5 diagram; 35 45 55 65R =R =R =R = -0.667EI , 15 25 75 85R =R =R =R = 0
From MR6 diagram; 26 76R =R = -0.667EI , 46 56R =R = 0.667EI , 16 36 66 86R =R =R =R = 0
Fig.4.5.7 Unit joint displacements and corresponding bending moment diagrams ( multi cell doubly symmetric section)
z
6
(k) Unit transverse displacement of joint
6
( Strain mode 6 )
1 2 3 4
5 7 8
z
1 2 3 4
5 6 7 8
(l) Bending moment diagram due to unit
displacement of joint 6 (Strain Mode 6)
0.667EI
0.667EI
0.667EI
0.667EI
0.667EI
R26
o
6
R36
R46
R56
R66
R76
R6R6R6R6MMMM
z
1 2 3 4
5 6 7 8
(j) Bending moment diagram due to unit
displacement of joint 5 (Strain Mode 5)
0.667EI
0.667EI
0.667EI
0.667EI
z
1 2 3 4
5 6 7 8
5
(i) Unit transverse displacement of joint
5
( Strain mode 5 )
o
R55
R45
R65
R35 R5R5R5R5MMMM
80
Since the moment diagrams in Fig. 4.5.7 are based on unit translation of the joints in
the direction of the joints displacements we now multiply these unit diagrams by the
actual displacements (distortion) of the joints obtained from the distortion diagram.
Table 4.3 shows the summation ofij i
R ∆ for the multi-cell doubly symmetric section.
The equations of equilibrium is given by eqn. (4.5.4)
Equilibrium equations
11 1 12 2 13 3 14 4 15 5 16 6 17 7 18 8 1r X +r X +r X +r X +r X +r X +r X +r X +R ∆ = 0
21 1 22 2 23 3 24 4 25 5 26 6 27 7 28 8 2r X +r X +r X +r X +r X +r X +r X +r X +R ∆ = 0
31 1 32 2 33 3 34 4 55 5 36 6 37 7 38 8 3r X +r X +r X +r X +r X +r X +r X +r X +R ∆ = 0
41 1 42 2 43 3 44 4 45 5 46 6 47 7 48 8 4r X +r X +r X +r X +r X +r X +r X +r X +R ∆ = 0
51 1 52 2 53 3 54 4 55 5 56 6 57 7 58 8 5r X +r X +r X +r X +r X +r X +r X +r X +R ∆ = 0
61 1 62 2 63 3 64 4 65 5 66 6 67 7 68 8 6r X +r X +r X +r X +r X +r X +r X +r X +R ∆ = 0
71 1 72 2 73 3 74 4 75 5 76 6 77 7 78 8 7r X +r X +r X +r X +r X +r X +r X +r X +R ∆ = 0
81 1 82 2 83 3 84 4 85 5 86 6 87 7 88 8 8r X +r X +r X +r X +r X +r X +r X +r X +R ∆ = 0 (4.5.4)
Substituting the coefficients ijr and ∑R ∆ij i into the equilibrium eqn. (4.5.4) and solving
we obtain; 1 4 5 8X = X = X = X = -0.1033 and 2 3 6 7X = X = X = X = 0.5166
The final bending moment Fig.4.5.8(a) is obtained by adding the products of
1 2 3M , M , M , 4M , 5 6 7 8M , M , M , M diagrams and 1 2 3X , X , X , 4X 5 6 7 8X , X , X , X
respectively to the products of R1 R2 R3M , M , M , R4M , R5 R6M , M diagrams and
1 2 3∆ , ∆ , ∆ , 4∆ , 5 6∆ , ∆ respectively as per eqn.(4.5.2).
Table 4.3 summation of work done on the joints due to distortion
Work done ( R ∆ij i )
Joint disp direction
Distorti0n ∆ R1j ∆1 R2j ∆2 R3j ∆3 R4j ∆4 R5j ∆5 R6j ∆6 R7j ∆7 R8j ∆8
1 1.929 0.667 0.667 0.667 0.667 0.667 0.667 0.667 0.667
2 5.871 -0.667 -0.667 0 0 0 0 -0.667 -0.667
3 2.014 0.667 0.00 -0.667
0 0 -0.667 0 0.667
4 1.929 0.6670 0.667 0.667 0.667 0.667 0.667 0.667 0.667
5 5.871 0 0 - -0.667 -0.667 -0.667 0 0
81
The distortional bending moment diagrams for multi cell and single cell doubly symmetric sections are shown in Fig.4.5.8 (a) and (b) respectively. From eqn.(4.5.1)
we obtain the hks coefficients for the doubly symmetric sections using Morh’s integral
chart (appendix A).
4.6 INTERACTION OF STRAIN MODES
Vlasov’s coefficients of differential equations of equilibrium involve a combination
of elementary displacements. In the general strain mode there are three strain fields
in the longitudinal direction and four in the transverse direction. These are;
(i) Out of plane displacement due to vertical load characterized by 1ϕ
(ii) Out of plane displacement due to horizontal load normal to bridge axis,
characterized by 2ϕ
(iii) Out of plane displacement due to warping of the cross section, characterized by
3ϕ
(iv) In-plane displacement due to vertical load, characterized by 1ψ
0.667
6 2.014 0 -0.667 0 0.667 0.667 0 -0.667 0
∑R ∆ =ij i 0.000 -2.686 -2.686
0.000 0.000 -2.686 -2.686 0.000
Fig.4.5.8 Bending moment diagrams due to distortion of doubly-symmetric sections
_ _ +
+
+ + + +
_ _
_ 2.366EI
2.366EI
2.366EI -3.606EI -3.606EI
_
1.654EI
2.366EI
1.953EI
-3.606EI
2.366EI
-3.606EI
1.654EI
-2.366EI
1.953EI
+
_ _
+ + +
+
1.953EI 1.273EI
3333M(S)M(S)M(S)M(S)
1.654EI 1.953EI
(a) Multi- cell section
3333M(S)M(S)M(S)M(S)
(b) Single cell section
0.484EI
0.484EI
0.484EI
0.484EI
82
(v) In-plane displacement due to horizontal load normal to bridge axis, characterized
by 2ψ
(vi) In-plane displacement due to distortion of the cross section, characterized by 3ψ
(vii) In-plane displacement due to pure rotation, characterized by 4ψ
Some or all of these strain modes may be present in a given frame depending on the
geometry of the cross section and the nature of loading.
In section 2.5 we noted that torsional loads consisting of opposing vertical forces
result from gravity loads that are eccentric to the centre line of the girder and they
give rise to bending and torsion. We also noted that a torsional load can be modeled
as a uniform (Saint Venant) torsion and a distortional component. Therefore an
eccentric load on a bridge structure introduces interaction between bending, pure
torsion and distortion. The torsional- distortional interaction is of great interest
particularly in a thin walled box structure where the geometry of the cross section
comes into play.
From the computations of Vlasov coefficients in section 4.4 it is observed that
certain coefficients have zero values in the case of mono-symmetric and doubly-
symmetric sections, the number of cells not withstanding. These are
12 13 12 13 12 13 12a , a , b , b , c , c , r and 13r for mono-symmetric section. For doubly
symmetric sections we have 23 23 23 24 14 23 24a , b , c , c , c , r , r and 14r , in addition to the
zero coefficients for mono-symmetric sections. These coefficients with zero values
indicate that there are no interactions between the displacement parameters
represented by the coefficients. For example, the value of the coefficient c23 indicates
the degree of interaction between the in-plane displacement 2ψ due to load normal
to bridge axis and the in-plane displacement 3ψ due to distortion of the cross section.
Where there is no interaction between one strain mode and another the value of the
relevant coefficient is zero indicating that the analysis of the girder for such strain
modes can be done independent of each other.
A doubly symmetric section has only one interaction of strain modes, i.e.
torsional strain mode and distortional strain mode interaction. A mono-symmetric
section on the other hand, has three strain modes interactions. Torsion interacts with
distortion and each of these interacts with flexure about the non axis of symmetry.
Thus we have torsional-distortional interaction, flexural-torsional interaction, and
83
flexural-distortional interaction. The analysis of such sections requires solution of
three coupled differential equations of equilibrium.
A non symmetric section has multiple strain modes interactions: Each of torsion
and distortion interacts with flexure about both axes of non-symmetry in addition to
the interaction between themselves. Thus, we have torsiona-distortional interaction,
torsional-flexural interaction in major and minor axes and distortional-flexural
interaction in major and minor axes. These strain modes interactions are quite
inseparable that it is not possible to examine one strain modes interaction in isolation
of others.
84
(3.4.18b)
85
CHAPTER FIVE
FORMULATION OF DIFFERENTIAL EQUATIONS OF EQUILIBRIUM FOR VARIOUS BOX GIRDER SECTIONS
5.1 PREAMBLE
General expressions for the differential equations of distortional equilibrium for
box girders were derived in chapter three while the coefficients for various cross
sections (doubly symmetric, mono symmetric and non symmetric sections) were
evaluated in chapter four
In this chapter, specific expressions for different conditions of equilibrium are
obtained for doubly symmetric, mono-symmetric and non symmetric box girder
structures. These equilibrium conditions cover all possible strain modes interactions
such as;
(a) Flexural-torsional interaction.
(b) Flexural-distortional interaction.
(c) Torsional-distortional interaction.
(d) Flexural-torsional-distortional interaction.
5.2 DOUBLY SYMMETRIC SECTIONS
5.2.1 Flexural - torsional equilibrium
The relevant coefficients for flexural – torsional equilibrium for doubly symmetric
sections are 22 22 22 22 24 42 44a , b , c , r , r = r , and r . For doubly symmetric sections there
are no interactions between flexure and torsion (strain modes 2 and 4).
Consequently, the coefficients 24 42 24 42c , c , r , r are zero.
Substituting these relevant coefficients into the matrix notation eqn (3.4.17) we
obtain
22 22
''U U10 0 0 0 1
''κ 0 0 U - 0 0 U2 2
0 0 0 0 0 0'' UU 33
0 0
a b -
'
1
'
222 '
3
'
4
V0 0 0 0
V0 c 0 0 = 0
V0 0 0 0
V
(a)
and
22
'0 0 U1
0 0 'U
20 0 0'
U0 0 0 3
0
c-
V0 0 0 0 1
V0 0 0 0 2κ0 0 0 0 V
30 0 0 0 V
4
+
22
44
''V10 0 0''
V0 0 0 2''0 0 0 0 V3
0 0 0 ''V
4
0
r
r
0
=
1
2
3
4
q
q1+
qG
q
(b)
86
By multiplying out matrix equations (a) and (b) above we obtain the following three
equations.
22 2 22 2 22 2ka U '' -b U -c V ' = 0 (5.2.1)
222 2 22 2
qc U '+r V '' = -
G (5.2.2)
444 4
qr V '' = -
G (5.2.3)
From eqn.(5.2.2) we obtain;
22 22 2
22 22
r qU ' = - V '' -
c c G (5.2.4)
Differentiating eqn.(5.2.4) twice gives,
iv22
2 2
22
rU ''' = - V
c (5.2.5)
Differentiating eqn (5.2.1) once we obtain,
22 2 22 2 22 2κa U ''' -b U ' - c V '' = 0 (5.2.6)
Substituting eqns (5.2.4) and (5.2.5) into eqn.(5.2.6) we obtain,
22
iv22 22 222 2 22 2 22 22 2
22 22
r r qκa - V -b - V '' -b - - c V '' = 0
c c c G (5.2.7)
Noting that 22 22 22b = c = r and k = E/G , we obtain from eqn.(5.2.7);
iv 22
22
qV =
a E (5.2.8)
From eqn.(5.2.3) we obtain that
44
44
qV '' = -
r G (5.2.9)
Equations (5.2.8 and 5.2.9) are the differential equations of equilibrium for flexural-
torsional response of a doubly symmetrical box girder section.
5.2.2 Flexural - distortional equilibrium (doubly symmetric sections)
The relevant coefficients for flexural-distortional behaviour of doubly symmetric
sections are those involving strain modes 2 and 3. These are
87
, ,22 22 22 33 33 33a , b c , a , b , c , 33s and 33r . All other interacting coefficients,
23 23 23 23a , b , c , r , are zero because of symmetry in major and minor axis.
Substituting these coefficients into the matrix notation eqn (3.4.17) we obtain;
''' 11 1 ''' 2
22 2 22 2 22 ''' 3
33 33 33 3 33 '
4
V0 0 0 U 0 0 0 U 0 0 0 0
Vκ 0 a 0 U - 0 b 0 U - 0 c 0 0 = 0
V0 0 a U 0 0 b U 0 0 c 0
V
and
'
1
22 '
2
33 '
3
0 0 0U
0 c 0U -
0 0 cU
0 0 0
1
2
333
4
V0 0 0 0
V0 0 0 0k
V0 0 s 0
V0 0 0 0
+
''
1
''22 2
''33 3
''
4
0 0 0 0 V
0 r 0 0 V
0 0 r 0 V
0 0 0 0 V
1
2
3
4
q
q1+ = 0
qG
q
Expanding gives,
22 2 22 2 22 2κa U '' -b U -c V ' = 0 (5.2.9)
33 3 33 3 33 3κa U '' -b U -c V ' = 0 (5.2.10)
c 222 2 22 2
qU '+r V '' = -
G (5.2.11)
333 3 33 3 33 3
qc U '+r V '' - ks V = -
G (5.2.12)
From eqn. (5.2.11) we have,
22 2
22
qU ' = -V '' -
c G (5.2.13)
⇒ iv
2 2U ''' = -V (5.2.14)
Differentiating eqn. (5.2.9) once and substituting eqns. (5.2.13 & 5.2.14) we obtain
iv 222 2 22 2 22 22 2
22
qκa (-V ) -b (-V '') -b - -C v '' = 0
c G
⇒
iv 22 222 2 22 2 22 2
22
b q- κa V +b V ''+ - C v '' = 0
c G
⇒ =iv 222 2
q ka V
G or =iv 2
2
22
qV
ka G
Substituting E
k =G
we obtain iv 22
22
qV =
a E (5.2.15)
From eqn. (5.2.12) we obtain
88
33 33 3 3
33 33
ks qU ' = -V '' + V -
c c G (5.2.16)
iv 333 3 3
33
ksU ''' = -V + V ''
c (5.2.17)
Differentiating eqns. (5.2.10) once we obtain;
33 3 33 3 33 3κa U ''' -b U ' - c V '' = 0 (5.2.18)
Substituting eqns.(5.2.16 & 5.2.17) into eqn.(5.2.18) we obtain
( ( )+ − −2
iv 33 33 22 33 333 3 3 33 3 3 33 33 3
33 33 33
a k s b ks qka V V '' -b V '') - V -b - c V '' = 0
c c c G
⇒ −2
iv 33 33 333 3 3 33 3
33
k a s qka V V ''+ks V =
c G
β⇒ −iv 2 4
3 3 3 3V α V ''+ 4 V = K
where, 2 333
33
ksα =
r
4 333
33
s4β =
a and 3
3
33
qK =
Ea
Thus, for flexural distortional equilibrium we have
iv 22
22
qV =
a E (5.2.19)
iv 2 4
3 3 3 3V -α V ''+ 4β V = K (5.2.20)
5.2.3 Torsional-distortional equilibrium (doubly symmetric section)
The relevant coefficients for torsional-distortional equilibrium are those involving
strain mode 4 ( for pure torsion), and strain mode 3 (for distortion). There will
however be additional coefficients due to the interaction between torsion and
distortion. These non-zero interacting coefficients modifying the displacement
quantities V3 (for distortion) and V4 (for torsion) are 34 43 34c ,c ,r , and 43r .Their values
are obtained as given in table 4.1. Thus, the relevant coefficients for torsional-
distortional behaviour of doubly symmetric sections are,
33 33 33 34 43 33 34 43 33a , b , c , c , c , r , r , r , s and 44r . Substituting these coefficients into
matrix notation eqn.(3.4.17) we obtain,
89
1
1 1
2
2 2
3
33 3 33 3 33 34
4
V '0 0 0 U '' 0 0 0 U 0 0 0 0
V 'k 0 0 0 U '' - 0 0 0 U - 0 0 0 0 = 0
V '0 0 a U '' 0 0 b U 0 0 c c
V '
1 1 1
1
2 2 2
2
33 33 34 3 3 333
3
43 43 44 4 4 4
0 0 0 0 0 0 0 V '' V q0 0 0 0U '
0 0 0 0 0 0 0 V '' V q0 0 0 0 1U ' + -k + = 0
0 0 c 0 0 r r V '' V q0 0 s 0 GU '
0 0 c 0 0 r r V '' V q0 0 0 0
Expanding the matrix notation we obtain the following equations.
33 3 33 3 33 3 34 4ka U '' -b U - c V ' - c V ' = 0 (5.2.21)
333 3 33 3 34 4 33 3
qc U '+r V ''+r V '' -ks V + = 0
G (5.2.22)
443 3 43 3 44 4
qc U '+r V ''+r V ''+ = 0
G (5.2.23)
From eqn. (5.2.23) we obtain
44 43 3 4
43 43
r qU ' = -V '' - V '' -
c c G (5.2.24)
From eqn.(5.2.22) we obtain
34 33 33 3 4 3
33 33 33
r ks qU ' = -V '' - V ''+ V - = 0
c c c G (5.2.25)
Subtracting eqn. (5.2.25) from eqn.(5.2.24) gives
+
34 33 344 44 3
33 43 33 43 33
r ks qr q- V '' - V - = 0
c c c c c G (5.2.26)
Let γ
34 33 344 41 1 1
33 43 33 33 43
r ks qr q- =β ; = ; - =K
c c c c G c G
Eqn. (5.2.26) becomes
γ1 4 1 3 1β V '' - V +K = 0 (5.2.27)
Differentiating eqn. (5.2.21) once with respect to (x) we obtain;
33 3 33 3 33 3 34 4ka U ''' -b U ' - c V '' - C V '' = 0 (5.2.28)
Differentiating eqn. (5.2.24) twice gives
iv iv443 3 4
43
rU ''' = -V - V
c (5.2.29)
Substituting eqn. (5.2.24) and (5.2.29) into eqn. (5.2.28) we obtain,
90
iv iv44 44 433 3 33 4 33 3 33 4 33 33 3 34 4
43 43 43
r r qka (-V )+ka - V -b (-V '') -b - V '' -b - - c V '' - c V '' = 0
c c c G
⇒ iv iv33 44 33 44 33 433 3 4 33 3 4 33 3 34 4
43 43 43
ka r b r b q-ka V - V +b V ''+ V '' - c V '' - c V '' = -
c c c G
⇒ + −
iv iv33 44 33 44 33 433 3 4 34 4
43 43 43
ka r b r b qka V V + c V '' =
c c c G
⇒ +
iv iv 34 43 33 44 3344 43 4 4
43 33 43 33 43
c c -b r br qV V + V '' =
c ka c ka c G
⇒ iv iv
3 1 4 2 4 2V +α V +β V '' = K
where,
441
43
rα = ;
c
34 441
33 43
r rβ = - ;
c c 34 43 33 44
2
33 43
c c -b rβ = ;
ka c γ 33
1
33
ks= ;
c
3 41
33 43
q qK = -
c G c G; 33 334 4
2
33 43 33 43
b bq qK = =
ka c G a c E
Thus the set of differential equations governing torsional-distortional behavior of
doubly symmetric sections is
γ1 4 1 3 1
iv iv
3 1 4 4 2
β V '' - V +K = 0
V +α V +βV '' =K (5.2.30)
5.2.4 Flexural-torsional-distortional equilibrium (doubly symmetric sections)
The relevant coefficients for flexural – torsional – distortional equilibrium are those
involving strain modes 2, 3, and 4. These are 22 23 33 22a , a , a , b ,
23b , 22c , 23 22c , r , 33 33b , c , 24 42 34 43 34 43c = c , c = c , r = r , 24 42r = r , 44 33r and s . Because of
symmetry we note that there is no interaction between strain modes 2 and 3 and
between strain modes 2 and 4. Consequently, 23 23 23 23a = b = c = r = 0 and
24 42 24 42c = c = r = r = 0 .
Substituting these coefficients into matrix eqn. (3.4.17) we obtain,
1 1
22 2 22 2
33 3 33 3
0 0 0 U '' 0 0 0 U
κ 0 a 0 U '' - 0 b 0 U
0 0 a U '' 0 0 b U
-
1
2
22
3
33 34
4
V '0 0 0 0
V '0 c 0 0 = 0
V '0 0 c c
V '
91
1
22
2
33
3
43
0 0 0U '
0 c 0U '
0 0 cU '
0 0 c
-
1
2
33
3
4
0 0 0 0V
0 0 0 0V
κ 0 0 s 0V
0 0 0 0V
+
1
22
2
33 34
3
43 44
4
0 0 0 0V ''
0 r 0 0V ''
0 0 r rV ''
0 0 r rV ''
1
2
3
4
q
q1+ = 0
qG
q
Multiplying out we obtain
22 2 22 2 22 2ka U '' -b U - c V ' = 0 (5.2.31)
33 3 33 3 33 3 34 4ka U '' -b U - c V ' - c V ' = 0 (5.2.32)
0=222 2 22 2
qc U '+r V ''+
G (5.2.33)
333 3 33 3 33 3 34 4
qc U ' -ks V +r V ''+r V '' = -
G (5.2.34)
443 3 43 3 44 4
qc U '+r V '' +r V '' = -
G (5.2.35)
Since there are no interactions between flexure and other strain modes (torsion and
distortion), equations 5.2.31 to 5.2.35 can be grouped into two according to their
strain modes and handled independently. Thus, for flexural strain mode 2 we have
22 2 22 2 22 2ka U '' -b U - c V ' = 0 (5.2.31)
0=222 2 22 2
qc U '+r V ''+
G (5.2.33)
From eqn. (5.2.33), 22 2
22
qU ' = -V '' -
c G
Differentiating twice we obtain;
22 2
22
qU ' = -V '' -
c G and iv
2 2U ''' = -V (5.2.36)
Differentiating eqn. (5.2.31) once we obtain
22 2 22 2 22 2ka U ''' -b U ' - c V '' = 0 (5.2.37)
Substituting eqn. (5.2.36) into eqn. (5.2.37) we obtain
G
iv 222 2 22 2 22 22 2
22
qka (-V ) -b (-V '') -b (- ) - c V '' = 0
c
⇒ iv '' 222 2 22 2 22 2
q-ka V +b V '' - c V + = 0
G
⇒ iv 222 2
qka V =
G or iv 2
2
22
qV =
ka G (5.2.38)
92
For torsional and distortional strain modes 4 and 3 we have
33 3 33 3 33 3 34 4ka U '' -b U - c V ' - c V ' = 0 (5.2.32)
333 3 33 3 33 3 34 4
qc U ' -ks V +r V ''+r V '' = -
G (5.2.34)
443 3 43 3 44 4
qc U '+r V '' +r V '' = -
G (5.2.35)
These equations are the same as the expanded matrix equations 5.2.21, 5.2.22, and
5.2.23 in torsional-distortional equilibrium considerations. Therefore, for doubly
symmetric sections where there are no interactions between flexural strain mode and
other strain modes the governing equations of equilibrium are those obtained for
torsional-distortional equilibrium considerations. That is
γ1 4 1 3 1
iv iv
3 1 4 2 4 2
β V '' - V +K = 0
V +α V +β V '' =K (5.2.39)
where
441
43
rα = ;
c
34 441
33 43
r rβ = - ;
c c 34 43 33 44
2
33 43
c c -b rβ = ;
ka c γ 33
1
33
ks= ;
c
3 41
33 43
q qK = -
c G c G; 33 334 4
2
33 43 33 43
b bq qK = =
ka c G a c E
5.3 MONO-SYMMETRIC SECTIONS
5.3.1 Flexural - torsional equilibrium
The relevant coefficients for flexural – torsional equilibrium are those involving
strain modes 3 and 4, i.e., 22 22 22a , b , c , 24 42 22 24 42c = c ; r , r = r , and 44r . Substituting
these coefficients into the matrix eqn (3.4.17) we obtain
93
22 22
''U U10 0 0 0 1
''κ 0 0 U - 0 0 U2 2
0 0 0 0 0 0'' UU 33
0 0
a b -
'
1
'
222 24 '
3
'
4
V0 0 0 0
V0 c 0 c = 0
V0 0 0 0
V
and
22
42
'0 0 U1
0 0 'U
20 0 0'
U0 0 3
0
c
c
-
V0 0 0 0 1
V0 0 0 0 2κ0 0 0 0 V
30 0 0 0 V
4
+
22 24
42 44
''V10 0 0''
V0 0 2''0 0 0 0 V3
0 0 ''V
4
0
r r
r r
1
2
3
4
q
q1+ = 0
qG
q
. Expanding the above matrix equation we obtain
22 2 22 2 22 2 24 4ka U '' -b U - c V ' - c V ' = 0 (5.3.1)
222 2 22 2 24 4
qc U '+r V ''+r V '' = -
G (5.3.2)
442 2 42 2 44 4
qc U '+r V '' +r V '' = -
G (5.3.3)
From eqn. (5.3.3) we obtain;
44 42 2 4
42 42
iv iv442 2 4
42
r qU ' = -V '' - V '' -
c c G
rU ''' = -V - V
c
(5.3.4)
Substituting eqn. (5.3.4) into eqn. (5.3.2) we obtain
22 44 22 4 222 2 4 22 2 24 4
42 42
c r c q qc (-V '') - V '' - +r V ''+r V '' = -
c c G G
⇒
22 44 22 4 224 4
42 42
c r c q qr - V '' = -
c c G G
⇒
24 42 22 44 22 4 24
42 42
r c - c r c q qV '' = -
c c G G
⇒
22 4 42 24
42 24 22 44 42 24 22 44
c q c qV '' = -
c r - c r G c r - c r G
⇒ 4 1V '' =K (5.3.6)
where,
22 4 42 21
42 24 22 44 42 24 22 44
c q c qK = -
c r - c r G c r - c r G
Differentiating eqn. (5.3.1) once we obtain,
94
22 2 22 2 22 2 24 4κa U ''' - b U ' - c V '' - c V '' = 0 (5.3.7)
Substituting eqns. (5.2.4) and (5.2.5) into eqn. (5.2.7) we obtain,
iv iv44 44 422 2 22 4 22 2 22 22 22 2 24 4
42 42 42
r r qκa (-V )+ka - V -b (-V '') -b - -b - - c V '' - c V '' = 0
c c c G
( )⇒ iv iv 422 42 2 22 44 4 22 44 24 42 4 22
q-κa c V +ka r V - b r - c c V '' = b
G
⇒1
iv iv
2 2 4 1 4 2α V +α V -β V '' =K (5.3.8)
Thus, the differential equations for flexural – torsional equilibrium are
1
4 1
iv iv
2 2 4 1 4 2
V '' =K
α V +α V -β V '' =K (5.3.9)
where,
( )
;
1 22 42 2 22 44 1 22 44 24 42
22 4 42 2 41 2 22
42 24 22 44 42 24 22 44
α =κa c ; α = ka r ; β = b r - c c ;
c q c q qK = - K = b
c r - c r G c r - c r G G
(5.3.10)
5.3.2 Flexural – distortional equilibrium (mono- symmetric Section)
The relevant coefficients are those involving strain modes 2 and 3. These are
22 23 33a , a ; a , 22,b 23 33 23 33 22 23 33 33b , b , c , c , r , r , r , s
Substituting these into matrix eqn. (3.4.17) we obtain
1 1
22 23 2 22 23 2
32 33 3 32 33 3
0 0 0 U '' 0 0 0 U
κ 0 a a U '' - 0 b b U
0 a a U '' 0 b b U
-
1
2
22 23
3
32 33
4
V '0 0 0 0
V '0 c c 0 = 0
V '0 c c 0
V '
1
22 23
2
32 33
3
U 'c c
U 'c c
U '
-
1
2
33
3
4
0 0 0 0V
0 0 0 0V
κ 0 0 s 0V
0 0 0 0V
+
1
22 23
2
32 33
3
4
0 0 0 0V ''
0 r r 0V ''
0 r r 0V ''
0 0 0 0V ''
1
2
3
4
q
q1+ = 0
qG
q
Multiplying out we obtain
22 2 23 3 22 2 23 3 22 2 23 3ka U ''+ka U '' -b U -b U -c V ' - c V ' = 0 (5.3.11)
U32 2 33 3 32 2 33 3 32 2 33 3ka U ''+ka U '' -b U -b - c V ' - c V ' = 0 (5.3.12)
222 2 23 3 22 2 23 3
qc U '+c U '+r V ''+r V '' = -
G (5.3.13)
95
332 2 33 3 33 3 32 2 33 3
qc U '+c U ' -ks V +r V ''+r V '' = -
G (5.3.14)
From eqn. (5.3.13) we obtain that
23 23 22 3 2 3
22 22 22
iv iv23 232 3 2 3
22 22
c r qU ' = - U ' - V '' - V '' -
c c c G
c rU ''' = - U ''' - V - V
c c
(5.3.15)
Substituting eqn. (5.3.15) into eqn. (5.3.14) we obtain
32 23 32 23 32 2 33 32 2 3 33 3 33 3 32 2 33 3
22 22 22
c c c r c q q- U ' - c V '' - V '' - +c U ' -ks V +r V ''+r V '' = -
c c c G G
⇒
2 2
32 32 32 2 333 3 33 3 33 3
22 22 22
c c c q qc - U '+ r - V '' -ks V = -
c c c G G
⇒
2 2
33 22 32 33 23 32 32 2 33 3 33 3
22 22 22
c c - c c c -c c q qU '+ V '' -ks V = -
c c c G G
⇒
22 33 32 23 32 223 3 32 2 2
33 22 32 33 22 32 33 22 32 22
c ks c c qq cU ' = -V ''+ V '+ -
c c - c c c - c G c c -c c G
⇒
22 333 3 3 12
33 22 32
c ksU ' = -V '' + V +K
c c -c (5.3.17)
where
=
32 23 32 221 2 2
33 22 32 33 22 32 22
c c qq cK -
c c - c G c c -c c G
Substituting eqn. (5.3.17) into eqn. (5.3.15) gives
−
23 23 22 33 32 23 22 3 32 2
22 22 33 22 32 33 22 32 22
c c c ks c c qU ' = - (-V '') - V
c c c c - c c c -c c G
23 322
2
33 22 32 22
c q-c-
c c -c c G23 2
2 3
22 22
r q-V '' - V '' -
c c G
⇒
23 23 23 33 23 32 3 3 2 32 2
22 22 33 22 32 33 22 32
c r c ks c q U ' = V '' - V '' - V '' - V +
c c c c -c c c - c G
32 23 22
2
22 33 22 32 22
c c qq- -c G c c -c c G
Let
23 32 2
33 22 32
c qK =
c c -c G
32 23 22
2
22 33 22 32 22
c c qq- -c G c c -c c G
96
∴
23 332 2 3 22
33 22 32
c ksU ' = -V '' - V +K
c c -c (5.3.18)
Differentiating eqn. (5.3.17 & 5.3.18) twice gives
iv 22 333 3 32
33 22 32
iv 23 332 2 32
33 22 32
c ksU ''' = -V + V
c c -c
c ksU ''' = -V - V
c c -c
(5.3.19)
Differentiating eqn. (5.3.11) once gives
22 2 23 3 22 2 23 3 22 2 23 3ka U '''+ka U ''' -b U ' -b U ' - c V '' - c V '' = 0 (5.3.20)
Substituting eqns. (5.3.17), ( 5.3.18) & (5.3.19) into eqn. (5.3.20) we obtain
( )
iv iv23 33 22 3322 2 22 3 23 3 23 32 2
33 22 32 33 22 32
c ks c kska (-V ) -ka V ''+ka -V +ka V ''
c c - c c c -c
23 33 22 3322 2 22 3 22 2 23 3 23 32 2
33 22 32 33 22 32
c ks c ks-b (-V '') -b - V -b K -b (-V '') -b V
c c -c c c - c
