of
Reinforced Concrete Buildings
Lecture Notes
August 2002
Shunsuke Otani
Department of Architecture
Graduate School of Engineering
University of Tokyo
ii
Preface
This note is intended to introduce the state of the art in the nonlinear response analysis of reinforced
concrete building structures under earthquake excitation to graduate students. The state of the
knowledge on the behavior of reinforced concrete members and structures and the art of nonlinear
response analysis are far form an established state. Therefore, this note will not provide any unique
solution to a problem.
The note was initially prepared for a special lecture on “nonlinear analysis of reinforced concrete
buildings” at Department of Civil Engineering, University of Canterbury, New Zealand, from February
to April, 1994. The note has been revised for use in Department of Architecture, University of Tokyo
since 1996; this course was given in English. The note was extensively revised for a series of lectures
on “nonlinear earthquake response analysis of reinforced concrete buildings” at European School for
Advanced Studies in Reduction of Seismic Risk, Universita degli Studi di Pavia, Italy, from February
to March, 2002.
The use of this note should be limited to personal use.
August 2002
Professor Shunsuke Otani
Department of Architecture, Graduate School of Engineering
University of Tokyo
otani@sake.t.utokyo.ac.jp
http://www.rcs.arch.t.utokyo.ac.jp/otani/
iii
Contents of Lecture
1. Introduction
2. Properties of Reinforced Concrete Materials
2.1 Concrete
2.2 Reinforcing Steel
2.3 Bond
3. Behavior of Reinforced Concrete Members
3.1 Behavior of Beams
3.2 Behavior of Columns
3.3 Behavior of Interior Beamcolumn Connections
3.4 Behavior of Exterior Beamcolumn Connections
3.5 Behavior of Structural Walls
4. Analysis of Reinforced Concrete Members
4.1 Flexural Analysis of Section
4.2 MomentCurvature Relation under Reversed Loading
4.3 Flexural Analysis of Members
4.4 Loaddeformation Relation of Beams
4.5 Analysis of Structural Walls
5. Structural Dynamics
5.1 Differential Equation of Motion
5.2 Mass of Inertia
5.3 Damping
5.4 Strainrate Effect
5.5 Properties of Earthquake Ground Motion
6. Numerical Integration Methods
6.1 Introduction
6.2 NigamJennings’ Direct Integration Method
6.3 Linear Acceleration Method
6.4 Newmark Beta Method
6.5 Wilson’s Theta Method
6.6 RungeKuttaGill Method (Fourth Order)
7. Matrix Analysis of Linearly Elastic Plane Frames
7.1 Assumptions
7.2 Member Stiffness Matrix in Local Coordinates
7.3 Coordinate Transformation
7.4 Member Stiffness Matrix in Global Coordinates
7.5 Continuity of Displacement at Joint
7.6 Equilibrium of Forces at Joint
7.7 Formulation of Structural Stiffness Matrix
7.8 Free Joint Displacements and Support Reactions
7.9 Member End Actions
8. Numerical Solution of Linear Equations
8.1 Incremental Formulation
8.2 Modified Cholesky Matrix Decomposition
8.3 Solution of Linear Algebraic Equations
8.4 Static Condensation
8.5 Damping Matrix
iv
9. Formulation of Member Stiffness Matrix
9.1 Introduction
9.2 Formulation of Member Stiffness Matrix
9.3 Member with Rigid Ends
9.4 Member with Flexible Ends
9.5 Simply Supported Member
10. Member Stiffness Models
10.1 Member Stiffness Model
10.2 Fiber Model
10.3 Discrete Element Models
10.4 Onecomponent Model
10.5 Multicomponent Model
10.6 Distributed Flexibility Model
10.7 Multispring Model (1)
10.8 Multispring Model (2)
10.9 Multispring Model (3)
10.10 Wall Models
11. Member Hysteresis Models
11.1 Introduction
11.2 Bilinear Model
11.3 RambergOsgood Model
11.4 Degrading Trilinear Model
11.5 Clough Degrading Model
11.6 Takeda Degrading Model
11.7 Pivot Model
11.8 Stable Hysteresis Models with Pinching
11.9 Sheartype Hysteresis Models
11.10 Axial ForceBending Moment Interaction
11.11 Special Purpose Models
12. Response of Different Models
12.1 Effect of Member Modeling
12.2 Effect of Damping Modeling
13. Response of Different Hysteresis Models
13.1 Analysis Method
13.2 Effect of Initial Stiffness (Takeda Model)
13.3 Effect of Cracking Force Level (Takeda Model)
13.4 Effect of Yield Resistance Level (Takeda Model)
13.5 Effect of PostYielding Stiffness (Takeda Model)
13.6 Effect of Unloading Stiffness Degradation Parameters (Takeda Model)
13.7 Effect of Hysteresis Energy Dissipation
13.8 Effect of Parameter of RambergOsgood Model
13.9 Response to Different Earthquake Motions
13.10 Response of Different Models
13.11 Response Waveforms and Hysteresis Relations
13.12 Effect of Hysteresis Shape on Frame Response
1４. Reliability of Nonlinear Response Analysis Methods
14.1 Introduction
14.2 Reinforced Concrete Column
14.3 Frame Structures
14.4 Framewall Structures
14.5 Wall Structures
v
15. U.S.Japan Fullscale Test
15.1 Test Program of Fullscale Sevenstory RC Building
15.2 Description of Test
15.3 Modeling of Structural Members
15.4 Stiffness of Member Models
15.5 Method of Response Analysis
15.6 Results of Analysis
15.7 Concluding Remarks
Suggested Reading
American Concrete Institute, Earthquake Resistant Concrete Structures  Inelastic Response and
Design, ACISP127, American Concrete Institute, Detroit, 1991.
Comite EuroInternational du Beton, RC Elements under Cyclic Loading  State of the Art Report,
Thomas Telford, 1996, 190 pp.
Comite EuroInternational du Beton, RC Frames under Earthquake Loading  State of the Art Report,
Thomas Telford, 1996, 303 pp.
1
Chapter 1. INTRODUCTION
Dynamic response of a structure can be caused by different loading conditions such as: (a)
earthquake ground motion; (b) wind pressure; (c) wave action; (d) blast; (e) machine vibration; and
(f) traffic movement. Among these, inelastic response is mainly caused by earthquake motions and
accidental blasts. Consequently, more research on nonlinear structural behaviour has been carried
out in relation to earthquake problems.
1.1 Structural Dynamics
Note that dynamic problems are different from static one in the following points: (a) inertial force;
(b) damping; (c) strain rate effect; and (d) oscillation (stress reversals). These need to be clarified in
order to analyze a structure under dynamic loading.
The equation of motion for a linearly elastic system under horizontal ground motion is normally
expressed in a form below;
[ ]{ } [ ]{ } [ ]{ } {0} m z c x k x + + =
where [ ] m , [ ] c , [ ] k : mass, damping coefficient and stiffness matrices, { } z : absolute acceleration
vector at mass level, { } x and { } x : velocity and displacement vector at mass level relative to the
structural foundation.
Dynamic characteristics up to failure cannot be identified solely through a dynamic test of a real
structure for the following reasons: (a) difficult to understand the behaviour due to complex
interactions of various parameters; (b) expensive to build a structure, as a specimen, for destructive
testing; and (c) capacity of loading devices insufficient to cause failure. Consequently, dynamic tests
of real buildings are rather aimed toward obtaining data (a) to confirm the validity of mathematical
modeling techniques for a linearly elastic structure; and (b) to obtain damping characteristics of
different types of structures. A specifically designed laboratory test becomes inevitable in order to
complement the weakness of fullscale tests and to study the effect of individual parameters.
Damping: Any mechanical system possesses some energydissipating mechanisms, for example:
(a) inelastic hysteretic energy dissipation; (b)
radiation of kinetic energy through foundation;
(c) kinetic friction; (d) viscosity in materials; and
(e) aerodynamic effect. Such capacity or energy
dissipation is vaguely termed “damping,” and is
most often assumed to be of viscous type
simply because of its mathematical simplicity;
i.e., resistance proportional to velocity. It should
be noted that the actual damping mechanism
may not be of viscous type.
Damping capacity is often determined by the
response curve (a plot of response amplitudes
at steady state oscillation with respect to
excitation frequencies) during a sinusoidal
steadystate test. Figure 1 shows such
acceleration response curves for a reinforced
concrete building at different excitation levels
（Jennings and Kuroiwa 1968). The frequency
corresponding to peak response indicates the
natural frequency of the structure, and the band
width of the response curve represent the
damping capacity. Note the shift of resonant
frequencies and the change in amplitudes of
damping with increase of excitation level despite
Fig. 1: Observed Acceleration amplitudes from
Steadystate test (Jennings and Kuroiwa, 1968
2
low response amplitudes.
Damping capacity, expressed in terms of viscous damping factor, is not a unique value of a
structure, but it depends on the level of excitation. The stateoftheart does not provide a method to
determine the damping capacity based on the material properties and geometrical characteristics of
a structure.
Strain Rate Effect: It is technically difficult to test structural members under dynamic conditions in a
laboratory. Before forcedeformation relations already obtained from thousands of static tests can be
studied for use in a dynamic analysis, the effect of strain rate on the forcedeformation relation
needs to be examined.
The speed of loading is known to influence the stiffness and strength of various materials (Cowell
1965, 1966). Some member test results are available (Mahin and Bertero l972). Important findings
from these investigations are as follows: (a) high strain rates increased the initial yield resistance,
but caused small differences in either stiffness or resistance in subsequent cycles at the same
displacement amplitudes; (b) strain rate effect on resistance diminished with increased deformation
in a strainhardening range; and (c) no substantial changes were observed in ductility and overall
energy absorption capacity.
Note that strain rate (velocity) during an oscillation is highest at low stress levels, and that the rate
gradually decreases toward a peak strain. Cracking and yielding of a reinforced concrete member
reduce the stiffness, elongating the period of oscillation. Furthermore, such damage is normally
caused by the lower modes of vibration having long periods. Therefore, the strain rate is small in the
case of earthquake response, and its effect on the response is small.
Consequently, the static hysteretic behaviour observed can be utilized in a nonlinear dynamic
analysis of reinforced concrete structures.
1.2 Stiffness Properties of Reinforced Concrete Members
It is not feasible to analyze an entire structure using microscopic material models. It is more
important to study the behaviour of isolated members and their subassemblies (beamcolumn,
slabcolumn, and slabwall
connections) so that their
analytical models can be
developed for use in the analysis
of a complete structure.
A typical forcedeflection curve
of a cantilever column is shown
in Fig. 2 (Otani et al. 1979). Note
the following observations: (a)
tensile cracking of concrete and
yielding of longitudinal
reinforcement reduced the
stiffness; (b) when a deflection
reversal was repeated at the
same newly attained maximum
amplitude (for example, cycles 3
and 4) the loading stiffness in the
second cycle was lower than that
in the first cycle, although the
resistances at the peak
displacement were almost
identical; and (c) average
FIG. 2. Hysteretic characteristics of reinforced
concrete member (Otani et al. 1979)
3
stiffness (peaktopeak) of a complete cycle decreased with a maximum displacement amplitude.
For example, the peaktopeak stiffness of cycle 5, after large amplitude displacement reversals,
was significantly reduced from that of cycle 2 at comparable displacement amplitudes. Therefore,
the hysteretic behaviour of the reinforced concrete is sensitive to loading history. Let us study some
typical stiffness characteristics of reinforced concrete.
Flexural Characteristics: The
flexural deformation index (average
curvature) is obtained from
longitudinal strain measurements at
two levels assuming that a plane
section remains plane. This flexural
deformation index does not
represent the flexural deformation
in a strict sense because a plane
section does not remain plane in a
region where an extensive shear
deformation occurs. However, the
index is useful for understanding
flexural deformation characteristics
qualitatively.
A typical momentflexural
deformation index curve obtained
from a simply supported beam test
(Celebi and Penzien 1973) is
shown.in Fig. 3. Note that the stiffness during loading gradually decreases with load, forming a fat
hysteresis loop, and absorbing a large amount of hysteretic energy. The hysteresis loops remain
almost identical even after several load reversals at the same displacement amplitude beyond
yielding. Consequently, vibration energy can be efficiently dissipated through flexural hysteresis
loops without a reduction in resistance. Many hysteretic models, as discussed later, are currently
available to represent the nexural behaviour.
The increase in axial force decreases the flexural ductility of a reinforced concrete member, but
increases force levels corresponding to (a) tensile cracking of concrete; and (b) tensile yielding of
longitudinal reinforcement.
Shear Characteristics: Similar to
the flexural deformation index, a
shear deformation index is defined
from strain measurements in the two
diagonal directions. Again, this index
does not represent the true shear
deformation because the
interference of shear and flexure
exists.
A typical lateral loadshear
deformation index curve (Celebi and
Penzien 1973) is shown in Fig. 4.
Unlike what occurs in flexure, the
stiffness during loading gradually
increases with load, exhibiting a
“pinching” in the curve. The
hysteretic energy dissipation is
smaller. The hysteresis loop decays
with the number of load reversals, resulting in a smaller resistance at the same peak displacement in
each repeated loading cycle. Although the curve shows a “yielding” phenomenon, it is important to
FIG. 4. Shear deformation characteristics
(Celebi and Penzien 1973)
FIG. 3. Flexural deformation characteristics
(Celebi and Penzien 1973)
4
recognize that the shear force of the member was limited by flexural yielding at the critical section
rather than by yielding in shear. This yielding clearly indicates the interaction of shear and bending.
The pinching in the forcedeformation curve is obviously less desirable. The shear span to
effective depth ratio is the most significant parameter. Decreasing the shear span to depth ratio
causes a more pronounced pinching in the curve, and a faster degradation of the hysteretic
energydissipating capacity. Considerable improvements in delaying and reducing the degrading
effects can be accomplished by using closely spaced ties. Existence of axial force tends to retard the
decrease in stiffness and resistance with cycles. However, it is hard to eliminate this undesirable
effect when high shear stress exists. Consequently, it becomes important to include this degrading
behaviour in a behavioural model for a short, deep reinforced concrete member. The current state of
knowledge is not sufficient to define the stiffness degrading parameters on the basis of the member
geometry and material properties.
Bar Slip and Bond Deterioration: When a structural element is framed into another element, some
deformation is initiated within the other element. Consider a beamcolumn subassembly. Bertero and
Popov (1977) reported a significant rotation at a beam end caused by the slippage (pullout) of the
beam's main longitudinal reinforcement within the beamcolumn joint (Fig. 5). The general shape of
the momentbar slip rotation curve is similar to that shown in Fig. 4, demonstrating a pronounced
pinching of a hysteresis loop. The contribution of bar slip to total deformation cannot be neglected,
especially in a stiff member (short or deep).
FIG. 5. Rotation due to bar slip (Bertero and Popov 1977).
Biaxial Lateral Load Reversals: During an earthquake, columns of a framed structure must resist
lateral forces simultaneously in longitudinal and transverse directions. Recent tests at the University
of Toronto (Otani et al. 1979) on reinforced concrete columns showed that the columns under lateral
load reversals in two perpendicular directions exhibited the decay of resistance and stiffness at a
faster rate than those under uniaxial lateral load reversals. This topic needs further study.
1.3 Hysteretic Models for Reinforced Concrete
Nonlinear dynamic analysis of a reinforced concrete structure requires two types of mathematical
modelling: (a) modelling for the distribution of stiffness along a member; and (b) modelling for the
forcedeformation relationship under stress reversals.
5
A hysteresis model must be able to provide the stiffness and resistance under any displacement
history. At the same time, the basic characteristics need to be defined by the member geometry and
material properties. The current state of knowledge is sufficient to define flexural hysteresis models.
However, it is not sufficient to determine the degree of stiffness degradation due to the deterioration
of shearresisting and rebarconcrete bond mechanisms.
Bilinear Model: The elasticperfectly plastic
hysteretic model was used by many investigators
because the model was simple. The maximum
displacement of an elastoplastic simple system
was found (Veletsos and Newmark 1960) to be
practically the same as that of an elastic system
having the same initial period of vibration as long
as the period was longer than 0.5 s.
A finite positive slope was assigned to the
postyield stiffness to account for the
strainhardening characteristic, and the model was
called a bilinear model. The bilinear model does
not represent the degradation of loading and
unloading stiffnesses with increasing displacement
amplitude reversals (Fig. 6), and the model is not
suited for a refined nonlinear analysis of a
reinforced concrete structure.
Clough 's Degrading Stiffness Model: A qualitative model for the reinforced concrete was
developed by Clough (1966), who incorporated the stiffness degradation in the elastoplastic model :
the response point during loading moved toward the previous maximum response point. The
unloading slope remained parallel to the initial elastic slope. This small modification improved the
capability to simulate the flexural behaviour of the reinforced concrete. Compared with the
elastoplastic model, less energy is absorbed per cycle beyond yielding by Clough's degrading
model.
From the response analysis of a series of
singledegreeoffreedom systems, Clough (1966)
concluded that (a) the degrading stiffness model did
not cause any significant change in the ductility
demand of longperiod structures (period longer
than 0.6 s) compared with the elastoplastic model;
on the other hand, (b) the degrading stiffness model
required significantly larger ductility from
shortperiod structures than the corresponding
elastoplastic systems; and (c) the response
waveform of a degrading stiffness model was
distinctly different from that of an ordinary
elastoplastic model.
The model is relatively simple, and has been
used extensively in nonlinear analysis with the
inclusion of strainhardening characteristics (Fig. 7).
Takeda's Degrading Stiffness Model: A more refined and sophisticated hysteresis model was
developed by Takeda et al. (l970) on the basis of experimental observation. This model included
stiffness changes at flexural cracking and yielding, and also strainhardening characteristics. The
unloading stiffness was reduced by an exponential function of the previous maximum deformation.
Takeda et al. also prepared a set of rules for load reversals within the outermost hysteresis loop.
FIG. 6. Bilinear hysteresis model
FIG. 7. Clough’s degrading stiffness model.
6
These are major improvements over the Clough
(1966) model.
Failure or extensive damage caused by shear
or bond deterioration was not considered in the
model. The Takeda model, similar to the Clough
model, simulates dominantly flexural behaviour
(Fig. 8). Simplified Takeda hysteresis models
were proposed by Otani and Sozen (1972) and
by Powell (1975), using a bilinear backbone
curve.
To test the goodness of the Takeda model,
cantilever columns tested on the University of
Illinois earthquake simulator were analyzed (Takeda et al.1970). Calculated waveforms were
favourably compared with the observed waveform as shown in Fig. 9.
FIG. 9. Takeda model applied to column analysis (Takeda et al. 1970)
TakedaTakayanagi Models: The amplitude of the exterior column axial load varies greatly due to
the earthquake overturning moment, and changes its momentcarrying capacity. Takayanagi and
Schnobrich (1976) incorporated the effect of axial force variation in the Takeda model by preparing
various backbone curves at different axial load levels (Fig. 10a).
A pinching action and strength decay are inevitable in a short and deep member due to bar slip
and deterioration in shear resistance. Takayanagi and Schnobrich (1976) introduced a pinching
action and strength decay in the Takeda model (Fig. 10b). Whenever a response point was located
in the positive rotationnegative moment range or the negative rotationpositive moment range, the
pinching was introduced. After the moment exceeded the yield level, strength decay was
incorporated. The values of guideline for strength decay and pinching stiffness were not related to
the member geometry and material properties.
FIG. 8. Takeda’s degrading stiffness models.
7
(a) axial force variation (b) pinching and strength decay
FIG. 10. TakedaTakayanagi models (Takayanagi and Schnobrich 1976):
Degrading Trilinear Hysteresis Model: A
model that simulates dominantly flexural
stiffness characteristics was developed in
Japan (Fukada 1969). The backbone curve is
a trilinear shape with stiffness changes at
cracking and yielding. Up to yielding, the
model behaves in the same way as the
bilinear model. Once deformation exceeds
the yield point, the model behaves as a
perfectly plastic system. Upon unloading, the
unloading point is treated as a new “yield”
point, and unloading stiffnesses
corresponding to pre and postcracking are
reduced proportionately so that the behaviour
becomes of the bilinear type in a range
between the positive and the negative yield
points (Fig. 11).
The degrading trilinear model can easily
include strainhardening characteristics. The hysteresis energy dissipation per cycle beyond the
initial yielding is proportional to the displacement, and the equivalent viscous damping factor
becomes constant. The fatness of a hysteresis loop is sensitive to the choice of a cracking point.
Comments
Many other hysteresis models have been proposed and used in the past. Figure 12 shows
attained ductility factors of singledegreeoffreedom systems with any of four flexural hysteresis
models: bilinear; Clough; Takeda; and degrading trilinear models. The four models have the same
backbone curve except the cracking point. The four models show similar variations of attained
ductility factors with periods, but attained ductility factors show a wide scatter from one model to
another, especially in a shortperiod range.
A reinforced concrete building is normally designed to behave dominantly in flexural mode, brittle
failure modes such as diagonal tension failure in shear being carefully prevented at the design stage.
Thus hysteretic models representing shear behavior were not studied.
FIG. 11. Degrading trilinear model.
8
1.4 Reinforced Concrete Member Model
Inelastic deformation of a reinforced concrete member does not concentrate in a critical location,
but rather spreads along the member (Fig. 13). Various member models have been proposed to
represent the distribution of stiffness within a reinforced concrete member. The effect of gravity load
on the beam behaviour and the contribution of slabs to the structural stiffness will not be discussed.
FIG. 13. Deformation of beam under gravity and earthquake loads.
Onecomponent Model: An elastoplastic frame structure was analyzed by placing a rigid plastic
spring at the location where yielding is expected. The part of a member between the two rigid plastic
springs remains perfectly elastic. All inelastic deformation is assumed to occur in these springs (Fig.
14). This onecomponent model was generalized by Giberson (1967).
Fig. 12. Response of various hysteresis models
9
FIG. 14. Onecomponent member model.
A major advantage of the model is that inelastic memberend deformation depends solely on the
moment acting at the end so that any momentrotation hysteretic model can be assigned to the
spring. This fact is also a weakness of the model because the memberend rotation should be
dependent on the curvature distribution along the member, hence dependent on moments at both
member ends. Consider two cases of moment distribution along a member AB with corresponding
curvature distributions as shown in Fig. 15. The inelastic rotations at the A end are given by the
shaded areas. For the same moments at the A end, case II causes larger inelastic rotation at the A
end. Consequently, this simple model does not simulate actual member behaviour. Furthermore, it is
not rational to lump all inelastic deformations at member ends.
FIG. 15. Inelastic rotation of beam: (a) moment; (b) curvature and inelastic rotation.
The stiffness of an inelastic spring is normally defined by assuming an asymmetric moment
distribution along a member with the infection point at midspan. The usage of the initial location of
the inflection point in evaluating spring properties was suggested by Suko and Adams (1971).
However, once yielding is developed at one member end, the moment at the other end must
increase to resist a higher stress, moving the inflection point toward the member centre. At the same
time, a large concentrated rotation starts to occur near the critical section. Despite rational criticisms
against this simple model, the performance of the onecomponent model is expected to be
reasonably good for a relatively lowrise frame structure, in which the inflection point of a column
locates reasonably close to midheight.
A specialpurpose computer program, SAKE (Otani 1974), for a regular rectangular reinforced
concrete frame structure and recent modifications (Powell 1975) to generalpurpose computer
program DRAIN 2D (Kanaan and Powell 1973) used the onecomponent model.
Multicomponent Model: In an effort to analyze frame structures well into the inelastic range under
earthquake excitation, an interesting model was proposed by Clough et al. (1965). A frame member
was divided into two imaginary parallel elements: an elastoplastic element to represent a yielding
phenomenon, and a fully elastic element to represent strainhardening behaviour. When the
10
memberend moment reaches the yield level, a plastic hinge is placed at the end of the elastoplastic
element. A memberend rotation depends on both memberend moments. Aoyama and Sugano
(1968) adapted the twocomponent model, creating the multicomponent model (Fig.16), using four
parallel beams to account for flexural cracking, different yield levels at two member ends, and
strainhardening. The deformation compatibility of the imaginary components is satisfied only at their
ends.
FIG. 16. Multicomponent member model.
The multicomponent model appears to have merit; rotation at one end of a member depends on
both memberend moments. In other words, the moment distribution along a member can be
approximately reflected in the analysis. However, the stiffness of the multiparallel components must
be evaluated under a certain assumed moment distribution. Therefore, the stiffness parameters are
valid only under such a moment distribution, and are bound to be approximate when the moment
distribution becomes drastically different.
Giberson (1967) discussed the advantages and disadvantages of the onecomponent and the
twocomponent models, and concluded that the onecomponent model was more versatile than the
twocomponent model because the twocomponent model was restricted to the bilineartype
hysteresis characteristics. This twocomponent model was used in a generalpurpose computer
program DRAIN 2D (Kanaan and Powell 1973) for plane structure. The interaction of the bending
moment and the axial force was easily incorporated by simply changing the yield value of the
elastoplastic component depending on the existing axial force.
Connected Twocantilever Model: When a frame is analyzed under lateral loads only, the member
moment distributes linearly. From the similarity of moment distribution, the member can be
considered to consist of two imaginary cantilevers, free at the point of contraflexure and fixed at the
member end, and connected at the inflection point, satisfying the continuity of displacement and
rotation (Otani and Sozen 1972).
The flexibility relation of a member was formulated by assuming: (a) the inflection point did not
shift much during a short time increment; (b) freeend rotation and displacement were proportional to
the beam length and the square of the beam length, respectively; and (c) instantaneous stiffness for
shearrotation and sheardisplacement curves of a unit length reference cantilever could be defined
by hysteretic models.
The weakness of this method is that the member flexibility matrix is a function of the location of the
inflection point, which tends to shift rapidly when the sign of a memberend moment changes. This
causes a numerical problem. Consequently, this method cannot be recommended for a general
dynamic analysis. However, the method is useful for incremental static load analysis of a structure.
Discrete Element Model: In order to overcome difficult problems of variable stiffness distribution
along a member, the member can be subdivided into short line segments along the length, with each
short segment assigned a nonlinear hysteretic characteristic. The nonlinear stiffness can be
assigned within a segment, or at the connection of two adjacent segments.
Wen and Janssen (1965) presented a method for dynamic analysis of a plane frame consisting of
elastoplastic segments. Consequently, the mass and flexibility of a member were lumped at the
11
connecting points on a tributary basis. Powell (1975) suggested using a degrading stiffness
hysteresis model for rigid inelastic connecting springs (Fig. 17a). Shorter segments were
recommended in a region of high moment, and longer segments in a lowmoment region.
An alternative method is to divide a member into short segments, each segment with a uniform
flexural rigidity that varies with a stress history of the segment (Fig. 17b). Local concentration of
inelastic action can be easily handled by arranging shorter segments at the location of high
concentration of inelastic deformation (Takayanagi and Schnobrich 1976).
These methods are useful when more accurate results are required, or in the analysis of walls.
More computational effort is required compared with the other simple models.
FIG. 17. Discrete element model: (a) lumped inelastic stiffness; (b) distributed inelastic stiffness.
Distributed Flexibility Model: Once cracks develop in a member, the stiffness becomes
nonuniform along the member length. Instead of dividing a member into short segments, Takizawa
(1973) developed a model that assumed a prescribed distribution pattern of crosssectional flexural
flexibility along member length. A parabolic distribution with an elastic flexibility at the infection point
was used (Fig. 18). The flexura1 flexibility at member ends was given by a hysteretic model
dependent on a stress history.
FIG. 18. Distributed flexibility model.
This is an interesting concept in analyzing an inelastic member. However, the parabolic flexibility
distribution may not describe the actual concentration of deformation at critical sections (normally at
member ends) due to flexural yielding and deformation attributable to slippage of longitudinal
reinforcement within a beamcolumn connection. The usage of inelastic springs at locations of
concentrated deformation in conjunction with this model may be a useful solution.
Summary
Various member models are reviewed, and their advantages and disadvantages are discussed.
12
These models have been developed specifically for earthquake response. Development of a simple
model for simultaneous gravity and earthquake situations is desired.
Member stiffness matrix, equilibrium of forces and continuity of displacement at joints are used to
formulate a series of linear equations under a given loading condition. Numerical integration
methods are used to solve the equation of motion under dynamic loading conditions. The response
of a structural model is evaluated by solving the set of linear equation incremental time steps.
1.5 Reliability of Analytical Models
Earthquake simulator tests provide interesting opportunities to examine the goodness of different
analytical models in simulating the observed response of small to mediumscale highly inelastic
model structures. This section reviews the reliability of different analytical models in relation to the
capability to simulate the observed behaviour.
These test structures were designed to behave dominantly in flexure, being prevented as much
as possible from failing in shear or anchorage because the two types of failure are not desirable in
real construction and are avoided in a design process.
Threestorey Onebay Frames (I): Smallscale threestorey onebay reinforced concrete frames
were tested on the University of Illinois earthquake simulator, and were analyzed using the
connected twocantilever member model (Otani and Sozen 1972). The stiffness properties of
individual members were calculated on the basis of the geometry and material properties. The
Takeda model was used to represent the forcedeflection of each cantilever model. Memberend
rotation due to bar slip was approximated by the simplified bilinear Takeda model, which did not
simulate the pinching behaviour. The damping matrix was assumed to be proportional to the
instantaneous stiffness matrix.
FIG. 19. Connected twocantilever model applied to threestorey frame analysis
(Otani and Sozen 1972): (a) measured; (b) calculated (h=0.0); (c) calculated (h=0.02)
The model structure was subjected to a base motion simulating the El Centro (NS) 1940
accelerogram. The firstfloor displacement was measured to be as much as four times the yield
displacement calculated under static lateral loads. The analytical models with and without viscous
13
damping favourably simulated the largeamplitude oscillations at 1.0, 2.0, and 5 s from the beginning
of the motion (Fig. 19). The analytical models, however, failed to simulate the medium and
lowamplitude oscillations. Note that the frequencies at the medium to lowamplitude oscillations are
higher for the analytical model, which indicates that the test structure was more flexible at low stress
levels than the analytical model. In order to reproduce lower amplitude oscillations of the observed
response waveforms, the pinching behaviour needs to be incorporated in a hysteretic model.
Threestorey Onebay Frames (II): Another set of threestorey onebay smallscale reinforced
concrete frame structures was tested on the University of Illinois earthquake simulator (Otani 1976).
The base motion is significantly more intense than a design earthquake motion.
A member was represented by the onecomponent model with two inelastic rotational springs at
each member end: one for the flexural deformation and the other for the memberend rotation due to
bar slip (Otani 1974). Takeda models with trilinear and bilinear backbone curves were assigned to
the two inelastic springs. Two types of damping were used in the analysis: (a) a damping matrix
proportional to the constant mass matrix; and (b) a damping matrix proportional to an instantaneous
stiffness matrix. The firstmode damping factor was 5% of critical at the initial elastic stage.
Observed and calculated thirdlevel displacement waveforms are compared in Fig. 20 (Otani 1976).
The comparison is fair for largeamplitude oscillations, and poor at lowamplitude oscillations. Again
in this analysis, the pinching characteristic was not incorporated. A fair agreement between the
computed and the observed may be attributable to the fact that the yielding was developed at most
member ends at the largeamplitude oscillations, and that the inflection point tended to be near the
midpoint of each member in such a lowrise frame structure.
FIG. 20. Onecomponent model applied to threestory frame analysis (Otani, 1976):
(a) measured; (b) calculated (mass proportional damping);
(c) calculated (stiffness proportional damping).
Twostorey Onebay Frame: A twostorey onebay mediumscale frame structure with slabs was
tested on the University of California earthquake simulator (Hidalgo and Clough 1974). The structure
was analyzed using the twocomponent model. In an effort to improve the correlation, the elastic
stiffness of the two parallel components was degraded as a function of the firstmoderesponse
14
amplitude history. The observed and the calculated secondfloor displacement waveforms are
satisfactorily compared in Fig. 21. However, the parameters controlling stiffness degradation could
not be determined from the theory.
FIG. 21. Twocomponent model applied to twostory frame analysis (Hidalgo and Clough 1974).
Tenstorey Coupled Shear Walls: Tenstorey coupled shear walls were tested on the University of
Illinois earthquake simulator (Aristizaba1Ochoa and Sozen 1976). Takayanagi and Schnobrich
(l976) divided a wall into short segments of uniform stiffness, and represented connecting beams by
the onecomponent model. The TakedaTakayanagi model with changing axial force was assigned to
a wall element, and the TakedaTakayanagi model with pinching action and strength decay was used
in a beam. It was judged that the usage of twodimensional plane stress elements for the walls was
less desirable because such an approach might cost more computational effort without any
compensating increase in accuracy.
(a) displacement at level 10, in inches (1 in. = 25.4 mm) (b)Acceleration at level 10, g.
FIG. 22. Analysis of tenstory coupled shear wall (Takayanagi and Schnobrich 1976):
The comparison of the measured and calculated displacement and acceleration is excellent, as
shown in Fig. 22. It is necessary to include the effects of inelastic axial rigidity of the wall section and
pinching action and strength decay of the connecting beams to reproduce the maximum
displacement response and the elongation of the period. Some stiffness parameters for the walls
and connecting beams were defined on the basis of static tests of connecting beamwall assemblies.
Summary
The favorable comparison of the measured and the calculated response waveforms encourages
the use of correct analytical and hysteretic models. It is desirable in developing a mathematical
model that all parameters of the proposed model should be evaluated on the basis of the geometry
of a structure and the properties of materials.
15
1.6 Threedimensional Building Analysis
The development of analytical methods has made it feasible to discuss nonlinear behaviour of
reinforced concrete plane structures with a certain confidence. However, we have not yet reached a
point to discuss, with any confidence, the nonlinear behaviour of threedimensional building
structures. Columns of a framed structure must resist lateral forces in two horizontal directions. The
stiffness is reduced significantly under biaxial lateral load reversals (Otani et al. 1979).
