This action might not be possible to undo. Are you sure you want to continue?
STABILITY ANALYSIS OF FRAME TUBE TALL BUILDINGS
by
Amit Urs
A thesis submitted in partial fulfillment of the requirements for the degree of
Masters of Science
Department of Civil and Environmental Engineering
Worcester Polytechnic Institute.
Worcester, MA. 01609.
Oct, 2002
Approved by ___________________________________________________
Thesis Advisor: Prof. P. Jayachandran
__________________________________________________
Prof. Leonard D. Albano
__________________________________________________
Prof. F. Hart
Head of the Department
Civil & Environmental engineering
DATE___________ Worcester Polytechnic Institute, MA.
i
Worcester Polytechnic Institute
Abstract
STABILITY ANALYSIS OF FRAME
TUBE TALL BUILDINGS
by Amit Urs
Thesis Advisor: Professor P. Jayachandran
Department of Civil Engineering
The frame tube buildings have been the most efficient structural system used for
building which is in the range of 40100storey. The soaring heights and the
demanding structural efficiency have led to them having smaller reserves of
stiffness and consequently stability.
In this thesis a Nonlinear analysis and stability check of frametube building is
done. Nonlinear analysis offers several options for addressing problems of
nonlinearity and in this work focus is on Geometric Nonlinearity. The main
sources can be identified as P∆ effect of gravity loading acting on a transversely
displaced structure due to lateral loading and can also be due to member
imperfections, such as member camber and out of plumb erection of the frame.
During analysis the element response keep continuously changing as a function
of the applied load so simple step computing methods have been employed
instead of direct analytical methods. The problem here is dealt in a piece wise
linear way and solved. In this thesis a program using the matrix approach has
been developed. The program developed can calculate the buckling load and can
do Linear and Nonlinear analysis using the Matlab as the computing platform.
ii
Numerical results obtained from the program have been compared with the
Finite Element software Mastan2. The comparative solutions presented later on
in the report clearly prove the accuracy of the program and go on to show, how
exploiting simple matrix equation can help solve the most complex structures in
fraction of seconds.
The program is modular in structure. It provides opportunity for user to make
minor manipulation or can append his own module to make it work for his
specific needs and will get reliable results.
ACKNO WLE DGME NTS
The author wishes to express sincere appreciation to Professors P. Jayachandran
for his thesis advising. The author also would like to thank Prof. William. W.
Durgin for the financial assistance and encouragement.
In addition, author wishes to thank his mom Mrs. Kaivalya V.K. Urs, Aunt
Meera, Aunt Sudha, his sister Ms. Geeta Urs and his Dad Mr. Venkatakrishna
Urs, for all the blessings and support they have provided him. Author wishes to
thank his friend Mr. Badri Krishna for all the help and valuable assistance.
Author would like to thank the members of Weidlinger Associates for their trust,
support and patience.
Mainly, Author wishes to express his humble gratitude to “PARAMATMA”
without whose mercy he could not have done anything.
iv
TABLE O F CO N TE N TS
Abstract .............................................................................................................i
Acknowledgement ............................................................................................iii
Lsit of figures....................................................................................................vi
Preface ............................................................................................................vii
Chapter I: Introduction......................................................................................1
Introduction................................................................................................1
Objective.....................................................................................................2
Scope ..........................................................................................................3
Previous Work.............................................................................................4
Chapter II: Conceptual Framework ....................................................................8
Matrix methods ...........................................................................................8
Stiffness method..........................................................................................8
Level of analysis ........................................................................................21
Different approach to nonlinear analysis ....................................................21
Load control method ................................................................................21
Chapter III: Tall buildings ................................................................................27
Introduction..............................................................................................27
Approaches to analysis...............................................................................31
Reduction modeling...................................................................................32
Frame tube building...................................................................................37
Mode of behavior ......................................................................................38
Structural Analysis .....................................................................................41
Reduction modeling of Frame tube building...............................................44
Chapter IV: Findings and Discussion ...............................................................51
Description of Findings .............................................................................51
Recommendations for future research........................................................55
v
Glossary ..........................................................................................................57
References.......................................................................................................58
Appendix A: ....................................................................................................61
Manual .....................................................................................................61
Algorithm..................................................................................................62
Tutorial 1...................................................................................................63
Tutorial2....................................................................................................68
Appendix B: ....................................................................................................72
Frame tube building reduction models .......................................................72
Frame tube building 40 Stories...................................................................88
vi
LI ST O F FI GURE S
N umber Page
1. Nodal forces and displacement.............................................................10
2. Nodal forces........................................................................................11
3. Load Deflection curve .........................................................................16
4. Finite strain of planar element ..............................................................19
5. Prototype rigid frame...........................................................................44
6. Equivalent lumped girder.....................................................................45
7. Axial stress distribution in column........................................................51
8. deformation of flange frame.................................................................52
9. Frame Tube structural plan..................................................................58
10. connection details................................................................................59
11. Eulers column ....................................................................................33
12. Side sway prevented.............................................................................33
13. Side sway permitted .............................................................................34
vii
PRE FACE
The high speed precise computing and increased memory of the computers has
made it possible to solve complex models. Finite Element Method and Matrix
method are the two methods which show great compatibility for computing
process and has become the most power tool in many engineering branches. In
this work an effort has been made to develop programs that will do an analysis of
highly indeterminate structures. In this work a frame tube building of 40 stories
high, which is highly indeterminate, has been selected.
Tall buildings like frame tube buildings are usually 40  100 stories high. There
analysis is extremely complex due to highly degree of indeterminacy and hand
calculation is very tough. In chapter 1, an introduction to work is given followed
by clearly laid out objectives of the study and its scope.
In Chapter 2, a conceptual frame work of the thesis has been laid out. As the
work is pivoted on the matrix method of structural analysis, different techniques
and approaches to analysis of planar frames have been presented. Linear elastic
analysis, nonlinear analysis and stability analysis have been discussed appropriate
enough, which form the basis for the programs developed.
In Chapter 3, an introduction to tall buildings followed by a detailed discussion
on structural action of frame tube buildings has been presented. Methods used
for mathematical modeling along with logics involved in the structural behavior
of frame tube building have been discussed.
In chapter 4, results and conclusion of the work have been presented.
viii
In Apendix A and Apendix B, Manual, results and comparative values have been
tabulated. Tutorial problem have been presented and theses problem should aid
for a proper understanding of the programs.
C h a p t e r 1
INTRODUCTION
Introduction:
In the past, designers had no need to accommodate for terror attacks from the
air, but now that this threat is a reality; designers have to look for sensible ways to
defend our highrises
All the evidence so far points that a combination of the impact and the fires
triggered the collapse of the World Trade Center buildings; we must now ask,
what we can do in the future to prevent another failure like this one. Some
believe that nothing can, or should, be done. While some opine that "nothing is
designed or will be designed to withstand that [kind of] fire.” Numbers of
theories have evolved to explain exact reason for the collapse of the Twin towers;
some believe the impact alone was the actual cause of the failure by providing
evidence that the structure failed quicker than the time required for the
fireproofed steel to fail while some argue that the fires where to blame, the steel
tube construction of the towers could not resist the intensity of the fires resulting
from the energy released by the 2400 gallons jet fuel that brought the fires to
higher than normal temperatures, as high as 2000 degrees Fahrenheit. But
combination of these two appears to be logical explanation to the tragedy on
9/ 11. Bottom line is one or combination of them may have caused the structure
to fail, from now on it would be necessary to do a thorough stability check of tall
buildings for all possible circumstances.
 2 
In “Limit state design”, buildings are designed for limit state of strength and limit
state of serviceability, leaving the structure with minimum reserve energy. If it is
the design case of a lowrise structure subjected to low horizontal loads,
deflections are small with insignificant change in geometry of the structure. Thus
the reserve energy of the building is sufficient to bring back the structure to
equilibrium state after the load is removed, rendering stability to the structure as a
whole. There by first order analysis of structures to satisfy the equilibrium
conditions is sufficient to verify the design. In case of tall buildings about 40 to
50 story’s high horizontal loads cause huge deflection of the building and a
significant change in the geometry of the structure. The large deformation and
low reserve energy can prove to be cataclysmic if this small energy in the
structure fails to resist loads. Thus failing to satisfy the equilibrium conditions the
structure could become highly unstable. To predict the exact behavior of the
structure a second order analysis of the structure has to be conducted. In this
thesis I have devolved a program to do first order, second order analysis and also
to calculate the elastic critical buckling loads of planar framed structures.
Objectives
The main objective of the thesis work is to develop software to check the stability
of structures and also to do a second order analysis of the 40 story FrameTube
Buildings. Matrix approach is used in the analysis. It is realized that the fulfillment
of the following subobjectives would in turn fulfill the main objective.
1. To understand the concepts of Matrix analysis of structure.
2. To study the behavior of Frame tube buildings.
3. To study various technique in nonlinear analysis of frametube buildings.
 3 
4. To develop a Mat Lab program using the principles of Matrix/ Finite
Element approach for both Linear and Nonlinear analysis. To develop
modular structured program versatile and should be able to analyze a
wide variety of problem.
5. Run the program for Non Linear analysis of Frame Tube building and
verify its structural stability as a whole and to compare these results with
commercial software called Mastan2 for accuracy.
Scope:
Focus of the work is on planar frames with the following characteristics.
1. The behavior of the structure is elastic.
2. Structure may behave like a truss or frame depending on the element
node connection. In frame element for example all member are
considered beam column elements.
3. The structures are considered to be an assembly of planar elements.
Different frames may interact through common columns and through
the floor diaphragms, which are considered to be rigid in their plane. If
the structure contains shear wall or cores it has to be idealized as a single
isolated column.
4. Static gravity loads and horizontal wind loads are applied as nodal forces.
5. During second order analysis geometric nonlinearity due to large
member deformation is the main focus and material nonlinearity due to
change in material property is not considered.
6. Connections are considered either fixed or hinged and it does not
account for partial fixity.
 4 
The program is based on Matrix analysis of structure. Each structural member is
idealized, i.e. is the depth, the width and elemental length are conventionally
reduced to line elements. The material property are represented by area (A),
young’s modulus of elasticity (E), torsion constant (J) and poisons ratio (ν). For
second order analysis simple step method has been employed
Previous work:
Refs (10 & 12) have introduced the topic of highrise structure building by
introducing the different structural forms, with the main intention of explaining
the different approximate methods for analysis.
A number of papers have been published for the analysis and design of tall
buildings. In case of tall buildings stability being the main criterion, selection of
structural form is important. Frame tube buildings have been considered to be
efficient structural form for building ranging from 40 stories to 100 stories height.
Refs (19) have attempted to analyze the frame tube building as accurately as
possible with the limited computing capability that computers of those times had
possessed. Importantly, each paper helps to understand the logic behind the
analysis of the frame tube building. Through time, computers have evolved so
has the complexity of the problem, various advanced techniques like the finite
element method have evolved in tantamount. In Ref (1), a simplified approach to
the analysis of frame tube buildings is made. The dominant modes of behaviors
of the frame tube buildings resting lateral loads are recognized as
1. Rigidly – jointed frame action of the shearresisting panels parallel to
the direction of the loads and
 5 
2. The axial deformation of the frame panels normal to the direction of
the load.
After which, they have proved that it is possible to simulate the behavior of the
structure using a reduced equivalent plane frame. The transfer of vertical shear
has been achieved by giving large values to appropriate shear transfer elements in
the stiffness matrix. Once the modification to stiffness matrix is made then the
process of analysis is as usual. Thus consequently reducing the amount of
computation required in a conventional threedimensional analysis to a simple 2
Dimensional planar frame analysis.
In Ref (2) author has taken consideration of the shear lag effects and has tried to
predict the possible response of the building so as to include this phenomenon.
The author accounts for the shear lag effect by assuming a parabolic stress
distribution. He further goes on to simplify the problem by assuming that the
spacing of columns and beams are uniform and they have all, uniform cross
section area thought out the height. The framework panel of the column and
beam are replaced by equivalent orthotropic plate, to form a closed tube
structure. The orthotropic plate is chosen that the both the modulus of elasticity
of the plate in both the horizontal and vertical direction be represent the axial
stiffness of the beam and the column respectively. Shear modulus representing
the shear stiffness of the framework. The structure then is analyzed as a plate.
The method of plate analysis is fairly accurate and provides design curves for
standard load cases for easy and quick reference; the assumption of a parabolic
stress distribution to simulate the shear lag effect is surely one of the limitations.
In next part ref (3), which is the 2nd of the 3 paper series published by the
authors, problem of torsion is tackled by replacing the discrete structure by
equivalent orthotropic tube and then a nalysis it as a plate. Here, it is assumed that
the shear rigidity of the structure to be low so as to simulate the shear racking
 6 
effect, which the ordinary elastic modulus would not. In Ref (4) third and the
final part authors tackles the deflection in frame tube buildings.
Papers by Fazlur Khan and Amin papers regarding approximate analysis of
framed tube buildings provide graphs for a quick and rapid analysis and design of
frame tube building. The assumption made for the hand method is its limitations.
In ref (8) author has devolved a computer program and has provided a number
of table to assist the analysis of frame tube buildings. He assumes the building to
be square in plan and all columns other than the corner columns to be of same
cross section. The corner column is assumed to have a twice the cross section
area. The columns are connected by equal cross section spandrel beam. For the
analysis he converts the structure into a planar model and for the corner column
interaction he introduces a unit force at each floor level along lines column lines(
for each frame he takes half of the corner column cross section area), one pair at
a time and he establishes the actual interaction forces based on the compatibity
condition. In his report he has presented tables for 40, 50, 60 story building. His
program does not do the stability check.
In Ref (9) authors has submitted his work as a partial fulfillment of his PhD .In
his dissertation author has presented a Macro element method to analyze the
frame tube building. In the process he has also developed two programs, one for
static and dynamic analysis and the other software for the optimization of frame
tube building. Ref (18) has used his software for his MS dissertation on the
analysis of frame tube buildings.
Ref (11) together with introducing the topic of tall buildings has also dealt with
important structural forms in detail the technique of structural analysis and
design. He has also shown the various reduction techniques that can be used for
transforming the structure so as to quickly verify the solutions. Reference (12)
 7 
covers the topic on the Matrix analysis, importance is laid for the solution of
planar framed structure. He also has developed software using the matrix analysis
principle for solving structures, which shall be used as reference software and will
be compared with the software being developed in this thesis work.
 8 
C h a p t e r 2
CONCEPTUAL FRAMEWORK
Matrix Methods:
Behaviors of all types of structures are described by differential equations. In
practice, it is common to represent frames as planar structures for which the
approximate or exact solution of the individual members are available. By using
these solutions, relationship between force and displacement in the end of the
members in combination with equilibrium and compatibility equations at the
joints and supports can be written which yields a system of algebraic equations
that describe the behavior of the structure. Thus these simultaneous equations
can be solved using matrix methods. Once the equations are written in Matrix
form taking advantages of the power computing capabilities of the computers the
large system of equations can be solved quickly. In matrix method there are
generally two approaches the flexibility and the stiffness approach and the former
approach is used in this thesis work
Stiffness Approach:
In matrix method there are generally two approaches the flexibility and the
stiffness approach and the later approach is used in this thesis work. Every
element in the structure is idealized as a line element with section properties (A)
area, (I) moment of Inertia and (J) torsion constant and material properties as (E)
modulus of elasticity, (G) shear rigidity, (NU) poisons ratio. The whole planar
 9 
structure is considered as mesh, formed by individual finite elements connected
together at joints called nodal points. Each element consists of 2 nodes and each
node is associated with corresponding degree of freedom or nodal degree of
freedom. The term Degree of freedom (DOF) is defined as “ the number of
independent coordinate necessary to define the configuration of the system” In
space Nodal DOF are 12 in number but in the case of framed structure with rigid
joints 3 DOF are considered. The frame of reference to identify entire structure is
called global coordinate system referred to as (XY axis). The orthogonal axis with
Xaxis along the member longitudinal axis refers to local coordinate axis. Forces
are applied to the nodes as point load and the corresponding displacement is
measured at nodal points. Thus relationship between nodal forces and
corresponding nodal displacement is written in terms of stiffness equation or
flexibility equations.
The procedure involved in the formation of member stiffness matrix is quite
simple can be summarized in 3 steps
1. Equilibrium.
2. Compatibility
3. Hooke’s law.
For illustration consider a 1D spring element which has 2 DOF, the element
force–displacement relationship relates 2 nodal forces (Fi,Fj) to 2 nodal
displacements through 2X2 stiffness matrix.
 10 
J
Fj ( υ j ) Fi ( υ i )
i
K
f i g 1
{f
e
}= [k]{u
e
}
Expanding, the above equation, we get
F
I
= k
11
u
I
+ k
12
u
J
F
J
= k
21
u
I
+ k
22
u
J
In matrix form, we get
1
]
1
¸
J
I
F
F
=
1
]
1
¸
22 21
12 11
K K
K K
1
]
1
¸
J
I
U
U
Thus k
11,
k
12
, k
11,
k
12
are to be defined
Apply springs law, force (F) Vs displacement (∆) is a linear curve, it can be
written as,
{f } = [k]{∆}
 11 
J
K
F j
F i
i
Fig.2
Nodal forces.
We need to find force (F ) in terms of end displacements u
j
and u
i.
Using the
compatibility conditions,
∆·u
j
u
i
f
I
= k
(u
J

u
I
)
Now applying the equilibrium equations
Σf
X
= 0;
f
I
+ f
J
= 0
Thus, f
I
= f
J
= (k
u
I
+ k
u
J
)
In matrix form can be written as,
1
]
1
¸
J
I
f
f
=
1
]
1
¸
−
−
K K
K K
1
]
1
¸
J
I
u
u
 12 
Thus this is the member stiffness matrix, the fundamental stiffness equations is
of the form
{F}=[k]{∆}
Where, [K] is the element stiffness matrix. In case of an axially loaded member, if
unit displacement is applied keeping all the other degrees of freedom fixed then
the force applied will be equal to the stiffness of the member or in other words,
nodal forces and nodal displacements are connected with each other by element
stiffness matrix. This is formulated depending upon the material and sectional
properties of the element. For information about the formulation of the complete
stiffness matrix please refer ref (13). In our program focus is on beamcolumn
and truss. So the element stiffness matrix will have 3 Dof at each node, thus for a
beam element stiffness matrix will be of the dimension [6X6].
1
1
1
1
1
1
1
1
]
1
¸
2
2
2
1
1
1
Z
Y
X
Z
Y
X
M
F
F
M
F
F
= E
1
1
1
1
1
1
1
1
1
1
1
1
1
]
1
¸
−
− − −
−
−
−
−
L
I
L
I
L
I
L
I
L
I
L
I
L
I
L
I
L
A
L
A
L
I
L
I
L
I
L
I
L
I
L
I
L
I
L
I
L
A
L
A
Z Z Z Z
Z Z Z Z
Z Z Z Z
Z Z Z Z
4 6
0
2 6
0
6 12
0
6 12
0
0 0 0 0
2 6
0
4 6
0
6 12
0
6 12
0
0 0 0 0
2 2
2 3 2 3
2 2
2 3 2 3
1
1
1
1
1
1
1
1
]
1
¸
2
2
2
1
1
1
θ
θ
v
u
v
u
If beam or truss is inclined at any angle θ w.r.t to horizontal the element stiffness
matrix, which is in the local coordinate system, has to be multiplied with a
 13 
transformation matrix to convert it into global coordinates. Thus the
transformation matrix is
A=
1
1
1
1
1
1
1
1
]
1
¸
−
−
1 0 0 0 0 0
0 cos sin 0 0 0
0 sin cos 0 0 0
0 0 0 1 0 0
0 0 0 0 cos sin
0 0 0 0 sin cos
θ θ
θ θ
θ θ
θ θ
Where θ is the angle made between the element and the horizontal. The matrix
operation involved is
[K]
global=
[A] [k]
local
[A]
T
Once the member stiffness matrix is formulated in Global coordinate sytem,
then they are assembled to form the combined global stiffness matrix [K].
Nonlinear analysis:
Most of the civil engineering structures fall in the category of lowrise structure
and the deflection are very small; there by do not exhibit significant nonlinearity. .
In structures like suspension bridges and very tall buildings that are slender,
before these structures reach limit of resistance they exhibit significant nonlinear
behavior. Thus even though most engineers are fluent in first order analysis it is
important to conduct an Nonlinear analysis which will help analytically simulate
 14 
more appropriate and better behavior of the structure than the first order
analysis. This will ultimately help in the optimal design of the structures.
In linear elastic structures, as seen before, it was assumed that the material is
unyielding and its properties invariable and the equations of equilibrium are
formulated on the geometry of the unloaded structures or, on the initial reference
of the configuration. In Nonlinear analysis the structure may lose its geometry
due to large deflections and the material can also exhibit nonlinearity. Nonlinear
analysis offers many options for solving the problems resulting from the above
cases. Geometric nonlinearity alone can be considered making an assumption
that the material is elastic including the effect of finite displacement and
deformation during the formulation of equilibrium equation. It is also possible
that material nonlinearity alone can be considered taking into account that the
member property changes under loads. Finally material nonlinearity and
geometric nonlinearity can be considered together in the analysis. Thus it is the
responsibility of the engineer to use his judgment and discretion while predicting
the most dominant behavior of the structure he wishes to simulate analytically.