23 1 22 2 23 3+b K -c V '' - c V '' = 0
( ) ⇒
2
33 23 22 22 23iv iv
22 2 23 3 3 22 2 33 12
33 22 32
k s a c -a c-ka V -ka V + V '' -b K +b K = 0
c c -c
⇒ iv iv
1 2 2 3 1 3 3α V +α V -β V '' = +K (5.3.21)
( ); ; ;
2
33 23 22 22 23
1 22 2 23 1 3 23 1 22 22
33 22 32
k s a c -a cα = ka α = ka β = K = b K -b K
c c -c
=
32 23 32 221 2 2
33 22 32 33 22 32 22
c c qq cK -
c c - c G c c -c c G
23 32 2
33 22 32
c qK =
c c -c G
32 23 22
2
22 33 22 32 22
c c qq- -c G c c -c c G
Differentiating eqn. (5.3.12) once we obtain
32 2 33 3 32 2 33 3 32 2 33 3ka U '''+ka U ''' -b U ' -b U ' - c V '' - c V '' = 0 (5.3.22)
Substituting eqns. (5.3.17), ( 5.3.18), & (5.3.19) into eqn. (5.3.22) we obtain
iv iv23 33 22 3332 2 32 3 33 3 33 32 2
33 22 32 33 22 32
c ks c kska (-V )+ka - V ''+ka (-V )+ka V ''
c c - c c c - c
23 33 22 3332 2 32 3 32 2 33 3 33 32 2
33 22 32 33 22 32
c ks c ks-b (-V '') -b - V -b K -b (V '') -b V
c c -c c c - c
97
33 1 32 2 33 3-b K -c V '' - c V '' = 0
( ) ⇒
2
33 33 22 32 23iv iv
32 2 33 3 32
33 22 32
k s a c -a c-ka V -ka V + V ''
c c - c
+( )
33 32 23 33 22
3 32 2 33 12
33 22 32
ks b c -b cV -b K -b K = 0
c c -c
γ⇒ iv iv
3 2 4 3 2 3 1 3 4α V +α V -β V '' - V = -K (5.3.23)
where, ( )
2
33 33 22 32 23
3 32 4 33 2 2
33 22 32
k s a c -a cα = ka ; α = ka ; β =
c c -c
( )
1γ
=
33 32 23 33 22
4 32 2 33 12
33 22 32
ks b c -b c; K = b K +b K
c c -c
Thus, the differential equations of equilibrium for flexural-distortional equilibrium of
mono-symmetric sections are:
γ
iv iv
1 2 2 3 1 3 3
iv iv
3 2 4 3 2 3 1 3 4
α V +α V -β V '' =K
α V +α V -β V '' - V = -K (5.3.24)
5.3.3 Torsional- distortional equilibrium (mono- symmetric Section)
The relevant coefficients for torsional-distortional equilibrium (strain modes 3 and
4), are a33, b33, c33, c34, r33, r34 = r 43, r44 and s33. Substituting these into the
matrix notation eqn. (3.4.17) we obtain;
1
2
33 3
0 0 0 U ''
κ 0 0 0 U ''
0 0 a U ''
-
1
2
33 3
0 0 0 U
0 0 0 U
0 0 b U
-
1
2
3
33 34
4
V '0 0 0 0
V '0 0 0 0
V '0 0 c c
V '
=0
33
43
0 0 0
0 0 0
0 0
0 0
c
c
1
2
3
U '
U '
U '
-
0 0 0
0 0 0
0 0
0 0
1
2
333
4
V0
V0κ
Vs 0
V0 0
+33 34
43 44
0 0 0 0
0 0 0 0
0 0
0 0
r r
r r
1
2
3
4
V ''
V ''
V ''
V ''
= −
1
2
3
4
1
G
qqqq
Multiplying out we obtain
33 3 33 3 33 3 34 4ka U '' -b U -c V ' -c V ' = 0 (5.3.25)
333 3 33 3 33 3 34 4
qc U ' -ks V +r V ''+r V '' = -
G (5.3.26)
98
443 3 43 3 44 4
qc U '+r V '' +r V '' = -
G (5.3.27)
From eqn. (5.3.27) we obtain,
44 43 3 4
43 43
iv iv443 3 4
43
r qU ' = -V '' - V '' -
c c G
rU ''' = -V - V
c
(5.3.28)
Substituting eqns. (5.3.28 ) into eqn. (5.3.26) we have
( ) 4''q
G
33 34433 3 33 4 33 3 33 3 34 4
43 43
c qrc -v - c V '' - - ks V +r V ''+r V '' = -
c c G
⇒
33 44 3 33 434 4 33 3
43 43
c r q c qr - V '' -ks V = - +
c G c G
⇒
34 43 33 44 3 33 44 33 3
43 43
r c - c r q c qV '' -ks V = - +
c G c G
( )⇒ 3 434 43 33 44 4 43 33 3 43 43
q qr c - c r V '' - c ks V = -c +c
G G
γ⇒ 1 4 1 3 1β V '' - V =K (5.3.29)
Differentiating eqn. (5.3.25) once we obtain
33 3 33 3 33 3 34 4ka U ''' - b U ' - c V '' - c V '' = 0 (5.3.30)
Substituting eqns. (5.3.28) into eqn. (5.3.30) we have
iv iv44 44 433 3 33 4 33 3 33 4 33 33 3 34 4
43 43 43
r r -qka (-V )+ka - V -b (-V '') -b - V '' -b - c V '' - c V '' = 0
c c c G
( )
⇒
⇒
iv iv33 44 33 44 33 433 3 4 34 4
43 43 43
iv iv 33 433 43 3 33 44 4 33 44 43 34 4
ka r b r -b q-ka V - V + -c V '' =
c c c G
b qka c V +ka r V - b r - c c c V '' =
G
iv iv
3 2 4 2 4 2V +α V -β V '' =K (5.3.31)
Thus, the coupled differential equations of torsional-distortional equilibrium for mono-symmetric sections are:
γ1 4 1 3 1
iv iv
3 2 4 2 4 2
β V '' - V =K
V +α V -β V '' =K (5.3.32)
where, r
c
442
43
α = , 1 34 43 33 44β = r c -c r ; 33 44 34 432
33 43
b r -c cβ = ,
ka c
99
1γ 43 33= c ks ; 341 33 43
qqK = c -c
G G;
33 42
33 43
b qK =
ka c G
5.3.4 Flexural-torsional-distortional equilibrium (mono- symmetric Section)
The relevant coefficients for flexural-torsional-distortional equilibrium are those
involving strain modes 2, 3 and 4. These are; a32, a33, b32, b33, c32, c33, c34, c42, r22,
r32, r33, r34, r42, s33, and r44. Substituting these into the matrix eqn. (3.4.17), noting
that ij ji ij jia = a , b = b , etc, we obtain;
1
22 23 2
32 33 3
0 0 0 U ''
κ 0 a a U ''
0 a a U ''
-
1
22 23 2
32 33 3
0 0 0 U
0 b b U
0 b b U
-
1
2
22 23 24
3
32 33 34
4
V '0 0 0 0
V '0 c c c
V '0 c c c
V '
=0
1
22 23
2
32 33
3
42 43
0 0 0U '
0 c cU '
0 c cU '
0 c c
-
1
2
333
4
V0 0 0 0
V0 0 0 0κ
V0 0 s 0
V0 0 0 0
+
1
22 23 24 2
32 33 34 3
42 43 44 4
0 0 0 0 V ''
0 r r r V ''
0 r r r V ''
0 r r r V ''
1
2
3
4
1
G+ = 0
q
q
q
q
(5.3.33)
Multiplying out we obtain:
22 2 23 3 22 2 23 3 22 2 23 3 24 4ka U ''+ka U '' -b U -b U -c V ' - c V ' - c V ' = 0 (5.3.34)
32 2 33 3 32 2 33 3 32 2 33 3 34 4ka U ''+ka U '' -b U -b U -c V ' -c V ' - c V ' = 0 (5.3.35)
222 2 23 3 22 2 23 3 24 4
qc U '+c U '+r V ''+r V '' +r V '' = -
G (5.3.36)
332 2 33 3 33 3 32 2 33 3 34 4
qc U '+c U ' -ks V +r V ''+r V '' +r V '' = -
G (5.3.37)
442 2 43 3 42 2 43 3 44 4
qc U '+c U '+r V ''+r V ''+r V '' = -
G (5.3.38)
From eqn. (5.3.36) we obtain; 23 23 24 22 3 2 3 4
22 22 22 22
c r r qU ' = - U' - V '' - V '' - V '' -
c c c c G (5.3.39)
From eqn. (5.3.37) we obtain
33 33 33 34 32 3 3 2 3 4
32 32 32 32 32
c ks r r qU ' = - U '+ V - V '' - V '' - V '' -
c c c c c G (5.3.40)
Equating eqns. (5.3.39 & 5.3.40) we obtain;
23 23 33 3324 23 2 3 4 3 3
22 22 22 22 32 32
c r c ksr qU' + V ''+ V '' + V'' + = U ' - V
c c c c G c c
100
33 34 32 3 4
32 32 32
r r q+V ''+ V ''+ V ''+
c c c G
⇒
23 33 33 23 34 33 324 23 3 4 3
22 32 32 22 32 22 32 32 22
c c r r r ks qr q- U ' = - V '' + - V'' - V + -
c c c c c c c c G c G
33 23 34 24
32 22 32 22 333 3 4 3
23 33 23 33 23 3332
22 32 22 32 22 32
r r r r- -
c c c c ksU ' = V ''+ V '' - V
c c c c c c- - c -
c c c c c c
3 2
23 33 23 3332 22
22 32 22 32
q q+ -
c c c cc G - c G -
c c c c
⇒ 3 1 3 2 4 3 3 1
iv iv
3 1 3 2 4 3 3
U ' =β V ''+β V '' -β V +K
U ''' =β V +β V -β V '' (5.3.41)
From eqn. (5.3.36) we obtain;
22 22 24 23 2 2 3 4
23 23 23 23
c r r qU ' = - U ' - V '' - V '' - V '' -
c c c c G (5.3.42)
From eqn. (5.3.37) we obtain
32 33 32 34 33 2 3 2 3 4
33 33 33 33 33
c ks r r qU ' = - U' + V - V '' - V '' - V '' -
c c c c c G (5.3.43)
Equating eqns. (5.3.42) & (5.3.43) we obtain
= −32 3322 22 24 22 2 3 4 3 2 3
23 23 23 23 33 33
c Ksc r r qU '+ V ''+ V '' + V ''+ U ' = U' V
c c c c G c c
+ 32 34 32 3 4
33 33 33
r r qV '' + V ''+ V ''+
c c c G
⇒ − − − + −
32 32 34 33 322 22 24 22 2 4 3
23 33 33 23 33 23 33 33 23
c r r Ks qc r r q- U ' = V ''+ V '' V
c c c c c c c c G c G
32 3422 24
33 23 33 23 33 3 22 2 4 3
32 32 32 32 3222 22 22 22 2233 33 23
23 33 23 33 23 33 23 33 23 33
r rr r- -
c c c c ks q qU ' = V ''+ V '' - V + -
c c c c cc c c c c- - c - c G - c G -
c c c c c c c c c c
⇒ 2 1 2 2 4 3 3 2
iv iv
2 1 2 2 4 3 3
U ' = α V ''+α V '' -α V +K
U ''' = α V +α V -α V '' (5.3.44)
101
Substituting eqns. (5.3.41) & (5.3.44) into eqn. (5.3.38) we have;
42 1 2 42 2 4 43 3 3 42 2 43 1 3 43 2 4c α V ''+c α V '' - c α V +c K +c β V ''+c β V ''
443 3 3 43 1 42 2 43 3 44 4
q-c β V +c K +r V ''+r V ''+r V '' = -
G
( ) ( ) ( )⇒ 42 1 42 2 43 1 43 3 42 2 43 2 44 4c α +r V ''+ c β +r V ''+ c α +c β +r V ''
( ) 442 3 43 3 3 42 2 43 1
q- c α +c β V = - - c K -C K
G
β γ⇒ 4 2 5 3 6 4 1 3 3β V ''+β V ''+ V '' - V = -K (5.3.45)
Differentiating eqn. (5.3.34) we obtain
22 2 23 3 22 2 23 3 22 2 23 3 24 4ka U '''+ka U ''' - b U ' -b U ' - c V '' - c V '' - c V '' = 0 (5.3.46)
Substituting eqns. (5.3.41) & (5.3.44) into eqn. (5.3.46) we obtain;
iv iv iv iv
22 1 2 22 2 4 22 3 3 23 1 3 23 2 4ka α V +ka α V -ka α V ''+ka β V +ka β V - '
23 3 3 22 1 2 22 2 4 22 3 3 22 2 23 1 3 23 2 4-ka β V '' -b α V '' -b α V ''+b α V -b K -b β V '' -b β V ''
23 3 3 23 1 22 2 23 3 24 4+b β V -b K -c V '' - c V '' - c V '' = 0
β β⇒ iv iv iv
22 1 2 23 1 3 22 2 23 3 4 22 1 22 2ka α V +ka V +k(a α +a )V - (b α +c )V ''
23 3 23 1 22 3 23 3 22 2 24 23 2 4-(ka β +b β +kα α +c )V '' - (b α +c +b β )V ''
22 3 23 3 3 22 2 23 1+(b α +b β )V = b K +b K
β γ⇒ iv iv iv
4 2 5 3 6 4 7 2 8 3 9 4 2 3 4 α V +α V +α V -β V '' -β V '' - V ''+ V =K (5.3.47)
Differentiating eqn. (5.3.35) once we obtain
32 2 33 3 32 2 33 3 32 2 33 3 34 4ka U '''+ka U ''' -b U ' -b U ' - c V '' - c V '' - c V '' = 0 (5.3.48)
Substituting eqns .(5.3.46) & (5.3.44) into eqn. (5.3.48) we obtain
iv iv iv iv
32 1 2 32 2 4 32 3 3 33 1 3 33 2 4ka α V +ka α V -ka α V ''+ka β V +ka β V -
33 3 3 32 1 2 32 2 4 32 3 3 32 2 33 1 3 33 2 4-Ka β V '' -b α V '' -b α V ''+b α V -b K -b β V '' -b β V ''
33 3 3 33 1 32 2 33 3 34 4+b β V -b K -c V '' - c V '' - c V '' = 0
⇒ iv iv iv
32 1 2 33 1 3 32 2 33 2 4 32 1 32 2ka α V +ka β V +k(a α +a β )V - (b α +c )V ''
32 3 33 1 33 3 33 3 32 2 34 33 2 4-(ka α +b β +kα β +c )V '' - (b α +c +b β )V ''
32 3 33 3 3 32 2 33 1+(b α +b β )V = b K +b K
γ⇒ iv iv iv
7 2 8 3 9 4 10 2 11 3 12 4 3 3 5α V +α V +α V -β V '' -β V '' -β V ''+ V =K (5.3.49)
The differential equations for flexural- torsional- distortional equilibrium are;
102
4 2 5 3 6 4 1 3 3
iv iv iv
4 2 5 3 6 4 7 2 8 3 9 4 2 3 4
iv iv iv
7 2 8 3 9 4 10 2 11 3 12 4 3 3 5
β V ''+β V ''+β V '' - γ V = -K
α V +α V +α V -β V '' -β V '' -β V ''+ γ V =K
α V +α V +α V -β V '' -β V '' -β V ''+ γ V =K
(5.3.50)
32 3422 24
33 23 33 23 331 2 3
32 32 3222 22 2233
23 33 23 33 23 33
r rr r- -
c c c c Ksα = ; α = ; α ;
c c cc c c- - c -
c c c c c c
4 22 1 5 23 1 6 22 2 23 3α = ka α ; α = ka β ; α = k(a α +a β );
7 32 1 8 33 1 9 32 2 33 3α = ka α ; α = ka β ; α = +k(a α +a β );
33 23 34 24
32 22 32 22 331 2 3
23 33 23 33 23 3332
22 32 22 32 22 32
r r r r- -
c c c c ksβ = ; β = ; β =
c c c c c c- - c -
c c c c c c
( ) ( ) ( )4 42 1 42 5 43 1 43 6 42 2 43 2 44β = c α +r ; β = c β +r ; β = c α +c β +r
7 22 1 22β = (b α +c ) 8 23 3 23 1 22 3 23 9 22 2 24 23 2β = (ka β +b β +kα α +c ); β = -(b α +c +b β )
10 32 1 32β = (b α +c ) 11 32 3 33 1 33 3 33 12 32 2 34 33 2β = (ka α +b β +kα β +c ); β = (b α +c +b β )
( )1γ = 42 3 43 3c α +c β ; γ 2 22 3 23 3= (b α +b β ); γ 3 32 3 33 3= (b α +b β );
3 21
23 33 23 3332 22
22 32 22 32
q qK = -
c c c cc G - c G -
c c c c
; = −
3 22
32 3222 2233 23
23 33 23 33
q qK
c cc cc G - c G -
c c c c
;
43 42 2 43 1
qK = +c K +C K
G; 4 22 2 23 1K = b K +b K ; 5 32 2 33 1K = b K +b K
5.4 NON-SYMMETRIC SECTIONS
The governing equations of equilibrium are,
∑ ∑ ∑3 3 4
ji i ji i jk Ki=1 i=1 k=1
k a U''(x) - b U (x) - c V '(x) = 0 (5.4.1a)
∑ ∑ ∑3 4 4
hhi i hk k hk k
i=1 i=1 i=1
qc U' + r V '' - k s V + = 0
G (5.4.1b)
In matrix notation we have,
103
''
11 12 13 1
''
21 22 23 2
''
31 32 33 3
a a a U
κ a a a U
a a a U
-
11 12 13 1
21 22 23 2
31 32 33 3
b b b U
b b b U
b b b U
-
'
111 12 13 14 '
221 22 23 24 '
331 32 33 34 '
4
Vc c c c
Vc c c c
Vc c c c
V
= (5.4.2a)
11 12 13 '
1
21 22 23 '
2
31 32 33 '
3
41 42 43
c c cU
c c cU
c c cU
c c c
-
1
2
333
4
V0 0 0 0
V0 0 0 0κ
V0 0 s 0
V0 0 0 0
+
''11 12 13 14 1
''21 22 23 24 2
''31 32 33 34 3
''41 42 43 44 4
r r r r V
r r r r V
r r r r V
r r r r V
1
2
3
4
1=
G
q
q
q
q
(5.4.2b)
We note that in non-symmetric sections there are interactions between all the four
strain modes such that the elements of the matrix eqns. (5.4.2a) and (5.4.2b) are non
zero except matrix S, where only S33 has value as explained in section 4.5.
Expanding matrix eqns. (5.4.2a) and (5.4.2b) we obtain
'' '' '' ' ' ' '
11 1 12 2 13 3 11 1 12 2 13 3 11 1 12 2 13 3 14 4ka U +ka U +ka U -b U -b U -b U -c V -c V -c V -c V = 0 (5.4.3)
'' '' '' ' ' ' '
21 1 22 2 23 3 21 1 22 2 23 3 21 1 22 2 23 3 24 4ka U +ka U +ka U -b U -b U -b U -c V -c V -c V -c V = 0 (5.4.4)
'' '' '' ' ' ' '
31 1 32 2 33 3 31 1 32 2 33 3 31 1 32 2 33 3 34 4ka U +ka U +ka U -b U -b U -b U -c V -c V -c V -c V = 0 (5.4.5)
' ' ' '' '' '' '' 111 1 12 2 13 3 11 1 12 2 13 3 14 4
qc U +c U +c U +r V +r V +r V +r V = -
G (5.4.6)
' ' ' '' '' '' '' 221 1 22 2 23 3 21 1 22 2 23 3 24 4
qc U +c U +c U +r V +r V +r V +r V = -
G (5.4.7)
' ' ' '' '' '' '' 331 1 32 2 33 3 31 1 32 2 33 3 34 4 33 3
qc U +c U +c U +r V +r V +r V +r V -ks V = -
G (5.4.8)
' ' ' '' '' '' '' 441 1 42 2 43 3 41 1 42 2 43 3 44 4
qc U +c U +c U +r V +r V +r V +r V = -
G (5.4.9)
Multiplying eqn. (5.4.4) by 11 21(a /a ) and subtracting from eqn. (5.4.3) we obtain
( )'' ''
12 21 11 22 2 13 21 11 33 3 21 11 21 11 1 11 22 21 12 2(a a -a ka )U +(a a -ka a )U +(b a -a b )U + a b -a b U
( ) ( ) ( ) ( )' ' '
11 23 21 13 3 11 21 21 11 1 11 22 21 12 2 11 23 21 13 3+ a b -a b U + a c -a c V + a c -a c V + a c -a c V
( ) '
11 24 21 14 4+ a c -a c V = 0 (5.4.10)
⇒
'' 23 11 21 1322 11 21 12 12 21 11 221 2 2 3
11 21 21 11 11 21 21 11 11 21 21 11
b a -a bb a -a b a a -ka aU = - U - U - U
a b -a b a b -a b a b -a b
'' ' '13 21 11 33 21 11 11 21 22 11 12 213 1 2
11 21 21 11 11 21 21 11 11 21 21 11
a a -ka a c a -c a c a -c a- U - V - V
a b -a b a b -a b a b -a b
104
' '11 23 21 13 11 24 21 143 4
11 21 21 11 11 21 21 11
a c -a c a c -a c- V - V
a b -a b a b -a b (5.4.11)
Multiplying eqn. (5.4.5) by 11 31(a /a ) and subtracting from eqn .(5.4.3) we obtain
( )'' ''
12 31 11 32 2 13 31 11 33 3 11 31 31 11 1 11 32 31 12 2(ka a -ka a )U +(ka a -ka a )U +(a b -a b )U + a b -a b U
( ) ( ) ( ) ( )' ' '
11 33 31 13 3 11 31 31 11 1 11 32 31 12 2 11 33 31 13 3+ a b -a b U + a c -a c V + a c -a c V + a c -a c V
( ) '
11 34 31 14 4+ a c -a c V = 0 (5.4.12)
⇒
''11 32 31 12 12 31 11 32 11 33 31 131 1 2 3
11 31 31 11 11 31 31 11 11 31 31 11
a b -a b ka a -ka a a b -a bU = - U - U - U
a b -a b a b -a b a b -a b
'' ' '13 31 11 33 11 31 31 11 11 32 31 123 1 2
11 31 31 11 11 31 31 11 11 31 31 11
ka a -ka a a c -a c a c -a c- U - V - V
a b -a b a b -a b a b -a b
' '11 33 31 13 11 34 31 143 4
11 31 31 11 11 31 31 11
a c -a c a c -a c- V - V
a b -a b a b -a b (5.4.13)
Combining eqns. (5.4.11) and (5.4.13) we obtain
11 33 21 12 11 33 31 122
11 21 21 11 11 31 31 11
a b -a b a b -a b- U +
a b -a b a b -a b
''12 31 11 3212 21 11 222
11 21 21 11 11 31 31 11
ka a -a aa a -ka a- U +
a b -a b a b -a b
11 23 21 13 11 33 31 133
11 21 21 11 11 31 31 11
a b -a b a b -a b- U +
a b -a b a b -a b
''11 21 11 33 13 31 111 333
11 21 21 11 11 31 31 11
a a -ka a ka a -ka a- U +
a b -a b a b -a b
''11 31 31 1111 21 21 111
11 21 21 11 11 31 31 11
a c -a ca c -a c- V +
a b -a b a b -a b
'11 32 31 1211 22 21 122
11 21 21 11 11 31 31 11
a c -a ca c -a c- U +
a b -a b a b -a b
'11 23 21 13 11 33 31 133
11 21 21 11 11 31 31 11
a c -a c a c -a c- V +
a b -a b a b -a b
'11 34 31 1411 24 21 144
11 21 21 11 11 31 31 11
a c -a ca c -a c- V = 0
a b -a b a b -a b
⇒ '' '' ' ' ' '
1 2 2 2 3 3 4 3 5 1 6 2 7 3 8 4αU +α U +α U +α U +α V +α V +α V +α V = 0 (5.4.14)
where,
11 33 21 12 11 33 31 121
11 21 21 11 11 31 31 11
a b -a b a b -a bα = -
a b -a b a b -a b (a)
12 31 11 3212 21 11 222
11 21 21 11 11 31 31 11
ka a -a aa a -ka aα = -
a b -a b a b -a b (b)
11 23 21 13 11 33 31 133
11 21 21 11 11 31 31 11
a b -a b a b -a bα = -
a b -a b a b -a b (c)
105
11 21 11 33 13 31 111 334
11 21 21 11 11 31 31 11
a a -ka a ka a -ka aα = -
a b -a b a b -a b (d)
11 31 31 1111 21 21 115
11 21 21 11 11 31 31 11
a c -a ca c -a cα = -
a b -a b a b -a b (e)
11 32 31 1211 22 21 126
11 21 21 11 11 31 31 11
a c -a ca c -a cα = -
a b -a b a b -a b (f)
11 23 21 13 11 33 31 137
11 21 21 11 11 31 31 11
a c -a c a c -a cα = -
a b -a b a b -a b (g)
11 34 31 1411 24 21 148
11 21 21 11 11 31 31 11
a c -a ca c -a cα = -
a b -a b a b -a b (h) (5.4.15)
Multiplying eqn. (5.4.5) by 21 31(a /a )and subtracting from eqn. (5.4.4) we obtain
( )'' ''
12 31 21 32 2 13 31 21 33 3 21 31 31 21 1 21 32 31 22 2(ka a -ka a )U +(ka a -ka a )U +(a b -a b )U + a b -a b U
( ) ( ) ( ) ( )' ' '
21 33 31 23 3 21 31 31 21 1 21 32 31 22 2 21 33 31 23 3+ a b -a b U + a c -a c V + a c -a c V + a c -a c V
( ) '
21 34 31 24 4+ a c -a c V = 0 (5.4.16)
⇒
''21 32 31 22 12 31 21 32 21 33 31 231 2 2 321 31 31 21 21 31 31 21 21 31 31 21a b -a b ka a -ka a a b - a bU = - U - U - U
a b -a b a b -a b a b - a b
'' ' '13 31 21 33 21 31 31 21 21 32 31 223 1 221 31 31 21 21 31 31 21 21 31 31 21ka a -ka a a c - a c a c - a c- U - V - V
a b -a b a b -a b a b -a b
' '21 33 31 23 21 34 31 243 4
21 31 31 21 21 31 31 21
a c -a c a c -a c- V - V
a b -a b a b -a b (5.4.17)
Combining eqns. (5.4.11 and 5.4.17) we obtain
21 32 31 2211 22 21 122
11 21 21 11 21 31 31 21
a b -a ba b -a b- U +
a b -a b a b -a b
''12 31 21 3212 21 11 222
11 21 21 11 21 31 31 21
ka a -a aa a -ka a- U +
a b -a b a b -a b
11 23 21 13 21 33 31 233
11 21 21 11 21 31 31 21
a b -a b a b -a b- U +
a b -a b a b -a b
''13 21 11 33 13 31 21 333
11 21 21 11 21 31 31 21
a a -ka a ka a -ka a- U +
a b -a b a b -a b
''21 31 31 2111 21 21 111
11 21 21 11 21 31 31 21
a c -a ca c -a c- V +
a b -a b a b -a b
'21 32 31 2211 22 21 122
11 21 21 11 21 31 31 21
a c -a ca c -a c- V +
a b -a b a b -a b
'11 23 21 13 21 33 31 233
11 21 21 11 21 31 31 21
a c -a c a c -a c- V +
a b -a b a b -a b
'21 34 31 2411 24 21 144
11 21 21 11 21 31 31 21
a c -a ca c -a c- V = 0
a b -a b a b -a b
106
⇒ '' '' ' ' ' '
1 2 2 2 3 3 4 3 5 1 6 2 7 3 8 4ρU +ρ U +ρ U +ρ U +ρ V +ρ V +ρ V +ρ V = 0 (5.4.18)
where,
21 32 31 2211 22 21 121 11 21 21 11 21 31 31 21a b - a ba b - a bρ = -
a b - a b a b - a b (a)
12 31 21 3212 21 11 222 11 21 21 11 21 31 31 21ka a - a aa a -ka aρ = -
a b - a b a b - a b (b)
11 23 21 13 21 33 31 233 11 21 21 11 21 31 31 21a b - a b a b - a bρ = -
a b - a b a b - a b (c)
13 21 11 33 13 31 21 334 11 21 21 11 21 31 31 21a a -ka a ka a -ka aρ = -
a b - a b a b - a b (d)
5ρ
=
21 31 31 2111 21 21 1111 21 21 11 21 31 31 21a c - a ca c - a c-
a b - a b a b - a b (e)
21 32 31 2211 22 21 126 11 21 21 11 21 31 31 21a c - a ca c - a cρ = -
a b - a b a b - a b (f )
7ρ
11 23 21 13 21 33 31 2311 21 21 11 21 31 31 21a c - a c a c - a c= -
a b - a b a b - a b (g)
8ρ
=
21 34 31 2411 24 21 1411 21 21 11 21 31 31 21a c - a ca c - a c-
a b - a b a b - a b (h) (5.4.19)
Combining eqns. (5.4.13) and (5.4.17) we obtain
11 32 31 12 21 32 31 22 211 31 31 11 21 31 31 21a b - a b a b - a b- U +
a b - a b a b - a b
''12 31 11 32 12 31 21 32 211 31 31 11 21 31 31 21a a - a a ka a -ka a- U +
a b - a b a b - a b
11 33 31 13 21 33 31 23 311 31 31 11 21 31 31 21a b - a b a b - a b- U
a b - a b a b - a b+
''11 31 11 33 13 31 21 33 311 31 31 11 21 31 31 21a a -ka a ka a -ka a- U +
a b - a b a b - a b
'11 31 31 11 21 31 31 21 111 31 31 11 21 31 31 21a c - a c a c - a c- V +
a b - a b a b - a b
'11 32 32 12 21 32 31 22 211 31 31 11 21 32 31 21a c - a c a c - a c- V +
a b - a b a b - a b
'11 33 31 13 21 33 31 23 311 31 31 11 21 31 31 21a c - a c a c - a c- V +
a b - a b a b - a b
'11 34 31 14 21 34 31 24 411 31 31 11 21 31 31 21a c - a c a c - a c- V = 0
a b - a b a b - a b
⇒ '' '' ' ' ' '1 2 2 2 3 3 4 3 5 1 6 2 7 3 8 4η U +η U +η U +η U +η V +η V +η V +η V =0 (5.4.20)
where,
11 32 31 12 21 32 31 221 11 31 31 11 21 31 31 21a b - a b a b - a bη = -
a b - a b a b - a b (a)
107
12 31 11 32 12 31 21 321 11 31 31 11 21 31 31 21a a - a a ka a -ka aη = -
a b - a b a b - a b (b)
11 33 31 13 21 33 31 233 11 31 31 11 21 31 31 21a b - a b a b - a bη = -
a b - a b a b - a b (c)
11 31 11 33 13 31 21 334 11 31 31 11 21 31 31 21a a -ka a ka a -ka aη = -
a b - a b a b - a b (d)
5η
11 31 31 11 21 31 31 2111 31 31 11 21 31 31 21a c - a c a c - a c= -
a b - a b a b - a b (e)
6η
11 32 32 12 21 32 31 2211 31 31 11 21 32 31 21a c - a c a c - a c= -
a b - a b a b - a b (f)
7η
11 33 31 13 21 33 31 2311 31 31 11 21 31 31 21a c - a c a c - a c= -
a b - a b a b - a b (g)
8η
11 34 31 14 21 34 31 2411 31 31 11 21 31 31 21a c - a c a c - a c= -
a b - a b a b - a b (h) (5.4.21)
Multiplying eqn. (5.4.7) by 11 21(c /c )and subtracting from eqn. (5.4.6) we obtain
( )' ' '' ''12 21 11 22 2 13 21 11 23 3 11 21 21 11 1 12 21 22 11 2(c c - c c )U +(c c - c c )U +(r c - r c )V + r c - r c V +
+ ( )
'' '' 11 2 21 113 21 23 11 3 14 21 24 11 4 c q c qr c - r c V +(r c - r c )V = -
G G
⇒ ' ' '' '' '' ''1 2 2 3 3 1 4 2 5 3 6 4 1β U +β U +β V +β V +β V +β V =K (5.4.22)
where, ( )1 12 21 11 22β = c c - c c ; ( )2 13 21 11 23β = c c - c c ; ( )3 21 11 11 21β = c r - c r
( )4 21 12 11 2β = c r - c r ; ( )5 21 13 11 23β = c r - c r ( )6 21 14 11 24β = c r - c r (5.4.23)
Multiplying eqn. (5.4.9) by 31 41( / )c c and subtracting from eqn. (5.4.8) we obtain
( )' ' '' ''32 41 31 42 2 33 41 31 43 3 41 31 31 41 1 41 31 31 42 2(c c - c c )U +(c c - c c )U +(c r - c r )V + c r - c r V +
+ ( )
'' '' 31 4 41 341 33 31 44 3 41 34 31 44 4 33 3 c q c qc r - c r V +(c r - c r )V -ks V = -
G G
⇒ ' ' '' '' '' ''
1 2 2 3 3 1 4 2 5 3 6 4 7 3 2λ U +λ U +λ V + λ V +λ V +λ V - λ v =K (5.4.24)
where, ( )1 32 41 31 42λ = c c - c c ; ( )2 33 41 31 43λ = c c - c c ; ( )3 41 31 31 41λ = c r - c r
( )4 41 32 31 42λ = c r - c r ; ( )5 41 33 31 43λ = c r - c r ( )6 41 34 31 44λ = c r - c r (5.4.25)
Multiplying eqn. (5.4.8) by 11 31( / )c c and subtracting from eqn. (5.4.6) we obtain
108
( )' ' '' ''11 32 12 31 2 11 33 13 31 3 11 31 31 11 1 11 32 31 12 2(c c - c c )U +(c c - c c )U +(c r - c r )V + c r - c r V +
+ ( )
'' '' 31 1 11 311 33 31 14 3 11 34 31 14 4 11 33 3 c q c qc r - c r V +(c r - c r )V +kc s V = -
G G
⇒ ' ' '' '' '' ''1 2 2 3 3 1 4 2 5 3 6 4 7 3 3ν U +ν U +ν V +ν V +ν V +ν V +ν v =K (5.4.26)
Thus we have '' '' ' ' ' '1 2 2 2 3 3 4 3 5 1 6 2 7 3 8 4α U +α U +α U +α U +α V +α V +α V +α V =0 (5.4.14) '' '' ' ' ' '1 2 2 2 3 3 4 3 5 1 6 2 7 3 8 4ρ U +ρ U +ρ U +ρ U +ρ V +ρ V +ρ V +ρ V =0 (5.4.18) '' '' ' ' ' '1 2 2 2 3 3 4 3 5 1 6 2 7 3 8 4η U +η U +η U +η U +η V +η V +η V +η V =0 (5.4.20) ' ' '' '' '' ''1 2 2 3 3 1 4 2 5 3 6 4 1β U +β U +β V +β V +β V +β V =K (5.4.22) ' ' '' '' '' ''1 2 2 3 3 1 4 2 5 3 6 4 7 3 2λ U +λ U +λ V +λ V +λ V +λ V - λ v =K (5.4.24) ' ' '' '' '' ''1 2 2 3 3 1 4 2 5 3 6 4 7 3 3ν U +ν U +ν V +ν V +ν V +ν V +ν v =K (5.4.26)
Multiplying eqn. (5.4.18) by 1 1( / )α ρ and subtracting eqn. (5.4.14) from it, we obtain
( )'' '' '1 2 2 1 2 1 3 3 1 3 1 4 4 1 3 1 5 5 1 1(α ρ -α ρ )U +(α ρ -β ρ )U +(α ρ - α ρ )U + α ρ - α ρ V +
+ ( ) ( )' ' '1 6 6 1 2 1 7 7 1 3 1 8 8 1 4α ρ -α ρ V +(α ρ - α ρ )V + α ρ -α ρ V =0
⇒ '' '' ' ' ' '1 2 2 3 3 3 4 1 5 2 6 3 7 4μU +μ U +μ U +μ V +μ V +μ V +μ v = 0 (5.4.27)
where, 1 1 2 2 1μ =(α ρ -α ρ ); 2 1 3 3 1μ =(α ρ -β ρ ); 3 1 4 4 1μ =(α ρ -α ρ )
( )4 1 5 5 1µ = α ρ -α ρ ; ( )5 1 6 6 1μ = α ρ - α ρ 6 1 7 7 1μ =(α ρ - α ρ )
( )7 1 8 8 1μ = α ρ - α ρ (5.4.28)
Multiplying eqn. (5.4.20) by 1 1( / )α η and subtracting eqn. (5.4.14) from it, we obtain
( )'' '' '1 2 2 1 2 1 3 3 1 3 1 4 4 1 3 1 5 5 1 1(α η -α η )U +(α η -β η )U +(α η - α η )U + α η -α η V +
+ ( ) ( )' ' '1 6 6 1 2 1 7 7 1 3 1 8 8 1 4α η - α η V +(α η -α η )V + α η -α η V =0
⇒ '' '' ' ' ' '1 2 2 3 3 3 4 1 5 2 6 3 7 4ε U +ε U +ε U +ε V +ε V +ε V +ε v = 0 (5.4.29)
where, 1 1 2 2 1ε =(α η - α η ); 2 1 3 3 1ε =(α η - α η ); 3 1 4 4 1ε =(α η - α η )
( )4 1 5 5 1ε = α η - α η ; ( )5 1 6 6 1ε = α η - α η 6 1 7 7 1ε =(α η -α η )
( )7 1 8 8 1ε = α η - α η (5.4.30)
Multiplying eqn. (5.4.20) by 1 1( / )ρ η and subtracting eqn. (5.4.18) from it, we obtain
( )'' '' '1 2 2 1 2 1 3 3 1 3 1 4 4 1 3 1 5 5 1 1(ρ η -ρ η )U +(ρ η -ρ η )U +(ρ η -ρ η )U + ρ η -ρ η V +