The first nonlinear analysis of a frame structure under horizontal biaxial ground motion was made
by Nigam (1967) with elastoplastic columns. Multistorey frames with rigid beams and floor slabs
under horizontal biaxial ground motion were studied by Pecknold (1974). Prager's kinematic
hardening theory as modified by Ziegler (1959) was used. The basic effect of biaxial inelastic
interaction is to produce a softer structure. Pecknold (1974) confirmed Nigam's finding that: (a)
horizontal biaxial ground motion increased ductility demand for stiff structures (initial natural period
less than 0.3 s); and (b) horizontal biaxial ground motion had little effect on the ductility demand of
flexible structures. Neither study included the stiffness degradation property.
Aktan et al. (1973) used the finite element technique to include the stiffness degradation through
the degradation of material properties. The biaxial ground motion was found to cause 20200%
larger response than the uniaxial ground motions from columns for which the calculated deflection
under uniaxial ground motion exceeded approximately twice the yield deflection.
Takizawa and Aoyama (1976) extended the onedimensional degrading trilinear hysteretic model
into a twodimensional model on the basis of plasticity theories (Ziegler 1959). The proposed model
was judged to predict the significant trends of the biaxial behaviour of reinforced concrete test
columns. The effect of biaxial response interaction was reported to be significant for degrading
stiffness models, and not so important for nondegrading type hysteretic models.
A brief review of nonlinear analysis of threedimensional structures indicates the necessity of
further study in this area.
1.7 Summary
The behaviour of reinforced concrete buildings, especially under earthquake motion, was briefly
reviewed. When a structure can be idealized as plane structures, the current stateoftheart
provides useful and reliable analytical methods.
However, more research is required to understand the effect of slabs, gravity loads, and biaxial
ground motion on nonlinear behaviour of a threedimensional reinforced concrete structure.
References:
Aktan, A. E., D. A. W. Pecknold, and M. A. Sozen, 1973. Effect of twodimensional earthquake
motion on a reinforced concrete column, University of Illinois, Urbana, IL, SRS No. 399.
Aoyama, H., and T. Sugano, l968. A generalized inelastic analysis of reinforced concrete structures
based on the tests of members. Recent researches of structural mechanics. Contribution in
Honor of the 60th Birthday of Professor Y. Tsuboi, UnoShoten, Tokyo, pp. 1530.
AristizabalOchoa, J. D., and M. A. Sozen, 1976. Behaviour of tenstorey reinforced concrete walls
subjected to earthquake motion, University of Illinois, Urbana, IL, SRS No. 431.
Bertero, V. V., and E. P. Popov, 1977. Seismic behaviour of momentresisting reinforced concrete
frames. In Reinforced concrete structures in seismic zones, American Concrete Institute, Special
Publication No. 53, pp. 247292.
Celebi, M., and J. Penzien, l973. Experimental investigation into the seismic behaviour of critical
region of reinforced concret components as influenced by moment and shear, Earthquake
Engineering Research Center, University of California, Berkeley, CA, EERC 734.
16
Clough, R. W., l966. Effect of stiffness degradation on earthquake ductility requirements, Structural
and Materials Research, Structural Engineering Laboratory, University of California, Berkeley, CA,
Report 6616.
Clough, R. W., K. L. Benuska and E. L. Wilson, l965. Inelastic earthquake response of tall buildings,
Proceedings, 3rd World Conference on Earthquake Engineering, New Zealand, Vo1. II, Session
II, pp. 6889.
Cowell, W. L., 1965. Dynamic tests of concrete reinforcing steels, U.S. Naval Civil Engineering
Laboratory, Port Hueneme, CA, Technical Report 394.
Cowell, W. L., 1966. Dynamic properties of plain Portland cement concrete, U.S. Naval Civil
Engineering Laboratory, Port Hueneme, CA, Technical Report 447.
Fukada, Y., 1969. Study on the restoring force characteristics of reinforced concrete buildings (in
Japanese), Proceedings, Kanto District Symposium, Architectural Institute of Japan, Tokyo,
Japan, No. 40.
Giberson, M. F., 1967. The response of nonlinear multistory structures subjected to earthquake
excitation, Earthquake Engineering Research Laboratory, California Institute of Technology,
Pasadena, CA, EERL Report.
Hidalgo, P., and R. W. Clough, l974. Earthquake simulator study of a reinforced concrete frame,
Earthquake Engineering Research Center, University of California, Berkeley, CA, EERC 7413.
Jennings, P. C., and J. H. Kuroiwa, 1968. Vibration and soilstructure interaction tests of a ninestory
reinforced concrete building, Bulletin of the Seismological Society of America, 58, pp. 891916.
Kanaan, A. E., and G. H. Powell, l973. DRAIN2D, A general purpose computer program for dynamic
analysis of inelastic plane structures, Earthquake Engineering Research Center, University of
California, Berkeley, CA, EERC 736.
Mahin, S. A., and V. V. Bertero, 1972. Rate of loading effect on uncracked and repaired reinforced
concrete members, Earthquake Engineering Research Center, University of California, Berkeley,
CA, EERC 729.
Nigam, N. C., 1967. Inelastic interactions in the dynamic response of structures, Ph.D. thesis,
California Institute of Technology, Pasadena, CA.
Otani, S., 1974. SAKE  A computer program for inelastic response of R/C frames to earthquakes,
University of Illinois, Urbana, IL, Structural Research Series, No. 413.
Otani, S., 1976. Earthquake tests of shear wallframe structures to failure, Proceedings, ASCE/EMD
(Engineering Mechanics Division) Specialty Conference, Dynamic Response of Structures,
University of California, Los Angeles, CA, Mar. 1976, pp. 298307.
Otani, S., V. W. T. Cheung and S. S. Lai, 1979. Behaviour and analytical models of reinforced
concrete columns under biaxial earthquake loads, Proceedings, 3rd Canadian Conference on
Earthquake Engineering, Montreal, P.Q., ,pp. 11411168.
Otani, S., and M. A. Sozen, 1972. Behaviour of multistory reinforced concrete frames during
earthquakes. University of Illinois, Urbana, IL, Structural Research Series, No. 392.
Pecknold, D. A. W., 1974. Inelastic structural response to 2D ground motion, ASCE Journal of the
Engineering Mechanics Division, 100(EM5), pp. 949963.
Powell, G. H., l975. Supplement to computer program DRAIN2D, Supplement to report, DRAIN2D
user's guide, University of California, Berkeley, CA.
Suko, M., and P. F. Adams, 197l. Dynamic analysis of mu1tibay multistory frames, ASCE Journal of
the Structural Division, 97(ST10), pp. 25192533.
Takayanagi, T., and W. C. Schnobrich, 1976. Computed behaviour of reinforced concrete coupled
shear walls, University of Illinois, Urbana, IL, Structural Research Series, No. 434.
Takeda, T., M. A. Sozen and N. N. Nielsen, 1970. Reinforced concrete response to simulated
earthquakes, ASCE Journal of the Structural Division, 96(ST12), pp. 25572573.
Takizawa, H., 1973. Strong motion response analysis of reinforced concrete buildings (in Japanese),
Concrete Journal, Japan National Council on Concrete, II (2), pp. l021.
Takizawa, H., and H. Aoyama, 1976. Biaxial effects in modelling earthquake response of R/C
structures, Earthquake Engineering & Structural Dynamics, 4, pp. 523552.
Veletsos, A. S., and N. M. Newmark, l960. Effect of inelastic behaviour on the response of simple
systems to earthquake motions, Proceedings, 2nd World Conference on Earthquake Engineering,
Tokyo and Kyoto, Vol. II, pp. 895912.
Wen, R. K., and J. G. Janssen, 1965. Dynamic analysis of elastoinelastic frames, Proceedings, 3rd
World Conference on Earthquake Engineering, Wellington, New Zealand, Jan. 1965, Vol. II. pp.
713729.
17
Ziegler, H. 1959. A modification of Prager's hardening rule. Quarterly of Applied Mechanics, 17(1), pp.
5565.
Chapter 2. Properties of Reinforced Concrete Materials
Material properties of reinforced concrete are briefly reviewed in this chapter. The following
references are recommended.
1. R. Park & T. Paulay: Reinforced Concrete Structures, John Wiley & Sons, Inc., 1975, 769 pp.
2. Comite EuroInternational du Beton, Structural Concrete under Seismic Action, AICAPCEB
Symposium, Rome, Vol. 1  State of the Art Report, Bulletin d’Information No. 131, May 1979, 286 pp.
3. Comite EuroInternational du Beton, Response of R.C. Critical Regions under Large Amplitude
Reversed Actions, Bulletin d’Information No. 161, August 1983, 306 pp. 4. Comite EuroInternational
du Beton: RC Elements under Cyclic Loading  State of the Art Report, Thomas Telford, 1996, 190 pp.
2.1 Concrete
Concrete is a hardened material obtained from a carefully proportioned mixture of (a) cement, (b)
sand, (c) gravel and (d) water in forms of the shape and dimensions of the desired structure. The
compressive strength of concrete is determined by static test on either standard cylinders or cubes.
The strength varies with (a) concrete mix, (b) age of testing, (c) curing method, (d) specimen shape
and size, and (e) loading speed. The watercement ratio is the main factor that controls the strength
of the concrete.
Compressive strength normally used in building construction ranges from 20 to 60 MPa. Higher
strength concrete is used in (a) columns of the lower part of a building to resist higher axial load, (b)
prestressed concrete members to balance the use of higher strength prestressing steel, and (c)
precast concrete for early removal of casting form.
Uniaxial StressStrain Relation in
Compression: The stressstrain relationship of
concrete under shortterm monotonically
increasing uniaxial compressive loading shows
gradual deterioration in stiffness with strain
even at a low stress level caused by
development of micro cracks. However, the
curve may be represented by a straight line up
to approximately 70 percent of the compressive
strength The secant stiffness at a point at
onethird of the compressive strength is often
used to represent this linear portion. Maximum
resistance is attained at approximately 0.002
strain, followed by descending branch.
The stressstrain relation of concrete
in compression varies with the strength
of concrete. Higher strength concrete
exhibits high initial stiffness and steep
descending slope after attaining the
compressive strength. The strain at
compressive strength does not change
appreciably with the concrete strength.
It should be noted that the damage
does not distribute uniformly over the
height of a specimen; the damage
normally develops in the middle part of
the specimen because the concrete
strength at the specimen ends is
normally enhanced by the confinement.
The lateral expansion by the Poisson’s
Stressstrain relation with concrete strengths
effect is resisted by the friction between the testing machine. The friction provides confining pressure
at the specimen ends, where the strength is enhanced; the damage concentrates in the middle part.
For a given stress level, a larger strain is measured in the middle part, and smaller strain near the
ends. The strain measurement is affected by the choice of gauge length especially in descending part
of the stressstrain curve.
The descending part of the loaddeformation relation of a concrete specimen is extremely difficult
to measure because the elastic strain energy stored in the testing machine is released abruptly
during the descending part of the test, causing sudden failure of the testing specimen; a stiff loading
machine is needed for the purpose.
Young’s Modulus: Elastic (Young's) modulus E
c
of concrete is normally defined as secant modulus
at approximately onethird of the compressive strength. The value of elastic modulus is often given by
the following empirical formula (Pauw, 1960):
0.5 1.5
1.5
1.35
33 '
c B
c c
E GPa
E w f psi
γ σ =
=
in which γ : unit mass density of air dried concrete (1000 kg/m
3
),
B
σ : compressive strength (MPa)
of concrete, w: airdry weight of the concrete at testing (pound per cubic feet), and '
c
f : concrete
strength at testing (pound per square inches). The unit mass density γ of concrete may be taken as
2.5 (x1000 kg/m
3
) for normal weight concrete. A wide scatter of data should be observed from the
Pauw's empirical expression.
ACI364 report on highstrength concrete
Young's modulus of highstrength concrete is influenced by the type of coarse aggregates. An
empirical expression was proposed for the elastic modulus E
c
concrete (Tomosawa, Noguchi, and
Onoyama, 1990), taking into account compressive strength and density of concrete, type of coarse
aggregates and mineral admixture;
4 1/ 3 2
1 2
3.35 10 ( / 60) ( / 2.4)
c B
E k k σ γ = × × × × × (MPa)
Reliability of Young’s Modulus Estimate (Pauw, 1960)
in which k
1
: factor representing type of coarse aggregates, k
2
: factor representing kind of mineral
admixture,
B
σ : observed concrete strength (MPa), γ : unit density of concrete (ton/m
3
). The factor k
1
is 0.95 for crushed quartzite, crushed andesite, basalt and clayslate aggregates, 1.0 for other coarse
aggregates, and 1.2 for crushed limestone and calcined bauxite aggregates. Factor k
2
is 0.95 for
silica fume, fine powder of blast furnace slag and fly ash fume, 1.00 for concrete without mineral
admixture or with other mineral admixture, 1.10 for fly ashes. Ninetyfive percent of test data are
shown to fall within 20 percent of the empirical expression. The modulus is important in defining the
elastic period of a structure. Therefore, if the dynamic analysis procedure is used in the design the
modulus should be controlled, especially in a large and important construction project, within an
acceptable range from the value specified by a structural engineer.
Poisson’s Ratio and Shear Modulus: Poisson's ratio µ of concrete in "elastic" range is
approximately 1/6. Shear modulus G
c
of concrete is defined by
) 1 ( 2 µ +
=
c
c
E
G
The shear modulus is used to estimate shear modulus of structural walls, where shear deformation
may not be neglected in comparison with flexural deformation.
Young's Modulus of Highstrength Concrete
(Tomosawa, Noguchi, Onoyama, 1990)
StressStrain Models under Monotonic
Loading: The parabolic expression up to the
maximum stress followed by straight
descending branch was used by E. Hognestad
(1951) to represent the uniaxial stressstrain
relation of the concrete. He suggested the use
of 0.85 f’
c
for the concrete strength
o
σ in a
structure, where f’
c
is the compressive strength
of concrete obtained from standard cylinder
tests. The strength of concrete in a real
structure is smaller than the strength of
concrete cylinders. In the descending branch,
the loss of resistance at ultimate strain of
0.0038 was assumed to be 15 % of the
maximum stress.
o c o c o c
o c
o
c
o
c
o c
for Z
for
ε ε ε ε σ σ
ε ε
ε
ε
ε
ε
σ σ
> − − =
≤ − =
)] ( 1 [
] ) ( 2 [
2
where
c
σ : concrete stress,
o
σ : compressive strength of concrete,
c
ε : concrete strain,
o
ε : strain at
compressive strength of concrete, defined as 2 /
o c
E σ .
Smith and Young (1955) proposed an exponential function for the stressstrain relation (
c c
σ ε − )
of concrete under monotonically increasing load;
( )
m c c
o o
K
σ ε
σ ε
=
where,
o
σ : compressive strength obtained from standard cylinders,
0
ε : strain at the peak stress.
Desayi and Krishnan (1964) proposed an expression;
2
0
1 ( )
c
c
c
E ε
σ
ε
ε
=
+
where E : initial tangent modulus,
0
ε : strain at maximum stress.
Saenz (1964) in his discussion to Desayi and Krishnan (1964) suggested the following expression
for ascending branch of the concrete under monotonic loading;
2
0 0
2
0 0
3 2
[1 ( 2) (1 )( ) ]
1 ( 2) ( )
o c o c
c c
c
c
c c
o
E E
E
E E
E
E
E
ε ε
σ ε
ε ε
ε
σ
ε ε
ε ε
= − − + −
=
+ − +
where
0
ε : strain at maximum stress,
o
E : secant modulus at maximum stress (=
0
o
σ
ε
), E : initial
tangent modulus.
Saenz (1964) also suggested an expression including descending branch;
0.15
o
σ
0.0038
0.002
o
ε =
2
[2( ) ( ) ]
c c
c o
o o
ε ε
σ σ
ε ε
= −
Linear
2
o
o
c
E
σ
ε =
tan
c
E α =
o
σ
c
σ
c
ε
Concrete Model (Hognestad, 1951)
2 3
2
0 0 0 0 0
0
2
0 0
1 2 2 1
( 1)
1
( 1)
c
c
c c c
E
E E E
E f f
o
E f o
o f
A B C D
where
R R R R
A B C D
E R R R
R R
E
R R R R E
R R E
ε
ε ε
ε
σ
ε ε ε
σ σ ε σ ε
ε
σ σ
σ ε ε
=
+ + +
+ − −
= = = − =
−
= − = = = =
−
,
f f
σ ε : stress and strain at failure. The parameters , , , A B C D were selected to satisfy strains and
stresses at the origin (0,0), maximum stress (
0 0
, ε σ ) and failure point ( ,
f f
ε σ ), initial tangent
stiffness
0
d
d
ε
σ
ε
=
(= E ), slope
o
d
d
ε ε
σ
ε
=
(=0.0) at the maximum stress point (
0 0
, ε σ ).
Kent and Park (1971) modified the model by Hognestad (1951), and proposed to vary the
stiffness of descending branch taking into account the confining effect of concrete by lateral
reinforcement;
o c
o c o c o c
o c
o
c
o
c
o c
but
for Z
for
σ σ
ε ε ε ε σ σ
ε ε
ε
ε
ε
ε
σ σ
2 . 0
)] ( 1 [
] ) ( 2 [
2
≥
> − − =
≤ − =
h
s h
o
o
u
o h u
s
b
p
Z
"
)
4
3
(
89 . 6
002 . 0 021 . 0
5 . 0
50
50
50 50
=
+
+
=
− +
=
ε
σ
σ
ε
ε ε ε
where
s
p : ratio of volume of transverse reinforcement to volume of concrete core measured to
outside of hoops, b": width of confined core measured to outside of hoops,
h
s : spacing of hoops. The
strain
o
ε at maximum stress
o
σ is taken as 0.002.
Attard and Setunge (1996) proposed a stressstrain curve model for concrete applicable for
concrete strength range from 20 to 130 MPa. The main parameters employed to establish the
equation are Young’s modulus
c
E , peak stress
o
σ , strain at peak stress
o
ε , and stress
i
σ and
i
ε at the inflection point on the descending branch of the stressstrain curve;
2
2
( ) ( )
1 ( 2)( ) ( 1)( )
c c
c o o
c c
o
o o
A B
A B
ε ε
σ ε ε
ε ε
σ
ε ε
+
=
+ − + +
To allow for the difference between the insitu uniaxial compressive strength and the cylinder strength,
20
ε
50c
ε
50u
ε
0.002
o
ε =
o
σ
0.5
o
σ
0.2
o
σ
50h
ε
c
ε
c
σ
Confined
Unconfined
Park and Kent Model, 1971
the peak stress
o
σ may be taken as 0.9 times the cylinder strength. For the ascending branch of the
stressstrain curve,
2
( 1)
1
0.55
c o
o
E
A
A
B
ε
σ
=
−
= −
and for descending branch;
2
( )
( )
0
i i o
o i o i
A
B
σ ε ε
ε ε σ σ
−
=
−
=
The values , , ,
c o i i
E ε σ ε may be determined from:
0.52
0.75
4370 ( )
4.11( ) /
1.41 0.17ln( )
2.50 0.30ln( )
c o
o o c
i
o
o
i
o
o
E
E
σ
ε σ
σ
σ
σ
ε
σ
ε
=
=
= −
= −
where stresses and Young’s modulus are in MPa.
There have been many research works leading to proposals of mathematical or phenomenological
models for concrete under shortterm uniaxial monotonic loading; e.g., Sargin (1971), Popovics
(1970), and Buyukozturk et al. (1971).
Creep and Shrinkage: Concrete, subjected
to constantamplitude longterm loading,
continues to deform with age after
instantaneous elastic strain, phenomenon of
which is called "creep." The rate of strain
decreases with time; creep strain is
stabilized in three to four years. The creep
strain is influenced by (a) stress level, age of
first loading, (b) curing condition, (c) mix of
concrete, (d) shape and size of specimens,
and (e) reinforcement. Concrete without any
loading shrinks with time with loss of water
content. Shrinkage and creep strains cannot
be separated.
Rusch (1960) conducted long term
loading tests on unconfined concrete. The
stressstrain relationship is influenced by
the loading rate. The strength of concrete
under longterm loading is less than that
under short tem loading. The tests
showed that sustained load strength of a
concentrically loaded concrete specimen
amounts to at least 75 percent, and on the
average to about 80 percent of the
strength determined in a short term test. If
the axial stress exceeding approximately
80 percent of the compressive strength is
applied over a long period, creep causes
compression failure of the concrete.
Strain rate effect increases resistance
and stiffness of concrete, but the effect is
relatively smaller compared with that on
steel.
Effect of Load Intensity and Duration (Rusch, 1960)
Strain, mm/mm
C
o
m
p
r
e
s
s
i
v
e
s
t
r
e
s
s
,
k
g
f
/
c
m
2
Quasistatic
Loading Rate Effect on Stressstrain
relation (Rusch, 1960)
Creep curve for concrete with
Behavior under Stress Reversals: Low cycle fatigue tests at the University of Colorado (Sinha,
Gerstle, and Tulin, 1964) on 6x12in.
(150x300mm) standard compression
cylinders led to the following
conclusions;
(a) The stressstrain relationships of
concrete under compressive load
histories possess an envelope curve,
which may be considered unique and
identical with the stressstrain curve
obtained under constantly increasing
strain.
(b) The stressstrain relationships of
concrete subjected to cyclic loading
possess a locus of common points
which are defined as the point where the
reloading portion of any cycle crosses the unloading portion. Stresses above the common points
produce additional strains, while stresses at or below these points will result in the stressstrain path
going into a loop, repeating the previous cycle
without further permanent strain. It was also
observed that the values of the common points
depended on the minimum stress in the cycle;
i.e., the stress amplitude.
Karsan and Jirsa (1969) reported the
cyclic compressive tests on 3x5in.
(76x127mm) prisms with flared ends;
(a) The envelope curve of stress strain
paths under cyclic loading coincided with the
stressstrain curve for a specimen under
monotonic loading to failure.
(b) The SmithYoung expression (Smith
and Young, 1955) was found to be a good
approximation of the envelope curves.
(c) The accumulation of strain under
constant maximum stress levels produced
failure when the envelope curve was reached;
the specimen, however, could be loaded to the
envelope curve regardless of the strain
accumulation prior to a given cycle. The strain
accumulation did not appear to reduce the
strength to a level below the envelope.
(d) The location of the common points
(intersection of unloading and reloading
curves) was dependent on the magnitude of
the maximum stress and strain of the previous
load cycle. The common points for loading
from nonzero levels were identical to the
common points corresponding to load cycles
starting at a stress level of zero.
(e) Examination of the location of the
common points shows that failure would be
produced under repeated loads with stresses
exceeding about 0.63 f ’c, the maximum of the
stability limit. This limit was independent of the
minimum stress levels in the cycles.
Stressstrain relationship under stress cycles
Peaks of cycles and envelope curve
(K d Ji 1969)
Variation of common points
(f) Loading and unloading curves starting from a point within the stressstrain domain were not
unique, and the value of stress and strain at the peak of the previous loading cycle must be known to
estimate the response.
Uniaxial Hysteresis Models for
Concrete: Karsan and Jirsa (1969)
proposed a model for the uniaxial cyclic
behavior of concrete based on 46 short
rectangular column tests under cyclical
loading. The envelope curve could be
defined as the stressstrain curve obtained
under monotonic loading to failure. The
monotonic loading curve may be
approximated by the expression proposed
by Smith and Young (1955);
(1 )
0.85
c
o c c
o o
e
ε
ε
σ ε
σ ε
−
=
where,
o
σ : compressive strength
obtained from standard cylinders,
o
ε :
strain at the peak stress.
The unloading curve is expressed by the seconddegree parabola, which passes the following
three points;
(a) Point ( , )
E E
σ ε on the envelope curve from which the unloading curve or its extension
terminates,
(b) Common point ( , )
C C
σ ε , and
(c) Plastic strain point (0.0, )
P
ε at which the unloading curve or its extension terminates.
The reloading curve is also expressed
by the second order parabola which
passes the following three points;
(a) Plastic strain point (0.0, )
P
ε at
which the reloading curve or its extension
starts,
(b) Common point ( , )
C C
σ ε , and
(c) Point ( , )
E E
σ ε on the envelope
curve at which the reloading curve or its
extension reaches the envelope curve.
The unloading and reloading curves
intersect at the common point ( , )
C C
σ ε ,
which lowers the number of loading
cycles. The upper and lower limit curves
of the locus of common points ( , )
C C
σ ε are called common point limit and the stability limit curves
as follows;
( )
[1 ]
0.315 0.77
( )
0.315 0.77
C
o
C
C o
o
e
ε
ε
β
ε
σ ε
β
σ β
−
+
=
+
where, 0.76 β = for the common point limit curve, and 0.63 β = for the stability limit point curve.
The plastic strain point is defined as follows;
Calculated stress strain relationship
Calculated stressstrain relationship
2
0.093 ( ) 0.091( )
P E E
o o o
ε ε ε
ε ε ε
= + for loading, and
2
0.145 ( ) 0.13 ( )
P E E
o o o
ε ε ε
ε ε ε
= + for unloading.
Park, Kent and Sampson (1971)
showed a simple model for the uniaxial
stressstrain relationship under cyclic loading
in the analysis of reinforced concrete
members. The stressstrain curve for
concrete is represented by a parabola for
ascending portion and straight line for
descending portion for monotonically
increasing strain (Kent and Park, 1971). A
linear stressstrain curve for concrete in
tension may be assumed to the tensile
strength. The curve under cyclic loading is
represented by straight lines. Upon
unloading from point E on the skeleton curve,
0.75 of the previous stress is lost without decrease in strain, whereupon a linear path of slope 0.25
c
E
is followed to point G. If the concrete has not cracked, it is capable of carrying tensile stress to point
K; but if the concrete has previously cracked, or if cracks form during this loading stage, the tensile
strains increase but no tensile stress develops. Upon reloading, the strain must regain the value at G
before compressive stress can be sustained again. If reloading commences before unloading
produces zero compressive stress, reloading follows one of the paths IJ. The average slope of the
assumed loop between E and G is parallel to the initial tangent modulus.
Darwin and Pecknold (1974) used Saenz’s equation (1964) for the monotonic ascending branch;
2
1 [ 2]( ) ( )
c
c
c c
o o o
E
E
E
ε
σ
ε ε
ε ε
=
+ − +
where E : tangent modulus of
elasticity at zero stress,
o
E :
secant modulus at the point of
maximum compressive stress,
o
ε : equivalent uniaxial strain at
the maximum compressive
stress. The falling branch after
attaining the maximum stress
expressed by a straight line
passing through maximum
stress point ( , )
o o
σ ε and point
(0.2 , 4.0 )
o o
σ ε .
Straight unloading stiffness
changes its slope at the turning
point; the initial unloading stiffness from the envelope curve is equal to the initial elastic tangent
stiffness. Straight reloading stiffness is parallel to the lower unloading stiffness and passes through
the common points. The stress levels of common points and turning points are given below;
P k K t d S 1974
Darwin and Pecknold model of concrete (1974)
Region 1:
1 1
1 1
5
6
1
2
cp en
tp en
σ σ
σ σ
=
=
Region 2:
2 2 2
2 2
1 1
min{ , }
6 6
1 1
min{ , }
2 2
cp en en B
tp en B
σ σ σ σ
σ σ σ
= −
=
Region 3:
3 3
3 3 3 3
3
1
6
2( )
1
3
cp en B
tp en en cp
en B
σ σ σ
σ σ σ σ
σ σ
= −
= − −
= −
Region 4:
4 4
4
2
3
1
3
cp en
tp en
σ σ
σ σ
=
=
Other models can be found in literatures by Blakely (1973) and Aoyama (1973).
Tensile Behavior: Tensile strength of concrete is
obtained by (a) splitting test of concrete cylinders
or (b) modulus of rupture test of concrete prism.
The splitting tensile test develops relatively
uniform tensile stress over the section of the
concrete cylinder. Tensile strength varies by
testing methods (pure tension, splitting and
modulus of rupture tests); splitting tensile
strength of concrete is often used in the
laboratory and is approximately 10 % of the
compressive strength for normal strength
concrete. Tensile strength of concrete in a real
structure or specimen is difficult to estimate
because the concrete is subjected to shrinkage
strain and accidental loading.
Stressstrain relation in tension is almost
linear up to the development of cracking; the
elastic strain is developed over the entire height
of the concrete specimen.
Large micro cracks concentrate near the
failure section. Once crack develops at the failure
plane, elastic deformation is released in slightly
damaged regions and crack width continues to
grow at the failure plane (Comite
EuroInternational du Beton, 1996).
Reinhardt and his research group presented
detailed description of stressstrain hysteresis
behavior of concrete under tensile stress reversal
(Reinhardt et al., 1986, Yankelevsky and Reinhardt,
1989). Various other hysteresis models have been
presented for concrete in tension (Hillerborg et al.,
1976 and Duda, 1990). These hysteresis models for
concrete in tension are important in the finite element
analysis of reinforced concrete structures.
Development of Tensile Cracks
(Comite EuroInternational du Beton, 1996)
Concrete under Tensile Stress Reversals
(Yankelevsky and Reinhardt, 1989)
Confining Effect: The strength
Confining Effect: The strength
and ductility of concrete increases
when subjected to lateral confining
pressure. Based on experimental
research, Richart, Brandtzaeg
and Brown (1928) proposed the
following expression for the
strength of concrete under
confining pressure;
' ' 4.1
cc c l
f f f = +
where '
cc
f : axial compressive
strength of confined concrete,
'
c
f : uniaxial compressive strength
of unconfined concrete, and
l
f :
lateral confining pressure.
Behavior Under Biaxial Stress Condition: Concrete compressive strength increases with lateral
compressive (confining) stress, but decreases with
lateral tensile stress under bidirectional stress state
(Kupfer, Hilsdorf and H. Rusch, 1969).
The modeling of stressstrain relationship under
biaxial and triaxial loading has been studied
extensively with the development of nonlinear finite
element analysis methods.
Mohr’s theory of failure is often used to estimate
the strength under combined normal and shear
stresses on a plane. The envelope of the failure
Mohr’s circle is often called failure envelope.
Modeling of Concrete Behavior under Multiaxial
Stresses: A mathematical model to represent
inelastic behavior of concrete under stress reversal
should be able to reproduce the following aspects
(Comite EuroInternational du Beton, 1996);
(a) The capacity of accounting for inelastic
nonproportional unloading and reloading,
(b) An adequate nonholonomic relationship between the state of stress and the stiffness of the
material.
(c) The capacity of accounting for the stress degradation as a function of the load history,
including postpeak behavior.
The above features are desired in the mathematical model, but the overall behavior of a reinforced
concrete structure may be insensitive to some aspects of the material behavior. The desirability of a
model is dependent on a type of problem in pursuit.
A large cooperative experimental research project (Gerstle et al. 1980) conducted under unified
testing procedures reported the following conclusions;
(a) Considerable scatter of measured strain were observed for concrete under applied multiaxial
stresses. Probabilistic methods are needed for more meaningful treatment of data.
(b) The behavior of concrete can be represented conveniently in terms of the octahedral normal
(hydrostatic) and shear (deviatoric) stresses
0
σ and
0
τ . The assumption of isotropy is not
substantially violated at a macroscopic level up to stress close to failure.
Biaxial Strength of Concrete
(Kupfer Hilsdorf and H Rusch 1969)
Axial StressStrain Curve with Lateral Confining Pressure
0 1 2 3
2 2 2
0 1 2 2 3 3 1
0 1 2 3
2 2 2
0 1 2 2 3 3 1
1
( )
3
1
( ) ( ) ( )
3
1
( )
3
1
( ) ( ) ( )
3
σ σ σ σ
τ σ σ σ σ σ σ
ε ε ε ε
γ ε ε ε ε ε ε
= + +
= − + − + −
= + +
= − + − + −
(c) The direct relationships between volumetric and deviatoric stresses and strains are governed
by the bulk modulus K
s
and shear modulus G
s
.
0 0
0
0
0 0
0
0
( )
( ) ( : )
3 3(1 2 )
( )
( ) ( : )
2 2(1 )
s
s
E
K elastic case
E
G elastic case
σ ε
ε
ε ν
τ γ
γ
γ ν
= =
−
= =
+
(4) A coupling effect between octahedral shear stress and volumetric strain is systematically
observed and can be described by introducing a coupling modulus H
s
.
0
0
s
H
τ
ε
=
Three stress and strain dependent moduli appear to be sufficient to describe the behavior of
concrete under monotonically increasing loads.
Models based on the theory of elasticity (Elwi and Murray, 1979, Stankowski and Gerstle, 1985,
Shafer and Ottosen, 1985, Buyukozturk and Shareef, 1985), models based on the theory of plasticity
(Han and Chen, 1987), and models based on the mechanism of microcracking or elastic damage
(Dougill, 1976, Resende and Martin, 1984, Krajcinovic and Fonseka, 1981, Mazars, 1984) are
introduced in Comite EuroInternational du Beton Report (1996).
References:
Attard, M. M., S. Setunge, “The Stress Strain Relationship of Confined and Unconfined Concrete,”
Material Journal, American Concrete Institute, Vol. 93, No. 5, 1996, pp. 432  442.
Blakeley, R. W. G., et al., “Prestressed Concrete Sections with Cyclic Flexure,” Journal, Structural
Division, ASCE, Vol. 99, No. ST8, August 1973, pp. 1717  1742.
Buyukozturk, O., et al., “Stressstrain Response and Fracture of a Concrete Model in Biaxial
Loading,” Journal, American Concrete Institute, Vol. 68, No. 8, August 1971, pp. 590  599.
Buyukozturk, O. and S. S. Shareef, “Constitutive Modeling of Concrete in Finite Element Analysis,”
Computers and Structures, Vol. 21, No. 3, 1985.
Comite EuroInternational du Beton, “Concrete under Multiaxial States of Stress, Constitutive
Equations for Practical Design, CEB, Lausanne, 1983, Bulletin d’Information No. 156.