According to W.Mcguire, R.H.Gallagher and R.D.Ziemian in their textbook
sources of nonlinearity are as follows.
Geometric effects:
1. Initial imperfections such as member camber and out of plumb erection
of a frame.
2. The P∆ effect, destabilizing moment equal to gravity load times the
horizontal displacement it undergoes as a result of lateral displacement.
3. The Pδ effect, the influence of the axial on the flexural stiffness of the
individual member.
 15 
Material effect :
1. plastic deformation of the steel structures.
2. Cracking or creep of reinforced concrete structure.
3. Inelastic interaction of axial force, bending, shear and torsion.
Combined effects:
1. Plastic deformation plus P∆ effect and/ or Pδ effects.
2. Connection deformation
3. Panel deformation
4. Contribution of secondary system to strength and stiffness.
Levels of analysis
The fig below shows loaddeflection behavior of plane frame building
analyzed using various refined and simplified models. The broken line is obtained
by I
st
order analysis ignoring the effects of both change in geometry and the
yielding of the material. The broken curved line is that of second order elastic
analysis that includes only the effect of change in geometry and instability of the
structure and this is the focus of thesis work. The curve approaches
asymptotically to the elastic critical limit generated by an Eigen value analysis. The
piecewise linear curve without a descending branch represents a first order
elastic–plastic hinge analysis when the effect of change in geometry has been
ignored. Its value is plastic limit load, which can be obtained by plastic analysis.
The piecewise linear curve with a descending branch shows the result of a
secondorder elasticplastic hinge analysis allowing for simple plastic analysis for
 16 
the change in the geometry associated with the sway deflection. The plastification,
initial imperfections, residual stresses and strain hardening are all accounted for,
the smooth continuous curve is obtained by doing an secondorder spread
plasticity analysis. The peak load is the true strength of the frame or the true load
carrying capacity of the structure.
Second Order elastic
First order Elastic
Second Order Elastic Plastic
Second  Order
SpreadofPlasticity
(plasticzone Theory)
L
o
a
d
i
n
t
e
n
s
i
t
y
Charecteristic Deformation
Plastic  Mechanism Load
Elastic Critical Load
First  Order  Elastic  Plastic
Fig 3
Load Deflection curve for a plane frame
In our thesis work the focus is laid on second order elastic analysis, taking into
consideration only the geometric nonlinearity ignoring the material nonlinearity.
The critical load is calculated by conducting a simple eigen value analysis. The
whole procedure has been converted in to a program which can be used to solve
problem.
Nonlinear analysis:
 17 
In II
nd
order analysis nonlinear equilibrium equations are reduced to a set of
simultaneous equations, which is written in matrix form. Once simultaneous
equations are written in matrix form, different methods can be devised to solve
there by displacement and deformations of the members can be traced
incrementally. Thus each method is a variation of global stiffness matrix and
symbolically written as
[K
t
] {d∆} = {dP}
[K
t
] = Tangent stiffness matrix.
{d∆}= incremental displacement.
{dP}= incremental loading.
The various levels differ in type of nonlinearity included in the types of
nonlinearity included [K
t
] , the way the equations are formulated and the details
of the equation solution, and the way members are subdivided.
Focusing on elastic II
nd
order analysis, Finite displacement and deformations are
accounted for in the equations of equilibrium and can be written as
[Ke + Kg] {d∆} · {dP}
[kg] is the geometric stiffness matrix. Many different ways are there to develop
the [Kg] matrix, dividing the members into sub elements yields very accurate
results. Many techniques for solving the equations are given later on in this
report. For the calculation of the elastic critical loads the global stiffness
equations is cast in the form of a generalized eigen value problem in which the
equations at the critical state is
 18 
{ } { } dp d
g
e
k
k
· ∆
1
]
1
¸
+
^
λ
Where K
^
g is called the geometric stiffness matrix computed for a reference load
{P
ref
}, {λ} (the eigen value) is the load w.r.t {P
ref
}, and {∆} the eigen vector is
the buckled shape. The lowest value λ that satisfies the equation of {∆} not equal
to 0 yields the Elastic buckling load.
Formulation of Kg:
As any standard text book Matrix analysis of structure would indicate the
technique involved in the derivation of Kg matrix, but we shall derive the Kg
matrix basic axial force member.
Let’s consider the segment of length dx as shown in the Fig4. Their initial
configuration is designated as ab. But after rigid body rotation and axial straining
its length is
dx
dx
dv
dx
du
dx
du
b a
2
1
2 2
2 1 ' '
,
_
¸
¸
,
_
¸
¸
+
,
_
¸
¸
+ + ·
Let’s designate,
,
_
¸
¸
+ + ·
dx
dv
dx
du
dx
du
dab 2
( )2
1
1
' '
ab d
dx
b a
+ ·
 19 
1
Mz1
Fy1
Fy1
Fy1
Fy2
E,A,Iz
L
2
Fy2
Mz2
X
x+u+dx+(du/dx)dx
b '
b
(v+(dv/dx)dx)
v1
u1
a
X dx
dx
1 '
x+u
a '
a '
b '
y
u2
2 '
x
Fig 4
Finite strains of planar element
or expanding a’b’ by binomial theorem and defining the extension per unit length,
(a’b’ – ab)/ dx , finite strain, e
fin
1
1
]
1
¸
,
_
¸
¸
,
_
¸
¸
+ ·
2 2
2
1
dx
dv
dx
du
dx
du
efin
Applying the principles of virtual displacement to the reference configuration
) ( int vol d e W fin x
vol
δ σ δ
∫
·
Integrating over the depth of the member, we get
 20 
,
_
¸
¸
1
1
]
1
¸
,
_
¸
¸
+
,
_
¸
¸
+
,
_
¸
¸
·
∫ ∫
dx
dx
dv
dx
du
A dx
dx
u d
A W
L
x
L
x
2 2
0 0
int
2
1
δ δ σ
δ
σ δ
let
,
_
¸
¸
dx
u dδ
= δ
,
_
¸
¸
dx
du
which is valid for infinite small displacement. Now using
the conventional elastic stress strain relationship in the first integral and letting
σ
x
A= F
x2
we get
,
_
¸
¸
1
1
]
1
¸
,
_
¸
¸
+
,
_
¸
¸
+
,
_
¸
¸
,
_
¸
¸
·
∫ ∫
dx
dx
dv
dx
du
F dx
dx
u d
A E
dx
du
W
L
x
L 2 2
0
2
0
int
2
1
δ δ
δ
δ
The first term in the equation above is for elastic stiffness matrix of the axial
force element and the second integral term is of great concern as it produces
geometric stiffness matrix [k
g
].Treating δ as differential operator w.r.t variables
du/ dx and dv/ dx, the internal virtual work can be written as
,
_
¸
¸
1
1
]
1
¸
,
_
¸
¸
+
,
_
¸
¸
·
∫
dx
dx
dv
dx
v d
dx
du
dx
u d
F W
L
x
δ δ
δ
0
2 int
Using the shape function the above equation can be written as
[K
g
] = F
x2
{ }
¸ ]
{ }
¸ ]
[ ]
∫
+
L
v v u u dx N N N N
0
' ' ' '
Where, {N’
u
} and {N’
v
} are the shape functions, in this case for axially loaded
members. Thus for axially loaded members displacement equation in the form of
shape function can be written as
u= (1ξ)u
1
+ ξu
2
and v= (1ξ)v
1
+ ξv
1
 21 
Therefore,
¸ ] ¸ ]
1
]
1
¸
−
·
1
]
1
¸
−
·
L L
N and
L L
N v u
1 1
'
1 1
'
Combining the equation suitably we get
[ ]
1
1
1
1
]
1
¸
−
−
−
−
·
1 0 1 0
0 1 0 1
1 0 1 0
0 1 0 1
2
L
F
K
x
g
Combining the bending and the axial forces the geometric stiffness matrix will be,
[ ]
1
1
1
1
1
1
1
1
1
1
1
1
]
1
¸
− −
− − −
− −
−
−
−
·
15
2
10
0
30 10
0
10 5
6
0
10 5
6
0
0 0 1 0 0 0
30 10
0
15
2
10
1
10 5
6
0
10 5
6
0
0 0 0 1 0 1
2 2
2 2
2
L L L L
L L
L L L L
L L
L
F
K
x
g
Different approaches to nonlinear analysis:
Basic principle involved in nonlinear analysis is to perform a set of matrix
structural operations to update the stiffness matrix constantly so that all nonlinear
effects of the structure are accounted. As equilibrium configuration of the
structure changes constantly, series of load increments are made to carry out the
analysis. The equilibrium and kinematics states of the structure from previous
cycle are used to formulate the stiffness relation for the solution of the next load
 22 
increment cycle. Thus, the solution of nonlinear analysis problem is got by a
series of linearized analysis. The size of the load increment has a profound
influence on the solution time and the convergence characteristics of the
problem. Smaller load steps yield accurate solution and in turn will take larger
number of load steps to cover the loads range. On the other hand if the
increment is too large, the number of iteration required to obtain a solution for a
given load step may be large and in some case the solution may even diverge. If a
noniterative scheme is followed the use of a large load increment may result in a
large drift off from the actual equilibrium path.
Thus nonlinear analysis is iterative or can be non iterative There are many
techniques and four major solution schemes for nonlinear problems are studied,
they are
1. Load control method.
2. Displacement control method
3. Arc control method
4. Work control method.
In applying these methods to trace the Nonlinear loaddeflection curve of the
structure, it is assumed that the equilibrium path to be followed is continuous and
unique.
It is assumed that the equilibrium and kinematics states of the structure at the (I
1)
t h
load step is know and it is now required to find the state of the structure or
the stiffness relation at i
t h
interval. Relation can be written as
K
i
j1
∆D
i
j
= ∆ R
i
j
+ Q
i
j
 23 
Where K
i
j1
= stiffness matrix for the load step I based on the equilibrium and
kinematics states at the end of the (j1)th load step are known.
∆D
i
j
= incremental displacement vector at the jth iteration load step i.
∆ R
i
j
· load increment at the jth iteration of load step I
Q
i
j
= force vector at the jth Iteration of the load step i.( difference between the
external and internal forces)
Suppose the solution converges after n iterations; the load at the end of the ith
load step is given by
R
i
= R
i1
+ Σ∆ R
i
j
And the displacement is given by
D
i
= D
i1
+ Σ ∆D
i
j
The updated stiffness matrix can then be built and the solution can move to next
increment. The solution schemes differ by size of the ∆R is determined and the
way iteration is carried out inside a load step.
Load control method:
Here the size of the incremental load at load step I, ∆R
i
, is obtained as a fraction
of the total applied load
 24 
∆R
i
· λ
i
R
i
In the above equation, λ
i
is the load increment factor and R is the total applied
load. The user determines the value of the λi by using personal judgment and
discretion. The load control method can be iterative or noniterative. Non
iterative method is referred to as simple step method and Iterative method
referred to as Load control NewtonRaphson method. In this thesis work to
calculate the nonlinear response of the frame tube building simple incremental
load method has been employed.
Simple incremental method:
In this method the unbalanced forces that exist in each load step is ignored. That
means no iteration is done. Thus stiffness relation that exits at load step I,
reduces to
K
i1
∆D
i
= ∆R
i
Here only one iteration not done, so the value of j= 1. Can be rewritten as
K
i1
∆D
i
= λ
i
R
From which ∆D
i
can be calculated. ∆D
i
calculated is added to the cumulative
sum ∆D
i1
from the previous cycles to get the new displacement D
i
D
i =
∆D
i1
+ ∆D
i
Thus updated stiffness matrix can be calculated.
 25 
Advantage of using simple incremental method is its simplicity. Its unloading
branch of the loaddeflection can be investigated by giving negative load
increments. The main disadvantage of this method is that the error tends to
accumulate and may prove to be disastrous if the structure is highly nonlinear in
nature. This is because iterations are not employed within a load step to bring the
calculated equilibrium path back to true equilibrium path. This is called drift off
error.
Load control Newton Raphsons method:
In this method iteration employed within the load cycle eliminate the unbalanced
force. The presence of unbalanced force indicate that the internal and the
external forces are not in equilibrium, and this errors occurs because of
linearization process in which current stiffness matrix is calculated based on the
passed configuration the structure. If this error is not corrected then the
calculated equilibrium path will drift off from the true equilibrium path. Lets take
ith load increment for the first iteration (j= 1)
K
0
i
∆D
1
i
= ∆R
1
i
K
0
i
∆D
1
i
= λ
1
i
R
Where ∆D
1
i
can be calculated from the above equation and added to the
cumulated displacement of the structure. Now the stiffness matrix can be
formed. Corrections are applied to the displacement by subjecting the structure
t o the unbalanced forces. Unbalanced forces are calculated as the difference
between internal and external applied load. Therefore subsequent iteration,
equation is
K
i
j1
∆D
i
j
= Q
i
j
 26 
The iteration stops when ∆D
i
j
or Q
i
j
is negligible. The advantage of this method is
that the drift off error is greatly reduced or controlled. The limitation of this
method is that it fails at limit points. This method fails also if the system exhibits
snap through behavior.
 27 
C h a p t e r 3
STRUCTURAL ANALYSIS OF TALL BUILDINGS
Introduction:
Is there an absolute solution to a problem? Hypothetically may be “yes”, in reality
we dare not to claim. Can this be our limitation? Or is it that the solution is
already in absolute state and our mind fails to perceive it. Have we to wait for
solution to evolve into absolute forms or rhetorical manner wait for the mind to
transcend into absolute form to get the solution. Whatever the answer may be, in
real world there is always a possibility for problems to arise even from well
accepted fact. When problem becomes conspicuous, complete understanding is
demanded for it. As problems get solved, they keep evolving, may be this is our
greatest problem. Question your self, are we really solving anything? Well, this is
our inspiration for the quest of perfection.
Rome saw its skyline pushed up by the construction of multistory buildings up to
4 stories high using wood, but eventually the great fire of Nero made
conspicuous its problems. As a solution, arches and barrel domes were
constructed of concrete and brick. Industrial revolution put lot of pressure on
construction industry to meet it demands, more and more different problems
became conspicuous and solution were demanded immediately. As a solution, tall
building were built to ease the space jams in urban areas, bridges and dams for
smooth communication and urban development and many more. All this could
be realized only by improved materials and structural innovations made during
the post industrial revolution. The tragic event of 9/ 11 has made conspicuous
 28 
another short come, which needs to be dealt with immediately. Solution for this
problem is inevitable, it just a matter of time until research will be done on every
aspect of tall buildings and solution will evolve.
There are a number of structural forms that are available for selection, selecting
an appropriate and optimal form is responsibility of a structural engineer. A range
of factors influence the selection of structural form they are internal planning,
material and the type of construction, the external architectural planning, the
planned location and the routing of service systems, the nature and the
magnitude of horizontal loading and the height and proportioning of the
structure. Taller the building slender it will be and becomes more dependant on
the structural forms. A major factor influencing the structural form is the
function of the building. Office building desire column free interiors so structural
form selected will be such that a lot of material is placed on the external
perimeter of the plan and internally placed around the elevators, stairs and service
shaft. For buildings taller than 10 stories, they need extra material to account for
resisting the wind loads.
In addition to the above factors, buildings structural forms are arranged such that
they support the gravity, dead and live loads and also to resist all levels of external
horizontal load shear. These necessities should be accomplished as economically
as possible. In the case of horizontal loading, tall buildings behave like a
cantilever. They may comprise of individual members acting in the form of a
cantilever, such as shear wall or cores, each bending about its own axis acting
unison through the horizontal in plane rigidity of the floor slab. It may also
comprise of a number of vertical columns and walls all acting compositely which
is made possible by using proper shear resistant connectors, there by each
member act as chord of a massive cantilever. Moreover, within the selected
structural form advantage can be taken by placing main vertical members in plan
 29 
so then the dead load of the structure can counter balance some all or a portion
of the tensile stresses developed in members due to horizontal load. This avoids
net tension occurring in vertical members and prevents t his uplifting force to rip
of the foundation.
Tall buildings are usually constructed of concrete, steel or composite. Steel
framing has played a prominent role in the development of tall buildings due to
its high strength to weight ratio. It is possible to have longer spans, partial
prefabrication leading to reduce on site work and more rapid erection, major
disadvantage being corrosion problem, fire resistance, expensive to clad, and
requiring expensive diagonal bracing or rigid frame connection. Concrete on the
other hand with shear walls resisting the horizontal loads and improved concrete
strength and innovation in the structural forms have allowed the height of the
structures to touch the 100story mark. The different structural forms are braced
frame structures, rigidframe structures, in filledframe structures, flatplate and
flatslab structures, shear wall structures, coupled wall frame structures, wall
frame structures, framedtube structure, tubeintube structures, bundledtube
structures, bracedtube structures, outriggerbraced structures, suspended
structures, space structure and hybrid structure.
Mathematical Modeling of a tall building:
A building's response to loading is governed by the components that are stressed
as the building deflects. Ideally for the ease of accuracy of structural analysis, the
participating components would include only the main structural elements, the
slabs, beams, columns girders, walls and cores. In reality, however, other non
structural elements are stressed and contribute to the building's behavior; these
include, for example, the staircase, partitions and cladding. To simplify the
problem it is usual in modeling a building for analysis to include only the main
 30 
structural members and to assume that the effects of the nonstructural
component are small and conservative.
To identify the main structural elements, it is necessary to recognize the dominant
modes of action of the proposed building structure and to access the extent of
the various member contributions to them. Then by neglecting consideration of
the nonstructural components, and less essential components, the problem of
analyzing a tall building structure gets reduced to viable sizes. In case of very tall
and complex building it is necessary to reduce even further the size of the analysis
problem by representing some of the structural elements with simpler analogous
components. Thus purpose of structural idealization is to represent all the
important features of the buildings for the purpose of deriving mathematical
equation’s governing the behavior of the system. The mathematical model should
include enough details to be able to describe the system in terms of the equations
without making it too complex. The mathematical model may be linear or
nonlinear, depending upon the behavior of the component system. Linear model
permit quick solution and are simple to handle, however nonlinear models
sometimes reveal certain characteristics of the systems that cannot be predicted
using linear model.
There by the mathematical model is gradually improved to obtain more accurate
results. Sometimes for approximate initial analysis, approaches crude or
elementary model are used to get a quick insight into the overall behavior of the
system. The following paragraphs to follow shall illustrate the modeling and
refining procedure.
 31 
Approaches to analysis:
The modeling of the tall building structure for analysis is dependent to some
extent on the approach to analysis which is in turn related to the type and the size
of the structure and the stage of the design or which the analysis is made .The
usual approach is to conduct approximate rapid analysis in preliminary design,
and a more detailed and accurate analysis for the final design stage. A hybrid
approach is also possible in which a simplified model of the total structure is
analyzed first after which the result are used to allow part by part detailed analysis
of the structure.
Preliminary Analysis:
The purpose of the preliminary analysis, that is analysis for early stage of the
design, may be to compare the performance of the alternative proposals for the
structure, or to determine the deflection and major member forces in a chosen
structure so as to allow it to be properly proportioned. The formation of the
model and the procedure for a preliminary analysis should be rapid and produce
results that are dependable approximates. The model and its analysis should
therefore represent fairly well, if not absolutely accurately, the principle modes of
action and interaction of major structural elements. It is important that the
solutions after preliminary analysis should yield deflection close to the exact
model and main member forces that are dependably within about 15% of the
value from an accurate analysis.
Intermediate and Final analysis:
The requirement of intermediate and the final analysis is that they should give, as
accurately as possible, result for deflection and member forces. The model
 32 
should, therefore, be as analysis program and computer capacity will allow for its
analysis. All the major modes of action and interaction and as many as possible
lesser modes, should be incorporated except where a structure is symmetrical in
plan and loading, the effect of the structure twisting should be included. If the
structure and loading are symmetrical, a three dimensional model would be
acceptable, or if repetitive regions up to the height of the structure can be
simplified by a lumping technique, this also would be acceptable.
In contrast to the reductions above, however certain final analysis may require
separate, more detailed analysis of particular parts, using forces or applied
displacement from the main analysis for example, in deep beams at transition
level of the structure around irregularities or holes in the shear wall. With easy
access to fast computer FEM model of the structure must be done as a part of a
final analysis.
Reduction technique:
When detailed modeling consumes lot of memory space and time, reduction
modeling can be used for simplicity and the ease of analysis. There are many ways
that a model can be reduced to simpler models, some of them are
1. Symmetry and Antisymmetry.
2. Twodimensional modeling of nontwisting model.
3. Lumping
a. Vertical lumping.
b. Horizontal lumping
The details for the above techniques can be found in ref (text book by Alex
Coull).A brief insight into lumped girder frame technique is presented below,
 33 
which is used as rapid analysis of frame tube building later on. In this method the
repetitive floor system are identified and is exploited to form a model with fewer
stories, which in turn benefits a rapid analysis. This method allows an accurate
estimate of the drift and a good estimate of the member forces. The girders are
usually lumped in sets of three girders or five girders into single girders thereby
reducing the number of stories considerably. For very accurate analysis it is
preferable not to include the first floor and roof girders during the lumping
process because the simulation of the frame behavior near the top and the base
of the building differ from the middle portion. In the figure below the first floor
and the top floor are untouched in order to maintain the boundary condition
same while the remaining floors are lumped up. In the building show below story
height is h and the column inertia is Ic and that of the girder is Ig.