109
+ ( ) ( )' ' '1 6 6 1 2 1 7 7 1 3 1 8 8 1 4ρ η -ρ η V +(ρ η -ρ η )V + ρ η -ρ η V =0
⇒ '' '' ' ' ' '1 2 2 3 3 3 4 1 5 2 6 3 7 4γ U +γ U +γ U +γ V +γ V +γ V +γ v = 0 (5.4.31)
where, 1 1 2 2 1γ =(ρ η -ρ η ); 2 1 3 3 1γ =(ρ η -ρ η ); 3 1 4 4 1γ =(ρ η -ρ η )
( )4 1 5 5 1γ = ρ η -ρ η ; ( )5 1 6 6 1γ = ρ η -ρ η 6 1 7 7 1γ =(ρ η -ρ η )
( )7 1 8 8 1γ = ρ η -ρ η (5.4.32)
Thus we have '' '' ' ' ' '1 2 2 3 3 3 4 1 5 2 6 3 7 4μU +μ U +μ U +μ V +μ V +μ V +μ v = 0 (5.4.27) '' '' ' ' ' '1 2 2 3 3 3 4 1 5 2 6 3 7 4ε U +ε U +ε U +ε V +ε V +ε V +ε v = 0 (5.4.29) '' '' ' ' ' '1 2 2 3 3 3 4 1 5 2 6 3 7 4γ U +γ U +γ U +γ V +γ V +γ V +γ v = 0 (5.4.31)
' ' '' '' '' ''1 2 2 3 3 1 4 2 5 3 6 4 1β U +β U +β V +β V +β V +β V =K (5.4.22) ' ' '' '' '' ''1 2 2 3 3 1 4 2 5 3 6 4 7 3 2λ U +λ U +λ V +λ V +λ V +λ V - λ v =K (5.4.24) ' ' '' '' '' ''1 2 2 3 3 1 4 2 5 3 6 4 7 3 3ν U +ν U +ν V +ν V +ν V +ν V +ν v =K (5.4.26)
Multiplying eqn. (5.4.22) by 1 1( / )λ β and subtracting eqn. (5.4.24) from it, we obtain
( )' '' '' ''1 2 2 1 3 1 3 3 1 1 1 4 4 1 2 1 5 5 1 3(λ β - λ β )U +(λ β - λ β )V +(λ β - λ β )V + λ β - λ β V +
( ) ( )''1 6 6 1 4 7 3 1 1 1 2+ λ β - λ β V +λ V = λ K -β K
⇒
' '' '' '' ''1 3 3 1 1 5 5 1 1 6 6 11 4 4 13 1 2 3 41 2 2 1 1 2 2 1 1 2 2 1 1 2 2 1λ β - λ β λ β - λ β λ β - λ βλ β - λ βU = - V - V - V - V
λ β - λ β λ β - λ β λ β - λ β λ β - λ β
7 1 1 1 231 2 1 1 1 2 2 1λ λ K -β K- V +
λ β - λ β λ β - λ β
' '' '' '' ''3 1 1 2 2 3 3 4 4 5 3 4⇒U = -ω V -ω V -ω V -ω V -ω V +K (5.4.33)
where
1 3 3 11 1 2 2 1λ β - λ βω =
λ β - λ β;
1 4 4 12 1 2 2 1λ β - λ βω =
λ β - λ β;
1 5 5 13 1 2 2 1λ β - λ βω =
λ β - λ β
1 6 6 14 1 2 2 1λ β - λ βω =
λ β - λ β;
75 1 2 1 1λω =
λ β - λ β;
1 1 1 24 1 2 2 1λ K -β KK =
λ β - λ β
Multiplying eqn .(5.4.24) by 2 2( / )β λ and subtracting eqn. (5.4.22) from it, we obtain
( )' '' '' ''1 2 2 1 2 3 3 2 3 1 4 2 2 4 2 5 2 2 5 3(λ β - λ β )U +(λ β - λ β )V +(λ β - λ β )V + λ β - λ β V +
( ) ( ) ( )'' ''5 2 2 5 3 6 2 2 6 4 7 2 3 2 2 2 2+ λ β - λ β V + λ β - λ β V - λ β V = β K - λ K
110
⇒
' '' '' '' ''3 2 2 3 5 2 2 5 6 2 2 64 2 2 42 1 2 3 41 2 2 1 1 2 2 1 1 2 2 1 1 2 2 1λ β - λ β λ β - λ β λ β - λ βλ β - λ βU = - V - V - V - V
λ β - λ β λ β - λ β λ β - λ β λ β - λ β
7 2 2 2 2 131 2 2 1 1 2 2 1λ β β K - λ K- V +
λ β - λ β λ β - λ β
' '' '' '' ''2 1 1 2 2 3 3 4 4 5 3 5⇒U = -π V -π V -π V -π V -π V +K (5.4.34)
Differentiating each of eqns. (5.4.27), ( 5.4.29) &( 5.4.31) once we obtain '' ' ''' '' '' '' ''1 2 2 3 3 3 4 1 5 2 6 3 7 4μU +μ U +μ U +μ V +μ V +μ V +μ v = 0 (5.4.35) ''' ' ''' '' '' '' ''1 2 2 3 3 3 4 1 5 2 6 3 7 4ε U +ε U +ε U +ε V +ε V +ε V +ε v = 0
(5.4.36) ''' ' ''' '' '' '' ''1 2 2 3 3 3 4 1 5 2 6 3 7 4γ U +γ U +γ U +γ V +γ V +γ V +γ v = 0 (5.4.37)
Differentiating eqn. ( 5.4.26) once we obtain ''''3333 = 0= 0= 0= 0
'' '' ''' ''' ''' '''1 2 2 3 3 1 4 2 5 3 6 4 7ν U +ν U +ν V +ν V +ν V +ν V +ν V (5.4.38)
But ' '' '' '' ''2 1 1 2 2 3 3 4 4 5 3 5U = -π V -π V -π V -π V -π V +K
'' ''' ''' ''' ''' '2 1 1 2 2 3 3 4 4 5 5U = -π V -π V -π V -π V -π V (5.4.39)
''' iv iv iv iv ''2 1 1 2 2 3 3 4 4 5 3U = -π V -π V -π V -π V -π V
Also,
' '' '' '' ''3 1 1 2 2 3 3 4 4 5 3 4U = -ω V -ω V -ω V -ω V -ω V +K
'' ''' ''' ''' ''' '3 1 1 2 2 3 3 4 4 5 3U = -ω V -ω V -ω V -ω V -ω V
''' iv iv iv iv ''3 1 1 2 2 3 3 4 4 5 3U = -ω V -ω V -ω V -ω V -ω V (5.4.40)
Substituting eqns. ( 5.4.39) and( 5.4.40) into eqn. (5.4.35) we obtain
( )iv IV iv iv ''1 1 1 2 2 3 3 4 4 5 3μ -π V -π V -π V +π V -π V + ( )'' '' '' ''2 1 1 2 2 3 3 4 4 5 3 4μ -ω V -ω V -ω V -ω V -ω V +K
( )iv iv iv iv ''3 1 1 2 2 3 3 4 4 5 3+μ -ω V -ω V -ω V -ω V -ω V '' '' '' ''4 1 5 2 6 3 7 4+μ V +μ V +μ V +μ V =0
⇒ ( ) ( ) ( ) ( )iv iv iv iv3 1 1 1 1 1 2 3 2 2 1 3 3 3 3 1 4 3 4 4- μ ω +μ π V - μ π +μ ω V - μ π +μ ω V - μ π +μ ω V
( ) ( ) ( )'' '' ''2 1 4 1 2 2 3 5 5 2 1 5 2 3 6 3- μ ω +μ V - μ ω +μ ω -μ V - μ π +μ ω -μ V
( ) ( )''3 4 7 4 2 5 3 4 2- μ ω -μ V - μ ω V +K μ =0
⇒⇒⇒⇒ iv iv iv iv '' '' '' ''1 1 2 2 3 3 4 4 5 1 6 2 7 3 8 4 9 3 2 4A V +A V +A V +A V +A V +A V +A V +A V +A V =μ K (5.4.41)
Also, substituting eqns. (5.4.39) and (5.4.40) into eqn. (5.4.36) we obtain
( )iv IV iv iv ''1 1 1 2 2 3 3 4 4 5 3ε -π V -π V -π V +π V -π V + ( )'' '' '' ''2 1 1 2 2 3 3 4 4 5 3 4+ε -ω V -ω V -ω V -ω V -ω V +K
( )iv iv iv iv ''3 1 1 2 2 3 3 4 4 5 3+ε -ω V -ω V -ω V -ω V -ω V '' '' '' ''4 1 5 2 6 3 7 4+ε V +ε V +ε V +ε V = 0
⇒ ( ) ( ) ( ) ( )iv iv iv iv1 1 3 1 1 1 2 3 2 2 1 3 3 3 3 1 4 3 4 4- ε π +ε ω V - ε π +ε ω V - ε π +ε ω V - ε π +ε ω V
( ) ( ) ( )'' '' ''2 1 4 1 2 2 5 2 1 5 2 3 6 3- ε ω +ε V - ε ω - ε V - ε π +ε ω - ε V
( ) ( )''2 4 7 4 2 5 3 2 4- ε ω - ε V - ε ω V +ε K =0
111
⇒⇒⇒⇒ iv iv iv iv '' '' '' ''1 1 2 2 3 3 4 4 5 1 6 2 7 3 8 4 9 3 2 4B V +B V +B V +B V +B V +B V +B V +B V +B V = ε K (5.4.42)
Substituting eqns. ( 5.4.39) and (5.4.40) into eqn. (5.4.37) we obtain
( )iv IV iv iv ''1 1 1 2 2 3 3 4 4 5 3γ -π V -π V -π V +π V -π V + ( )'' '' '' ''2 1 1 2 2 3 3 4 4 5 3 4γ -ω V -ω V -ω V -ω V -ω V +K
( )iv iv iv iv ''3 1 1 2 2 3 3 4 4 5 3+γ -ω V -ω V -ω V -ω V -ω V '' '' '' ''4 1 5 2 6 3 7 4+γ V +γ V +γ V +γ V =0
⇒ ( ) ( ) ( ) ( )iv iv iv iv1 1 3 1 1 1 2 3 2 2 1 3 3 3 3 1 4 3 4 4- γ π + γ ω V - γ π +γ ω V - γ π +γ ω V - γ π +γ ω V
( ) ( ) ( )'' '' ''2 1 4 1 2 2 5 2 1 5 2 3 6 3- γ ω +γ V - γ ω - γ V - γ π +γ ω -μ V
( ) ( )''2 4 7 4 2 5 3 2 4- γ ω - γ V - γ ω V +γ K =0
⇒⇒⇒⇒ iv iv iv iv '' '' '' ''1 1 2 2 3 3 4 4 5 1 6 2 7 3 8 4 9 3 2 4C V +C V +C V +C V +C V +C V +C V +C V +C V = γ K (5.4.43)
Substituting eqns. (5.4.39) and (5.4.40) into eqn. (5.4.38) we obtain
( )υ1111- +- +- +- +''' ''' ''' ''' '1 1 2 2 3 3 4 4 5 3π V +π V +π V +π V -π V ( )υ2222----
''' ''' ''' ''' '1 1 2 2 3 3 4 4 5 3ω V +ω V +ω V +ω V +ω V ''' ''' ''' ''' '3 1 4 2 5 3 6 4 7 3+υ V +υ V +υ V +υ V -υ V =0
(5.4.44)
Differentiating eqn. (5.4.44) once we obtain
( )iv iv iv iv ''
1 1 1 2 2 3 3 4 4 5 3-υ π V +π V +π V +π V -π V + ( )iv iv iv iv ''
2 1 1 2 2 3 3 4 4 5 3-υ ω V +ω V +ω V +ω V +ω V
iv iv iv iv ''
3 1 4 2 5 3 6 4 7 3-υ V -υ V -υ V -υ V -υ V = 0
⇒ ( ) ( ) ( )iv iv iv1 1 3 1 3 1 1 2 2 2 4 2 1 3 2 3 5 3υ π +υ ω -υ V + υ π +υ ω -υ V + υ π +υ ω -υ V
( ) ( )iv ''1 4 2 4 6 4 2 5 1 5 7 3+ υ π +υ ω -υ V + υ ω -υ π -υ V = 0
⇒⇒⇒⇒ iv iv iv iv ''
1 1 2 2 3 3 4 4 5 3D V +D V +D V +D V +D V = 0 (5.4.45)
Thus the coupled differential equations of equilibrium for non symmetric sections are
iv iv iv iv '' '' '' ''
1 1 2 2 3 3 4 4 5 1 6 2 7 3 8 4 9 3 2 4A V + A V + A V + A V + A V + A V + A V + A V + A V =µ K (5.4.41)
iv iv iv iv '' '' '' ''
1 1 2 2 3 3 4 4 5 1 6 2 7 3 8 4 9 3 2 4B V +B V +B V +B V +B V +B V +B V +B V +B V = ε K (5.4.42)
iv iv iv iv '' '' '' ''
1 1 2 2 3 3 4 4 5 1 6 2 7 3 8 4 9 3 2 4C V +C V +C V +C V +C V +C V +C V +C V +C V = γ K (5.4.43)
iv iv iv iv ''
1 1 2 2 3 3 4 4 5 3D V +D V +D V +D V +D V = 0 (5.4.45)
where
( )1 3 1 1 1A = µ ω +µ π ( )2 1 2 3 2A = µ π +µ ω
( )3 1 3 3 3A = µ π +µ ω ( )4 1 4 3 4A = µ π +µ ω
( )5 2 1 4A = µ ω +µ ( )6 2 3 3 5A = µ ω+µ ω -µ
( ) ( )7 1 5 2 3 6 8 3 4 7A = µ π +µ ω -µ A = µ ω -µ ( )9 2 5A = µ ω
112
( )
( )
( )
( )
1 1 1 3 1
2 1 2 3 2
3 1 3 3 3
4 1 4 3 4
B = ε π + ε ω
B = ε π + ε ω
B = ε π + ε ω
B = ε π + ε ω
( )
( )
( )
( )
5 2 1 4
6 2 2 5
7 1 5 2 3 6
8 2 4 7
B = ε ω + ε
B = ε ω - ε
B = ε π + ε ω - ε
B = ε ω - ε
( )9 2 5B = ε ω (5.4.47)
( )
( )
( )
( )
1 1 1 3 1
2 1 2 3 2
3 1 3 3 3
4 1 4 3 4
C = γ π +γ ω
C = γ π + γ ω
C = γ π + γ ω
C = γ π +γ ω
( )
( )
( )
( )
5 2 1 4
6 2 2 5
7 1 5 2 3 6
8 2 4 7
C = γ ω + γ
C = γ ω - γ
C = γ π + γ ω -µ
C = γ ω - γ
( )9 2 5C = γ ω (5.4.48)
( )
( )
( )
1 1 1 3 1 3
2 1 2 2 2 4
3 1 3 2 3 5
D = υ π +υ ω -υ
D = υ π +υ ω -υ
D = υ π +υ ω -υ
( )
( )4 1 4 2 4 6
5 2 5 1 5 7
D = υ π +υ ω -υ
D = υ ω -υ π -υ (5.4.49)
Table 5.1 Summary of Differential Equations of Equilibrium
Equilibriu
m
Condition
(a)
Doubly
symmetric
Sections
(b)
Mono-symmetric Sections
(c) Non-symmetric sections
Flexural-
torsional
equilibriu
m
qiv 2V =2
Ea22
qiv 4V =4
r G44
iv ivα V + α V - β V '' = K4 42 2 1 21
V '' = K4 1
iv iv iv iv '' ''A V + A V + A V + A V + A V + A V +4 4 51 1 2 2 3 3 1 6 2
'' ''+A V + A V + A V = µK7 4 43 8 9 3
iv iv iv iv '' ''B V + B V + B V + B V + B V + B V +4 4 51 1 2 2 3 3 1 6 2
'' ''+B V + B V + B V = ε K7 4 43 8 9 3 2
Flexural-
distortion
al
equilibriu
m
qiv 2V =2
Ea22
iv 2 4V - α V '' + 4β V = K3 3 3 3
γ
iv ivα V + α V - β V '' = K1 2 2 3 1 3 3
iv ivα V + α V - β V '' - V = -K4 43 2 3 2 3 1 3
113
Torsional
-
distortion
al
equilibriu
m
β V '' - γ V = -K41 1 3 1
iv ivV + α V + β V '' = K4 43 1 2 2
γβ V '' - V = K41 1 3 1
iv ivV + α V - β V '' = K4 43 2 2 2
iv iv iv iv '' ''C V + C V + C V + C V + C V + C V +4 4 51 1 2 2 3 3 1 6 2
'' ''+C V + C V + C V = τ K7 4 43 8 9 3 2
iv iv iv iv ''D V + D V + D V + D V + D V = 04 4 51 1 2 2 3 3 3
Flexural-
torsional-
distortion
al
equilibriu
m
γ
ivV = K2 0
β V '' - V = - K41 1 3 1
iv ivV - α V + β V '' = K4 43 1 2 2
β γβ V '' + β V '' + V '' - V = - K4 5 42 3 6 1 3 3
(2)γ
iv iv ivα V + α V + α V - β V '' - β4 5 4 72 3 6 2 8 3
-β V '' + V = K 4 49 2 3
(3)γ
iv iv ivα V +α V +α V -β V ''-β V ''7 42 8 3 9 10 2 11 3-β V ''+ V = K 4 512 3 3
114
CHAPTER SIX
TORSIONAL – DISTORTIONAL ANALYSIS OF BOX GIRDER SECTIONS 6.1 PREAMBLE
In this chapter the solutions of the differential equations of equilibrium derived in
chapter five are obtained and used to analyze the various cross sections of box
girder structures whose Vlasov’s coefficients were evaluated in chapter four.
Specifically, the following cross sections are considered.
(a) Single cell doubly symmetric section
(b) Multi cell doubly symmetric section
(c) Single cell mono-symmetric section
(d) Double cell mono-symmetric section
Each of these sections was analyzed for
(i) flexural- torsional interaction
(ii) flexural-distortional interaction
(iii) torsional-distortional interaction
(iv) flexural-torsional-distortional interaction.
It should be pointed out here that closed form solutions of the differential
equations require knowledge of the boundary conditions of the bridge girder. To this
effect, a three span simply supported bridge deck structure, Fig.6.2.1, was
considered. The torsional loads were obtained from AASHTO-LRFD(1998)
specifications. The dead load on the structure is not considered as they do not
contribute to torsional load (except where there is eccentricity in construction). The
flexural displacements concern strain mode 2 (horizontal bending) which has
interaction with strain modes 3 and 4. The vertical bending displacements concern
strain mode 1 which has no interaction whatsoever with other strain modes (except in
non symmetric sections) and is therefore not considered in the analysis.
Some of the governing equations of equilibrium were integrated by method of
trigonometric series with accelerated convergence, and others by Laplace transform
and traditional methods.
115
6.2 EVALUATION OF LOADS
Live loads are considered according to AASHTO-LRFD (AASHTO 1998),
following the HL-93 loading. Uniform lane load of 9.3N/mm distributed over a
3000mm width plus tandem load of two 110 KN axles. The loads are positioned at
the outermost possible location to generate the maximum torsional effects, Fig.6.2.2b
and Fig.6.2.4b. The tandem load is swept along the longitudinal direction of the
bridge axis, while the uniform lane load is positioned on each span separately. A
three span simply supported box girder bridge, Fig.6.2.1, is considered in the
analysis.
6.2.1 Doubly-symmetric sections
50m 50m 50m
Fig 6.2.1 Three Span Simply Supported Box Girder Bridge
e2 1500
600
100
110KN CL
(b) Tandem load (one axle shown)
Fig.6.2.2 Positioning of live loads for single lane bridge (AASHTO 1998)
3000mm 600mm
100mm
e1
CL
b
9.3KN/m
(a) Lane load
e2 2200m
P=138KN CL
b
h
(a) Torsional Load
P/2 CL
b
h
(b) Bending Component
P/2
116
6.2.2 Mono-symmetric sections
Fig.6.2.3 Components of bridge loads for doubly symmetric section
b
h=3m
(e) Pure Rotation
Tb,rot
Mq =
2h
Tt,rot
Mq =
2h
Tw,rot
Mq =
2b
Tw,rot
Mq =
2b
b
h=3m
(c) Torsionl Component
MT/b MT/b
b
h=3m
(d) Distortional
Component
Tw,dist
Mq =
2b
Tb,dist
Mq =
2h
Tw,dist
Mq =
2b
Tt,dist
Mq =
2h
1800
(a) Lane load (b) Tandem load (one axle shown)
600
b
110KN
3000 9.3 N/mm
CE
b 1800
Fig. 6.2.4 Positioning of Live Loads for Single Lane Bridge
2745 2745
2200 e 138KN
(C) Lane load + tandem load
2200 e 138KN
(a) Bridge eccentric Load
3.05m ⇒⇒⇒⇒
69KN 69KN
b
h = 3.05m
(b) Bending component (c) Torsional component
27.52KN
b
h = 3.05m ++++
Fig. 6.2.5 Components of Bridge Eccentric Load
117
Table 6.1 Distortional and rotational forces in the plates
S/No Sections Distortional and rotational forces (kN)
t,distq b,distq w,distq t,rotq b,rotq w,rotq
1 Single cell
doubly
symmetric
33.58 33.58 13.76 33.58 33.58 13.76
2 Multi-cell
doubly
symmetric
52.90 52.90 17.63 52.90 52.90 17.63
3 Single cell
mono-
symmetric
11.01 22.02 10.70 44.05 20.00 21.40
4 Doule cell
mono-
symmetric
11.01 22.02 10.70 44.05 20.00 21.40
t,distq = top flange plate distortional force, b,distq = bottom flange plate distortional force
w,distq = distortional force in the web plate, t,rotq = top flange plate rotational force
b,rotq = bottom flange plate rotational force, w,rotq = rotational force in the web plate
The distortional and rotational forces in the box girder plate elements are as shown in
Table 6.1. Details of the computations are shown in Appendix Four.
(c) Distortional component
(warping)
wDq wDq
tDq
bDq
Fig. 6.2.6 Components of the Torsional Load
(b) Pure torsional
component (rotation)
wRq
tRq
bRq
wRq
(a) Torsional load
27.52KN
b
h = 3.05m
27.52KN
⇒⇒⇒⇒ ++++
118
Table 6.2 Cross sectional dimensions and bridge loads
S/No Sections Dimensions Box girder loads
b (m) h (m) MT (kNm) 2q (kN) 3q (kN) 4q (kN)
1 Single cell
doubly
symmetric
7.32 3.05 201.48 0.00 620.45 1462.62
2 Multi-cell
doubly
symmetric
9.00 3.05 317.40 0.00 6495.23 2856.33
3 Single cell
mono-
symmetric
7.32 3.05 201.45 0.00 157.16 1446.51
4 Doule cell
mono-
symmetric
7.32 3.05 201.45 0.00 157.16 1446.51
MT = torsional moment, 2q = load causing bending about minor axis
3q = distortional load, 4q = rotational load
The computed box girder loadings are given in Table 6.2. Details of the computation
are given in Appendix Four
6.3 ANALYSIS OF DOUBLY SYMMETRIC SECTIONS
6.3.1 Flexural- torsional analysis: single cell doubly symmetric section
The governing differential equations for flexural-torsional equilibrium are
iv 22
22
qV =
Ea (6.3.1)
iv 44
44
qV = -
r G (6.3.2)
We note that for doubly symmetric section there is no interaction between flexure and
torsion. Consequently, eqns. (6.3.1 & 6.3.2) are independent of one another.
For single cell doubly symmetric section we have, from Table 4.1(a) and (e),
22 44a = 27.994, r = 22.032
For concrete we have 9 2 9 2G = 9.6X10 N/m ; E = 24x10 N/m , υ= 0.25
119
From analysis of bridge loads in Table 6.2, we obtained that q2 = 0KN;
4q = 1462.63KN
Thus, the solution to eqn. (6.3.1) is V2(x) = 0
Let 41
44
qK =
r G, Eqn. (6.3.2) can be written as
1 1v '' = -K (6.3.3)
The Laplace transform of the terms in eqn. (6.3.3) are,
{ }iv 4 3 2
1 1 0 1 2 3v = s v - s C - s C - sC -C
ℓ{ } 11
KK =
s
where 0 1 1 1 2 1 3 1C = v (0); C = v '(0); C = v ''(0); C = v '''(0)
Substituting into eqn. (6.3.3) we obtain;
2
4 0 3 1s v - sC -C -K /s = 0
2
4 1 0 1s v = +sC +C -K /s
2 3 14 2
sC +C -K /sv =
s
⇒ 1 11 0 2
C Kv = C + -
s s (6.3.4)
Taking the inverse transform of eqn (6.3.4) we obtain,
2
1 0 1 1
x xv = C +C -K
1! 2!
211 0 1
Kv = C +C x - x
2 (6.3.5)
where , 0 1 2 3C , C , C , C are constants determined by the boundary conditions.
Boundary conditions
⇒4 0V (0) = 0; C = 0 ; ⇒2
44 1
K LV (L) = 0; C L - = 2
2 or 4
1
K LC =
2
Substituting into eqn. (6.3.5) gives
2
1 14
K Lx K xV = -
2 2
For L = 50m we have
120
2
4 1 1V = 25K x -0.5K x ; (a)
Where, 3
-641 9
44
q 1462.63 *10K = - = = 6.9153 *10
r G 22.032* 9.6 *10
Substituting into (a) above we obtain the expression for variation of torsional
displacement along the length of the girder. Thus,
(6.3.6)
Table 6.3 Variation of torsional
displacement along the girder length
(single cell doubly symmetric section)
Distance x,
along the
length of
the girder
(m)
Torsional displacement ,
V4(x) x 10 -3(m)
0 0.000
5 -0.778
10 -1.400
15 -1.816
20 -2.075
25 -2.161
30 -2.075
35 -1.816
40 -1.400
45 -0.778
50 -0.000
-6 2 -4
4V (x) = 3.4577 *10 x -1.729 *10 x
121
6.3.2 Flexural-Torsional Analysis: Multi- cell doubly symmetric section
The governing differential equations for flexural-torsional equilibrium are the
same as for single cell doubly symmetric sections but the coefficients are different.
Thus we have,
iv 22
22
iv 44
44
qV =
Ea
qV = -
r G
(6.3.7)
22 44a = 27.994, r = 22.032; 9 2 9 2G = 9.6X10 N/m ; E = 24x10 N/m , υ= 0.25
From analysis of bridge loads we obtained that 2q = 0KN; 4q = 1462.63KN
Applying the same procedure as in single cell doubly symmetric sections we obtain
V2(x) = 0
2
4 2 2V (x) = 25K (x) -0.5K x
where , 3
-642 9
44
q 2856.33 *10K = - = - - 8.4768 *10
r G 35.100 * 9.6 *10
Therefore,
-4 -6 2
4V (x) = - 2.119 *10 x + 4.239 *10 x
Table 6.4 Variation of torsional
displacement along the girder length
(multi- cell doubly symmetric section)
Distance x,
along the
length of
the girder
(m)
Torsional displacement
V4(x) x 10 -3(m)
0 0.000
5 -0.954
10 -1.695
15 -2.225
20 -2.542
25 -2.648
122
30 -2.542
35 -1.225
40 -1.695
45 -0.954
50 -0.000
The variation of torsional displacement V4, along the length of the girder is shown in
Fig.6.3.2
6.3.3 Flexural-distortional analysis: single cell doubly symmetric section
The governing equations for flexural-distortional equilibrium for doubly symmetric
section are
( )
( )
a
b
iv 22
22
iv 2 4
3 3 3 3 3 3
qV =
Ea
V -α V ''+ 4β V =K
(6.3.8)
where
-3
33 22 33 33a = 6.721, a = 27.994, c = 3.736, s =1.062*10
υ9 2 9 2E = 24 *10 N/m , G = 9.6 *10 N/m , = 0.25
From analysis of bridge loading 2 3q = 0.0KN and q = 620KN
Consequently, V2 (x) = 0
Let the solution for eqn. (6.3.8b) be sought in the form; λx3V (x) = e
∴ λx3V '(x) = λe , 2 λx
3V ''(x) = λ e , iv 4 λx3V (x) = λ e (6.3.9)
Substituting eqn. (6.3.9) into eqn. (6.3.8) we obtain the auxiliary equation for
homogeneous condition as follows.
( )λx 4 2 2 4e λ -α λ + 4β = 0 ⇒ 4 2 2 4λ -α λ + 4β = 0
Let 2υ = λ ∴ 2 2 4υ -α υ+ 4β = 0
⇒2 4 4
1,2
α ± α -16β=
2υ ⇒
2 44
1
α αυ = + - 4β
2 4,
2 44
2
α αυ = - - 4β
2 4
∴2 4
4
1,2 1
α αλ = υ = ± + - 4β
2 4
123
2 4
4
3,4 2
α αλ = υ = ± - - 4β
2 4
where 2 333
33
ksα =
c, 4 33
33
s4β =
a
-32 -4333
33
ks 2.5x1.0625x10α = = = 7.11x10
c 3.736
-34 -433
33
s 1.0625x104β = = =1.581x10
a 6.721
-4 -7-4
1,2
7.11x10 5.0552x10λ = ± + -1.581x10
2 4
= -4 -7 -4± 3.555x10 + 1.2638x10 -1.581x10 4 43.555 10 1.5797 10x i x− −= ± +
Similarly, -4 -4
3,4λ = ± 3.555x10 - i 1.5797x10
Let -4a = 3.555x10 and -4 -2b = 1.5797x10 =1.2569x10
∴ ( )1
21λ = a+ib ( )
1
22λ = - a+ib ( )
1
23λ = a - ib ( )
1
24λ = - a - ib
Substituting in eqn. (6.3.8) we obtain
( ) ( ) ( ) ( )1 1 1 1
2 2 2 2a+ib - a+ib a-ib - a-ib
3 1 2 3 4V (x) = A e + A e + A e + A e
⇒ 3 1 2 3 4V (x) = A coshωxcosγx + A coshωxsinγx + A sinhωxcosγx + A sinhωxsinγx
where =ω real value of
1-4 2 23.555x10 +i1.2569x10
λ = imaginary value of 1
4 2 23.555 10 1.2569 10x i x− +
For λ = a+ib ,ω = Re a + ib , γ =Im a+ib
where -4a = 3.555x10 , -2b =1.2569x10
∴ a+ib =ω+iγ ⇒ ( )2 2 2a+ib = ω+iγ =ω +2iωγ - γ
( )2 2a+ib = ω - γ +2iωγ (6.3.12)
Equating equivalent terms on both sides of eqn. (6.3.10) we obtain
2 2ω - γ = a ……….(a) and 2ωγ = b ……..(b)
Squaring both sides of expressions (a) and (b) and adding up we have
( )2
2 2 2 2 2 2ω - γ + 4ω γ = a +b ⇒ 4 2 2 4 2 2ω - 2ω γ + γ + 4ω γ = a+b
124
( )⇒ 4 2 2 4 2 2ω +2ω γ +γ = a +b
( ) ( )⇒2
2 2 2 2ω +γ = a +b ∴ 2 2 2 2ω +γ = a +b ……….(c)
Adding expressions (a) and (c) we obtain
2 2 22ω = a+ a +b
⇒2 2a+ a +b
ω =2
( ) ( )
=
2 2-4 -4 -23.555x10 + 3.555x10 + 1.2569x10
2
=-4 -23.555x10 +1.2572x10
= 0.08042
Subtracting (a) from (c) we obtain
2 2 22γ = a +b -a
⇒2 2a +b -a
γ =2
( ) ( )
=
2 2-4 -2 -43.555x10 + 1.2569x10 -3.555x10
2
⇒ -36.10926x10 = 0.0782γ =
Thus, ω = 0.0804 , γ = 0.0782
3 homogeneous 1 2 3 4V (x) = A coshωxcosλx + A coshωxsinγx + A sinhωxcosγx + A sinhωxsinγx
Let the particular integral be sought in the form; 3V (x) = B (6.3.13)
This solution must satisfy eqn. (6.3.8b)
i.e., iv 2 4
3 3 3 3 3 3V -α V ''+ 4β V = K
From eqn. (6.3.13) we obtain iv
3 3V '' = 0, V = 0
Substituting into eqn. (6.3.8b) we obtain
4
44β B = K ∴ 3
4
KB =
4β ⇒ ( ) 3
3 4particular
KV x =
4β
where -123 33 39
33
q qK = = = 6.1995x10 q
a E 6.721x24x10
Boundary Conditions;
V(0) = 0; V(L) = 0; V''(0) = 0;
V''(L) = 0
33 1 2 3 4 4
KV (x) = A coshωxcosλx + A coshωxsinγx + A sinhωxcosγx + A sinhωxsinγx +
4β
125
[ ]3 1 2 3 4V ''(x) = 2 -A coshωxcosλx + A coshωxsinγx - A sinhωxcosγx + A sinhωxsinγx
∴ ⇒ 31 4
KV(0) = 0 : A = -
4β (a)
⇒''
3 4V (0) = 0 : 2A = 0 ⇒ 4A = 0 (b)
0 : ⇒ 31 2 3 4
KV(L) = -20.0315A -19.632A +20.019A + = 0
4β (c)
V''(L) = 0 : ⇒ 1 2 3-38.699A - 40.032A +38.724A = 0 (d)
Solving expressions (a), (b), (c), and (d) we obtain
31 4
KA = -
4β; ; 3
2 4
KA = 0.0256
4β 3
3 4
KA = 0.9649
4β; 4A = 0
Substituting values of 1 2 3 4A , A , A , & A into eqn. (6.3.14) we obtain
[ ]33 4
KV (x) = -coshωxcosλx +0.0256coshωxsinγx +1.0258sinhωxcosγx +1
4β
or [ ]33 4
KV (x) = 1.0258sinhωxcosγx - coshωxcosλx +0.0256coshωxsinγx +1
4β
where, 4 -44β =1.581x10 and -12
3 3K = 6.1995x10 q (6.3.15)
Substituting value of 3q into eqn. (6.3.15) we obtain
-12 33
4 -4
K 6.1995 *10 * 620.45 *10= = 0.02433
4β 1.581*10
∴ [ ]3V (x) = 0.02433 1.0258 * SinhωxCosγx -CoshωxCosγx +0.0256 * CoshωxSinγx +1
(6.3.16)
where, ω = 0.0804radian; γ = 0.0782radian
Table 6.5 Variation of distortional displacement along the girder length
(Single cell doubly symmetric section)
Distance (x)
from Left
Support (m)
1+ 1.0258 x
SinhwxCosvx -CoshwxCosvx 0.0256 x
CoshwxSinvx
Distortional displacement V3(x)
0 1.000 -1.0000 0.0000 0.000 x 103 (m)
5 1.3916 -1.0002 0.0106 9.78
10 1.6503 -0.9514 0.0242 17.39
15 1.6042 -0.7049 0.0429 22.92
126
20 1.0167 -0.0176 0.0665 25.93
25 -0.4090 1.4238 0.0902 26.88
30 -2.9725 3.9353 0.1028 25.93
35 -6.8345 7.6926 0.0843 22.93
40 -11.7639 124830 -0.0043 17.39
45 -16.7543 17.3327 -0.1759 9.79
50 -19.5351 20.0315 -0.4956 0.00
6.3.4 Flexural distortional analysis: multi- cell doubly symmetric section
The governing equations of flexural-distortional equilibrium for doubly symmetric
sections are
( )
( )
a
b
iv 22
22
iv 2 4
3 3 3 3 3 3
qV =
Ea
V -α V ''+ 4β V =K
(6.3.17)
where 2 433 33 33 3 3
33 33 33
ks s qα = ; 4β = ; K =
c a a E
For multi-cell doubly symmetric section we have;
c-2
33 33 33a =127.752, = 55.727, s = 4.338 *10
υ9 2 9 2E = 24 *10 N/m , G = 9.6 *10 N/m , = 0.25
From analysis of bridge loading 2 3q = 0.0KN and q = 6495.24KN
Consequently, V2 (x) = 0
Let the solution for eqn. (6.3.17b) be sought in the form; λx3V (x) = e (6.3.18)
∴ λx3V '(x) = λe , 2 λx
3V ''(x) = λ e , iv 4 λx3V (x) = λ e
Substituting derivatives of eqn. (6.3.18) into eqn. (6.3.17) we obtain the
characteristic equation for homogeneous condition as follows.