Comite EuroInternational du Beton: RC Elements under Cyclic Loading  State of the Art Report,
Thomas Telford, 1996, 190 pp.
Darwin, D. and D. A. W. Pecknold, “Inelastic Model for Cyclic Biaxial Loading of Reinforced
Concrete,” Structural Research Series No. 409, Department of Civil Engineering, University of
Illinois, UrbanaChampaign, July 1974.
Desayi, P., and S. Krishnan, “Equation for the Stressstrain Curve of Concrete“, Journal, American
Concrete Institute, Vol. 61, No.3, March 1964, pp. 345  .
Dougil, J. W., On Stable Progressively Fracturing Solids,” Zeitschrift fur Angewandte Mathematik und
Physik, Vol. 27, Fasc. 4, 1976, pp. 423  437.
Duda, H., “Bruchmechanincsche Verhalten von Beton unter monotoner und zyklischer
Zugbeanstruchung,” Doctoral Thesis, Technical Hochschule Darmstadt, 1990.
Elwi, A. A., and D. W. Murray, “A 3D Hypoelastic Concrete Constitutive Relationship, Journal,
Engineering Mechanics Division, ASCE, Vol. 105, August 1979.
Gerstle, K. H., H. Aschl, R. Bellotti, P. Bertacchi, M. D. Kotsovos, H. Y. Ko, D. Linse, J. B. Newman, P.
Rossi, G. Schickert, M. A. Taylor, L. A. Traina, H. Winkler and R. M. Zimmerman, “Behavior of
Concrete under Multiaxial Stress States,” Journal, Engineering Mechanics, ASCE, Vol. 106, No.
6, December 1980, pp. 1383  1404.
Han, D. J., and W. F. Chen, “A Nonuniform Hardening Plasticity Model for Concrete Materials,”
Journal, Mech. Mat., Vol. 4, 1985.
Hillerborg, A. et al., “Analysis of Crack Formation and Crack Growth in concrete by Means of Fracture
Mechanics and Finite Elements,” Cement and Concrete Research, Vol. 6, 1976, pp. 773  782.
Hognestad, E., "A Study of Combined Bending and Axial Load in Reinforced Concrete Members,"
Bulletin No. 399, Engineering Experimental Station, University of Illinois, 1951.
Karsan, I. D., and J. O. Jirsa, “Behavior of Concrete under Compressive Loadings,” Journal,
Structures Division, ASCE, Vol. 95, No. ST12, December 1969, pp. 2543  2563.
Kent, D. C., and R. Park, "Flexural Members with Confined Concrete," Journal, Structural Division,
ASCE, Vol. 97, ST 7, July 1971, pp. 19691990.
Krajcinovic, D., and G. U. Fonseka, “The Continuous Damage Theory of Brittle Materials,” Journal,
Applied Mechanics, ASME, Vol. 48, 1981.
Kupfer, H., H. K. Hilsdorf, and H. Rusch, "Behavior of Concrete Under Biaxial Stress," Journal,
American Concrete Institute, Vol. 66, No. 8, pp. 656666, August 1969.
Mander, J. M., N. M. Priestley and R. Park, “Theoretical StressStrain Model for Confined Concrete,”
Journal, Structural Division, ASCE, Vol. 114, No. 8, 1988, pp. 1804  1826.
Mazars, J., “Description of the Multiaxial Behavior of Concrete under Multiaxial Conditions,”
INSAOPS Toulouse, May 1984.
Park, R., and T. Paulay, Reinforced Concrete Structures, John Wiley & Sons, 1975, 769 pp.
Pauw, A., "Static Modulus of Elasticity of Concrete as Affected by Density," Journal, American
Concrete Institute, No. 57, No. 6, December 1960, pp. 679687.
Popovics, S., “Stressstrain Relations for Concrete under Compression,” Journal, American Concrete
Institute, Vol. 67, No. 3, March 1970, pp. 243  248.
Popovics, S., “A Review of StressStrain Curve of Concrete,” Cement and Concrete Research, Vo. 3,
No. 4, September 1973, pp. 583  599.
Reinhardt, H. W. et al., “Tensile Tests and Failure Analysis of Concrete,” Journal, Structural
Engineering, ASCE, Vol. 112, No. 11, November 1986, pp. 2462  2477.
Resende, L., and J. B. Martin, “A Progressive Damage Continuum Model for Granular Materials,”
Comp. Meth. Appl. Mech. Engng, Vol. 42, 1984.
Richart, F. E., A. Brandtzaeg, and R. L. Brown, “A Study of the Failure of Concrete Under Combined
Compressive Stresses,” University of Illinois Engineering Experimental Station, Bulletin No. 185,
1928, 104 pp.
Rusch, H., “Researches Toward a General Flexural Theory for Structural Concrete,” Journal,
American Concrete Institute, Vol. 57, No. 1, July 1960, pp. 1  28.
Saenz, L. P., “Discussion of the paper by Desayi, P., and S. Krishnan entitled ‘Equation for the
Stressstrain Curve of Concrete’,” Journal, American Concrete Institute, Vol. 61, No. 9,
September 1964, pp. 1229  1235.
Sargin, M., “Stressstrain Relationships for Concrete and the Analysis of Structural Concrete
Sections,” Study No. 4, Solid Mechanics Division, University of Waterloo, Ontario, 1971, 167 pp.
Sinha, B. P., K. H. Gerstle, and L. G. Tulin, "StressStrain Relationships for Concrete Under Cyclic
Loading," Journal, American Concrete Institute, Vol. 61, No. 2, pp. 195211, February 1964.
Smith, G. M., and L. E. Young, “Ultimate Theory in Flexure by Exponential Function,” Journal,
American Concrete Institute, Vol. 52, No. 3, November 1955, pp. 349  359.
Shafer, G. S., and N. S. Ottosen, “An Invariantbased Constitutive Model, Structural Research Series
No. 8506, Department of Civil Environmental and Architectural Engineering, University of
Colorado, Boulder, 1985.
Stankowski, T., and K. H. Gerstle, “Simple Formulation of Concrete Behavior under Multiaxial Load
Histories,” Journal, American Concrete Institute, Vol. 82, No. 2., MarchApril 1985.
Tomosawa, F., T. Noguchi, and K. Onoyama, "Investigation on Fundamental Mechanical Properties of
Highstrength and Super High Strength Concrete (in Japanese)," Summaries of Technical Papers
of Annual Meeting of Architectural Institute of Japan, Vol. A, 1990, pp. 497498.
Yankelevsky, D. Z. and H. W. Reinhardt, “Uniaxial Behavior of Concrete in Cyclic Tension,” Journal,
Structural Engineering, ASCE, Vol. 115, No. 1, 1989, pp. 166182.
2.2 Reinforcing Steel
Stressstrain relation of a reinforcing bar under monotonically increasing tensile load shows a
linearly elastic region, yield plateau, strainhardening region before fracture. Elastic modulus E
s
of
steel is approximately 210 GPa and Poisson's ratio µ is approximately 1/3. High strength steel or
workhardened steel does not exhibit clear yield point nor yield plateau; a yield point is defined as the
intersection of a straight line parallel to the initial stiffness at 0.2 % strain and the observed
stressstrain curve.
Stressstrain relation of reinforcing steel
In reinforced concrete building construction,
yield strength of reinforcing bars ranges from
300 to 500 MPa, but higher strength of the
order of 700 to 1200 MPa is used for the lateral
reinforcement. Higher strength steel is
commonly used in prestressed concrete
construction.
With the strength of steel, yield plateau
becomes narrower, strain at fracture becomes
smaller, and the yield ratio (the ratio of the yield
stress to the tensile strength) decreases.
Properties of steel in compression are similar to
that in tension unless buckling of a bar does not
occur.
High rate of loading can significantly
increase the upper yield stress especially at a
rate of loading for blast. The stress increase in
the strainhardening is relatively small. The
strainrate properties were studied for blast
loading cases. Until the servocontrolled
actuators were developed, the wellcontrolled
testing and instrumentation was difficult for
reliable data.
When stress is reversed after plastic deformation in a direction, the range of linearly elastic
behavior is reduced, and stressstrain relation departs from linearly elastic response. This
phenomenon is called "Bauschinger effect." The stress after a stress reversal beyond initial yield
stress is increased compared to the stress at the same strain; the phenomenon is called “isotropic
strain hardening.”.
The range of cyclic strain history, to which reinforcing bars are likely to be subjected, differs
significantly from that of structural steel members; i.e., compressive strains are not as large as tensile
strains. Under a flexural condition, reinforcing bars resist tensile stresses after concrete cracking,
while compressive stress is resisted by concrete. When concrete cover spalls after crushing of
concrete, reinforcing bars may yield in compression. The Bauschinger effect is important in
simulating the behavior of reinforced concrete especially when the compression stresses are resisted
solely by reinforcing bars during crack opening immediately after unloading.
Any stressstrain relation under
repeated and reversed loading can be
decomposed into three parts, (a) skeleton
part, (b) unloading part and (c) softening
part. Kato Akiyama and Yamanouchi
(1973) demonstrated that if a series of
progressively larger hysteretic loops are
known for a given material, a monotonic
stressstrain curve can be constructed
with a reasonable degree of accuracy. by
plotting the stressstrain relations for
positive and negative directions
separately. In a cluster of cyclic hysteresis
loops, the bold lines are defined as the
skeleton parts, which are the portions of
the curves at stresses of the same sign
larger than the ones during the previous
cycles. Fine straight lines are the
unloading parts and the dashed lines are
softening parts in which the Bauschinger
effect is dominant. Connecting the
skeleton lines end to start, either above or below the horizontal axis given curves similar to those
found from monotonic experiments except for the first yielding region; i.e., any skeleton line is always
a part of the monotonic stressstrain curve. Any unloading line can be approximated by a straight line.
Reloading curve after unloading from the skeleton curve in the opposite direction can be expressed
by
Reloading stiffness after unloading from the skeleton
T
e
n
s
i
l
e
s
t
r
e
s
s
,
k
g
f
/
c
m
2
Room temperature
Strain, mm/mm
( )( 1) (1 ) 0
B
s s
B
E
a a a a
E
a
E E
σ
ε
σ σ
− + − − − =
=
−
where E : Young’s modulus,
B
E : secant modulus on the monotonic loading curve at stress level
s
σ and strain
B
ε as shown below. The expression for a was obtained by equating the tangent
stiffness at zero stress level to be equal to the Young’s modulus. The secant stiffness
B
E can be
empirically expressed as
10
log 10
6
B s
E
E ε = −
where
s
ε : accumulated skeleton strain.
The stressstrain relationship of metal is simulated by Ramberg and Osgood (1943) for the
skeleton and reloading cases;
1
1
0 0 0
(1 )
(1 )
2 2 2
r
y y y
r
y y y
ε σ σ
ε σ σ
ε ε σ σ σ σ
ε σ σ
−
−
= +
− − −
= +
Park, Kent and Sampson (1972) proposed the modification of the RambergOsgood model as
follows:
1
1000
(1 )
0.744 0.71
{ 0.2411}
ln(1 1000 ) 1
ip
r
s s
s si
s ch
ch sy
ip
E
e
ε
σ σ
ε ε
σ
σ σ
ε
−
− = +
= − +
+ −
n even for
e n
r
n odd for
e n
r
n
n
04 . 3
) 1 (
469 . 0
) 1 ln(
20 . 2
297 . 0
) 1 (
03 . 6
) 1 ln(
49 . 4
+
−
−
+
=
+
−
−
+
=
where
s
ε : steel strain,
si
ε : steel strain at zero
stress,
s
σ : steel stress, E
s
: elastic modulus of
steel,
ch
σ : stress dependent on the yield
strength and plastic strain in the steel produced in
the previous loading run, and γ : parameter of
the RambergOsgood model,
ip
ε : plastic strain
in steel produced in previous loading run, n:
number of post yield loading runs with n = 0 for the first yielding. Unloading stiffness was taken equal
to the initial elastic stiffness.
A simple nonlinear stressstrain model of reinforcing bars is proposed by Zulfiqar and Filippou
(1990). The proposed steel stressstrain relation under monotonically increasing load consists of
three regions; (a) a linearly elastic region, (b) a plastic yield plateau, and (c) a nonlinear
strainhardening range.
0
1 1
2 2
( )
( )
s s s y
s y s y y s sh
s y s sh sh s
E for
E for
E for
σ ε ε ε
σ σ ε ε ε ε ε
σ σ ε ε ε ε
= ⋅ ≤
= + ⋅ − < ≤
= + ⋅ − <
The behavior under load reversals is expressed as
follows;
*
*
* *
*
0
1
1
1
ln[1 (1 ) ]
1
S
E S
S
E S
E S
E S
e
e
e
E E
e
λε
λ
λ
λ
σ σ
σ
σ σ
ε ε
ε σ
ε ε λ
ε ε λ
σ σ
−
−
−
−
− −
= =
− −
−
= = − − −
−
−
= =
− −
where, parameter λ may be calculated for the
third equation for given stresses
S
σ ,
E
σ and
strains
S
ε ,
E
ε at the two outer most unloading
points.
A summary of steel behavior is reported in
Bulletin d’Information No. 161 from Comite EuroInternational du Beton, August 1983. Additional
model can be found in literatures (Peterson and Popov, 1977, Ma, Bertero and Popov, 1976).
References:
Aktan, A. E., B. I. Karlsson, and M. A. Sozen, “Stressstrain Relationships of Reinforcing Bars
subjected to Large Strain Reversals,” Civil Engineering Studies, Structural Engineering Series No.
397, University of Illinois at UrbanaChampaign, 1973.
Comite EuroInternational du Beton, “Response of RC Critical Regions under Large Amplitude
Reversed Actions,” Paris, Bulletin d’Information No. 161, August 1983.
Kato, B., H. Akiyama and H. Yamanouchi, “Predictable Properties of Material under Incremental
Cyclic Loading,” Symposium on Resistance Ultimate Deformability of Structures Acted on by
Stress strain relation under monotonic loading
Stressstrain relation under stress reversals
WellDefined Repeated Loads, Reports of Working Commissions, Vol. 13, International
Association for Bridge and Structural Engineering, Lisbon, 1973.
Kent, D. C., and R. Park, “Cyclic Load Behavior of Reinforcing Steel,” Strain, British Society for Strain
Measurement, Vol. 9, No. 3, July 1973, pp. 98  103.
Ma, S. Y., V. V. Bertero and E. P. Popov, “Experimental and Analytical Studies on the Hysteretic
Behavior of Reinforced Concrete Rectangular and TBeams,” Report No. EERC 762, Earthquake
Engineering Research Center, University of California at Berkeley, May 1976.
Park, R., D. C. Kent and R. A. Sampson, "Reinforced Concrete Members with Cyclic Loading,"
Journal, Structural Division, ASCE, Vol. 98, ST 7, pp. 13411360, July 1972.
Peterson, H., and E. P. Popov, “Constitutive Relations for Generalized Loadings,” Journal,
Engineering Mechanics Division, ASCE, August 1977.
Ramberg, W. and W. R. Osgood, "Description of StressStrain Curves by Three Parameters,"
Technical Note No. 902, National Advisory Committee for Aeronautics, July 1943.
Zulfiqar, N., and F. C. Filippou, “Models of Critical Regions in Reinforced Concrete Frames under
Earthquake Excitations,” Earthquake Engineering Research Center, Report EERC 9006,
University of California, Berkeley, May 1990.
2.3 Bond
A mechanism of stress transfer between a reinforcing bar and surrounding concrete in the form
shear stress on reinforcement surface is called "bond," developed by (a) chemical adhesion between
mortar paste and bar surface, (b) friction on rough steel surface and (c) mechanical interlocking of
ribs with surrounding concrete.
Adhesion bond is known to break at an earlier
stage of relative deformation (slippage) of a bar
and concrete. Major source of resistance after
initial bar slippage is the mechanical interlocking.
At the inclined face of bar deformation, stress is
transferred by direct contact (bearing) stress
normal to the face and friction stress along the
face.
When the ribs on a deformed bar is high and
spaced too closely, the shear failure occurs along
the cylindrical concrete surface connecting the
top of adjacent ribs.
When the rib spacing is larger, the concrete crushes at the rib face, and crushed concrete forms a
wedge in front of the rib; hence surrounding
concrete is pushed outward by force in the
reinforcing bar, and tensile force, called "ring (or
circumferential) tension", is developed in the
circumferential direction of the bar face.
Therefore, surrounding concrete is split by this
ring tension stress.
Bond stress (average shear stress over bar
surface) freeend slip relation is shown below
for a pullout test, in which concrete strength
was 25 MPa. Deformed bar D16 and round bar
16φ were used in the test. Maximum bond
stress is developed at approximately 0.1 to 0.3
mm slip deformation.
The bond strength may be affected by (a)
concrete strength, (b) deformation of a
reinforcing bar surface, (c) bar size (Poisson's
effect), (d) bar stress (tension or compression),
(e) bar location, (f) concrete cover depth over a
bar, (g) lateral reinforcement, (h) lateral
pressure, and (i) anchorage length.
Eligehausen et al. (1983) suggests bond stressbar slip relation under monotonically increasing
loading;
1 1
1
1 1 2
3 1
1 2 2
3 2
( )
( )
s
for s s
s
for s s s
s s for s s
s s
α
τ τ
τ τ
τ τ
τ τ
= ≤
= < ≤
−
= + − <
−
The parameters are defined in the figure and table.
Cycles with reversed loading produce degradation of bond strength and bond stiffness. The rate of
degradation is more pronounced under reversed loading than under unidirectional loading.
Degradation depends on the previous peak slip in either direction, the number of cycles and the
difference of the peak slips in the two directions. Significant strength decay is observed with slip
amplitude under reversed loading. If the peak bond stress during cycling does not exceed
approximately 70 percent of monotonic bond strength, the bond stressslip relation is stable.
Bond stressslip relation under onesided slip
Bond stress slip relation under
monotonic loading
The deterioration of stiffness under
monotonically increasing loading is
described by Eligehausen et al.
(1983). At low bond stress (Point A),
inclined cracks begin to propagate
from the top of the ribs; their growth
and size are controlled by the
confining pressure provided by
transverse reinforcement. Transfer of
forces is more by bearing with a
shallow angle of inclination.
Increasing the load, local crushing of
concrete in front of the lugs produces
reduction of the tangent of the bond
stressslip curve (Point B). When the
maximum bond stress is attained the
concrete key is sheared off, forming a
cone with a length of about four times
the lug height. With increasing slip the
bond stress begins to drop slowly. As
the bond shear cracks reach the
bottom of the adjacent lug (Point D)
the bond stress begins to drop, and by
the time the slip reached the lug
spacing only the frictional component
remains (Point E).
The deterioration of stiffness under
reversed loading is also described by
Eligehausen et al (1983). The initial
loading follows the monotonic curve,
but the cyclic load behavior is
sensitive to the level of slip at which
the reversal occurs.
Case 1: If a slip reversal is imposed before inclined cracking, unloading branch exhibits high
stiffness (path AF) because only a small part of the slip is caused by inelastic concrete deformation.
As soon as slip in the opposite direction is imposed, the friction branch is reached (path FH). The
slope of this portion of the curve is small because the surface of the concrete surrounding the bar is
smooth. As soon as the cracks close, the stiffness differs little from that of the monotonic envelope
(point I). Unloading from point I, where the slip in the two directions is about equal, the curve (path
IKL) is very similar to that from the initial unloading curve (path AFH). The major difference is that due
to previous cracking and crushing of the concrete in front of the ribs, the point where the bond
stresses begin to pick up again (point L) will be shifted to the right of the origin. The lug will not be
bearing fully until point M is reached. Further loading follows the bondslip curve up to the monotonic
envelope.
Case 2: If unloading occurs after the inclined cracks were formed, and therefore near the slip at
which ultimate bond stress has been attained, the unloading path is similar to that of the first case up
to point F. Since there is more damage to the concrete, a higher frictional resistance is mobilized
(point G). When the loading is reversed the lug presses against a key whose resistance has been
lowered by inclined cracks over a part of its length that were induced by the first halfcycle. The
splitting cracks created in the first halfcycle close at a higher load than those of the first case (point
H), and lead to an earlier formation of splitting cracks in the opposite directions. Splitting cracks,
combined with the existing inclined cracks along the bar, result in a reduced envelope (path HI) and a
reduction of bond capacity in the second direction (point I). Unloading from this peak (path IKLMN)
and reversing the load results in a reduced stiffness and strength because only the remaining
uncrushed concrete between the lugs must be sheared off. The bond strength (point N) is
substantially lower than that of point C, and lower than that of point I.
Bond stressslip relation with respect to damage
(Eligehausen, et al., 1983)
Case 3: If unloading occurs after the
slip reached a value much larger than
the slip at maximum strength (point C),
the behavior is very poor. Since more
damage has occurred, the friction
resistance (point G) is larger than for
either of the previous cases. However,
since the concrete between the lugs is
completely sheared very little force can
be transmitted by bond when direction
of loading is reversed (path HIJ).
Unloading and reloading in the opposite
direction (path JKLMN) results in very
little additional bond capacity beyond
that provided by friction since most of
the mechanical anchorage is lost.
The first analytical model of the bond
stressslip relationship for reversed
cyclic loading was presented by Morita
and Kaku (1973). The monotonic
loading envelopes of bilinear type were
varied for loading in compression and
tension. The deterioration of bond
resistance at peak slip and frictional
bond resistance with the number of
cycles is not considered in the model.
Bond stressslip relation with respect to damage
(Reversed loading) (Eligehausen, et al., 1983)
Additional models for bond stressbar slip hysteresis models can be found in Refs. Viwathanatepa,
S. and et al. (1979), Hawkins et al. (1982) and Balaz (1989).
References:
Balazs, G. L., “Bond Softening under Reversed Load Cycles,” Stui e Ricerche  Corso Flli. Presenti,
Pollitecnico di Milano, No. 11, 1989, pp. 503  524.
Eligehausen, R., V. V. Bertero, and E. Popov, “Local Bond Stressslip Relationships of Deformed
Bars under Generalized Excitations,” Earthquake Engineering Research Center, Report No.
EERC 8323, University of California, Berkeley, October 1983.
Hawkins, H. M., et al., “Local Bond Strength of Concrete for Cyclic Reversed Loadings,” Bond in
Concrete, P. Bartos (ed.), Applied Science Publishers Ltd., London, 1982, pp. 151  161.
Morita, S. and T. Kaku, “Local Bond Stressslip Relationship under Repeated Loading,” Proceedings,
IABSE Symposium on Resistance and Ultimate Deformability on Structures Acted on by Well
Defined Repeated Loads, Lisbon, 1973, pp. 221  227.
Viwathanatepa, S. and et al., “Effect of Generalized loadings on Bond of Reinforcing Bars embedded
in Confined Concrete Blocks,” Earthquake Engineering Research Center, Report No. EERC 79/22,
University of California, Berkeley, August 1979.
Assignment No. 1
20020222
S. Otani
When a concrete specimen is tested under monotonically increasing deformation, the descending
part of the stressstrain relationship is difficult to obtain due to the sudden release of elastic energy by
the frame of a testing machine.
In order to understand this phenomenon, consider the following mathematical model, in which a
concrete cylinder is subjected to forced displacement through a linearly elastic spring representing
the stiffness of the testing machine.
The forced displacement is applied to the system at a constant rate of 0.1 mm/sec. The
forcedeformation of the concrete specimen is given below. Plot the loaddeformation relation of the
concrete specimen and of the loading point at 0.1 sec interval.
Discuss the force deformation relation of the concrete for the two cases for the stiffness of the
testing machine; elastic stiffness of (a) 100 kN/mm, (b) 150 kN/mm, and (c) 300 kN/mm.
0.5
1.0 3.0
Displacement, mm
300
F
o
r
c
e
,
k
N
Testing
Machine
Concrete
Specimen
Constant
displacement
Solution:
1) Equilibrium of forces: Forces in the concrete cylinder
concrete
P and the frame of the testing
machine
spring
P should be the same, and equal to the force P in the testing machine.
concrete spring
P P P = =
2) Displacement compatibility: The applied displacement D by the testing machine should be equal
to the sum of displacement
concrete
D in the concrete cylinder and displacement
spring
D of the frame.
concrete spring
D D D = +
The stressstrain relation of the concrete is piecewise linear and the stiffness of the testing machine
is linearly elastic. Therefore, we need to consider the three points.
Point A:
300
0.5
300 /
0.5 (300/ )
concrete spring
concrete
spring spring
spring
P P P kN
D mm
D kN k
D k
= = =
=
=
= +
Point B:
300
1.0
300 /
1.0 (300/ )
concrete spring
concrete
spring spring
spring
P P P kN
D mm
D kN k
D k
= = =
=
=
= +
Point C:
0.0
3.0
0.0 /
3.0
concrete spring
concrete
spring spring
P P P kN
D mm
D kN k
D mm
= = =
=
=
=
Displacement at the loading head.
Point A B C
Spring 100 kN/mm 3.5 mm 4.0 mm 3.0 mm (not controllable)
Spring 150 kN/mm 2.5 mm 3.0 mm 3.0 mm (sudden failure)
Spring 300 kN/mm 1.5 mm 2.0 mm 3.0 mm
0.5 1.0
3.0
A
B
C
Concrete
Displacement, mm
300
F
o
r
c
e
,
k
N
Chapter 3 Behavior of Reinforced Concrete Members
3.1 Behavior of Beams
A beam member, normally subjected to bending and shear without axial force, can develop large
deformation capacity and dissipate ample hysteresis energy during an earthquake well in the inelastic
range as long as brittle shear and bond splitting failures are prevented. Four modes of failure should
be considered when the behavior of reinforced concrete members is discussed; i.e., (a) flexural
failure caused by crushing of concrete or buckling of longitudinal reinforcement in compression or
tensile fracture of longitudinal reinforcement, (b) shear failure in diagonal tension mode or shear
compression mode or shear tension mode, (c) bond splitting failure along the longitudinal
reinforcement and, and (d) anchorage failure of longitudinal reinforcement at the member end. The
building code places design requirements to prevent these modes of failure.
The behavior of beams is influenced by (a) dimensions of section, (b) member length, (c) amount
of longitudinal reinforcement, (d) anchorage of longitudinal reinforcement at member ends, (e)
concrete strength, (f) amount and detailing of lateral reinforcement, (g) contribution of slabs, and (h)
loading (oneway monotonic, oneway cyclic, or reversal).
The beam is normally designed to develop flexural yielding at the member ends under earthquake
loading. The amount of tensile reinforcement and the effective depth control the flexural resistance of
reinforced concrete section. The use of compressive reinforcement does not increase the flexural
strength, but it contributes to the deformation capacity after flexural yielding. The concrete strength
has little influence on the flexural strength of girders without axial forces. A large deformation capacity
and stable hysteresis energy dissipation can be expected from the beam dominantly behaving in
flexure if a member is properly designed to prevent shear failure, bond splitting failure and anchorage
failure of longitudinal reinforcement at the formation of flexural yield hinges at the member ends; the
gravity load should be included in the evaluation of design forces. The use of large diameter bars as
longitudinal reinforcement may result in bond splitting failure along the longitudinal reinforcement.
Lateral reinforcement is effective (a) improving shear resistance, (b) confining core concrete to
enhance resistance and deformation capacity, (c) improving bond splitting resistance, and (d)
improving resistance against buckling of longitudinal reinforcement. The spacing and detailing of
lateral reinforcement is important. The concrete strength is important in defining shear strength of a
reinforced concrete member. It should be noted that the shear strength is not a unique value, but that
shear resistance deteriorates with damage in concrete after flexural yielding.
The slab monolithically cast with a girder
contributes to the stiffness and flexural resistance
of the girder. This increase in flexural resistance
should be considered in determining the design
shear of a member in preventing shear failure.
An uncountable number of beam specimens
have been tested, under monotonically increasing
loading and under load reversals, in structures
laboratories throughout the world. These
specimens are generally tested under simulated
earthquake loading without vertical gravity loading
acting on slabs; i.e., the inflection point was
assumed to locate at midspan of the beam.
In old days, specimens were tested under a
simple support condition or in a cantilever
configuration, but more recently, beam specimens
have been tested under antisymmetric bending
conditions. Some specimens were tested with adjacent slabs.
Examples of Antisymmetric Bending Test
Flexural Deformation and Shear Deformation: The flexural deformation is associated with the
Bernoulli’s hypothesis that the plane section normal to the member axis before bending remains
plane after bending. The flexural deformation is attributed to the normal stress acting on the section.
Shear force causes deformation in the direction normal to the member axis. There exists the
interaction of flexural and shear deformation, and the two deformations cannot be clearly separated
after the formation of cracking. Additional rotational deformation of a beam is resulted near the critical
section of bending from the pullout of the longitudinal reinforcement from its anchorage zone.
The flexural deformation (average curvature) is obtained from longitudinal strain measurements at
the top and bottom chords assuming that a plane section remains plane after bending. Similar to the
flexural deformation, a shear deformation index is defined from strain measurements in the two
diagonal directions. This index does not represent the true shear deformation because the
deformation cannot be simply divided into the classical flexure and shear deformations once a shear
cracking occurs in a member.
1 1 2 2
2 2
1
5 6 1
1 1
1 2
1
6 7
1
( ' ) ( ' )
:
( )
:
2
2
2
flexural deformation index
h
h
shear deformation index
h
h h
h
θ
δ δ
γ
− + −
=
− +
=
+
=
+
=
A A A A
A
A
A A
A
A typical momentaverage curvature curve obtained from a simply supported beam test (Celebi
and Penzien, 1973) shows that the stiffness during loading gradually decreases with load, forming a
fat hysteresis loop and absorbing a large hysteresis energy. The hysteresis loops remain almost
identical even after several load reversals at the same displacement amplitude beyond yielding. A
typical lateral loadshear deformation index curve obtained in the same specimen shows the stiffness
during loading gradually increases with load, exhibiting a "pinching" in the curve. The hysteretic
energy dissipation is smaller. The hysteresis loop decays with the number of load reversals, resulting
in a smaller resistance at the same peak displacement in each repeated loading cycle. Although the
curve shows a "yielding" phenomenon, it is important to recognize that the shear force of the member
was limited by flexural yielding at the critical section rather than by yielding in shear. This yielding
clearly indicate the interaction of shear and bending.
Flexural deformation index Shear deformation index
Flexure Dominant Behavior: In the virgin loading branch, flexural cracking near the crosssection of
maximum moment reduces the initial elastic stiffness. The specimen continues to soften with loading
after cracking due to the spreading of cracking along the specimen length, crack opening associated
with tension softening of concrete and bondslip of the reinforcement between cracks. Yielding of
tensile reinforcement causes an abrupt and sharp reduction in stiffness. Even after the flexural
yielding, the resistance continues to increase due to the shift of neutral axis and later due to strain
hardening of reinforcement. Spalling of the concrete in compression has a negative effect on the
resistance.
Upon unloading after postyielding, the
unloading stiffness is generally high, but gradually
softens at lower loading level. A significant residual
deformation exists even after the removal of loads
caused by permanent strain in longitudinal
reinforcement and residual bar slip. Cracks remain
open at the removal of loads due to the residual
bar slip. The overall unloading stiffness degrades
with increasing plastic deformation amplitudes.
Reloading stiffness immediately after load
reversal is generally low until opened cracks close;
the compression by bending moment must be
resisted by the compressive reinforcement. The
reloading stiffness gradually recovers with the closing of cracks. Although the first postelastic
excursion may be considered as virgin loading in the reloading direction, the softening is more
gradual than in the initial yielding direction partly attributable to the Bauschinger effect of the steel.
The resistance at the previous maximum displacement reaches the level of the previous maximum
resistance.
When the reloading branch reaches the previous maximum response point, further loading
proceeds along the continuation of the virgin loading branch. Flexural failure of beams due to cyclic
loading is gradual, controlled by progressive deterioration in the compressive zone such as spalling of
concrete and local buckling of the longitudinal reinforcement. The reinforcement sometimes fracture
in tension in the subsequent half cycle after bucking.
Clear definition of failure is difficult in flexure dominated members unless tensile fracture of
longitudinal reinforcement is observed. Therefore, failure is often defined in the experiment as a point
where the resistance can not recover approximately 80 percent of the maximum resistance. It should
be noted that this definition of failure point is affected by the loading history.
Member deformation capacity is defined as the ratio (ductility ratio) of the deformation at failure to
Dominant flexural behavior of beam
(Celebi and Penzien, 1973)
the deformation at flexural yielding. The definition of yield point is sometimes difficult to determine
especially when the tensile reinforcement is placed in double layers.
Role of Compressive Reinforcement:
The compressive reinforcement placed in
section will not contribute to the flexural
strength of the section. Gaston, Siess and
Newmark (1952) tested five beams of
152x274 mm under monotonically
increasing load. Main parameter in the test
was the amount of tensile reinforcement.
The tensile reinforcement ratio ρ was
increased from 0.34% (Specimen T1La) to
1.90% (Specimen T4La) in the first four
specimens. The test results showed that
the bending resistance increased
proportional to the tensile reinforcement
ratio, but the deformation capacity
decreased with increasing amount of
tensile reinforcement. Concrete must resist
compressive force equal to the yielding
force in tensile reinforcement, and failed at
early stage when larger compressive forces
were demanded by the tensile
reinforcement.
In the last specimen (Specimen C4xna), the tensile reinforcement ratio was 1.90%, but 0.98%
compressive reinforcement was added in the section. The result showed small increase in flexural
resistance of Specimen C4xna compared with the resistance of Specimen T4La, but the deformation
capacity was significantly improved by the use of compressive reinforcement.
Flexural Member with Unsymmetric Cross Section: Beams are often not symmetrically reinforced
at the top and bottom. Even if the top and bottom are equally reinforced, the contribution of slab
reinforcement to beam flexural resistance will result in bending resistances different in positive and
negative bending. During the reloading in the weak direction with open cracks on the compression
side, tensile yielding is not enough to cause the compression bars to yield and cracks to close.