3 girders lumped
3 girders lumped
3 girders lumped
Fig 5
Prototype rigid frame
 34 
Ige= Equivalent moment of inertia or the girder
Ice= Equivalent moment of inertia of the column.
Ige = ( I
g1
+ I
g2
+ I
g3
)
Fig6
Equivalent lumped girder frame
The requirement of a substitute frame is that, for horizontal loading, joint
translations should be the same as those of the original structure. For translations
caused by girder flexure, equation below shows the requirements to be satisfied
 35 
by assigning the inertia of the equivalent girder to be equal to the sum of the
lumped ngirder inertias of the original frame, that is,
∑
·
n
Ig Ige
To determine the properties of the columns in lumped girder method, the
component drift caused by the double curvature column bending story height nh
in the equivalent frame to the corresponding drift over n stories in the original
frame to the corresponding drift over n stories in the original frame are equated
to get the equivalent column inertias. Or in other words the flexural strengths of
the equivalent frame to original frame are equated to displacements to get the
equivalent column inertias.
( )
∑
∑
∑ ∑
,
_
¸
¸
·
,
_
¸
¸
·
n
i
n
i
Ic
n
Ice
h
Ic E
Qh
nh
Ice
E
nh Q
1
1
12
12
3
2 2
The column sectional areas, which control the cantilever deflection, must have
the same second moment about their common centroid in the lumped and the
original structure. Consequently, the areas of the equivalent and the original
structure remain the same as the original structure. The horizontal loading on the
equivalent structure frame is applied as equivalent concentrated loads at the
lumped girder levels, taking half new storyheight regions above and below the
lumped girder as tributary areas.
 36 
When the lumped girder frame has been analyzed the , the results are
transformed back to the original frame The moment in the original girder a t the
same level will be 1/ n
t h
of the resulting moment and the moments in the girders
between the lumped girder level shall be obtained by interpolating the moment.
Assumptions:
Following simplifying assumptions are necessary to reduce the problem to viable
sizes. The assumptions were made keeping in mind the frame tube building, its
anticipated mode of behavior, and the type of analysis.
The common assumptions are
• Participation components: Only the primary structural components
participate in the overall behavior. The effect of secondary structural
components and nonstructural components are assumed to negligible and
conservative. Although the assumption is generally valid, heavy cladding may
not be negligible and be significantly stiffening the structure. Similarly
masonry infill may significantly change the behavior and increase the forces
unconservatively in a surrounding frame.
• Floor slab: Floor slab is assumed to be infinitely rigid in plane. This
assumption causes horizontal plane displacement of all vertical elements at a
floor level to be definable in terms of the horizontal plane rigid body rotation
& translation of the floor slab. Thus the number of unknown displacements
to be determined in the analysis is greatly reduced.
• Negligible stiffness: components stiffness of relatively small stiffness is
assumed negligible. These include for example, the transverse bending
stiffness of slab, the minor axis stiffness of shear wall, and the torsion
stiffness of column beams, and walls.
 37 
• Negligible deformation: deformations, which are relatively small and of little
significance are neglected like the axial deformation of the beams, the
previously discussed in plane bending and shear of the floor slab.
Once the whole structure has being modeled for analysis, it has to be furthered
modeled for individual member and then joined together back into the structure,
satisfying all the compatibility the equilibrium equations.
Mathematical modeling begins with structural idealization members which have
depth and width as well as length, are conveniently reduced to line elements.
Their resistance to deformation is represented by material properties such as
young’s modulus ( E ) and Poisson ratio ( ν ), and by geometric properties of the
cross section such as are ( A ), moment of inertia ( I ), and torsion constant ( J ) .
The behavior of the connection, that is, whether pinned, rigid, semi rigid,
yielding, etc., must be stipulated. How these idealization decisions are made, or
the necessary property determined, is extremely important and involves
considerable judgment.
The different types of Idealization used in mathematical modeling for a frame
structure is explained in the following paragraphs. For the purpose of
mathematical modeling the state of stress in the member is represented by forces
at the element ends. The corresponding displacements of these nodes  the
degree of freedom – are employed in the characterization of the displaced state of
the element. The complete frame work has 12 nodal degrees of freedom and 12
nodal force components as shown in figure.
If the size problem still persists then a reduction analysis can be used for the
analysis of the structure. The technique is explained below. Our work
concentrates on the modeling and the analysis of frametube building. All of the
above assumptions are included for the analysis of frametube building.
 38 
Frame tube buildings:
Frame tube buildings consist of closely spaced column connected by deep
spandrel beams placed on the perimeter in plan. FramedTube buildings are
generally suited for story height from (40100). The close spacing of the columns
are viewed with mixed reaction with respect to aesthetics. Tubeform though
original developed for rectangular plans, it is now used for different shapes and
sometimes used for circular and triangular pans too.
In frame tube building columns are arranged unlike the frame structure with the
strong bending direction of the columns aligned along the face of the building.
Main reason do so is to make sure large amount of material is placed on the
periphery of the building to maximize the inertia of the building. In many
buildings this is used to resist entire lateral load on the building to be resisted by
the exterior columns while the gravity loads are shared between exterior columns
and the interior structure in the form of columns and load inner core. The panel’s
normal to the direction of the wind is considered as flange and the panel
perpendicular is considered as web of the perforated cantilever. In the lower floor
it is desired to have columns spaced wide apart for which an deep transfer girder
is provided which connects all the loads from the upper floors and transfers to
the columns below.
The uniformity of the structure makes construction simpler and enables the
industrialized production of the components. For steel buildings large elements
can be prefabricated in the factory and easily erected in the work site. In case of
concrete construction use of gang forms raised enables speedy construction.
 39 
Mode of behavior:
Although the structure has a tubelike structure, its behavior is more complex
than a plain imperforated tube as the stiffness may be less. When subjected to
lateral load the flange portion of the structure, on the windward is subject to
tension force and the flange portion on the leeward side is subjected to
compression, as indicated in the fig 7. In addition to this, the parallel frames to
the wind are subjected to the usual inplane bending and the shearing or racking
action associated with an independent rigid frame. This action gets problematical
by the fact that flexibility of the spandrel beams produces a shear lag effect,
thereby increasing the compressive stresses in the corner column and decreasing
the stresses in the interior columns of both the flange and the web portion of the
structure. This can be seen in the fig 8.
When a frame tube building is subjected to lateral loading it is resisted by web
side panel, which deform such that the columns A and B are in tension and
columns C and D are in compression. The principle interaction between the web
and the flange would be through the vertical displacement of the corner columns.
These displacements correspond to the vert ical shear in the girder of the flange
frame, which mobilizes
 40 
Axial Stresses in the web
columns
W
e
b
F
r
a
m
e
s
Y N Axis
W
e
b
F
r
a
m
e
s
Axial stress distrabution in columns of
laterally loaded framed tube
Flange Frames
A
Wind
X
B
Tension
D
Flange Frames
Compression
Compression C
Axial Stresses in Flange
Column
Fig 7
the column forces in the flange column. For example if the column C is in
compression it will try to compress the adjacent column since the two are
connected by spandrel beam, the compressive deformation will not be
identical since the spandrel beam will bend. Thus the axial deformation in
adjacent columns will be less, by an amount equal to the stiffness of the
 41 
connecting beam. Hypothetically, beams of infinite stiffness will develop a
pure tubular action. Thus the deformations in successive adjacent columns in
frame will be lesser than the previous one. Since the external applied moment
is to be resisted by internal couple produced by the compressive and the
tension stress produced on either side of neutral axis of structure, it follows
that the stresses in the corner column will be greater than those in the interior
column.
C2 C1 C3
∆
Deformation of the flange frame causing shear
Fig 8
Structural analysis:
Ref (12 ) During analysis has assumed that the horizontal diaphragm slab has
infinite in plane stiffness. This maintains the cross section shape of the structure
at each level and the crosssection at these positions under go only rigid body
 42 
movements in plane. Horizontal displacement of the building can be expressed in
terms of orthogonal translation and rotation. The out of plane stiffness of the
floor slab is assumed low so that they do not resist twisting and bending. The
floor slab is assumed unable to provide the coupling between opposite normal to
wind which act as flange of the side frame. Both the side and normal frames are
therefore subjected largely to in plane action, and the out of plane actions are
generally negligible. When a frametube building is subjected to lateral loading the
action of the floor system is mainly to transmit the horizontal forces between the
different vertical members of the structure. As floor systems does not participate
otherwise in the lateral load resistance of the structure, then, provided that the
floor loading is constant through out the structure, a repetitive floor loading can
be economically used for the design and the analysis of the structure.
If the plan of the frame tube building happens to be a square i.e. is doubly
symmetrical about the plan, any applied load can be resolved into two orthogonal
force components about the axis, and a twisting moment about the central axis of
the structure. The response of the structure can be obtained by a superposition of
the separate bending action about the XX axis and the YY axis and a pure
torsion action. Further the individual action may be simplified by utilizing the
double symmetry of the structure in plan there by ½ structure analyses or ¼
structure analyses can be made.
Usually a three dimensional structural analysis of frame tube buildings can be
carried out using stiffness approach, but its feasibility depends on the computer
capacity. Though the analysis is possible it is time consuming and very costly,
even though it is theoretically straightforward. Analysis of FrameTube buildings
is a complex problem as they are highly indeterminate structure, consisting of
series of rigidly joined frame works connected together at the corners of the
buildings. Each frame consists of large number of elements and number of
 43 
degree of freedom may run into thousands. Consequently, the size of the stiffness
matrix can be considerably reduced and hence the computation, by doing an
analysis of ½ structure or even ¼ structure if the structure is symmetrical in plan
about one or two central axes. The applied load may be treated as a combination
of symmetric or skew symmetric system, acting on either a ½ or ¼ structure.
In the building, the high in plane rigidity of the floor slabs will have a
considerable influence on the structural behavior, by ensuring that the out of
plane deformation of the panels will be effectively restrained at each floor level.
The main action of the panels will be effectively restrained at each floor level.
Main action will then be in planes of the frame panels. As a result of the in plane
rigidity, it may be assumed that the cross section of the building will undergo only
rigid body displacements, translation and rotation, in the horizontal plane at each
floor level. To obtain an accurate estimate of the structural behavior, it is essential
to include this constraining action in the structural analysis of the three
dimensional formwork.
The constraining action can be achieved in a number of different ways. One way
is using the option provided by the program, which provide an option for end
constraints. If not a fictitious axially rigid joint pinended horizontal diagonal
bracing members connecting nodes on the opposite corners at each floor level.
The restrained corners then remain fixed relative to each other during any
translation or rotation under applied loads. In addition, the axial stiffness of the
beams at each story level can be assigned to be so large that any in plane axial
deformations, or relative displacements between nodes, are negligible. The
introduction of such diagonal bracing member has the disadvantage that it
increases the bandwidth of the stiffness matrix and increases the solution time.
 44 
Reduction of ThreeDimensional Framed Tube to an Equivalent Plane
Frame
In this section is presented a simplified yet accurate approximate method for the
analysis of symmetrical framedtube structures subjected to bending produced by
lateral forces. The method is intuitively appealing to the engineer since, by
recognizing the dominant mode of behavior of the structure, it is possible to
reduce the analysis to that of an equivalent plane frame, with a consequently large
reduction in the amount of computation required for a conventional full three
dimensional analysis.
Consider initially the framed tube of Fig.7 subjected to bending by lateral forces
in the X direction. The lateral load is resisted primarily by the following actions.
1. The shearing actions in the web panels AD and BC parallel to the direction of
the applied load.
2. The axial deformations of the normal frame panels AB and DC action
effectively as flanges to the web panels.
Due to the symmetry of the structure about the XX axis, and the very high
inplane stiffness of the floor slabs, it may be assumed that outofplane actions of
the web frames are negligible, and the frames are subjected only to planar actions.
It is also assumed that the torsion rigidities of the girders are negligible.
The axial displacements of the corner columns in the web frames are r estrained
by the vertical rigidity of the two normal frames. Consequently, the interaction
between the flange and web panels consists mainly of vertical interactive forces
through the common corner columns, A,B,C, and D. As a result of these
interactive forces, the flange panels AB and DC are subjected primarily to axial
 45 
deformation, the uniformity of which across the panel will depend on the
stiffness, that is, on the spans and flexural rigidities, of the connecting spandrel
beams at each floor level.
Under the applied lateral loading, the shear forces will this be resisted mainly by
the web frames, while the bending moments will be resisted by the moments and
axial forces in the columns of the web frames and the axial forces in the columns
of the flange frames. By virtue of the large lever arm that exists between these
flange panels, the wind moments will be resisted most effectively if the maximum
amount of axial force can be induced in the columns of frames AB and CD.
All other torsion and outofplane actions may be considered to be secondary,
apart from the outofplane bending of the columns in the flange frames whose
horizontal deflections will be the same as those of the web frames. This action
may be of significance in the lower levels of the building since bending then
occurs about the weaker axis of the columns.
In analyzing the primary mode of behavior, the fundamental compatibility
condition that must be established is that of equal vertical displacements at the
corner where the orthogonal panels meet. In the analytical model, a mechanism
is required that will allow vertical shear forces, but not horizontal forces or
bending moments, to be transmitted from the web panels to the flange panels
through the corner columns.
In addition, for the web frames, the joints must be free rotate in the plane of the
frame, to displace vertically, and to displace horizontally in unison in the plane of
the frame at each floor level, sue to the inplane rigidity of the floor slabs. For
the flange frames, the joints must be free to rotate in the plane of the frame, and
 46 
to displace vertically. But flange joints on the line of symmetry must, and
preferably all flange joints should, be constrained against horizontal displacement
in the plane of the frame, as a result of the slab inplane rigidity.
For example, consider the simple framed tube shown in plan in Fig.9. Since the
structure is symmetrical about both center lines XX and YY, only one quarter, for
example, EBH, need be considered in the analysis. The required boundary
conditions at the lines of symmetry and skew symmetry are then introduced as
described because of symmetry about the XX axis, the shear force in the beams,
and the slope in the Y direction, of panel AB must be zero at the line of
symmetry (E). Conditions of skew symmetry (H) must be zero. If, on the other
hand, the web frame BC contains an even number of columns, so that no column
is situated on the center line YY, the bending moment at the line of skew
symmetry must also be zero. Appropriate support systems to simulate the
required boundary conditions for the quadrant EBH of the structure of Fig.9 are
shown in Fig10.
 47 
Y
W
e
b
F
r
a
m
e
s
N Axis
Fig 9
Frame tube structure. Structural plan
Fig 10
Connection Details
 48 
fig10
The detail of a connection
The equivalent planar system is obtained by “rotating” the normal halfpanel EB
through 90° into the plane of the webhalfpanel BH. The inplane stiffness of
the floor slabs constrains all members of the two web bents to have the same
horizontal deflection in the X direction; therefore it can be assumed that the one
quarter of the total lateral forces acting on the faces of the flange frames of the
building can applied in the plane of the halfweb frame, as indicated in Fig.6.
Since the beams are assumed axially rigid, the forces may be applied at any
convenient nodes.
The desired vertical interaction between the web and flange panel may be
achieved in various ways.
Most comprehensive modern generalpurpose structural analysis programs
include an inter nodal constraint option. This allows the displacement relating to
specified degrees of freedom at two or more nodes in structure to be constrained
 49 
to be identical. The a ppropriate nodes at the intersections of the web and flange
frames may then be specified directly in the analysis to have equal vertical
displacements.
If this option is not available, some other device must be used to achieve the
required vertical compatibility at the junctions.
One simple technique is to displace horizontally the intersection column of each
flange frame by a small distance of, say, onehundredth of the span of the
adjacent beams, so that each common intersection joint is represented twice,
once on each of the web and flange joints(B and B’ in Fig.6). In numbering the
nodes, the two nodes representing each intersection joint are numbered
separately. The duplicate nodes are then joined by a fictitious stiff beam with a
flexural rigidity of say 10,000 times that of the larger adjacent girder and with one
end assigned to be released for moment and axial force (Fig.7). By this device,
vertical compatibility is established and the required shear transmitted between
the web and flange frames, while decoupling the rotation and horizontal
displacement. However, it has been found that the results may be sensitive to the
stiffness assumed for the fictitious beams.
An alternative technique has been devised to improve the disconnection of the
rotations and lateral displacements between the frame panels, and to reduce the
sensitivity of the results to the stiffness of the connecting fictitious members.
The technique again involves the rotation of the flange frames into the plane of
the web. The intersection line column of each frame is shown superimposed on
the distance; say less than one hundredth of the story height. Thus each
intersection joint is duplicated, once on the web column, and, immediately above,
on the flange column. These common nodes should again be numbered
separately. Each pair of intersection joint nodes is then connected by a fictitious
 50 
stiff vertical link with a large sectional area of, say, 10,000 times that of the
intersection line column. The stiff links ensure vertical compatibility and transfer
vertical shear between the web and flange frames while disconnecting rotations
and vertical displacements.
In each model, the corner column is assigned its true inertia’s in the
corresponding planes of the web and flange frames, but its area should be
assigned wholly to the column B’ in the web frame with a zero area assigned to
the column B in the flange frame. Horizontal and rotational constraints are
applied to the flange frame nodes on the vertical line of symmetry and preferably,
as a means of reducing the total number of degrees of freedom, horizontal
constraints are also applied to all other flange frame nodes.
The resulting planar model does not include the outofplane of columns in the
normal frames, which may be of significance in the lower levels. These columns
suffer the same horizontal deflections about their weaker axis as the columns in
the side frames do about their stronger axis. The effect of the outofplane
bending may be included in the basic model by adding an equivalent column (RR
in Fig.6), whose flexural rigidity is equal to the sum of the outofplane flexural
rigidities of onequarter of the flange columns. The additional column is
connected by the pinended axially stiff links to the existing basic plane frame
system. The links constrain the column to have the same horizontal deflection as
the side panel members, and allow it to carry its share of the lateral forces. Once
the total force and consequent moment has been determined for the effective
column, it may be distributed to the individual column in proportion to their
flexural rigidities.
 51 
C h a p t e r 4
FINDINGS AND DISCUSSION
The frame program developed in this thesis is intended to focuses on planar
structures. It can handle both trusses and frames. Each element has 2 nodes
and each node has 3 degrees of freedom. The program has been tested for
many varieties of problems and is fairly stable and the results are very accurate.
The program is based on the matrix methods of structural analysis and is
developed using Matlab. Care should be taken during the input process as it
has a great impact on the result. Input data for 40stories structure and its
reduction model, both by program and Mastan2 are as follows,
Frame tube building reduction model:
Data:
Number of nodes = 108 nos.
Number of elements= 184 nos.
Width of the bay = 12 ft
Building dimension in plan = 132ft X 132ft
Story height = 60 ft
Total height = 480 ft
Wind load = 87.5 Kips/ node
Floor load = 218.75 kips/ column
Actually structureFrame tube building 40 stories:
Data:
Number of bays = 11 nos
Width of the bay = 12 ft
 52 
Building dimension in plan = 132ft X 132ft
Story height = 12 ft
Total height = 480 ft
Column modulus of elasticity (E ) = 3150 Ksi, 453600 ksf
Spandrel Beam (E) = 3150 Ksi
Wind load = 17.4 Kips/ node
Floor load = 43.5 kips / column
Mastan2 model for 8 story building:
Data:
Number of Nodes = 108
Number of Elements = 184
Number of Sections = 6
Number of Materials = 1
Number of Supports = 12
Applied Loads
Lateral load= 87.5 Kips/ node
Floor Load= 218.75 Kips/ node
Mastan Model 40 story building:
Data:
Number of Nodes = 492
Number of Elements = 920
Number of Sections = 6
Number of Materials = 1
Number of Supports = 12
Applied Loads
Lateral load= 17.5 kips/ node
 53 
Floor load= 43.5 Kips/ node
Comparative result: [reduction model Vs actual model]
40 Story Model [Program Vs. Mastan]
Program Mastan Program
X
Displace
ment
(in)
Y
Displace
ment
(in)
θ
Rotation
(rad)
X
Displace
ment
(in)
Y
Displace
ment
(in)
θ
Rotation
(rad)
I
order
Analy
sis
2.8934 1.93430 0.00041 3.792 2.2752 0.00041
II
Order
Analy
sis
2.95 0.00041 3.8676 2.2836 0.00054
Reduction Model [Program Vs. Mastan]
Program Mastan Program
X
Displace
ment
(in)
Y
Displace
ment
(in)
θ
Rotation
X
Displacem
ent
(in)
Y
Displace
ment
(in)
θ
Rotation
I
order
Analy
sis
3.30 2.18 0.00054 3.30 2.18 0.00054
II
Order
Analy
sis
3.84 0.00054 3.36 2.19 0.00055
From the values of comparative results table we can calculate the drift index of
the building for both program and Mastan2. In case of matlab program drift
index for first order analysis it is 0.000503 less than the allowable drift index of
 54 
0.002., while the drift index for second order analysis of the building we get
0.0006713, which is less than 0.002, the limit value of the drift index. The drift
Index calculated for Matan2 model is, first order analysis it is 0.00066 and for
second order analysis it is 0.00067, both less than the allowable limits. This
confirms the stability of the building for deflection criterion. Comparing the
results of actual model with reduction model we see the values are quite close
to one another.