( )λx 4 2 2 4e λ -α λ + 4β = 0 ⇒ 4 2 2 4λ -α λ + 4β = 0
Let 2υ = λ ∴ 2 2 4υ -α υ+ 4β = 0
⇒2 4 4
1,2
α ± α -16β=
2υ ⇒
2 44
1
α αυ = + - 4β
2 4,
2 44
2
α αυ = - - 4β
2 4
127
∴2 4
4
1,2 1
α αλ = υ = ± + - 4β
2 4
2 4
4
3,4 2
α αλ = υ = ± - - 4β
2 4
where 2 333
33
ksα =
c, 4 33
33
s4β =
a
-32 -3333
33
ks 2.5 * 4.338 *10α = = =1.946 *10
c 55.727
-24 -433
33
s 4.338 *104β = = = 3.396 *10
a 127.752
-3 -6-4
1,2
1.946 *10 3.787 *10λ = ± + -3.396 *10
2 4
4 49.73*10 3.3865*10i− −= ± +
Similarly, -4 -4
3,4λ = ± 9.73 *10 - i 3.3865 *10
Let -4a = 9.73 *10 and -4 -2b = 3.3865 *10 =1.840210
∴ ( )1
21λ = a+ib ; ( )
1
22λ = - a+ib ; ( )
1
23λ = a - ib ; ( )
1
24λ = - a - ib
Substituting in eqn. (6.3.18) we obtain
( ) ( ) ( ) ( )1 1 1 1
2 2 2 2a+ib - a+ib a-ib - a-ib
3 1 2 3 4V (x) = A e + A e + A e + A e
⇒ 3 1 2 3 4V (x) = A coshωxcosγx + A coshωxsinγx + A sinhωxcosγx + A sinhωxsinγx
where =ω real value of
1-3 2 29.73 *10 +i1.8402*10
λ = imaginary value of
1-3 2 29.73 *10 +i1.8402*10
For λ = a+ib ,ω = Re a + ib , γ =Im a+ib
where -4a = 9.73 *10 and -2b =1.840210
∴ a+ib =ω+iγ ⇒ ( )2 2 2a+ib = ω+iγ =ω +2iωγ - γ
( )2 2a+ib = ω - γ +2iωγ (6.3.19)
Equating equivalent terms on both sides of eqn. (6.3.19) we obtain
2 2ω - γ = a ……….(a) and 2ωγ = b ……..(b)
128
Squaring both sides of expressions (a) and (b) and adding up we have
( )2
2 2 2 2 2 2ω - γ + 4ω γ = a +b ⇒ 4 2 2 4 2 2ω - 2ω γ + γ + 4ω γ = a+b
( )⇒ 4 2 2 4 2 2ω +2ω γ +γ = a +b
( ) ( )⇒2
2 2 2 2ω +γ = a +b ∴ 2 2 2 2ω +γ = a +b ……….(c)
Adding expressions (a) and (c) we obtain
2 2 22ω = a+ a +b ⇒2 2a+ a +b
ω =2
= -3 -2ω 4.865 *10 +1.04076 *10 = 0.12358
Subtracting (a) from (c) we obtain
2 2 22γ = a +b -a
⇒2 2a +b -a
γ =2
( ) ( )
=
-5 -4 -39.467 *10 + 3.386 *10 -9.73 *10
2
34.865*10−= −-21.04076 *10 = 0.074449
Thus, ω = 0.12358 , γ = 0.074449
3 homogeneous 1 2 3 4V (x) = A coshωxcosλx + A coshωxsinγx + A sinhωxcosγx + A sinhωxsinγx
Let the particular integral be sought in the form; 3V (x) =B (6.3.20)
This solution must satisfy eqn. (6.3.17b)
i.e, iv 2 4
3 3 3 3 3 3V -α V ''+ 4β V =K
From eqn. (6.3.20) we obtain iv
3 3V '' = 0, V = 0
Substituting into eqn. (6.3.17b) we obtain
4
44β B =K ∴ 3
4
KB =
4β ⇒ ( ) 3
3 4particular
KV x =
4β
where -133 33 39
33
q qK = = = 3.2615 *10 q
a E 127.752* 24 *10
Boundary Conditions;
V(0) = 0; V(L) = 0; V''(0) = 0; V''(L) = 0
33 1 2 3 4 4
KV (x) = A coshωxcosλx + A coshωxsinγx + A sinhωxcosγx + A sinhωxsinγx +
4β
(6.3.21)
129
∴ ⇒ 31 4
KV(0) = 0 : A = -
4β (a)
⇒''
3 4V (0) = 0 : 2A = 0 ⇒ 4A = 0 (b)
⇒ + 31 2 3 4 4
KV(L) = 0 : 240.8736A +15.6631A +240.744A +15.663A = 0
4β (c)
V''(L) = 0 : ⇒ 1 2 3 4-15.663A +240.744A -15.66314A +240.74665A = 0 (d)
Solving expressions (a), (b), (c), and (d) we obtain
31 4
KA = -
4β; -4 3
2 4
KA = - 2.34 *10
4β; 3
3 4
KA = 0.9964
4β; 4A = 0
Substituting values of 1 2 3 4A , A , A , & A into eqn. (6.3.21) we obtain
-43
3 4
KV (x) = -CoshωxCosλx - 2.34 *10 * CoshωxSinγx +1.0258SinhωxSosγx +1
4β
or [ ]33 4
KV (x) = 0.9964SinhωxCosγx -CoshωxCosλx +0.063576CoshωxSinγx +1
4β
where, 4 -44β = 3.396 *10 , -13
3 3K = 3.2615 *10 q and 3q = 6495.24KN
-13 3-33
4 -4
K 3.2615 *10 * 6495.24 *10= = 6.238 *10
4β 3.396 *10
∴ [ ]V (x) = 0.00624 0.9964 * SinhωxCosγx - CoshωxCosγx + 0.06358 * CoshωxSinγx +13
(6.3.22)
where, ω = 0.12358 radian ; γ = 0.074449 radian
Table 6.6 Variation of distortional displacement along the girder length
(Multi- cell doubly symmetric section)
Distance (x)
from Left
Support (m)
1+ 1.0258 x
SinhwxCosvx -CoshwxCosvx 0.0256 x
CoshwxSinvx
Distortional displacement V3(x)
0 1.000 -1.0000 0.0000 0.000 x 103 (m)
5 1.6556 -1.1970 -0.0001 2.863
10 2.5695 -1.8657 -0.0002 4.390
15 4.1016 -3.2694 -0.0003 5.195
20 6.855 -5.9609 -0.0013 5.579
25 11.915 -11.000 -0.0030 5.851
30 21.2729 -20.386 -0.0070 5.403
35 38.6125 -37.7616 -0.0143 5.220
130
40 70.7588 -70.0180 -0.0303 4.433
45 130.3628 -129.8341 -0.0632 2.900
50 240.877 -240.7468 -0.1302 0.000
6.3.5 Torsional - distortional analysis: single cell doubly symmetric section
The governing equations of equilibrium are
γ1 4 1 3 1β V '' - V = -K (6.3.23)
iv iv
3 1 4 2 4 2V +α V +β V '' =K (6.3.24)
γ
2
34 43 33 44 3344 441 1 2 1
43 33 43 33 43 33
r c -b r ksr rα = β = - ; β = ; =
c c c ka c c
3 33 441 2
33 43 33 43
q b qqK = - ; K =
c G c G a c E
-3
33 33 33 43 44 33a = 6.721, b = c = 3.736; c = 3.732; r = 22.032; s =1.062*10
υ 9 2 9 2= 0.25; E = 24 *10 N/m ; G = 9.6 *10 N/mm
From analysis of loads in section 6.2 we obtained that
3 4q = 620KN; q =1462.62KN .
The coefficients for the governing equations are evaluated using the expressions
above. Thus,
γ -4
1 1 1α = 5.904; β = -4.905; β = -1.0905; = 7.106 *10
Substituting the coefficients into eqns.(6.3.23 & 6.3.24) we obtain
-4
4 3 1 4.905V '' + 7.106 *10 V = K (6.3.25)
iv iv
3 4 4 2V +5.904V -1.0905V '' =K
The solution of these coupled differential equations is obtained using the method of
trigonometric series with accelerated convergence
Boundary conditions; 3 3 3 3V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
4 4 4 4V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
We seek the solution in the form;
3 3 3
4 4 4
V (x) = f (x)+g (x)
V (x) = f (x)+g (x) (6.3.26)
131
Substituting eqn. (6.3.26) into eqn. (6.3.25) we obtain
-4 -4
4 3 4 14.905f ''+7.106 *10 f + 4.905g ''+7.106 *10 g =K
⇒ -4 -4
4 3 4 3 14.905g ''+7.106 *10 g = - 4.905f '' - 7.106 *10 f +K
⇒ -4
4 3 14.905g ''+7.106 *10 g =P (x)
where, -4
1 4 3 1P (x) = - 4.905f '' - 7.106 *10 f +K
Similarly,
iv iv iv iv
3 4 4 3 4 4 2f +5.904f -1.0905f ''+g +5.904g -1.0905g '' =K
⇒ iv iv iv iv
3 4 4 3 4 4 2g +5.904g -1.0905g '' = - f - 5.904f +1.0905f ''+K
⇒ iv iv
3 4 4 2g +5.904g -1.0905g '' =P (x)
where, iv iv
2 3 4 4 2P (x) = - f - 5.904g +1.0905g ''+K
∴ ( )
( )
a
b
-4
1 4 3 1
iv iv
2 3 4 4 2
P (x) = - 4.905f '' - 7.106 *10 f + K
P (x) = - f - 5.904g +1.0905g ''+ K (6.3.27)
We seek the complimentary function 3g (x)and 4g (x) in the form of sine series as
follows.
∞
∞
∑
∑
n3 3n
n=1
n4 4n
n=1
α xg (x) = a sin
L
α xg (x) = a sin
L
(6.3.28)
where nα = nπ
The auxiliary function can be sought in terms of algebraic function as follows.
2 3
3 0 1 2 3f (x) = A + A x + A x + A x (a)
2 3
4 0 1 2 3f (x) =B +B x +B x +B x (b) (6.3.29)
The constants 0 0 2 2A , B , A & B are easily defined using some of the boundary
conditions.
i.e. ⇒3 0V (0) = 0 A = 0
∑ n3 2 3 3n
α xV ''(x) = 2A +6A x - a sin
L
⇒3 2V ''(0) = 0 2A = 0 or 2A = 0
⇒4 0V (0) = 0 B = 0
∑ n4 2 3 4n
α xV ''(x) = 2B +6B x - a sin
L
132
⇒4 2V ''(0) = 0 2B = 0 or 2B = 0
Substituting values of 0, 0 1 1A B , A & B in eqn. (6.3.29) we obtain;
3
3 1 3f (x) = A x + A x (a)
3
4 1 3f (x) =B x +B x (b) (6.3.30)
The constants 1 3 1 3A , A , B & B are obtained using the remaining boundary
conditions. Thus, 3 3V (L) = f (L) = 0 ⇒ 3
1 3A L + A L = 0
3 3V ''(L) = f ''(L) = 0 ; ⇒ 36A L = 0 ∴ 3A = 0 and 1A = 0
4 4V (L) = f (L) = 0 ⇒ 3
1 3B L +B L = 0
4 4V ''(L) = f ''(L) = 0 ; 36B L = 0 ∴ B3 = 0 and B1= 0
Hence, 3 3
4 4
f (x) =P (x) = 0
f (x) =P (x) = 0 } ⇒
3 3
4 4
V (x) = g (x)
V (x) = g (x) (6.3.31)
Consequently eqn. (6.3.26) becomes
0=-4
4 14.905g ''+7.106 *10 g-K (a)
0=iv iv
3 4 4 2g +5.904g -1.0905g '' -K (b) (6.3.32)
But ∞
∑ n3 3n
n=1
α xg (x) = a sin
L (a)
and ∞
∑ n4 4n
n=1
α xg (x) = a sin
L (b)
∞
∑ 2 n3 3n n
n=1
α xg ''(x) = - a α sin
L (c)
∞
∑iv 4 n3 3n n
n=1
α xg (x) = a α sin
L (d)
∞
∑ 2 n4 4n n
n=1
α xg ''(x) = - a α sin
L (e)
∞
∑iv 4 n4 4n n
n=1
α xg (x) = a α sin
L (f) (6.3.33)
Substituting eqn. (6.3.33) into eqn. (6.3.32) we obtain;
∑ ∑
∑ ∑ ∑
2 -4 n4n n 3n 1
4 4 2 n3n n 4n n 4n n 2
α x-4.905 a α +7.106 *10 a sin =K
L
α xa α +5.904 a α +1.0905 a α sin =K
L
(6.3.34)
Equations (6.3.34) will always be satisfied if
133
( )α 2 -4
n 4n 3n 1-4.905 a +7.106 *10 a =K (6.3.35)
and ( )4 4 2
n 3n n n 4n 2α a + 5.904α +1.0905α a =K (6.3.36)
From eqn. (6.3.35), we have
-4
3n14n 2 2
n n
7.106 *10 aKa = - +
4.905α 4.905α (6.3.37)
For n = 1, ,2 -3 4 -5
n n n
π πα = = , α = 3.9478 *10 α =1.5585 *10
L 50
∴-4
-231 141 31 1-2 -2
7.106 *10 α Ka = - = 3.6697 *10 a - 51.642K
1.9364 *10 1.9364 *10
Substituting into eqn. (6.3.36) we have
( )-5 -3 -2
31 1 21.5585 *10 a + 4.397 *10 3.6697 *10 -51.642K =K
⇒
⇒
-4
31 2 1
31 2 1
1.7694 *10 a =K +0.22707K
a = 5651.63K +1283.32K
⇒ 41 2 1a = 207.397K - 4.623K
where = -53 41
33 43
q qK = - -2.2856 *10
c G c G
-633 42
33 43
b qK = = 8.634 *10
a c E
∴ -2
31a =1.946 *10 , -3
41a =1.896 *10
Substituting values of a31, a32 and eqn. (6.3.28) into eqn. (6.3.31) we obtain
-2
3
-3
4
πxV (x) =1.946 *10 Sin
50
πxV (x) =1.896 *10 Sin
50
Table 6.7 Variation of torsional and distortional
displacements along the length of the girder
(Single cell doubly symmetric section)
Distance (x)
from Left
Support (m)
πxSin
50
Distortional
displacement
V3(X)
X 10-3
(m)
Torsional
displacement
V4(X)
X 10-3
(m)
0 0.000 0.00 0.000
134
5 0.309 6.013 0.586
10 0.588 11.442 1.115
15 0.809 15.743 1.534
20 0.951 18.506 1.803
25 1.000 19.460 1.896
30 0.951 18.506 1.803
35 0.809 15.743 1.534
40 0.588 11.442 1.115
45 0.309 6.013 0.586
50 0.000 0.000 0.000
6.3.6 Torsional-distortional analysis: multi cell doubly symmetric section
The governing equations of equilibrium are
γ1 4 1 3 1β V '' - V = -K (6.3.38)
iv iv
3 1 4 2 4 2V +α V +β V '' =K
; γ
2
34 43 33 44 3344 441 1 2 1
43 33 43 33 43 33
r c -b r ksr rα = β = - ; β = ; =
c c c ka c c
3 4 22 41 2
33 43 33 43
q q b qK = - ; K =
c G c G a c E
33 33 33 43 43a =127.752, b = c = 55.752; c = r = 3.732;
-2
44 33r = 35.100; s = 4.338 *10
υ 9 2 9 2= 0.25; E = 24 *10 N/m ; G = 9.6 *10 N/mm
From analysis of loads in section 6.2 we obtained that;
3 4q = 6495KN; q = 2856.33KN .
The coefficients for the governing equations are evaluated using the expressions
above. Thus,
γ -3
1 1 2 1α =1.418; β = - 0.973; β = - 0.1698; =1.946 *10
-7 -6
1 2K =1.2278 *10 ; K = 2.097 *10
Substituting the coefficients into eqns. (6.3.38) we obtain
135
-3
4 3 1
iv iv
3 4 4 2
-0.973V '' -1.946 *10 V = -K
V +1.418V -0.170V '' =K (6.3.39)
The solution of these coupled differential equations is obtained using the method of
trigonometric series with accelerated convergence
Boundary conditions; 3 3 3 3V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
4 4 4 4V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
We seek the solution in the form;
3 3 3
4 4 4
V (x) = f (x)+g (x)
V (x) = f (x)+g (x) (6.3.40)
Substituting eqns. (6.3.39) into eqns. (6.3.40) we obtain
-3 -3
4 3 4 3 1-0.973f '' -1.946 *10 f -0.973g '' -1.946 *10 g = -K
iv iv iv iv
3 4 4 3 4 4 2f +1.418f -0.170f ''+g +1.418g -0.170g '' =K
⇒ -3
4 3 1-0.973f '' -1.946 *10 f =P (x)
⇒ iv iv
3 4 4 2f +1,418f -0.170f '' =P (x)
-3
1 4 1P (x) = -0.973g '' -1.946 *10 g+K (a)
iv iv
2 3 4 4 2P (x) = g +1.418g -0.170g '' -K (b) (6.3.41)
We seek the complimentary function 3g (x)and 4g (x) in the form of sine series as
follows.
∞
∞
∑
∑
n3 3n
n=1
n4 4n
n=1
α xg (x) = a sin
L
α xg (x) = a sin
L
(6.3.42)
where nα = nπ
The auxiliary function can be sought in terms of algebraic function as follows.
2 3
3 0 1 2 3f (x) = A + A x + A x + A x (a)
2 3
4 0 1 2 3f (x) =B +B x +B x +B x (b) (6.3.43)
The constants 0 0 1 1 2 2 3 3
A , B , A , B , A , B , A & B are easily defined using the boundary conditions.
From section 6.3.5 we note that for simply supported box girder the boundary
conditions are such that all the constants are zero.
Hence, 3 3
4 4
f (x) =P (x) = 0
f (x) =P (x) = 0 } ⇒
3 3
4 4
V (x) = g (x)
V (x) = g (x)
Consequently eqn. (6.3.41) becomes
136
0
0
=
=
-3
4 3 1
iv iv
3 4 4 2
-0.973g '' -1.946 *10 g +K
g +1.418g -0.170g '' -K (6.3.45)
But ∞
∑ n3 3n
n=1
α xg (x) = a sin
L (a)
and ∞
∑ n4 4n
n=1
α xg (x) = a sin
L (b)
∞
∑ 2 n3 3n n
n=1
α xg ''(x) = - a α sin
L (c)
∞
∑iv 4 n3 3n n
n=1
α xg (x) = a α sin
L (d)
∑¥
2 n4 4n n
n=1
α xg ''(x) = - a α sin
L (e)
∞
∑iv 4 n4 4n n
n=1
α xg (x) = a α sin
L (f) (6.3.46)
Substituting eqn. (6.3.46) into eqn. (6.3.45) we obtain;
0.973
∑ ∑
∑ ∑ ∑
2 -3 n4n n 3n 1
4 4 2 n3n n 4n n 4n n 2
α xa α -1.946 *10 a sin = -K
L
α xa α +1.418 a α +0.170 a α sin =K
L
(6.3.47)
Equations (6.3.47) will always be satisfied if
( )α 2 -3
n 4n 3n 10.973 a -1.946 *10 a = - K (6.3.48)
and ( )4 4 2
n 3n n n 4n 2α a + 1.418α +0.170α a =K (6.3.49)
From eqn. (6.3.48), we have
-3
14n 3n2 2
n n
K 1.946 *10a = - + a
0.973 α 0.973 α (6.3.50)
For n = 1, 2 -3 4 -5
1 1 1
π πα = = , α = 3.9478 *10 , α =1.5585 *10
L 50
∴-3 -7
-53141 31-3 -3
1.946 *10 α 1.2346 *10a = - = 0.5066a - 3.196 *10
0.973 * 3.9478 *10 0.973 * 3.9478 *10
Substituting into eqn. (6.3.49) we have
( )-5 -4 -5 -6
31 311.5585 *10 a +6.9323 *10 0.5066a -3.196 *10 = 2.097 *10
⇒ -4 -8 -6
313.6677 *10 a = 2.2158 *10 +2.097 *10
137
3
45.779*10
3.667*10
−
−⇒ =
-6
31
2.119 *10a =
⇒ -3 -5 -3
41a = 0.5066 * 5.779 *10 -3.196 *10 = 2.896 *10
Substituting values of a31, a41 and eqn. (6.3.42) into eqn. (6.3.40) we obtain
-3
3
-3
4
πxV (x) = 5.779 *10 Sin
50
πxV (x) = 2.896 *10 Sin
50
Table 6.8 Variation of torsional and distortional
displacements along the length of the girder
( Multi cell doubly symmetric section)
Distance (x)
from Left
Support (m)
πx
Sin50
Distortional
displacement
V3(X)
X 10-3
(m)
Torsional
displacement
V4(X)
X 10-3
(m)
0 0.000 0.000 0.000
5 0.309 1.786 0.895
10 0.588 3.398 1.703
15 0.809 4.675 2.343
20 0.951 5.496 2.754
25 1.000 5.779 2.896
30 0.951 5.496 2.754
35 0.809 4.675 2.343
40 0.588 3.398 1.703
45 0.309 1.786 0.895
50 0.000 0.000 0.000
6.3.7 Flexural-torsional-distortional analysis: single cell doubly symmetric
section
The differential equation governing flexural- torsional-distortional equilibrium for
doubly symmetric sections are
138
γ
iv
2 0
1 4 1 3 1
iv iv
3 1 4 2 4 2
V =K
β V '' - V = -K
V +α V +β V '' =K
(6.3.51)
The coefficients are as follows.
22 22 22 22a = 27.994, b = c = r = 2.880
33 33 33 43 43a = 6.721, b = c = 3.737; c = r = 3.732
-3
44 33r = 22.032; s =1.062*10
υ 9 2 9 2= 0.25; E = 24 *10 N/m ; G = 9.6 *10 N/mm
2 3 4q = 0.00KN, q = 620KN, q =1462.62KN
3444 441 1
43 33 43
rr rα = = 5.904; β = - = - 4.905
c c c
47.106*10γ −=2
43 33 44 332 1
33 43 33
c -b r ks β = = -1.0905 =
ka c c
-5 -63 33 441 2
33 43 33 43
q b qqK = - = - 2.2856 *10 ; K = = 8.634 *10
c G c G a c E
Substituting the coefficients into eqns. (6.3.51 ) we obtain
∴iv
2 2V = 0 V = 0
-4
4 3 14.905V ''+7.106 *10 V =K (6.3.52)
iv iv
3 4 4 2V +5.904V -1.0905V '' =K
These coupled equations are the same as eqns. (6.3.25) in torsional-distortional
analysis of section 6.3.5., and their coefficients are the same. Their solutions are
-2
3
-3
4
πxV (x) =1.946 *10 Sin
50
πxV (x) =1.896 *10 Sin
50
(6.3.53)
Therefore, flexural-torsional-distortional analysis of single cell doubly symmetrical
section results to torsional- distortional analysis of the same section because there
are no interactions between flexure and torsion or distortion in doubly symmetric
section.
139
Table 6.9 Variation of flexural, torsional and distortional
displacements along the length of the girder ( single cell doubly
symmetric section)
Distance (x)
from Left
Support (m)
πx
Sin50
Flexural
Displacement
V2(X) X 10-3
(m)
Distortional
Displacement
V3(X) X 10-3
(m)
Torsional
displacement
V4(X)
X 10-3
(m)
0 0.000 0.000 0.000 0.000
5 0.309 0.000 10.366 0.799
10 0.588 0.000 18.480 1.460
15 0.809 0.000 22.782 1.879
20 0.951 0.000 22.859 2.016
25 1.000 0.000 19.46 1.896
30 0.951 0.000 22.859 2.016
35 0.809 0.000 22.783 1.879
40 0.588 0.000 18.480 1.479
45 0.309 0.000 10.370 0.799
50 0.000 0.000 0.000 0.000
6.3.8 Flexural-torsional-distortional analysis: multi- cell doubly symmetric
section
The differential equation governing flexural- torsional-distortional equilibrium for
doubly symmetric sections are;
γ
∴iv
2 2
1 4 1 3 1
iv iv
3 1 4 2 4 2
V = 0; V (x) = 0
β V '' - V = -K
V +α V +β V '' =K
(6.3.54)
where,
; γ
2
34 43 33 44 3344 441 1 2 1
43 33 43 33 43 33
r c -b r ksr rα = β = - ; β = ; =
c c c ka c c
3 4 22 41 2
33 43 33 43
q q b qK = - ; K =
c G c G a c E
140
It has been established in section 6.3.7 that flexural –torsional-distortional analysis of
doubly symmetric sections does not differ from their torsional-distortional analysis.
Consequently, eqns. (6.3.54) have the same solutions as eqns. (6.3.38), i.e.
-3
3
-3
4
πxV (x) = 5.779 *10 Sin
50
πxV (x) = 2.896 *10 Sin
50
(6.3.55)
6.4 ANALYSIS OF MONO- SYMMETRIC SECTIONS 6.4.1 Flexural-torsional analysis: single cell mono symmetric section The governing equations of equilibrium are
iv iv
1 2 2 4 1 4 1
4 2
α V +α V - β V '' = - K
V '' =K (6.4.1)
where,
1 22 42 2 22 44α = ka c ; α = ka r
41 22 44 24 42 1 22
qβ = b r - c c ; K = b
G
42 2 22 42
42 24 22 44 42 24 22 44
c q c qK = - +
c r - c r G c r - c r G
The relevant coefficients are
22 22 22 22
24 42 24 42
-4 -4
44 33
a = 25.05; b = c = r = 2.982
c = c = r = r = -2.515
r =14.616; s = 0.261* 6.667 *10 =1.740 *10
2 4
9 2 9 2
q = 0.00KN; q =1446.505KN
E = 24 *10 N/m ; G = 9.6 *10 N/m , k = 2.5
The coefficients of the governing equations are
1 2 1α = -157.502; α = 915.327; β = 49.910
-4 -5
1 2K = 4.4932*10 ; K = -1.206 *10
Substituting the coefficients into eqns. (6.4.1 ) we obtain
iv iv -4
2 4 4
-5
4
-157.502V +915.327V - 49.910V '' = 4.4932*10
V '' = -1.206x10 (6.4.2)
Integrating by method of trigonometric series with accelerated convergence we
have,
Interval= span L=50m. ∴ ≤ ≤0 x 50
141
Boundary conditions; 3 3 3 3V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
4 4 4 4V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
We seek the solution in the form;
3 3 3
4 4 4
V (x) = f (x)+g (x)
V (x) = f (x)+g (x) (6.4.3)
Substituting eqn. (6.4.3) into eqn. (6.4.2) we obtain
iv iv iv iv -4
2 4 4 2 4 4
-5
4 4
-157.502f +915.327f - 49.910f '' -157.502g +915.327g - 49.910g '' = 4.4932*10
f ''+g '' = -1.206 *10
⇒iv iv
2 4 4 1
4 2
-157.502f +915.327f - 49.910f '' =P (x)
f '' =P (x)
iv iv -4
1 2 4 4
-5
2 4
P (x) = -157.502g +915.327g - 49.910g '' - 4.4932x10
P (x) = g ''+1.206 *10 (6.4.4)
We seek the complimentary function 2g (x)and 4g (x) in the form of sine series as
follows.
∞
∞
∑
∑
n2 2n
n=1
n4 4n
n=1
α xg (x) = a sin
L
α xg (x) = a sin
L
(6.4.5)
where nα = nπ
The auxiliary function can be sought in terms of algebraic function as follows.
2 3
3 0 1 2 3f (x) = A + A x + A x + A x (a)
2 3
4 0 1 2 3f (x) =B +B x +B x +B x (b) (6.4.6)
The constants 0 0 1 1 2 2 3 3A , B , A , B , A , B , A & B are easily defined using the boundary
conditions.
From section 6.3.5 we note that for simply supported span the boundary conditions
are such that all the constants are zero.
Hence, 2 1
4 2
f (x) =P (x) = 0
f (x) =P (x) = 0 } ⇒
2 2
4 4
V (x) = g (x)
V (x) = g (x) (6.4.7)
Consequently eqn. (6.4.4) becomes
0=iv iv -4
2 4 4
-5
4
-157.502g +915.327g - 49.910g '' - 4.4932*10
g ''+1.206 *10 = 0 (6.4.8)
142
But ∞
∑ n2 2n
n=1
α xg (x) = a sin
L (a)
and ∞
∑ n4 4n
n=1
α xg (x) = a sin
L (b)
∞
∑ 2 n2 2n n
n=1
α xg ''(x) = - a α sin
L (c)
∞
∑iv 4 n2 2n n
n=1
α xg (x) = a α sin
L (d) (6.4.9)
∑¥
'' 2 n4 4n n
n=1
α xg (x) = - a α sin
L (e)
∞
∑iv 4 n4 4n n
n=1
α xg (x) = a α sin
L (f)
Substituting eqn. (6.4.9) into eqn. (6.4.8) we obtain;
∑ ∑ ∑
∑
4 4 2 -4nn 2n n 4n n 4n
2 -5nn 4n
α x-157.502 α a +915.327 α a + 49.910 α a sin = 4.4932*10
L
α x- α a Sin = -1.206 *10
L
(6.4.10)
Equations (6.4,10) will always be satisfied if
( ) ( )+4 4 2 -4
n 2n n n 4n-157.502α a 915.327α + 49.910α a = 4.4932*10 (6.4.11)
and ( )− 2 -5
n 4nα a = -1.206x10 ⇒-5
4n 2
n
1.206 *10a =
α (6.4.12)
For n = 1, 2 4
1 1 1α =π, α = 9.8696 α = 97.409
∴-5
-3
41 -3
1.206 *10a = = 3.055 *10
3.9478 *10
Substituting into eqn.(6.4.11) we have
( )
⇒
-3 -3 -4
21
-4-2
21 -3
-2.4547 *10 a +0.2113 3.055 *10 = 4.4932*10
-1.962*10a = = 7.993 *10
-2.4546 *10
Substituting values of a21, a41, and eqn. (6.4.5) into eqn. (6.4.7) we obtain
-3
2
-3
4
πxV (x) = 79.73 *10 Sin
50
πxV (x) = 3.055 *10 Sin
50
143
Table 6.10 Variation of flexural and torsional
displacements along the length of the girder
(Single cell mono symmetric section)
Distance (x)
from Left
Support (m)
πxSin
50
Flexural
displacement
V2(x) x 10-3
(m)
Torsional
displacement
V4(x) x 10-3
(m)
0 0.000 0.000 0.000
5 0.309 24.637 0.944
10 0.588 46.881 1.796
15 0.809 64.502 2.471
20 0.951 75.823 2.905
25 1.000 79.73 3.055
30 0.951 75.823 2.905
35 0.809 64.502 2.471
40 0.588 46.881 1.796
45 0.309 24.637 0.944
50 0.000 0.000 0.000
6.4.2 Flexural-torsional analysis: double cell mono symmetric section The governing equations of equilibrium are
iv iv
1 2 2 4 1 4 1
4 2
α V +α V -β V '' =K
V '' =K (6.4.13)
where,
1 22 42 2 22 44α = ka c ; α = ka r
41 22 44 24 42 2 22
qβ = b r - c c ; K = b
G
42 2 22 41
42 24 22 44 42 24 22 44
c q c qK = - +
c r - c r G c r - c r G
The relevant coefficients are
22 22 22 22
24 42 24 42
-4 -4
44 33
a = 25.073; b = c = r = 2.982
c = c = r = r =1.136
r =14.485; s = 0.723 * 6.667 *10 = 4.82*10
144
2 4
9 2 9 2
q = 0.0KN; q =1446.505KN
E = 24 *10 N/m ; G = 9.6 *10 N/m , k = 2.5
The coefficients of the governing equations are
1 2 1α = 71.209; α = 907.956; β = 41.903
;-4 -5
1 2K = 4.4932*10 K = -1.0709 *10 ;
Substituting the coefficients into eqns. (6.4.13 ) we obtain
iv iv -4
2 4 4
-5
4
71.209V +907.956V - 41.958V '' = 4.4932*10
V '' = -1.0723x10 (6.4.14)
Integrating by method of trigonometric series with accelerated convergence we have,
Interval = span, L = 50m. ∴ ≤ ≤0 x 50
Boundary conditions; 3 3 3 3V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
4 4 4 4V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
We seek the solution in the form;
3 3 3
4 4 4
V (x) = f (x)+g (x)
V (x) = f (x)+g (x) (6.4.15)
Substituting eqn. (6.4.15) into eqn. (6.4.14) we obtain
iv iv iv iv -4
2 4 4 2 4 4
-5
4 4
71.209f +907.956f - 41.958f ''+71.209g +907.956g - 41.958V '' = 4.4932*10
f ''+g '' = -1.0723 *10
⇒iv iv
2 4 4 1
4 2
71.209f +907.956f - 41.958f '' =P (x)
f '' =P (x)
iv iv -4
1 2 4 4
-5
2 4
P (x) = 71.209g +907.956g - 41.958g '' - 4.4932x10
P (x) = g ''+1.0723 *10 (6.4.16)
We seek the complimentary function 2g (x)and 4g (x) in the form of sine series as
follows. ∞
∑ n2 2n
n=1
α xg (x) = a sin
L;
∞
∑ n4 4n
n=1
α xg (x) = a sin
L (6.4.17)
where nα = nπ
The auxiliary function can be sought in terms of algebraic function as follows.
2 3
3 0 1 2 3f (x) = A + A x + A x + A x (a)
2 3
4 0 1 2 3f (x) =B +B x +B x +B x (b) (6.4.18)
145
The constants 0 0 1 1 2 2 3 3A , B , A , B , A , B , A & B are easily defined using the boundary
conditions.
From section 6.3.5 we note that for simply supported box girder the boundary
conditions are such that all the constants are zero.
Hence, 3 1
4 2
f (x) =P (x) = 0
f (x) =P (x) = 0 } ⇒
3 3
4 4
V (x) = g (x)
V (x) = g (x) (6.4.19)
Consequently eqn. (6.4.16) becomes
iv iv -4
2 4 4
-5
4
71.209g +907.956g - 41.958g '' - 4.4932*10 = 0
g ''+1.0723 *10 = 0 (6.4.20)
But ∞
∑ n3 3n
n=1
α xg (x) = a sin
L (a)
and ∞
∑ n4 4n
n=1
α xg (x) = a sin
L (b)
∞
∑ 2 n3 3n n
n=1
α xg ''(x) = - a α sin
L (c)
∞
∑iv 4 n3 3n n
n=1
α xg (x) = a α sin
L (d) (6.4.21)
∑¥
'' 2 n4 4n n
n=1
α xg (x) = - a α sin
L (e)
∞
∑iv 4 n4 4n n
n=1
α xg (x) = a α sin
L (f)
Substituting eqn. (6.4.21) into eqn. (6.4.20) we obtain;
71.209
∑ ∑ ∑
∑
4 4 2 -4nn 2n n 4n n 4n
2 -5nn 4n
α xα a +907.956 α a + 41.958 α a sin = 4.4932*10
L
α x- α a Sin = -1.0723 *10
L
(6.4.22)
Equations (6.4,22) will always be satisfied if
( ) ( )+4 4 2 -4n 2n n n 4n71.209α a 907.956α + 41.903α a = 4.4932*10 (6.4.23)
and ( )− 2 -5
n 4nα a = -1.0723x10 ⇒-5
4n 2
n
1.0723 *10a =
α (6.4.24)
For n = 1, 2 -3 4 -5
1 1 1
πα = , α = 3.9478 *10 , α =1.5585 *10
50
∴-5
-3
41 -3
1.0723 *10a = = 2.716 *10
3.9478 *10
146
Substituting into eqn. (6.4.23) we have
( )
⇒
-3 -3 -4
21
-5-2
21 -3
1.0942*10 a +0.1798 2.716 *10 = 4.4932*10
- 3.90168 *10a = = - 3.5658 *10
1.0942*10
Substituting values of a21, a41, and eqn. (6.4.17) into eqn. (6.4.19) we obtain
-2
2
-3
4
πxV (x) = - 3.566 *10 Sin
50
πxV (x) = 2.716 *10 Sin
50
Table 6.11 Variation of flexural and torsional-
displacements along the length of the girder
(Double cell mono symmetric section)
Distance (x)
from Left
Support (m)
πx
Sin50
Flexural
displacement
V2(x) x 10-3
(m)
Torsional
displacement
V4(x) x 10-3
(m)
0 0.000 0.000 0.000
5 0.309 11.019 0.839
10 0.588 20.968 1.597
15 0.809 28.849 2.197
20 0.951 33.913 2.583
25 1.000 35.660 2.716
30 0.951 33.913 2.583
35 0.809 28.849 2.197
40 0.588 20.968 1.597
45 0.309 11.019 0.839
50 0.000 0.000 0.000
6.4.3 Flexural-distortional analysis: single cell mono symmetric section The governing equations of equilibrium are
γ
iv iv
1 2 2 3 1 3 3
iv iv
3 2 4 3 2 3 1 3 4
α V +α V -β V '' =K
α V +α V -β V '' - V = - K (6.4.25)
147
The relevant coefficients are as follows.
22 23 32 33
22 22 22 33 33 33
a = 25.05; a = a = -0.270, a = 0.757
b = c = r = 2.982 b = c = r =1.407
23 32 23 32 23 32
-4 -4
44 33
b = b = c = c = r = r = -0.153
r =14.616; s = 0.261* 6.9712*10 =1.8195 *10
2 3 4
9 2 9 2
q = 0.0KN; q =196.46KN q =1446.505KN
E = 24 *10 N/m ; G = 9.6 *10 N/m , k = 2.5
The coefficients of the governing equations are as follows
1 22 2 23α = ka = 62.625; α = ka = -0.675
3 32 4 33α =Ka = -0.675; α =Ka =1.893
( ) 2
23 22 22 23 33 -4
1 2
33 22 32
a c -a c k sβ = = 8.252*10
c c - c
( )2
33 33 22 32 23 -4
2 2
33 22 32
k s a c -a cβ = = 6.04 *10
c c - c
( )γ 32 23 33 22 33 -4
1 2
33 22 32
a c -b c ks= = -4.5487 *10
c c - c
c
-53221 2
33 22 32
qcK = - = -1.4626 *10
c -c G
c
-723 32 2
33 22 32
c qK = - = -7.504 *10
c -c G
-6
3 23 1 22 2
-5
4 32 2 33 1
K = b K -b K = 4.47545 *10
K = b K +b K = -2.0463 *10
Substituting the coefficients into eqns. (6.4.25 ) we obtain
iv iv -4 -6
2 3 3
iv iv -4 -4 -5
2 3 3
62.625V -0.675V -8.2517 *10 V '' = 4.47545 *10
-0.675V +1.893V - 6.04 *10 V ''+ 4.5487 *10 = 2.0463 *10 (6.4.26)
Integrating by method of trigonometric series with accelerated convergence we have,
Interval= span, L = 50m. ∴ ≤ ≤0 x 50
Boundary conditions; 3 3 3 3V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
4 4 4 4V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
We seek the solution in the form;
2 2 2
3 3 3
V (x) = f (x)+g (x)
V (x) = f (x)+g (x) (6.4.27)
148
Substituting eqn. (6.4.27) into eqn. (6.4.26) we obtain
iv iv -4 iv iv
2 3 3 2 3
-4 -6
3
iv iv -4 -4
2 3 3
iv iv -4 -4 -5
2 3 3 3
62.625f -0.675f -8.2517 *10 f ''+62.625g -0.675g
-8.2517 *10 g '' = 4.47545 *10
-0.675f +1.893f - 6.04 *10 f ''+ 4.5487 *10 f
-0.675g +1.893g - 6.04 *10 g ''+ 4.5487 *10 g = 2.0463 *10
⇒ iv iv -4
2 3 3 1
iv iv -4 -4
2 3 3 2
62.625f -0.675f -8.2517 *10 f '' =P (x)
- 0.675f +1.893f - 6.04 *10 f '' + 4.5487 *10 f =P (x)
⇒iv iv -4 -6
1 2 3 3
iv iv -4 -4 -5
2 2 3 3 3
P (x) = 62.625g -0.675g -8.2517 *10 g '' - 4.47545 *10
P (x) = -0.675g +1.893g - 6.04 *10 g ''+ 4.5487 *10 g - 2.0463 *10 (6.4.28)
We seek the complimentary function 2g (x)and 3g (x) in the form of sine series as
follows. ∞
∑ n2 2n
n=1
α xg (x) = a sin
L
∞
∑ n3 3n
n=1
α xg (x) = a sin
L (6.4.29)
where nα = nπ
The auxiliary function can be sought in terms of algebraic function as follows.