Therefore, during the reloading in the weak direction after yielding in the strong direction, cracks
remain open over fulldepth near the critical section, the compression stress under bending must be
Behavior of flexure dominated beam under load reversals
Effect of tensile and compressive reinforcement
Gaston, Siess, and Newmark, 1951
resisted by the compression reinforcement. The pinching of hysteresis shape occurs upon reloading
in the strong direction because high compression stress must be resisted by small amount of
longitudinal reinforcement before cracks closing.
Failure of unsymmetrically reinforced section occurs in two forms; (a) gradual progressive
compressive failure of the weaker side, and (b) abrupt fracture of tensile reinforcement in the weak
side.
Test of fullscale sevenstory RC structure (Hiraishi et al., 1985)
Effect of Shearspantodepth Ratio: Shearspan to depth ratio is the most significant parameter
that influences the shear deformation characteristics. In a beam of small shearspantodepth ratio,
shear deformation becomes appreciable compared with bending deformation. Dominant shear
response causes a more pronounced pinching in the
forcedeformation (hysteresis) curve, and a faster
degradation of the hysteresis energy dissipating capacity.
Considerable improvements in delaying and reducing the
degrading effects can be accomplished by using closely
spaced ties. The state of the knowledge is not sufficient to
define the stiffness degrading parameters on the basis of
the member geometry and material properties.
When the shear spandepth ratio becomes less than
two, the effect of shear becomes important. Garstka et al.
(1993) demonstrated the effect of shearspandepth ratio
on the forcedeformation curves under monotonically
increasing loading. The compression zone of the concrete
is severely damaged with a decreasing shear spandepth
ratio by the diagonal strut action.
Shear Failure Modes: Shear failure of reinforced concrete members takes place in the from of (a)
diagonal tension failure, (b) shear compression failure, and (c) shear tension failure.
Shear failure became notorious by the experience of abrupt shear failure of members in diagonal
tension mode. This mode of shear failure occurs abruptly in a relatively slender member with light
amount of lateral reinforcement; i.e., the amount of lateral reinforcement is not sufficient to resist
tensile forces previously carried by concrete before diagonal shear cracking.
Failure in shear compression mode is relatively ductile, in which concrete in diagonal direction fails
in compression after tensile yielding of lateral reinforcement.
Failure in shear tension mode takes place in the form of bond splitting failure along longitudinal
reinforcement after formation of diagonal shear cracks.
Shear strength may be affected by (a) tensile reinforcement ratio, (b) shear spantodepth ratio, (c)
shear reinforcement ratio, and (d) arrangement of shear reinforcement.
The shear reinforcement is inactive up to the occurrence of inclined cracking. Shear reinforcement,
upon yielding, develops large plastic strain. During unloading stage, large residual strain remains due
to relative movement along inclined shear cracking. Unloading and reloading after yielding of lateral
reinforcement leads to a gradual buildup of permanent tensile strain in lateral reinforcement. During
reloading, slippage along the inclined cracking takes place at low stiffness exhibiting a pinching effect.
Effect of shear spandepth ratio on
behavior (Garstka et al., 1993)
Bond Splitting Failure: Another failure mode is bond
splitting failure along the longitudinal reinforcement.
The stress in the longitudinal reinforcement must be
transferred to surround concrete within the beam.
This is a critical when large diameter bar or high
strength bars are used for longitudinal reinforcement.
Koda, Otani and Aoyama (1987) tested Tshape
beam with largediameter bars (2D25 and 1D19,
tensile reinforcement ratio
t
p =0.0144). The lateral
reinforcement ratio at the end was 0.98 % (3D10 at
87 mm) in the two specimens; Specimen No. 9 had a
constant lateral reinforcement but bond along the
longitudinal reinforcement was removed by placing
wax around the longitudinal reinforcement. The
lateral reinforcement ratio was reduced to 0.65 %
(2D10 at 87 mm) in the middle span in Specimen No.
3.
Specimen No. 9 without bond resistance along the
longitudinal reinforcement clearly shows pinching
hysteresis with lower initial stiffness; significant decay
in the second loading cycle at each deflection
amplitude is observed. Bond splitting cracks
developed in Specimen No. 3 at a member deflection
angle of 4/100.
Effect of Lateral Reinforcement on Bond Splitting
Failure: Reinforced concrete beams are normally
designed to develop flexural yielding at the member
ends, and resistance sufficient to prevent shear
failure or bond splitting failure is provided. A series of
tests on onehalf scale reinforced concrete beams
(175 x 270 mm, effective depth was 243 mm) with
slab were conducted by Fujisawa et al. (1988) to
study the effect of lateral reinforcement under load
reversals. Shear spanto(overall) depth ratio was 2.5 for all specimens. Tensile reinforcement ratio
t
p was 1.20 % (4D13) in all specimens. Slab thickness was 50 mm, and width was 500 mm on
each side reinforced with 6D6 bars.
The test results of four specimens are compared to study the effect of lateral reinforcement on the
bondsplitting behavior. The lateral reinforcement ratio
w
p was 0.30 % in Specimen No. 1
(2D6@123 mm), 0.60 % in Specimen No. 2 (2D6@61 mm), and 0.90 % in Specimen No.3
(3D6@61 mm). In Specimen No. 4, lateral reinforcement ratio within distance D (overall beam
depth=270 mm) from the end was 0.60 %, but the ratio was reduced to 0.30 % in the middle region.
Specimen No. 1 with light lateral reinforcement ratio failed in bond splitting mode along the
longitudinal reinforcement before flexural yielding. With an increase in lateral reinforcement ratio in
Specimen No. 2, the resistance after flexural yielding was maintained to a larger deformation.
Although significant damage occurred near the member ends, Specimen No. 3 with high lateral
reinforcement ratio could develop flexural yielding and stable hysteresis loops, and maintained the
yield resistance to deflection angle R of 5/100 rad. The lateral resistance decayed due to the
compression failure of concrete and buckling of longitudinal reinforcement. Specimen No. 4 failed in
bond splitting mode in the middle part where the amount of lateral reinforcement was reduced to the
level the same as Specimen No. 1.
試験体
Specimen No 3
Specimen No.9
Deformation, mm
p
w
=0.3% p
w
=0.6%
p
w
=0.9%
p
w
=0.6% p
w
=0.3%
Displ.
Displ.
Displ.
Displ.
pw=0.3% pw=0.6%
pw=0.9%
Member deflection angle R is defined as lateral deflection divided by the member length. A
reinforced concrete member designed in accordance with Japanese seismic force is known to yield in
flexure approximately at a member deflection angle of 1/200 rad. Shear dominated specimens
exhibited thin and Sshaped hysteresis loops compared with flexure dominated specimens.
Effective Width of Tbeam: A girder and slabs, cast monolithically, act integrally as a Tbeam; a part
of the slab acts as the flange of the girder under bending, increasing stiffness and flexural resistance
when the girder is subjected to negative bending. Therefore, slabs on either side of the girder should
be considered in evaluating the stiffness and flexural resistance of the girder.
Under positive bending causing, compression stress at the top fiber, the stress at extreme
compressive fiber of the slab decay with distance from the girder face due to shear deformation in the
flange (shear lag) (Park and Paulay, 1975). The slab increases flexural stiffness, but does not
increase the resistance appreciably.
Under negative bending, slab reinforcement parallel to the girder increases the bending resistance
of the girder. Width of slab in which the longitudinal reinforcement is effective to girder flexural
resistance, increases with lateral deformation.
Suzuki et al. (1984) tested halfscale threedimensional beamcolumn subassemblages under
bidirectional horizontal load reversals (. The beam dimensions were 200 x 300 mm, reinforced by
4D13 bars at the top and 4D13 bars at the bottom. The column dimensions were 300 mm square,
reinforced by 8D13 bars. The
yield stress of D13 bars was 366
MPa, and concrete strength in
beams was 19.5 MPa. The slab
was 70 mm thick and 2,440 mm
wide, reinforced by D6 bars at
200 mm on centers.
The strain in slab longitudinal
reinforcement was shown to
increase with lateral deformation
applied to the specimen, and the
region of yielding slab
reinforcement became wider. The
resistance at the final stage was
close to the resistance calculated
using entire slab tensile
reinforcement. The spread of the
effective width of slab affects the
stiffness after yielding.
A sevenstory fullscale building specimen was tested under pseudodynamic loading (Yoshimura
and Kurose, 1985). Strains in slab reinforcement were measured during the test. The sum of slab
stress divided by the yield stress of slab reinforcement gave equivalent number of yielded slab
reinforcing bars. They indicated that the slab reinforcement in a width equal to three to four times the
normally assumed effective slab width (slab width equal to onetenth of the girder span) yielded at a
story drift angle of 1/75 rad.
References:
Celebi, M., and J. Penzien, "Experimental Investigation into the Seismic Behavior of Critical Region of
Reinforced Concrete Components as Influenced by Moment and Shear," Report EERC No. 734,
Earthquake Engineering Research Center, University of California at Berkeley, 1973.
Comite EuroInternationa du Beton, “RC Frames under Earthquake Loading  State of the Art
Report,” Thomas Telford, 1996, 303 pp.
French, C. W., and A. Boroojerdi, "Contribution of RC Floor Slabs in Resisting Lateral Loads," Journal,
Structural Engineering, ASCE, Vol. 115, No. 1, January 1989, pp. 118.
Fujisawa, M., et al., "Study on Ductility of Girders, Pilot Tests (in Japanese)," Report on Development
of Highrise Frame Wall Structures, Building Research Institute, Ministry of Construction, 1988.
Garstka, B., et al., “Damage Assessment in Cyclically Loaded Reinforced Concrete Members,”
Cyclically Loaded Reinforced Concrete Members, Structural Dynamics, Balkema, Rotterdam, Vol.
Strain in slab reinforcement
Suzuki et al., 1984
1, 1993, pp. 121  128.
Hiraishi, H., S. Nakata, Y. Kitagawa and T. Kaminosono, “Static Tests on Shear Walls and
Beamcolumn Assemblies and Study on Correlation between Shaking Table Tests and
Psuedodynamic Tests,” ACI SP84, Earthquake Effects n Reinforced Concrete Structures, 
U.S.Japan Research, American Concrete Institute, 1985, pp. 11  48.
Koda, S., S. Otani and H. Aoyama, “Reinforcement Details of Tshape beams and Ductility (in
Japanese),” Transactions, Architectural Institute of Japan Annual Convention, October 1987, pp.
209  210.
Ma, S. M., V. V. Bertero and E. P. Popov, "Experimental and Analytical Studies of the Hysteretic
Behavior of Reinforced Concrete Members under Bidirectional Reversed Lateral Loading," Report
No. EERC 762, University of California at Berkeley, 1976.
Park, R. and T. Paulay, Reinforced Concrete Structures, John Wiley & Sons, Inc., pp. 99100, 1975.
Suzuki, N., S. Otani and H. Aoyama, "Threedimensional Beamcolumn Subassemblages under
Bidirectional Earthquake Loading," Proceedings, Eighth World Conference on Earthquake
Engineering, Vol. VI, San Francisco, July 1984, pp. 453460.
Yoshimura, M., and Y. Kurose, "Inelastic Behavior of the Building," ACI SP84, Earthquake Effects on
Reinforced Concrete Structure, USJapan Research, American Concrete Institute, Detroit, 1985,
pp. 163202.
3.2 Behavior of Columns
A column is an important structural element that supports the weight of a structure and resists
earthquake story shear. The column height is normally determined by the story height of a building.
Section dimensions are controlled by the amount of axial force and shear force used in design.
During an earthquake, exterior columns and corner columns are subjected to varying axial load in
addition to bidirectional lateral load reversals, while interior columns are subjected to almost constant
axial loads.
Yielding is generally permitted at the top of topstory columns and at the base of firststory
columns, and often at the ends of exterior columns subjected to tension force under earthquake
induced overturning moment. The other columns are normally provided with flexural strength
sufficient to prevent flexural yielding using the capacity design consideration. The failure of columns
may lead to the collapse of a building; hence, the brittle failure is carefully avoided in the design of
columns. Brittle failure of a column causes an increase in shear in the other columns of the story, and
might lead to the progressive failure of all columns in the story. Furthermore, the axial force carried by
the failing column must be transferred to the adjacent columns and walls by girders.
The behavior of independent columns (without wing walls nor structurally attached nonstructural
elements) is influenced by (a) dimensions of section, (b) amount of longitudinal reinforcement, (c)
amount of lateral reinforcement, (d) level of axial load, (e) concrete strength, and (f) loading (oneway,
cyclic, or reversal). Four modes of failure should be considered when the behavior of reinforced
concrete members is discussed; i.e., (a) flexural failure caused by crushing of concrete or buckling of
compressive reinforcement, (b) shear failure, (c) bond splitting failure along the longitudinal
reinforcement, and (d) anchorage failure of longitudinal reinforcement at the member end.
Columns are normally tested (a) simple support conditions, (b) cantilever form, and (c)
antisymmetric bending conditions.
Hysteresis Characteristics of Columns under Flexure: A typical forcedeflection curve of a
cantilever column is shown with no axial force (Otani et al. 1979). The column section was 305x205
mm, and reinforced with 8No. 7 bars (yield stress of 438 MPa). The concrete strength was 23 MPa.
The height of loading point from the column base was 1,372 mm.
Note the following observations:
(a) tensile cracking of concrete and
yielding of longitudinal
reinforcement reduced the
stiffness; (b) when a deflection
reversal was repeated at the same
newly attained maximum amplitude
(for example, cycles 3 and 4) the
loading stiffness in the second
cycle was lower than that in the first
cycle, although the resistances at
the peak displacement were almost
identical; and (c) average stiffness
(peaktopeak) of a complete cycle
decreased with a maximum
displacement amplitude. For
example, the peaktopeak
stiffness of cycle 5, after large
amplitude displacement reversals,
was significantly reduced from that
of cycle 2 at comparable
displacement amplitudes. Therefore, the hysteretic behaviour of the reinforced concrete is sensitive
to loading history.
Column lateral force deflection relationship
Otani et al., 1979
Higashi et a. (1977) showed the influence of loading history on the response of shear failing
beams under oneway loading and reversal loading. The reduction in resistance after yielding is
significant when the specimen fails in shear. It is difficult to generalize the hysteresis relation for this
type of failure.
Monotonicaly loading
Cyclic or reversed loading
Effect of Loading History on Shear Failing Members
Higashi et al., 1977
Columns with Confining Reinforcement: Lateral
reinforcement is known to confine the core concrete and
enhance the compressive strength and deformation capacity
of the confined concrete. The effect of confining reinforcement
is important; a column with closely spaced stirrups and
welldistributed longitudinal reinforcement shows very little
strength decay even subjected to high axial force above the
balance point.
Rabbat et al. (1986) showed the improvement of
deformation capacity with additional lateral reinforcement. The
axial force level was 0.30 times the compressive strength of
concrete area. The lateral reinforcement ratios of the two
specimens were 0.74% and 1.60%. The column with small
amount of lateral reinforcement failed before flexural yielding,
while the column with large lateral reinforcement maintained
stable hysteresis. The lateral reinforcement can confine the
core concrete and delay the compression failure of concrete
and also increase shear resistance.
Sugano et al. (1985) tested a series of columns of 210 x
210 mm with different amount of lateral reinforcement
subjected to monotonically increasing axial loads. Lateral
reinforcement used were round lateral reinforcement (yield
stress
wy
σ = 421 MPa) used in Specimen S06 (lateral
reinforcement ratio p
w
= 0.56 %) and Specimen S12 (p
w
=
1.12 %), plain 4 φ welded wire mesh lateral reinforcement
(
wy
σ = 554 MPa) in Specimen M06 (p
w
= 0.60 %) and
Specimen M12 (p
w
= 1.20 %), highstrength lateral reinforcement (
wy
σ = 1,397 MPa) in Specimen
U06 (p
w
= 0.56 %) and Specimen U12 (p
w
= 1.12 %), and band plate of 25 x 2.3 mm at 50 mm on
centers in Specimen BP12. Longitudinal reinforcement was common among the specimens; 16D10
bars (
wy
σ = 400 MPa) were used in a section with gross reinforcement ratio of 2.58 %. Concrete
strength was 25 to 28 MPa.
Tie Fracture
Tie Fracture
Welding Failure
Welding Failure
Strain
No Reinf.
L
o
a
d
Welded Wire Fabric Rectangular Spiral
Band Plate
Sugano et al., 1985
Effect of lateral reinforcement
(Rabbat et al., 1986)
Lateral reinforcement ratio = 0.0074
Deformation of each column was measured over middle 375 mm length on opposite two faces.
All specimens reached maximum resistance at strain between 0.3 to 0.5 %, and the resistance
deteriorated by fracture of lateral reinforcement. The use of lateral reinforcement is seen to enhance
the strength and deformation capacity of a reinforced concrete column. The following can be
observed in the figure;
(a) For the same type of lateral reinforcement, the axial deformation capacity can be observed to
increase with increasing amount of lateral reinforcement.
(b) High strength lateral reinforcement (Specimens U06 and U12) can develop larger deformation
capacity than normal strength lateral reinforcement (Specimens S06 and S12).
(c) Steel band plate did not perform well compared.
(d) Welded wire fabric appears to be most effective in the series.
(e) Thin but closely placed lateral reinforcement is more effective than large but far spaced lateral
reinforcement.
Another example is shown to demonstrate the effect of lateral (confining) reinforcement on the
deformation capacity of reinforced concrete columns. The two specimens have the same dimensions
and the same amount of longitudinal reinforcement. The amount of lateral reinforcement was varied
from 1.2% (Specimen A2) to 2.0% (Specimen A4) in the two specimens. Specimen A2 could develop
flexural yielding and maintain the resistance to a story drift angle of 1.5/100, while Specimen A4 could
maintain the resistance to a story drift angle of 4/100.
Column subjected to Axial Load and Bending: The yield and ultimate moments under monotonic
loading increase with the compressive axial force below the balanced load level. The stiffness in
virgin loading, unloading and reloading increases with the level of axial force. It is generally known
that a column subjected to compression force above the balanced point cannot develop large plastic
deformation beyond yielding because compressed concrete deteriorates at a faster rate. Axial force
closes cracks at a low lateral force in a column. This phenomenon is different from the behavior of a
beam.
The strength degrades considerably with cycling when the axial force level is near or above the
balance point. Rabbat et al. (1986) tested columns with different levels of axial force and
demonstrates the effect of axial force level on the column deformation capacity. The lateral
reinforcement ratio was 1.74 percent in the specimens. The axial force was 10, 20 and 30 percent of
the compressive strength of the concrete area. The column subjected to low axial force level
developed flexural yielding and exhibited reasonable deformation capacity under lateral load
reversals until failure occurred in compression side. The column subjected to higher axial load, on the
other hand, did not develop flexural yielding.
A small scale column was tested under constant tensile axial force and lateral load reversals. The
column was 225 x 225 mm, reinforced with grade SD395 (nominal yield stress of 395 MPa)
longitudinal reinforcement and grade SBPD1300 highstrength lateral reinforcement. The tensile
reinforcement ratio
t
p was quite large and 4.7%. Lateral reinforcement ratio
w
p was 0.7%.
Concrete strength was 60.7 MPa. The axial force level was 0.18 times the compressive strength of
concrete area. Note a stable hysteresis behavior to a large lateral deformation under tensile axial
force.
Column subjected to tensile axial force
Effect of axial force levels (N/Ac f’c=0.10, 0.20 and 0.30)
Column subjected to Varying Axial Load: Exterior columns and corner columns are subjected to
varying axial load and bidirectional lateral load reversals during an actual earthquake, while interior
columns are subjected to almost constant axial loads. The level of varying axial forces is limited by
the formation of yield hinges at the beams in the structure; the upper limit is the sum of shears in the
exterior beams at the formation of yield hinges at the two ends. Variation of axial force is more closely
related to shear acting in the first story column.
Sakaguchi et al. (1985) tested two columns of 450 x 450 mm section subjected to varying axial
load (0.6 N
uc
to 0.7 N
ut
for an exterior column) or constant axial load (0.35
B
σ A
e
for an interior
column) and lateral load reversals, where
B
σ : compressive strength of concrete (=43 MPa), A
e
:
transformed section area of column, N
uc
: compressive strength of column (= 0.85
B
σ A
c
+ A
g
sy
σ ),
N
ut
: tensile strength of column (= A
g
sy
σ ). Gross reinforcement ratio was 2.29 % (12D22, yield stress
sy
σ = 411 MPa). Lateral reinforcement ratio was 1.27 % (exterior and interior spiral reinforcement of
4D10 at 50 or 60 mm spacing, yield stress
wy
σ = 390 MPa). Clear height was 1,640 mm. Steel
Hsection was embedded at the centroid of section to resist high tensile force. During the loading, the
inflection was maintained at column midheight.
Both specimens showed good performance, dominated by flexure, up to member rotational angle
(lateral deformation divided by clear member length) of 1/50 rad. However, the resistance started to
decay at a larger deformation in the exterior column due to high level of axial force. The loading
condition of constant axial force is severer to a specimen than that of varying axial load as long as the
axial force amplitudes of the two loading cases are comparable.
Maximum Load
Maximum Load
Calculated Strength
Calculated
Strength
Maximum Load
Maximum Load
Li et al. (1986) tested a cantilever column (200 x 200 x 570 mm) under varying axial load and
uniaxial lateral load reversal. Eight D10 bars were used as longitudinal reinforcement (yield stress
sy
σ = 426 MPa). Lateral reinforcement ratio was 0.64 % (2D6@50 mm, yield stress
wy
σ = 394 MPa).
Concrete strength was 27.6 MPa.
Axial stress was varied by 2 MPa from the static value of 2 MPa proportional to lateral resistance
of the column. As expected, the flexural resistance increased with increased axial load in positive
direction: "interaction of axial force and flexural resistance". The resistance started to decay after
buckling of longitudinal reinforcement in the positive direction.
Additional test results were reported by Gibertsen and Moehle (1980), Kreger and Linbeck (1986),
Abrams (1987) and Ristic et al. (1986, 1988).
Test Specimen (unit: mm) Lateral loaddeformation relation
Li et al., 1986
Columns subjected to Bidirectional Lateral Load Reversals: Columns in a structure are
subjected to constant or varying axial forces and bidirectional shear and bending moment reversals
during an earthquake. The test results under bidirectional lateral load reversals are rather scarce.
The first column test was conducted by Fujii, Aoyama and Umemura (1974).
Otani and Cheung (1981) reported tests of concrete columns subjected to bidirectional loading.
No axial force was applied to specimens. Column section was 305x305 mm, and reinforced by 8No.
7 deformed bars (yield stress of 438 MPa). Lateral reinforcement was No. 2 square outer ties and No.
2 (yield stress of 236 MPa and 300 MPa) diamond inner ties both spaced at 88 mm. Three pairs of
specimens were tested. Compressive strength was 34 MPa in specimens SP3 and Sp4, 23 MPa in
specimens SP5 and SP6, and 27 MPa in specimens SP7 and SP8.
Specimen SP4 was loaded NS direction first and then EW direction, separately, and finally in NS
direction. The influence in loading in the orthogonal direction can be clearly observed in the
loaddisplacement relation in NS direction. Cycles 13 and 22 were loaded to the same displacement
amplitude before and after loading in the orthogonal direction. Significant reduction in resistance was
caused by the loading in the orthogonal direction. Specimens failed after heavy flexural and inclined
cracking, spalling of the concrete, disintegration of the core concrete, mainly due to the propagation
of inclined cracks and grinding along them, followed by buckling of longitudinal reinforcement.
Specimen SP5 was loaded only in NS direction, while specimen SP6 was initially loaded in NS
direction and then in EW direction. The total displacement history was made same in the two
specimens. Specimen SP5 subjected to loading in NS direction failed at smaller deflection.
Effect of transverse loading, specimen SP4
Otani, Cheung, 1981
Cantilever column specimens,
constructed under the same specifications,
were tested under constant and varying axial
load and bidirectional lateral load reversals
(Li et al., 1986).
Column section was 200x200 mm, and
the height from the base to the lateral loading
point was 600 mm. A column section was
reinforced with 8D10 deformed bars (gross
reinforcement ratio of 1.43 %, yield stress of
417 MPa); lateral reinforcement was square
D6 bars spaced at 50 mm (lateral
reinforcement ratio of 0.64 %, yield stress of 386 MPa). Concrete strength at testing was 27 to 58
MPa.
For Forced displacement path under bidirectional loading is shown right; unidirectional
displacement was applied to a specimen at a time to clarify the effect of loading in each direction.
Initial axial force level was 0.07 times bDf ’c, where b and D: width and depth of section, and f ’
c
:
compressive strength of concrete.
The level of axial force was maintained constant in Specimen B80. An interaction of resistance
was observed in the test; i.e., when the displacement amplitude was kept at peak value in one
direction and displacement was increased in the orthogonal direction, the resistance in the first
direction decayed with displacement in the orthogonal direction: "interaction of bidirectional
Effect of orthogonal loading
Otani and Cheung, 1981
resistance."
Column subjected to Varying Axial Load and Bidirectional Lateral Load Reversals: An exterior
column is subjected to constant axial force and varying axial force depending of the loading direction
in a structure. A corner column is subjected to varying axial load and lateral load reversals in the two
direction.
Specimen B81 was subjected to varying axial force proportional to shear acting in EW direction.
Compressive strength of concrete was 31.4 MPa, otherwise the specimen was fabricated under the
same specification as the previous two specimens. The axial stress was varied from the static level of
0 to 0.13 times bDf ’
c
, proportional to lateral resistance in EW direction.
Interaction of bidirectional resistance
Li et al., 1986
Lateral loadbiaxial horizontal displacement relation under constant axial force
Li et al. 1986
Corner column specimen B82 was loaded varying axial force in the two directions. As the axial
force was varied proportional to the lateral resistance in each direction, the axial force was varied
from 0.21 A
g
f
y
to 0.19 bDf ’
c
. The decay in resistance after reaching maximum resistance was faster
under bidirectional lateral load reversals.
References:
Abrams, D. P., “Influence of Axial Force Variation on Flexural Behavior of Reinforced Concrete
Columns,” Structural Journal, American Concrete Institute, Vol. 84, MayJune 1987, pp. 246  254.
Abrams, D. P., "Laboratory Definitions of Behavior for Structural Components and Buildings
Systems," ACISP127, Earthquake Resistant Concrete Structures  Inelastic Response and
Design, American Concrete Institute, Detroit, 1991, pp. 91152.
Bousias, S. N., et al., “RC Columns in Cyclic Biaxial Bending and Axial Load,” Proceedings, Tenth
World Conference on Earthquake Engineering, Madrid, 1992, pp. 3041  3046.
Bousias, S. N., “Experimental and Analytical Study of Reinforced Concrete Columns in Cyclic Biaxial
Bending with Axial Force, “ Ph. D. Thesis, Department of Civil Engineering, University of Patras,
Corner column specimen subjected to varying axial stress proportional
to the lateral resistance in each direction
Li et al., 1986
Exterior column specimen subjected to varying axial force in EW direction and
constant axial force in NS direction
Li et al., 1986
Greece, 1993.
Fujii, S., “Study on Reinforced Concrete Columns subjected to Biaxial Flexure (in Japanese),” M. eng.
Thesis, Department of Architecture, University of Tokyo, 1974.
Gibertsen, N. and J. P. Moehle, “Experimental Study of Small Scale R/C Columns subjected to Axial
Load and Shear Reversals,” Structural Research Series No. 481, Department of Civil Engineering,
University of Illinois at Urbana, 1980.
Higashi, Y., M. Ohkubo and M. Ohtsuka, “Influence of Loading Excursions on Restoring Force
Characteristics and Failure Modes of Reinforced Concrete Columns,” Paper No. 1123, Sixth
World Conference on Earthquake Engineering, New Delhi, January 1977.
Kreger, M., and L. Linbeck, “Behavior of Reinforced Concrete Columns subjected to Lateral and Axial
Loading Reversals,” Proceedings, Third U.S. National Conference on Earthquake Engineering,
Charleston, South Carolina, Vol. “, pp. 1475  1486, 1986.
Li, K.N., et al., "Behavior of Reinforced Concrete Columns subjected to Varying Axial Load and
Bidirectional Horizontal Earthquake Loads (in Japanese)," Proceedings, Eighth Annual
Conference, Japan Concrete Institute, pp. 489492, 1986.
Low, S., and J. P. Moehle, “Experimental Study of Reinforced Concrete Columns subjected to
Multiaxial Cyclic Loading,” Earthquake Engineering Research Center, University of California at
Berkeley, EERC Report 8714,1987.
Otani, S, "Nonlinear Dynamic Analysis of Reinforced Concrete Building Structures," Canadian Journal
of Civil Engineering, Vol. 7, National Research Council of Canada, pp. 333344, 1980.
Otani, S., and V. W.T. Cheung, “Behavior of Reinforced Concrete Columns under Biaxial Lateral
Load Reversals,  (II) Test without Axial Loads,” Department of Civil Engineering Publication 8102,
University of Toronto, February 1981, 127 pp.
Rabbat, B., et al., “Seismic Behavior of Lightweight and Normalweight Concrete Columns,” Journal,
American Concrete Institute, Vol. 83, Jan.Feb. 1986, pp. 69  78..
Ristic, D., et al., “Effect of Variation of Axial Forces to Hysteretic Earthquake Response of Reinforced
Concrete Structures,” Proceedings, Eighth European Conference on Earthquake engineering,
Lisbon, Vol. 4, 7.4, 1986, pp. 49  56.
Ristic, D., et al., “Inelastic Stressstrain based Seismic Response Prediction of RC Structures
considering Dynamically Varying Axial Forces,” Proceedings, Ninth World Conference on
Earthquake Engineering, TokyoKyoto, 1988, Vol. VI, pp. 531  536.
Saatcioglu, M., "Modeling Hysteretic ForceDeformation Relations for Reinforced Concrete
Elements," ACI SP127, EarthquakeResistant Concrete Structures  Inelastic Response and
Design, American Concrete Institute, Detroit, 1991, pp. 153198.
Saatcioglu, M., and G. Azcebe, “Response of Reinforced Concrete Columns to Simulated Seismic
Loading,” Structural Journal, American Concrete Institute, Vol. 86, Jan.Feb. 1989, pp. 3 12.
Sakaguchi, N., et al., "Study on Structural Characteristics of Highrise Reinforced Concrete
Residential Buildings (Part 2: Strength Test of Columns) (in Japanese)," Summary Report, Annual
Meeting, Architectural Institute of Japan, Structures 2, pp. 153154, October 1985.
Sugano, S., et al., "Experimental Study on Highrise Reinforced Concrete Buildings (Part 1: Outline of
Study and Compression Test of Columns) (in Japanese)," Summary Report, Annual Meeting,
Architectural Institute of Japan, Structures 2, pp. 145146, October 1985.
3.3 Behavior of Interior Beamcolumn Connections
A beamcolumn connection (joint), a region common for columns and girders, is often treated as a
rigid panel or a shear panel in a structural analysis; shear deformation may be considered in the
frame analysis. The beamcolumn connections are classified into (a) interior beamcolumn
connection, (b) exterior beam column connection and (c) corner beamcolumn connection, depending
on the location of the column in a threedimensional frame. The beamcolumn connection is designed
to sustain maximum resistance of all members connected to the joint.
The essential requirements for the satisfactory performance of a beamcolumn joint in a reinforced
concrete structure can be summed up as follows (Park and Paulay, 1975);
(a) A joint should exhibit a service load performance equal in quality to that of the members it
joints,
(b) The strength of the joint should not govern the strength of the structure, and its behavior should
not impede the development of the full strength of the adjoining member, and
(c) Ease of construction and access for depositing and compacting concrete.
Failure Modes: The beamcolumn joint is subjected to normal
stresses and high shear stresses at the boundary of adjacent
beams and columns in addition to bond stress acting along the
beam and column longitudinal reinforcement. Concrete flexural
compression from beams and columns develops at opposite
corners in laterally loaded frames, suggesting the formation of a
diagonal compression strut.
Beam and column reinforcement passing through joints is
subjected to tension at one boundary and compression at the
other; significant amount of forces must be transmitted from
reinforcement to concrete through bond resistance within the
connection.
Parameters to affect the joint performance are (a) amount of joint hoop, (b) column axial force, (c)
input intensity of joint shear, (d) concrete compressive strength, (e) bond demand along beam bars
through joint, and (f) presence of transverse beams.
The beamcolumn subassemblage may fail in three modes; i.e., (a) shear failure in the connection
before flexural yielding at girder ends, (b) shear failure after flexural yielding at girder ends, (c)
flexural yielding at girder ends, and (d) bond failure along the beam longitudinal reinforcement. There
exist significant interaction of shear resistance and bond resistance along the beam longitudinal
reinforcement.
Examples of crack patterns and shearshear distortion relations in the interior beamcolumn
connection are shown. In the first specimen (a), large distortion (large diagonal extension but small
diagonal compression) is observed in the beamcolumn connection with the reduction in shear
resistance; extensive damage was observed in the connection. In the second specimen (b), the
distortion in the beamcolumn connection was small before beam flexural yielding, but the connection
Interior and exterior beamcolumn connections
T
s1
C
s1
C
c2
C
s2
T
s2
C
c2
V
c1
V
c2
V
jh
=T
s1
+T
s2
V
c1
Actions in interior beamcolumn
connection
distortion increased significantly after the beam yielding. When the connection is reinforced with
sufficient lateral reinforcement, the shear distortion of the connection becomes small (the third
specimen). The beamcolumn joint sometimes fails after developing flexural yielding at beam ends
although the shear input into the connection was controlled by the beam yielding. The shear strength
is not a unique value of the joint, but the resistance deteriorates with the damage within the joint.