When first order analysis is done using the developed program, it is observed
that the values when compared between actual and reduction model, a
difference of 0.4” (actual model – reduction model) is got. While that for
second order analysis it is 0.9”. When First order analysis was done using
Mastan2, we see that error between models are, + 0.49” (actual model –
reduction model) and that for second order analysis is + 0.5”. This is because in
40 story building, story shear causes shear deformations, which is not
accounted for in the program developed. The story deflection we get is less
than the values calculated using Mastan2 for 40 story building. When number
of stories is less, shear deformation is also less, thereby solutions obtained
from the program is close to Mastan2 solutions. It is observed that shear
deformation causes error to accumulate as the number of stories is increased,
It is observed that the in reduction model, doing a first order analysis the
solutions from the program and Mastan are exactly same up to 2 decimal
places. The second order analysis results between the models differ by 0.48”
this is because drift off error is not controlled in the program unlike the
Mastan2.
 55 
The results from the buckling load analysis in the case of very tall buildings
have to be sorted out and the least positive value is the buckling load and
remaining values are the mode shapes. The buckling load ratio from the
program is 61. Calculations using Mastan2 yielded approximately 50. The
values differ because the program developed calculates buckling load by
applying load in one step unlike Mastan2 where load is applied in increments.
For small structures with story height below 8 errors is the same as Mastan2
up to 2 decimal places. The member forces, member stress obtained from this
program were very accurate and reliable. Nodal displacement and rotation are
also accurate. Program run time is very small in fraction of seconds and the
results obtained are accurate. For solving building over 50 stories good
computer is a necessity as total degree of freedom of the whole structure
increases and we encounter memory problem.
Recommendation for future research:
The program developed is modular program and therefore in future research
appending dynamic analysis modules to this program would be a good start.
Iterative program employing Newton Raphsons method has been developed
but could not be checked for numerical stability and so has not been included
in this wok so a future research to append the remaining different second
order methods like the work control method, arc length control method and
the displacement control method could be attempted.
Finite element method is the most widely used method to solve engineering
problem, as matrix method is a back bone of FEM programs, this program
would serve as a skeleton to build a advanced Finite element software.
 56 
The algorithm still needs to be fine tuned for numerical stability and checked
for unique cases.
Finally, I think this program has all that one needs for the development of a
basic and simple FEM program, so an effort in this direction is worth a try.
 57 
GLO SSARY
First order analysis. An analysis carried out for structures whose deformation is
small and exhibit elastic behaviour.
Second order analysis. A iterative method which takes into account the material
and geometric nonlinearity.
Elastic buckling load. The load at which the deformation of a slightly
imperfect system increases without bound.
Stable equilibrium. A state of equilibrium where in a body is slightly displaced
from its original position of equilibrium, will return back to that position
subsequent to the removal of the disturbing force.
Unstable equilibrium: A state of equilibrium in which, when a body is displaced
slightly from its equilibrium position of rest, does not return, but instead
continues to move further away from original equilibrium position.
Neutral equilibrium: A state of equilibrium in which the body remains in the
position to which the disturbing force has moved it.
Mastan2. A Matrix structural analysis commercial soft ware.
 58 
R e f e r e n c e s
1. Coull, A., and Subedi, N. K.,” Framed–Tube Structures for Highrise
Buildings” Journal of the Structural Division, ASCE, Vol. 97, No. ST8,
Proc. Paper 8301, Aug., 1971, pp, 20972105.
2. Coull, A., and Bose,B.,” Simplified Analysis of FramedTube structures,”
Journal of the Structural Division, Proceedings of the American Society
of Civil Engineers, Vol.101, No. ST.11, Proc. Paper 11696, Nov
1975,pp.22232240
3. Coull, A., and Bose,B.,” Torsion of FramedTube structures,” Journal of
the Structural Division, ASCE, Vol.102,No.ST.12,December, 1976,
pp.23662370.
4. Coull, A., and Ahmed, A, K., “Deflections of Framed Tube Structures,”
Journal of Structural Division, Proceedings of the American society of
civil Engineers, Vol. 104, No.St5, May, 1978, pp 857 861
5. Khan F.R and Amin N.R., “Analysis and Design of Framed Tube
structure for all tall buildings”, ACI Publication SP36: Response of
multistory Concrete structures to Lateral Forces, 1973, pp3960.
6. Khan, F, R., ”Current Trends in Concrete High Rise Buildings,”
Proceedings, Symposium on Tall buildings, University of Southampton,
Pergamon Press, London, England, Coull and Stafford Smith, eds., 1967,
pp.571590.
7. J. J. Connor and C.C. Pouangare, “A simple model for the Analysis and
Design of Frame Tube Structures.”
 59 
8. Schwaighofer, J., and Ast, P.A., “Tables for the Analysis of FramedTube
Buildings,” Publication 7201, Department of Civil Engineering,
University of Toronto, Toronto, Cannada, Mar., 1972
9. H. De Clercq, “Analysis and design of tubetype buildings structure,”
PhD dissertation, earth quake engineering research center, college of
engineering research center, University of California at Berkeley, ca.
10. Taranath, B. S. “Structural analysis and design of tall buildings,” New
York, McGrawHill, c1988..
11. Alexander Chajes, “Principles of Structural Stability theory ”Civil
Engineering Mechanics series, 1974, PretenceHall, Inc., Engle Cliffs,
New jersey.
12. B. S. Smith and A. Coull, Tall Building Structures Analysis and Design,
Tall Buildings StructuresAnalysis and Design, John Wiley & sons
Publication.
13. W. McGuire, R. H. Gallagher, and R. D. Ziemian, Matrix Structural
Analysis, Second Edition, John Wiley & Sons, Inc.
14. W. F. Chen, and E. M. Lui, Stability Design of Steel frames, CRC press,
Boca Raton, Florida.
15. W. F. Chen, and E. M. Lui, Structural Stability  Theory and
Implementation. Elsevier Science Publication Co., Inc.
16. H. Grandin, Jr., Fundamentals of the Finite Element Method, Waveland
press, Illinois.
 60 
17. P. Jayachandran, M.ASCE, “Structural collapse of the World trade center
Towers Subjected to Explosions caused by Aircraft”.
18. Carrie, P.F. “Approximate Analysis and optimization of tall framed tube
buildings” Ms Thesis dissertation, Dept of Civil engineering, Worcester
Polytechnic Institute, MA.
 61 
A P P E N D I X  A
MANUAL
Matrix Program is simple to use and userfriendly. No prior knowledge of the
language of the program is necessary to get started, but to develop complex
models knowledge of Matlab is necessary. The programs have a modular
structure and user adaptive. This dynamic nature of the algorithm is the start to
more advanced programming using matrix methods.
The objective of this manual only to make the user an insight into the program
developed and to understand the different variables he inputs into the program
with the aid tutorial problems. Section 1 explains the program using the flow
chart and algorithms. This manual presents two tutorials. Section 2 explains the
input file of a simple 2D Truss, which can be quickly modeled using these
programs. This tutorial is recommended as a quick start. New users should go
through this tutorial first.
Section 3 presents an input file for frame tube building in two parts. In first part
modified frame tube structures are presented. They are developed as RC
Framed structure using the GUIinput program devolved which later runs to
give the results. They are compared with other commercial software and errors
have been presented. Lastly these tutorials shall help to understand the basis for
the analysis of the 40 storey structure.
The above tutorial problems include only Analysis and the output are plotted
and displayed on the screen as well as in an output file.
 62 
Algorithm:
Here is given a series of Algorithm to give a brief insight into the logic of the
program.
ALGORITHM
1. Input Structure details: READ:
o Coordinates
o Applied Forces
o Boundary Conditions
o Element group and Material properties
o Element Connectivity
2. Construct Stiffness Matrix and Force Vector
o Construct Element stiffness matrix
o Group to form global structure stiffness matrix [ K
e
]
o Construct Force Vector { F }
3. OPTION : a) Linear b)Nonlinearc)Critial buckling
a) LINEAR
o Compute displacement Vector { U } = [ K
e
]
1
{ F }
o Calculate Internal Stresses and Forces
o Display results
o STOP
b) NONLINEAR
o Construct element Geometric stiffness matrix initially with zero
internal stress
o Group to form initial global structure geometric stiffness matrix
[ K
g
]
1
o Total Structure Stiffness matrix [ K
T
] = [ K
e
] + [ K
g
]
1
o Construct ∆F = [ F ] / n ; n = number of iterations
o Loop i to n
§
Compute { ∆U } = [K
T
]
1
{ ∆F }
§
Update coordinate displacement Vector { U } = { U } +
{ ∆ U }
§
Compute internal stresses and forces in each element
 63 
§
Construct new geometric stiffness matrix [ K
g
]
i
with
computed internal stresses
§
Update Total Structure stiffness matrix [ K
T
] = [ K
e
] + [
K
g
]
i
o End loop
o Plot Loaddisplacement Curve for specified node and DOF
o STOP
c) CRITICAL
o Construct Element stiffness matrix
o Group to form global structure stiffness matrix [ K
e
]
o Construct Force Vector { F }
o Compute internal stresses and Forces
o With these forces, construct element geometric stiffness matrix
o Construct Structure Geometric Stiffness matrix [ K
g
]
o Solve [ K
g
+ λ K
g
] { ∆ } = 0 for λ and { ∆ } which represent the
eigenvalues (critical load factor corresponding to the specified force
vector) and the corresponding mode shapes.
o Display results
o STOP
TUTORIAL 1
Description of the Tutorial Problem
The structure for this project is a single bay, two story RC frame that will be
analyzed. A first order analysis and Elastic critical buckling load analysis is done
and results are presented. Comparative results are also tabulated in the end. The
figure below shows the structure.
F 3 6
F 2 5
1 4
 64 
A n input file called " I N T R O.tx t" containing the input data for the above structure has been
provided with the program. T his file contains input data.
Basic Data
CONTROL INFORMATION
Number of Nodal Points = 6
Number of Spatial Dimensions = 2
Number of Degrees of Freedom = 3
Number of Element Groups = 1
Number of Material = 1
NODAL COORDINATES:
Node X Y Z
1 0.00000000 0.0000000 0.00000000
2 0.00000000 50.0000000 0.00000000
3 0.00000000 100.0000000 0.00000000
4 5.00000000 0.0000000 0.00000000
5 5.00000000 50.0000000 0.00000000
6 5.00000000 100.0000000 0.00000000
BOUNDARY CONDITIONS
Node X Y Z
1 1 1 1
2 0 0 0
3 0 0 0
4 1 1 1
5 0 0 0
6 0 0 0
1= Fixed, 0= Free
 65 
APPLIED FORCES
Node DOFs
1 0.00000000 0.00000000 0.00000000
2 1000.00000000 0.00000000 0.00000000
3 100.00000000 10000.0000000 0.00000000
4 0.00000000 0.00000000 0.00000000
5 1000.00000000 0.00000000 0.00000000
6 0.00000000 10000.0000000 0.00000000
ELEMENT INFORMATION
Element Group Number 1
A E I G
10 200000 100 0
CONNECTIVITY:
1 1 1 2 0 0 0 0 0 0
2 1 2 3 0 0 0 0 0 0
3 1 4 5 0 0 0 0 0 0
4 1 5 6 0 0 0 0 0 0
5 1 2 5 0 0 0 0 0 0
6 1 3 6 0 0 0 0 0 0
Refer the intro.txt file. The input data is explained below in the following
paragraphs.
 66 
In the second line we 6 2 3 1 1 all separated by space represent number of
elements, number spatial degree of freedom( 2D planar), Degree of freedom per
node, Element group type ( possibilities are 2, i.e either beamcolumn or truss),
type of analysis (1 = Linear Elastic analysis, 2 = Nonlinear analysis, 3 = critical
buckling load analysis ).
Next 7 lines is for coordinate positions. For example, 1 0 0 0 all separated by
spaces represent node number, co ordinate position in x, coordinate position in y
and lastly interpolation decision variable (1 invokes interpolation loop, if there are
series of nodes then instead of entering one by one it can be done in 1 step).On
the 9 line we have 0 which represents end of data for read_coord module
program.
Lines 10 11 12 are for support condition. On line 10 we have 1 1 1 1 0 which
represent node number, fixed x Dof, fixed y Dof, fixed rotation about z axis, no
interpolation. Note 0 represent free and 1 represents fixed condition.
Lines 1317 are for force. On line 13 we have 2 1000 0 0 0 represent node
number, Force in the xdirection, Force in the YDirection, rotation about the z
axis and no interpolation. Note ve sign represent force towards left of node and
downwards direction. On 17 line 0 represent termination of the respective
program module.
On Line 18 we have 1 4 6 1 representing element group number, element type (3
represent truss type and 4 represents beam type. Don’t think what happened to 1
and 2 that’s not your problem), number of elements, number of materials used.
On Line 19 we have 1 10 200000 100 0 representing Material number, area of
cross section, Modulus of elasticity, moment of inertia and Shear modulus. The
remaining input data is for node connectivity. On line 20 we have 1 1 1 2 0
representing element number, material number, i
t h
node, j
t h
node and no
interpolation.
 67 
This completes the input file. Fem_main is the main program which needs to be
invoked then type the input file name and then program run to finish. For more
logical understanding refer the algorithm and flow chart.
OUTPUT
PHI OUTPUT
Node No. Phi
X Y θ
1 0.00000 0.00000 0.00000
2 1.65683 0.14009 0.04440
3 3.94695 0.38465 0.04616
4 0.00000 0.00000 0.00000
5 1.65683 0.35991 0.04440
6 3.94683 0.61535 0.04616
Elastic Critical Buckling Load: (λ)
0.78
5.79
19.01
39.71
82.95
163.89
200.15
200.00
 68 
Comparative results:
TUTORIAL 2
PROBLEM DICRIPTION:
The structure for this project is a 2D planar truss that will be analyzed. A first
order analysis is done and results are presented. Comparative results are also
tabulated in the end. The figure below shows the structure.
Program
M astan
X
Y
θ
X
Y
θ
0.00000
1.65683
3.94695
0.00000
1.65683
3.94683
0.00000
1.4009
3.8465
0.00000
3.5991
6.1535
0.00000
0.04440
0.04616
0.00000
0.04440
0.04616
0.0000
1.6568
3.9470
0.0000
1.6568
3.9468
0.0000
1.4009
3.8465
0.0000
3.5991
6.1535
0.0000
0.044398
0.046163
0.0000
0.044398
0.046162
 69 
2D Planar Truss
A
B
C
D
Fh
Fv
CONTROL INFORMATION:
Number of Nodal Points = 4
Number of Spatial Dimensions = 2
Number of Degrees of Freedom = 2
Number of Element Groups = 1
NODAL COORDINATES
Node X Y Z
1 0.00000000 0.00000000 0.00000000
2 5000.00000 0.00000000 0.00000000
3 5000.00000 8670.000000 0.00000000
4 10000.0000 8670.000000 0.00000000
BOUNDARY CONDITIONS
Node DOFs
1 1 1
2 0 1
3 0 0
4 0 0
APPLIED FORCES AND DISPLACEMENTS
Node DOFs
 70 
1 0.00000000 0.00000000
2 0.00000000 0.00000000
3 0.00000000 0.00000000
4 282840.0000 282840.0000
ELEMENT INFORMATION
Element Group Number 1
CONNECTIVITY
1 1 1 2 0 0 0 0 0 0
2 2 1 3 0 0 0 0 0 0
3 2 2 3 0 0 0 0 0 0
4 1 3 4 0 0 0 0 0 0
5 2 2 4 0 0 0 0 0 0
Referring to the input file truss.txt, Input basically is the same with changes
as to Dof per node which is 2 in this case.
Output: Nodal displacement
PHI OUTPUT
Node No. Phi
X Y
1 0.00000 0.00000
2 0.40779 0.00000
3 9.83626 2.23479
4 10.95114 7.80813
 71 
Comparative results:
Program
M astan
X
Y
θ
X
Y
θ
0.00000
0.40779
9.83626
10.95114
0.00000
0.00000
2.23479
7.80813
0.00000
0.00000
0.00000
0.00000
0.0000
0.40779
9.8363
1.0951e+ 001
0.0000
0.0000
2.2348
7.8081
0.0000
0.0000
0.0000
0.0000
 72 
A P P E N D I X  B
Frame Tube Reduction model of the building:
Data:
Number of bays = 11 nos.
Width of the bay = 12 ft
Building dimension in plan = 132ft X 132ft
Number story = 8 nos.
Story height = 60 ft
Total height = 480 ft
Column modulus of elasticity (E ) = 3150 ksi, 453600 ksf
Spandrel Beam (E) = 3150 ksi
Wind load = 87.5Kips/ node
Floor load = 218.75 kips / node
Strength of concrete = 3000 ksi
Frame tube building reduction modeling technique
Lumped girder Method:
In this case reduction modeling is done to reduce the structure
from 40 story to 8 story for ease of analysis and data
manipulation. The wind load is 87.5kips/ node and the floor
load is 218.75.
120 STORIES,
For columns having Ic = 182250in
4
and area Ac= 1080in
2
and
spandrel beam having Ig= 182250in
4
And area Ag= 1080 in
2
the equivalent inertias are
 73 
∑ ∑
∑ ∑
· · ·
·
,
_
¸
¸
·
,
_
¸
¸
·
n
n
in Ig Ige
in
Ic
n
Ice
1
4
5
1
4
5
1
3
1
3
911250 182250
4556250
182250
1
5
1
2030 STORIES,
For columns having Ic= 129654in
4
and area Ac= 882in
2
and
spandrel beam having Ig= 129654in
4
And area Ag= 882in
2
the
equivalent inertias are
∑ ∑
∑ ∑
· · ·
·
,
_
¸
¸
·
,
_
¸
¸
·
n
n
in Ig Ige
in
Ic
n
Ice
1
4
5
1
4
5
1
3
1
3
648270 129654
3241350
129654
1
5
1
3040 STORIES,
For columns having Ic= 111132in
4
and area Ac= 756in
2
and
spandrel beam having Ig= 111132in
4
And area Ag= 756in
2
the
equivalent inertias are
 74 
∑ ∑
∑ ∑
· · ·
·
,
_
¸
¸
·
,
_
¸
¸
·
n
n
in Ig Ige
in
Ic
n
Ice
1
4
5
1
4
5
1
3
1
3
555660 111132
2778300
111132
1
5
1
Reduction model section properties,
Stories Ige (in
4
) Ice (in
4
) Ag (in
2
) Ac (in
2
)
120 14 911250 4556250 1080 1080
2030 46 648270 3241350 882 882
3040 68 555660 2778300 756 756
Reduction model section properties,
Stories Ige (ft
4
) Ice (ft
4
) Ag (ft
2
) Ac (ft
2
)
120 14 43.9 219.7 7.5 7.5
2030 46 31.26 156.31 6.125 6.125
3040 68 26.79 133.98 5.25 756
The results of the reduction analysis are presented in tabulated
form in end. The results of the program are compared with the
Mastan results, and the comparative charts are presented in the
succeeding pages to follows.
The plots of story height vs. displacement clearly show that the
displacement at the top of the building is 0.2751 ft or 3.30in
 75 
and the result from the program and Mastan is 0.275ft or
3.30in.
Plot 1
The plot of rotation Vs height of building is presented below.
The plot of rotation gives the cumulative rotation along the
height of the building. The plot below shows that the moment
at the bottom is high and gradually at the top becomes zero,
which is indicated by high values of rotation at the bottom and
very low increase in the values of rotations in the middle
portion and no rotation at the top portion of the building. The
results match with Mastan results.
 76 
Plot 4
Second order analysis:
A simple step method is employed for the analysis. Max
iteration specified is 10, greater the iteration greater will be the
accuracy. Load increment is 0.1*200 (the reference load).The
plot indicates a slight nonlinearity with max displacement being
0.32 ft or 3.84 inches. Mastan program gives the deflection
value to be 0.28ft or3.36.
 77 
Mastan2 model for 8 story building:
Input for Structural Analysis
(i) Number of Nodes = 108
(ii) Number of Elements = 184
(iii) Number of Sections = 6
(iv) Number of Materials = 1
(v) Number of Supports = 12
(vi) Applied Loads
lateral load= 87.5 Kips/ node
Floor Load= 218.75 Kips/ node
The Mastan model diagram is given below and
comparative solution is also presented.