2 3
2 0 1 2 3f (x) = A + A x + A x + A x (a)
2 3
3 0 1 2 3f (x) =B +B x +B x +B x (b) (6.4.30)
The constants 0 0 1 1 2 2 3 3A , B , A , B , A , B , A & B are easily defined using the boundary
conditions. From section 6.3.5 we note that for simply supported span the boundary
conditions are such that all these constants are zero.
Hence, 2 1
3 2
f (x) =P (x) = 0
f (x) =P (x) = 0 } ⇒
2 2
3 3
V (x) = g (x)
V (x) = g (x) (6.4.31)
Consequently eqn. (6.4.28) becomes
iv iv -4 -6
2 3 3
iv iv -4 -4 -5
2 3 3 3
62.625g -0.675g -8.2517 *10 g '' - 4.47545 *10 = 0
-0.675g +1.893g - 6.04 *10 g ''+ 4.5487 *10 g - 2.0463 *10 = 0 (6.4.32)
But ∞
∑ n2 2n
n=1
α xg (x) = a sin
L (a)
and ∞
∑ n3 3n
n=1
α xg (x) = a sin
L (b)
∞
∑ 2 n2 2n n
n=1
α xg ''(x) = - a α sin
L (c)
∞
∑iv 4 n2 2n n
n=1
α xg (x) = a α sin
L (d) (6.4.33)
149
∞
∑ 2 n3 3n n
n=1
α xg ''(x) = - a α sin
L (e)
∞
∑iv 4 n3 3n n
n=1
α xg (x) = a α sin
L (f)
Substituting eqn. (6.4.33) into eqn. (6.4.32) we obtain;
∑ ∑ ∑
∑ ∑ ∑ ∑
4 4 -4 2 -6n2n n 3n n 3n n
4 -4 2 -4 n2n 3n n 3n n 3n
α x62.625 a α - 0.675 a α +8.2517 *10 a α Sin = 4.47545 *10
L
α x-0.675 a +1.893 a α +6.04 *10 a α + 4.5487 *10 a Sin
L
-5 = 2.0463 *10
(6.4.34)
Equations (6.4.34) will always be satisfied if;
4 4 -4 2 -6
n 2n n n 3n(62.625α )a +(-0.675α +8.2517 *10 α )a = 4.47545 *10 (6.4.35)
4 4 -4 2 -4 -5
n 2n n n 3n(-0.675α )a +(1.893α +6.04 *10 α + 4.5487 *10 )a = 2.0463 *10 (6.4.36)
From eqn. (6.4.35) we obtain that
( )4 -4 2 -6n n
2n 3n4 4
n n
-0.675α +8.2517 *10 α 4.4754 *10a = - a +
62.625α 62.625α (6.4.37)
For n = 1, 2 -3 4 -5
1 1 1
πα = , α = 3.948 *10 , α =1.5585 *10
50
Substituting into eqn. (6.4.37) gives
-3 -3
21 3na = -7.4414 *10 a + 4.5858 *10
Substituting into eqn. (6.4.36) we have
( )-5 -4 -3 -5 -5
31 31-1.05199 *10 -7.4414 *10 a + 4.4558 *10 + 4.8676 *10 a = 2.0463 *10
⇒ -4 -5 -8
3n4.8684 *10 a = 2.0463 *10 + 4.68746 *10
⇒-5
-2
31 -4
2.05099 *10a = = 4.213 *10
4.8684 *10
∴ -3
21 a = 4.272*10
Substituting values of a21, a31, and eqn. (6.4.29) into eqn. (6.4.31) we obtain
-3
2
-2
3
πxV (x) = 4.272*10 Sin
50
πxV (x) = 4.213 *10 Sin
50
150
Table 6.12 Variation of flexural and distortional
displacements along the girder length (Single cell mono
symmetric section)
Distance (x)
from Left
Support (m)
πx
Sin50
V2(x)
X 10-3
V3(x)
X 10-3
0 0.000 0.000 0.000
5 0.309 1.320 13.018
10 0.588 2.512 24.772
15 0.809 3.456 34.083
20 0.951 4.063 40.065
25 1.000 4.272 42.130
30 0.951 4.063 40.065
35 0.809 3.456 34.083
40 0.588 2.512 24.772
45 0.309 1.320 13.018
50 0.000 0.000 0.000
6.4.4 Flexural-distortional analysis: double cell mono symmetric section
The governing equations of equilibrium are
γ
iv iv
1 2 2 3 1 4 3
iv iv
3 2 4 3 2 3 1 3 4
α V +α V -β V '' =K
α V +α V -β V '' - V = -K (6.4.41)
The relevant coefficients are as follows.
22 23 32 33
22 22 22 33 33 33
a = 25.073; a = a = 0.425, a = 0.750
b = c = r = 2.982, b = c = r =1.533
23 32 23 32 23 32
-4 -4
33
b = b = c = c = r = r = -0.449
s = 0.261* 6.9712*10 =1.8195 *10
2 3 4
9 2 9 2
q = 0.0KN, q =196.46KN q =1446.505KN
E = 24 *10 N/m , G = 9.6 *10 N/m , k = 2.5
The coefficients of the governing equations are as follows
1 22
2 23
α = ka = 62.6825
α = ka =1.0625
151
3 32
4 33
α = Ka =1.0625
α = Ka =1.875
( ) 2
23 22 22 23 33
1 2
33 22 32
a c -a c k sβ = = 0.0283
c c - c
( )2
33 33 22 32 23 -3
2 2
33 22 32
k s a c - a cβ = =1.750 *10
c c - c
( )γ 32 23 33 22 33 -3
1 2
33 22 32
a c -b c ks= = -1.260 *10
c c - c
c
-53221 2
33 22 32
qcK = - = -1.115 *10
c -c G
c
-623 32 2
33 22 32
c qK = - = -1.684 *10
c -c G
-5
3 23 1 22 2
-5
4 32 2 33 1
K = b K -b K =1.003 *10
K = b K +b K = -1.634 *10
Substituting the coefficients into eqns. (6.4.41 ) we obtain
iv iv -5
2 3 3
iv iv -3 -3 -5
2 3 3 3
62.683V +1.0625V -0.0283V '' = -1.003 *10
1.0625V +1.875V -1.750 *10 V ''+1.26 *10 V =1.634 *10 (6.4.42)
Integrating by method of trigonometric series with accelerated convergence we
have,
Interval = span, L = 50m. ∴ ≤ ≤0 x 50
Boundary conditions; 2 2 2 2V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
3 3 3 3V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
We seek the solution in the form;
2 2 2
3 3 3
V (x) = f (x)+ g (x)
V (x) = f (x)+ g (x) (6.4.43)
Substituting eqn. (6.4.43) into eqn. (6.4.42) we obtain
( )a
iv iv iv iv
2 3 3 2 3
-5
3
iv iv -3 -3
2 3 3
iv iv -3
2 3
62.683f +1.0625f - 0.0283f ''+ 62.683g +1.0625g
-0.0283g '' = -1.003 *10
1.0625f +1.875f -1.750 *10 f ''+1.26 *10 f
+1.0625g +1.875g -1.70 *10 ( )b-3 -5
3 3g '' +1.26 *10 g = -1.634 *10
(6.4.44)
⇒ iv iv
2 3 3 1
iv iv -3 -3
2 3 3 2
62.683f +1.0625f -0.0283f '' =P (x)
1.0625f +1.875f -1.750 *10 f ''+1.260 *10 f =P (x)
152
where,
iv iv -5
1 2 3 3
iv iv -3 -3 -5
2 2 3 3 3
P (x) = 62.683g +1.0625g -0.0283g ''+1.003 *10
P (x) =1.0625g +1.875g -1.750 *10 g ''+1.260 *10 g -1.634 *10 (6.4.45)
We seek the complimentary function 2g (x)and 3g (x) in the form of sine series as
follows.
∞
∞
∑
∑
n2 2n
n=1
n3 3n
n=1
α xg (x) = a sin
L
α xg (x) = a sin
L
(6.4.46)
where nα = nπ
The auxiliary function can be sought in terms of algebraic function as follows.
2 3
2 0 1 2 3
2 3
3 0 1 2 3
f (x) = A + A x + A x + A x
f (x) =B +B x +B x +B x (6.4.47)
The constants 0 0 1 1 2 2 3 3A , B , A , B , A , B , A & B are easily defined using the boundary
conditions.
From section 6.3.5 we note that for simply supported span the boundary conditions
are such that all the constants are zero.
Hence, 2 1
3 2
f (x) = P (x) = 0
f (x) = P (x) = 0 } ⇒
2 2
3 3
V (x) = g (x)
V (x) = g (x) (6.4.48)
Consequently eqn (6.4.45) becomes
iv iv -5
2 3 3
iv iv -3 -3 -5
2 3 3 3
62.683g +1.0625g -0.0283g ''+1.003 *10 = 0
1.0625g +1.875g -1.750 *10 g ''+1.260 *10 g -1.634 *10 = 0 (6.4.49)
But ∞
∑ n2 2n
n=1
α xg (x) = a sin
L (a)
and ∞
∑ n3 3n
n=1
α xg (x) = a sin
L (b)
∞
∑ 2 n2 2n n
n=1
α xg ''(x) = - a α sin
L (c)
∞
∑iv 4 n2 2n n
n=1
α xg (x) = a α sin
L (d) (6.4.50)
∞
∑ 2 n3 3n n
n=1
α xg ''(x) = - a α sin
L (e)
∞
∑iv 4 n3 3n n
n=1
α xg (x) = a α sin
L (f)
153
Substituting eqn. (6.4.50) into eqn. (6.4.49) we obtain;
4
1.0625 ( )
1.0625 n
a
α
+
∑ ∑ ∑
∑ ∑ ∑ ∑
4 4 2 -5n2n n 3n n 3n n
4 -3 2 -3 n2n 3n n 3n n 3n
α x62.682 a α a α +0.0283 a α Sin = -1.003 *10
L
α xa +1.875 a α +1.75 *10 a α +1.260 *10 a Sin
L
( )b-5 =1.634 *10
(6.4.51)
Equations (6.4.51) will always be satisfied if
4 4 2 -5
n 2n n n 3n(62.682α )a +(1.0625α +0.0283α )a = -1.003 *10 (6.4.52)
4 4 -3 2 -3 -5
n 2n n n 3n(1.0625α )a +(1.875α +1.75 *10 α +1.260 *10 )a =1.634 *10 (6.4.53)
From eqn. (6.4.52) we obtain that
( )4 2 -5n n
2n 3n4 4
n n
1.0625α +0.0283α 1.003 *10a = - a -
62.682α 62.682α (6.4.54)
For n = 1, 2 -3 4 -5
1 1 1
πα = , α = 3.948 *10 , α =1.5585 *10
50
Substituting into eqn. (6.4.54) gives
-2
21 31a = -0.1313a +1.02672*10
Substituting into eqn. (6.4.53) we have
( )-5 -2 -3 -5
31 311.656 *10 -0.1313a -1.0264 *1o +1.296 *10 a =1.634 *10
-3 -5
311.294 *10 a =1.616998 *10
⇒ -2
31
-2 -2
21
a = 1.2496 *10
a = - 0.1313 *1.2496 *10 +1.0267 *10
⇒ -3
21a = 8.626 *10
Substituting values of a21, a31 and eqn.(6.4.46) into eqn.(6.4.48) we obtain
-3
2
πxV (x) = 8.626 *10 Sin
50; -2
3
πxV (x) =1.250 *10 Sin
50
Table 6.13 Variation of flexural and distortional
displacements along the length of the girder
(Double cell mono symmetric section)
Distance (x)
from Left
Support (m)
πx
SinL
V2(x)
X 10-3
V3(x)
X 10-3
0 0.000 0.000 0.000
154
6.4.5 Torsional-distortional analysis: single cell mono-symmetric section The governing equations of equilibrium are
γ1 4 1 3 1
iv iv
1 3 2 4 2 4 2
β V '' - V =K
α V +α V -β V '' =K (6.4.55)
The relevant coefficients are as follows.
33 33 33 33
34 43 34 43 44
a = 0.750; b = c = r =1.407
c = c = r = r =1.265; r =14.616
-4 -4
33s = 0.261* 6.9712*10 =1.8195 *10
2 3 4q = 0.0KN; q =196.46KN; q =1446.505KN
9 2 9 2E = 24 *10 N/m ; G = 9.6 *10 N/m ; k = 2.5
The coefficients for the governing equations are as follows
1 33 43
2 33 44
α = ka c = 2.372
α = ka r = 27.405
1 34 43 33 44β = r c - c r = -18.964
2 33 44 34 43β = b r - c c =18.964
9 2 9 2E = 24 *10 N/m , G = 9.6 *10 N/m , k = 2.5
1γ -4
43 33= c ks = 5.503 *10
-441 43 33
qK = - c +C =1.9163 *10
G G3333qqqq
5 0.309 2.665 3.863
10 0.588 5.072 7.350
15 0.809 6.978 10.113
20 0.951 8.203 11.888
25 1.000 8.626 12.500
30 0.951 8.203 11.888
35 0.809 6.978 10.113
40 0.588 5.072 7.350
45 0.309 2.665 3.863
50 0.000 0.000 0.000
155
-442 33
qK = b = 2.120 *10
G
Substituting the coefficients into eqns (6.4.55 ) we obtain
iv iv -6
3 4 4
iv -4 -4
4 3
2.371V +27.405V -18.963V '' = 2.120 *10
-18.964V -5.503 *10 V =1.9163 *10 (6.4.56)
Integrating by method of trigonometric series with accelerated convergence we have,
Interval = span, L=50m. ∴ ≤ ≤0 x 50
Boundary conditions; 3 3 3 3V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
4 4 4 4V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
We seek the solution in the form;
3 3 3
4 4 4
V (x) = f (x)+g (x)
V (x) = f (x)+g (x) (6.4.57)
Substituting eqn (6.4.57) into eqn (6.4.56) we obtain
iv iv iv iv -4
3 4 4 3 4 3
'' -4 '' -4 -4
4 3 4 3
2.371f +27.405f -18.964f '' +2.371g +27.405g -18.964g '' = 2.120 *10
-18.964f -5.503 *10 f -18.964g -5.503 *10 g =1.9163 *10
2.371⇒ iv iv
3 4 4 1
iv -4
4 3 2
f +27.405f -18.964f '' =P (x)
-18.964f -5.503 *10 f =P (x)
where, iv iv -4
1 3 4 4P (x) = 2.371g +27.405g -18.964g '' - 2.12*10 (a)
-4 -4
2 4 3P (x) = -18.964g '' - 5.503 *10 g -1.9163 *10 (b)
(6.4.58)
We seek the complimentary function 3g (x)and 4g (x) in the form of sine series as
follows.
∞
∞
∑
∑
n3 3n
n=1
n4 4n
n=1
α xg (x) = a sin
L
α xg (x) = a sin
L
(6.4.59)
where nα = nπ
The auxiliary function can be sought in terms of algebraic function as follows.
2 3
2 0 1 2 3f (x) = A + A x + A x + A x (a)
2 3
3 0 1 2 3f (x) =B +B x +B x +B x (b) (6.4.60)
156
The constants 0 0 1 1 2 2 3 3A , B , A , B , A , B , A & B are easily defined using the boundary
conditions.
From section 6.3.5 we note that for simply supported span the boundary conditions
are such that all the constants are zero.
Hence, 3 3
4 4
f (x) =P (x) = 0
f (x) =P (x) = 0 } ⇒
3 3
4 4
V (x) = g (x)
V (x) = g (x) (6.4.61)
Consequently eqn (6.4.58) becomes
iv iv -4
3 4 4
-4 -4
4 3
2.371g +27.405g -18.964g '' - 2.12 *10 = 0
-18.964g '' - 5.503 *10 g -1.9163 *10 = 0 (6.4.62)
But ∞
∑ n3 3n
n=1
α xg (x) = a sin
L (a)
and ∞
∑ n3 3n
n=1
α xg (x) = a sin
L (b)
∞
∑ 2 n3 3n n
n=1
α xg ''(x) = - a α sin
L (c)
∞
∑iv 4 n3 3n n
n=1
α xg (x) = a α sin
L (d) (6.4.63)
∞
∑ 2 n4 4n n
n=1
α xg ''(x) = - a α sin
L (e)
∞
∑iv 4 n4 4n n
n=1
α xg (x) = a α sin
L (f)
Substituting eqn. (6.4.63) into eqn. (6.4.62) we obtain;
∑ ∑ ∑
∑ ∑
4 4 2 -4n3n n 4n n 4n n
2 -4 n4n n 3n
α x2.371 a α +27.405 a α +18.964 a α Sin = 2.120 *10
L
α x18.964 a α - 5.503 *10 a Sin =1.9163 *10 - 4
L
(6.4.64)
Equations (6.4,64) will always be satisfied if
4 4 2 -4
n 3n n n 4n(2.371α )a +(27.405α +18.964α )a = 2.120 *10 (6.4.65)
2 -4 -4
n 4n 3n(18.964α )a - (5.503 *10 )a =1.9163 *10 (6.4.66)
From eqn. (6.4.66) we obtain that
( )−
2 -4n
3n 4n-4 -4
18.964α 1.9163 *10a = - a
5.503 *10 5.503 *10 (6.4.67)
157
⇒ 2
3n n 4na = 34461.203α a -0.3482
For n = 1, 2 -3 4 -5
1 1 1
πα = , α = 3.948 *10 , α =1.5585 *10
50
Substituting into eqn. (6.4.67) gives
31 41a =136.053a -0.3482
Substituting into eqn.(6.4.66) we have
( )3.695 -5 -2 -4
41 412*10 136.053a -0.3482 +7.5297 *10 a = 2.120 *10
⇒ -2 -4 -5 -4
41
-3
41
8.0324 *10 a = 2.120 *10 +1.28667 *10 = 2.24667 *10
a = 2.7995 *10
( )⇒ -2
31a = 0.380879 -0.3482 = 3.268 *10
Substituting values of a31, a42 and eqn. (6.4.59) into eqn. (6.4.61) we obtain
-2
3
-3
4
πxV (x) = 3.268 *10 Sin
50
πxV (x) = 2.80 *10 Sin
50
Table 6.14 Variation of torsional and distortional
displacements along the length of the girder
(Single cell mono symmetric section)
Distance (x)
from Left
Support (m)
πx
Sin50
V3(x)
X 10-3
V4(x)
X 10-3
0 0.000 0.000 0.000
5 0.309 10.098 0.865
10 0.588 19.216 1.646
15 0.809 26.438 2.265
20 0.951 31.079 2.663
25 1.000 32.680 2.800
30 0.951 31.079 2.663
35 0.809 26.438 2.265
40 0.588 19.216 1.646
45 0.309 10.098 0.865
158
50 0.000 0.000 0.000
6.4.6 Torsional-distortional analysis: double cell mono-symmetric section
The governing equations of equilibrium are
γ1 4 1 3 1
iv iv
1 3 2 4 2 4 2
β V '' - V =K
α V +α V -β V '' =K (6.4.70)
The relevant coefficients are as follows.
33 33 33 33
34 43 34 43 44
a = 0.750; b = c = r =1.533
c = c = r = r =1.295; r =14.485
-4 -4
33s = 0.723 * 6.9712*10 = 5.04 *10
2 3 4
9 2 9 2
q = 0.0KN; q =196.46KN; q =1446.505KN
E = 24 *10 N/m ; G = 9.6 *10 N/m ; k = 2.5
The coefficients of the governing equations are as follows
1 33 43
2 33 44
α = ka c = 2.428
α = ka r = 27.160
1 34 43 33 44β = r c - c r = -20.528
2 33 44 34 43β = b r -c c = 20.528
1γ -3
43 33= c ks =1.632*10
-43 41 43 33
q qK = - c +C = 2.613 *10
G G
-442 33
qK = b = 2.87845 *10
G
Substituting the coefficients into eqns. (6.4.70 ) we obtain
iv iv -4
3 4 4
iv -3 -4
4 3
2.428V +27.16V - 20.528V '' = 2.87845 *10
-20.528V +1.632*10 V = 2.613 *10 (6.4.71)
Integrating by method of trigonometric series with accelerated convergence we have,
Interval = span, L=50m. ∴ ≤ ≤0 x 50
Boundary conditions; 3 3 3 3V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
4 4 4 4V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
We seek the solution in the form;
159
3 3 3
4 4 4
V (x) = f (x)+g (x)
V (x) = f (x)+g (x) (6.4.72)
Substituting eqn. (6.4.72) into eqn. (6.4.71) we obtain
( )
( )
a
b
iv iv iv iv -4
3 4 4 3 4 3
iv -3 iv -3 -4
4 3 4 3
2.428f +27.16f - 20.528f ''+2.428g +27.16g - 20.5284g '' = 2.87845 *10
-20.528f +1.632*10 f - 20.528g +1.632*10 g = 2.613 *10
⇒ iv iv
3 4 4 1
iv -3
4 3 2
2.428f +27.16f - 20.528f '' =P (x)
- 20.528f +1.632*10 f =P (x)
where,
iv iv -4
1 3 4 4
-3 -4
2 4 3
P (x) = 2.428g +27.16g - 20.528g '' - 2.87845 *10
P (x) = -20.528g +1.632*10 g - 2.613 *10 (6.4.73)
We seek the complimentary function 3g (x)and 4g (x) in the form of sine series as
follows.
∞
∞
∑
∑
n3 3n
n=1
n4 4n
n=1
α xg (x) = a sin
L
α xg (x) = a sin
L
(6.4.74)
where nα = nπ
The auxiliary function can be sought in terms of algebraic function as follows.
2 3
3 0 1 2 3f (x) = A + A x + A x + A x (a)
2 3
4 0 1 2 3f (x) =B +B x +B x +B x (b) (6.4.75)
The constants 0 0 1 1 2 2 3 3A , B , A , B , A , B , A & B are easily defined using the boundary
conditions.
From section 6.3.5 we noted that for simply supported span the boundary conditions
are such that all the constants A0, B0, A1, B1, etc, are zero.
Hence, 3 3
4 4
f (x) =P (x) = 0
f (x) =P (x) = 0 } ⇒
3 3
4 4
V (x) = g (x)
V (x) = g (x) (6.4.76)
Consequently eqn. (6.4.73) becomes
iv iv -4
3 4 4
-3 -4
4 3
2.428g +27.16g - 20.528g '' - 2.87845 *10 = 0
-20.528g ''+1.632*10 g - 2.613 *10 = 0 (6.4.77)
But ∞
∑ n3 3n
n=1
α xg (x) = a sin
L (a)
and ∞
∑ n4 4n
n=1
α xg (x) = a sin
L (b)
160
∞
∑ 2 n3 3n n
n=1
α xg ''(x) = - a α sin
L (c)
∞
∑iv 4 n3 3n n
n=1
α xg (x) = a α sin
L (d) (6.4.78)
∞
∑ 2 n4 4n n
n=1
α xg ''(x) = - a α sin
L (e)
∞
∑iv 4 n4 4n n
n=1
α xg (x) = a α sin
L (f)
Substituting eqn. (6.4.78) into eqn. (6.4.77) we obtain;
220.528 α
−
∑ ∑ ∑
∑ ∑
4 4 2 -4n3n n 4n n 4n n
-3 -4n4n 3n
α x2.428 a α +27.16 a α +20.528 a α Sin = 2.87845 *10
L
α xa +1.632*10 a Sin = 2.613 *10
L
(6.4.79)
Equations (6.4.79) will always be satisfied if
4 4 2 -4
n 3n n n 4n(2.428α )a +(27.16α +20.528α )a = 2.87845 *10 (6.4.80)
-3 -4
3n 4n(1.632*10 )a - (20.528)a = 2.613 *10 (6.4.81)
From eqn. (6.4.81) we obtain that
( )−
2 -4n
3n 4n-3 -3
20.528α 2.613 *10a = a
1.632*10 1.632*10
⇒ 2
3n n 4na =12578.43 α a -0.16011 (6.4.82)
For n = 1, 2 -3 4 -5
1 1 1
πα = , α = 3.948 *10 α =1.5585 *10
50
Substituting into eqn. (6.4.82) gives
31 41a = 49.6596 a -0.16011
Substituting into eqn. (6.4.80) we have
( )3.785 -5 -2 -4
41 41*10 49.6596 a -0.16011 +8.147 *10 a = 2.87845 *10
⇒
⇒
-2 -4 -6 -4
41
-3
41
8.3347 *10 a = 2.87845 *10 +6.0602*10 = 2.939 *10
a = 3.526 *10
( )∴ -2
31a = 0.1751-0.16011 = 1.500 *10
Substituting values of a31, a41 and eqn.(6.4.74) into eqn.(6.4.72) we obtain
161
-2
3
-3
4
πxV (x) =1.500 *10 Sin
50
πxV (x) = 3.526 *10 Sin
50
Table 6.15 Variation of torsional and distortional
displacements along the length of the girder
(Double cell mono symmetric section)
Distance (x)
from Left
Support (m)
πx
Sin50
V3(x)
X 10-3
V4(x)
X 10-3
0 0.000 0.000 0.000
5 0.309 4.635 1.090
10 0.588 8.820 2.073
15 0.809 12.135 2.852
20 0.951 14.265 3.353
25 1.000 15.000 3.526
30 0.951 14.265 3.353
35 0.809 12.135 2.852
40 0.588 8.820 2.073
45 0.309 4.635 1.090
50 0.000 0.000 0.000
6. 4.7 Flexural-torsional-distortional analysis: single cell mono-symmetric section The governing equations of equilibrium are
γ
γ
γ
4 2 5 3 6 4 1 3 3
iv iv iv iv
3 2 5 3 6 4 7 2 8 3 9 4 2 3 4
iv iv iv
7 2 8 3 9 4 10 2 11 3 12 4 3 3 5
β V ''+β V ''+β V '' - V = -K
α V +α V +α V -β V '' -β V '' -β V ''+ V =K
α V +α V +α V -β V '' -β V '' -β V ''+ V =K
(6.4.85)
The relevant coefficients are as follows.
22 23 33
22 22 22
23 23 23
a = 25.05, a = -0.270, a = 0.757,
b = c = r = 2.982
b = c = r = -0.153
162
24 42 24 42 2.515c c r r= = = = −
33 33 33 44
34 43 34 43
b = c = r =1.407, r =14.616
c = c = r = r =1.265
-4 -4
33s = 0.261* 6.9712*10 =1.8195 *10
2q 3 4
9 2 9 2
= 0.0KN, q =154.58KN, q =1446.505KN
E = 24 *10 N/m , G = 9.6 *10 N/m , k = 2.5
The coefficients of the governing equations are as follows.
32 22
33 231
3222
23 33
r r-
c c 19.38α = = = -1
cc -19.38-
c c
34 24
33 232
3222
23 33
r r-
c c -15.54α = = = 0.802
cc -19.38-
c c
-4-533
3
322233
23 33
ks 2.5 *1.74 *10α = = = -1.6679 *10
1.407 * (-19.38)cc- c
c c
33 23
221
23 33
22 32
r r-
c2 c -9.14477β = = = -1
c c 9.14477-
c c
34 24
32 222
23 33
22 32
r r-
c c -7.4246β = = = -0.812
c c 9.14477-
c c
-4-433
3
23 3332
22 32
ks 2.5 *1.74 *10β = = = -3.251*10
-0.153 * 9.14477c c- c
c c
;
( ) ( )4 42 1 42β = c α +r = -2.515 * -1 + -2.515 = 0 ; ( )5 43 1 43β = c β +r =1.265 * -1 +1265 = 0
( )6 42 2 43 2 44
7 22 1 22
β = c α +c β +r =11572
β = b α +c = 2.982* -1 +2.982 = 0
( )
( )
-4
8 23 3 23 1 22 3 23
9 22 2 33 2 24
β = ka β +b β +ka α +c = -7.89 *10
β = b α +b β +c = -1.266
163
( )
( )10 32 1 32
-4
11 32 33 3 33 1 33
β = b α +c = 0
β = ka +ka β +b β +c = -5.820 *10
( ) -4
12 32 2 33 2 34β = b α +b β +c = -1.90 *10
( )6 22 2 23 2
7 32 1
α = k a α +a β = 50.7734
α = ka α = 0.675
( )8 33 1
9 32 2 33 2
α = ka β = -1.893
α = k a α +a β = -2.078
( )
( )
γ
γ
-4
1 42 3 43 3
-9
2 22 3 23 3
= c α +c β = -3.532*10
= b α +b β = 4.8 *10 »0
( )3γ -4
32 3 33 3= b α +b β = -4.350 *10
-53 21
23 33 23 3332 22
22 32 22 32
q qK = - = -1.4625 *10
c c c c- c G - c G
c c c c
-73 22
32 3222 2233 23
23 33 23 33
q qK = - = -7.505 *10
c cc c- c G - c G
c c c c
-443 42 2 43 1
qK = +c K +c K = 1.7115 *10
G
( )
( )
-6
4 22 2 23 1
-5
5 32 2 33 1
K = b K +b K = -2.123 *10
K = b K +b K = -2.046 *10
Substituting the coefficients and the constants into eqn.(6.4.85) we obtain,
( )
( )
( )
a
b
c
-4 -4
4 3
iv iv iv -4 -6
2 3 4 3 4
iv iv iv -4 -4
2 3 4 3 4
-4 -5
3
11.572V ''+3.532*10 V = - 1.7115 *10
- 62.625V +0.675V +50.773V +7.89 *10 V +1.266V '' = -2.123 *10
0.675V -1.893V - 2.078V +5.820 *10 V ''+1.90 *10 V '' -
- 4.35 *10 V = - 2.046 *10
(6.4.86 )
Integrating by method of trigonometric series with accelerated convergence we have, Interval = span, L = 50m. ∴ ≤ ≤0 x 50
Boundary conditions; 2 2 2 2V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
3 3 3 3V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
4 4 4 4V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
We seek the solution in the form;
4 22 1
5 23 1
α = ka α = -62.625
α = ka β = 0.675
164
2 2 2
3 3 3
4 4 4
V (x) = f (x)+g (x)
V (x) = f (x)+g (x)
V (x) = f (x)+g (x)
(6.4.87)
Substituting eqn (6.4.87) into eqn (6.4.486) we obtain
-4 -4 -4
4 3 4 311.572f ''+3.532*10 f +11.572g ''+3.532*10 g = -1.7115 *10 (a)
iv iv iv -4 iv
2 3 4 3 4 2
iv iv -4 -6
3 4 3 4
-62.625f +0.675f +50.225f +7.89 *10 f +1.266f '' - 62.625g
+0.675g +50.225g +7.89 *10 g +1.266g '' = -2.123 *10 (b)
iv iv iv -4 -4
2 3 4 3 4
-4 iv iv iv -4 -4
3 2 3 4 3 4
-4 -5
3
0.675f -1.875f - 2.064f +5.768 *10 f ''+1.90 *10 f '' -
-4.35 *10 f +0.675g -1.875g - 2.064g +5.768 *10 g ''+1.90 *10 g '' -
-4.35 *10 g = -2.046 *10 (c)
⇒ -4
4 3 1 11.572f ''+3.532*10 f =P (x) (a)
iv iv iv -4
2 3 4 3 4 2-62.625f +0.675f +50.225f +7.89 *10 f +1.266f '' =P (x) (b)
iv iv iv -4 -4
2 3 4 3 4
-4
3 3
0.675f -1.875f - 2.064f +5.768 *10 f ''+1.90 *10 f '' -
-4.35 *10 f =P (x) (c)
where,
-4 -4
1 4 3P (x) =11.572g '' +3.532*10 g +1.7115 *10 (a)
iv iv iv -4 -6
2 2 3 4 3 4P (x) = -62.625g +0.675g +50.225g +7.89 *10 g +1.266g ''+2.123 *10 (b)
iv iv iv -4 -4
3 2 3 4 3 4
-4 -5
3
P (x) = 0.675g -1.875g - 2.064g +5.768 *10 g ''+1.90 *10 g '' -
-4.35 *10 g +2.046 *10 (c) (6.4.88)
We seek the complimentary function 2g (x) , g3(x) and g4(x) in the form of sine series
as follows.
∞
∞
∞
∑
∑
∑
n2 2n
n=1
n3 3n
n=1
n4 4n
n=1
α xg (x) = a sin
L
α xg (x) = a sin
L
α xg (x) = a sin
L
(6.4.89)
where nα = nπ
The auxiliary function can be sought in terms of algebraic function as follows.
2 3
2 0 1 2 3f (x) = A + A x + A x + A x (a)
165
2 3
3 0 1 2 3f (x) =B +B x +B x +B x (b) (6.4.90)
2 3
4 0 1 2 3f (x) = C +C x +C x + x (c)
The constants 0 0 1 1 2 2 3 3A , B , A , B , A , B , A & B are easily defined using the boundary
conditions.
From section 6.3.5 we noted that for simply supported span the boundary conditions
are such that all the constants A0, B0, C0, etc, are zero.