Crack pattern
(a) Shear failure of joint
Shear stressdistortion relation
Shear distortion angle x10
2
rad
Crack Pattern
Story drift
S
h
e
a
r
s
t
r
e
s
s
/
C
o
n
c
r
e
t
e
s
t
r
e
n
g
t
h
(b) Shear failure after beam flexural yielding
Shear stressdistortion relation
Shear distortion angle x102 rad
S
h
e
a
r
s
t
r
e
s
s
/
C
o
n
c
r
e
t
e
s
t
r
e
n
g
t
h
Crack pattern
Shear distortion angle x102 rad
Shear stressdistortion relation
S
h
e
a
r
s
t
r
e
s
s
/
C
o
n
c
r
e
t
e
s
t
r
e
n
g
t
h
(c) Beam flexural yielding
Shear Resisting Mechanisms: Two major
shear resisting mechanisms are generally
considered; (a) diagonal compression strut
mechanism and (b) truss mechanism. The
diagonal compression strut mechanism is
formed by normal concrete stresses at the
boundary of the connection. The truss
mechanism is formed by the bond stress
acting along the beam and column
longitudinal reinforcement, the tensile stress
in lateral reinforcement, and compression
struts uniformly distributed in the connection.
Therefore, the truss mechanism is strongly
related to the condition of bond resistance
Diagonal compression strut
Truss
Shear resisting mechanisms of interior
beamcolumn connection
along the beam longitudinal reinforcement especially after flexural yielding at the beam ends.
Paulay et al. (1978) postulated that a diagonal compression strut develops at the initial load stage
before significant cycles of flexural hinging cause residual reinforcement strain and fulldepth
cracking so that shear forces are introduced to the joint by normal stresses acting in the compression
zones of the framing members. A minimum amount of joint hoop reinforcement is necessary for
confinement purpose at this stage. Following bar yield penetration and bond deterioration during
repeated inelastic excursion, the shear input is resisted by a selfequilibrating truss mechanism that
consists of a network of small compressive struts in the core concrete and of tensile forces in the
horizontal and vertical joint reinforcement (including the longitudinal reinforcement of the column).
Kitayama et al. (1991) compared the behavior of two beamcolumn assembly specimens J1 and
C1. The overall dimensions and loading methods of the two specimens are the same. Beams
(200x300 mm) of Specimen J1 were reinforced with 8D13 deformed bars (cross sectional area
A
s
=1016 mm
2
, and yield stress f
y
=402 MPa) and 4D13 bars at the bottom (A
s
=508 mm
2
), while those
of Specimen C1 were reinforced with 12D10 bars (A
s
= 856 mm
2
and f
y
= 323 MPa) at the top and
6D10 bars (A
s
=428 mm
2
). The amount of lateral reinforcement in the connection was the same
(lateral reinforcement ratio of 0.38%). Note that the tensile reinforcement ratio was quite high
compared to that commonly used in a frame structure. Both specimens were designed to yield at the
end of beams. Bond stress transfer along the beam reinforcement within the connection was severer
in Specimen J1 by the use of larger diameter and higher strength reinforcement.
Specimen J1 was judged
to fail in shear in the joint at a
story drift angle of 1/23 rad by
crushing and spalling of shell
concrete; Xshaped cracks
gradually opened along the
main diagonal with
deformation. Specimen C1
developed beam yielding with
fine diagonal cracks uniformly
developed in the connection.
Specimen J1 exhibited a
pinching hysteretic shape
especially after a story drift
angle of 1/46 rad, while
Specimen C1 developed a
good spindleshape
hysteresis. It should be
pointed out that the shear
stress level in Specimen J1 was approximately 1.25 times larger than that in Specimen C1.
The crack patterns in the joint support the modeling of the shear resistance of a beamcolumn
connection as the diagonal strut and truss mechanisms. Note that Specimen J1 developed shear
cracks initially caused by the truss mechanism, but the shear cracks in the mail diagonal became
dominant at a larger deformation; i.e., the truss mechanism deteriorated with the bond deterioration
along the beam longitudinal reinforcement, and the principal stress concentrated along the main
diagonal strut to cause shear failure. On the other hand, the diagonal strut and truss mechanisms
were maintained in Specimen C1; diagonal compression stresses distributed uniformly in the panel
concrete. The diagonal strut mechanism can exist without any bond stress transfer along the beam
reinforcement within the connection, while the truss mechanism can exist only when a good bond
stress transfer is maintained. The bond along the beam reinforcement inevitably deteriorates
especially after the beam flexural yielding unless the strength and size of the reinforcement are
strictly controlled.
Liande and Jirsa (1982) discussed the shear resistance of interior beamcolumn joints as the
diagonal compressive strut mechanism. This concept is accepted in ACIASCE Committee 352 report
(1985).
Otani (1991) suggests that the truss mechanism can be effective only when the bond resistance
along the longitudinal reinforcement is maintained. After load reversals of flexural hinging, it is difficult
to maintain the bond resistance along the beam reinforcement; hence, the shear must be resisted by
the main diagonal strut. Lateral reinforcement is necessary to confine cracked core concrete.
Lateral Reinforcement: Lateral reinforcement participates in the truss mechanism and confines the
core concrete in the beamcolumn connection. Noguchi and Kurusu (1987) reported that the strain in
lateral reinforcement was much larger if the bond resistance along the longitudinal reinforcement was
better. The lateral reinforcement may confine the core concrete in the connection, enhancing the
concrete compressive strength and ductility in the diagonal compression strut mechanism.
Otani et al. (1986) reported three halfscale beamcolumn subassemblage tests. Beams were
200x300 mm, and were reinforced with 12D10 bars (tensile reinforcement ratio
t
p =1.59%, yield
stress
y
σ =333 MPa) at the top and 6D10 bars (
t
p =0.79%). Columns were 300x300 mm, and
reinforced with 16D13 bars (
y
σ =439 MPa). The amount of lateral reinforcement was varied in the
three specimens; 2D6 bars (lateral reinforcement ratio
w
p =0.27%,
y
σ =337 MPa) in Specimen C1,
4D6 (
w
p =0.90%) and 4D10 bars (
w
p =2.01%) in Specimen C3. The concrete strength was 26.6
MPa. The column was subjected to a constant axial stress of 20 kgf/cm
2
.
Specimen C1 provided with less joint lateral
reinforcement developed diagonal shear cracks wider
than specimen C3 at a story drift angle of 1/23 rad.
The shell concrete of the connection swelled out
slightly at a story drift angle 1/15 rad. On the other
hand, specimen C3 did not widen diagonal shear
cracks after a story drift angle of 1/46 rad. The amount
of joint lateral reinforcement affected the states of
cracking. However, it is important to note that the
difference was small up to a story drift angle less than
1/46 rad.
The story shearstory drift relations showed fat
spindleshaped hysteretic shape without a decay in
resistance even at a story drift angle of 1/15 rad. The
resistance at peak deflection amplitudes was almost
the same for the three specimens. The hysteretic loop
area in the second cycles at the same displacement
amplitude was slightly greater for specimen C3 than
for specimen C1. From the fat hysteretic loop, it is
concluded that the bond of beam bars within a
connection did not deteriorate much with the number
of load reversals.
The story shear is compared with diagonal deformation of a joint panel. The diagonal deformation
was measured over a gauge length of 333 and 310 mm for specimens C1 and C3, respectively. Little
shear distortion of a joint panel was observed in two specimens up to a story drift angle of 1/46 rad.
Specimen C1 started to increase the width of shear cracks at a story drift angle of 1/46 rad, while
specimen C3 did not increase the diagonal deformation even at a story drift angle 1/15 rad. Observe
that the amount of joint lateral reinforcement little affected the shear distortion for a deflection range
expected from an intense earthquake motion. Beyond such deflection level, the lateral reinforcement
started to show the difference in confining the joint core concrete.
0.0 50.0 100.0 50.0
100.0
Story Drift, mm
S
t
o
r
y
S
h
e
a
r
,
t
o
n
f
S
t
o
r
y
S
h
e
a
r
,
t
o
n
f
Story shearstory drift relation
The strain of joint lateral
reinforcement is studied at peaks of each
positive loading cycle. Strains in
Specimen C1 exceeded the yield strain at
a story drift angle of 1/92 rad when the
beam started to yield. Strains in
Specimen C3 stayed less than 0.1
percent; in other words, the amount of
joint lateral reinforcement might not be
required as much as that in Specimen C3,
but the amount could be reduced to one
half even if the structure is expected to
deform at a story drift angle of 1/23 rad.
Bond and Beam Bar Slip: Additional
source of deformation in the connection
is deformation caused by bar slip within the connection. Consider a girder reinforcement passing
through a connection; girder reinforcement is subjected to high tensile stress at one end and
compression stress at
the other end, and the
stresses must be
transferred to concrete
within the connection.
However, the width of
column may not be
sufficient to allow the
stress transfer. This
situation is much easier
in the exterior
beamcolumn
connection.
The bond
deterioration of beam
bars within a joint is said
to be undesirable
because (a) the energy
dissipation at beam ends
is reduced by pinching in
a hysteresis shape, (b)
the diagonal
compressive stresses increase with a change in the joint shear transfer mechanism after beam
T
s1
C
s1
C
c2
C
s2
T
s2
C
c2
V
c1
V
c2
V
jh
=T
s1
+T
s2
V
c1
T
s1
C
s2
Bond stress
Rotation due to bar slip
(Bertero and Popov 1977)
Diagonal Deformation of BeamColumn Connection
Strains in Joint Lateral Reinforcement
yielding, and (c) the beam deformation increases due to the bar slip within a joint.
Examples of lateral loaddeformation relation of beamcolumn subassemblages are shown with
strain distribution of girder reinforcement within the connection.
If the column width to bar diameter ratio is small as shown in case (a), the bond resistance along
the reinforcement within the connection deteriorates at a faster rate. The stress supposedly in
compression in flexure changes its sign to tension, and beam bars slip in the connection. The
hysteresis of the beamcolumn assembly exhibits the pinching shape even at a low stress stage. If
the column width to bar diameter ratio is large and low strength steel is used for beam bars, bond
stress can be kept low along the beam bars within the connection as shown in case (b) , and fat
spindleshape hysteresis shape is observed even after beam yielding. If the column width to bar
diameter ratio is large, but the strength of the beam bars is high as shown in case (c), bond stress
increases near the beam yielding, and bond resistance deteriorates within the connection, developing
a pinching hysteresis shape.
It is, therefore, desirable to use lowstrength smalldiameter bars for the beam longitudinal
reinforcement to develop good performance in the beamcolumn joint. However, it is not practical
from the construction point of view. The use of largediameter and highstrength bars is demanded to
lower the construction cost. A reasonable compromise is necessary.
Story drift, mm
Strain distribution in beam bars
Column
(a) Specimen with Small column width/bar diameter ratio
S
t
o
r
y
s
h
e
a
r
,
t
f
(b) Specimen with large column width/bar diameter ratio (low yield stress bars)
Column
Strain distribution in beam brs
Story drift, mm
S
t
o
r
y
s
h
e
a
r
,
t
f
S
t
r
a
i
n
,
%
S
t
r
a
i
n
,
%
Strain distribution in beam bars
Column
Story drift, mm
S
t
o
r
y
s
h
e
a
r
,
t
o
n
f
S
t
r
a
i
n
,
%
(c) Specimen with large column width/bar diameter ratio (high yield stress bars)
Effect of column width to bar diameter ratio on bond resistance deterioration
Kitayama et al.
Transverse Beams: Joint shear failure accompanies the increase in the volume of the connection.
To the extent that transverse beams can restrain this volume increase, the transverse beams are
expected to enhance the joint shear resistance.
Kitayama et al. (1991) tested a threedimensional
interior beamcolumn joint and compared with the
joint strength of a planer joint. Both specimens
have the same dimension and reinforcement. The
threedimensional specimen was preloaded in the
transverse direction causing flexural cracks at the
faces of the column before loading in the principal
direction. The threedimensional specimen yielded
in flexure and no failure was observed in the
beamcolumn connection. The planer specimen
failed in shear compression in the connection. The
transverse beams confine the connection and
enhance the strength of the joint even after the
formation of cracks at the joint boundary.
Slab Effect: Floor slabs, whose concrete is cast
monolithically with beam concrete, have two
effects; (a) slab reinforcement contributes to the
flexural resistance of the beam, and (b) slabs
confine the interior joint.
Column Axial Force: Park and Paulay (1975) expect
beneficial effects of axial force on the joint performance and
joint shear reinforcement. A steeper diagonal compression
strut may form as a result of an enlarged compression block
across the column section. The horizontal bond force along
the beam bars can develop within the wider diagonal
compression strut.
Kitayama, Otani and Aoyama (1987) studied the
influence of column axial load on the bond stress transfer
along the beam reinforcement in a joint and reported the
relation of column compressive stresses normalized by the concrete compressive strength and
energy dissipation expressed in the form of equivalent viscous damping ratio h
eq
at a story drift angle
of 1/50 rad for planar interior beamcolumn joint specimens tested in Japan. Solid circles represent
specimen with beam bar bond index u
b
/f ’
c
less than 4.5. Test results are scattered widely regardless
Story drift
Stroy drift angle, rad
J
o
i
n
t
s
h
e
a
r
s
t
r
e
s
s
/
C
o
n
c
r
e
t
e
S
t
r
e
n
g
t
h
Effect of transverse beams on joint strength
Kitayama et al., 1989
Effect of axial force on joint behavior
Effect of column axial force on energy dissipation (Kitayama, Otani and Aoyama, 1987)
of column axial stress level. Therefore,. it is considered that column axial stress smaller than 0.3 f ’
c
does not exhibit beneficial effect on the bond resistance along the beam bar within a joint.
Column axial stress level is compared with the maximum joint shear stress normalized by concrete
compressive strength for plane beamcolumn joint specimens, failed in joint shear. These specimens
were tested in Japan and U.S. Column axial load does not seem to influence the joint shear strength.
High axial compression load in a column, however, accelerates the strength decay in the diagonal
compression failure of the joint core concrete after beam flexural yielding.
References:
ACIASCE Committee 352, “Recommendations for Design of Beamcolumn Joints in Monolithic
Reinforced Concrete Structures, Journal., American Concrete Institute, Vol. 82, No. 3, MayJune
1985, pp. 266  283.
Bertero, V. V., and E. P. POPOV, “Seismic Behaviour of Momentresisting Reinforced Concrete
Frames,” ACI Sp53, Reinforced Concrete Structures in Seismic Zones, American Concrete
Institute, 1977, pp. 247292.
Bonacci, J and S. Pantazopoulou, “Parametric Investigation of Joint Mechanics,” Structural Journal,
American Concrete Institute, Vol. 90, No. 1, JanuaryFebruary 1993, pp. 61  71.
Cheung, P. C., et al., “Mechanism of Slab Contribution in Beamcolumn Subassemblages,”
ACISP123, Design of Beamcolumn Joints for Seismic Resistance, American Concrete Institute,
Detroit, Michigan, 1991, pp. 259 0 289.
Durrani, A. J., and J. K. Wight, “Behavior of Interior Beamtocolumn Connections under Earthquake
Type Loading,” Journal, American Concrete Institute, Vol. 82, No. 3, MayJune 1985, pp. 343 
350.
Ehsani, M. R., and J. K. Wight, “Exterior Reinforced Concrete Beamtocolumn Connections
subjected to Earthquaketype Loading,” Journal, American Concrete Institute, Vol. 82, JulyAugust
1985, pp. 492  499.
Fujii, S., and S. Morita, “Comparison between Interior and Exterior RC Beamcolumn joint Behavior,”
ACISP123, Design of Beamcolumn Joints for Seismic Resistance, American Concrete Institute,
Detroit, Michigan, 1991, pp. 145  165.
Hanson, N. W., and H. W. Connor, “Seismic Resistance of Reinforced Concrete Beamcolumn
Joints,” Journal, Structural Division, ASCE, Vol. 93, No. 5, October 1967, pp. 533  560.
Effect of column axial force on joint shear resistance
Kitayama, Otani and Aoyama, 1987
Joh, O., et al., “Influence of Transverse Joint and Beam Reinforcement and Relocation of Plastic
Hinge Region on Beamcolumn Joint Stiffness Deterioration,” ACISP123, Design of Beamcolumn
Joints for Seismic Resistance, American Concrete Institute, Detroit, Michigan, 1991, pp. 187  223.
Kitayama, K., S. Otani and H. Aoyama, “Development of Design Criteria for RC Interior Beamcolumn
Joints,” ACISP123, Design of Beamcolumn Joints for Seismic Resistance, American Concrete
Institute, Detroit, Michigan, 1991, pp. 97  123.
Kurose, Y., “Recent Studies on Reinforced Concrete Beam Column Joints in Japan,” Report 878, Phil
M. Ferguson Structural Engineering Laboratory, Department of Civil Engineering, University of
Texas at Austin, Texas, 1987.
Liande, Z., and J. O. Jirsa, “A Study of Shear Behavior of Reinforced Concrete Beamcolumn Joints,”
Report No. 821, P. M. Ferguson Structural Engineering Laboratory, Department of Civil
Engineering, University of Texas at Austin, February 1982.
Meinheit, D. F., and J. O. Jirsa, “Shear Strength of R. C. Beamcolumn Connections,” Journal,
Structural Division, ASCE, Vol. 107, No. 11, November 1981, pp. 2227 2244.
Noguchi, H. and Kurusu, “Experimental Study on Seismic Resistance of Reinforced Concrete
Beamcolumn Connections (in Japanese),” Report CII, Annual Meeting, Architectural Institute of
Japan, Kyusyu, October 1997, pp. 627  628.
Otani, S., “The Architectural Institute of Japan Proposal of Ultimate Strength Design Requirements for
RC Buildings with emphasis on Beamcolumn Joints,” ACISP123, Design of Beamcolumn Joints
for Seismic Resistance, American Concrete Institute, Detroit, Michigan, 1991, pp. 125  144.
Pantazopoulou, S., and J. Bonacci, “Consideration of Questions about Beamcolumn Joints,”
Structural Journal, American Concrete Institute, Vol. 89, No. 1, JanuaryFebruary 1992, pp. 27 
36.
Paulay, T., “Equilibrium Criteria for Reinforced Concrete Beamcolumn Joints,” Journal, Structural
Engineering Division, ASCE, Vol. 86, No. 6, Nov.Dec. 1989, pp. 635  643.
Paulay, T., and R. Park, “Joints in Reinforced Concrete Frames Designed for Earthquake
Resistance,” Research Report 849, Department of Civil Engineering, University of Canterbury,
Christchurch, New Zealand.
Paulay, T., et al., “Reinforced Concrete Beamcolumn Joints under Seismic Actions,” Journal,
American Concrete Institute, Vol. 75, No. 11. November 1978, pp. 585  593.
Uzumeri, S. M., “Strength and Ductility of Cast in Place Beam Column Joints,” ACISP53, Reinforced
Concrete Structures in Seismic Zones, American Concrete Institute, Detroit, Michigan, 1977, pp.
293  350.
Sugano, S., et al., “Behavior of Beamcolumn Joints using Highstrength Materials,” ACISP123,
Design of Beamcolumn Joints for Seismic Resistance, American Concrete Institute, Detroit,
Michigan, 1991, pp. 359  377.
Zhang, S., and J. O. Jirsa, “A Study of shear Behavior of Reinforced concrete Beamcolumn Joints,”
Report No. 821, P. M. Ferguson Structural Engineering Laboratory, Department of Civil
Engineering, University of Texas at Austin, 1982.
Zhu, S., and J. O. Jirsa, “A Study of Bond Deterioration in Reinforced Concrete Beamcolumn Joints,”
Report No. 831, P. M. Ferguson Structural Engineering Laboratory, Department of Civil
Engineering, University of Texas at Austin, 1983.
3.4 Behavior of Exterior Beamcolumn Connections
The shear acting in an exterior beam column joint is normally smaller than that in an interior beam
column joint because only one beam is connected to the joint. The anchorage detail of beam
longitudinal reinforcement affects the performance of the exterior joint such as (a) horizontal
development length, radius of 90degree hook and vertical length after the hook.
Park and Paulay (1975) pointed out that diagonal tension and compression stresses are induced
in the exterior beamcolumn joint panel. The diagonal tension may be high when the ultimate capacity
of the adjoining members is developed, and this can lead to extensive diagonal cracking. The severity
of diagonal tension is influenced by the amount of flexural steel and the magnitude of the axial force
on the column.
The beam top reinforcement is anchored in relatively weak concrete; the surrounding concrete is
subjected to sedimentation. A splitting crack can easily form along these bars at a relative early stage
of loading, and the bond resistance may be lost in the horizontal anchorage portion. Major anchorage
resistance arises from the high bearing stress in the corner of the bent. Therefore, the horizontal
portion of anchorage length may be ignored in design. The condition of the beam bottom
reinforcement is better. The outer column bars are also subjected to high bond stress due to the
change of stress required to resist bending at the top and bottom boundary. of the joint.
The role of lateral reinforcement is important in the exterior beam column joint in resisting diagonal
tension and confining the core concrete. When the transverse shear reinforcement across diagonal
cracks of joints commences to yield, disintegration of the concrete begins because of the repeated
opening and closing of cracks and sliding movement along the diagonal shear cracks. Full shear
expected in the joint must be resisted by the lateral reinforcement. Column width is important effect in
shear resistance of the exterior joint; i.e., the column width is necessary to increase the angle of
diagonal strut.
The level of axial force is important on the performance of an exterior joint.
References:
Paulay, T., and R. Park, “Joints in Reinforced Concrete Frames Designed for Earthquake
Resistance,” Research Report 849, Department of Civil Engineering, University of Canterbury,
Christchurch, New Zealand.
Seckin, M., and S. M. Uzumeri, “Examination of Design Criteria for Beamcolumn Joints,”
Proceedings, Sixth European Conference on Earthquake Engineering, Dubrovnik, 1978.
Actions in exterior beamcolumn joint
(Park and Paulay, 1975)
3.5 Behavior of Structural Walls
Introduction: A structural wall is a stiff element and attracts large lateral load, especially shear,
during an earthquake. Therefore, a structural wall was often called as a shear wall. The structural wall
provides large lateral load resistance to a structure. The effectiveness of structural walls in
earthquake resistance has been demonstrated in many earthquakes; e.g., the 1985 Mexico
Earthquake, the 1985 Chile Earthquake, the 1992 Erzincan, Turkey, Earthquake. The basic roles of
structural walls in earthquake resistant structures are listed by Paulay (1981);
(a) To provide adequate stiffness to protect the structure against damage particularly in
nonstructural components during moderate seismic disturbances.
(b) To provide adequate strength to limit structural damage to superficial level under
codespecified design forces even though some nonstructural damage is expected.
(c) To provide adequate structural ductility and capability to dissipate energy under the largest
disturbance expected in the region.
In order to fulfill the roles, Paulay strongly advocates the design of structural walls to develop flexural
yielding at the base.
A structural wall, even of rectangular cross section, has boundary elements, where large amount
of vertical longitudinal reinforcement is placed to resist bending moment. The longitudinal
reinforcement in a boundary element must be confined by properly detailed lateral reinforcement. The
role of the lateral reinforcement is (a) to maintain the effectiveness of core concrete to a large plastic
strain, (b) to support the vertical reinforcement against inelastic buckling, and (c) to enhance shear
carrying capacity of the boundary element.
The existence of structural walls sometimes causes eccentricity of stiffness relative to the mass
center. Proper arrangement of structural walls must be considered at the structural planning stage.
Failure Modes: The failure of structural walls take place in the following mode, (a) flexure, (b)
diagonal tension, (c) sliding shear, (d) hinge sliding, (e) inadequate anchorage or splicing of
reinforcement, (f) buckling of compressive reinforcement. Concrete, being a relatively brittle material
that show s rapid strength degradation in both compression and shear when subjected to repeated
inelastic strains and multidirectional cracking, should not be considered in structural walls as a
significant source of energy dissipation. An obvious source of hysteretic energy dissipation should be
the yielding of the principal flexural reinforcement (Paulay, 1981).
A structural wall in a lowrise building tends to fail in shear, and more research was conducted on
the shear behavior of structural walls. With a development of highrise reinforced concrete
construction, flexural behavior of a structural wall was studied. Typical crack patterns at failure are
illustrated for a firststory structural wall. Shear crack pattern in (a); shear dominant failure in (b) to
(d); flexural crack pattern in (e); and flexure dominant failure in (f) and (g).
Failure modes of structural walls （Paulay, 1981）
Although it is desirable for a structural wall to behave in dominantly the flexure mode, it is difficult
to expect such behavior from a squat wall with height to width ration less than 2.0. A significant
portion of horizontal shear is transmitted directly to the foundation by diagonal compression.
In a shear dominated failure, diagonal cracks are spaced regularly at an angle of approximately 45
degrees. When the diagonal shear cracks develop in a structural wall, the entire shear must be
resisted by the horizontal shear reinforcement. If the amount of shear reinforcement is not sufficient,
a diagonal tension failure plane may develop, and the widened diagonal shear cracks lead to
diagonal tension failure. When the average shear stress in the wall section is large and when
adequate horizontal shear reinforcement has been provided, failure may be cause by crushing of
concrete under diagonal compression. The diagonal compression failure is observed when the
flexural capacity is large or when the diagonal shear cracks develop under load reversals. Diagonal
compression failure sometimes develops suddenly with a rapid propagation of the crushing zone of
the concrete. A few cycles of reversed loading sometimes cause significant yielding of flexural
reinforcement; horizontal cracks may penetrate full depth of the wall base section. Sliding
(a) Shear cracks
(e) Flexural crakcks
(b)Shear compression failure
(f) Flexural failure
(c)Shear compression failure (g) Sliding shear failure
(d) Sliding shear failure
Failure modes of structural walls
displacement can develop along the flexural cracks. The process of this sliding shear mechanism is
described by Paulay (1981). In the first cycle, involving large flexural yielding, the major part of the
shear force at the base of a cantilever squat wall is transmitted across the flexural compression zone.
This compression zone is small due to light axial load in the lowrise wall. Upon load reversal, cracks
develop across the previous flexural compression zone, while bars are subjected to compressive
stresses. Until the base moment reaches a level sufficient to yield these bars, in compression, a
continuous wide crack develop along the base. The shear must be resisted by dowel action of the
vertical reinforcement. Because of the relatively flexible nature of this mechanism, large sliding
displacement takes place until yielding of the compression steel occurs, closing the crack at the
compression end. With subsequent inelastic load reversals, further deterioration of the shear friction
mechanism along the planes of potential sliding develops. Due to a deterioration of the bond transfer
along the vertical bars and due to the Bauschinger effect, the stiffness of the dowel shear mechanism
also reduces. Eventually the principal mode of shear transfer along the base is by kinking of the
vertical bars.
In a flexure dominated failure, initially horizontal flexural cracks change their direction toward
flexural compression zone at the base of the structural wall, forming a fanshape crack pattern. Note
that compression force acts in the direction of cracks. Compression struts are formed uniformly at an
angle of 45 degrees in a shear dominated wall, whereas compression struts are formed toward
flexural compression zone. Failure may result either by yielding of vertical web reinforcement or by
crushing of concrete.
Shear failure modes of squat wall Sliding shear failure of squat wall
Paulay (1981)
Deformation: Lateral deformation of a member is often divided into flexural and shear deformation,
similar to elastic problem. However, the two deformation cannot be clearly separated after cracking;
e.g., a diagonal crack developed by shear stress causes rotational deformation.
Flexural deformation can be evaluated by integrating rotational deformation along a member.
Therefore, shear deformation is often defined as the difference of the total deformation and flexural
deformation.
Testing of a structural wall in a laboratory is difficult because the resistance is large. Most
laboratory test data have been obtained for singlestory singlespan walls to study shear
characteristics.
Shear Stressdeformation Relation: A typical average shear stressshear deformation relation
shows stiffness reduction at the formation of large inclined crack along the main diagonal; and further
stiffness reduction with opening of the main diagonal crack and formation of additional inclined cracks.
The structural wall fails when the main diagonal crack penetrates through the boundary columns,
leading either shear compression failure in the compression zone of the wall panel and the boundary
column or sliding failure along a horizontal plane of the wall panel accompanied by shear failure of the
boundary columns. The average shear stress may be calculated for the area defined by
centertocenter distance of the boundary columns and thickness of the wall.
Shear Cracking: A diagonal shear crack may be assumed to develop when the principal tensile
stress of section reaches the tensile strength of concrete in the elastic analysis.
Shear stress reported in a structural wall test does not agree with the result from the elastic theory
by the following reason;
(a) the stressstrain relationship of concrete is not linearly elastic before the formation of a crack,
(b) shrinkage stress exists in a reinforced concrete member before the test,
(c) flexural crack may form prior to shear crack, and
(d) a crack is normally reported after it becomes visible.
Sugano (1970) developed an
empirical formula to estimate average
shear stress at shear cracking for wall
test data using sheartype loading:
B g cr
p σ τ ) 051 . 0 3 . 4 ( + =
where
cr
τ : average shear stress at
shear cracking,
B
σ : compressive
strength of concrete, p
g
: gross vertical
reinforcement ratio of boundary columns.
The coefficient of correlation is 0.73.
Axial stress is known to increase the
resistance against cracking; The shear
cracking stress
*
cr
τ may be modified to
include the effect of axial stress acts
o
σ in a wall section;
2 / 1 2 *
) (
cr o cr cr
τ σ τ τ + =
Deformation at shear cracking may be calculated on the basis of the elastic theory using shear
modulus G
c
) 1 ( 2 µ +
=
c
c
E
G
Column gross reinforcement ratio, %
Shear cracking stress and column reinforcement ratio
(Sugano, 1970)
where E
c
: elastic (Young's) modulus of concrete, µ : Poisson's ratio (approximately 1/6 for concrete).
Secant modulus is normally used for the elastic modulus of concrete.
Shear deformation (angle) γ is given as
c
c
G
κτ
γ =
where κ : shape factor for shear deformation. The shape factor κ may be obtained by equating the
external work done by external shear V and the internal work done by shear stress obtained by the
elastic theory. The shear angle is defined as lateral deformation at wall centerline divided by the wall
height.
Ultimate Shear Strength:
Cracking dilates a wall panel,
but boundary girders above and
below the panel and boundary
columns on both side give
confinement to the panel.
Therefore, the shear strength
increases with the amount of
longitudinal reinforcement in
the girders and boundary
columns.
Sugano (1970) plotted the
average shear stress
u
τ
observed at failure in cantilever
wall tests and calculated at
ultimate flexural strength.
Ultimate flexural strength was
calculation at an extreme
compressive fiber strain of
0.004 assuming plane section
to remain plane after
deformation. Different symbols
were used for the amount of wall reinforcement ratio p
w
(%). A strong correlation can be observed in
the figure; flexural yielding of boundary members reduce the confining effect on the panel and
triggers shear failure in the boundary columns. A lightly reinforced shear panel tends to fail at
approximately 80 percent of average shear stress
mu
τ calculated at flexural strength.
Sugano (1970) also plotted the relation between the average shear stress observed at shear
failure and the product of lateral reinforcement ratio of a wall and yield stress of wall reinforcement.
The ultimate shear stress increases with the amount of lateral reinforcement, but the rate of increase
in shear resistance is smaller than the amount of shear reinforcement (p
w
wy
σ ).
Hysteresis Behavior: The hysteresis shape of reinforced concrete structural walls, which behavior is
dominated by flexure, can be stable and fat (Paulay and Spurr, 1977).
Reinforced concrete members can be relatively ductile in shear during monotonic loading, but they
generally do not develop suitable performance under inelastic load reversals. The hysteresis shape
exhibits pinching phenomenon. The shear resistance can be attained only when the subsequent
largest displacement is attained and the stiffness decreases with displacement amplitudes. Note that
the envelope curve of the hysteresis curves follows closely the loaddisplacement curve under
monotonically increasing displacement. Inelastic tensile strains in lateral reinforcement cannot
recovered at complete unloading and accumulates with reversed loading. Special reinforcement to
Calculated shear stress at flexural resistance, kgf/cm
2
S
h
e
a
r
s
t
r
e
s
s
o
b
s
e
r
v
e
d
a
t
s
h
e
a
r
f
a
i
l
u
r
e
,
k
g
f
/
c
m
2
Average shear stress observed at shear failure
and calculated flexural resistance (Sugano, 1970)
suppress inelastic shear distortion is desired such as diagonal shear reinforcement.
Threedimensional Effect of Wall: Flexural
behavior of a structural wall is similar to a girder or
a column; i. e., flexural crack and flexural yielding
take place before failure in compression zone. Due
to the flexural cracking, the neutral axis of a
section shifts well to compression side, causing a
significant elongation in the boundary column in
tension, while deformation in the boundary column
in compression is small. Consequently, a girder
end connected to the tensile boundary column is
lifted upward, and girders in the plane of the wall
and orthogonal to the wall resist the upward
movement by introducing downward force to the
boundary column. Due to this additional vertical
force in the tensile boundary column, the yielding
of longitudinal reinforcement in the tensile
boundary column is delayed. This behavior was
first reported in the test of a fullscale sevenstory
reinforced concrete building (Yoshimura and
Kurose, 1985).
Base Uplifting Rotation at Wall Base: A
structural wall attracts large lateral load due to its
high lateral stiffness. A large overturning moment
at the base of a structural wall sometimes causes
tensile stress at its footing and uplifts the tensile
end. The rocking at the base of structural walls is
known to be beneficial in dissipating kinetic energy of the structure without much damage (Priestley et
al., 1978).
Three fifthscale twostory threebay reinforced concrete frame with a structural wall in the central
bay were tested. The base of the wall was supported on footing foundation. Variables in the test were
failure modes (flexural yielding of shear failure) of connecting girders and support conditions (flexible
support or rigid support) at the wall base. Vertical loads were assumed to act on columns and lateral
loads were concentrated at the second floor level. The wall was designed to uplift at the tension side
under the seismic forces prior to the formation of yield mechanism of the system.