 78 
Comparative results:
Program Mastan Program
X
Displace
ment
(ft)
Y
Displace
ment
(ft)
θ
Rotation
X
Displace
ment
(ft)
Y
Displace
ment
(ft)
θ
Rotation
I order
Analysis
0.275 0.1815 0.00054 0.275 0.1815 0.00054
II Order
Analysis
0.320 0.00054 0.280 0.1822 0.00055
 79 
Mastan Plot of Displacement Vs Load Ratio
Actually structureFrame tube building 40 stories:
Data:
L
o
a
d
R
a
t
i
o
Displacement
 80 
Number of bays = 11 nos
Width of the bay = 12 ft
Building dimension in plan = 132ft X 132ft
Story height = 12 ft
Total height = 480 ft
Column modulus of elasticity (E ) = 3150 Ksi, 453600 ksf
Spandrel Beam (E) = 3150 Ksi
Wind load = 17.4 Kips/ node
Floor load = 43.5 kips / column
Floor Column dimension
(in)
Spandrel beam
(in)
Concrete
(Ksi)
B D H B D H F’
c
1 19 24 45 144 24 45 144 3000
2029 21 42 144 21 42 144 3000
3040 18 42 144 18 36 144 3000
Floor Column dimension (in) Spandrel Beam (in) Concrete
B D H B D H F’c(ksi)
119 24 45 144 24 45 144 3000
2029 21 42 144 21 42 144 3000
3040 18 42 144 18 42 144 3000
Flexural strength:
Columns and spandrel beam Moment of inertia and Area
Story Strong axis I (in
4
) Weak Axis I (in
4
) Area (in
2
)
119 182248.7 51840 1008
2029 129662.21 332410.368 882
3040 111186.43 20404.224 756
 81 
The results of the reduction analysis are presented in tabulated
form in end. The results of the program are compared with the
Mastan results, and the comparative charts are presented in the
succeeding pages to follows.
The plots of story height vs. displacement clearly show that the
displacement at the top of the building is 2.89348in and the
result from the program and Mastan is exactly the same up to
4
t h
decimal place. The drift factor of the buildings is
0.0005017.The drift factor for 40 stories
Plot of Drift factor
0
0.0002
0.0004
0.0006
1 6 11 16 21 26 31 36 41
story number
d
r
i
f
t
f
a
c
t
o
r
Drift factor
 82 
The plot of Vertical displacement Vs distance along the
perimeter is presented below. The vertical displacement of the
columns along the perimeter in the plan is considered. The
results obtained from Mastan are compared with the program
and plot shows that the results match with one another.
 83 
The plot below is plot of rotation along the story height. The
rotations are in radians and for complete rotation of all the
nodal points refer the out put file.
 84 
Second order analysis:
Results from second order analysis compared are presented in
the form of a graph of Load ratio versus displacement. The
graph indicates that the deflection is around 2.95. The linear
nature of the curve and the proximity of the result close to
2.89 obtained from Ist order analysis clearly indicates that there
is not much of a second order effect taking place. The
structure there is stable. The curve obtained from Mastan2 is
also presented and from the graph and it is clear that the
deflection is 3.8 in. The drift factor of the building is
 85 
0.0005121. The limit value for drift factor is 0.002.The
buckling load factor for the building calculated using the
program is 61.6. The elastic critical load ratio from Mastan is
47.
Mastan Model 40 story building:
General Information Categories:
(i) Number of Nodes = 492
(ii) Number of Elements = 920
 86 
(iii) Number of Sections = 6
(iv) Number of Materials = 1
(v) Number of Supports = 12
(vi) Applied Loads
Lateral load= 17.5 kips/ node
Floor load= 43.5 Kips/ node
 87 
Fig Mastan Load Ratio Vs Displacement
Comparative result: [reduction model Vs actual model]
40 Story Model [Program Vs. Mastan]
Program Mastan Program
X
Displace
ment
Y
Displace
ment
θ
Rotation
X
Displace
ment
Y
Displace
ment
θ
Rotation
I order
Analysis
2.8934 1.93430 0.00041 3.792 2.2752 0.00041
II
Order
Analysis
2.95 0.00041 3.8676 2.2836 0.00054
Reduction Model [Program Vs. Mastan]
Program Mastan Program
X
Displace
ment
(in)
Y
Displace
ment
(in)
θ
Rotation
X
Displace
ment
(in)
Y
Displace
ment
(in)
θ
Rotation
I order
Analysis
3.30 2.18 0.00054 3.30 2.18 0.00054
II Order
Analysis
3.84 0.00054 3.36 2.19 0.00055
Displacement in X
L
o
a
d
R
a
t
i
o
Worcester Polytechnic Institute Abstract STABILITY ANALYSIS OF FRAME TUBE TALL BUILDINGS by Amit Urs Thesis Advisor: Professor P. Jayachandran Department of Civil Engineering
The frame tube buildings have been the most efficient structural system used for building which is in the range of 40100storey. The soaring heights and the demanding structural efficiency have led to them having smaller reserves of stiffness and consequently stability. In this thesis a Nonlinear analysis and stability check of frametube building is done. Nonlinear analysis offers several options for addressing problems of nonlinearity and in this work focus is on Geometric Nonlinearity. The main sources can be identified as P∆ effect of gravity loading acting on a transversely displaced structure d to lateral loading and can also be due to member ue imperfections, such as member camber and out of plumb erection of the frame. During analysis the element response keep continuously changing as a function of the applied load so simple step computing methods have been employed instead of direct analytical methods. The problem here is dealt in a piece wise linear way and solved. In this thesis a program using the matrix approach has been developed. The program developed can calculate the buckling load and can do Linear and Nonlinear analysis using the Matlab as the computing platform.
i
Numerical results obtained from the program have been compared with the Finite Element software Mastan2. The comparative solutions presented later on in the report clearly prove the accuracy of the program and go on to show, how exploiting simple matrix equation can help solve the most complex structures in fraction of seconds. The program is modular in structure. It provides opportunity for user to make minor manipulation or can append his own module to make it work for his specific needs and will get reliable results.
ii
Aunt Sudha. Aunt Meera. Kaivalya V. author wishes to thank his mom Mrs. In addition. The author also would like to thank Prof. for all the blessings and support they have provided him. Badri Krishna for all the help and valuable assistance. William. Author wishes to express his humble gratitude to “PARAMATMA” without whose mercy he could not have done anything. support and patience. Author wishes to thank his friend Mr. Jayachandran for his thesis advising. W. Urs.ACKNOWLEDGMENTS The author wishes to express sincere appreciation to Professors P. Durgin for the financial assistance and encouragement. Geeta Urs and his Dad Mr. .K. his sister Ms. Mainly. Venkatakrishna Urs. Author would like to thank the members of Weidlinger Associates for their trust.
..............................................................................................................................................................................................................................................................................................................................38 Structural Analysis ...............................................................................................................................3 Previous Work ...............27 Approaches to analysis......................vi Preface .......................................................21 Chapter III: Tall buildings ....................55 iv ...................37 Mode of behavior ................................................................................................51 Recommendations for future research...........................2 Scope .................................................4 Chapter II: Conceptual Framework .............................................................................51 Description of Findings .............................8 Stiffness method...........................................31 Reduction modeling......................................................32 Frame tube building....44 Chapter IV: Findings and Discussion ........................................................1 Objective...........................................................................................................................................................................8 Level of analysis ...............27 Introduction...................................................................................TABLE OF CONTENTS Abstract ...................1 Introduction...................................i Acknowledgement.................................................................................................................41 Reduction modeling of Frame tube building...................................................21 Load control method .....................................................................iii Lsit of figures...........................................vii Chapter I: Introduction..................................................................................................................................................................................................................................................................................................................................................21 Different approach to nonlinear analysis ...............8 Matrix methods .......................................................................................................................................................................................................................................
.................................................................................................................68 Appendix B: ........................................72 Frame tube building 40 Stories.......61 Algorithm............................................................................................................58 Appendix A: ................................................62 Tutorial 1.........................................................................................................................61 Manual ............................................................57 References....................................................................................................................88 v ..............................................................63 Tutorial2..........................................................................................................................................................72 Frame tube building reduction models ...............................................Glossary ......................................................
Finite strain of planar element .......... Eulers column .........34 vi ..................................45 7............................................................ Prototype rigid frame.................................................................................................44 6........................................ Axial stress distribution in column............................................... deformation of flange frame..........59 11............................10 2..................................... Side sway prevented........ Nodal forces and displacement............................................................................33 13....................... Load Deflection curve .. Side sway permitted ..................................................11 3........................................ Nodal forces................19 5.. connection details...............51 8.......................................................................................................................................58 10..................... Frame Tube structural plan...........................................................................52 9...................................................33 12..16 4...LIST OF FIGURES Number Page 1........................ Equivalent lumped girder................................
In chapter 4. As the work is pivoted on the matrix method of structural analysis. vii . a conceptual frame work of the thesis has been laid out. Methods used for mathematical modeling along with logics involved in the structural behavior of frame tube building have been discussed. In this work a frame tube building of 40 stories high. an introduction to work is given followed by clearly laid out objectives of the study and its scope. Linear elastic analysis. In Chapter 2. In Chapter 3. In this work an effort has been made to develop programs that will do an analysis of highly indeterminate structures. Finite Element Method and Matrix method are the two methods which show great compatibility for computing process and has become the most power tool in many engineering branches. an introduction to tall buildings followed by a detailed discussion on structural action of frame tube buildings has been presented.PREFACE The high speed precise computing and increased memory of the computers has made it possible to solve complex models. There analysis is extremely complex due to highly degree of indeterminacy and hand calculation is very tough. which form the basis for the programs developed. nonlinear analysis and stability analysis have been discussed appropriate enough.100 stories high. Tall buildings like frame tube buildings are usually 40 . has been selected. results and conclusion of the work have been presented. which is highly indeterminate. different techniques and approaches to analysis of planar frames have been presented. In chapter 1.
Manual. results and comparative values have been tabulated. Tutorial problem have been presented and theses problem should aid for a proper understanding of the programs.In Apendix A and Apendix B. viii .
While some opine that "nothing is designed or will be designed to withstand that [kind of] fire. But combination of these two appears to be logical explanation to the tragedy on 9/11. Bottom line is one or combination of them may have caused the structure to fail. Some believe that nothing can. what we can do in the future to prevent another failure like this one.Chapter 1 INTRODUCTION Introduction: In the past. but now that this threat is a reality. designers have to look for sensible ways to defend our highrises All the evidence so far points that a combination of the impact and the fires triggered the collapse of the World Trade Center buildings. as high as 2000 degrees Fahrenheit. or should. designers had no need to accommodate for terror attacks from the air. some believe the impact alone was the actual cause of the failure by providing evidence that the structure failed quicker than the time required for the fireproofed steel to fail while some argue that the fires where to blame. we must now ask. .” Numbers of theories have evolved to explain exact reason for the collapse of the Twin towers. from now on it would be necessary to do a thorough stability check of tall buildings for all possible circumstances. the steeltube construction of the towers could not resist the intensity of the fires resulting from the energy released by the 2400 gallons jet fuel that brought the fires to higher than normal temperatures. be done.
Matrix approach is used in the analysis. leaving the structure with minimum reserve energy. rendering stability to the structure as a whole. deflections are small with insignificant change in geometry of the structure. second order analysis and also to calculate the elastic critical buckling loads of planar framed structures. Thus failing to satisfy the equilibrium conditions the structure could become highly unstable. Objectives The main objective of the thesis work is to develop software to check the stability of structures and also to do a second order analysis of the 40 story FrameTube Buildings. buildings are designed for limit state of strength and limit state of serviceability. 3. The large deformation and low reserve energy can prove to be cataclysmic if this small energy in the structure fails to resist loads.In “Limit state design”. To predict the exact behavior of the structure a second order analysis of the structure has to be conducted. It is realized that the fulfillment of the following subobjectives would in turn fulfill the main objective. Thus the reserve energy of the building is sufficient to bring back the structure to equilibrium state after the load is removed. To study the behavior of Frame tube buildings. In case of tall buildings about 40 to 50 story’s high horizontal loads cause huge deflection of the building and a significant change in the geometry of the structure. In this thesis I have devolved a program to do first order. 2 . There by first order analysis of structures to satisfy the equilibrium conditions is sufficient to verify the design. To study various technique in nonlinear analysis of frametube buildings. 1. To understand the concepts of Matrix analysis of structure. If it is the design case of a lowrise structure subjected to low horizontal loads. 2.
The behavior of the structure is elastic. To develop modular structured program versatile and should be able to analyze a wide variety of problem. Connections are considered either fixed or hinged and it does not account for partial fixity. 5.4. If the structure contains shear wall or cores it has to be idealized as a single isolated column. which are considered to be rigid in their plane. 4. Run the program for Non Linear analysis of Frame Tube building and verify its structural stability as a whole and to compare these results with commercial software called Mastan2 for accuracy. Scope: Focus of the work is on planar frames with the following characteristics. During second order analysis geometric nonlinearity due to large member deformation is the main focus and material nonlinearity due to change in material property is not considered. 6. 3. Different frames may interact through common columns and through the floor diaphragms. 2. 3 . 5. 1. In frame element for example all member are considered beam column elements. Static gravity loads and horizontal wind loads are applied as nodal forces. To develop a Mat Lab program using the principles of Matrix/Finite Element approach for both Linear and Nonlinear analysis. Structure may behave like a truss or frame depending on the element node connection. The structures are considered to be an assembly of planar elements.
selection of structural form is important. a simplified approach to the analysis of frame tube buildings is made. A number of papers have been published for the analysis and design of tall buildings. each paper helps to understand the logic behind the analysis of the frame tube building. In case of tall buildings stability being the main criterion. various advanced techniques like the finite element method have evolved in tantamount. Frame tube buildings have been considered to be efficient structural form for building ranging from 40 stories to 100 stories height. Refs (19) have attempted to analyze the frame tube building as accurately as possible with the limited computing capability that computers of those times had possessed. For second order analysis simple step method has been employed Previous work: Refs (10 & 12) have introduced the topic of highrise structure building by introducing the different structural forms. In Ref (1). computers have evolved so has the complexity of the problem.e. with the main intention of explaining the different approximate methods for analysis. The material property are represented by area (A). Each structural member is idealized. Importantly. Rigidly – jointed frame action of the shearresisting panels parallel to the direction of the loads and 4 . Through time. young’s modulus of elasticity (E). i. the width and elemental length are conventionally reduced to line elements. torsion constant (J) and poisons ratio (ν).The program is based on Matrix analysis of structure. The dominant modes of behaviors of the frame tube buildings resting lateral loads are recognized as 1. is the depth.
which is the 2nd of the 3 paper series published by the authors. problem of torsion is tackled by replacing the discrete structure by equivalent orthotropic tube and then a nalysis it as a plate. In next part ref (3). Thus consequently reducing the amount of computation required in a conventional threedimensional analysis to a simple 2Dimensional planar frame analysis. to form a closed tube structure. uniform cross section area thought out the height. they have proved that it is possible to simulate the behavior of the structure using a reduced equivalent plane frame. The orthotropic plate is chosen that the both the modulus of elasticity of the plate in both the horizontal and vertical direction be represent the axial stiffness of the beam and the column respectively. it is assumed that the shear rigidity of the structure to be low so as to simulate the shear racking 5 . The axial deformation of the frame panels normal to the direction of the load. Here. In Ref (2) author has taken consideration of the shear lag effects and has tried to predict the possible response of the building so as to include this phenomenon. The transfer of vertical shear has been achieved by giving large values to appropriate shear transfer elements in the stiffness matrix. the assumption of a parabolic stress distribution to simulate the shear lag effect is surely one of the limitations. Once the modification to stiffness matrix is made then the process of analysis is as usual. After which. The structure then is analyzed as a plate. Shear modulus representing the shear stiffness of the framework. The author accounts for the shear lag effect by assuming a parabolic stress distribution. He further goes on to simplify the problem by assuming that the spacing of columns and beams are uniform and they have all.2. The framework panel of the column and beam are replaced by equivalent orthotropic plate. The method of plate analysis is fairly accurate and provides design curves for standard load cases for easy and quick reference.
He assumes the building to be square in plan and all columns other than the corner columns to be of same cross section. Ref (11) together with introducing the topic of tall buildings has also dealt with important structural forms in detail. Reference (12) 6 . Papers by Fazlur Khan and Amin papers regarding approximate analysis of framed tube buildings provide graphs for a quick and rapid analysis and design of frame tube building. In Ref (9) authors has submitted his work as a partial fulfillment of his PhD . one for static and dynamic analysis and the other software for the optimization of frame tube building. In Ref (4) third and the final part authors tackles the deflection in frame tube buildings. Ref (18) has used his software for his MS dissertation on the analysis of frame tube buildings. He has also shown the various reduction techniques that can be used for transforming the structure so as to quickly verify the solutions.the technique of structural analysis and design. The corner column is assumed to have a twice the cross section area. which the ordinary elastic modulus would not. His program does not do the stability check.In his dissertation author has presented a Macro element method to analyze the frame tube building. one pair at a time and he establishes the actual interaction forces based on the compatibity condition. The columns are connected by equal cross section spandrel beam. In the process he has also developed two programs. For the analysis he converts the structure into a planar model and for the corner column interaction he introduces a unit force at each floor level along lines column lines( for each frame he takes half of the corner column cross section area). 60 story building.effect. 50. In ref (8) author has devolved a computer program and has provided a number of table to assist the analysis of frame tube buildings. The assumption made for the hand method is its limitations. In his report he has presented tables for 40.
He also has developed software using the matrix analysis principle for solving structures.covers the topic on the Matrix analysis. which shall be used as reference software and will be compared with the software being developed in this thesis work. importance is laid for the solution of planar framed structure. 7 .
(G) shear rigidity. Every element in the structure is idealized as a line element with section properties (A) area. Thus these simultaneous equations can be solved using matrix methods. By using these solutions. In matrix method there are generally two approaches the flexibility and the stiffness approach and the former approach is used in this thesis work Stiffness Approach: In matrix method there are generally two approaches the flexibility and the stiffness approach and the later approach is used in this thesis work. (NU) poisons ratio. The whole planar 8 . Once the equations are written in Matrix form taking advantages of the power computing capabilities of the computers the large system of equations can be solved quickly.Chapter 2 CONCEPTUAL FRAMEWORK Matrix Methods: Behaviors of all types of structures are described by differential equations. it is common to represent frames as planar structures for which the approximate or exact solution of the individual members are available. In practice. relationship between force and displacement in the end of the members in combination with equilibrium and compatibility equations at the joints and supports can be written which yields a system of algebraic equations that describe the behavior of the structure. (I) moment of Inertia and (J) torsion constant and material properties as (E) modulus of elasticity.
formed by individual finite elements connected together at joints called nodal points. The procedure involved in the formation of member stiffness matrix is quite simple can be summarized in 3 steps 1. For illustration consider a 1D spring element which has 2 DOF. Thus relationship between nodal forces and corresponding nodal displacement is written in terms of stiffness equation or flexibility equations. The orthogonal axis with Xaxis along the member longitudinal axis refers to local coordinate axis.structure is considered as mesh.Fj) to 2 nodal displacements through 2X2 stiffness matrix. Each element consists of 2 nodes and each node is associated with corresponding degree of freedom or nodal degree of freedom. Compatibility 3. Hooke’s law. The frame of reference to identify entire structure is called global coordinate system referred to as (XY axis). 2. the element force–displacement relationship relates 2 nodal forces (Fi. 9 . The term Degree of freedom (DOF) is defined as “ the number of independent coordinate necessary to define the configuration of the system” In space Nodal DOF are 12 in number but in the case of framed structure with rigid joints 3 DOF are considered. Equilibrium. Forces are applied to the nodes as point load and the corresponding displacement is measured at nodal points.
we get FI K11 K12 UI FJ = K 21 K 22 UJ Thus k11.10  . {f } = [k]{∆} . the above equation.F i ( υ i) i K Fj (υ j ) J fig 1 {fe}= [k]{ue} Expanding. force (F) Vs displacement (∆) is a linear curve. we get F I = k11 uI + k12 uJ F J = k21 uI + k22 uJ In matrix form. k 11. k 12. k 12 are to be defined Apply springs law. it can be written as.
Using the compatibility conditions. We need to find force (F ) in terms of end displacements uj and ui. ∆=ujui f I = k (uJ . f I = f J = (k uI + k uJ ) In matrix form can be written as.11  . fI + fJ = 0 Thus.uI) Now applying the equilibrium equations ΣfX=0. fI K f J = − K − K K uI uJ .2 Nodal forces.Fi i K J Fj Fig.
has to be multiplied with a . thus for a beam element stiffness matrix will be of the dimension [6X6]. the fundamental stiffness equations is of the form {F}=[k]{∆} Where.12  .r. In case of an axially loaded member. In our program focus is on beamcolumn and truss. So the element stiffness matrix will have 3 Dof at each node. A L FX 1 0 FY 1 0 M Z1 =E − A FX 2 FY 2 L 0 M Z 2 0 0 12 IZ L3 6 IZ L2 0 − 12 IZ L3 6 IZ L2 0 6I Z L2 4I Z L 0 − 6 IZ L2 2I Z L −A L 0 0 A L 0 0 0 − 12 IZ L3 − 6I Z L2 0 12 IZ L3 − 6I Z L2 0 6IZ L2 2I Z L 0 − 6 IZ L2 4I Z L u1 v1 θ 1 u 2 v2 θ 2 If beam or truss is inclined at any angle θ w.Thus this is the member stiffness matrix.t to horizontal the element stiffness matrix. if unit displacement is applied keeping all the other degrees of freedom fixed then the force applied will be equal to the stiffness of the member or in other words. This is formulated depending upon the material and sectional properties of the element. For information about the formulation of the complete stiffness matrix please refer ref (13). which is in the local coordinate system. nodal forces and nodal displacements are connected with each other by element stiffness matrix. [K] is the element stiffness matrix.