Hence,
2 2
3 3
4 4
f (x) =P (x) = 0
f (x) =P (x) = 0
f (x) =P (x) = 0
} ⇒
2 2
3 3
4 4
V (x) = g (x)
V (x) = g (x)
V (x) = g (x)
(6.4.91)
Consequently eqn. (6.4.91) becomes
-4 -4
4 311.572g ''+3.532*10 g +1.7115 *10 = 0 (a)
iv iv iv -4 -6
2 3 4 3 4-62.625g +0.675g +50.225g +7.89 *10 g +1.266g ''+2.123 *10 = 0 (b) (6.4.92)
iv iv iv -4 -4
2 3 4 3 4
-4 -5
3
0.675g -1.875g - 2.064g +5.768 *10 g ''+1.90 *10 g '' -
-4.35 *10 g +2.046 *10 = 0 (c)
But,
∞
∞
∞
∑
∑
∑
n2 2n
n=1
2 n2 2n n
n=1
iv 4 n2 2n n
n=1
α xg (x) = a sin
L
α xg ''(x) = - a α sin
L
α xg (x) = a α sin
L
(a)
and
∞
∞
∞
∑
∑
∑
n3 3n
n=1
2 n3 3n n
n=1
iv 4 n3 3n n
n=1
α xg (x) = a sin
L
α xg ''(x) = - a α sin
L
α xg (x) = a α sin
L
(b) (6.4.93)
166
Also,
∞
∞
∞
∑
∑
∑
4 n4 4n n
n=1
2 n4 4n n
n=1
iv 4 n4 4n n
n=1
α xg (x) = a α sin
L
α xg ''(x) = - a α sin
L
α xg (x) = a α sin
L
(c)
Substituting eqn (6.4.93) into eqn (6.4.92) we obtain;
( )a∑ ∑
α x2 -4 -4n-11.572 a α + 3.532 *10 a Sin = -1.7115 *10n4n 3n L
∑ ∑ ∑ ∑ ∑
α x4 4 4 -4 2 n-62.625 a α + 0.675 a α + 50.773 a α + 7.89 * 10 a - 1.266 a α Sinn n n n4n 4n2n 3n 3n
L
( )b-6
= -2.123 * 10
(6.4.94)
∑ ∑ ∑ ∑ ∑
∑
α x4 4 4 -4 2 -4 2 n0.675 a α - 1.893 a α - 2.078 a α - 5.82 * 10 a α - 1.90 * 10 a α Sinn n n n n4n 4n2n 3n 3n
L
α xn-4 - 4.35*10 a Sin 3nL
( )c-5
= - 2.046 * 10
Equations (6.4.94) will always be satisfied if
2 -4 -4
n 4n 3n(-11.572α )a +(3.532*10 )a = -1.7115 *10 (6.4.96)
( )4 4 -4 2 4 2 -6
n 2n n n 3n n n 4n(-62.625α )a +(0.675α +7.89 *10 α )a + 50.773α -1.266α a = -2.123 *10
(6.4.97)
( )
−
−
4 4 -4 2 -4
n 2n n n 3n
4 -4 2 -5
n n 4n
(0.675α )a - (1.893α +5.82*10 α + 4.35 *10 )a
2.078α +1.90 *10 α a = -2.046 *10 (6.4.98)
For n = 1, ,2 -3 4 -5
1 1 1
πα = , α = 3.948 *10 α =1.5585 *10
50
Substituting into eqn.(6.4.96, 6.4. 97, & 6.4.98) gives
( )
( )
( )
a
b
c
-2 -4 -4
41 31
-4 -5 -3 -6
21 31 41
-5 -4 -5 -5
21 31 41
-4.5684 *10 a +3.532*10 a = -1.7115 *10
-9.760 *10 a +1.3635 *10 a - 4.2066 *10 a = -2.123 *10
1.052*10 a - 4.658 *10 a -3.3136 *10 a = -2.046 *10
(6.4.99)
Solving eqns.[6.4.99 (a), (b), & (c)) we obtain
-2
21a = -3.287 *10 ; -2
31a = 4.282*10 ; -3
41a = 4.077 *10
Substituting values of a21, a31, a41 and eqn.(6.4.89) into eqn.(6.4.91) we obtain
167
-2
2
-2
3
-3
4
πx V (x) = -3.287 *10 Sin
50
πx V (x) = 4.282*10 Sin
50
πx V (x) = 4.077 *10 Sin
50
Table 6.16 Variation of flexural, torsional and distortional displacements along the length of the girder (Single cell mono symmetric section) Distance (x)
from Left
Support (m)
πx
Sin50
V2(x)
X 10-3
V3(x)
X 10-3
V4(x)
X 10-3
0 0.000 0.000 0.000 0.000
5 0.309 -10.156 13.231 1.260
10 0.588 -19.328 25.178 2.397
15 0.809 -26.592 34.641 3.298
20 0.951 -31.259 40.722 3.877
25 1.000 -32.870 42.820 4.077
30 0.951 -31.259 40.722 3.877
35 0.809 -26.592 34.641 3.298
40 0.588 -19.328 25.178 2.397
45 0.309 -10.156 13.231 1.260
50 0.000 0.000 0.000 0.000
6.4.8 Flexural-torsional-distortional analysis: double cell mono-symmetric section The governing equations of equilibrium are
γ
γ
γ
4 2 5 3 6 4 1 3 3
iv iv iv iv
3 2 5 3 6 4 7 2 8 3 9 4 2 3 4
iv iv iv
7 2 8 3 9 4 10 2 11 3 12 4 3 3 5
β V ''+β V ''+β V '' - V = -K
α V +α V +α V -β V '' -β V '' -β V ''+ V =K
α V +α V +α V -β V '' -β V '' -β V ''+ V =K
(6.4.101)
The relevant coefficients are as follows.
22 23 33
22 22 22
23 23 23
a = 25.073, a = 0.425, a = 0.750,
b = c = r = 2.982
b = c = r = -0.449
168
24 42 24 42 1.112c c r r= = = =
33 33 33 44
34 43 34 43
b = c = r =1.533, r =14.488
c = c = r = r =1.295
-4 -4
33s = 0.723 * 6.9712*10 = 5.040 *10
2 3 4
9 2 9 2
q = 0.00KN, q =154.58KN, q =1446.505KN
E = 24 *10 N/m , G = 9.6 *10 N/m , k = 2.5
The coefficients of the governing equations are as follows.
32 22
33 231
3222
23 33
r r-
c c 6.4385α = = = -1
cc -6.3485-
c c
34 24
33 232
3222
23 33
r r-
c c 3.321α = = = -0.523
cc -6.3485-
c c
-4-433
3
322233
23 33
ks 2.5 * 5.04 *10α = = = -1.295 *10
1.533 * (-6.3485)cc- c
c c
33 23
221
23 33
22 32
r r-
c2 c -3.2637β = = = -1
c c 3.2637-
c c
34 24
32 222
23 33
22 32
r r-
c c -3.257β = = = -0.998
c c 3.2637-
c c
-4333
23 3332
22 32
ksβ = = -8.5983 *10
c c- c
c c
( )
( )4 42 1 42
5 43 1 43
β = c α +r =1.112* -1 +1.112 = 0
β = c β +r =1.295 * -1 +1295 = 0
( )6 42 2 43 2 44
7 22 1 22
β = c α +c β +r =12.811
β = b α +c = 2.982* -1 +2.982 = 0
169
( )
( )
-3
8 23 3 23 1 22 3 23
9 22 2 33 2 24
β = ka β +b β +ka α +c = -9.031*10
β = b α +b β +c = -1.978
( )
( )10 32 1 32
-3
11 32 33 3 33 1 33
β = b α +c = 0
β = ka +ka β +b β +c = -1.7498 *10
( ) -4
12 32 2 33 2 34β = b α +b β +c = -1.07 *10
4 22 1α = ka α = -62.6825
5 23 1α = ka β = -1.063
( )6 22 2 23 2
7 32 1
α = k a α +a β = -33.8433
α = ka α = -1.063
( )8 33 1
9 32 2 33 2
α = ka β = -1.875
α = k a α +a β = -2.427
( )
( )
γ
γ
-3
1 42 3 43 3
-7
2 22 3 23 3
= c α +c β =1.2575 *10
= b α +b β =1.0533 *10
( )3γ -3
32 3 33 3= b α +b β = -1.260 *10
-53 21
23 33 23 3332 22
22 32 22 32
q qK = - = -1.3964 *10
c c c c- c G - c G
c c c c
-63 22
32 3222 2233 23
23 33 23 33
q qK = - = 4.09 *10
c cc c- c G - c G
c c c c
-443 42 2 43 1
qK = +c K +c K =1.7424 *10
G
( )
( )
-5
4 22 2 23 1
-5
5 32 2 33 1
K = b K +b K =1.8464 *10
K = b K +b K = 2.3197 *10
Substituting the coefficients and the constants into eqn.(6.4.101) we obtain,
( )
( )
a
b
-3 -4
4 3
iv iv iv -3
2 3 4 3 4
-7 -5
3
iv iv iv -3 -4
2 3 4 3 4
-4
12.611V '' -1.2573 *10 V = -1.400 *10
-62.6825V -1.063V -33.8433V +9.031*10 V ''+1.978V ''
-1.0533 *10 V =1.453 *10
-1.063V -1.875V - 2.427V +1.7498 *10 V ''+1.070 *10 V '' -
-1.260 *10 V ( )c-5
3 = -1.829 *10
(6.4.102)
Integrating by method of trigonometric series with accelerated convergence we have,
Interval = span, L=50m. ∴ ≤ ≤0 x 50
Boundary conditions;
2 2 2 2V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
170
3 3 3 3V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
4 4 4 4V (0) = 0; V ''(0) = 0; V (L) = 0; V ''(L) = 0
We seek the solution in the form;
2 2 2
3 3 3
V (x) = f (x)+g (x)
V (x) = f (x)+g (x) (6.4.103)
4 4 4V (x) = f (x)+g (x) (6.4.103)
Substituting eqn (6.4.103) into eqn (6.4.102) we obtain
-3 -3 -4
4 3 4 312.611f '' -1.2573 *10 f +12.611g '' -1.2573 *10 g = -1.400 *10 (a)
iv iv iv -3
2 3 4 3 4
-7 iv iv iv -3
3 2 3 4 3 4
-7 -5
3
-62.6825f -1.063f -33.8433f +9.031*10 f ''+1.978f ''
-1.0533 *10 f - 62.6825g -1.063g -33.8433g +9.031*10 g ''+1.978g ''
-1.0533 *10 g =1.453 *10 (b)
iv iv iv -3 -4
2 3 4 3 4
-4 iv iv iv -3 -4
3 2 3 4 3 4
-4 -5
3
-1.063f -1.875f - 2.427f +1.7498 *10 f ''+1.070 *10 f '' -
-1.260 *10 f -1.063g -1.875g - 2.427g +1.7498 *10 g ''+1.070 *10 g '' -
-1.260 *10 g = -1.829 *10 (c)
⇒ -3
4 3 12.611f '' -1.2573 *10 f =P(x) (a)
iv iv iv -3
2 3 4 3 4
-7
3 2
-62.6825f -1.063f -33.8433f +9.031*10 f ''+1.978f ''
-1.0533 *10 f =P (x) (b)
iv iv iv -3 -4
2 3 4 3 4
-4
3 3
-1.063f -1.875f - 2.427f +1.7498 *10 f '' +1.070 *10 f '' -
-1.260 *10 f =P (x) (c)
where,
-3 -4
1 4 3P (x) =12.611g '' -1.2573 *10 g +1.400 *10 (a)
iv iv iv -3
2 2 3 4 3 4
-7 -5
3
P (x) = -62.6825g -1.063g -33.8433g +9.031*10 g ''+1.978g ''
-1.0533 *10 g -1.453 *10 (b)(6.4.104)
iv iv iv -3 -4
3 2 3 4 3 4
-4 -5
3
P (x) = -1.063g -1.875g - 2.427g +1.7498 *10 g ''+1.070 *10 g '' -
-1.260 *10 g +1.829 *10 (c)
We seek the complimentary function 2g (x) , g3(x) and g4(x) in the form of sine series
as follows.
171
∞
∞
∞
∑
∑
∑
n2 2n
n=1
n3 3n
n=1
n4 4n
n=1
α xg (x) = a sin
L
α xg (x) = a sin
L
α xg (x) = a sin
L
(6.4.105)
where nα = nπ
The auxiliary function can be sought in terms of algebraic function as follows.
2 3
2 0 1 2 3f (x) = A + A x + A x + A x (a)
2 3
3 0 1 2 3f (x) =B +B x +B x +B x (b) (6.4.106)
2 3
4 0 1 2 3f (x) = C +C x +C x +C x (c) (6.4.106)
The constants 0 0 0 1 1 1 2 2 2 3 3 3A , B , C , A , B , C , A , B , C , A , B ,C are easily defined using
the boundary conditions.
In section 6.3.5 we noted that for simply supported span the boundary conditions are
such that all the above constants are zero.
Hence,
2 2
3 3
4 4
f (x) =P (x) = 0
f (x) =P (x) = 0
f (x) =P (x) = 0
} ⇒
2 2
3 3
4 4
V (x) = g (x)
V (x) = g (x)
V (x) = g (x)
(6.4.107)
Consequently eqn (6.4.104) becomes
-3 -4
4 312.611g '' -1.2573 *10 g +1.400 *10 = 0 (a)
iv iv iv -3
2 3 4 3 4
-7 -5
3
-62.6825g -1.063g -33.8433g +9.031*10 g ''+1.978g ''
-1.0533 *10 g -1.453 *10 = 0 (b) (6.4.108)
iv iv iv -3 -4
2 3 4 3 4
-4 -5
3
-1.063g -1.875g - 2.427g +1.7498 *10 g ''+1.070 *10 g '' -
-1.260 *10 g +1.829 *10 = 0 (c)
But,
∞
∞
∞
∑
∑
∑
n2 2n
n=1
2 n2 2n n
n=1
iv 4 n2 2n n
n=1
α xg (x) = a sin
L
α xg ''(x) = - a α sin
L
α xg (x) = a α sin
L
(a)
and
172
∞
∞
∞
∑
∑
∑
n3 3n
n=1
2 n3 3n n
n=1
iv 4 n3 3n n
n=1
α xg (x) = a sin
L
α xg ''(x) = - a α sin
L
α xg (x) = a α sin
L
(b) (6.4.109)
Also,
;
∞
∞ ∞
∑
∑ ∑
4 n4 4n n
n=1
'' 2 iv 4n n4 4n n 4 4n n
n=1 n=1
α xg (x) = a α sin
L
α x α xg (x) = - a α sin g (x) = a α sin
L L
(c)
Substituting eqn. (6.4.109) into eqn. (6.4.108) we obtain;
( )a∑ ∑
- 1.2573α x2 -3 -4n-12.611 a α *10 a Sin = -1.400 *10n4n 3n L
1.063 33.8433
( )b
∑ ∑ ∑ ∑ ∑− −
4 4 4 -3 2 2 -7-62.6825 a α a α a α - 9.031* 10 a α - 1.978 a α - 1.0533 * 10 a *n n n n4n 4n2n 3n 3n n 3n
α x -6nSin = -2.123 * 10
L
( )c
∑ ∑ ∑ ∑ ∑ ∑
4 4 4 -3 2 -4 2 -3-1.063 a α - 1.875 a α - 2.427 a α - 1.7498 * 10 a α - 1.07 * 10 a α - 1.260 * 10 a *n n n n n4n 4n2n 3n 3n 3n
α x -5nSin = - 2.046 * 10
L (6.4.110)
Equations (6.4,110) will always be satisfied if
2 -3 -4
n 4n 3n(-12.611α )a - (1.2573 *10 )a = -1.400 *10 (6.4.111)
( )−
4 4 -3 2 -7
n 2n n n 3n
4 2 -5
n n 4n
(-62.6825α )a - (1.063α +9.031*10 α +1.0533 *10 )a
33.8433α +1.978α a = -1.453 *10 (6.4.112)
( )−
4 4 -3 2 -3
n 2n n n 3n
4 -4 2 -5
n n 4n
-(1.063α )a - (1.875α +1.7498 *10 α +1.260 *10 )a
2.427α +1.07 *10 α a = -1.829 *10 (6.4.113)
For n = 1, 2 -3 4 -5
1 1 1α =π, α = 3.948 *10 α =1.5585 *10
Substituting into eqn.(6.4.111, 6.4.112, & 6.4.113) gives
( )
( )
( )
a
b
c
-2 -3 -4
41 31
-4 -5 -3 -5
21 31 41
-5 -3 -5 -5
21 31 41
-4.9786 *10 a -1.2573 *10 a = -1.400 *10
-9.7691*10 a -5.2325 *10 a -8.3362*10 a =1.453 *10
-1.6567 *10 a -1.2961*10 a -3.8247 *10 a = -1.829 *10
(6.4.114)
Solving eqns.[6.4.114 (a), (b) & (c)] we obtain
, ,-2 -2 -3
21 31 41a = -3.557 *10 a =1.449 *10 a = 2.446 *10
173
Substituting values of a21, a31, a41, and eqn.(6.4.5) into eqn.(6.4.7) we obtain
∴ -2
2
-2
3
-3
4
πx V (x) = -3.557 *10 Sin
50
πx V (x) =1.449 *10 Sin
50
πx V (x) = 2.446 *10 Sin
50
Table 6.17 Variation of flexural, torsional & distortional
displacements along the length of the girder
(Double cell mono symmetric section)
Distance (x)
from Left
Support (m)
πx
Sin50
V2(x)
X 10-3
V3(x)
X 10-3
V4(x) X 10-3
0 0.000 0.000 0.000 0.000
5 0.309 -10.991 4.477 0.756
10 0.588 -20.915 8.520 1.438
15 0.809 -28.776 11.722 1.979
20 0.951 -33.827 13.780 2.326
25 1.000 -35.570 14.490 2.446
30 0.951 -33.827 13.780 2.326
35 0.809 -28.776 11.722 1.979
40 0.588 -20.915 8.520 1.438
45 0.309 -10.991 4.477 0.756
50 0.000 0.000 0.000 0.000
174
6.5 DISTORTIONAL ANALYSIS OF A THIN WALLED BOX GIRDER: ( Numerical Example ) 6.5.1 Introduction
In this section, numerical analysis of a single cell doubly symmetric box girder
subjected to distortional loading is carried out using the present formulation. In order
to compare the results with those obtained by using other methods of analysis, e.g.,
beam on elastic foundation ( B.E.F) analogy, the example problem used by Osadebe
and Mbajiogu (2006) was chosen.
The box structure is a single cell doubly symmetric, simply supported box girder
with the following parameters.
Length, L = 10m; depth (height of section) h = 1.25m; width b = 2.50m; flange
thickness
tf = 0.200m; web thickness, tw = 0.250m. E = 213 x 106 KN/m2 ; v = 0.23. Distortional
load q = 50KN.
6.5.2 Strain mode diagrams and evaluation of Vlasov coefficients
Figure 6.5.1 shows the strain mode diagrams for the numerical example problem,
from where the Vlasov coefficients are evaluated using the procedures discussed in
chapter four. The summary of the coefficients is given on Table 6.16.
0.208
0.208 0.4166
0.4166
(b) Distortion
ψ3
0.2604
0.2604
0.2604 0.2604
0.2604
0.2604
(a) Warping function
x
y
ϕMω =3
(c) Bending moment due to distortion
1.079 EI
1.079 EI
1.079 EI 1.079 EI
1.079 EI
1.079 EI
M3
Fig. 6.5.1 Strain mode diagrams for numerical example problem
175
Table 6.18 Coefficients for numerical example problem
A ij = a ji b ij = b ji ckj = c jk r kh = r hk
11
12
13
22
23
33
a = 0.472
a = 0.000
a = -0.034
a =1.497
a = 0.000
a = 0.40
-3
33s = 2.07x10
11
12
13
22
23
33
b = 0.625
b = 0.000
b = 0.000
b =1.000
b = 0.000
b = 0.173
11
12
13
22
23
33
34
c = 0.635
c = 0.000
c = 0.000
c =1.000
c = 0.000
c = 0.173
c = 0.033
11
12
13
22
23
33
34
44
c = 0.635
c = 0.000
c = 0.000
c =1.000
c = 0.000
c = 0.173
c = 0.033
r =1.367
6.5.3 Differential equation for distortional equilibrium (numerical example)
From the formulation of flexural-distortional equilibrium equation for doubly
symmetric section in section 5.2.2 we noted that there was no interaction between
flexural strain mode 2 and distortional strain mode 3. Therefore, the two equations
obtained for flexure and distortion eqns.(5.2.19 and 5.2,20) are uncoupled and can
be solved independent of each other. The equation obtained for distortional
equilibrium for single cell doubly symmetric section is given by eqn. (5.2.20), i.e.,
iv 2 4
3 3 3 3V -α V ''+ 4β V =K (6.5.1)
where, 2 333
33
ksα =
r, 4 33
3
33
s4β =
a and 3
3
33
qK =
Ea
6.5.4 Distortional analysis ( numerical example problem) The governing equation for distortional equilibrium for the single cell doubly
symmetric section is
iv 2 4
3 3 3 3 3 3V -α V ''+ 4β V =K ( 6.5.1)
Where, 2 433 33 33 3 3
33 33 33
ks s qα = ; 4β = ; K =
c a a E
For the single cell doubly symmetric section we have;
-3
33 33 33 34 44 33a = 0.0396, c = r = 0.173, r = 0.1954, r =1.367, s = 2.07 *10
υ9 2 9 2E = 24 *10 N/m , G = 9.6 *10 N/m , = 0.25
176
3 3q = q x b = 50 x 2.5 =125KN
2 -2
3
4 -2
α = 3.111*10
4β = 5.227 *10
Let the solution be sought in the form; λx3V (x) = e (6.5.2)
∴ 2 λx3V (x) = λe , 2 2 λx
3V (x) = λ e , iv 4 λx3V (x) = λ e
Substituting eqn.(6.5.2) and its derivatives into eqn.(6.5.1) we obtain the auxiliary
equation for homogeneous condition as follows.
( )λx 4 2 2 4e λ -α λ + 4β = 0 ⇒ 4 2 2 4λ -α λ + 4β = 0
Let 2υ= λ ∴ 2 2 4υ -α υ+ 4β = 0
⇒2 4 4
1,2
α ± α -16β=
2υ ⇒
2 44
1
α αυ = + - 4β
2 4,
2 44
2
α αυ = - - 4β
2 4
∴2 4
4
1,2 1
α αλ = υ = ± + - 4β
2 4 ;
2 44
3,4 2
α αλ = υ = ± - - 4β
2 4
where 2 -233
33
ksα = = 3.11*10
c,
4 -233
33
s4β = = 5.227 *10
a
-2 -4-2
1,2
3.111*10 9.678 *10λ = ± + -5.227 *10
2 4
-2 -2= ± 1.555 *10 +i 5.2028 *10 = -2± 1.555 *10 +i2.281
Similarly, -23,4λ = ± 1.555 *10 - i * 0.2281
Let -2a =1.555 *10 and b = 0.2281
∴ ( )1
21λ = a+ib ( )
1
22λ = - a+ib ( )
1
23λ = a - ib ( )
1
24λ = - a - ib
Substituting into eqn.(6.5.2) we obtain
( ) ( ) ( ) ( )1 1 1 1
2 2 2 2a+ib - a+ib a-ib - a-ib
3 1 2 3 4V (x) = A e + A e + A e + A e
⇒3 1 2 3 4
V (x) = A coshωx cosφx + A coshωx sinφx + A sinhωx cosφx + A sinhωx sinφx (6.5.3)
177
where =ω real value of
1-2 21.555 *10 + i * 0.2281*
=φ imaginary value of
1-2 21.555 *10 + i * 0.2281*
For φ = a+ib ,ω =Re a+ib , φ =Im a+ib
∴ a+ib =ω+iφ ⇒ ( )2 2 2a+ib = ω+iφ =ω +2iωφ -φ
( )2 2a+ib = ω -φ +2iωφ
Equating equivalent terms we obtain
2 2ω -φ = a ……….(a) and 2ωφ = b ……..(b)
Squaring both sides of expressions (a) and (b) and adding up we have
( )2
2 2 2 2 2 2ω -φ + 4ω φ = a +b ⇒ 4 2 2 4 2 2ω - 2ω φ +φ + 4ω φ = a+b
( )⇒ 4 2 2 4 2 2ω +2ω φ +φ = a +b
( ) ( )⇒2
2 2 2 2ω +φ = a +b ∴ 2 2 2 2ω +φ = a +b ……….(c)
Adding expressions (a) and (c) we obtain
2 2 22ω = a+ a +b
⇒2 2a+ a +b
ω =2
= -2 -4 -21.555 *10 2.418 *10 +5.2029 *10
+ = 0.349412 2
Subtracting (a) from (c) we obtain
2 2 22φ = a +b -a
⇒2 2a +b -a
φ =2
( ) ( )
22-2 -21.555 *10 + 0.2282 -1.555 *10
= = 0.32642
Thus, ω = 0.34941, φ = 0.3264
3 homo. 1 2 3 4V (x) = A coshωx cosφx + A coshωx sinφx + A sinhωx cosφx + A sinhωx sinφx
Let the particular integral be sought in the form; 3V (x) =B (6.5.4)
This solution must satisfy eqn.(6.5.1)
i.e, iv 2 4
3 3 3 3V -α V ''+ 4β V =K
From eqn.(6.5.4) we obtain iv
3 3V '' = 0, V = 0
Substituting into eqn.(6.5.1) we obtain
178
4
44β B =K ∴ 3
4
KB =
4β ⇒ ( ) 3
3 4particular
KV x =
4β
where 3
-433 9
33
q 125 *10K = = =1.3152*10
a E 0.0396 * 24 *10
Boundary Conditions;
V(0) = 0; V(L) = 0; V''(0) = 0; V''(L) = 0
33 1 2 3 4 4
KV (x) = A coshωx cosφx + A coshωx sinφx + A sinhωx cosφx + A sinhωx sinφx +
4β
(6.5.5)
[ ]3 1 2 3 4V ''(x) = 2 -A sinhωx sinφx + A sinhωx cosφx - A coshωx sinφx + A coshωx cosφx
∴ ⇒ 31 4
KV(0) = 0 : A = -
4β (a)
⇒3 4V ''(0) = 0 : 2A = 0 ⇒ 4A = 0 (b)
⇒ 31 2 3 4 4
KV(L) = 0 : -16.3522A - 2.0117A -16.3221A - 2.008A + = 0
4β (c)
V''(L) = 0 : ⇒ 1 2 32.008A -16.3221A +2.0117A = 0 (d)
Solving expressions (a), (b), (c), and (d) we obtain
31 4
KA = -
4β; -3 3
2 4
KA = 7.873 *10
4β; 3
3 4
KA =1.0621
4β; 4A = 0
Substituting values of 1 2 3 4A , A , A , & A into eqn.(6.5.5) we obtain
-33
3 4
KV (x) = -Coshωx Cosφx +7.873 *10 * Coshωx Sinφx +1.062* Sinhωx Cosφx +1
4β
where, 4 -24β = 5.227 *10 , -4
3K =1.3153 *10 , ⇒ -33
4
K= 2.5163 *10
4β
ω = 0.3494, φ = 0.3264
∴
-33 -3
1.0621Sinhωx * Cosφx -Coshωx * CosφxV (x) = 2.5163 *10
+7.873 *10 * Coshωx * Sinφx +1
Table 6.19 Variation of distortional displacement along the girder
length:(numerical example problem)
Distance (x)
from Left
Support (m)
1+ 1.0621 *
sinhωx * cosφx
coshωx * cosφx- 7.873 x 10-3
*
coshωx * sinφx
Distortional displacement
V3(x)
179
0 1.000 -1.0000 0.0000 0.000 x 103 (m)
2 1.6388 -0.9964 0.0060 1.631
4 1.5287 -0.5626 0.0163 2.472
5 0.8193 0.1808 0.0232 2.575
6 -0.6085 1.5610 0.0301 2.473
8 -6.4690 7.0850 0.0327 1.632
10 -16.3340 16.3506 -0.01580. 0.000
6.5.5 Results From Other Methods of Analysis
Osadebe and Mbajiogu (2006) carried out a distortional analysis of the numerical
example problem using
(a) Vlasov’s theory with provision for shear deformations
(b) Beam on elastic foundation (BEF) analogy.
Table 6.18 shows the results they obtained for the two methods of analysis.
Substituting the actual values of q = q*b = 50 * 2.5 =125KN, and E = 24 *106 KN/m2
(for concrete), we obtain values in Table 6.19 which are comparable with the values
from the present formulation.
Table 6.20:Distortional Displacement V(x) (m) (Osadebe & Mbajiogu (2006))
X (m) 0 2 4 5 6 8 10
V(x)
(BEF)
0 31.654759
q
E 48.070156
q
E 50.074227
q
E 48.070217
q
E 31.654892
q
E
0
V(x)
Osadebe
&
Mbajiogu
0 32.479423
q
E 48.32075
q
E 49.86876
q
E 48.03229
q
E 32.479381
q
E
0
180
Table 6.21 Comparison of results from present formulation with
results from other methods of analysis (Numerical example)
X (m) Distortional displacement V3(x)
Present formulation x 10-3 (m)
Osadebe & Mbajiogus’ formulation x 10-3 (m)
BEF Analogy method x 10-3 (m)
0 0.000 0.000 0.000
2 1.631 1.691 1.649
4 2.472 2.516 2.505
5 2.575 2.597 2.608
6 2.473 2.516 2.503
8 1.632 1.691 1.649
10 0.000 0.000 0.000
181
CHAPTER SEVEN
7.0 DISCUSSION OF RESULTS, CONCLUSIONS AND RECOMMENDATIONS
7.1 General
In this work the torsional-distortional response of different cross sectional box
girder profiles were studied. They include (a) Single cell doubly symmetric section,
(b) multi cell doubly symmetric section, (c) single cell mono symmetric section and
double cell mono symmetric section.
Governing differential equations of equilibrium were derived for analysis of the
structure under the following modes of interaction.
(i) Flexural- torsional interaction
(II) Flexural distortional interaction
(iii) Torsional-distortional interaction
(iv) Flexural-torsional-distortional interaction
The results obtained from the analysis of a single cell doubly symmetric section,
numerical example problem, are also presented for the purpose of comparison.
7.2 Discussion of Results
7.2.1 Governing differential equations of equilibrium
The summary of the differential equations of equilibrium derived for each of these
modes of interaction is presented in Table 5.1, from where it can be observed that
the equilibrium equations for all doubly symmetric sections are the same for every
strain mode interaction. For example, the same equilibrium equation was obtained for
both single cell and multi cell doubly symmetric section for torsional-distortional
analysis. Similarly, the same equilibrium equations were derived for mono symmetric
sections for every strain modes interaction.
In doubly symmetric section it was observed that there was no interaction
between strain mode 2, (minor axis bending) and strain modes 3 and 4 (distortion
and torsion). Thus, the differential equations of equilibrium governing flexure (in
minor axis) is independent of distortion and torsion.
In mono symmetric sections, on the other hand, all the three strain modes interact
with each other and the equations of equilibrium governing any of the modes of
interaction were always coupled.
182
For doubly symmetric and mono symmetric sections the governing equations of
equilibrium are independent of strain mode 1 (major axis bending). In other words,
the major axis strain mode does not interact with any other strain modes in either
doubly symmetric or mono symmetric sections. It is for this reason that torsional
loads can be separated from the general loads on the bridge structure for the
purpose of torsional-distortional analysis.
For non-symmetric sections all the four strain modes interact in such a way that
the equilibrium conditions are inseparable. This means that a consideration of
torsional-distortional analysis of a non symmetric section can not be accomplished
without a consideration of all the strain modes interacting on the structure giving rise
to displacement in major axis V1(x), displacement in minor axis V2(x), distortion
V3(x) and torsion V4(x). It is for this reason that torsional-distortional analysis of non-
symmetric sections is most challenging, requiring a consideration of the general
loads on the box girder structure.
The governing differential equations of equilibrium presented in Table 5.1(a), (b)
and (c) are for analysis of deformable cross sections where distortion is not
prevented by the use of diaphragms or intermediate stiffeners, which is outside the
scope of this work.
7.2.2 Single cell doubly symmetric section
Fig. 7.2.1 shows the variation of torsional displacement along the length of the
girder while Fig.7.2.2 shows the variation of distortional displacement along the
length of the girder. Fig.7.2.1 is the result obtained for flexural-torsional analysis
while Fig.7.2.2 is for flexural distortional analysis. However, since there was no
interaction between flexure & torsion and between flexure & distortion in doubly
symmetric sections, both results represent the results for flexural analysis and
torsional analysis of single cell doubly symmetric section respectively. In the two
analysis , the flexural displacements are zero as a result of absence of component of
torsional load in minor axes plane and as a result of non interaction of strain mode 2
with other strain modes.
Fig.7.2.3 is the result for flexural-torsionl-distortional analysis of single cell doubly
symmetric section. It shows the variation of torsional displacement and distortional
displacement along the length of the girder, the flexural displacement being also zero
in this case.
183
0 5 10 15 20 25 30 35 40 45 500
20
40
Distance Along the Length of the Girder (m)
Dis
tort
ional D
ispla
cem
ent
(mm
)
0 5 10 15 20 25 30 35 40 45 500
1
2
3
4
5
Distance Along the Length of the Girder (m)Tors
ional D
ispla
cem
ent
(mm
)
Fig.7.2.1: Variation of torsional displacement along the length of the girder (Single cell doubly symmetric section)
Cross section
Fig.7.2.2: Variation of distortional displacement along the length of the girder (Single cell doubly symmetric section)
Cross section
184
0 5 10 15 20 25 30 35 40 45 500
10
20
30
40
50
Distance Along the Length of the Girder (m)
Dis
pla
cem
ent
(mm
)
Distortional displacement
Torsional displacement
A closer look at Figs 7.2.1, 7.2.2 and 7.2.3 shows that a maximum (mid span)
torsional displacement of about 2mm was obtained for torsional analysis while 27mm
deplanation was obtained for distortional analysis. These values remain relatively
unchanged in the case of torsional distortional analysis as could be seen from
Fig.7.2.3. This shows that the effect of interaction between torsion and distortion is
negligible in single cell doubly symmetric section.
7.2.3 Multi-cell doubly symmetric section
Figures 7.2.4 and 7.2.5 show the variation of torsional deformation and distortional
deformation respectively, along the length of the girder, for multi-cell doubly
symmetric section. The mid span torsional deformation is 2.6mm, while the mid span
deplanation is 5.9mm.
The result of torsional distortional analysis for multi cell doubly symmetric section
is graphically presented in Fig 7.2.6. The maximum (mid san ) deformation was 6mm
while torsional deformation was 3mm. These values are very close to those obtained
by separate analysis for torsion and distorsion. Consequently, the equations of
equilibrium
Fig.7.2.3: Variation of torsional and distortional displacements along the length of the girder (Single cell doubly symmetric section)
Cross section
185
governing torsional-distortional response of doubly symmetric sections can be
0 5 10 15 20 25 30 35 40 45 500
1
2
3
4
5
Distance Along the Length of the Girder (m)
Tors
ional dis
pla
cem
ent
(mm
)
0 5 10 15 20 25 30 35 40 45 500
2
4
6
8
10
Distance Along the Length of the Girder (m)
Dis
tort
ional D
ispla
cem
ent
(mm
)
Fig.7.2.4: Variation of torsional displacement along the length of the girder (Multi- cell doubly symmetric section)
Cross section
Fig.7.2.5: Variation of distortional displacement along the length of the girder (Multi- cell doubly symmetric section)
Cross section
186
0 5 10 15 20 25 30 35 40 45 500
5
10
15
Distance Along the Length of the Girder (m)
Dis
pla
cem
ent
(mm
)
Distortional Displacement
Tortional Displacement
decoupled to obtain two independent equations for distorsional displacement V3(x)
and torsional displacement V4(x). This implies that the weak interaction between
torsion and distortion can be ignored in all doubly symmetric sections, and that
analysis for torsion and distortion can be independently carried out.
7.2.4 Single cell mono symmetric section
The results for the analysis of single cell mono symmetric section are presented
as follows.
Flexural- torsional analysis; Fig. 7.2.7
Flexural-distortional analysis; Fig.7.2.8
Torsional-distortional analysis; Fig.7.2.9
Flexural-torsional-distortional analysis; Fig.7.2.10
For flexural torsional analysis, the mid span flexural displacement was 80mm
which dropped to 4mm in flexural distortional analysis. This shows that distortional
stresses tend to minimize flexural stresses. A comparison of maximum (mid span)
torsional values in Figs. 7.2.7 and 7.2.9 also shows that distortional stresses do not
encourage torsion in torsional- distortional analysis. However, when the three
modes of interaction
Fig.7.2.6: Variation of torsional and distortional displacements along the length of the girder (Multi- cell doubly symmetric section)
Cross section
187
0 5 10 15 20 25 30 35 40 45 500
20
40
60
80
100
Distance Along the Length of the Girder (m)
Dis
pla
cem
ent
(mm
)
Flexural displacement
Torsional displacement
0 5 10 15 20 25 30 35 40 45 500
20
40
60
80
Distance Along the Length of the Girder (m)
Dis
pla
cem
ent
(mm
) Flexural displacement
Distortional displacement
Fig.7.2.7: Variation of flexural and torsional displacements along the length of the girder (Single cell mono symmetric section)
Fig.7.2.8: Variation of flexural and distortional displacements along the length of the girder (Single cell mono symmetric section)
cross section
cross section
188
0 5 10 15 20 25 30 35 40 45 500
10
20
30
40
50
60
Distance Along the Length of the Girder (m)
Dis
pla
cem
ent (m
m)
Distortional displacement
Torsional displacement
0 5 10 15 20 25 30 35 40 45 500
10
20
30
40
50
60
70
Distance Along the Length of the Girder (m)
Dis
pla
cem
ent
(mm
)
Flexural displacement
Distorsional displacement
Torsional displacement
interplay, we notice from Fig.7.2.10 that distortional deformation (43mm) is higher
than torsional deformation (33mm) which in turn is higher than torsional displacement
(4mm).