Vertical loads were applied at the top of columns and wall simulating the gravity loads. Horizontal
reversal force was applied at the top of the wall simulating seismic actions.
Flexure dominated wall (Paulay and Spurr, 1977) Shear dominated wall (Oesterle et al, 1976)
Threedimensional effect of structural walls
Charney, 1991
The lateral loaddisplacement relation of the three specimens is compared. The three specimens
developed “yielding behavior” by the formation of a yield mechanism of the system. A steep stiffness
upon unloading and small residual displacement at complete unloading was observed. After the
unloading was completed, the high stiffness was recovered during the reloading. Small energy was
dissipated by the specimen with shear failing girders; energy dissipation of the system was controlled
by the frame. The uplifting of the wall commenced at a story drift as small as 1/400. Shear input to the
wall was limited by the uplifting and the damage in the wall was relatively small compared with a wall
with the fixed base.
References:
Bertero, V.V., E. P. Popov, T. Y. Wang, and J. Vallenas, “Seismic Design Implication of Hysteretic
Behavior of Reinforced Concrete Structural Walls,” Proceedings, Sixth World Conference on
Earthquake Engineering, New Delhi, 1977, Reprints 5, pp. 159  165.
Cardenas, A. E., and D. D. Mugura, “Strength of High Rise Shear Walls  Rectangular Cross
Sections,” ACI SP36, Response of Multistory Concrete Structures to Lateral Forces, American
Concrete Institute, 1973, pp. 119  150.
Charney, F. A., "Correlation of the Analytical and Experimental Inelastic Response of a 1/5Scale
SevenStory Reinforced Concrete FrameWall Structure," ACI SP127, EarthquakeResistant
Concrete Structures, Inelastic Response and Design, American Concrete Institute, Detroit, 1991.
Kato, D., S. Otani, H. Katsumata and H. Aoyama, "Effect of Wall Base Rotation Behavior on
Reinforced Concrete FrameWall Building," Proceedings, Third Pacific Regional Conference on
Earthquake Engineering, Victoria University of Wellington, New Zealand, May 1983.
Oesterle, R. G., A. E. Fiorato, L. S. Johal, J. E. Carpenter, H. G. Russell, and W. G Corley,
“Earthquake Resistant Structural Walls,  Tests of Isolated Walls,” Report to National Science
Foundation, Portland Cement Association, Illinois, November 1976, 44 pp.
Paulay, T., “The Design of Reinforced Concrete Ductile Shear Walls for Earthquake Resistance,”
Research Report 811, Department of Civil Engineering, University of Canterbury, New Zealand,
February 1981, 72 pp.
Priestley, M. J. N., R. J. Evison, and A. J. Carr, “Seismic Response of Structures Free to Rock on their
Foundation.” Bulletin, New Zealand National Society for Earthquake Engineering, Vol. 11, No. 3,
September 1978, pp. 141  150.
Sugano, S, "Experimental Study on Restoring Force Characteristics of Reinforced Concrete
Members (in Japanese)," Thesis submitted to fulfill the requirements of Doctor of Philosophy,
University of Tokyo, March 1970.
Synge, A. T., T. Paulay, and M. J. N. Priestley, “Ductility of Squat Shear Walls,” Research Report 808,
Department of Civil Engineering, University of Canterbury, 1980, 150 pp.
Yoshimura, M. and Y. Kurose, "Inelastic Behavior of the Building," ACI SP84, Earthquake Effects on
Reinforced Concrete Structures, U.SJapan Research, American Concrete Institute, Detroit, 1985,
pp. 163202.
Home Assignment No. 2
20011024 Otani, S.
Discuss the following topics on typical behavior of reinforced concrete members:
(1) Loaddeflection relation of a beam with different amount of longitudinal reinforcement at the top
and at the bottom, typically seen in a Tshaped beam
Reference: Ma, S. M., V. V. Bertero and E. P. Popov, "Experimental and Analytical Studies of the
Hysteretic Behavior of Reinforced Concrete Members under Bidirectional Reversed Lateral Loading,"
Report No. EERC 762, University of California at Berkeley, 1976.
(2) Effect of slab on the stiffness and flexural resistance of a beam.
Reference: French, C. W., and A. Boroojerdi, "Contribution of RC Floor Slabs in Resisting Lateral
Loads," Journal, Structural Engineering, ASCE, Vol. 115, No. 1, January 1989, pp. 118.
(3) Role of compressive reinforcement on flexural resistance and deformation capacity.
1
Chapter 4 Analysis of Reinforced Concrete Members
The minimum modeling element of nonlinear analysis of a structure can be selected as (a)
material level modeling (Finite Element Method), (b) member level modeling (frame modeling), (c)
story level modeling (massandspring modeling), and (d) onedegreeoffreedom model for an entire
structure. The finite element method (FEM) is used for material level modeling. The story level
modeling by a massandstick model was extensively used to understand the overall behavior of a
structure. Recent capacity and demand spectra method idealizes a structure by a
singledegreeoffreedom system.
The FEM has been developed for a half century, and is popularly used in the analysis of
reinforced concrete member and subassemblage even in a nonlinear range to failure. The nonlinear
FEM generally requires (a) modeling of constitutive relations for plain or reinforced concrete and
reinforcement under generalized multiaxial loading and loading reversals, (b) modeling of bar slip
with respect to surrounding concrete, (c) modeling of crack initiation and propagation, and (d)
modeling of stress transfer at crack interface. The method can represent the detailed geometry of the
structure, the history of stresses and strains at every point. The constitutive modeling of materials
developed significantly in the past decades. The FEM, however, is not practical for use in the
nonlinear earthquake response analysis of an entire building structure due to the limitation in
computational and memory requirements.
The member level modeling may not represent the detailed geometry of reinforcement in a
section, but it can reflect the basic features of member characteristics. The stiffness analysis of a
member is normally based on flexural response, and ignores the shear and bond stiffness
degradation because the two modes of response are prevented in design process. The method can
describe the damage distribution and intensity within a structure. A fiber model of section can be
combined with the member level modeling to represent the geometry of section to a limited extent.
A closecoupled shearbeam model or farcoupled bending beam model is used to represent
stiffness at each story in the massandstick model analysis. This type of modeling is sometimes
useful to estimate the story drift of a structural system from different earthquake motions.
Therefore, the nonlinear earthquake response analysis of a building is normally carried out on the
basis of memberbymember modeling. This chapter reviews the analysis of reinforced concrete
section and members.
2
4.1 Flexural Analysis of Section
The momentcurvature relation of a cross section provides the basis of understanding the
nonlinear behavior of reinforced concrete members, such as cracking, crushing and spalling of the
concrete, and yielding and buckling of longitudinal reinforcement.
Assumptions: Reinforced concrete section under bending is normally analyzed using the following
simplifying assumptions;
(a) Plane section before bending remains plane and normal to the member axis after bending,
(b) Normal stressstrain relation (constitutive model) of materials is known, and
(c) External forces are equal to internal forces of the section.
The first assumption, often called the Bernoulli's hypothesis, simplifies the analysis and gives
linear distribution of longitudinal strain across the section with null strain at the neutral axis. The
location of the neutral axis of the section is determined by the equilibrium of axial force acting in the
section.
Bernoulli’s Assumption: Eivind Hognestad (1951)
reported the results of strain measurement in a
reinforced concrete section near failure region of tied
columns (254 mm square) and spiral columns (305
mm round) under the combined axial load and
bending moment. Strains in the reinforcement and on
the concrete surface were measured by SR4 wire
strain gages.
He noted some deviation of observed strains from
the linearity due to inaccuracy in individual strain
measurements and to small errors in the location of
gage lines. He concluded that a reasonable
agreement existed between strains measured in
reinforcement and on the concrete surface. He noted
that the departures from linearity appeared to be
inconsistent, indicating accidental or local rather
than systematic variations. Therefore, he
assumed that “Bernoulli’s hypothesis is valid.”
The neutral axis of a section is defined as a
level where no extension nor compressive
deformation takes place. It should be noted that
the neutral axis moves to the compression side
with loading especially after initial cracking. The
slope of linear strain distribution is equal to the
curvature at the section, which is the second
derivative of deflection.
Local bond stresses between the
reinforcement and concrete and local bar slips
relative to the surrounding concrete exist near
tension cracks. Therefore, the Bernoulli’s
assumption is not valid, in the exact sense, near a
crack, shifting the neutral axis away from the
geometrical centroid. The curvature also varies
along a member with the varying depth of the
neutral axis. However, it is believed that the
assumption holds on the average over a finite
region. Slip between concrete and reinforcing
steel may be ignored if deformed bars are used as
E. Hognestad (1951)
o z
y ε φ −
y
z
o
ε
z
φ
( , ) y z ε
Section
Strain
o
ε
y
φ
S
t
r
a
i
n
o y
z ε φ +
Coordinate system and strain in section
3
the longitudinal reinforcement.
Let us take the righthand coordinate system with xaxis in the direction of a horizontal member,
and yaxis in the vertical direction and zaxis in the horizontal direction.
Strain: For the small deformation, the normal strain ( , ) y z ε at coordinate ( , ) y z in the section can
be expressed as
0
( , )
z y
y z y z ε ε φ φ = − +
where,
0
ε : normal strain at the geometrical centroid of section, ,
y z
φ φ : curvature about the
centroidal axes y and z .
Stress: Using the stressstrain relations of constitutive
materials, normal stress ( , )
x
y z σ can be determined
for a given normal strain ( , )
x
y z ε at coordinate
( , ) y z from the origin.
( , ) ( ( , ))
x x
y z f y z σ ε =
Equilibrium of Forces: Knowing stresses within a
section, resultant axial force and bending moments
about the geometric centroid are calculated by
integrating the stress and moments of stress over the
section:
( ) ( , )
( ) ( , )
( ) ( , )
x
Cross Sectional Area
z x
Cross Sectional Area
y x
Cross Sectional Area
N x y z dA
M x y z y dA
M x y z z dA
σ
σ
σ
=
= −
=
∫
∫
∫
( )
( ) ( )
( )
y
R z
M x
S x M x
N x
¦ ¹
¦ ¦
=
´ `
¦ ¦
¹ )
It is important to note that all structural members are represented by line elements at the
geometric centroid section in the frame analysis; i.e., moment of a member in the structural analysis
and bending moment resistance of the section must be calculated about the common axis.
MomentCurvature Relation under Monotonically Increasing Load: Uniaxial momentcurvature
relationship of section under a constant axial load N is calculated for a given curvature
z
φ rather
than for a given moment
z
M . Calculation of moment for a given curvature requires an iterative
procedure because the strain at the center of the section is not known.
The following procedure is normally taken to evaluate moment
z
M of a section under existing
axial load N and a given curvature
z
φ ;
(a) Assume strain
cent
ε at the center of the section; which enables us to define strain profile
across the section;
Strain
S
t
r
e
s
s
( ) f σ ε =
Stressstrain relation of material
4
( )
cent
y y ε ε φ = −
in which ) ( y ε : strain at distance y from the section center.
(b) Determine stress profile across the section on the basis of stressstrain relation of materials;
)) ( ( ) ( y y ε σ σ =
(c) Resultant axial force
cal
N is estimated by integrating normal stress over the section;
∫
=
Area Sectional Cross
cal
dA y N ) ( σ
(d) This resultant axial force must be equal to the existing axial load of the section if the choice of
the assumed strain
o
ε at the center is correct. Trial and error method with interpolation is used to
iterate steps (a) to (c) until the resultant axial load becomes practically equal to the existing axial load.
(e) Bending moment
z
M is calculated for the correctly assumed strain at the center of the
section;
( )
z
Cross Sectional Area
M y y dA σ = −
∫
Stressstrain Relation of Concrete: Various stressstrain relations have been used in the flexural
analysis of reinforced concrete section.
Kent and Park (1971), for example, used the following expression for concrete stressstrain
relation under monotonically increasing compression load;
o c
o c o c o c
o c
o
c
o
c
o c
but
for Z
for
σ σ
ε ε ε ε σ σ
ε ε
ε
ε
ε
ε
σ σ
2 . 0
)] ( 1 [
] ) ( 2 [
2
≥
> − − =
≤ − =
where
c
σ : concrete stress,
o
σ : compressive
strength (MPa) of concrete,
c
ε : concrete strain,
o
ε : strain at compressive strength of concrete
(=0.002). The slope of descending part was
controlled by a parameter Z in the following
manner;
20
ε
50c
ε
50u
ε
0.002
o
ε =
o
σ
0.5
o
σ
0.2
o
σ
50h
ε
c
ε
c
σ
Confined
Unconfined
Stressstrain relation for concrete
under monotonic loading
Park and Kent Model, 1971
5
h
s h
o
o
u
o h u
s
b
p
Z
"
)
4
3
(
89 . 6
002 . 0 021 . 0
5 . 0
50
50
50 50
=
+
+
=
− +
=
ε
σ
σ
ε
ε ε ε
where
s
p : ratio of volume of transverse reinforcement to volume of concrete core measured to
outside of hoops, b": width of confined core measured to outside of hoops,
h
s : spacing of hoops.
Stiffness of descending branch was varied taking into account the confining effect of concrete by
lateral reinforcement.
Tensile stress of the concrete may be considered before the first cracking in the section, but the
tensile stress may be ignored after the initial cracking.
Stressstrain Relation of Reinforcing Steel:
Stressstrain relation of reinforcing steel is normally
represented by elasticperfectly plastic relation
with/without strain hardening branch. Young’s
modulus E
s
of steel is approximately 205 GPa.
( )
s s s s sy
s sy sy s sh
s sy sh s sh sh s
E
E
σ ε ε ε
σ σ ε ε ε
σ σ ε ε ε ε
= ≤
= ≤ ≤
= + − ≤
where,
s
E : elastic modulus of steel, ,
y sh
ε ε : strains
of reinforcement at yielding at the initiation of strain
hardening. The stressstrain relation is normally
assumed to be the same in tension and
compression.
Numerical Procedure: The bending momentcurvature relation is analytically obtained by a
numerical method. It is not practical to calculate the curvature for a given bending moment. Bending
moment is normally calculated for a given curvature.
For a given curvature and axial force in the section, the strain at the geometrical centroid of the
section is assumed. With the curvature and the strain at the geometrical centroid, the strain
distribution is uniquely defined across the depth of the section. Stress at each point in the section is
determined from the stressstrain relations of materials for the strain at the point. Normal stresses are
summed up to calculate an axial force corresponding to the assumed strain at the geometrical
centroid. If the calculated axial force is not equal to the existing (given) axial force, the assumed strain
at the geometric centroid is not correct. An iterative procedure is used until the assumed strain gives
the resultant axial force equal to the existing axial force in the section. It is convenient to plot the
resultant axial force in the vertical axis and the assumed strain in the horizontal axis, and use linear
interpolation or extrapolation to estimate the strain in the next iteration.
Depth to the neutral axis may be chosen as a variable instead of the strain a the geometrical
centroid.
Momentcurvature Relation: Moment curvature relationship of a reinforced concrete section
changes its stiffness at (a) tensile cracking of concrete, (b) tensile yielding of longitudinal
reinforcement.
6
Before tensile cracking of concrete, the
neutral axis of the section without axial load
lies at the geometric centroid of the section,
and area of concrete in tension cannot be
neglected for bending resistance.
After initial cracking, the neutral axis
shift to compression side, and the
contribution of uncracked concrete in
tension to bending resistance becomes
negligible. For a reinforced concrete section
satisfying design requirements, the
concrete in compression can be assumed
to act linearly elastic even after cracking,
and cracked transformed section may be
used after cracking.
Significant stiffness change occurs at the tensile yielding of longitudinal reinforcement, and very
small increase in resistance takes place after flexural yielding.
In design, the concrete compressive strain at the ultimate stage is normally used as 0.003 to 0.004.
However, this ultimate strain is not intended for use in evaluating ultimate deformation (curvature,
rotation nor deflection), but is used to estimate the ultimate moment, which is not affected by the
amplitude of ultimate strain. The concrete strain at the flexural failure is believed to be much larger
especially when the concrete is properly confined by lateral reinforcement.
Resisting moments at cracking and yielding
are significantly influenced by existing axial
force in the section. This is called the
"interaction of axial force and bending moment".
Compressive force delays the tensile cracking
of concrete and tensile yielding of longitudinal
reinforcement.
The balance point is defined as a point in the
axial forcemoment diagram where the
compression failure of concrete develops at the
same time as the tensile yielding of longitudinal
bars in the section. A compression failure
occurs if the axial force is larger than the axial
force at the balance point; no tensile
reinforcement develops before failure.
Therefore, there is no ductility. A tension failure
occurs if the axial force is lower than the axial force at the balance point, where the concrete fails in
compression after the tensile yielding in the longitudinal reinforcement. The presence of axial force
significantly reduce the ductility of the section.
References:
Blume, J. A., N. M. Newmark and L. H. Corning, “Design of Multistory Reinforced Concrete Buildings
for Earthquake Motions, Portland Cement Association, Chicago, 1961, 318 pp.
Hognestad, E., "A Study of Combined Bending and Axial Load in Reinforced Concrete Members,"
Bulletin No. 399, Engineering Experimental Station, University of Illinois, 1951.
Kent, D. C., and R. Park, "Flexural Members with Confined Concrete," Journal, Structural Division,
ASCE, Vol. 97, ST7. pp. 19691990, July 1971.
Park, R., and T. Paulay, Reinforced Concrete Structures, WileyInterScience Publication, 1975.
Yielding
Cracking
Ultimate
Curvature
B
e
n
d
i
n
g
M
o
m
e
n
t
E
c
I
e
Momentcurvature relation
Bending Moment
A
x
i
a
l
F
o
r
c
e
Compression
Failure
Tension
Failure
Yielding
Balance
Point
Interaction of bending moment and axial force
7
Pfrang, E. O., C. P. Siess and M. A. Sozen, “LoadMomentCurvature Characteristics of Reinforced
Concrete Cross Sections,“ Journal, American concrete Institute, Vol. 61, No. 7, July 1964, pp. 763
 778.
8
Example of Flexural Analysis
Standard Dimensions and Material Properties
 Dimensions
Width b=400 mm
Depth D=800 mm
Effective depth d= 750 mm
d
c
=50 mm (concrete cover depth to center of longitudinal reinforcement)
 Reinforcement
Tensile reinforcement area A
s
=1435 mm
2
(5Ｄ19)
Compressive reinforcement area A’
s
=861 mm
2
(3Ｄ19)
 Material Properties
Yield stress of longitudinal reinforcement
y
σ =345 N/mm
2
(SD345)
y
ε =0.00167
Concrete strength
B
σ =20.6 N/mm
2
(C20)
Case studies:
 Case 1: Tensile reinforcement area A
s
was reduced to 2871 mm
2
(3D35).
 Case 2: Compressive reinforcement was totally removed (A’
s
=0.0 mm
2
).
 Case 3: Yield stress of steel was reduced to
y
σ =235 N/mm
2
(SD235).
y
ε =0.00114
 Case 4: Concrete strength was increased to 31 N/mm
2
(C31).
 Case 5: Width was reduced to 200 mm.
 Case 6: Effective depth was reduced to 400 mm.
 Case 7: Concrete cover depth was increased to 100 mm.
Yield Moment and Sectional Parameters
The flexural yield point is defined at the tensile yielding of longitudinal reinforcement. The
stressstrain relation of concrete is assumed to be linear with concrete elastic modulus (secant
stiffness at onethird compressive strength). It should be examined if the concrete stress
max c
σ at
the extreme compressive fiber should be less than 70 percent of the compressive strength
B
σ . If the
stress exceeds the limit value, more realistic stressstrain curve should be used for the concrete. It
was also examined if the strain in compressive reinforcement exceeds the yield strain. Modular ratio
(Modular ratio =E
s
/E
c
) n=15 was assumed in the analysis.
Standard Case 1 Case 2 Case 3 Case 4 Case 5 Case 6 Case 7
d
k , mm
217 268 235 217 217 274 147 213
y
φ , x10
6
/mm
3.13 3.46 3.24 2.15 3.13 3.51 6.58 3.11
y
M , kNm
335 625 331 230 335 330 172 326
cr
I , x10
5
mm
4
7.81 13.2 7.44 7.81 7.81 6.85 1.92 7.66
max
/
c B
σ σ
0.45 0.62 0.51 0.31 0.30 0.64 0.64 0.44
/
sc y
ε ε
0.31 0.45  0.36 0.31 0.47 0.38 0.21
d
k : Depth of the neutral axis from the extreme compressive fiber;
y
φ : curvature at yielding;
y
M : yield moment;
cr
I : moment of inertia of cracked transformed section.
9
Ultimate moment and section parameters
The ultimate moment of sections was calculated using equivalent rectangular stress block when
the strain at the extreme compressive fiber reached 0.003. The depth and amplitude of the
rectangular stress block were 0.85 times the depth c to the neutral axis and 0.85 times the
compressive strength of concrete
B
σ . The resultant of compressive stress was assumed to act at
the center of the rectangular stress block.
Standard Case 1 Case 2 Case 3 Case 4 Case 5 Case 6 Case 7
c , mm 63.6 116 82.7 52.9 53.2 88.2 63.6 91.3
u
φ , x10
6
/mm
47.2 25.9 36.2 56.7 56.4 34.0 47.2 32.9
u
M , kNm
353 690 352 249 358 348 180 353
/
st y
ε ε
19.4 9.85 14.0 34.5 23.5 13.5 9.53 13.0
/
sc y
ε ε
0.39 1.02  0.14 0.11 0.78 0.39 0.2
c : Depth of the neutral axis from the extreme compressive fiber;
u
φ : curvature at yielding;
u
M : yield moment.
10
4.2 Momentcurvature Relation under Reversed Loading
Analytical procedure is identical to the flexural analysis of section under monotonically increasing
moment. Three basic assumptions are used; i.e., (a) Bernoulli's hypothesis, (b) normal stressstrain
relations of the concrete and reinforcing steel, and (c) equilibrium of internal and external forces.
Stressstrain hysteresis relation must be prepared for the concrete and reinforcing steel under stress
reversals. Iterative procedure is normally used to determine the strain at a preselected point of the
section for a given curvature until the equilibrium of axial forces is reached. The integration of normal
stresses over the section is not simple because the stress distribution cannot be defined by a simple
function. Therefore, the lamina or fiber model is generally used in the analysis.
The first effort to calculate the momentcurvature relation for reinforced concrete beams under
reversed loading was made by H. Aoyama (1964), followed by Agrawal et al. (1965), Bertero and
Bresler (1969), and Brown and Jirsa (1971). Aoyama (1964) assumed elastoplastic stressstrain
relation for both steel and concrete in the analysis of reinforced concrete section under constant axial
force and reversal bending. He reported that the level of axial force and the plastic deformation in the
previous loading made drastic change in the momentcurvature relations.
Lamina Model Analysis: A reinforced section is sliced to horizontal pieces parallel to the neutral axis.
For each layer of concrete and reinforcement, stress and strain are represented by the values at the
midheight of the layer and stress is assumed to be uniform within each layer element.
Park, Kent and Sampson (1972) analyzed a rectangular girder section under moment reversals.
Stressstrain relations of concrete and reinforcing steel are modeled from the observed behavior in
the laboratory. The amount of top and bottom reinforcement is quite different in the section.
Loading part of stressstrain relation of reinforcing bars is represented by modified
RambergOsgood Model (1943):
} 2411 . 0
1
71 . 0
) 1000 1 ln(
744 . 0
{
1 (
1000
1
+
−
−
+
=
+ = −
−
ip
e
E
ip
sy ch
r
ch
s
s
s
si s
ε
ε
σ σ
σ
σ σ
ε ε
n even for
e n
r
n odd for
e n
r
n
n
04 . 3
) 1 (
469 . 0
) 1 ln(
20 . 2
297 . 0
) 1 (
03 . 6
) 1 ln(
49 . 4
+
−
−
+
=
+
−
−
+
=
where
s
ε : steel strain,
si
ε : steel strain at zero stress,
s
σ : steel stress, Es: elastic modulus of steel,
Strain Section
Stress
Resultants
M
N
Lamina model
11
ch
σ : stress dependent on the yield strength and plastic strain in the steel produced in the previous
loading run, and γ : parameter of the RambergOsgood Model,
ip
ε : plastic strain in steel produced
in previous loading run, n: number of post yield loading runs with n = 0 for the first yielding. Unloading
stiffness was taken equal to the initial elastic stiffness.
The stressstrain curve for concrete is represented by a parabola for ascending portion and
straight line for descending portion for monotonically increasing strain (Kent and Park, 1971). A linear
stressstrain curve for concrete in tension may be assumed to the tensile strength. The curve under
cyclic loading is represented by straight lines. Upon unloading from point E on the skeleton curve,
0.75 of the previous stress is lost without decrease in strain, whereupon a linear path of slope 0.25
c
E
is followed to point G. If the concrete has not cracked, it is capable of carrying tensile stress to point
K; but if the concrete has previously cracked, or if cracks form during this loading stage, the tensile
strains increase but no tensile stress develops. Upon reloading, the strain must regain the value at G
before compressive stress can be sustained again. If reloading commences before unloading
produces zero compressive stress, reloading follows one of the paths IJ. The average slope of the
assumed loop between E and G is parallel to the initial tangent modulus.
The stressstrain curve for the cover concrete in compression may be assumed to follow the curve
for the confined core concrete at strains less than 0.004. The cover concrete at strains greater than
0.004 may be considered to have spalled and to have zero strength.
An iterative technique may be used to calculate the depth of the neutral axis at each loading
stage.
Doubly reinforced concrete beams had a rectangular section of 125 x 203 mm, and simply support
span of 1830 mm. Strains were measured on the top and bottom reinforcement over a 51 mm gauge
length in the critical region and the curvature was calculated from the strain measurement.
The comparison of the observed and the calculated was reasonably good. Note that the observed
curvature history was given to the analysis model and that the resistance after yielding is limited by
yield moment. Therefore, the behavior prior to yielding needs be carefully examined. The calculated
hysteresis loops before flexural yielding were fat compared with the observed. General behavior
during crack opening and closing was simulated well. When the cracks were open in the theoretical
curves, the moment is carried by a steel couple alone.
12
The filament model was used in the analysis of section and members under bidirectional
bending; i.e., Monegotto and Pinto (1973), Aktan et al. (1974), and Zeris and Mahin (1988).
References:
Agrawal, G. L., L. G. Tulin and K. H. Gerstle, “Response of Doubly Reinforced Concrete Beams to
Cyclic Loading,” Journal, American concrete Institute, Vol. 62, No. 7, July 1965, pp. 823  836.
Aktan, A. E., et al., “R/C Column Earthquake Response in Two Dimensions,” Journal, Structural
Division, ASCE, Vol. 100, No. ST10, October 1974, pp. 1999  2015.
Aoyama, H., “MomentCurvature Characteristics of Reinforced Concrete Members subjected to Axial
Load and Reversal of Bending,” Proceedings, International Symposium on the Flexural
Mechanics of Reinforced Concrete, ASCEACI, Miami, November 1964, pp. 183  212.
Bertero, V. V., and B. Bresler, “Seismic Behavior of Reinforced Concrete Framed Structures,”
Proceedings, Fourth World Conference on Earthquake Engineering, Vol. 1, Session B2, Chile,
1969, pp. 109  124.
Brown, R. H., and J. O. Jirsa, “Reinforced Concrete Beams under Reversed Loading,” Journal,
American Concrete Institute, Vol. 68, No. 5, May 1971, pp. 380  390.
CEB, “RC Elements under Cyclic Loading,” Bulletin d’Information 210, Thomas Telford, London,
1991.
Fujii, S., H. Aoyama and H. Umemura, “Momentcurvature Relation Calculated on the Basis of
Material Properties (in Japanese),” Report (Structural Engineering), Annual Meeting, Architectural
Institute of Japan, October 1973, pp. 1261  1262.
Kent, D. C., and R. Park, "Flexural Members with Confined Concrete," Journal, Structural Division,
ASCE, Vol. 97, ST7. pp. 19691990, July 1971.
Menegotto, M., and P. E. Pinto, “Method of Analysis for Cyclically Loaded RC Plane Frames including
Changes in Geometry and Nonelastic Behavior of Elements under Combined Normal force and
Bending,” Preliminary Report, IABSE, Vol. 13, 1973, pp. 15  22.
Park, R., D. C. Kent and R. A. Sampson, "Reinforced Concrete Members with Cyclic Loading,"
Journal, Structural Division, ASCE, Vol. 98, ST 7, pp. 13411360, July 1972.
Ramberg, W. and W. R. Osgood, "Description of StressStrain Curves by Three Parameters,"
Technical Note No. 902, National Advisory Committee for Aeronautics, July 1943.
Zeris, C., and S. A. Mahin, “Analysis of Reinforced Concrete Beamcolumns under uniaxial
excitation,” Journal, Structural Division, ASCE, Vol. 114, No. ST4, April 1988, pp. 804  820.
13
4.3 Flexural Analysis of Members
In the analysis of prismatic reinforced concrete members, it is convenient to work with the member
axis and with cross sections normal to the member axis. The resultant axial and shear forces and
bending and torsional moments of the normal and shear stresses acting on the section are
considered. The interaction between shear and bending deformations as well as between the shear
and bending moment is important especially under nonlinear loading reversals. However, the
interaction is generally neglected because the shear dominated behavior is generally prevented in
the design process.
The local behavior of a member is described by the momentrelative rotation relations, in which
the relative rotation
AB
θ of two neighboring crosssections A and B is the integral of curvature
between the two sections,
( )
B
A
x
AB
x
x dx θ φ =
∫
where ( ) x φ : curvature at x. The relative rotation is defined as the angle between their planes of
cross section or the corresponding tangents to the member axis.
The chord rotation is the angle between the
chord connecting the member axis and the tangent
to the member axis; for member AB, chord rotations
A
θ and
B
θ are
( )
( )
B
A
B
A
x
B
A
x
B A
x
A
B
x
B A
x x
x dx
x x
x x
x dx
x x
θ φ
θ φ
−
=
−
−
=
−
∫
∫
The relative rotation
AB
θ , therefore, is expressed as
AB A B
θ θ θ = −
The deflection of point A from the tangent to the member axis at point B due to curvature between
the two points is calculated by
( )
B
A
x
AB
x
x x dx φ ∆ =
∫
where x : distance from point A.
The following effects should be considered in evaluating member deformation; (a) the increase in
stiffness of members due to tension carried by the concrete between the cracks (tension stiffening) ,
(b) deformation caused by diagonal tension cracks, and (c) deformation caused by bond slip of the
reinforcement.
When diagonal tension cracks are present in a member, the tension in the flexural reinforcement
at sections away from the section of maximum moment may be larger than that computed from the
bending moment diagram. The internal tension remains nearly constant at the maximum value over a
distance nearly equal to the effective depth from the critical section. The region of a yield hinge zone
is wider than the bending moment diagram implies.
Bond slip of reinforcement in anchorage zones increases the deformation.
A
θ
B
θ
AB
θ
A
B
14
Deformation of Simply Supported Member:
Member end displacement expressed for a simply
supported member can be transformed to
displacements at member ends in the local
coordinate system. Consider member end
displacements of member 12 in a plane; i.e.,
member end displacements in the member axis
direction, orthogonal to the member axis and
rotation. The member end displacements of a simply
supported member are given by member end
rotations ,
A B
θ θ and axial deformation e These
displacements are related in the following;
L d d
L d d
d d e
y y z B
y y z A
x x
/ ) (
/ ) (
2 1 2
2 1 1
2 1
− + =
− + =
+ − =
θ θ
θ θ
in a matrix form;
¦
)
¦
`
¹
¦
¹
¦
´
¦
−
− +
¦
)
¦
`
¹
¦
¹
¦
´
¦
−
=
¦
)
¦
`
¹
¦
¹
¦
´
¦
2
2
2
1
1
1
0
1
0
0
1
0
0 0 1
0
1
0
1
1
0
0 0 1
z
y
x
z
y
x
B
A
d
d
L
L
d
d
L
L
e
θ θ θ
θ
It is often convenient to study the deformation relation of simply supported beams under midpoint
loading and antisymmetric loading. The loading cases will cause triangular bending moment
distribution in onehalf member, but maximum moment at member end under antisymmetric bending
and maximum moment at the midspan in midpoint loading. The chord rotation
A
θ of the
antisymmetric loading beam is defined as the angle between the tangent at the end and the straight
line connecting two member ends, which is the ratio of deflection δ at the midspan measured from
the tangent line divided by onehalf member length / 2 l . The same angle can be obtained in the
midspan loading beam by dividing the deflection at midspan by onehalf member length.
R
A
θ
B
θ
1
2 1 y
d
1 x
d
2 x
d
2 y
d
2 1 y y
d d
R
−
=
l
Deformation of Member in Local Coordinates
B
θ
A
θ
A B
e
Deformation of simple Beam
A
m
B A
m m =
δ
A
θ
B A
θ θ =
A AB
R θ =
/( / 2)
AB
R δ = l
δ
A
m
P
Relation between chord rotation at beam ends of antisymmetric bending and
member rotation of simply supported beam
15
Simply Supported Beams:
Consider a simply supported straight
member i subjected to member end
moments m
A
and m
B
at the two ends
A and B. No intermediate loads act
within the member. Moment
distribution within the member can
be determined from the member end
forces m
A
and m
B
; curvature ) (x φ
at point x along the member can be
determined for the existing bending
moment m(x) at the point on the
basis of momentcurvature relation.
Memberend rotations θ
A
and θ
B
may be calculated by using the
dummy (unit) load method; bending
moment m
uj
(x) due to unit moment m
j
= 1.0 (j= A or B) applied at jend,
then memberend rotation
j
θ is
calculated as
∫
=
L
uj j
dx x m x
0
) ( ) ( φ θ
In other words,
dx
L
x
x
dx
L
x
x
L
B
L
A
) ( ) (
) 1 ( ) (
0
0
− =
− =
∫
∫
φ θ
φ θ
Note that a member end rotation is
not solely dependent on the moment at
the end.