In structures like suspension bridges and very tall buildings that are slender. Thus even though most engineers are fluent in first order analysis it is important to conduct an Nonlinear analysis which will help analytically simulate . then they are assembled to form the combined global stiffness matrix [K]. there by do not exhibit significant nonlinearity. before these structures reach limit of resistance they exhibit significant nonlinear behavior.13  . Thus the transformation matrix is cos θ − sin θ 0 A= 0 0 0 sin θ cos θ 0 0 0 0 0 0 0 0 1 0 0 cos θ 0 − sin θ 0 0 0 0 0 sin θ cos θ 0 0 0 0 0 0 1 Where θ is the angle made between the element and the horizontal. The matrix operation involved is [K]global=[A] [k]local [A]T Once the member stiffness matrix is formulated in Global coordinate sytem. Nonlinear analysis: Most of the civil engineering structures fall in the category of lowrise structure and the deflection are very small. .transformation matrix to convert it into global coordinates.
Gallagher and R. According to W. 2. R. on the initial reference of the configuration. Geometric nonlinearity alone can be considered making an assumption that the material is elastic including the effect of finite displacement and deformation during the formulation of equilibrium equation. destabilizing moment equal to gravity load times the horizontal displacement it undergoes as a result of lateral displacement. as seen before. In linear elastic structures.Mcguire.D. The P∆ effect. Geometric effects: 1.more appropriate and better behavior of the structure than the first order analysis.H. Thus it is the responsibility of the engineer to use his judgment and discretion while predicting the most dominant behavior of the structure he wishes to simulate analytically. Nonlinear analysis offers many options for solving the problems resulting from the above cases. In Nonlinear analysis the structure may lose its geometry due to large deflections and the material can also exhibit nonlinearity. the influence of the axial on the flexural stiffness of the individual member. 3. The Pδ effect.14  . This will ultimately help in the optimal design of the structures. it was assumed that the material is unyielding and its properties invariable and the equations of equilibrium are formulated on the geometry of the unloaded structures or. . It is also possible that material nonlinearity alone can be considered taking into account that the member property changes under loads.Ziemian in their textbook sources of nonlinearity are as follows. Initial imperfections such as member camber and out of plumb erection of a frame. Finally material nonlinearity and geometric nonlinearity can be considered together in the analysis.
Contribution of secondary system to strength and stiffness. Combined effects: 1. Connection deformation 3. bending. Levels of analysis The fig below shows loaddeflection behavior of plane frame building analyzed using various refined and simplified models. The broken curved line is that of second order elastic analysis that includes only the effect of change in geometry and instability of the structure and this is the focus of thesis work.Material effect : 1. 2. 3. Its value is plastic limit load.15  . The piecewise linear curve with a descending branch shows the result of a secondorder elasticplastic hinge analysis allowing for simple plastic analysis for . which can be obtained by plastic analysis. shear and torsion. 2. Cracking or creep of reinforced concrete structure. plastic deformation of the steel structures. The piecewise linear curve without a descending branch represents a first order elastic–plastic hinge analysis when the effect of change in geometry has been ignored. The curve approaches asymptotically to the elastic critical limit generated by an Eigen value analysis. The broken line is obtained by Ist order analysis ignoring the effects of both change in geometry and the yielding of the material. Panel deformation 4. Plastic deformation plus P∆ effect and/ or Pδ effects. Inelastic interaction of axial force.
16  .Plastic Second. The critical load is calculated by conducting a simple eigen value analysis. residual stresses and strain hardening are all accounted for.the change in the geometry associated with the sway deflection. the smooth continuous curve is obtained by doing an secondorder spread plasticity analysis. The plastification.Elastic .Order SpreadofPlasticity (plasticzone Theory) Charecteristic Deformation Fig 3 Load Deflection curve for a plane frame In our thesis work the focus is laid on second order elastic analysis. The peak load is the true strength of the frame or the true load carrying capacity of the structure.Order .Mechanism Load First . First order Elastic Elastic Critical Load Second Order elastic Load intensity Plastic . The whole procedure has been converted in to a program which can be used to solve problem.Order Elastic Plastic Second . initial imperfections. taking into consideration only the geometric nonlinearity ignoring the material nonlinearity. Nonlinear analysis: .
which is written in matrix form. Many different ways are there to develop the [Kg] matrix. For the calculation of the elastic critical loads the global stiffness equations is cast in the form of a generalized eigen value problem in which the equations at the critical state is .17  . Many techniques for solving the equations are given later on in this report. the way the equations are formulated and the details of the equation solution. Focusing on elastic IIndorder analysis. dividing the members into sub elements yields very accurate results. Thus each method is a variation of global stiffness matrix and symbolically written as [Kt] {d∆} = {dP} [Kt] = Tangent stiffness matrix. different methods can be devised to solve there by displacement and deformations of the members can be traced incrementally. The various levels differ in type of nonlinearity included in the types of nonlinearity included [Kt] . {d∆}= incremental displacement. {dP}= incremental loading. and the way members are subdivided. Finite displacement and deformations are accounted for in the equations of equilibrium and can be written as [Ke + Kg] {d∆} = {dP} [kg] is the geometric stiffness matrix. Once simultaneous equations are written in matrix form.In IInd order analysis nonlinear equilibrium equations are reduced to a set of simultaneous equations.
But after rigid body rotation and axial straining its length is a ' b ' = 1 + 2 du du dv 2 + + dx dx dx dx 2 2 1 du Let’s designate.t {Pref}. Let’s consider the segment of length dx as shown in the Fig4.^ e + λ kg {d∆} = {dp} k Where K ^ g is called the geometric stiffness matrix computed for a reference load {Pref}. Their initial configuration is designated as ab. dab = 2 + dx 1 a' b' = (1 + dab ) 2 dx du + dx dv dx . and {∆} the eigen vector is the buckled shape. but we shall derive the Kg matrix basic axial force member. {λ} (the eigen value) is the load w.r. The lowest value λ that satisfies the equation of {∆} not equal to 0 yields the Elastic buckling load. Formulation of Kg: As any standard text book Matrix analysis of structure would indicate the technique involved in the derivation of Kg matrix.18  .
A. finite strain. (a’b’ – ab)/dx .19  .Iz L y x+u+dx+(du/dx)dx x+u b' b' 2' 2 Mz2 Fy2 X a' (v+(dv/dx)dx) v1 u1 X 1' a' a dx b dx u2 x Fig 4 Finite strains of planar element or expanding a’b’ by binomial theorem and defining the extension per unit length. e fin 2 2 du 1 du dv + dx 2 dx dx efin = Applying the principles of virtual displacement to the reference configuration δW int = vol ∫σ xδ e fin d ( vol ) Integrating over the depth of the member. we get .Fy1 Fy1 Fy2 Mz1 Fy1 1 E.
{N’ u } and {N’ v} are the shape functions.r.20  .strain relationship in the first integral and letting σx A= Fx2 we get L L du dδu 1 δW int = ∫ E A dx + Fx2 ∫ 0 dx dx 2 0 2 du 2 dv δ + δ dx dx dx The first term in the equation above is for elastic stiffness matrix of the axial force element and the second integral term is of great concern as it produces geometric stiffness matrix [k g]. Now using dx dx the conventional elastic stress. Thus for axially loaded members displacement equation in the form of shape function can be written as u= (1ξ)u1 +ξu2 and v= (1ξ)v 1+ξv 1 .2 2 L L σ A dδu dx + 1 σ Aδ du + δ dv dx δW int = ∫ x x 0 2∫ dx dx dx 0 dδ u du let = δ which is valid for infinite small displacement.Treating δ as differential operator w.t variables du/dx and dv/dx. the internal virtual work can be written as L Fx2 dδu du + dδv dv dx δW int = ∫ dx dx dx dx 0 Using the shape function the above equation can be written as L [Kg] = Fx2 ∫ [{N ' u }N ' u + {N ' v }N ' v ]dx 0 Where. in this case for axially loaded members.
N ' u = − 1 L 1 −1 and N ' v = L L 1 L Combining the equation suitably we get 0 −1 0 1 0 1 0 − 1 Fx 2 [Kg ] = L − 1 0 1 0 0 −1 0 1 Combining the bending and the axial forces the geometric stiffness matrix will be.Therefore. series of load increments are made to carry out the analysis. As equilibrium configuration of the structure changes constantly. 0 1 6 0 5 L − 1 10 [Kg ] = Fx2 0 L 0 −6 0 5 L 0 10 −1 L 10 2 L2 15 0 −L 10 − L2 30 0 0 0 1 0 0 0 −6 5 −L 10 0 6 5 −L 10 0 L 10 − L2 30 0 −L 10 2 L2 15 Different approaches to nonlinear analysis: Basic principle involved in nonlinear analysis is to perform a set of matrixstructural operations to update the stiffness matrix constantly so that all nonlinear effects of the structure are accounted.21  . The equilibrium and kinematics states of the structure from previous cycle are used to formulate the stiffness relation for the solution of the next load .
Displacement control method 3. the number of iteration required to obtain a solution for a given load step may be large and in some case the solution may even diverge. The size of the load increment has a profound influence on the solution time and the convergence characteristics of the problem.increment cycle. In applying these methods to trace the Nonlinear loaddeflection curve of the structure. Work control method. 2. it is assumed that the equilibrium path to be followed is continuous and unique.22  . they are 1. the solution of nonlinear analysis problem is got by a series of linearized analysis. It is assumed that the equilibrium and kinematics states of the structure at the (I1)th load step is know and it is now required to find the state of the structure or the stiffness relation at ith interval. On the other hand if the increment is too large. If a noniterative scheme is followed the use of a large load increment may result in a large drift off from the actual equilibrium path. Relation can be written as Kij1 ∆Dij = ∆ Rij + Qij . Thus. Arc control method 4. Load control method. Thus nonlinear analysis is iterative or can be non iterative There are many techniques and four major solution schemes for nonlinear problems are studied. Smaller load steps yield accurate solution and in turn will take larger number of load steps to cover the loads range.
the load at the end of the ith load step is given by Ri= Ri1+Σ∆ Rij And the displacement is given by Di = Di1 + Σ ∆Dij The updated stiffness matrix can then be built and the solution can move to next increment. ∆Ri. The solution schemes differ by size of the ∆R is determined and the way iteration is carried out inside a load step. ∆Dij= incremental displacement vector at the jth iteration load step i. ∆ Rij = load increment at the jth iteration of load step I Qij= force vector at the jth Iteration of the load step i.Where Kij1= stiffness matrix for the load step I based on the equilibrium and kinematics states at the end of the (j1)th load step are known.23  . Load control method: Here the size of the incremental load at load step I.( difference between the external and internal forces) Suppose the solution converges after n iterations. is obtained as a fraction of the total applied load .
Simple incremental method: In this method the unbalanced forces that exist in each load step is ignored. In this thesis work to calculate the nonlinear response of the frame tube building simple incremental load method has been employed. Can be rewritten as Ki1 ∆Di = λiR From which ∆Di can be calculated. ∆Di calculated is added to the cumulative sum ∆Di1 from the previous cycles to get the new displacement Di Di = ∆Di1 + ∆Di Thus updated stiffness matrix can be calculated. Noniterative method is referred to as simple step method and Iterative method referred to as Load control NewtonRaphson method.24  . The user determines the value of the λi by using personal judgment and discretion. reduces to Ki1 ∆Di = ∆Ri Here only one iteration not done. That means no iteration is done. λi is the load increment factor and R is the total applied load. Thus stiffness relation that exits at load step I.∆Ri = λi Ri In the above equation. . The load control method can be iterative or noniterative. so the value of j=1.
25  . If this error is not corrected then the calculated equilibrium path will drift off from the true equilibrium path. This is because iterations are not employed within a load step to bring the calculated equilibrium path back to true equilibrium path. Its unloading branch of the loaddeflection can be investigated by giving negative load increments. Unbalanced forces are calculated as the difference between internal and external applied load. The main disadvantage of this method is that the error tends to accumulate and may prove to be disastrous if the structure is highly nonlinear in nature. and this errors occurs because of linearization process in which current stiffness matrix is calculated based on the passed configuration the structure. Corrections are applied to the displacement by subjecting the structure to the unbalanced forces. equation is Kij1 ∆Dij = Qij . Load control Newton Raphsons method: In this method iteration employed within the load cycle eliminate the unbalanced force. The presence of unbalanced force indicate that the internal and the external forces are not in equilibrium. Now the stiffness matrix can be formed. Therefore subsequent iteration. Lets take ith load increment for the first iteration (j=1) K0i ∆D1i = ∆R1i K0i ∆D1i = λ1iR Where ∆D1i can be calculated from the above equation and added to the cumulated displacement of the structure.Advantage of using simple incremental method is its simplicity. This is called drift off error.
The limitation of this method is that it fails at limit points. The advantage of this method is that the drift off error is greatly reduced or controlled.The iteration stops when ∆Dij or Qij is negligible.26  . . This method fails also if the system exhibits snap through behavior.
As a solution. bridges and dams for smooth communication and urban development and many more. The tragic event of 9/11 has made conspicuous . Question your self. may be this is our greatest problem. Industrial revolution put lot of pressure on construction industry to meet it demands. As problems get solved. they keep evolving. Have we to wait for solution to evolve into absolute forms or rhetorical manner wait for the mind to transcend into absolute form to get the solution. As a solution. All this could be realized only by improved materials and structural innovations made during the post industrial revolution. are we really solving anything? Well. Can this be our limitation? Or is it that the solution is already in absolute state and our mind fails to perceive it. in reality we dare not to claim. Whatever the answer may be. in real world there is always a possibility for problems to arise even from well accepted fact. but eventually the great fire of Nero made conspicuous its problems. tall building were built to ease the space jams in urban areas. When problem becomes conspicuous.27  . complete understanding is demanded for it. Rome saw its skyline pushed up by the construction of multistory buildings up to 4 stories high using wood. this is our inspiration for the quest of perfection. arches and barrel domes were constructed of concrete and brick.Chapter 3 STRUCTURAL ANALYSIS OF TALL BUILDINGS Introduction: Is there an absolute solution to a problem? Hypothetically may be “yes”. more and more different problems became conspicuous and solution were demanded immediately.
another short come. material and the type of construction. the nature and the magnitude of horizontal loading and the height and proportioning of the structure. In the case of horizontal loading.28  . In addition to the above factors. Solution for this problem is inevitable. within the selected structural form advantage can be taken by placing main vertical members in plan . dead and live loads and also to resist all levels of external horizontal load shear. It may also comprise of a number of vertical columns and walls all acting compositely which is made possible by using proper shear resistant connectors. the planned location and the routing of service systems. Office building desire column free interiors so structural form selected will be such that a lot of material is placed on the external perimeter of the plan and internally placed around the elevators. A major factor influencing the structural form is the function of the building. Taller the building slender it will be and becomes more dependant on the structural forms. they need extra material to account for resisting the wind loads. There are a number of structural forms that are available for selection. For buildings taller than 10 stories. each bending about its own axis acting unison through the horizontal in plane rigidity of the floor slab. buildings structural forms are arranged such that they support the gravity. which needs to be dealt with immediately. Moreover. such as shear wall or cores. They may comprise of individual members acting in the form of a cantilever. there by each member act as chord of a massive cantilever. it just a matter of time until research will be done on every aspect of tall buildings and solution will evolve. A range of factors influence the selection of structural form they are internal planning. selecting an appropriate and optimal form is responsibility of a structural engineer. the external architectural planning. These necessities should be accomplished as economically as possible. tall buildings behave like a cantilever. stairs and service shaft.
outriggerbraced structures. It is possible to have longer spans. The different structural forms are bracedframe structures. in filledframe structures. partitions and cladding. space structure and hybrid structure. fire resistance. This avoids net tension occurring in vertical members and prevents this uplifting force to rip of the foundation. bundledtube structures. coupled wall frame structures. columns girders. Steel framing has played a prominent role in the development of tall buildings due to its high strength to weight ratio. Tall buildings are usually constructed of concrete. tubeintube structures. suspended structures. Mathematical Modeling of a tall building: A building's response to loading is governed by the components that are stressed as the building deflects. steel or composite. rigidframe structures. Concrete on the other hand with shear walls resisting the horizontal loads and improved concrete strength and innovation in the structural forms have allowed the height of the structures to touch the 100story mark. flatplate and flatslab structures. these include. wallframe structures. however. To simplify the problem it is usual in modeling a building for analysis to include only the main . the staircase. the slabs. the participating components would include only the main structural elements. bracedtube structures. In reality. and requiring expensive diagonal bracing or rigid frame connection. Ideally for the ease of accuracy of structural analysis. expensive to clad. partial prefabrication leading to reduce on site work and more rapid erection. framedtube structure. shear wall structures.so then the dead load of the structure can counter balance some all or a portion of the tensile stresses developed in members due to horizontal load. other nonstructural elements are stressed and contribute to the building's behavior. walls and cores. for example. major disadvantage being corrosion problem.29  . beams.
The following paragraphs to follow shall illustrate the modeling and refining procedure.structural members and to assume that the effects of the nonstructural component are small and conservative. the problem of analyzing a tall building structure gets reduced to viable sizes. Sometimes for approximate initial analysis. There by the mathematical model is gradually improved to obtain more accurate results. however nonlinear models sometimes reveal certain characteristics of the systems that cannot be predicted using linear model. In case of very tall and complex building it is necessary to reduce even further the size of the analysis problem by representing some of the structural elements with simpler analogous components. depending upon the behavior of the component system. and less essential components. it is necessary to recognize the dominant modes of action of the proposed building structure and to access the extent of the various member contributions to them. The mathematical model may be linear or nonlinear. Linear model permit quick solution and are simple to handle. The mathematical model should include enough details to be able to describe the system in terms of the equations without making it too complex. To identify the main structural elements. approaches crude or elementary model are used to get a quick insight into the overall behavior of the system. Thus purpose of structural idealization is to represent all the important features of the buildings for the purpose of deriving mathematical equation’s governing the behavior of the system.30  . . Then by neglecting consideration of the nonstructural components.
the principle modes of action and interaction of major structural elements. may be to compare the performance of the alternative proposals for the structure. result for deflection and member forces.Approaches to analysis: The modeling of the tall building structure for analysis is dependent to some extent on the approach to analysis which is in turn related to the type and the size of the structure and the stage of the design or which the analysis is made . It is important that the solutions after preliminary analysis should yield deflection close to the exact model and main member forces that are dependably within about 15% of the value from an accurate analysis.31  . The formation of the model and the procedure for a preliminary analysis should be rapid and produce results that are dependable approximates. The model and its analysis should therefore represent fairly well. that is analysis for early stage of the design. A hybrid approach is also possible in which a simplified model of the total structure is analyzed first after which the result are used to allow part by part detailed analysis of the structure. as accurately as possible. Intermediate and Final analysis: The requirement of intermediate and the final analysis is that they should give.The usual approach is to conduct approximate rapid analysis in preliminary design. The model . Preliminary Analysis: The purpose of the preliminary analysis. if not absolutely accurately. or to determine the deflection and major member forces in a chosen structure so as to allow it to be properly proportioned. and a more detailed and accurate analysis for the final design stage.
the effect of the structure twisting should be included. All the major modes of action and interaction and as many as possible lesser modes.should. Horizontal lumping The details for the above techniques can be found in ref (text book by Alex Coull). therefore. in deep beams at transition level of the structure around irregularities or holes in the shear wall.32  . Lumping a. . Vertical lumping. In contrast to the reductions above. this also would be acceptable. Symmetry and Antisymmetry. Reduction technique: When detailed modeling consumes lot of memory space and time. more detailed analysis of particular parts. however certain final analysis may require separate. b. be as analysis program and computer capacity will allow for its analysis. There are many ways that a model can be reduced to simpler models. a three dimensional model would be acceptable. using forces or applied displacement from the main analysis for example. 2. If the structure and loading are symmetrical. should be incorporated except where a structure is symmetrical in plan and loading. Twodimensional modeling of nontwisting model. With easy access to fast computer FEM model of the structure must be done as a part of a final analysis. some of them are 1. or if repetitive regions up to the height of the structure can be simplified by a lumping technique.A brief insight into lumped girder frame technique is presented below. 3. reduction modeling can be used for simplicity and the ease of analysis.
which is used as rapid analysis of frame tube building later on. In this method the repetitive floor system are identified and is exploited to form a model with fewer stories, which in turn benefits a rapid analysis. This method allows an accurate estimate of the drift and a good estimate of the member forces. The girders are usually lumped in sets of three girders or five girders into single girders thereby reducing the number of stories considerably. For very accurate analysis it is preferable not to include the first floor and roof girders during the lumping process because the simulation of the frame behavior near the top and the base of the building differ from the middle portion. In the figure below the first floor and the top floor are untouched in order to maintain the boundary condition same while the remaining floors are lumped up. In the building show below story height is h and the column inertia is Ic and that of the girder is Ig.
3 girders lumped
3 girders lumped
3 girders lumped
Fig 5 Prototype rigid frame  33 
Ige= Equivalent moment of inertia or the girder Ice= Equivalent moment of inertia of the column.