Fig.7.2.9: Variation of torsional and distortional displacements along the length of the girder (Single cell mono symmetric section)
cross section
Fig.7.2.10: Variation of flexural, torsional and distortional displacements along the length of the girder (single cell mono symmetric section)
cross section
189
Generally we observe that distortional deformations (and hence stresses) were
consistently higher than higher than flexural and torsional stresses where ever either
or both of flexure and torsion interact with distortion. Also flexural deformations were
consistently higher than torsional deformations when ever both strain modes interact.
7.2.5 Double cell mono symmetric section
The results for the analysis of double cell mono symmetric section are presented
as follows. Fig. 7.2.11 is for flexural- torsional analysis, Fig.7.2.12 is for flexural
distortional analysis, Fig.7.2.13 is for torsional-distortional analysis, while Fig.7.2.14
is for flexural torsional- distortional analysis.
As in the case of single cell mono symmetric section, comparison of Figs. 7.2.11
and 7.2.12 shows that distortional stresses do not encourage deflection of the box
girder structure since mid span flexural deformation for flexural-torsional interaction
analysis was 33mm but for flexural- distortional interaction analysis it came down to
9mm.
When the three modes of interaction interplay we notice from Fig. 7.2.14 that
flexural deformations are significantly higher than distortional and torsional
deformations. This appears to represent the true picture of state of stress in bridge
structures where by more attention is given to flexural stresses than distortional and
torsional stresses in design.
0 5 10 15 20 25 30 35 40 45 500
10
20
30
40
50
60
Distance Along the Length of the Girder (m)
Dis
pla
cem
ent (m
m)
Flexural displacement
Torsional displacement
Fig.7.2.11: Variation of flexural and torsional displacements along the length of the girder (Double cell mono symmetric section)
cross section
190
0 5 10 15 20 25 30 35 40 45 500
5
10
15
20
25
30
Distance Along the Length of the Girder(m)
Dis
pla
cem
ent
(mm
)
Distortional displacement
Flexural Displacement
0 5 10 15 20 25 30 35 40 45 500
5
10
15
20
25
30
35
40
45
50
Distance Along the Length of the Girder (m)
Dis
pla
cem
ent (m
m)
Distortional displacement
Torsional displacement
Fig.7.2.12: Variation of flexural and distortional displacements along the length of the girder (Double cell mono symmetric section)
cross section
Fig.7.2.13: Variation of torsional and distortional displacements along the length of the girder (Double cell mono symmetric section)
cross section
191
0 5 10 15 20 25 30 35 40 45 500
10
20
30
40
50
60
Distance Along the Length of the Girder (m)
Dis
pla
cem
ent
(mm
)
Flexural displacement
Distorsional displacement
Torsional displacement
7.3 COMPARISON OF RESULTS
Fig.7.3.1 shows the results obtained for distortional analysis of the numerical
example problem in section 6.5.1, using the present formulation. Table 6.19 shows
the results obtained from the present formulation and by Osadebe and Mbajiogu
(2006) for the same numerical example problem using (a) Vlasov’s theory with
modification for shear and (b) Beam on elastic foundation (BEF) analogy method.
The three results are graphically represented in Fig.7.3.2 from where it can be seen
that they are in agreement.
Fig.7.2.14: Variation of flexural, torsional and distortional displacements along the length of the girder (Double cell mono symmetric section)
cross section
192
0 1 2 3 4 5 6 7 8 9 100
0.5
1
1.5
2
2.5
3
3.5
Distance Along the Length of the Girder (mm)
Dis
tort
ional D
ispla
cem
ent
(mm
)
0 1 2 3 4 5 6 7 8 9 100
0.5
1
1.5
2
2.5
3
3.5
Distance Along the Length of the Girder (m)
Dis
tort
ional D
ispla
cem
ent
(mm
) Present formulation
Osadebe & Mbajiogu
BEF analogy
Fig. 7.3.1: Variation of distortional displacement along the length of the girder (numerical example problem)
cross section
Fig.7.3.2: Variation of distortional displacements along the length of the girder ( comparison of results)
cross section
193
7.4 CONCLUSIONS
In doubly symmetric profiles distortional deformations were found to be higher
than torsional deformations and the effect of interaction between torsion and
distortion is quite negligible. In fact, there was no interaction between (i) major axis
bending and distortion (ii) major axis bending and torsion (iii) minor axis bending and
distortion (iv) minor axis bending and torsion.
In mono symmetric section it was found that distortional stresses tent to reduce
flexural deformation more than torsional deformation. However, when the three strain
modes of flexure, torsion and distortion interplay, it was found that distortional
deformation was higher than flexural deformations only in single cell mono symmetric
section. In double cell mono symmetric profile flexural deformation was appreciably
higher than distortional and torsional deformations.
Application of the derived governing differential equations of equilibrium in the
analysis of doubly symmetric and mono symmetric sections gave good results.
Comparison of the results with those obtained using other methods of analysis show
that the derived equations of equilibrium are reliable.
For doubly symmetric and mono symmetric sections the governing equations of
equilibrium are independent of strain mode 1. In other words, the major axis strain
mode does not interact with any other strain modes in either doubly symmetric or
mono symmetric profiles. In non symmetric section all the four strain modes interact
in such a way that the equations of equilibrium are inseparable.
In real life structure with generalized loading, torsional- distortional interaction
analysis can not be separated from flexural- torsional- distortional interaction
analysis. Thus, in other to obtain reasonable results flexural–torsional-distortional
analysis were incorporated in the analysis.
As would be expected, multi cell sections offered greater resistance to flexural,
torsional and distortional deformations than single cell sections. This is because of
increase in the number of vertical members in multi cell sections contributing to the
rigidity of the structure. The introduction of middle member in double cell mono
symmetric cross section decreased torsional deformation and distortional
deformation by 40% and 67% respectively, while the lateral flexural deformation was
increased by 9%. The torsional-distortional deformations on mono symmetric
sections can not be compared with that of doubly symmetric sections because the
torsional moments are not the same even though the loadings are the same.
194
7.5 CONTRIBUTION TO KNOWLEDGE
The major contributions of this work to knowledge are as follows.
• For box girders possessing doubly symmetric sections, it was found that
neither torsion nor distortion interacts with flexure. Differential equations of
equilibrium were therefore derived for independent analysis of the structure in
flexure, torsion and distortion.
• For box girders possessing mono symmetric sections, it was ascertained that
both torsion and distortion interact with bending about the non symmetric axis
only. An independent equation for flexural analysis is therefore derived for the
single axis of symmetry.
• On the basis of (2) above differential equations of equilibrium for the following
cases were derived for box girders with mono symmetric section.
Case 1; flexural –torsional analysis (eqn.5.3.9),
Case 2; flexural-distortional analysis (eqn.5.3.24),
Case 3; torsional-distortional analysis (eqn.5.3.32),
Case 4; flexural-torsional-distortional analysis (eqn. 5.3.50 )
• For non symmetric sections, it was found that torsion, distortion and bending
interact and are therefore inseparable. Hence, a set of coupled differential
equations of equilibrium for generalized analysis of non symmetric sections in
flexure, torsion and distortion was derived.
• From all the above, the obtained differential equations of equilibrium can be
used to analyze box girders possessing doubly symmetric sections, mono
symmetric sections and non symmetric sections. Numerical examples
demonstrating the applicability of the obtained equations are given for box
girders possessing doubly symmetric section and mono symmetric section.
• Generalized equations for the flexural-torsional-distortional analysis of box
girder structures possessing non symmetric section, to the best of knowledge
of the author, is not available in technical literature. This makes this work
unique as it provides a holistic approach to torsional-distortional analysis of
thin walled box girder structures.
195
• The study has provided foundation for ramification of research work in elastic
analysis of thin walled structures made of concrete, steel, timber and other
structural materials. It only suffices to use the appropriate physical properties
(E, G, υ ) of a chosen material.
7.6 RECOMMENDATIONS FOR FURTHER WORK
Knowledge of higher engineering mathematics plays a vital role not only in the
formulation of the differential equations of equilibrium but also in the integration of the
derived equations to obtain solutions. In the case of non-symmetric sections, four 4th
order coupled differential equations of equilibrium were derived but not applied to
solution of problems. It is recommended that further work be carried out on non
symmetric sections so as to demonstrate the applicability of the derived equations
for flexural-torsional-distortional analysis of such profiles.
During the process of evaluation of warping functions it was found that material
thickness does not affect the warping function. However, the bending moments due
to distortion of cross sections vary with the thickness of the cross sectional members.
A parametric study on bridge deck arrangements is recommended. Such study would
high light the effect of cross sectional parameters on the distortional bending moment
of box girder sections and invariably on the torsional-distortional response of box
girder structures.
The present work considered only simply supported box girders in the analysis,
even though the derived governing differential equations are not limited to simply
supported bridge decks. The extension of this work to cover the analysis of bridge
decks with a combination of support conditions (rigid support, continuous support,
pinned support etc.) will bring the concept of this work to a logical conclusion.
196
REFERENCES
Al-Rifaie, W.N., and Evans, H.R. (1979).”An approximate method for the analysis
of box girder bridges that are curved in plan”. Proc., Int. Assoc. of Bridges
and Struct. Eng. (IABSE),P 1-21.
American Association of State Highway and Transportation Officials (AASHTO),
(1996). “Standard specifications for highway bridges” Washington, D.C.
American Association of State Highway and Transportation Officials (AASHTO),
(1998).“Load and Resistance Factor Design, LRFD, bridge design
specifications” Washington , D.C.
Bakht, B. and Jaeger, L.G. (1985). Bridge deck simplified, Mc Graw-Hill New
York.
Bakht, B. and Jaeger, L.G.(1992),”Simplified method of bridge analysis for the
third edition of OHBDC”. Can.J. Engrg., Ottawa, vol.19 no.4, p551-559.
Bakht, B., Jaeger, L.G., and Cheung, M.S. 91981). “State of the art in analysis of
cellular and voided slab bridges” Can. J. Civ. Engrg., vol.8 no.3, p376-391.
Bazant, Z.P., and El-Nimeiru, M. (9174) “ stiffness method for curved box girders
at initial stress” Journal of Structural Division, vol.100, issue10, p.2071-2090.
Benscoter,S.U. (1954). “A theory of torsion bending for multi-cell beams” Journal
of Applied Mechanics, vol,21, issue 1, p.25-43.
Canadian High way bridge design code (CHBDC 2000).
Canadian Standards Association CSA(1988). “Design of High way bridges”
CAN/CSA- S6-88, Rexdale, Ont.
Chueng,M.S.,Bakht,B., And Jaeger, L.G.(1982). Analysis of box girder gridges by
grillage and orthotropic plate methods” Can. J. Civ. Engrg vol.9 no. 4, p595-
601.
Chueng, M.S., AND Foo, S.H.C. (1995) “Design of horizontally curved composite
box girder bridges: A simplified approach”. Can J. Engrg., Ottawa, vol.22,
no.1, p93-105.
Dabrowski, R. (1968). “Curved thin-walled girders” Translatioon No. 144, Cement
and Concrete Association, London.
Davis, R.E., Bon, V.D., and Semans, F.M. (1982). “Live load distribution in
concrete box girder bridges” Transp. Res. Rec.871, Transportation research
Board, Washington, D.C., P47-52.
197
Dean, D.L. (1994). “Torsion of regular multi-cellular members” Journal of
Structural Engrg., ASCE, vol.120 no.12, p3675-3678.
Evans, H.R.(1984). “Simplified method for the analysis and design of bridge of
regular cross-section”. Proc. NATO Advanced Study Institute on analysis and
design of gridges, Cesme, Izmir, Turkey, vol.74 p95-115.
Evans , H.R., and Shanmugam, N.E. (1984). “Simplified analysis for cellular
structures”. J. Struct. Div., vol.110 no.3, p531-543.
Fan, Z., and Helwig, T.A. (1999).”Behaviour of steel box girders with top flange
bracing”. Journal of Struct, Engrg., vol.125, issue 8, p.829-837.
Fan, Z., and Helwig, T.A. (2002). “Distortional loads and brace forces in steel box
girders” J.Struct. Engrg vol.128 issue 6, p710-718.
Foinquinos, R., Kuzmanovic, B., and Vargas, L.M. (1977). “Influence of
diaphragms on live load distribution in straight multiple steel box girder
bridges”. Proc. 15th Struct. Congr. Vol. 1, ACE, New York.
Fountain, R.S., and Mattock, R.S. (1968). “Composite steel-concrete multi box-
girder bridges” Proc. Can. Struct. Engrg. Conf., Canadian Civil Engineering
Society, Montreal, Quebec, p19-30.
Fu, C.C. and Yang, D. (1996). “Design of concrete bridges with multiple box cells
due to torsion using softened truss model”. ACI Struct J., vol. 93, no. 6, p696-
702.
Goodall, J.K. (1971). “Torsional stiffness of multi-cellular box sections”. Proc.
Conf. on devel. Of bridge design and construction, Lockwood, London
Grant,C. (1993). “Shear flow in multicell sections”. Proceedings Institute of
Mechanical Engineers, London, vol.207 no.4, p247-253.
Hambly, E.C., and Pennells, E. (1975). “Grillage analysis applied to cellular
bridge decks”. Structural Engineer,vol53 no.7, p276-275.
Heins, C.P. “Bending and torsion design in structural members” Lexinton Books,
Massachusetts.
Heins, C.P. (1978). “Box girder bridge design: state -of -the –art”.AISC Engrg
Journal, vol.15 no.4, p 126-142.
Helwig, T.A., and Fan, Z. (2000). Field and computational studies of steel
trapezoidal box girder bridges”. TXDOT report NO.0-1395-3F.
Ho, S., Cheang, M.S., NG, S.F., and Yu, T. (1989). “Longitudinal girder moments
in simply supported bridges by finite strip method” Canadian Journal of Civil
198
Engineering, Ottawa, vol.16 no.5, p.698-703.
Hsu, Y.T., Fu, C.C. and Shelling, D.R. (1995).“EBEF method for distortional
analysis of steel box girder bridges”. Journal of Structural Engineering,
vol.121 issue 3, p.557-566.
Johanston, S.B. and Mattock, A.H. (1967). “ Lateral distribution of loads in
composite box girder bridges”. Transp. Res. Rec. 167, Transportation
Research Board, Washington D.C., P.25-33.
Kissane, R. and Beal, D.B. (1975). “Field testing of horizontally curved steel
girder bridges”. Research report 27, U.S. Dept. of Transportation,
Washington, D.C.
Kollbruner, C.F., and Basler, K(1969), “Torsion in Structures: an Engineering
Approach” Springer, New York.
Meyer, C., and Scordelis, A.C.(1971). ” Analysis of curved folded plate
structures” Journal of Structural Division, vol.97 no.10, p.2459-2480.
Nakai, H, and Yoo, C.H. (1998) “ Analysis and design of curved steel bridges” Mc
Graw –Hill, New York.
Normandin, P., and Massicotte, B (1994). “ Load distribution in composite steel
box girder bridges” Proceedings 4th International conference on short- and
medium-span bridges, Canadian Society for Civil Engineering, Montreal,
Quebec, p.165-176.
Okeil, A.M., and El-Tawil, S. (2004). “Warping stresses in curved box girder
bridges: Case study”. Journal of Bridge Engineering, vol.9, issue 5, p.487-
496.
Ontario Ministry of Transport and Communication OMTC (1992). Ontario
highway bridge deign code, OHBDC, 3rd Ed., Downsview, Ontario.
Ontario Ministry of Transport and Communication OMTC (1983). Ontario
highway bridge deign code, OHBDC, 2nd Ed., Downsview, Ontario.
Osadebe, N.N., and Mbajiogu M.S.W. (2006) “ Distortional analysis of thin-
walled box girders” Nigerian journal of technology, vol.25 no.2, p.36-45
Rekach. V.G. (1978) “Static theory of thin walled space structures” Mir
publishers, Moscow
Salaheldin, M., and Schmidt, L.C. (1991). “Prediction of transverse flange
stresses in a simple box girder”. Journal of Constructional Steel Research,
vol.19 no.3,p.183-202.
199
Sennah, K.M. and Kennedy, J.B.,(2001). “State of the art in design of curved box
girder bridges” Journal of Bridge Engineering, vol.6, issue 3 p.159-167.
Sennah, K.M. and Kennedy, J.B.,(2003). “Literature review in analysis of box
girder bridges” Journal of Bridge Engineering, vol.7 issue 2, p.134-140.
Sennah, K.M., Khaled---- and Kennedy J.B., (2003). Design for shear in curved
composite multiple steel box girder bridges” Journal of Bridge Engineering,
vol.8, issue 3, p144-152.
Trikha, D.N., and Edward, A.D. (1992). “Analysis of concrete box girder before
and after cracking”. Proceeding Institution of Civil Engineers Part 2, London,
no. 59, p.743-762.
Tung,D.H., and Fountain, R.S. (1970). “Approximate torsional analysis of curved
box girders by the M/R method” American Institution of Steel Construction
Engineering Journal, vol.7, no.3, p65-74.
Turkstra, C.J. and Fam, A.R.M. (1978). “Behaviour study of curved box bridges.”
Journal of structural Division, ASCE, vol. 103 issue 3.p453-462.
Varbernov, C.P.(1976), Theory of Elasticity, Technika Press Sofia, 4th edition,
p.254-271
Vlasov, V.Z.(1961). “Thin walled elastic beams”, 2nd Ed., National Science
Foundation, Washington, D.C.
Vlasov V.Z. (1958). “Tjin walled space structures” Gosstrojizdat, Moscow.
Yabuki, T., Arizumi, Y., and Vinnakota, S. (1995).”Strength of thin walled box
girders curved in plan” Journal of Structural Engineering, vol.121, issue 5, p.
907-914.
200
APPENDIX ONE
COMPUTATION OF WARPING FUNCTIONS
A1(a) Single cell mono-symmetric box girder section
Referring to Fig.A1 we obtain the following:
Cross sectional area F = 4.0446*106 mm2
Enclosed area A= 16,744,500mm2
Fy = 0, symmetry on y-axis
Fz = 2(3557 x 203 x 3050/2) + (3660 x 203 x 3050) = 4.648x10 9mm3
Integrating from point v to point 1, Fig.4.3.3b, we obtain; Bρ = 0, s = 3660, t = 203 mm
∫
22A 2x16,744500ψ = = = 375718mm
ds 7230 2x3557 3660+ +
t 203 203 203
6
1
44722ω = (0 - )(3660) = - 6.774x10
203mm2
From point 1 to point 2; 6
2ω = - 6.774x10
From point 1 to point 3;
o
Bρ = 3660sin59.04 = 3138.4mm
6 6 2
3
185083ω = (3138.4 - )(3556.88) - 6.744x10 = -2.194x10 mm
203
From point 3 to point 4; Bρ = 3050, s = 3660, t =203
∴66 2
4
185083ω = (3050 - )(3660) - 2.194x10 = 2.194x10 mm
203
By reason of symmetry we have,
915 7320
2745 2745 3660
FIG. A1: Single cell box girder section
3050
915
t = 203
201
6
5 1ω = -ω = 6.774x10 and 6 2
6 2ω = - ω = - 6.774X10 mm
The warping function B
ω with respect to the pole B for this profile is shown in
Fig.4.3.3 (d). The _
x and _
y diagrams are plotted in Figs. 4.3.3(b) and (c) respectively.
The moment of areas and product moments of areas are obtained using Morh’s
integral for displacement computations. Thus,
F = 4.045 x 106 mm
2 ; _
yF = 0
_
9
z
F = 4.468x10 ; __
13
yyF = 2.508x10 mm4
--
13
zzF =1.1388x10 mm4; --
yzF = 0
--
16
yωF = - 4.242x10 mm6, --
zωF = 0
--
19
ωωF = 7.253x10 mm6
The product integrals based on the centroidal axis are obtained from eqn.(4.3.5).
Thus,
113
yyF = 2.508x10 ; 12
zzF = 6.453x10
yzF = 0 ; 16
yωF = - 2.242x10 mm6
zωF = 0 ; 19
ωωF = 7.253X10 mm6
From eqn. (4.3.6) and (4.3.7) the coordinates of the shear centre are obtained as
below.
M B(y - y ) = 0 and M B(z - z ) =1692mm
From eqn. (4.3.8) the expression for the warping function is obtained:
M Bω (s) =ω +1692x (4.3.9)
The warping function for the single cell mono-symmetric section with respect to the
shear centre is shown in Fig. 4.3.8(g).
A1(b) Double cell mono-symmetric section
From Fig.A2 we obtain the following properties of the section.
Cross sectional area F =4,663,722 mm2
Enclosed area A1 = A2 = A3 = 8,372,250 mm2
Fx = 0, symmetry about y-y axis
Fy = (3660 x 203 x 3050) + (3050 x 203 x 1525)+2(3557x203x1525) = 5.4126 x
109mm3
-
202
∫B Bω (s) = ρ (s)ds for open section of the profile
∫B B
ψω (s) = ρ - ds
t for closed section of the profile
For closed sections we have, from eqn.(2.6.19)
∫ ∫ ∫B D
i i-1 i i+1
A C
ds ds ds2A = -ψ +ψ -ψ
t t t
For; Cell 1 we obtain; 6
1 259.59ψ -15.025ψ =16.7445x10 (1)
Cell 2 ⇒ 6
1 2 315.02ψ - 59.59ψ +35.57ψ = -16.7445x10 (2)
Solving eqns (1, & 2) above gives;
5
1 2ψ =ψ = 3.7572x10
For closed Cells 1 & 2, ( )∫B Bω = ρ -1850.84 ds
Integrating from point V to points 1(closed section) we obtain;
Bρ = 0, ds = 3660, t = 203, ∴ 6 2
1ω = -1850.84 * 3660 = - 6.774 *10 mm mm2
From point 1 to point 2 (open section) we have;
Bρ = 0, ds = 915, t = 203, ∴ 6 6
2ω = (0 * 915) - 6.774 *10 = - 6.774 *10 mm2
From point 1 to point 3 (closed section) we have;
Bρ = 3660 * sin59.036°= 3138.42, t = 203, ds = 3557,
( )∴ 6 6
3ω = -6.774 *10 + 3138.42 -1850.84)(3557 = -2.194 *10 mm2
By reason of symmetry we observe that;
6
4 3ω = -ω = 2.194 *10 mm2
and 6
5 1ω = -ω = 6.774 *10 mm2
3050
2745 2745 1830 1830
915 915 3660
203
203
203
1889.4
1160.58
3660
y
x
Fig. A2: Double cell mono symmetric box section
203
6
6 2ω = -ω = 6.774 *10 mm2
The section properties are evaluated using the same procedure used in section
4.3.4.1(a). The relevant values obtained are,
xy ωyF =F = 0
13
xxF = 2.5073 *10
16
ωxF = -4.242*10
Substituting these properties into eqns.(4.3.6 & 4.3.7) the coordinates of the shear
centre are obtained. Thus,
( ) 0M Bx x− = and ( )M By - y =1691.84mm
From eqn.(4.3.8) we obtain the expression for the warping function with respect to
the shear centre as,
M Bω =ω +1691.84x
where Bω is obtained from Fig.4.3.4(d).
The warping function Mω with respect to shear centre is on Fig.4.3.9 (g).
A1(c) Non-symmetric section
Fig.A3 shows the cross section of a single cell non symmetric section.
Cross sectional area F = 3.1514x106mm2
Enclosed area A= 13.954x106mm2
Fx = 4.2477x109mm3
Fy = 7.608x109mm3
∫B B
ψω (s) = ρ - ds
t for closed section of the profile
Fig.A3: Single cell non-symmetric section
1830
200
5490
200
200
3660
x
204
∫
22A 2x13,953,750x200ψ = = = 354230mm
ds 5490+3557+3600+3050
t
Integrating from point v to point 1; Bρ = 0, ds = 5490, t =100 mm
6
1
354230ω = (0 - )(5490) = -9.723x10
200mm2
From point 1 to point 2; Bρ = 5490sin59.036 = 4708mm
( ) 6
2ω = 4708 -1771 x3557 -9.723x10 = 6 20.724x10 mm
From point 2 to point 3;
Bρ = 3050, ds -3050, t =100
( )∴ 6
3ω = 3050 -1771 x3660+0.724x10 = 6 25.405x10 mm
The location of the centroidal axes of the section is given by,
9
x
6
F 7.608 *10x = = = 2414mm
F 3.1514 *10;
9y
6
F 4.2477 *10y = = =1348mm
F 3.1514 *10
The warping function with respect to the centroidal axes is given by,
C B 0ω =ω -ω , where Bω0
Fω =
F
The cross sectional parameters are evaluated following the procedures given in
section 4.3.4. The values obtained are as follows.
F = 3.151 *106 mm
2 , _
9
x
F = 7.608 *10
_
9
yF = 4.248 *10 , _
12
ωF = -4.647 *10
__
13
xx
F = 2.936 *10 mm4, 13
yyF =1.091*10 mm4
10
xyF = 4.359 *10 , 16
ωxF = -3.226 *10 mm6
15
ωyF = 7.202*10 ,
The product integrals based on the centroidal axis are obtained from eqn.(4.3.5).
Thus,
13
xxF = -1.099 *10 ; 12
yyF = 5.183 *10
12
xyF = -1.539 *10 ; 16
ωxF = -2.104 *10 mm6
16
ωyF =1.347 *10 ;
205
From eqn (4.3.6) and (4.3.7) the coordinates of the shear centre are obtained as
follows.
ωy xx ωx xy
M B 2
xx yy xy
F .F -F F(x - x ) = = 2119mm
F .F - (F )
and ωy xy ωx yy
M B 2
xx yy xy
F .F -F F(y - y ) = =1618mm
F .F - (F )
From eqn.(4.3.8) the warping function with respect to the shear center is obtained:
( ) ( )M C M B M Bω =ω + y - y x - x - x y
C=ω +1618x - 2119y
Bω 6
C B B
Fω =ω - =ω +1.475 *10
F
∴ 6
M Bω =ω +1.475 *10 +1618x - 2119y
The warping function with respect to the centroidal axes for this profile is shown in
Fig. 4.3.6(b) while that with respect to the shear centre is on Fig.4.3.12 (g).
206
APPENDIX TWO:
COMPUTATION OF VLASOV’S COEFFICIENTS
A2.1 Coefficients for single cell mono-symmetric section
ϕ ϕ∫a =a = (s) (s)dAij ji i j
' '( ). ( )b b s s dAi jij jisϕ ϕ= = ∫
11 1 1( ). ( ) 3.179s
a s s dAϕ ϕ= =∫ ' '
11 1 1( ). ( ) 0.522s
b s s dAϕ ϕ= =∫
12 21 1 2( ). ( ) 0.000s
a a s s dAϕ ϕ= = =∫ ' '
12 21 1 2( ). ( ) 0.000s
b b s s dAϕ ϕ= = =∫
13 31 1 2( ) ( ) 0.000s
a a s s dAϕ ϕ= = =∫ ' '
13 31 1 3( ). ( ) 0.000s
b b s s dAϕ ϕ= = =∫
22 2 2( ) ( ) 12.341s
a s s dAϕ ϕ= =∫ ' '
22 2 2( ). ( ) 1.469s
b s s dAϕ ϕ= =∫
23 32 2 3( ). ( ) 0.133s
a a s s dAϕ ϕ= = = −∫ ' '
23 32 2 3( ). ( ) 0.525s
b b s s dAϕ ϕ= = = −∫
33 3 3( ). ( ) 0.373s
a s s dAϕ ϕ= =∫ ' '
33 3 3( ). ( ) 0.693s
b s s dAϕ ϕ= =∫
'( ). ( )c c s s dAjkj jk ks
ψ ϕ= = ∫ '( ). ( )i h
c c s s dAih hi s
ϕ ψ= = ∫
'
11 1 1( ). ( ) 0.522c s s dAψ ϕ= =∫ '
11 1 1( ). ( ) 0.522c s s dAψ ϕ= =∫
( ) '
12 21 1 2. ( ) 0.000s
c c s s dAψ ϕ= = =∫ ( ) '
12 21 1 2. ( ) 0.000s
c c s s dAψ ϕ= = =∫
'
13 31 1 3( ). ( ) 0.000s
c c s s dAψ ϕ= = =∫ '
13 31 1 3( ). ( ) 0.000s
c c s s dAψ ϕ= = =∫
'
22 2 2( ). ( ) 1.469s
c s s dAψ ϕ= =∫ '
22 2 2( ). ( ) 1.469s
c s s dAψ ϕ= =∫
'
23 32 2 3( ). ( ) 0.525s
c c s s dAψ ϕ= = = −∫ '
23 32 2 3( ). ( ) 0.525s
c c s s dAψ ϕ= = = −∫
'
33 3 3( ). ( ) 0.693s
c s s dAψ ϕ= =∫ '
33 3 3( ). ( ) 0.693s
c s s dAψ ϕ= =∫
'
41 4 1( ). ( ) 0.000s
c s s dAψ ϕ= =∫ '
41 4 1( ). ( ) 0.000s
c s s dAψ ϕ= =∫
'
42 4 2( ). ( ) 0.769S
c s s dAψ ϕ= = −∫ '
42 4 2( ). ( ) 0.769S
c s s dAψ ϕ= = −∫
207
'
43 4 3( ). ( ) 0.623s
c s s dAψ ϕ= =∫ '
43 4 3( ). ( ) 0.623s
c s s dAψ ϕ= =∫
( ). ( )r r s s dAkh hk k hs
ψ ψ= = ∫
11 1 1( ). ( ) 0.525s
r s s dAψ ψ= =∫
12 21 1 2( ). ( ) 0.000s
r r s s dAψ ψ= = =∫
13 31 1 3( ). ( ) 0.000s
r r s s dAψ ψ= = =∫
14 41 1 4( ). ( ) 0.000s
r r s s dAψ ψ= = =∫
22 2 2( ). ( ) 1.469s
r s s dAψ ψ= =∫
23 32 2 3( ). ( ) 0.525s
r r s s dAψ ψ= = = −∫
24 42 2 4( ). ( ) 0.769s
r r s s dAψ ψ= = = −∫
33 3 3( ). ( ) 0.693s
r s s dAψ ψ= =∫
34 43 3 4( ). ( ) 0.623s
r r s s dAψ ψ= = =∫
44 4 4( ). ( ) 7.200s
r s s dAψ ψ= =∫
A2.2 Coefficients for double cell mono-symmetric section
( ) ( )a a s s dAij ji i j
ϕ ϕ= = ∫ ' '( ). ( )b b s s dAi jij jisϕ ϕ= = ∫
11 1 1( ). ( ) 7.023s
a s s dAϕ ϕ= =∫ ' '
11 1 1( ). ( ) 1.680s
b s s dAϕ ϕ= =∫
12 21 1 2( ). ( ) 0.000s
a a s s dAϕ ϕ= = =∫ ' '
12 21 1 2( ). ( ) 0.000s
b b s s dAϕ ϕ= = =∫
13 31 1 2( ) ( ) 0.000s
a a s s dAϕ ϕ= = =∫ ' '
13 31 1 3( ). ( ) 0s
b b s s dAϕ ϕ= = =∫
22 2 2( ). ( ) 25.073s
a s s dAϕ ϕ= =∫ ' '
22 2 2( ). ( ) 2.982s
b s s dAϕ ϕ= =∫
208
23 32 2 3( ) ( ) 0.425a a s s dAϕ ϕ= = =∫ ' '
23 32 2 3( ). ( ) 0.449s
b b s s dAϕ ϕ= = = −∫
33 3 3( ). ( ) 0.750s
a s s dAϕ ϕ= =∫ ' '
33 3 3( ). ( ) 1.533s
b s s dAϕ ϕ= =∫
'( ). ( )c c s s dAjkj jk ks
ψ ϕ= = ∫ '( ). ( )i h
c c s s dAih hi s
ϕ ψ= = ∫
'
11 1 1( ). ( ) 1.680s
c s s dAψ ϕ= =∫ '
11 1 1( ). ( ) 1.680s
c s s dAψ ϕ= =∫
( ) '
12 21 1 2. ( ) 0.000s
c c s s dAψ ϕ= = =∫ ( ) '
12 21 1 2. ( ) 0.000s
c c s s dAψ ϕ= = =∫
'
13 31 1 3( ). ( ) 0.000s
c c s s dAψ ϕ= = =∫ '
13 31 1 3( ). ( ) 0.000s
c c s s dAψ ϕ= = =∫
'
22 2 2( ). ( ) 2.982s
c s s dAψ ϕ= =∫ '
22 2 2( ). ( ) 2.982s
c s s dAψ ϕ= =∫
'
23 32 2 3( ). ( ) 0.449s
c c s s dAψ ϕ= = = −∫ '
23 32 2 3( ). ( ) 0.449s
c c s s dAψ ϕ= = = −∫
'
33 3 3( ). ( ) 1.533s
c s s dAψ ϕ= =∫ '
33 3 3( ). ( ) 1.533s
c s s dAψ ϕ= =∫
'
41 4 1( ). ( ) 0.000s
c s s dAψ ϕ= =∫ '
41 4 1( ). ( ) 0.000s
c s s dAψ ϕ= =∫
'
42 4 2( ). ( ) 3.404S
c s s dAψ ϕ= =∫ '
42 4 2( ). ( ) 3.404S
c s s dAψ ϕ= =∫
'
43 4 3( ). ( ) 0.886s
c s s dAψ ϕ= =∫ '
43 4 3( ). ( ) 0.886s
c s s dAψ ϕ= =∫
( ). ( )r r s s dAkh hk k hs
ψ ψ= = ∫
11 1 1( ). ( ) 1.680s
r s s dAψ ψ= =∫
12 21 1 2( ). ( ) 0.000s
r r s s dAψ ψ= = =∫
13 31 1 3( ). ( ) 0.000s
r r s s dAψ ψ= = =∫
14 41 1 4( ). ( ) 0.000s
r r s s dAψ ψ= = =∫
22 2 2( ). ( ) 2.982s
r s s dAψ ψ= =∫
209
23 32 2 3( ). ( ) 0.449s
r r s s dAψ ψ= = = −∫
24 42 2 4( ). ( ) 3.404s
r r s s dAψ ψ= = =∫
33 3 3( ). ( ) 1.533s
r s s dAψ ψ= =∫
34 43 3 4( ). ( ) 0.886s
r r s s dAψ ψ= = =∫
44 4 4( ) ( ) 12.011r s s dAψ ψ= =∫
A2.3 Coefficients for single cell doubly symmetric section
( ) ( )a a s s dAij ji i j
ϕ ϕ= = ∫ ' '( ). ( )b b s s dAi jij jisϕ ϕ= = ∫
11 1 1( ). ( ) 10.980s
a s s dAϕ ϕ= =∫ ' '
11 1 1( ). ( ) 1.200s
b s s dAϕ ϕ= =∫
12 21 1 2( ). ( ) 0.000s
a a s s dAϕ ϕ= = =∫ ' '
12 21 1 2( ). ( ) 0.000s
b b s s dAϕ ϕ= = =∫
13 31 1 2( ) ( ) 0s
a a s s dAϕ ϕ= = =∫ ' '
13 31 1 3( ). ( ) 0.000s
b b s s dAϕ ϕ= = =∫
22 2 2( ). ( ) 27.994s
a s s dAϕ ϕ= =∫ ' '
22 2 2( ). ( ) 2.880s
b s s dAϕ ϕ= =∫
23 32 2 3( ) ( ) 0.000s
a a s sϕ ϕ= = =∫ ' '
23 32 2 3( ). ( ) 0.000s
b b s s dAϕ ϕ= = =∫
33 3 3( ). ( ) 7.709s
a s s dAϕ ϕ= =∫ ' '
33 3 3( ). ( ) 3.736s
b s s dAϕ ϕ= =∫
'( ). ( )c c s s dAjkj kj ks
ψ ϕ= = ∫ '( ). ( )i h
c c s s dAih hi s
ϕ ψ= = ∫
'
11 1 1( ). ( ) 1.200s
c s s dAψ ϕ= =∫ '
11 1 1( ). ( ) 1.200s
c s s dAϕ ψ= =∫
( ) '
12 21 1 2. ( ) 0.000s
c c s s dAψ ϕ= = =∫ '
12 21 1 2( ). ( ) 0.000s
c c s s dAϕ ψ= = =∫
'
13 31 1 3( ). ( ) 0.000s
c c s s dAψ ϕ= = =∫ '
13 31 1 3( ). ( ) 0.000s
c c s s dAϕ ψ= = =∫
'
22 2 2( ). ( ) 2.880s
c s s dAψ ϕ= =∫ '
14 41 1 4( ). ( ) 0.000s
c c s s dAϕ ψ= = =∫
210
'
23 32 2 3( ) ( ) 0.000s
c c s s dAψ ϕ= = =∫ '
22 2 2( ). ( ) 2.880s
c s s dAϕ ψ= =∫
'
33 3 3( ). ( ) 3.736s
c s s dAψ ϕ= =∫ '
23 32 2 3( ). ( ) 0.000s
c c s s dAϕ ψ= = =∫
'
41 4 1( ). ( ) 0.000s
c s s dAψ ϕ= =∫ '
24 42 2 4( ). ( ) 0.000s
c c s s dAϕ ψ= = =∫
'
42 4 2( ). ( ) 0.000S
c s s dAψ ϕ= =∫ '
33 3 3( ). ( ) 3.736s
c s s dAϕ ψ= =∫
'
43 4 3( ). ( ) 3.732s
c s s dAψ ϕ= =∫ '
34 43 3 4( ). ( ) 3.732s
c c s s dAϕ ψ= = =∫
( ). ( )r r s s dAkh hk k hs
ψ ψ= = ∫
11 1 1( ). ( ) 1.200s
r s s dAψ ψ= =∫
12 21 1 2( ). ( ) 0.000s
r r s s dAψ ψ= = =∫
13 31 1 3( ). ( ) 0.000s
r r s s dAψ ψ= = =∫
14 41 1 4( ). ( ) 0.000s
r r s s dAψ ψ= = =∫
22 2 2( ). ( ) 2.880s
r s s dAψ ψ= =∫
23 32 2 3( ). ( ) 0.000s
r r s s dAψ ψ= = =∫
24 42 2 4( ). ( ) 0.00s
r r s s dAψ ψ= = =∫
33 3 3( ). ( ) 3.736s
r s s dAψ ψ= =∫
34 43 3 4( ). ( ) 3.732s
r r s s dAψ ψ= = =∫
44 4 4( ). ( ) 22.032s
r s s dAψ ψ= =∫
A2.4 Coefficients for multi- cell doubly symmetric section
( ) ( )a a s s dAji ij i j
ϕ ϕ= = ∫ ' '( ). ( )b b s s dAi jij jisϕ ϕ= = ∫
211
11 1 1( ). ( ) 9.900s
a s s dAϕ ϕ= =∫ ' '
11 1 1( ). ( ) 2.400s
b s s dAϕ ϕ= =∫
12 21 1 2( ). ( ) 0.000s
a a s s dAϕ ϕ= = =∫ ' '
12 21 1 2( ). ( ) 0.000s
b b s s dAϕ ϕ= = =∫
13 31 1 2( ) ( ) 0.000s
a a s s dAϕ ϕ= = =∫ ' '
13 31 1 3( ). ( ) 0.000s
b b s s dAϕ ϕ= = =∫
22 2 2( ). ( ) 27.000s
a s s dAϕ ϕ= =∫ ' '
22 2 2( ). ( ) 3.600s
b s s dAϕ ϕ= =∫
23 32 2 3( ) ( ) 0.000s
a a s sϕ ϕ= = =∫ ' '
23 32 2 3( ). ( ) 0.000s
b b s s dAϕ ϕ= = =∫
33 3 3( ). ( ) 127.752s
a s s dAϕ ϕ= =∫ ' '
33 3 3( ). ( ) 55.727s
b s s dAϕ ϕ= =∫
'( ). ( )c c s s dAjkj kj ks
ψ ϕ= = ∫ '( ). ( )i h
c c s s dAih hi s
ϕ ψ= = ∫
'
11 1 1( ). ( ) 2.400s
c s s dAψ ϕ= =∫ '
11 1 1( ). ( ) 2.400s
c s s dAψ ϕ= =∫
( ) '
12 21 1 2. ( ) 0.000s
c c s s dAψ ϕ= = =∫ ( ) '
12 21 1 2. ( ) 0.000s
c c s s dAψ ϕ= = =∫
'
13 31 1 3( ). ( ) 0.000s
c c s s dAψ ϕ= = =∫ '
13 31 1 3( ). ( ) 0.000s
c c s s dAψ ϕ= = =∫
'
22 2 2( ). ( ) 3.600s
c s s dAψ ϕ= =∫ '
14 4 1( ). ( ) 0.000s
c s s dAψ ϕ= =∫
'
23 32 2 3( ) ( ) 0.000s
c c s s dAψ ϕ= = =∫ '
22 2 2( ). ( ) 3.600s
c s s dAψ ϕ= =∫
'
33 3 3( ). ( ) 55.727s
c s s dAψ ϕ= =∫ '
23 32 2 3( ) ( ) 0.000s
c c s s dAψ ϕ= = =∫
'
41 4 1( ). ( ) 0.000s
c s s dAψ ϕ= =∫ '
42 4 2( ). ( ) 0.000S
c s s dAψ ϕ= =∫
'
42 4 2( ). ( ) 0.000S
c s s dAψ ϕ= =∫ '
33 3 3( ). ( ) 55.727s
c s s dAψ ϕ= =∫
'
43 4 3( ). ( ) 24.759s
c s s dAψ ϕ= =∫ '
34 4 3( ). ( ) 24.759s
c s s dAψ ϕ= =∫
( ). ( )r r s s dAkh hk k hs
ψ ψ= = ∫
11 1 1( ). ( ) 1.200s
r s s dAψ ψ= =∫
212
12 21 1 2( ). ( ) 0.000s
r r s s dAψ ψ= = =∫
13 31 1 3( ). ( ) 0.000s
r r s s dAψ ψ= = =∫
14 41 1 4( ). ( ) 0.000s
r r s s dAψ ψ= = =∫
22 2 2( ). ( ) 3.600s
r s s dAψ ψ= =∫
23 32 2 3( ). ( ) 0.000s
r r s s dAψ ψ= = =∫
∫24 42 2 4
s
r = r = ψ (s).ψ (s)dA = 0.00
∫33 3 3
s
r = ψ (s).ψ (s)dA = 57.727
∫34 43 3 4
s
r = r = ψ (s).ψ (s)dA = 24.759
∫44 4 4r = ψ (s)ψ (s)dA =35.100
A2.5 Coefficients for single cell non- symmetric section
( ) ( )a a s s dAji ij i j
ϕ ϕ= = ∫ ' '( ). ( )b b s s dAi jij jisϕ ϕ= = ∫
11 1 1( ). ( ) 5.264s
a s s dAϕ ϕ= =∫ ' '
11 1 1( ). ( ) 1.132s
b s s dAϕ ϕ= =∫
12 21 1 2( ). ( ) 1.536s
a a s s dAϕ ϕ= = =∫ ' '
12 21 1 2( ). ( ) 0.314s
b b s s dAϕ ϕ= = =∫
13 31 1 2( ) ( ) 0.001s
a a s s dAϕ ϕ= = = −∫ ' '
13 31 1 3( ). ( ) 0.037s
b b s s dAϕ ϕ= = = −∫
22 2 2( ). ( ) 11.018s
a s s dAϕ ϕ= =∫ ' '
22 2 2( ). ( ) 2.018s
b s s dAϕ ϕ= =∫
23 32 2 3( ) ( ) 0.038s
a a s sϕ ϕ= = = −∫ ' '
23 32 2 3( ). ( ) 0.028s
b b s s dAϕ ϕ= = =∫
33 3 3( ). ( ) 0.293s
a s s dAϕ ϕ= =∫ ' '
33 3 3( ). ( ) 0.244s
b s s dAϕ ϕ= =∫
'( ). ( )c c s s dAjkj kj ks
ψ ϕ= = ∫ '( ). ( )i h
c c s s dAih hi s
ϕ ψ= = ∫
'
11 1 1( ). ( ) 1.132s
c s s dAψ ϕ= =∫ '
11 1 1( ). ( ) 1.132s
c s s dAψ ϕ= =∫
213
( ) '
12 21 1 2. ( ) 0.314s
c c s s dAψ ϕ= = =∫ ( ) '
12 21 1 2. ( ) 0.314s
c c s s dAψ ϕ= = =∫
'
13 31 1 3( ). ( ) 0.037s
c c s s dAψ ϕ= = = −∫ '
13 31 1 3( ). ( ) 0.037s
c c s s dAψ ϕ= = = −∫
'
22 2 2( ). ( ) 2.018s
c s s dAψ ϕ= =∫ '
14 4 1( ). ( ) 0.148s
c s s dAψ ϕ= =∫
'
23 32 2 3( ) ( ) 0.028s
c c s s dAψ ϕ= = =∫ '
22 2 2( ). ( ) 2.018s
c s s dAψ ϕ= =∫
'
33 3 3( ). ( ) 0.244s
c s s dAψ ϕ= =∫ '
23 32 2 3( ) ( ) 0.028s
c c s s dAψ ϕ= = =∫
'
41 4 1( ). ( ) 0.148s
c s s dAψ ϕ= =∫ '
24 4 2( ). ( ) 0.476S
c s s dAψ ϕ= =∫
'
42 4 2( ). ( ) 0.476S
c s s dAψ ϕ= =∫ '
33 3 3( ). ( ) 0.244s
c s s dAψ ϕ= =∫
'
43 4 3( ). ( ) 0.262s
c s s dAψ ϕ= =∫ '
34 4 3( ). ( ) 0.262s
c s s dAψ ϕ= =∫
( ). ( )r r s s dAkh hk k hs
ψ ψ= = ∫
11 1 1( ). ( ) 1.132s
r s s dAψ ψ= =∫
12 21 1 2( ). ( ) 0.314s
r r s s dAψ ψ= = =∫
13 31 1 3( ). ( ) 0.037s
r r s s dAψ ψ= = = −∫
14 41 1 4( ). ( ) 0.148s
r r s s dAψ ψ= = =∫
22 2 2( ). ( ) 2.018s
r s s dAψ ψ= =∫
23 32 2 3( ). ( ) 0.028s
r r s s dAψ ψ= = =∫
∫24 42 2 4
s
r = r = ψ (s).ψ (s)dA = 0.476
∫33 3 3
s
r = ψ (s).ψ (s)dA = 0.244 ∫34 43 3 4
s
r = r = ψ (s).ψ (s)dA = 0.262
∫44 4 4r = ψ (s)ψ (s)dA =10.358
214
APPENDIX THREE
MORH’S INTEGRAL FOR DISPLACEMENT COMPUTATION
Let 1f (x) = a+bx be any linear function and 2f (x) be any arbitrary curve or function as
shown in Fig.A1.