The elongation e of the member is
calculated by integrating axial strain at
the geometrical centroid of the section.
No elongation occurs under bending at
initial elastic stage. However, a
significant elongation occurs after initial
cracking even at the centroid due to
bending because the neutral axis of a section shifts to compressive side.
∫
=
L
dx x e
0
) ( ε
The memberend momentrotation relation is calculated for antisymmetric moment distribution
having the inflection point at the midspan; rotations at the two ends are the same. The memberend
momentrotation relation under monotonically increasing load may be approximated by a trilinear
relation with stiffness changes at initial cracking and at tensile yielding at the member ends.
At an elastic stage with elastic modulus
c
E and equivalent moment of inertia
e
I of transformed
section is given as
x
A
B
B
θ
A
θ
( ) m x
( ) x φ
A
m
B
m
Memberend Moment and Memberend Rotation
( ) m x
( ) x φ
A
m
B
m
x
1.0
A
m =
1.0
B
m =
Calculation of Memberend Rotation
16
A
e c
A
L
I E
m θ
6
=
in which L: member length.
Cracking Moment: Cracking moment may be estimated as
6
ND
Z m
e t c cr
+ = σ
where
t c
σ : tensile strength of concrete,
e
Z : section modulus of transformed section, N : axial load
(positive in compression), D: overall depth of section. Sugano (1970) evaluated the tensile strength
of concrete in test specimens from observed cracking moments and section modulus of concrete. A
wide scatter can be observed in data, partially attributable to shrinkage strain, partially to late noticing
the cracking in a specimen during the test.
Member end rotation at flexural yielding and ultimate stages can be calculated from the curvature
distribution along the member;
Rotation,
A
θ
M
o
m
e
n
t
Cracking
Yielding
Ultimate
6EI
S =
l
y
S α
Memberend momentrotation relation
A
m
B A
m m =
Antisymmetric moment distribution
2
( / )
B
kgf cm σ
2
( / )
c t
kgf cm σ
Concrete tensile strength evaluated from test results (Sugano, 1970)
First Yieldng
Ultimate Stage
y
m
y
m
y
φ
y
φ
c
φ
c
φ
u
m
u
m
y
m
y
m
c
m
c
m
c
m
c
m
u
φ
u
φ
u
φ
y
φ
y
φ
c
φ
c
φ
Moment and curvature distribution at critical stages
17
Yield Deformation: The memberend rotation at flexural yielding calculated from curvature
distribution is known to underestimate the yield rotation observed in the test because the analysis
does not consider (a) shear deformation, (b) deformation caused by bar slip within anchorage zone,
and because the Bernoulli's hypothesis holds only in an approximate sense. Therefore, an empirical
expression
y
α was formulated for the ratio of secant stiffness at yielding to the initial stiffness
(Sugano, 1970):
2
) }( 33 . 0 043 . 0 64 . 1 043 . 0 {
D
d
bD
N
D
a
p n
B
t y
σ
α + + + =
where n : modular ratio of steel to
concrete,
t
p : tensile reinforcement
ratio calculated for overall cross
sectional area, / a D : shear
spantodepth ratio, N : Axial force,
b D: cross sectional area of section,
B
σ : compressive strength of
concrete, d: effective depth of section,
D: overall depth of section.
The expression is applicable for
the following range of parameters:
t
p = 0.4 to 2.8 %, / a D = 2.0 to 5.0,
/
B
N b Dσ = 0.0 to 0.55.
Although the stiffness degradation
ratio has been used in the nonlinear
earthquake analysis of buildings
under design in Japan, the yield
stiffness is determined with respect
to the initial stiffness, which cannot
be determined with accuracy; e.g.,
note the reliability of an expression
for the elastic modulus
c
E of
concrete.
Another simplifying method to estimate a
member end rotation at yielding of
antisymmetrically loaded member is to
assume a triangular distribution of curvature
with a calculated yield curvature at a
member end and null curvature at the
inflection point. This sometimes gives a
reasonable value. This method sometimes
underestimate the deformation for a short
deep beam.
Ultimate Deformation: The deformation at
the ultimate stage is one of the most
important deformation indeces of a member.
However, the method of estimating this
deformation has drawn little attention in the
past. The plastic rotation in a yield hinge region at the ultimate stage may be estimated by assuming
Beams
Columns
Calculated
2
( )
y
D
d
α
O
b
s
e
r
v
e
d
Calculated and Observed Stiffness Reduction Factor at Yielding
Sugano, 1970
y
m
y
m
y
φ
y
φ
Simplified method to evaluate member end rotation
Triangular curvature distribution
18
the plastic curvature (
u y
φ φ − ) distributes uniformly over
hinge width
p
l ;
( )
p u y p
θ φ φ = − l
Baker (1964) purposes an expression for plastic hinge
length
p
l as
1 2
0.8
p
z
k k c
d
 
=

\ .
l
where
1
k =0.7 for mild steel and 0.9 for coldworked steel,
2
k =0.6 for '
c
f =35 MPa and 0.9 for '
c
f =12 MPa, z :
distance of critical section to the inflection point, d :
effective depth of section, c : neutral axis depth at the
ultimate moment with ultimate strain of concrete given in the following form;
0.0015{1 150 (0.7 10 ) } 0.01
cu s s
d
c
ε ρ ρ = + + − ≤
where
s
ρ : ratio of the volume of the transverse confining reinforcement to the volume of the core
concrete. Concrete stressstrain relation is a parabola for ascending region and a constant value to
ultimate strain. The maximum concrete stress "
c
f is given by
" (0.8 0.1 ) ' '
c c c
d
f f f
c
= + ≤
The test results show a significant scatter, but Baker claims that the expression furnishes a
reasonable safe prediction.
Corley (1964) proposed the following expressions for the equivalent plastic hinge length
p
l and
the maximum concrete strain
cu
ε at the ultimate stage;
2
0.5 0.2
0.003 0.02
20
p
s y
cu
z
d d
d
f
b
z
ρ
ε
 
= +

\ .
 
= + +

\ .
l
where z : distance of critical section to the inflection point, b : width of beam in inches (=25.4 mm),
d : effective depth of section,
s
ρ : ratio of the volume of the transverse confining reinforcement and
compressive longitudinal bars to the volume of the core concrete,
y
f : yield strength of confining
steel in kips (=6.89 MPa)
References:
Baker, A. L., and A. M. N. Amarakone, “Inelastic Hyperstatic Frame Analysis,” Proceedings,
International Symposium on the Flexural Mechanics of Reinforced Concrete, ASCEACI, Miami,
November 1964, pp. 85  142.
Corley, G. W., “Rotational Capacity of Reinforced Concrete Beams,” Journal, Structural Division,
ASCE, Vol. 92, No. ST5, October 1964, pp. 121  146,
Sugano, S, "Experimental Study on Restoring Force Characteristics of Reinforced Concrete
Members (in Japanese)," Thesis submitted to fulfill the requirements of Doctor of Philosophy,
University of Tokyo, March 1970.
u
φ
y
φ
p
l
Plastic hinge length
19
4.4 LoadDeformation Relation of Beams
Data Base of RC Beams: Test data of beams, yielding in flexure before failure, were searched from
literature (Ref. 122) between 1982 and 1992 in Proceedings of Japan Concrete Institute (JCI),
Summaries of Technical Papers of Annual Meeting of Architectural Institute of Japan (AIJ), Annual
Reports of the New RC Project, and reports of research institutes of Japanese construction
companies. The specimens must satisfy the following conditions; (a) rectangular cross section, (b)
same amount of top and bottom longitudinal reinforcement, (c) width wider than 150 mm, and (d)
overall depth deeper than 225 mm. Among 105 beam specimens obtained, the concrete strength
ranged from 26 to 98 MPa; the shear spantodepth ratio from 1.0 to 3.6; the tensile reinforcement
ratio from 0.44 to 2.70 percent; the yield strength of longitudinal reinforcement from 261 to 976 MPa.
The specimens were tested under two types of loading methods (Fig. 1) simulating the conditions
during earthquake excitation; i.e., (a) TypeA test (31 specimens): statically indeterminate beams with
two stiff end stubs subjected to lateral displacement at the two ends maintaining the end stubs in
parallel during loading, and (b) Type B test (74 specimens): simply supported beams, normally with
two loading stubs, subjected to twopoint loading causing the point of inflection at the center of the
middle span. In Type A test, moment distribution is not known due to static indeterminacy; linear
moment distribution is assumed with an inflection point at the midspan. In Type B test, the damage
within the test span tends to concentrate at an end during the test. Simply supported specimens
subjected to midspan loading were not selected for the study.
The loaddeformation relation curves under reversed cyclic loading were obtained from the generous
researchers, and were digitized at the University of Tokyo for the sequence of parts where the load
exceeded the maximum load of the previous loading cycles. The error of digitization is less than 0.3
percent, on the average 0.2 percent, of the full scale.
MomentRotation Relation: The
memberend moment was obtained
from shear force measured in the test
span by assuming the inflection point
at the midspan. The memberend
rotation of Type A specimens was
calculated from the measured relative
lateral displacement divided by the
clear span, but the member end
rotation of Type B specimens was
used as the researchers reported.
The observed momentrotation curve
was idealized into a trilinear
relationship (Fig. 2).
Although many specimens failed in
shear or bondsplitting modes after
flexural yielding, the initial stiffness
(a) Type A test (b) Type B test
Fig. 1  Loading methods in laboratory
Fig. 2  Idealization of observed momentrotation relation
20
may not be affected by the failure modes. Therefore, all 105 specimens were used in the study of
initial stiffness and flexural cracking moment. On the other hand, the yield deflection is increased by
the damage associated with the failure modes. Therefore, those specimens failing in shear or
bondsplitting modes within a deflection equal to three times flexural yielding deflection were
excluded from the examination of yield and ultimate points; 38 specimens were used to study the
yield deflection.
The initial stiffness in a test was defined as a secant slope at a load equal to onehalf of the reported
cracking load. The cracking load was not reported in five specimens; hence the cracking moment was
evaluated by assuming the tensile strength of concrete to be 0 56 . σ
B
(Ref. 23). Note that the
cracking load is normally reported at a loading step when cracking is detected for the first time; i.e.,
the reported cracking load is normally higher than the actual cracking load. Therefore, the cracking
point was determined from the shape of forcedeformation relation by the method described in the
following paragraph.
The stiffness of reinforced concrete section changes drastically at the yielding of tensile
reinforcement. If tensile reinforcement is placed in double layers in the section, the stiffness changes
at the yielding of the outer layer reinforcement and then of the inner layer reinforcement. In order to
select a single yield point, the yield point and cracking moment were defined such that the energy
stored at the ultimate deformation should be the same for the test and the model making the absolute
difference in the energy to be minimum (Fig. 2).
Resistance at the ultimate point was taken as the observed maximum resistance. The determination
of an ultimate deformation is an important but impossible issue; the ultimate deformation is not a
unique value but is highly dependent on the progress of concrete deterioration dictated by loading
history and failure modes. Small stiffness after yielding will not change appreciably by the choice of
an ultimate deformation. Therefore, the ultimate deformation was selected to be an arbitrary
deformation at a deformation ductility factor of four.
An iterative procedure was used to define the yield point and cracking moment for the established
initial stiffness and ultimate point.; i.e., (a) a trial yield displacement was assumed, (b) an ultimate
displacement was selected at four times yield deformation, (c) postyield stiffness was determined by
connecting the ultimate point and a point on the observed curve at 2.5 times yield deformation, and
(d) the cracking moment and yield deformation were determined for equal absorbed energy at the
ultimate deformation and minimum absolute difference.
Initial Elastic Stiffness: The methods to evaluate stiffness parameters and their reliability with
respect to the observed values are discussed. The initial elastic stiffness K
E
was evaluated by the
elastic theory of a lineal member considering flexural and shear deformation;
1 1 1
K K K E B S
= + (Eq. 1)
where K
B :
flexural stiffness (= 6 E
c
I
e
/ L), K
S
: shear stiffness (= G
c
A L/ 2κ), E
c
: elastic modulus of
concrete, I
e
: moment of inertia of uncracked transformed section, L: member length, G
c
: shear
modulus of concrete, A: cross sectional area,
κ
:shape factor for shear deformation (=1.2). The
shear modulus of concrete was estimated from the elastic modulus E
c
and assumed Poisson's ratio
of 0.20. Elastic modulus E
s
of steel was assumed to be 206 GPa.
The initial stiffness was calculated using the observed elastic modulus of concrete and the clear span.
The observed initial stiffness (Fig. 3) was notably low and, on the average, 0.53 times that calculated
with a large coefficient of variation (=0.51) for 73 specimens with reported concrete elastic moduli.
The large coefficient of variation and discrepancy between the test and calculation was probably
attributed to (a) technical difficulty in measuring accurate initial stiffness in the test and (b) formation
of accidental and shrinkage cracks prior to the test. In a real structure, flexural cracks under gravity
21
loading, shrinkage cracks, cracks after
medium intensity earthquake excitation
may exist, and the initial stiffness for the
analysis is difficult to estimate. However,
the initial stiffness and cracking force level
of a singledegreeoffreedom system do
not influence the maximum response
amplitude as long as the attained
response ductility factor reached more
than 4 (Ref. 24). The response in this
range is more sensitive to the secant
stiffness and the resistance at yielding
rather than the initial stiffness.
Furthermore, the actual elastic modulus in
a structure, as built, is not known at the
time of structural design although the initial
stiffness of a member is directly dependent
on the modulus. The elastic modulus of concrete is normally difficult to control in construction.
Therefore, the initial stiffness of a structure can be significantly different from the value assumed by a
structural engineer.
An empirical expression was proposed for the elastic modulus E
c
(Ref. 25), taking into account
compressive strength and density of concrete, type of coarse aggregates and mineral admixture;
Ec k k
B
= × × × × ×
1 2
4 1 3 2
3 35 10 60 2 4 . ( / ) ( / . )
/
σ γ (MPa) (Eq. 2)
in which k
1
: factor representing type
of coarse aggregates, k
2:
factor
representing kind of mineral
admixture, σ
B
: observed concrete
strength (MPa),
γ
: unit density of
concrete (ton/m
3
). The factor k
1
is
0.95 for crushed quartzite, crushed
andesite, basalt and clayslate
aggregates, 1.0 for other coarse
aggregates, and 1.2 for crushed
limestone and calcined bauxite
aggregates. Factor k
2
is 0.95 for silica
fume; fine powder of blast furnace
slag and fly ash fume, 1.00 for
concrete without mineral admixture or
with other mineral admixture, 1.10 for
fly ashes. Ninetyfive percent of test
data are shown to fall within 20
percent of the empirical expression (Fig. 4). The modulus should be controlled in construction within
an acceptable range from the value specified by a structural engineer.
Cracking Moment: Cracking moment M
cr
is calculated on the basis of the observed splitting tensile
strength σ
t
of concrete and the section modulus Z
e
of the uncracked transformed section. The ratio
of the reported to the calculated cracking moments is compared in Fig. 5 for 59 specimens; the
average ratio was unexpectedly good at 1.03, with a significantly large coefficient of variation of 0.50.
Cracking tensile strength σ
cr
of concrete was determined by dividing the cracking moment of the
Fig. 3  Reliability of calculated initial stiffness
Fig. 4  Elastic modulus and strength of concrete (Noguchi)
22
trilinear idealization of the momentrotation relation by the section modulus Z
e
of uncracked
transformed section. The cracking tensile strength and compressive strength of concrete are
compared in Fig. 6 for 68 specimens. A wide scatter of data can be observed, but a tendency is
observed for the cracking tensile strength to increase with the compressive strength. Following
empirical relation was derived:
σ σ
cr B
= × 1 26
0 45
.
.
(Eq. 3)
The cracking tensile strength should not be used as the tensile strength of concrete, but is intended to
evaluate a moment level at which the initial elastic stiffness of the trilinear idealization changes in the
memberend momentrotation relationship.
Yield Moment and Rotation: Yield moment at the critical section was calculated for the yielding at an
imaginary centroid of tensile reinforcement. The amount of tensile reinforcement is normally limited
well below the balanced tensile reinforcement ratio; hence, the stressstrain relation of the concrete
was assumed to remain linearly elastic when the tensile reinforcement first yielded under bending. In
addition, the following assumptions were made in calculating yield moment; i.e., (a) plane section
remained plane after deformation, and (b) the concrete in tension did not resist tensile stresses. The
elastic modulus of concrete was determined by Eq. 2 with k
1
=k
2
=1.0 and
γ
=2.4 ton/m
3
, while the
compressive strength of concrete and the yield stress of reinforcement were obtained from the
reported material tests. In 2 specimens out of
38, the calculated stress at the extreme
compressive fiber exceeded the compressive
strength of concrete; these specimens were
removed from the study.
The calculated yield moment and the estimated
yield moment of the trilinear idealization are
compared in Fig. 7 for 36 specimens. The
average ratio of the estimated to the calculated
yield moment is 1.12 with a coefficient of
variation of 0.078; only one estimated yield
moment was smaller (0.94) than the calculated
value. The yield moment at which the stiffness
of an RC member changes drastically may be
calculated conservatively by the flexural
analysis.
Fig. 5  Reported to calculated cracking moment
Fig. 6  Cracking tensile strength and
compressive strength of concrete
Fig. 7  Calculated and observed yield moment
23
Member end rotation at flexural yielding was calculated using cracked transformed section for flexural
deformation (θ
f y c cr
M L E I = / 6 ), elastic stiffness for shear deformation (θ κ
s y
M GAL = ⋅ 2 / ) and
pullout deformation θ
slip
of longitudinal reinforcement from the anchorage. The pullout deformation
δ
slip
was calculated by the following expression (Ref. 26):
δ ε ε σ
slip y y b B
d = + ⋅ ( , ) / ( / )
/
2 3 500 20
2 3
(Eq. 4)
where, ε
y
: yield strain of longitudinal reinforcement, d
b
: diameter of longitudinal reinforcement, σ
B
:
concrete strength (MPa). The center of rotation at the critical section was assumed to be at the
centroid of compressive reinforcement; i.e., θ δ
slip slip c
d d = − / ( ), where d: effective depth and d
c
:
depth to the centroid of compressive reinforcement.
The calculated yield rotation and the estimated yield rotation of the trilinear idealization are compared
in Fig. 8.a for 36 specimens. The estimated rotation was, on the average, 2.54 times larger than the
calculated rotation with a coefficient of variation of 0.22. Calculated yield rotation significantly
underestimates the estimated rotation. This discrepancy is attributable to the additional rotation
caused by shear cracking and the error in evaluating the pullout deformation of longitudinal
reinforcement from the anchorage.
An empirical expression was derived for the yield deformation by assuming the yield rotation at a
member end consists of the rotations from flexural deformation, θ
f
, shear deformation, θ
s
and
deformation, θ
slip
due to the pullout of longitudinal reinforcement from the anchorage zone. A
regression analysis with respect to the estimated yield rotation was carried out to determine
coefficients;
θ θ θ θ
y f s slip
= + + 115 12 6 3 89 . . . (Eq. 5)
The calculated yield deflection and the observed yield rotation of the trilinear idealization are
compared in Fig. 8.b for 36 specimens. A coefficient of variation of the ratio was 0.20 with a mean of
1.0.
Ultimate Moment: Flexural strength of beam section is not sensitive to the shape of stressstrain
relationship nor the compressive strength of concrete because the neutral axis depth is so small that
the distance between the resultant compressive and tensile forces cannot change appreciably within
the section. The ultimate moment was calculated using the plasticity theory suggested by Eberhard
(a) Calculated yield rotation (b) Empirical expression
Fig. 8  Calculated and observed yield rotation
24
and Sozen (Ref. 27), in which the flexural mechanism was assumed to form by the yielding of tensile
reinforcement followed by the compressive failure of concrete. Instead of strain compatibility,
equilibrium conditions for axial force and bending moment of section were used based on the lower
bound theorem. The maximum bending resistance was sought by satisfying the yield criteria of the
materials and is given by
M k b d a d d
u B c t y c
= ⋅ ⋅ ⋅ + ⋅ − σ σ
2
( ) (Eq. 6)
in which, σ
B
: compressive strength of concrete, b: width of section, d
c
: distance from the extreme
compressive fiber to the centroid of compressive reinforcement, a
t
: area of tensile reinforcement, σ
y
:
yield stress of tensile reinforcement, d:
effective depth of section. Values of k are
(3/8) and (1/2) for triangular and rectangular
stress distribution of compressive concrete
with maximum stress of σ
B
, respectively.
The stress in the compressive reinforcement
must be checked not to exceed the yielding
stress.
The ratio of the observed to the calculated
ultimate moments is compared with respect
to compressive strength of concrete in Fig. 9
for 38 specimens; the average ratio is 1.15
with a coefficient of variation of 0.074 for the
triangular concrete stress block, and the
average of 1.13 with a coefficient of variation
of 0.069 for the rectangular stress block. The
shape of a stress block shape should be
carefully selected in the evaluation of
ultimate moment of a column.
References:
1. Korenaga, T., T. Mogami, et al., "Test of structural members and frames in tall buildings utilizing the
R. C. layered construction system, Part 1: Test on short beams," Taisei Technical Research
Report, No. 18, March 1986, pp. 111126.
2. Tanaka, N., N. Sakaguchi, et al., "Flexural and shear strength of short span beams using ultra high
strength reinforced concrete, (Part 1: Flexural behavior of beams, Part 2: Shear behavior of
beams)," Summaries of Technical Papers of Annual Meeting, Architectural Institute of Japan,
Structures, 1987, pp. 67  70.
3. Kobayashi, J., K. Kanada, S. Yoshizaki and T. Yamada, "Test of structural members and frames in
tall buildings utilizing the R. C. layered construction system, Part 4: Test of shear reinforcing
methods of beams," Taisei Technical Research Report, No. 18, 1987, pp. 73  88.
4. Taga, A., K. Kawasaki, et al., "Development of MAEDA highrise reinforced concrete building
system (MARC system), (Part 4: Experimental study on structural members and
subassemblages)," Report of Technical Research Institute, Maeda Corporation, Vol. 292, 1988,
pp. 31  55.
5. Yanagisawa, N., Y. Shimizu, K. Tsumura and M. Fujiwara, "Strength and ductility of reinforced
concrete Tbeam with high strength concrete," Proceedings, Japan Concrete Institute, Vol. 10, No.
3, 1988, pp. 681  684.
6. Matsutani, T., J. Ishida, et al., "Experiments for development of highrise reinforced concrete
structure," Technical Research Reports, Konoike Construction Co., Ltd., 1988, pp. 71  84.
7. Honda, Y., T. Iwakura, S. Hakuto and H. Maie, "Experimental study on RC beams with web
openings, ultimate shear strength and deformability of RC beams reinforced using ring materials
of high tensile strength steel," Technical Reports, Tokyu Construction, No. 15, 1989, pp. 67  72.
8. Sugano, S., T. Nagashima, H. Kimura and A. Ichikawa, "Experimental study on high strength
concrete beam using high strength main bar," Proceedings, Japan Concrete Institute, Vol. 12, No.
2, 1990, pp. 215  220.
Fig. 9  Observed and calculated ultimate moment
(Triangular stress block)
25
9. Taga, A., K. Kawasaki, T. Watanabe, and K. Tsujita, "Study on a seismic design of a highrise
reinforced concrete building (Flexureshear loading tests and shear loading tests on beams with
web openings)," Report of Technical Research Institute, Maeda Corporation, Vol. 30, 1990, pp.
131  144.
10. Nakamura, M., S. Bessho, T. Kato and A. Zan, "Bendingshear test of beams with high strength
concrete and rebars for high rised R/C building," Proceedings, Japan Concrete Institute, Vol. 14,
No. 2, 1992, pp. 529  534.
11. Ishikawa, Y., M. Hamamoto, S. Otani and H. Aoyama, "Experimental study on deformation
capacity of reinforced concrete beams," Proceedings, Japan Concrete Institute, Vol. 14, No. 2,
1992, pp. 255  260.
12. Kamura, T., T. Ohmizu, S. Otani and H. Aoyama, "Experimental study on deformation capacity of
reinforced concrete beams," Proceedings, Japan Concrete Institute, Vol. 15, No. 2, 1993, pp. 335
 340.
13. Sumi, A., T. Segawa, et al., "An experimental study on flexural performance of reinforced concrete
beams using high tensile strength shear reinforcement," Summaries of Technical Papers of
Annual Meeting, Architectural Institute of Japan, Structures, 1984, pp. 1681  1682.
14. Muguruma, H., A. Sumi, T. Segawa and T. Hisatoku, "An experimental study of reinforced
concrete beams laterally confined by high strength reinforcement," Summaries of Technical
Papers of Annual Meeting, Architectural Institute of Japan, Structures, 1988, pp. 229  230.
15. Sumi, A., K. Masuo, et al., "An experimental study on flexural performance of short span RC
beams using high strength shear reinforcement," Proceedings of Architectural Research Meetings,
Kinki Chapter, Architectural Institute of Japan, 1992, pp. 73  76.
16. Muguruma, H., F. Watanabe, "Study on shear design of R/C ductile beams subjected to combined
bending and shear (Part 1, 2)," Summaries of Technical Papers of Annual Meeting, Architectural
Institute of Japan, Structures 2C, 1988, pp. 183  184.
17. Iwai, Y., Y. Kakita, F. Watanabe and H. Muguruma, "Study on shear design of RC ductile beams
subjected to combined bending and shear (Part 3)," Summaries of Technical Papers of Annual
Meeting, Architectural Institute of Japan, Structures 2C, 1990, pp. 305  306.
18. Minami, K., H. Kuramoto, N. Tsukamoto and A. Nakazawa, "Shear and bond strength behavior of
R/C beams with grade 13,000kgf/cm2 shear reinforcement under cyclic bending and shear,"
Proceedings, Japan Concrete Institute, Vol. 12, No. 2, 1990, pp. 221  226.
19. Fujisawa, M., T. Kaminosono, M. Takeuchi and H. Murakami, "Study on bond splitting failure of
beams after flexural yielding," Report of Structures Committee, New RC Project, Japan Institute of
Construction Engineering, March 1990, pp. 441  4415.
20. Fujisawa, M., T. Kaminosono, M. Takeuchi and H. Murakami, "Study on ductility of highstrength
reinforced concrete beams," Summaries of Technical Papers of Annual Meeting, Architectural
Institute of Japan, Structures, 1990, pp. 277  278.
21. Fujisawa, M., T. Kaminosono, et al., "Study on effect of slab on flexural performance of beams,"
Report of Structures Committee, New RC Project, Japan Institute of Construction Engineering,
March 1991, pp. 441  4419.
22. Fujii, S., H. Fujitani, et al., " Study on bond splitting failure of beams after flexural yielding," Report
of Structures Committee, New RC Project, Japan Institute of Construction Engineering, March
1992, pp. 411  4410.
23. Sugano, S., "Experimental study on restoring force characteristics of reinforced concrete
members (in Japanese)," a thesis submitted to the University of Tokyo for a partial fulfillment of the
requirements for doctor of engineering degree, University of Tokyo, December 1970.
24. Otani, S., "Hysteresis models of reinforced concrete for earthquake response analysis," Journal
(B), Faculty of Engineering, University of Tokyo, Vol. 36, No. 2, 1981, pp. 125159.
25. Tomosawa, F., T. Noguchi, and K. Onoyama, "Investigation on fundamental mechanical
properties of highstrength and super high strength concrete," Summaries of Technical Papers of
Annual Meeting of Architectural Institute of Japan, Vol. A, 1990, pp. 497498.
26. Shima, H., L.L. Chou, and H. Okamura, "Micro and macro models for bond in reinforced
concrete," Journal, Faculty of Engineering, University of Tokyo, Series (B), Vol. 39, No. 2, 1987, pp.
133  194.
27. Eberhard, M. O., and M. A. Sozen, "Behaviorbased Method to determine design shear in
earthquakeresistant walls," Journal, Structural Engineering, ASCE, Vol. 119, No. 2, February,
1993, pp. 619  640.
26
4.5 Analysis of Structural Walls
T. Paulay (1981) presents analysis procedures of reinforced concrete structural walls for design
purpose. The following deformation should be considered for a structural wall; (a) flexural
deformation, (b) shear deformation after diagonal cracking, (c) anchorage deformation, (d)
deformation of foundation and supporting ground.
The equivalent second moment
e
I of area may be taken as 60 % of the value
g
I based on the
uncracked gross concrete section. A more accurate estimate of flexural deformations may be made if
the ratio of the moment causing cracking to the maximum applied moment is evaluated;
3 3
1
cr cr
e g cr
a a
M M
I I I
M M
   
= + −
 
\ . \ .
(1)
where,
cr
M : cracking moment,
a
M : maximum moment at which deflection is computed,
g
I :
second moment of area of the gross concrete section,
cr
I : moment of inertia of cracked transformed
section. The contribution of longitudinal reinforcement can be neglected because the reinforcement
ratio is normally less than 0.1 % in a wall.
The elongation of the vertical bars within the foundation structure and slip due to high local bond
stresses along the development length will result in an apparent pullout of such bars at the base of
the wall. This can significantly increase the wall deflection.
For cantilever walls with aspect ratios, /
w w
h l , larger than 4, the shear deformation may be
neglected. After diagonal cracking, however, the shear stiffness is reduced to 10  30 % of the
uncracked stiffness. The lateral reinforcement influences the shear deformation after diagonal
cracking.
When the aspect ratio is less than 4, the second moment of area for a structural wall may be
assumed that
2
1.2
30
e
w
e
w w w
I
I
F
I
F
h b
=
+
=
l
(2)
In this expression, some allowance is made for shear distortions and deflections due to anchorage
(pullout) deformation at the base of a wall.
The structural wall tests at Portland Cement Association were examined. The aspect ratio of the
specimens was 2.4. Cross sections were flanged (Hshape), barbell (with boundary columns) and
rectangular. Equation (1) tends to overestimate the initial stiffness.
For a cantilever wall, the wall may be represented by a straight line passing through the centroid
of the gross section..
Reference:
Paulay, T., “The Design of Reinforced Concrete Ductile Shear Walls for Earthquake
Resistance,“ Research Report 811, Department of Civil Engineering, University of Canterbury,
New Zealand, February 1981, 72 pp.
27
Home Assignment No. 2
20020227 Otani, S.
A simply supported beam was tested at two point loading. The cross section was 102 x 152 mm;
support distance was 2.60 m with each load point at 1.00 m from the support. The section was
reinforced singly by 2D12 bars at the bottom (cover thickness of 15 mm). Concrete strength was
22.5 MPa; load at yielding of a D12 reinforcing bar was 37.1 kN. The beam was reinforced laterally by
6 φ bars at 100 mm spacing.
Assume concrete and reinforcing bars behave linearly elastic until flexural yielding. The concrete
in tension carries tensile stress before cracking (tensile strength of concrete is onetenth of
compressive strength). Ignore the stress carried by tensile region of concrete after cracking. Use
rectangular stress block for concrete in calculating the moment and curvature at an ultimate state,
where a strain of concrete at the extreme compression fiber reaches a limiting value of 0.004 and
stress of tensile reinforcement reaches the yield stress.
Calculate (a) momentcurvature relation of section at cracking, flexural yielding and ultimate
stages, (b) loaddeformation relation of the specimen on the basis of curvature distribution at cracking,
flexural yielding and ultimate stages at the loading point.
Compare the calculated and the observed loaddeformation relation. The total load (kN) and
displacement (mm) at midspan relation is shown below:
10.0 20.0 30.0
Deflection at midspan, mm
Total Load, kN
20.0
10.0
∆
P
28
(1) Linearly elastic stage:
Young’s modulus of concrete
1.5 0.5
3 0.5
1.35
1.35 2.5 10 9.8 (22.5)
24.6
c B
E
GPa
γ σ = × ×
= × × × ×
=
Young’s modulus of steel
210
s
E GPa =
Total area of longitudinal reinforcement
2 2 4 2
2 2 3.14 (0.006) 2.26 10
s
A r m π
−
= = × × = ×
Effective depth
125 d mm =
Modulus ratio
8.55
s
c
E
n
E
= =
Moment of inertia of transformed section
5 4
0
3.35 10 I m
−
= ×
Neutral axis from the extreme tensile fiber
80.9
n
x mm =
Cracking moment
1061
( )
o
c ct
n
I
M Nm
D x
σ = =
−
Cracking curvature
3
0
1.29 10 (1/ )
c
c
c
M
m
E I
φ
−
= = ×
(2) Moment and curvature at yielding
Tensile reinforcement ratio
4
2.26 10
0.01776
0.102 0.125
s
t
A
bd
ρ
−
×
= = =
×
Depth of neutral axis kd of cracked section
2
( ) 2 0.420
t t t
k n n n ρ ρ ρ = + − =
Yield moment
3
2 (1 )
3
0.420
2 37.1 10 0.125 (1 )
3
7980
y y
k
M P d
Nm
= × × × −
= × × × × −
=
Moment of inertia of cracked section
3
2 2
3
4 2 2
5 4
( )
(1 )
3
0.102 (0.420 0.125)
8.55 2.26 10 (1 0.420) 0.125
3
1.506 10
cr s
b kd
I nA k d
m
−
−
= + −
× ×
= + × × × − ×
= ×
Yield curvature
29
0.0215 (1/ )
y
y
c cr
M
m
E I
φ = =
(c) Ultimate moment and curvature
The tensile reinforcement ratio is less than the balanced reinforcement ratio.