Ige = ( Ig1+Ig2+Ig3)
Fig6 Equivalent lumped girder frame The requirement of a substitute frame is that, for horizontal loading, joint translations should be the same as those of the original structure. For translations caused by girder flexure, equation below shows the requirements to be satisfied  34 
by assigning the inertia of the equivalent girder to be equal to the sum of the lumped ngirder inertias of the original frame, that is,
n
Ige = ∑ Ig
To determine the properties of the columns in lumped girder method, the component drift caused by the double curvature column bending story height nh in the equivalent frame to the corresponding drift over n stories in the original frame to the corresponding drift over n stories in the original frame are equated to get the equivalent column inertias. Or in other words the flexural strengths of the equivalent frame to original frame are equated to displacements to get the equivalent column inertias. Q(nh )2 Qh 2 = Ice 12E 12 E ∑ nh
∑
n
1 Ic ∑ h i
n3 Ice = n 1 ∑ Ic i The column sectional areas, which control the cantilever deflection, must have the same second moment about their common centroid in the lumped and the original structure. Consequently, the areas of the equivalent and the original structure remain the same as the original structure. The horizontal loading on the equivalent structure frame is applied as equivalent concentrated loads at the lumped girder levels, taking half new storyheight regions above and below the lumped girder as tributary areas.  35 
. heavy cladding may not be negligible and be significantly stiffening the structure. • Negligible stiffness: components stiffness of relatively small stiffness is assumed negligible. This assumption causes horizontal plane displacement of all vertical elements at a floor level to be definable in terms of the horizontal plane rigid body rotation & translation of the floor slab.When the lumped girder frame has been analyzed the . Similarly masonry infill may significantly change the behavior and increase the forces unconservatively in a surrounding frame. Although the assumption is generally valid. the results are transformed back to the original frame The moment in the original girder a t the same level will be 1/nth of the resulting moment and the moments in the girders between the lumped girder level shall be obtained by interpolating the moment. The common assumptions are • Participation components: Only the primary structural components participate in the overall behavior. • Floor slab: Floor slab is assumed to be infinitely rigid in plane. and the type of analysis. and the torsion stiffness of column beams. the minor axis stiffness of shear wall. Assumptions: Following simplifying assumptions are necessary to reduce the problem to viable sizes. These include for example. and walls.36  . The effect of secondary structural components and nonstructural components are assumed to negligible and conservative. The assumptions were made keeping in mind the frame tube building. the transverse bending stiffness of slab. Thus the number of unknown displacements to be determined in the analysis is greatly reduced. its anticipated mode of behavior.
semi rigid. rigid.. Their resistance to deformation is represented by material properties such as young’s modulus ( E ) and Poisson ratio ( ν ).• Negligible deformation: deformations. If the size problem still persists then a reduction analysis can be used for the analysis of the structure. and by geometric properties of the cross section such as are ( A ). The technique is explained below. How these idealization decisions are made. The complete frame work has 12 nodal degrees of freedom and 12 nodal force components as shown in figure. The corresponding displacements of these nodes .37  . which are relatively small and of little significance are neglected like the axial deformation of the beams. moment of inertia ( I ). Mathematical modeling begins with structural idealization members which have depth and width as well as length. All of the above assumptions are included for the analysis of frametube building. that is. is extremely important and involves considerable judgment. must be stipulated. satisfying all the compatibility the equilibrium equations. .the degree of freedom – are employed in the characterization of the displaced state of the element. The different types of Idealization used in mathematical modeling for a frame structure is explained in the following paragraphs. the previously discussed in plane bending and shear of the floor slab. and torsion constant ( J ) . it has to be furthered modeled for individual member and then joined together back into the structure. are conveniently reduced to line elements. etc. whether pinned. Our work concentrates on the modeling and the analysis of frametube building. Once the whole structure has being modeled for analysis. yielding. The behavior of the connection. or the necessary property determined. For the purpose of mathematical modeling the state of stress in the member is represented by forces at the element ends.
it is now used for different shapes and sometimes used for circular and triangular pans too. The uniformity of the structure makes construction simpler and enables the industrialized production of the components. Main reason do so is to make sure large amount of material is placed on the periphery of the building to maximize the inertia of the building.Frame tube buildings: Frame tube buildings consist of closely spaced column connected by deep spandrel beams placed on the perimeter in plan. The close spacing of the columns are viewed with mixed reaction with respect to aesthetics. In many buildings this is used to resist entire lateral load on the building to be resisted by the exterior columns while the gravity loads are shared between exterior columns and the interior structure in the form of columns and load inner core. For steel buildings large elements can be prefabricated in the factory and easily erected in the work site. In the lower floor it is desired to have columns spaced wide apart for which an deep transfer girder is provided which connects all the loads from the upper floors and transfers to the columns below. FramedTube buildings are generally suited for story height from (40100). In case of concrete construction use of gang forms raised enables speedy construction. Tubeform though original developed for rectangular plans. The panel’s normal to the direction of the wind is considered as flange and the panel perpendicular is considered as web of the perforated cantilever. .38  . In frame tube building columns are arranged unlike the frame structure with the strong bending direction of the columns aligned along the face of the building.
When subjected to lateral load the flange portion of the structure. on the windward is subject to tension force and the flange portion on the leeward side is subjected to compression.Mode of behavior: Although the structure has a tubelike structure. which deform such that the columns A and B are in tension and columns C and D are in compression. which mobilizes . The principle interaction between the web and the flange would be through the vertical displacement of the corner columns.39  . In addition to this. as indicated in the fig 7. When a frame tube building is subjected to lateral loading it is resisted by webside panel. This can be seen in the fig 8. These displacements correspond to the vertical shear in the girder of the flange frame. its behavior is more complex than a plain imperforated tube as the stiffness may be less. thereby increasing the compressive stresses in the corner column and decreasing the stresses in the interior columns of both the flange and the web portion of the structure. the parallel frames to the wind are subjected to the usual inplane bending and the shearing or racking action associated with an independent rigid frame. This action gets problematical by the fact that flexibility of the spandrel beams produces a shear lag effect.
the compressive deformation will not be identical since the spandrel beam will bend. by an amount equal to the stiffness of the .Compression Axial Stresses in Flange Column Flange Frames D C Compression Web Frames Web Frames Y N. Thus the axial deformation in adjacent columns will be less.40  . For example if the column C is in compression it will try to compress the adjacent column since the two are connected by spandrel beam.Axis Axial Stresses in the web columns A X B Tension Flange Frames Wind Axial stress distrabution in columns of laterally loaded framed tube Fig 7 the column forces in the flange column.
C1 C2 C3 ∆ Deformation of the flange frame causing shear Fig 8 Structural analysis: Ref (12 ) During analysis has assumed that the horizontal diaphragm slab has infinite in plane stiffness. Thus the deformations in successive adjacent columns in frame will be lesser than the previous one. This maintains the cross section shape of the structure at each level and the crosssection at these positions under go only rigid body . beams of infinite stiffness will develop a pure tubular action.41  . Hypothetically. Since the external applied moment is to be resisted by internal couple produced by the compressive and the tension stress produced on either side of neutral axis of structure.connecting beam. it follows that the stresses in the corner column will be greater than those in the interior column.
Horizontal displacement of the building can be expressed in terms of orthogonal translation and rotation.42  . provided that the floor loading is constant through out the structure. and the out of plane actions are generally negligible. and a twisting moment about the central axis of the structure. consisting of series of rigidly joined frame works connected together at the corners of the buildings. Though the analysis is possible it is time consuming and very costly. The response of the structure can be obtained by a superposition of the separate bending action about the XX axis and the YY axis and a pure torsion action. is doubly symmetrical about the plan. Usually a three dimensional structural analysis of frame tube buildings can be carried out using stiffness approach. but its feasibility depends on the computer capacity. The out of plane stiffness of the floor slab is assumed low so that they do not resist twisting and bending. If the plan of the frame tube building happens to be a square i.movements in plane. even though it is theoretically straightforward. The floor slab is assumed unable to provide the coupling between opposite normal to wind which act as flange of the side frame.e. When a frametube building is subjected to lateral loading the action of the floor system is mainly to transmit the horizontal forces between the different vertical members of the structure. Both the side and normal frames are therefore subjected largely to in plane action. a repetitive floor loading can be economically used for the design and the analysis of the structure. Analysis of FrameTube buildings is a complex problem as they are highly indeterminate structure. As floor systems does not participate otherwise in the lateral load resistance of the structure. Each frame consists of large number of elements and number of . Further the individual action may be simplified by utilizing the double symmetry of the structure in plan there by ½ structure analyses or ¼ structure analyses can be made. any applied load can be resolved into two orthogonal force components about the axis. then.
or relative displacements between nodes.symmetric system. Consequently. The restrained corners then remain fixed relative to each other during any translation or rotation under applied loads. If not a fictitious axially rigid joint pinended horizontal diagonal bracing members connecting nodes on the opposite corners at each floor level. acting on either a ½ or ¼ structure. The applied load may be treated as a combination of symmetric or skew.degree of freedom may run into thousands. the axial stiffness of the beams at each story level can be assigned to be so large that any in plane axial deformations. the high in plane rigidity of the floor slabs will have a considerable influence on the structural behavior. which provide an option for end constraints. The constraining action can be achieved in a number of different ways. by doing an analysis of ½ structure or even ¼ structure if the structure is symmetrical in plan about one or two central axes. . To obtain an accurate estimate of the structural behavior. The main action of the panels will be effectively restrained at each floor level. it is essential to include this constraining action in the structural analysis of the threedimensional formwork. In the building. Main action will then be in planes of the frame panels. the size of the stiffness matrix can be considerably reduced and hence the computation. it may be assumed that the cross section of the building will undergo only rigid body displacements.43  . in the horizontal plane at each floor level. In addition. As a result of the in plane rigidity. One way is using the option provided by the program. by ensuring that the out of plane deformation of the panels will be effectively restrained at each floor level. are negligible. The introduction of such diagonal bracing member has the disadvantage that it increases the bandwidth of the stiffness matrix and increases the solution time. translation and rotation.
7 subjected to bending by lateral forces in the X direction. with a consequently large reduction in the amount of computation required for a conventional full threedimensional analysis. The lateral load is resisted primarily by the following actions. Consequently. Consider initially the framed tube of Fig. it may be assumed that outofplane actions of the web frames are negligible. 2.C.44  . A. it is possible to reduce the analysis to that of an equivalent plane frame. It is also assumed that the torsion rigidities of the girders are negligible. The method is intuitively appealing to the engineer since.B. The shearing actions in the web panels AD and BC parallel to the direction of the applied load. and D. the interaction between the flange and web panels consists mainly of vertical interactive forces through the common corner columns. the flange panels AB and DC are subjected primarily to axial . by recognizing the dominant mode of behavior of the structure. The axial deformations of the normal frame panels AB and DC action effectively as flanges to the web panels. 1. and the frames are subjected only to planar actions. The axial displacements of the corner columns in the web frames are restrained by the vertical rigidity of the two normal frames. Due to the symmetry of the structure about the XX axis. As a result of these interactive forces. and the very high inplane stiffness of the floor slabs.Reduction of ThreeDimensional Framed Tube to an Equivalent Plane Frame In this section is presented a simplified yet accurate approximate method for the analysis of symmetrical framedtube structures subjected to bending produced by lateral forces.
for the web frames.45  . Under the applied lateral loading. This action may be of significance in the lower levels of the building since bending then occurs about the weaker axis of the columns. By virtue of the large lever arm that exists between these flange panels. In analyzing the primary mode of behavior. In the analytical model. the shear forces will this be resisted mainly by the web frames. All other torsion and outofplane actions may be considered to be secondary. a mechanism is required that will allow vertical shear forces. For the flange frames. the uniformity of which across the panel will depend on the stiffness. In addition. the fundamental compatibility condition that must be established is that of equal vertical displacements at the corner where the orthogonal panels meet. sue to the inplane rigidity of the floor slabs. the wind moments will be resisted most effectively if the maximum amount of axial force can be induced in the columns of frames AB and CD. on the spans and flexural rigidities.deformation. to be transmitted from the web panels to the flange panels through the corner columns. the joints must be free to rotate in the plane of the frame. but not horizontal forces or bending moments. of the connecting spandrel beams at each floor level. and to displace horizontally in unison in the plane of the frame at each floor level. while the bending moments will be resisted by the moments and axial forces in the columns of the web frames and the axial forces in the columns of the flange frames. the joints must be free rotate in the plane of the frame. and . apart from the outofplane bending of the columns in the flange frames whose horizontal deflections will be the same as those of the web frames. to displace vertically. that is.
on the other hand. EBH. so that no column is situated on the center line YY. only one quarter. of panel AB must be zero at the line of symmetry (E).to displace vertically. as a result of the slab inplane rigidity. the shear force in the beams. . But flange joints on the line of symmetry must. Since the structure is symmetrical about both center lines XX and YY. If. The required boundary conditions at the lines of symmetry and skew symmetry are then introduced as described because of symmetry about the XX axis. the bending moment at the line of skew symmetry must also be zero.9 are shown in Fig10. Conditions of skew symmetry (H) must be zero.9.46  . and the slope in the Y direction. the web frame BC contains an even number of columns. Appropriate support systems to simulate the required boundary conditions for the quadrant EBH of the structure of Fig. and preferably all flange joints should. For example. be constrained against horizontal displacement in the plane of the frame. need be considered in the analysis. consider the simple framed tube shown in plan in Fig. for example.
Axis Fig 9 Frame tube structure.Web Frames Y N. Structural plan Fig 10 Connection Details .47  .
The desired vertical interaction between the web and flange panel may be achieved in various ways.48  . This allows the displacement relating to specified degrees of freedom at two or more nodes in structure to be constrained .6. the forces may be applied at any convenient nodes. Since the beams are assumed axially rigid. Most comprehensive modern generalpurpose structural analysis programs include an inter nodal constraint option. therefore it can be assumed that the onequarter of the total lateral forces acting on the faces of the flange frames of the building can applied in the plane of the halfweb frame. The inplane stiffness of the floor slabs constrains all members of the two web bents to have the same horizontal deflection in the X direction. as indicated in Fig.fig10 The detail of a connection The equivalent planar system is obtained by “rotating” the normal halfpanel EB through 90° into the plane of the webhalfpanel BH.
immediately above. The duplicate nodes are then joined by a fictitious stiff beam with a flexural rigidity of say 10. These common nodes should again be numbered separately. and to reduce the sensitivity of the results to the stiffness of the connecting fictitious members. vertical compatibility is established and the required shear transmitted between the web and flange frames.6). once on each of the web and flange joints(B and B’ in Fig. An alternative technique has been devised to improve the disconnection of the rotations and lateral displacements between the frame panels. However. once on the web column. on the flange column. One simple technique is to displace horizontally the intersection column of each flange frame by a small distance of. say.000 times that of the larger adjacent girder and with one end assigned to be released for moment and axial force (Fig.7). The technique again involves the rotation of the flange frames into the plane of the web. The a ppropriate nodes at the intersections of the web and flange frames may then be specified directly in the analysis to have equal vertical displacements. it has been found that the results may be sensitive to the stiffness assumed for the fictitious beams. say less than one hundredth of the story height. onehundredth of the span of the adjacent beams. Each pair of intersection joint nodes is then connected by a fictitious . The intersection line column of each frame is shown superimposed on the distance. so that each common intersection joint is represented twice. If this option is not available. In numbering the nodes.49  . some other device must be used to achieve the required vertical compatibility at the junctions.to be identical. Thus each intersection joint is duplicated. the two nodes representing each intersection joint are numbered separately. and. while decoupling the rotation and horizontal displacement. By this device.
The additional column is connected by the pinended axially stiff links to the existing basic plane frame system. The links constrain the column to have the same horizontal deflection as the side panel members. The stiff links ensure vertical compatibility and transfer vertical shear between the web and flange frames while disconnecting rotations and vertical displacements. which may be of significance in the lower levels. the corner column is assigned its true inertia’s in the corresponding planes of the web and flange frames.000 times that of the intersection line column. say. In each model. 10. These columns suffer the same horizontal deflections about their weaker axis as the columns in the side frames do about their stronger axis. but its area should be assigned wholly to the column B’ in the web frame with a zero area assigned to the column B in the flange frame. it may be distributed to the individual column in proportion to their flexural rigidities. Horizontal and rotational constraints are applied to the flange frame nodes on the vertical line of symmetry and preferably. The effect of the outofplane bending may be included in the basic model by adding an equivalent column (RR in Fig.stiff vertical link with a large sectional area of. Once the total force and consequent moment has been determined for the effective column. horizontal constraints are also applied to all other flange frame nodes.50  .6). as a means of reducing the total number of degrees of freedom. . and allow it to carry its share of the lateral forces. The resulting planar model does not include the outofplane of columns in the normal frames. whose flexural rigidity is equal to the sum of the outofplane flexural rigidities of onequarter of the flange columns.
Each element has 2 nodes and each node has 3 degrees of freedom. The program has been tested for many varieties of problems and is fairly stable and the results are very accurate. Input data for 40stories structure and its reduction model. Care should be taken during the input process as it has a great impact on the result. The program is based on the matrix methods of structural analysis and is developed using Matlab. Frame tube building reduction model: Data: Number of nodes = 108 nos.51  .75 kips/column Actually structureFrame tube building 40 stories: Data: Number of bays = 11 nos Width of the bay = 12 ft . Width of the bay = 12 ft Building dimension in plan = 132ft X 132ft Story height = 60 ft Total height = 480 ft Wind load =87.Chapter 4 FINDINGS AND DISCUSSION The frame program developed in this thesis is intended to focuses on planar structures. Number of elements=184 nos. It can handle both trusses and frames.5 Kips/node Floor load =218. both by program and Mastan2 are as follows.
5 kips /column Mastan2 model for 8 story building: Data: Number of Nodes = 108 Number of Elements = 184 Number of Sections = 6 Number of Materials = 1 Number of Supports = 12 Applied Loads Lateral load=87.Building dimension in plan = 132ft X 132ft Story height = 12 ft Total height = 480 ft Column modulus of elasticity (E ) = 3150 Ksi.52  .4 Kips/node Floor load =43.5 kips/node . 453600 ksf Spandrel Beam (E) = 3150 Ksi Wind load =17.5 Kips/node Floor Load=218.75 Kips/node Mastan Model 40 story building: Data: Number of Nodes = 492 Number of Elements = 920 Number of Sections = 6 Number of Materials = 1 Number of Supports = 12 Applied Loads Lateral load= 17.
95 0.Floor load= 43.2752 0.00054 Reduction Model [Program Vs.36 2. Mastan] X Displace ment (in) 2.00041 3.53  .00055 From the values of comparative results table we can calculate the drift index of the building for both program and Mastan2. In case of matlab program drift index for first order analysis it is 0.792 2.000503 less than the allowable drift index of .00054 3.00054 I order Analy sis II Order Analy sis 3.84 0. Mastan] Program X Y θ Displace Displace Rotation ment ment (in) (in) 3.18 0.8934 Program Mastan Program Y X Y θ θ Displace Rotation Displace Displace Rotation ment ment ment (rad) (rad) (in) (in) (in) 1.19 0.2836 0.00041 I order Analy sis II Order Analy sis 2.30 2.8676 2.93430 0.00041 3.18 0.5 Kips/node Comparative result: [reduction model Vs actual model] 40 Story Model [Program Vs.30 2.00054 Mastan Program X Y θ Displacem Displace Rotation ent ment (in) (in) 3.
which is less than 0. a difference of 0.5”.002. Comparing the results of actual model with reduction model we see the values are quite close to one another.00066 and for second order analysis it is 0.54  . shear deformation is also less. we see that error between models are. thereby solutions obtained from the program is close to Mastan2 solutions. which is not accounted for in the program developed. the limit value of the drift index. The story deflection we get is less than the values calculated using Mastan2 for 40 story building. When First order analysis was done using Mastan2. story shear causes shear deformations. both less than the allowable limits. This is because in 40 story building. first order analysis it is 0. It is observed that the in reduction model.00067.49” (actual model – reduction model) and that for second order analysis is +0.48” this is because drift off error is not controlled in the program unlike the Mastan2. doing a first order analysis the solutions from the program and Mastan are exactly same up to 2 decimal places. .0. While that for second order analysis it is 0. it is observed that the values when compared between actual and reduction model.0006713.4” (actual model – reduction model) is got.9”. When first order analysis is done using the developed program. The drift Index calculated for Matan2 model is.002.. This confirms the stability of the building for deflection criterion. It is observed that shear deformation causes error to accumulate as the number of stories is increased. The second order analysis results between the models differ by 0. while the drift index for second order analysis of the building we get 0. When number of stories is less. +0.
member stress obtained from this program were very accurate and reliable. as matrix method is a back bone of FEM programs. For solving building over 50 stories good computer is a necessity as total degree of freedom of the whole structure increases and we encounter memory problem.55  . The member forces. For small structures with story height below 8 errors is the same as Mastan2 up to 2 decimal places. this program would serve as a skeleton to build a advanced Finite element software. . The values differ because the program developed calculates buckling load by applying load in one step unlike Mastan2 where load is applied in increments.The results from the buckling load analysis in the case of very tall buildings have to be sorted out and the least positive value is the buckling load and remaining values are the mode shapes. Nodal displacement and rotation are also accurate. arc length control method and the displacement control method could be attempted. The buckling load ratio from the program is 61. Iterative program employing Newton Raphsons method has been developed but could not be checked for numerical stability and so has not been included in this wok so a future research to append the remaining different second order methods like the work control method. Recommendation for future research: The program developed is modular program and therefore in future research appending dynamic analysis modules to this program would be a good start. Program run time is very small in fraction of seconds and the results obtained are accurate. Finite element method is the most widely used method to solve engineering problem. Calculations using Mastan2 yielded approximately 50.