( )I = ∫ ∫1 2 2L Lf (x) * f (x)dx = a+bx f (x)dx
∫ ∫2 2L L
= a f (x)dx +b xf (x)dx
2 2= aA +bxA
where x is the centroid of area A2
( )I∴ 2 2 1= A a+bx = A xf (x)dx
2 1= A * ordinate of f (x) at x = x
Notes
(a) The function 1f (x) is a linear algebraic function
Example; 1 0 1 1 1 1 0f (x) = b +b x; f (x) = b x; f (x) = b
(b) The function 2f (x) is an arbitrary algebraic function
Example; 2 3
2 0 1 2f = a +a x +a x
(c) It is required that within the boundary, ≤ ≤0 x L , both functions must be
continuous. If however, there is discontinuity within the boundary, then L is broken up
into segments of piece-wise continuous functions.
common length L
1f (x) = a+bx
2f (x)
f (x) = a + bx1
x
• Area A2
Fig.A3: Functions f1(x) and f2(x)
x
y
215
Table A1: Morh’s Integral Chart - ∫L
1 20
f (x)f (x)dx
2a L
21a L
3
2 21
a +ab+b L3
acL
1acL
2 ( )
1c a+b L
2
1acL
2
1acL
3 ( )
1c 2a+b L
6
1acL
2
1acL
6 ( )
1c a+2b L
6
( )
1a c + d L
2 ( )
1a 2c + d L
6 ( ) ( )
1ac + a+b c + d +bd L
6
c
L
c
L
c
L
c L
d
a
L
1f (x)
2f (x)
a
L
a
L
b
216
APPENDIX FOUR
COMPUTATION OF DISTORTIONAL AND ROTATIONAL FORCES IN THE BOX
GIRDER PLATES
A4.1 Doubly symmetric section
Referring to Fig. 6.2.3a, the torsional load on the bridge girder is 138KN acting at
2.2m from the edge of the slab. and at (b/2 -1.46) m eccentricity. Thus, the torsional
moment MT on the girder is
T
bM =138 * -1.46
2KNm/m run. The torsional load is
decomposed into bending component Fig.6.2.3b, and torsional component
Fig.6.2.3c, which comprises of distortional component, Fig.6.2.3d and pure rotation,
Fig.6.2.3e.
A4.2 Single cell doubly symmetric section
For single cell mono-symmetric section, b = 7.32m, h = 3.00m, MT =
201.48KNm/m
From Fig.6.2.3 (d and e) we obtain the following.
t,dist Tq = M /2h = 33.58KN/m
b,dist t,distq = q = 33.58KN/m
w,dist Tq = M /2b =13.76KN/m
t,rot Tq = M /2h = 33.58KN/m
b,rot t,rot.q = q = 33.58KN/m
w,rot. Tq = M /2b =13.76KN/m
Work done on various plates of the cross section is given by
∫ i iq = qψds (A4.2.1)
where iψ is the transverse strain mode given in the generalized strain mode diagram
for single cell doubly symmetric section, Fig.4.3.10b.
e.g. for distortional strain mode 3:
3ψ = ±0.618 for bottom and top flanges
3ψ = ±1.482 for webs
For rotational strain mode 4:
4ψ = ±1.5 for the flanges and ± 3.6 for the webs
217
Substituting appropriate values of iψ and iq into eqn. (6.2.1) we obtain
( )3q = 2x33.58x0.618+2x13.76x1.482 x7.32 = 620.45KN
( )4q = 33.58x1.5x2+13.76x3.6x2 x7.32 =1462.62KN
A4.3 Multi-cell doubly symmetric section
For multi- cell doubly-symmetric section, b = 9.00m, h = 3.00m, MT =
317.40KNm/m
From Fig.6.2.2 we obtain the following.
t,dist Tq = M /2h = 52.90KN/m
b,dist t,distq = q = 52.90KN/m
w,dist Tq = M /2b =17.63KN/m
t,rot Tq = M /2h = 52.90KN/m
b,rot t,rot.q = q = 52.90KN/m
w,rot. Tq = M /2b =17.63KN/m
Work done on various plates of the cross section is given by eqn.(6.2.1)
Where iψ is the transverse strain mode given in the generalized strain mode diagram
for multi- cell doubly symmetric section, Fig. 4.3.11.
e. g., for distortional strain mode 3:
3ψ =1.929 for outer top and bottom flanges, 2.014 for middle top and bottom
flanges, 5.871 for the two outer webs and 2.014 for the two inner webs.
For rotational strain mode 4:
4ψ =1.5 0 for top and bottom flanges, 4.50 for the two outer webs and 1.50 for the
two inner webs.
Substituting appropriate values of iψ and iq into eqn. (6.2.1) we obtain
( )3q = 52.90x1.929x4+52.90x2.014x2+17.63x5.87x2 x9.00 = 6495.23KN
( )4q = 52.90x1.5x2+17.63x4.5x2 x9.00 = 2856.33KN
218
A4.4 Mono-symmetric sections
The torsional load on the bridge girder is 138KN, acting at 2.2m from the edge of the
slab, and at (b/2 -1.46) m eccentricity. Thus, the torsional moment MT on the girder is
T
bM =138 * -1.46
2KNm/m run. The torsional load is decomposed into bending
component Fig.6.2.5b, and torsional component Fig 6.2.5c, which comprises of
distortional component, Fig.6.2.6c and pure rotation, Fig.6.2.6b.
A4.5 Single cell mono-symmetric section
For single cell mono-symmetric section, b = 7.32m, h = 3.05m and hence, MT =
201.45KNm/m run.
The various components of pure torsional and distortional loads are evaluated using
Table 2.1.
( )
2
tR
Pbq = = 44.05KN/m
h a+b
( )bR
Pabq = = 20.0KN/m
h a+b
( )wR
Pbcq = = 21.40KN/m
h a+b
( )
2
tD
Paq = =11.01KN/m
h a+b
( )bD
Pabq = = 22.02KN/m
h a+b
( )wD
Pac= =10.702KN/m
h a+bq
Work done on the various components (plates) of the cross section is given by
∫i i iq = qψds (A4.5.1)
Values of iψ for various plates of the cross section are obtained from strain made
diagrams shown in Fig.4.3.8.
For bending strain mode 1: 1ψ = 0.857 for the webs and 0.0 for top and bottom
flanges.
∴ 1q = 69x0.857x2 =118.27KN/m
⇒ 1 q = qxb =118.27x9.15 =1082.17KN
For distortional strain mode 3: 3ψ = 0.159 for top flange, 0.417 for bottom flange and
0.123 for the webs.
∴
( ) ( ) ( )3 tD bD wD
3 3
q = q x1.105 + q x1.945 + 2 q x2.57
=11.01x0.159+ 22.02x0.492+10.702x0.417x2 = 21.47KN/m
q = q xb = 21.47x7.320 =157.16KN
219
For rotational strain mode 4: 4ψ =1.105 for top flange, 1.945 for bottom flange and
2.57 for the webs.
( ) ( ) ( )4 tD bD wD
4 4
q = q *1.105 + q *1.945 + 2 q * 2.57
= 44.05x1.105+ 20.02x1.945+ 21.4x2.57x2 =197.61
q = q xb =197.61x7.32 =1446.505KN
2q = 0 , since horizontal loads are not involved
A4.6 Double cell mono-symmetric section
For double cell mono-symmetric section, b = 7.32m, h = 3.05m and hence, MT =
201.45KNm/m run (same as in single cell).
Therefore 2q = 0.0KNm/m run,
3q =157.16KNm,
4q = 1446.505KNm, as in single cell mono-symmetric section
220
APPENDIX FIVE
LAPLACE SOLUTION OF TORSIONAL-DISTORTIONAL EQUATIONS OF
EQUILIBRIUM FOR MONO-SYMMETRIC SECTIONS
The Laplace transforms of the terms in eqn (6.3.8) are;
ℓ{ } ( ) ( ) ( )_
4 3 20 ' 0 '' 0 '''ivv s v s v s v sv v= − − − −
ℓ{ } ( ) ( )_
2'' 0 ' 0v s v sv v= − −
ℓ{ }_
v v=
ℓ{ } /K K s=
Substituting these expressions into eqn (6.3.8) we obtain
_ _ _4 3 2 2 2 2 2 4
0 1 2 3 0 1 4 /s v s C s C sC C s v sC C v K sα α α β− − − − − + + + = (A5.1)
where, 0 (0),C v= 1 '(0)C v= , ( )2 '' 0C v= , 3 '''(0)C v=
⇒ ( )_
4 2 2 44s s vα β− + = 3 2 2 2
0 1 0 2 3 1( ) /s C s C C C s C C K sα α+ + − + + − +
( )3 2 2 2_
0 1 2 0 3 1
4 2 2 4 4 2 2 4
1
4 4
s C s C C C s C CKv
s s s s s
α α
α β α β
+ + − + −= +
− + − + (A5.2)
Determination of partial factors for terms in eqn (A5.2)
We consider the denominator 4 2 2 44s sα β− +
Case I: 244 αβ ≥
If 4 24β α≥ then we have real and imaginary roots that can be expressed in the form
( )( )4 2 2 4 2 24s s s ps q s rs zα β− + = − + − + (A5.3)
where, , , ,p q r z are constants.
Multiplying out we have,
( ) ( )2 2 4 3 2 3 2 2s pz q s rs z s ps qs rs prs qrs zs pzs qz− + − + = − + − + − + − +
⇒ ( ) ( ) ( )4 2 2 2 4 3 24s s s p r s q pr z s pz qr s qzα β− + = − + + + + − + + (A5.4)
Comparing coefficients in eqn (6.2.16) we obtain that,
( ) 0p r− + = or p r= − (a)
221
2pr q z α+ + = − (b)
0pz qr+ = (c)
44qz β= ⇒ 44 /q zβ= (d)
Substituting (d) into (c) gives;
44 / 0pz r zβ+ =
∴ 44 / 0rz r zβ− + = [since p r= − ]
or ( )44 / 0r z zβ − = ( )4 24 0r zβ⇒ − = 4 24 zβ⇒ =
∴ 22z β=
From (d), 4 4 2 24 / 4 / 2 2q zβ β β β= = =
From (b); 2 2 22 2pr β β α+ + = − 2 2 24r β α⇒ − + = −
∴ 2 2 24r β α= +
2 24r β α⇒ = + ; 2 24p β α= − − ; 22q β= ; 22z β=
Thus, from eqn (6.2.15) we obtain,
( ) ( )4 2 2 4 2 2 2 2 2 2 2 24 4 2 4 2s s s s s sα β β α β β α β− + = + + + − + + (A5.5)
Let 2
2 2 214
2 4
αλ β α β= + = +
∴Eqn (A5.5) becomes,
( ) ( )4 2 2 4 2 2 2 24 2 2 2 2s s s s s sα β λ β λ β− + = + + − + (A5.6)
The partial fractions of the terms in eqn (A5.2) can be obtained using eqn (A5.6).
Thus,4 2 2 4 2 2 2 2
1
( 4 ) 2 2 2 2
E As B Cs D
s s s s s s s sα β λ β λ β
+ += + +
− + − + + + (A5.7)
⇒ ( ) ( )( ) ( ) ( )4 2 2 4 2 2 2 21 4 2 2 2E s s s As B s s s s Cs D s sα β λ β λ β= − + + + + + + + − +
4 2 2 4 4 3 3 2 2 2 21 4 2 2 2 2Es E s E As Bs A s s A s B sα β λ βλ β β= − + + + + + + +
4 3 3 2 2 2 22 2 2 2Cs Ds C s D s C s D sλ λ β β+ + − − + +
Comparing coefficients we obtain
0E A C+ + = (a)
( )2 0B D A Cλ+ + − = A C⇒ = (b)
( ) ( )2 22 2 0B D A C Eλ β α− + + − = (c)
222
( )22 0B Dβ + = B D⇒ = − (d)
41 4Eβ= 41/ 4E β⇒ = (e)
Solving eqns a,b,c,d and e above yields;
4
1
8A
β= − ;
2 2
4
2
8B
λ β
λβ
−= ;
4
1
8C
β= − ;
4 2
4
2
8D
β λ
λβ
−= ;
2
1
4E
β=
Substituting into eqn (A5.7) we obtain;
( ) ( )2 2 2 4 4 2 2 4
4 2 2 4 4 2 2 2 2
/ 8 2 /8 /8 2 /81 1 1.
( 4 ) 4 2 2 2 2
s s
s s s s s s s s
β λ β λβ β λ β λβ
α β β λ β λ β
− + − − − − = + +
− + − + + +
2 2 2 2
4 2 2 4 4 4 2 2 4 2 2
1 1 1 1 2 1 (2 ).
( 4 ) 4 8 2 2 8 2 2
s s
s s s s s s s s
λ λ β λ λ β
α β β λβ λ β λβ λ β
− + − − + −= + +
− + − + + +
4 4 2 2 2 2 2
1 1 1 2 1 1.
4 8 2 2 8 2 2
s
s s s s s
λ
β β λ β λβ λ β
−= − − − + − +
_
4 2 2 2 2 2
1 2 1 1
8 2 2 8 2 2
s
s s s s
λ
β λ β λβ λ β
+− + + + + +
(A5.8)
Considering the second term on the right hand side of eqn (A5.2) we obtain,
( )3 2 2 2
0 1 2 0 3 1
4 2 2 4 2 2 2 24 2 2 2 2
C s C s C C s C C Gs H Ks M
s s s s s s
α α
α β λ β λ β
+ + − + − + += +
− + − + + + (A5.9)
( ) ( ) ( )3 2 2 2 2 2
0 1 2 0 3 1 2 2C s C s C C s C C Gs H s sα α λ β⇒ + + − + − = + + + +
( ) ( )2 22 2Ks M s sλ β+ + − +
3 2 2 2 22 2 2 2Gs Hs G s H s G s Hλ λ β β= + + + + + +
3 2 2 2 22 2 2 2Ks Ms K s M s K s Mλ λ β β+ + − − + + (A5.10)
Comparing coefficients in eqn (A5.10) we have,
0 0G K C G C K+ = ⇒ = −
1( ) 2 ( )H M G K Cλ+ + − = (b)
2 2
2 02 ( ) 2 ( )G K H M C Cβ λ α+ + − = − (c)
2 2
3 1( ) ( ) / 2H M C Cα β+ = − (d)
223
Substituting (d) into (b) gives,
2 2
3 1 0 1( ) / 2 2 ( 2 )C C C K Cα β λ− + − = 2 2
3 1 0 1( ) / 2 2 4C C C C Kα β λ λ⇒ − + − =
2 2 2 2
3 1 0 3 1 0 11
2 2
4 2
8 2 4 8
C C C C C C CCK
α α λβ β
λβ λ λβ
− − + −∴ = + − =
⇒ ( )2 2 2
0 1 3
2
4 2
8
C C CK
λβ β α
λβ
− + += (A5.11)
From (a), ( )2 2 2 2
0 0 1 3
0 2
8 4 2
8
C C C CG C K
λβ λβ β α
λβ
− + + += − =
( )2 2 2
0 1 3
2
4 2
8
C C CG
λβ β α
λβ
+ + +⇒ = (A5.12)
( )2 2
1
2
2 2( )
8
CG K
β α
λβ
+− = ; ( )
2
0 3
2
4
4
C CG K
λβ
λβ
++ = (A5.13)
From (d), 2
3 1
22
C CH M
α
β
−= −
Substituting into expression (c) we obtain
2 22 20 3 3 1
2 02 2
42 2 2
4 2
C C C CM C C
λβ αβ λ α
λβ β
+ −+ − = −
2 2 220 3 3 1
2 02 2
44
2
C C C CM C C
λβ β λ α λλ α
λβ β
+ −⇒ + − = −
2 220 3 3 1
2 02
44
2
C C C CM C C
λ λ α λλ α
λ β
+ −⇒ + − = −
2 2 2
0 3 3 1 0 2
2 2
4
8 4 4
C C C C C CM
β λ λ α λ α
λ λβ λ
+ − −= + +
4 2 2 2 2 2 2 2
0 3 3 1 0 2
2 2
4 2 2 2 2
8
C C C C C Cλβ β λ α λ α λβ λβ
λ β
+ + − + −=
( ) ( )2 2 2 2 2 2 2 2
0 1 2 3
2 2
4 2 2 2 2
8
C C C CM
β λ α β λ α λ λβ β λ
λ β
+ − − + += (A5.14)
2
3 1
22
C CH M
α
β
−= −
( ) ( )2 2 2 2 2 2 2 2 2 2 2
3 1 0 1 2 3
2 2
4 4 4 2 2 2 2
8
C C C C C Cλ α λ β λ α β λ α λ β λ β λ
λ β
− − + + + − +=
224
( ) ( )2 2 2 2 2 2 2 2
3 2 1 0
2 2
2 2 2 4 2
8
C C C CH
λ β β λ λ α β λ β α λ
λ β
− + − − +⇒ = (A5.15)
Substituting eqns (A5.8) and (A5.9) into eqn (A5.2) while noting eqns (A5.11) to
(A5.15) we obtain;
3 3
4 4 2 2 2 2 2
1 2 1
4 8 2 2 8 2 2
K K s Kv
s s s s s
λ
β β λ β λβ λ β
−= − − − + − +
3 3
4 2 2 2 2 2
2 1
8 2 2 8 2 2
K Ks
s s s s
λ
β λ β λβ λ β
+− + + + + +
2 22 2
Gs
s sλ β+
− +
2 22 2
H
s sλ β+
− + 2 2 2 22 2 2 2
Ks M
s s s sλ β λ β+ +
+ + + + (A5.16)
Now, ( )22 2 2 22 2 2s s sλ β λ λ β− + = − − +
But 2 214
2λ β α= +
22 2
4
αλ β
⇒ = +
( )2
22 2 2 22 2 24
s s sα
λ β λ β β
∴ − + = − − + +
( )2
22 2 22 24
s s sα
λ β λ β
⇒ − + = − + −
( )2 2s λ ω= − − (A5.17)
where, 2
2 2
4
αω β
= −
(A5.18)
Similarly, ( )22 2 22 2s s sλ β λ ω+ + = + − (A5.19)
Noting that ( )2s sλ λ λ+ = + + and ( )2s sλ λ λ− = − − , eqn (A5.16) becomes,
( ) ( ) ( )3 3 3
2 2 24 4 22 2 2
1 1
4 8 8
K K Ksv
s s s s
λ λ
β β λβλ ω λ ω λ ω
−= − − −
− + − + − +
( ) ( ) ( )3 3
2 2 24 22 2 2
1
8 8
K Ks
s s s
λ λ
β λβλ ω λ ω λ ω
+− + +
+ + + + + +
( ) ( )
2 22 2
Gs H
s sλ ω λ ω+ +
− + − + ( ) ( )2 2 22
Ks M
ss λ ωλ ω+ +
+ ++ + (A5.20)
( ) ( ) ( )3 3 3 3
2 2 24 4 4 22 2 2
1 1
4 8 8 8
K K K Ksv
s s s s
λ λ
β β β λβλ ω λ ω λ ω
−⇒ = − + −
− + − + − +
225
( ) ( ) ( )
3 3 3
2 2 24 4 22 2 2
1
8 8 8
K K Ks
s s s
λ λ
β β λβλ ω λ ω λ ω
+− − +
+ + + + + +
( )
( ) ( )2 22 2
G s G
s s
λ λ
λ ω λ ω
−+ +
− + − + ( )
( )( )2 2 22
K sH
ss
λ
λ ωλ ω
++ +
+ +− +
( ) ( )2 2 22
K M
ss
λ
λ ωλ ω− +
+ ++ + (A5.21)
Taking the inverse transform of eqn (A5.21) we obtain;
3 3 3 3 3
4 4 4 2 4cos sin sin cos
4 8 8 8 8
x x x xK K K K Kv e x e x e x e xλ λ λ λλ
ω ω ω ωβ β β ω λβ ω β
− = − + − −
3 3
4 2sin sin cos sin
8 8
x x x xK K Ge x e x Ge x e xλ λ λ λλ λ
ω ω ω ωβ ω λβ ω ω
− −− + + +
sin cos sin sinx x x xH K Me x Ke x e x e x
λ λ λ λλω ω ω ω
ω ω ω− − −+ − + (A5.22)
( ) ( )3 3
4 4sin cos
4 8
x x x xK Kv x e e x e e
λ λ λ λλω ω
β β ω− −
⇒ = + − − +
( )3
4
1sin
8
x xKx e e
λ λλω
β ω ω−
+ −
( )cos x xx Ge Keλ λω −+ +
( ) ( ) 1sin sinx x x x
x Ge Ke x He Meλ λ λ λλ
ω ωω ω
− −+ − + + (A5.23)
3 3
4 4sinh .sinh cos .cosh
4 8
K Kv x x x x
λω λ ω λ
β β ω
⇒ = + −
-
3
2
1sinh .sinh
4
Kx xω λ
λβ ω
−
( ) ( )cos cosh sinh cosh sinhx G x x K x xω λ λ λ γ+ + + −
( ) ( )sin cosh sinh cosh sinhG H M K
x x x x xλ λ
ω λ λ λ λω ω
+ − + + + −
(A5.24)
3 3
4 4
1sin .sinh cos .cosh
4 4
K KV x x x x
λω λ ω λ
β ω λω β
⇒ = − − +
cos .cosh cos .sinhG x x G x xω λ ω λ+ + cos .cosh cos .sinhK x x K x xω λ ω λ+ −
sin .cosh sin .sinhG H G H
x x x xλ λ
ω λ ω λω ω
+ ++ +
226
sin .cosh sin .sinhM K M K
x x x xλ λ
ω λ ω λω ω
− −+ − (A5.25)
2
3 3
4 4
1sin .sinh cos .cosh
4 4
K KV x x x x
λω λ ω λ
β β ωλ
−⇒ = + −
( ) ( )cos .cosh cos .sinhG K x x G K x xω λ ω λ+ + + −
sin .coshG H M K
x xλ λ
ω λω
+ + − +
sin .sinhG H M K
x xλ λ
ω λω
+ − + +
(A5.26)
Addition of eqns.(A5.14) and (A5.15) gives
( )2
3 1
22
C CH M
α
β
−+ = (A5.27)
Subtracting eqn.(A5.13) from eqn.(A5.14) gives
( )2
2 3 0
2
2 2 2
4
C C CH M
λ λ α
λ
− − +− = (A5.28)
Substituting eqns.(A5.12),( A5.27) and (A5.28) into eqn.(A5.23) we obtain
2
3
4
11 sin .sinh cos .cosh
4
KV x x x x
λω λ ω λ
β ωλ
−= + −
2
0 3
2
4cos .cosh
4
C Cx x
λβω λ
λβ ω
++
( )2 2
1
2
2cos .sinh
4
Cx x
β αω λ
λβ ω
+ +
( )2 2
3 1
2
2 2sin .cosh
4
C Cx x
β αω λ
ωβ
+ − +
( ) ( )( )2 2 2 3 2 2 2
2 3 0
2 2
2 4 2 2sin .sinh
4
C C Cx x
λβ λ β λ β λβ αω λ
ωλ β
+ − + − + +
(A5.29)
⇒2
3 0 3
4 2
4cos .cosh
4 4
K C CV x x
λβω λ
β λβ ω
+= +
3
4cos .cosh
4
Kx xω λ
β−
( )2
3
4
1sin .sinh
4
Kx x
λω λ
ωλβ
− +
( )2 2
1
2
2cos .sinh
4
Cx x
β αω λ
ωλβ
+ +
( )2 2
3 1
2
2 2sin .cosh
4
C Cx x
β αω λ
ωβ
+ − +
227
( ) ( )( )2 2 2 3 2 2 2
2 3 0
2 2
2 4 2 2sin .sinh
4
C C Cx x
λβ λ β λ β λβ αω λ
ωλ β
+ − + − + +
(A5.30)
⇒2
3 0 3 33 4 2 4
4cos .cosh
4 4 4
K C C KV x x
λβω λ
β λβ ω β
+= + −
( )2 2
1
2
2cos .sinh
4
Cx x
β αω λ
ωλβ
+ +
( )2 2
3 1
2
2 2sin .
4
C Cx cos x
β αω λ
ωβ
+ − +
( ) ( ) ( )2 2 2 3 2 2 2 2
2 3 0 3
2 2 4
2 4 2 2 1sin .sinh
4 4
C C C Kx x
λβ λ β λ β λβ α λω λ
ωλ β ωλβ
+ − + − + − + +
(A5.31)
3 0 1 2cos .cosh cos .sinhV A A x x A x xω λ ω λ= + + + 3 4sin .cosh sin .sinhA x x A x xω λ ω λ+
(A5.32)
where
30 44
KA
β=
2
0 3 31 2 4
4
4 4
C C KA
λβ
λβ ω β
+= −
( )2 2
1
2 2
2
4
CA
β α
ωλβ
+ =
( )2 2
3 1
3 2
2 2
4
C CA
β α
ωβ
+ − =
(A5.33)
( ) ( ) ( )2 2 2 3 2 2 2 2
2 3 0 3
4 2 2 4
2 4 2 2 1
4 4
C C C KA
λβ λ β λ β λβ α λ
ωλ β ωλβ
+ − + − + − = +
Case II; 4 24 24 24 24444ββββ ≤ α≤ α≤ α≤ α
If 244 αβ ≤ , roots are real and the denominator of eqn (A5.2) can factorize as follows
442
22
24224 4422
4 βααα
βα +−
−
−=+− ssss
−−
−= 4
42
22 4
42β
ααs
2 4 2 42 4 2 44 * 4
2 4 2 4s s
α α α αβ β
= − + − − − −
⇒ 4 2 2 44s sα β− +2 4 2 4
2 4 2 44 * 42 4 2 4
s sα α α α
β β = − − − − + −
(A5.34)
228
Let 442
2 442
βαα
−−=u and 442
2 442
βαα
ω −+=
Eqn (A5.34) becomes
( )( )22224224 4 ωβα −−=+− susss
The partial factors of the first term in eqn (6.2.4) are as follows;
( ) 22224224 4
1
ωβα −
++
−
++=
+− s
EDs
us
CBs
s
A
sss
( )( ) ( ) ( ) ( ) ( )222222221 ussEDsssCBssusA −++−++−−=⇒ ωω (A5.35)