Finally. I think this program has all that one needs for the development of a basic and simple FEM program. .56  .The algorithm still needs to be fine tuned for numerical stability and checked for unique cases. so an effort in this direction is worth a try.
GLOSSARY First order analysis. Elastic buckling load. Second order analysis. will return back to that position subsequent to the removal of the disturbing force. does not return. Neutral equilibrium: A state of equilibrium in which the body remains in the position to which the disturbing force has moved it. An analysis carried out for structures whose deformation is small and exhibit elastic behaviour.57  . Stable equilibrium. The load at which the deformation of a slightly imperfect system increases without bound. A Matrix structural analysis commercial soft ware. A state of equilibrium where in a body is slightly displaced from its original position of equilibrium. . Mastan2. Unstable equilibrium: A state of equilibrium in which. A iterative method which takes into account the material and geometric nonlinearity. but instead continues to move further away from original equilibrium position. when a body is displaced slightly from its equilibrium position of rest.
A.Rise Buildings. and Bose. Nov 1975..58  . pp.. K.St5.C.. Aug. N. Proc. “Analysis and Design of Framed Tube structure for all tall buildings”.No. Pergamon Press. Pouangare. 104. No. Vol.B. Coull. 7. London. K.. ASCE. pp 857.102. 6. England. eds. J.” Journal of the Structural Division..R. Khan F. pp. A.pp. Vol.” Journal of Structural Division. 1973.. May.B. 4. “A simple model for the Analysis and Design of Frame Tube Structures. Coull. A.. ASCE.” Framed–Tube Structures for Highrise Buildings” Journal of the Structural Division.22232240 3. 1976. ST.” Torsion of FramedTube structures. and Bose. Paper 11696.23662370.” . Vol. 1978. 20972105.11.. A. Symposium on Tall buildings.” Simplified Analysis of FramedTube structures. Proceedings of the American Society of Civil Engineers. 1971. A.861 5. pp. Coull. Proceedings of the American society of civil Engineers.” Proceedings.December.12. Khan.References 1. and Ahmed. 1967. “Deflections of Framed Tube Structures. 2.R and Amin N. Paper 8301. and Subedi. University of Southampton.571590. 97. J. Coull and Stafford Smith.ST. No. Vol.101. R.. ST8. No. F.. Proc. ”Current Trends in Concrete High.. ACI Publication SP36: Response of multistory Concrete structures to Lateral Forces. Connor and C.. pp3960.” Journal of the Structural Division. Coull.
Chen. 12. Tall Buildings StructuresAnalysis and Design. Grandin. John Wiley & sons Publication.” PhD dissertation. ca.. W. Tall Building Structures Analysis and Design. Smith and A.” New York. 10. Illinois. H. Fundamentals of the Finite Element Method. Taranath. Inc.. and E. 1972 9. 15. “Analysis and design of tubetype buildings structure. Inc. S. S. Toronto. Matrix Structural Analysis. B. Department of Civil Engineering. H. college of engineering research center. earth quake engineering research center. M. Structural Stability . McGrawHill. Chen. Alexander Chajes. McGuire. University of California at Berkeley. F. New jersey. 1974.” Publication 7201. Engle Cliffs. “Structural analysis and design of tall buildings. and R. 16. 11.59  .. W.8. Elsevier Science Publication Co. W. Stability Design of Steel frames. J. D. and E. Jr. Coull. c1988. . P. Cannada.. Second Edition. Lui. M. “Principles of Structural Stability theory ”Civil Engineering Mechanics series. R. B.A. Lui.. PretenceHall. Ziemian. 13. Waveland press. Schwaighofer. 14. F.. De Clercq. and Ast. CRC press. University of Toronto. Inc.. “Tables for the Analysis of FramedTube Buildings. John Wiley & Sons. Mar. Boca Raton. Gallagher. Florida. H.Theory and Implementation.
Carrie. “Structural collapse of the World trade center Towers Subjected to Explosions caused by Aircraft”. Worcester Polytechnic Institute. “Approximate Analysis and optimization of tall framed tube buildings” Ms Thesis dissertation. Dept of Civil engineering.F.17. P. . P.ASCE. 18. MA.60  . Jayachandran. M.
Section 2 explains the input file of a simple 2D Truss. which can be quickly modeled using these programs. The programs have a modular structure and user adaptive. This tutorial is recommended as a quick start.61  .APPENDIXA MANUAL Matrix Program is simple to use and userfriendly. but to develop complex models knowledge of Matlab is necessary. New users should go through this tutorial first. They are compared with other commercial software and errors have been presented. They are developed as RC Framed structure using the GUIinput program devolved which later runs to give the results. . Section 3 presents an input file for frame tube building in two parts. The objective of this manual only to make the user an insight into the program developed and to understand the different variables he inputs into the program with the aid tutorial problems. Lastly these tutorials shall help to understand the basis for the analysis of the 40 storey structure. This dynamic nature of the algorithm is the start to more advanced programming using matrix methods. No prior knowledge of the language of the program is necessary to get started. In first part modified frame tube structures are presented. Section 1 explains the program using the flow chart and algorithms. The above tutorial problems include only Analysis and the output are plotted and displayed on the screen as well as in an output file. This manual presents two tutorials.
Algorithm: Here is given a series of Algorithm to give a brief insight into the logic of the program. Construct Stiffness Matrix and Force Vector o Construct Element stiffness matrix o Group to form global structure stiffness matrix [ Ke ] o Construct Force Vector { F } 3. n = number of iterations o Loop i to n § Compute { ∆U } = [KT]1 { ∆F } § Update coordinate displacement Vector { U } = { U } + {∆ U} § Compute internal stresses and forces in each element .62  . Input Structure details: READ: o Coordinates o Applied Forces o Boundary Conditions o Element group and Material properties o Element Connectivity 2. OPTION : a) Linear b)Nonlinearc)Critial buckling a) LINEAR o Compute displacement Vector { U } = [ Ke ] 1 { F } o Calculate Internal Stresses and Forces o Display results o STOP b) NONLINEAR o Construct element Geometric stiffness matrix initially with zero internal stress o Group to form initial global structure geometric stiffness matrix [ Kg ]1 o Total Structure Stiffness matrix [ KT ] = [ Ke ] + [ Kg ]1 o Construct ∆F = [ F ] / n . ALGORITHM 1.
two story RC frame that will be analyzed. construct element geometric stiffness matrix o Construct Structure Geometric Stiffness matrix [ Kg ] o Solve [ Kg + λ Kg ] { ∆ } = 0 for λ and { ∆ } which represent the eigenvalues (critical load factor corresponding to the specified force vector) and the corresponding mode shapes. o Display results o STOP TUTORIAL 1 Description of the Tutorial Problem The structure for this project is a single bay. F3 F 2 1 6 5 4 . The figure below shows the structure.§ § o End loop o Plot Loaddisplacement Curve for specified node and DOF o STOP Construct new geometric stiffness matrix [ Kg ]i with computed internal stresses Update Total Structure stiffness matrix [ K T ] = [ K e ] + [ Kg ]i c) CRITICAL o Construct Element stiffness matrix o Group to form global structure stiffness matrix [ Ke ] o Construct Force Vector { F } o Compute internal stresses and Forces o With these forces.63  . Comparative results are also tabulated in the end. A first order analysis and Elastic critical buckling load analysis is done and results are presented.
0000000 0.0000000 0.00000000 2 0.00000000 100. 0=Free .txt" containing the input data for the above structure has been provided with the program.0000000 0.An input file called "INTRO. This file contains input data.0000000 0.00000000 50.00000000 50.00000000 3 0.00000000 0.0000000 0.0000000 0.00000000 6 5.00000000 0.00000000 4 5.64  .00000000 BOUNDARY CONDITIONS Node 1 1 2 0 3 0 4 1 5 0 6 0 X 1 0 0 1 0 0 Y Z 1 0 0 1 0 0 1= Fixed.00000000 100.00000000 5 5. Basic Data CONTROL INFORMATION Number of Nodal Points Number of Spatial Dimensions Number of Degrees of Freedom Number of Element Groups Number of Material =6 =2 =3 =1 =1 NODAL COORDINATES: Node X Y Z 1 0.
txt file.00000000 1000.APPLIED FORCES Node DOFs 1 2 3 4 5 6 0. The input data is explained below in the following paragraphs. .00000000 0.00000000 ELEMENT INFORMATION Element Group Number 1 A 10 E 200000 I 100 G 0 CONNECTIVITY: 1 2 3 4 5 6 1 1 1 1 1 1 1 2 4 5 2 3 2 3 5 6 5 6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Refer the intro.0000000 0.00000000 0.00000000 1000.00000000 0.00000000 10000.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 100.00000000 10000.0000000 0.00000000 0.00000000 0.65  .
Element group type ( possibilities are 2. fixed x Dof. On line 10 we have 1 1 1 1 0 which represent node number. number spatial degree of freedom( 2D planar).In the second line we 6 2 3 1 1 all separated by space represent number of elements. On line 13 we have 2 1000 0 0 0 represent node number. fixed rotation about z axis. Lines 10 11 12 are for support condition. Lines 1317 are for force. area of cross section. Force in the Y Direction. On Line 19 we have 1 10 200000 100 0 representing Material number. no interpolation.On the 9 line we have 0 which represents end of data for read_coord module program. number of elements. moment of inertia and Shear modulus. material number. Force in the xdirection. Note 0 represent free and 1 represents fixed condition. Modulus of elasticity. rotation about the z axis and no interpolation. ith node. jth node and no interpolation. On 17 line 0 represent termination of the respective program module. 2 = Nonlinear analysis. i. On Line 18 we have 1 4 6 1 representing element group number. Don’t think what happened to 1 and 2 that’s not your problem). . The remaining input data is for node connectivity. On line 20 we have 1 1 1 2 0 representing element number.e either beamcolumn or truss). 1 0 0 0 all separated by spaces represent node number. if there are series of nodes then instead of entering one by one it can be done in 1 step). number of materials used. Degree of freedom per node. type of analysis (1 = Linear Elastic analysis. co ordinate position in x. element type (3 represent truss type and 4 represents beam type. 3 = critical buckling load analysis ). For example. coordinate position in y and lastly interpolation decision variable (1 invokes interpolation loop. fixed y Dof.66  . Note ve sign represent force towards left of node and downwards direction. Next 7 lines is for coordinate positions.
79 19.95 163.This completes the input file.38465 4 0.04616 0.14009 3 3.65683 0.89 200.00000 0.04616 Elastic Critical Buckling Load: (λ) 0.00000 0.04440 0.00000 0. For more logical understanding refer the algorithm and flow chart.71 82.61535 θ 0.94683 0. Phi X Y 1 0. Fem_main is the main program which needs to be invoked then type the input file name and then program run to finish.78 5.00000 2 1.01 39.65683 0.94695 0. OUTPUT PHI OUTPUT Node No.00 .15 200.00000 0.00000 5 1.67  .04440 0.35991 6 3.
0000 3.0000 0. A first order analysis is done and results are presented. Comparative results are also tabulated in the end.9468 0.1535 0.0000 1.9470 0.044398 0.00000 1.8465 0.00000 0.6568 3.0000 1.00000 1.94695 0.4009 3.0000 0.5991 6.8465 0.68  . .046162 TUTORIAL 2 PROBLEM DICRIPTION: The structure for this project is a 2D planar truss that will be analyzed.1535 0.6568 3.04440 0.04616 0.046163 0.04616 0.044398 0.65683 3.00000 1.00000 3.00000 0. The figure below shows the structure.4009 3.5991 6.04440 0.0000 1.Comparative results: Program Mastan X Y θ X Y θ 0.65683 3.94683 0.
000000 Z 0.69  .00000 4 10000.00000 3 5000.00000000 0.0000 Y 0.00000000 0.00000000 0.000000 8670.00000000 BOUNDARY CONDITIONS Node 1 1 2 0 3 0 4 0 DOFs 1 1 0 0 APPLIED FORCES AND DISPLACEMENTS Node DOFs .00000000 0.00000000 2 5000.00000000 8670.Fv D C Fh A B 2D Planar Truss CONTROL INFORMATION: Number of Nodal Points =4 Number of Spatial Dimensions = 2 Number of Degrees of Freedom = 2 Number of Element Groups = 1 NODAL COORDINATES Node X 1 0.
80813 .83626 2. Input basically is the same with changes as to Dof per node which is 2 in this case.00000 2 0.95114 7.00000 3 9.23479 4 10.00000000 0.00000 0.0000 ELEMENT INFORMATION Element Group Number 1 CONNECTIVITY 1 2 3 4 5 1 2 2 1 2 1 1 2 3 2 2 3 3 4 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Referring to the input file truss.00000000 0.00000000 0. Output: Nodal displacement PHI OUTPUT Node No.00000000 0.0000 282840.1 2 3 4 0.70  .40779 0.00000000 0.00000000 282840. Phi X Y 1 0.txt.
80813 0.71  .0000 .40779 9.0000 0.00000 0.83626 10.40779 9.0000 2.Comparative results: Program Mastan X Y θ X Y θ 0.95114 0.8363 1.00000 0.0951e+001 0.00000 0.00000 0.00000 0.23479 7.2348 7.8081 0.0000 0.0000 0.0000 0.00000 2.00000 0.0000 0.
5Kips/node Floor load = 218.75 kips /node Strength of concrete = 3000 ksi Frame tube building reduction modeling technique Lumped girder Method: In this case reduction modeling is done to reduce the structure from 40 story to 8 story for ease of analysis and data manipulation. Width of the bay = 12 ft Building dimension in plan = 132ft X 132ft Number story = 8 nos. Story height = 60 ft Total height = 480 ft Column modulus of elasticity (E ) = 3150 ksi. 120 STORIES.5kips/ node and the floor load is 218.72  .75. For columns having Ic =182250in4 and area Ac= 1080in2 and spandrel beam having Ig=182250in4 And area Ag=1080 in2 the equivalent inertias are .APPENDIXB Frame Tube Reduction model of the building: Data: Number of bays = 11 nos. The wind load is 87. 453600 ksf Spandrel Beam (E) = 3150 ksi Wind load = 87.
For columns having Ic=111132in4 and area Ac= 756in2 and spandrel beam having Ig=111132in4And area Ag=756in2 the equivalent inertias are .Ice = n3 53 = 5 = 4556250in 4 n 1 1 ∑ Ic ∑ 182250 1 1 n 1 5 1 Ige = ∑ Ig = ∑ 182250 = 911250in 4 2030 STORIES. For columns having Ic=129654in4 and area Ac= 882in2 and spandrel beam having Ig=129654in4And area Ag=882in2 the equivalent inertias are Ice = n3 53 = 5 = 3241350in 4 n 1 1 ∑ Ic ∑ 129654 1 1 n 1 5 1 Ige = ∑ Ig = ∑ 129654 = 648270in 4 3040 STORIES.73  .
and the comparative charts are presented in the succeeding pages to follows. Stories Ige (ft4) 120 14 43. The results of the program are compared with the Mastan results.79 Ice (ft4) 219.31 133.2751 ft or 3.9 2030 46 31.26 3040 68 26.25 Ac (ft2) 7. displacement clearly show that the displacement at the top of the building is 0.30in .5 6.125 756 The results of the reduction analysis are presented in tabulated form in end.125 5. The plots of story height vs.7 156.74  . Stories Ige (in4) 120 14 911250 2030 46 648270 3040 68 555660 Ice (in4) 4556250 3241350 2778300 Ag (in2) 1080 882 756 Ac (in2) 1080 882 756 Reduction model section properties.98 Ag (ft2) 7.Ice = n3 53 = 5 = 2778300in 4 n 1 1 ∑ Ic ∑ 111132 1 1 n 1 5 1 Ige = ∑ Ig = ∑ 111132 = 555660in 4 Reduction model section properties.5 6.
. The plot below shows that the moment at the bottom is high and gradually at the top becomes zero. The results match with Mastan results.30in. which is indicated by high values of rotation at the bottom and very low increase in the values of rotations in the middle portion and no rotation at the top portion of the building.75  .275ft or 3. The plot of rotation gives the cumulative rotation along the height of the building.and the result from the program and Mastan is 0. Plot 1 The plot of rotation Vs height of building is presented below.
28ft or3. Load increment is 0. Max iteration specified is 10. greater the iteration greater will be the accuracy.32 ft or 3.The plot indicates a slight nonlinearity with max displacement being 0. Mastan program gives the deflection value to be 0.76  . .84 inches.36.1*200 (the reference load).Plot 4 Second order analysis: A simple step method is employed for the analysis.
.Mastan2 model for 8 story building: Input for Structural Analysis (i) Number of Nodes = 108 (ii) Number of Elements = 184 (iii) Number of Sections = 6 (iv) Number of Materials = 1 (v) Number of Supports = 12 (vi) Applied Loads lateral load=87.75 Kips/node The Mastan model diagram is given below and comparative solution is also presented.77  .5 Kips/node Floor Load=218.
Comparative results: X Displace ment (ft) 0.00054 Mastan Program X Y θ Displace Displace Rotation ment ment (ft) (ft) 0.320 Program Y Displace ment (ft) 0.275 0.78  .00054 0.00054 0.280 0.1815 0.1822 0.275 0.1815 θ Rotation 0.00055 I order Analysis II Order Analysis .
Load Ratio
Displacement Mastan Plot of Displacement Vs Load Ratio Actually structureFrame tube building 40 stories: Data:
 79 
Number of bays = 11 nos Width of the bay = 12 ft Building dimension in plan = 132ft X 132ft Story height = 12 ft Total height = 480 ft Column modulus of elasticity (E ) = 3150 Ksi, 453600 ksf Spandrel Beam (E) = 3150 Ksi Wind load =17.4 Kips/node Floor load =43.5 kips /column
Floor 1 19 2029 3040
Column dimension (in) B D H 24 45 144 21 42 144 18 42 144
Spandrel beam (in) B D H 24 45 144 21 42 144 18 36 144
Concrete (Ksi) F’ c 3000 3000 3000
Floor 119 2029 3040
Column dimension (in) B D H 24 45 144 21 42 144 18 42 144
Spandrel Beam (in) B D H 24 45 144 21 42 144 18 42 144
Concrete F’c(ksi) 3000 3000 3000
Flexural strength: Story 119 2029 3040 Columns and spandrel beam Moment of inertia and Area Strong axis I (in4) Weak Axis I (in4) Area (in2) 182248.7 51840 1008 129662.21 332410.368 882 111186.43 20404.224 756
 80 
The results of the reduction analysis are presented in tabulated form in end. The results of the program are compared with the Mastan results, and the comparative charts are presented in the succeeding pages to follows. The plots of story height vs. displacement clearly show that the displacement at the top of the building is 2.89348in and the result from the program and Mastan is exactly the same up to 4 th decimal place. The drift factor of the buildings is 0.0005017.The drift factor for 40 stories
Plot of Drift factor
drift factor 0.0006 0.0004 0.0002 0 1 6 11 16 21 26 31 36 41 story number Drift factor
 81 
The plot of Vertical displacement Vs distance along the perimeter is presented below. .82  . The results obtained from Mastan are compared with the program and plot shows that the results match with one another. The vertical displacement of the columns along the perimeter in the plan is considered.
The rotations are in radians and for complete rotation of all the nodal points refer the out put file.83  . .The plot below is plot of rotation along the story height.
The graph indicates that the deflection is around 2.8 in.95. The drift factor of the building is . The curve obtained from Mastan2 is also presented and from the graph and it is clear that the deflection is 3. The linear nature of the curve and the proximity of the result close to 2.89 obtained from Ist order analysis clearly indicates that there is not much of a second order effect taking place. The structure there is stable.84  .Second order analysis: Results from second order analysis compared are presented in the form of a graph of Load ratio versus displacement.
The limit value for drift factor is 0. The elastic critical load ratio from Mastan is 47.0.The buckling load factor for the building calculated using the program is 61.85  .6. Mastan Model 40 story building: General Information Categories: (i) Number of Nodes = 492 (ii) Number of Elements = 920 .0005121.002.
86  .5 kips/node Floor load= 43.5 Kips/node .(iii) Number of Sections = 6 (iv) Number of Materials = 1 (v) Number of Supports = 12 (vi) Applied Loads Lateral load= 17.
00054 3.792 2.8934 2.2752 0. Mastan] X Displace ment 2.30 2.00054 3.00054 3.93430 θ Rotation 0.18 0.2836 0.Load Ratio Displacement in X Fig Mastan Load Ratio Vs Displacement Comparative result: [reduction model Vs actual model] 40 Story Model [Program Vs.36 2.87  . Mastan] Program Mastan Program X Y X Y θ θ Displace Displace Rotation Displace Displace Rotation ment ment ment ment (in) (in) (in) (in) I order 3.30 2.00041 Mastan Program X Y θ Displace Displace Rotation ment ment 3.00055 .8676 2.18 0.00054 I order Analysis II Order Analysis Analysis II Order Analysis Reduction Model [Program Vs.00041 0.95 Program Y Displace ment 1.84 0.19 0.00041 3.