From the equilibrium of axial forces,
'
'
3
6
0.85
0.85
37.1 2 10
0.038
0.85 22.5 10 0.102
c s y
s y
c
f ab A f
A f
a
f b
m
=
=
× ×
= =
× × ×
Ultimate moment
3
( )
2
0.038
37.1 10 2 (0.125 )
2
7870
u s y
a
M A f d
Nm
= −
= × × × −
=
Depth of neutral axis
0.0447
0.85
a
c m = =
Ultimate curvature
0.004
0.0894 (1/ )
u
m
c
φ = =
30
Assignment No. 2
20020227
S. Otani
A cantilever reinforced concrete beam was tested under lateral load reversals. The shear span
(distance between loading point and the end of the beam specimen) was 750 mm. The section is
200x300 mm and is reinforced with 4D13 bars (SD345, nominal area of a bar= 127 mm
2
, tensile
reinforcement ratio of 0.85%) at the top and bottom. The lateral reinforcement was 2 4 φ plain bars
placed at 40 mm on centers (lateral reinforcement ratio of 0.32 %, nominal area of a bar=12.6 mm
2
).
The concrete cover to the center of the longitudinal reinforcement was 30 mm.
The standard tensile test shows that
the yield stress of Grade SD345 steel
was 361 N/mm
2
, and tensile strength
was 500 N/mm
2
. Yield stress of a plain
4 φ bar was 478 N/mm
2
and the tensile
strength was 509 N/mm
2
. The
compressive strength of concrete was
28.2 N/mm
2
, tensile strength 2.50
N/mm
2
, and secant elastic modulus
28.6 GPa.
The loaddeformation relation of the
specimen is shown below;
Estimate the resistance and deformation under monotonically increasing loading at flexural
cracking, yielding and ultimate points from the given information and compare the calculated results
with the test results. The observed relation is enlarged below.
80
60
40
20
0
20
40
60
80
8 6 4 2 0 2 4 6 8
Displacement, mm
R
e
s
i
s
t
a
n
c
e
,
k
N
80
60
40
20
0
20
40
60
80
30 20 10 0 10 20 30
Displacement, mm
R
e
s
i
s
t
a
n
c
e
,
k
N
1
Chapter 5. Structural Dynamics
5.1 Differential Equation of Motion
Newton's Law of Motion: The relationship between force and motion was formulated by Sir Isaac
Newton (16421727) in "Principia (1687)."
Law 1: The absolute velocity } {v of a particle remains constant if there is no net external force
applied to the particle.
Law 2: When a force { } f acts on a particle, the absolute acceleration } {a of the particle is
directly proportional to the force.
} { } { f c a =
Law 3: When two particles A and B are in contact, the force applied to particle A by particle B, at
the contact point, is equal in magnitude but opposite in direction from the force applied to particle B
by particle A.
"Newton's second law of motion" expresses the motion of a particle under a given force history:
} { } {
2
2
f c x
dt
d
=
By solving the differential equation, the displacement } {x of the particle can be uniquely defined for
a given set of initial conditions.
D'Alembert's Principle: Newton’s law of motion can be rewritten in the form:
} {
} { } {
2
2
x
dt
d
m
a m f
=
=
If a particle moves with an absolute acceleration of )} ( { t a at time t , force )} ( { t f must act on the
particle in the direction of the acceleration. The proportionality constant is called "mass of inertia."
A fictitious force commonly known as the "inertia force" was introduced by D'Alembert (Traite de
Dynamique, 1743) in order to express a state of dynamic equilibrium:
2
2
{ '} { }
d
f m x
dt
= −
The equilibrium of (static) forces was extended to a dynamic problem by the introduction of an inertia
force:
} 0 { } ' { } { = + f f
Equation of Motion for MassSpringDamping System:
A system consisting of a mass and a spring oscillates with
a constant period. For simplicity, let us consider a system
consisting of one translational mass of inertia m and one
linear spring k , where the particle can move in the
direction of the lineal spring.
Writing an equilibrium equation of forces (D'Alembert's
Principle) at a given displacement;
Mathematical Model of MassSpring System
m
k
x
2
) (
0 ) (
2
2
2
2
t f kx
dt
x d
m
t f kx
dt
x d
m
= +
= + − −
The solution ) (t x for a differential equation is the sum of a particular solution ) (t x
p
, which
satisfies the original equation at any time, and a complementary solution ) (t x
c
, which is a general
solution for the differential equation for the right hand side of the equation to be null (the state of free
vibration).
0 ) ( ) (
) ( ) ( ) (
) ( ) ( ) (
2
2
2
2
= +
= +
+ =
t x k t x
dt
d
m
t f t x k t x
dt
d
m
t x t x t x
c c
p p
c p
The particular solution ) (t x
p
must be found for a given forcing function ) (t f . The particular
solutions have been found for simple forcing functions ) (t f , but a general closed form solution is
not available.
The complementary solution ) (t x
c
can be found as follows;
2 2
2 2
2 2
cos
sin
) sin(
sin cos ) (
B A
B
B A
A
B A
t B t A t x
n
n n c
+
=
+
=
+ + =
+ =
φ
φ
φ ω
ω ω
where
m
k
n
= ω . A and B are integration constants dependent on the initial conditions. For given
initial displacement
o
x and velocity
o
v at 0 = t : the complementary solution is derived as
t
v
t x t x
n
n
o
n o c
ω
ω
ω sin cos ) ( + =
Note that the system oscillates with a constant period of
k
m
T
n
n
π
ω
π
2
2
= = .
This period
n
T of oscillation is called
"natural period" of the system;
n
ω is
called "circular (or angular) (natural)
frequency" of the system. Natural
frequency
n
f is the reciprocal of the
natural period.
3
5.2 Mass of Inertia
The dynamic equilibrium of forces is normally formulated for each displacement degree of
freedom. The inertia force is associated with mass and acceleration; the mass of a particle is
constant, but the mass must be defined for each degree of freedom. It should be noted that there
exists mass associated with rotational acceleration.
Distributed Mass: The mass distributes in a solid body, the displacement is not uniform in a
deformable body. The system of a deformable body has an infinite number of degrees of freedom.
For a practical analysis, the number of degrees of freedom must be reduced to a finite number.
Consider a lineal member of length L . The (axial/ lateral/ rotational) displacement is ) , ( t x y at
coordinate x at time t . Suppose the member oscillates in a single shape ) (x φ with the
timevarying amplitude (coordinate) ) (t Y ;
) ( ) ( ) , ( t Y x t x y φ =
The kinetic energy T of the entire member under this oscillation is defined as;
2
0
2
0
2 2
0
1 ( , )
( ){ }
2
1 ( )
( ){ ( ) }
2
1 ( )
( ){ ( )} { }
2
L
L
L
dy x t
T x dx
dt
dY t
x x dx
dt
dY t
x x dx
dt
ρ
ρ φ
ρ φ
=
=
=
∫
∫
∫
where, ) (x ρ : mass (translational or rotational) per unit length at location x .
Using the displacement coordinate ) (t Y , the kinetic energy is defined as
2 *
}
) (
{
2
1
dt
t dY
M T =
where M*: equivalent lumped mass associated with the deformed shape. Therefore, the equivalent
mass is defined as
dx x x M
L
∫
=
0
2 *
)} ( ){ ( φ ρ
The shape function ) (x φ is sometimes called a generalized coordinate, and
*
M is a generalized
mass associated with the coordinate.
Consistent Mass: Mass coefficients corresponding to the nodal coordinates of a beam element may
be defined by a procedure similar to the determination of element stiffness coefficients; the mass
coefficient
ij
m is the force at nodal coordinate i due to a unit acceleration at nodal coordinate j
while all other nodal coordinates are maintained at zero acceleration.
It is assumed that the deflection resulting from unit dynamic displacement 0 . 1 =
i
d at the nodal
coordinate i of a beam element is given by the consistent function ) (x
i
φ obtained from static
consideration.
4
For a linearly elastic prismatic uniform
beam element, the deflection due to a
member end displacement is expressed by a
cubic function if no external load acts within
the member. The deflected shape ) (x
i
φ
due to unit displacement
i
d at an end is
given in Reference (Paz, 1985).
(1) Axial displacement ) (
1
x φ due to unit
displacement 0 . 1
1
= d at the starting end,
(2) Lateral displacement ) (
2
x φ due to unit displacement 0 . 1
2
= d at the starting end,
(3) Lateral displacement ) (
3
x φ due to unit displacement 0 . 1
3
= d at the starting end,
(4) Axial displacement ) (
4
x φ due to unit displacement 0 . 1
4
= d at the starting end,
(5) Lateral displacement ) (
5
x φ due to unit displacement 0 . 1
5
= d at the starting end,
(6) Lateral displacement ) (
6
x φ due to unit displacement 0 . 1
6
= d at the starting end,
0 . 1 } 1 ) ){( ( ) (
0 . 1 ) ( 2 ) ( 3 ) (
0 . 1 ) ( ) (
0 . 1 )} ( 1 { ) (
0 . 1 ) ( 2 ) ( 3 1 ) (
0 . 1 ) ( 1 ) (
2 6
2
3 2
5
2 4
1
2
3
1
3 2
2
1 1
= − =
= − =
= =
= − =
= + − =
= − =
z
y
x
z
y
x
for
L
x
L
x
x x
d for
L
x
L
x
x
d for
L
x
x
for
L
x
x x
d for
L
x
L
x
x
d for
L
x
x
θ φ
φ
φ
θ φ
φ
φ
If acceleration is assumed to be proportional to
displacement, the inertia force ) (x f
i
per unit
length due to a member end acceleration ) (t a
i
is
) ( ) ( ) ( ) ( t a x x x f
i i i
φ ρ − =
The inertia force
ji
f in j direction caused by member end acceleration ) (t a
i
in i direction
using a consistent mass
ji
m is
i ji ji
a m f − =
For a virtual displacement
j
δ at member
end in j direction, the virtual work WE of
the external force at coordinate i is
j i ji
a m WE δ − =
and the corresponding virtual work WI of
internal forces is
3
3
d
p
6
6
d
p
1
1
d
p
2
2
d
p
5
5
d
p
4
4
d
p
Coordinate Systems for Forces and Displacements
d
1x
d
1y
d
2x
d
2y
1z
θ
2z
θ
d
1x
=1.0
d
1y
=1.0
1
1.0
z
θ =
d
2x
=1.0
d
2y
=1.0
2
1.0
z
θ =
5
∫
∫
− =
− =
L
j i i
L
j j i i
dx x x x a
dx x x a x WI
j
0
0
) ( ) ( ) (
) ( ) ( ) (
φ φ ρ δ
φ δ φ ρ
Therefore,
∫
=
L
j i ij
dx x x x m
0
) ( ) ( ) ( φ φ ρ
The consistent mass matrix associated with the member end acceleration is given (Ref. 1) by
¦
¦
¦
)
¦
¦
¦
`
¹
¦
¦
¦
¹
¦
¦
¦
´
¦
− − −
−
−
−
=
¦
¦
¦
)
¦
¦
¦
`
¹
¦
¦
¦
¹
¦
¦
¦
´
¦
z
y
x
z
y
xx
z
y
x
z
y
x
d
d
d
d
dt
d
L L L L
L L
L L L L
L L
L
m
p
p
m
p
p
2
2
2
1
1
1
2
2
2 2
2 2
2
2
2
1
1
1
4 22 0 3 13 0
22 156 0 13 54 0
0 0 140 0 0 70
3 13 0 4 22 0
13 54 0 22 156 0
0 0 70 0 0 140
420
θ
θ
ρ
The deflection shape function changes in an inelastic stage. It is normally difficult to evaluate the
distributed mass and the deflected shape consistent with member stiffness matrix. The member
mass matrix must be reformulated consistent with the stiffness matrix.
The lateral deflection shape functions for a member with a
rotational plastic hinge at the start end are given (Ref. 1) as
0 . 1 } 1 ) {(
2
1
) (
0 . 1 ) (
2
1
) (
2
3
) (
0 . 1 0 ) (
0 . 1 ) (
2
1
) (
2
3
1 ) (
2
2
6
2
3
5
1 3
1
3
2
= − =
= − =
= =
= + − =
z
y
z
y
for
L
x
x x
d for
L
x
L
x
x
for x
d for
L
x
L
x
x
θ φ
φ
θ φ
φ
The consistent mass matrix for this case is given (Ref. 1) by
¦
¦
¦
¦
)
¦
¦
¦
¦
`
¹
¦
¦
¦
¦
¹
¦
¦
¦
¦
´
¦
− −
−
−
=
¦
¦
¦
¦
)
¦
¦
¦
¦
`
¹
¦
¦
¦
¦
¹
¦
¦
¦
¦
´
¦
z
y
x
z
y
x
z
y
x
z
y
x
d
d
d
d
dt
d
L L L
L
L
L
m
p
p
m
p
p
2
2
2
1
1
1
2
2
2
2
2
2
1
1
1
8 36 0 0 5 . 16 0
36 99 0 0 5 . 58 0
0 0 140 0 0 70
0 0 0 0 0 0
5 . 16 5 . 58 0 0 204 0
0 0 70 0 0 140
420
θ
θ
ρ
The deflection shape functions for a member with a
rotational plastic hinge at the terminal end are given (Ref. 1) as
1
2
1 2
6
0 . 1 0 ) (
0 . 1 ) (
2
1
) (
2
3
) (
0 . 1 ) (
2
1
) (
2
3
) (
0 . 1 ) (
2
1
) (
2
3
1 ) (
2 6
2
3 2
5
1
2
3
1
3
2
= =
= − =
= + − =
= + − =
z
y
z
y
for x
d for
L
x
L
x
x
for
L
x
x
L
x
x x x
d for
L
x
L
x
x
θ φ
φ
θ φ
φ
The consistent mass matrix associated with the member end acceleration is given (Paz, 1985) by
¦
¦
¦
¦
)
¦
¦
¦
¦
`
¹
¦
¦
¦
¦
¹
¦
¦
¦
¦
´
¦
=
¦
¦
¦
¦
)
¦
¦
¦
¦
`
¹
¦
¦
¦
¦
¹
¦
¦
¦
¦
´
¦
z
y
x
z
y
x
z
y
x
z
y
x
d
d
d
d
dt
d
L
L L L
L
L
m
p
p
m
p
p
2
2
2
1
1
1
2
2
2
2
2
2
1
1
1
0 0 0 0 0 0
0 204 0 5 . 16 5 . 58 0
0 0 140 0 0 70
0 5 . 16 0 8 36 0
0 5 . 58 0 36 99 0
0 0 70 0 0 140
420
θ
θ
ρ
The lateral deflection shape functions for a member with
rotational plastic hinges at the two ends are given (Ref. 1) as
0 . 1 0 ) (
0 . 1 ) ( ) (
0 . 1 0 ) (
0 . 1 ) ( 1 ) (
2 6
2 5
1 3
1 2
= =
= =
= =
= − =
z
y
z
y
for x
d for
L
x
x
for x
d for
L
x
x
θ φ
φ
θ φ
φ
The consistent mass matrix associated with the member end acceleration is given by
¦
¦
¦
¦
)
¦
¦
¦
¦
`
¹
¦
¦
¦
¦
¹
¦
¦
¦
¦
´
¦
=
¦
¦
¦
¦
)
¦
¦
¦
¦
`
¹
¦
¦
¦
¦
¹
¦
¦
¦
¦
´
¦
z
y
x
z
y
x
z
y
x
z
y
x
d
d
d
d
dt
d L
m
p
p
m
p
p
2
2
2
1
1
1
2
2
2
2
2
1
1
1
0 0 0 0 0 0
0 2 0 0 1 0
0 0 2 0 0 1
0 0 0 0 0 0
0 1 0 0 2 0
0 0 1 0 0 2
6
θ
θ
ρ
It should be noted that (a) the displacement functions due to unit member end displacement is not
the same as the deflected shape due to member end displacement and inertia force acting along the
member, (b) the stiffness distribution changes along a member in the nonlinear response due to the
progress in damage during an oscillation. Therefore, the deflected shape must be evaluated at each
instance during response analysis. The consistent mass is not constant with time, but is affected by
the stiffness (damage) distribution.
Lumped Mass: A simple method to consider the inertial properties for a dynamic system is to
assume that the mass of the structure is lumped at the nodal coordinates where translational
displacements are defined. The inertia effect associated with any rotational degree of freedom may
be assumed to be zero.
For a member AB of length L and distributed mass ) (x ρ per unit length, the lumped mass may
be defined by considering the rigid body motion;
1
2
7
dx x
L
x
m
dx x
L
x
m
L
B
L
A
) ( ) (
) ( )} ( 1 {
0
0
ρ
ρ
∫
∫
=
− =
The mass matrix may be formed by adding the contribution of lumped masses at the nodal
coordinates defined for translations.
A regular building structure oscillates in a horizontal direction under a horizontal earthquake
motion, causing horizontal inertia forces associated with the mass of floors and beams. The floor
diaphragm may be assumed to be rigid in its own plane in a castinsitu reinforced concrete building,
yielding the same horizontal displacement at a floor. Therefore, the mass of a building may be
assumed to concentrate at the floor level; the mass of interstory structural and nonstructural
elements may be included in the floor mass.
The total mass of a structure may be assumed to concentrate at the levels of the floors, and no
other mass may be ignored as the secondary effect. The mass matrix becomes diagonal with
nonzero elements only at the locations associated with horizontal floor displacements. For
horizontal inertia forces
Xi
P at floor level i , the lumped mass matrix may be written as follows;
¦
¦
¦
¦
)
¦
¦
¦
¦
`
¹
¦
¦
¦
¦
¹
¦
¦
¦
¦
´
¦
=
¦
¦
¦
¦
)
¦
¦
¦
¦
`
¹
¦
¦
¦
¦
¹
¦
¦
¦
¦
´
¦
− − −
Xn
Xn
X
X
X
n
n
Xn
Xn
X
X
X
D
D
D
D
D
dt
d
M
M
M
M
M
P
P
P
P
P
1
3
2
1
2
2
1
3
2
1
1
3
2
1
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
M
L
L
M M O M M M
L
L
K
M
where } {
2
2
X
D
dt
d
: horizontal acceleration at floor level i , and
i
M : total mass of floor level i .
Reference:
1. Mario Paz, Structural Dynamics, Theory & Computation, Van Nostrand Reinhold Co. Ltd.,
1985.
8
5.3 Damping
A massspring (oscillatory) system under free vibration does not oscillate forever, but the
amplitude of oscillation in a real system is known to diminish with time. In other words, kinetic energy
of motion must decrease with time indicating there exists an energy dissipating mechanism in a real
oscillatory system. Such a mechanism of energy dissipation is vaguely termed as "damping."
"Damping" is a mechanism to dissipate kinetic energy; some mechanisms have been suggested
for damping such as;
(a) inelastic hysteresis energy dissipation,
(b) radiation of kinematic energy through foundation,
(c) kinetic friction,
(d) viscosity in materials, or
(e) aerodynamic effect.
Unfortunately, the stateoftheart cannot define the characteristics of damping on the basis of
material properties and geometrical configuration of a structure.
"Because damping mechanism is not clearly understood, we may use a crude model to represent
the energy dissipating feature."
Looking at the equation of motion, there is not a linear term associated with velocity. Therefore, it
is mathematically convenient and beautiful to introduce a velocity related term; i.e., a resistance
proportional to velocity called "viscous damping":
) (
2
2
t f kx
dt
dx
c
dt
x d
m = + +
where c: damping coefficient.
Solution (complementary solution) for free vibration with damping yields
t p t p
c
e C e C x
2 1
2 1
+ =
where
m
mk c c
p
2
4
2
2 , 1
− ± −
=
Critical damping coefficient
cr
c is defined as
mk c
mk c
cr
cr
2
0 4
2
=
= −
for which damping is so large that the oscillation will not be developed. In a normal structural system,
damping coefficient is much smaller than the critical value.
Ratio of a damping coefficient c of a system to the critical damping coefficient
cr
c is called a
damping factor (or ratio) h;
mk
c
h
2
=
The value of a damping factor is normally determined by dynamic tests of a structure rather than
calculated from the material properties and geometry of the structure. Damping capacity is often
determined by the bandwidth at the half power level of the response curve during a sinusoidal
"steadystate test" or by "logarithmic decrement" of the response record during a free vibration test.
Note that the damping capacity is not a unique value of a structure, but is known to vary with the
level of excitation. Acceleration response amplitude was plotted with respect to exciting frequency in
a steadystate test of a reinforced concrete building at different excitation levels (Jennings and
9
Kuroiwa, 1968). The resonant frequency
decreased with the level of excitation and
the damping value increased despite low
response amplitude.
Damping factor is known to be
approximately 0.03 to 0.05 for reinforced
concrete buildings. Radiation damping is
dominant source of energy dissipation in a
heavy reinforced concrete building.
Free vibration or steadystate vibration
test data show a discrepancy from the
theoretical response of systems derived
for viscous damping naturally because the
mechanism of damping in a real structure
is not of viscous type.
n
f
f f
h
1 2
2
1 −
≅
The complementary solution ) (t x
c
using damping coefficient is
) sin cos ( t B t A e x
d d
t h
c
n
ω ω
ω
+ =
−
where
d
ω : circular frequency of a damped system:
2
1 h
n d
− = ω ω
and A, B: integration constants dependent on the initial conditions.
For a given set of initial displacement
o
x and velocity
o
v at 0 . 0 = t , the solution under free
vibration is
] sin cos [ ) ( t
h x v
t x e t x
d
d
n o o
d o
t h
n
ω
ω
ω
ω
ω
+
+ =
−
Damping factor h may be determined by the logarithmic decrement, defined as the natural
logarithm of the amplitude ratio of two consecutive peaks in free vibration;
1
ln
2
1
+
=
m
m
x
x
h
π
10
Viscous damping is the only means to dissipate kinetic energy in a linear system, and damping
has a considerable influence on response amplitude of a linear system.
Damping Matrix: The mechanism of damping in a structure is not of viscous type. The damping
matrix cannot be formulated by the material properties and geometry of a structure. Therefore, it is
not reasonable to assume complicated damping. The damping matrix may be formulated for each
story rather than for each member.
Rayleigh type damping （Lord Rayleigh, 18421919）of the following form is often assumed in the
response analysis of a multidegreeoffreedom system;
] [ ] [ ] [
1
K a M a C
o
+ =
where ] [C : damping matrix, ] [M : mass matrix, and ] [K : stiffness matrix, and a
0
and a
1
are
proportionality constants. The Rayleigh damping is used because the general coordinates (mode
shapes), obtained for an undamped system, are also orthogonal with respect to the damping matrix.
For n th mode, damping factor h
n
is expressed as
n
n
o
n
a
a
h ω
ω
1
+ =
If the damping matrix is made proportional to the mass matrix, damping factors corresponding to
vibration modes decreases with mode number; i.e., higher mode response can be more easily
excited. If the damping matrix is proportional to stiffness matrix, damping factors increases with
mode number; i.e., higher mode response can be suppressed in the response. The constants
o
a
and
1
a are determined by damping factors and frequencies of vibration in arbitrary two modes.
In general, the mode shapes can be made orthogonal if a damping matrix is selected in the
following form (Caughey damping);
∑
−
=
i
i
i
K M a M C ]) [ ] ([ ] [ ] [
1
The Rayleigh damping is a special case of the Caughey damping, in which 0 = i and 1 = i are
considered. A damping factor
n
h of n th mode is given as
i
n
i
i
n
n
a h
2
1
ω
ω
∑
=
Note that damping factors are normally estimated from a series of steadystate dynamic tests of
a structure, but the accuracy of the values is limited due to, first of all, unknown mechanism of
damping, and difficulty in obtaining higher mode modal characteristics.
With the reduction of stiffness caused by damage in en inelastic response, the constant damping
matrix (proportional to constant mass matrix or proportional to initial elastic stiffness matrix) tends to
increase effective damping factors. The additional energy dissipation by damping with damage is
hard to rationalize because hysteretic energy dissipation is considered in the hysteresis of stiffness
models. Therefore, it is recommended to make the damping matrix proportional to instantaneous
stiffness in an in elastic response analysis. Damping matrix proportional to instantaneous stiffness is
reported to be favorable in simulation of earthquake response of test structures (Otani and Sozen,
1972, Omote and Takeda, 1974).
References:
Jennings, P. C., and J. H. Kuroiwa, "Vibration and SoilStructure Interaction Tests of a Ninestory
Reinforced Concrete Building," Bulletin, Seismological Society of America, No. 58, 1968, pp.
891916.
Omote, Y., and T. Takeda, "Study on Elastoplastic Response of Reinforced Concrete Chimney
11
(Part 1: Model Test) (in Japanese)," Transactions, Architectural Institute of Japan, No. 215,
January 1974, pp. 2132.
Otani, S., and M. A. Sozen, "Behavior of Multistory Reinforced Concrete Frame Structures," Civil
Engineering Studies, Structural Research Series No. 392, University of Illinois, 1972.
12
5.4 Strainrate Effect
The speed of loading has been known to influence the stiffness and strength of various materials
since 19th century; however, the difficulty in testing materials under controlled loading rate and in
instrumenting the response delayed the understanding of the characteristics.
Material, subjected to fast loading (Cowell, 1965, 1966),
(a) increases initial stiffness,
(b) increases resistance at yielding, but
(c) does not change hysteresis characteristics, and
(d) does not change the behavior after strain hardening,
including ductility and energy dissipation characteristics. For example, strain rate of 0.3 mm/mm/sec
was reported to increase the upper yield point stress by 1.60 times from the yield stress at
quasistatic loading; such loading rate was intended for explosion rather than for earthquake
response. The strain rate expected in concrete material during an earthquake is in the order of 0.001
to 0.10 mm/mm/sec (Mahin and Bertero, 1972).
Strain rate effect on material stressstrain relations
(Manjoine,1944, and Mahin and Bertero, 1972)
Strain, mm/mm
Strain, mm/mm
T
e
n
s
i
l
e
s
t
r
e
s
s
,
k
g
f
/
c
m
2
C
o
m
p
r
e
s
s
i
v
e
s
t
r
e
s
s
,
k
g
f
/
c
m
2
Quasistatic
Reinforcing bars were tested under constant
strain rate by Mahin and Bertero (1972).
Stressstrain relation under stress reversals in
plastic region will not be affected by the strain
rate.
Average bond stressbar slip relation (gauge
length of 30 mm) of deformed and plain bars was
obtained under impact pullout loading by Vos and
Reinhardt (1982); the bond stress rate was
varied from 0.3 MPa/sec to 105 MPa/sec. The
loading rate did not influence the bond
characteristics of plain bars; bond resistance
remained constant after 0.02 mm bar slip.
Deformed bars increased the initial stiffness and
average bond stress with loading rate in
deformed bars; the bond resistance increased
with bar slip. The loading rate effect decreases
Strain, mm/mm
T
e
n
s
i
l
e
s
t
r
e
s
s
o
f
s
t
e
e
l
,
k
g
f
/
c
m
2
Stressstrain relation of steel under load reversals
13
with strength of concrete. The difference
in the loading rate effect on deformed
bars and plain bars might be interpreted
that chemical adhesion bond resistance
is not influenced by loading rate, but
mechanical interlocking bond resistance
is influenced by the loading rate.
Tests of reinforced concrete members
at constant velocity of 0.002 rad/sec and
0.2 rad/sec of member rotational velocity
(Mahin and Bertero, 1972) indicated that
(a) high strain rates increased the
initial yield resistance, but caused small
differences in either stiffness or
resistance in subsequent cycles at the
same displacement amplitudes,
(b) strain rate effect on resistance
diminished with increased deformation in
a strainhardening range, and
(c) no substantial changes were
observed in ductility and overall energy
absorption capacity.
Earthquake simulator tests of
reinforced concrete members and
structural models have been successfully
simulated by analysis on the basis of
hysteresis model developed on the basis
of forcedeformation relationship
observed in static laboratory tests
(Takeda, Nielsen and Sozen, 1970,
Otani and Sozen, 1972, Koike, Omote and Takeda, 1980,).
Displacement, cm Displacement, cm
R
e
s
i
s
t
a
n
c
e
,
t
o
n
f
R
e
s
i
s
t
a
n
c
e
,
t
o
n
f
Note that strain rate (velocity) during an oscillation is highest at low stress levels, and that the rate
gradually decreases toward a peak strain (displacement). Damage in the reinforced concrete
reduces the stiffness, elongating the period of oscillation. Furthermore, such damage is caused
normally by lower modes of oscillation having longer periods. Therefore, the strain rate effect can be
judged small on the earthquake response of a normal reinforced concrete structure.
Bar slip
Bar slip, mm
A
v
e
r
a
g
e
b
o
n
d
s
t
r
e
s
s
,
k
g
f
/
c
m
2
Bond stress speed
Bar slip, mm
A
v
e
r
a
g
e
b
o
n
d
s
t
r
e
s
s
,
k
g
f
/
c
m
2
Average bond stressbar slip relation for D10 bars
(vos and Reinhardt, 1982)
Bond stress speed
14
References:
Cowell, W. L., "Dynamic Tests of Concrete Reinforcing Steels," Technical Report No. 394, U.S.
Naval Civil Engineering Laboratory, Port Hueneme, California, 1965.
Cowell, W. L., "Dynamic Properties of Plain Portland Cement Concrete," Technical Report No. 447,
U.S. Naval Civil Engineering Laboratory, Port Hueneme, California, 1966.
Koike, K., Y. Omote and T. Takeda, "Reinforced Concrete WallFrame Structures subjected to
Dynamic and Static Loading," Proceedings, Eighth World Conference on Earthquake Engineering,
Vol. 6, September 1980, 419426.
Mahin, M. A., and V. V. Bertero, "Rate of Loading Effect on Uncracked and Repaired Reinforced
Concrete Members," EERC No. 729, Earthquake Engineering Research Center, University of
California at Berkeley, 1972.
Manjoine, M. J., "Influence of Rate of Strain and Temperature on Yield Stresses of Mild Steel,"
Journal of Applied Mechanics, Transactions, American Society of Mechanical Engineers, Vol. 11,
December 1944, pp. A211A218.
Otani, S., and M. A. Sozen, "Behavior of Multistory Reinforced Concrete Frames during
Earthquakes," Civil engineering Studies, Structural Research Series No. 392, University of Illinois,
November 1972.
Takeda, T., N. N. Nielsen and M. A. Sozen, "Reinforced Concrete Response to Simulated
Earthquake," Proceedings, Structural Engineering Division, ASCE, Vol. 96, No. St12, December
1970, pp. 25572573.
Vos, E., and H. W. Reinhardt, "Influence of Loading Rate on bond Behavior of Reinforcing Steel and
Prestressing Strands," Materials and Structures, Vol. 15, No. 85, March 1982, pp. 310.
Further References on Strain Rate Effect
1. Atchley, B. L., H. L. Furr, "Strength and Energy Absorption Capabilities of Plain Concrete under
Dynamic and Static Loadings," Journal, American Concrete Institute, Vol. 64, No. 11, November
1967, pp. 7457556.
2. Birkimer, D. L., and R. Lindemann, "Dynamic Tensile Strength of Concrete Materials," Journal,
American Concrete Institute, Vol. 68, No. 1, January 1971, pp. 4749.
3. Clark, D. S., and P. E. Duwez, "The Influence of Strain Rate on Some Tensile Properties of Steel,"
Proceedings, American Society of Testing Materials, Vol. 50, 1950, pp. 560575.
4. Criswell, M. E., "Static and Dynamic Response of Reinforced Concrete SlabColumn Connections,
ACI SP42, Shear in Reinforced Concrete, Vol. II, 1974, pp. 721746.
5. Crum, R. G., "Tensile Impact Tests for Concrete Reinforcing Steel," Journal, American Concrete
Institute, Vol. 56, No. 1, July 1959, p. 59.
6. Kaplan, S. A., "Factors Affecting the Relationship between Rate of Loading and Measured
Compressive Strength of Concrete," Magazine of Concrete Research, Vol. 32, No. 111, June
1980, pp. 7988.
7. Manjoine, M. J., and A. Nadai, "High Speed Tension Tests at Elevated Temperatures, Part 1,"
Proceedings, American Society of Testing Materials, Vol. 40, 1940, pp. 822837.
8. Seabolt, R. H., "Dynamic Shear Strength of Reinforced Concrete Beams, Part II," Technical
Report R502, U. S. Naval Civil Engineering Laboratory, California, January 1967, pp. 5256.
9. Shiga, T. and J. Ogawa, "The Experimental Study on the Dynamic Behavior of Reinforced
Concrete Frames," Proceedings, Fifth World Conference on Earthquake Engineering, Santiago,
Chile, 1969, B2, pp. 166176.
10. Sparks, P. E., and J. R. Menzies, "The Effect of Rate of Loading upon the Static and Fatigue
Strengths of Plain Concrete in Compression," Magazine of Concrete Research, Vol. 25, No. 83,
June 1973, pp. 7380.
11. Takeda, J, and H. Tachikawa, "Deformation and Fracture of Concrete subjected to Dynamic
Load," Proceedings, International Conference on the Mechanical Behavior of Materials, Kyoto,
Vol. IV, 1971.
12. Watstein, D. "Effect of Strain Rate on the Compressive Strength and Elastic Properties of
Concrete," Journal, American Concrete Institute, Vol. 49, No. 8, April 1953, pp. 729744.
13. Zielinski, A. J., H. W. Reinhardt and H. A. Kormeling, "Experiments on Concrete under Uniaxial
Impact Tensile Loading," Materials and Structures, Vol. 14, No. 80, MarchApril 1981, pp.
103112.
15
5.5 Properties of Earthquake Ground Motion
The earthquake risk is not uniform on the earth, but varies significantly from a region to another.
The recent development in seismology is fascinating; plate tectonics developed since the 1960s can
explain the occurrence of earthquakes along the boundaries of tectonic plates. Most major
earthquakes occur along the boundaries of tectonic plates due to their relative movement, which can
be monitored with the use of the global positioning system (GPS). Seismically blank regions where
next large earthquakes may occur can be identified by the observation and historical records. These
plate boundary earthquakes occur at a relatively uniform interval (50 to a few hundred years) for a
given region with accumulation of strain energy.
Some faults have been identified on ground surface, but others are buried under the ground.
Some earthquakes of lesser magnitudes occur by the fracture of active faults within a tectonic plate
caused by stresses developed by the plate movement. Some active faults have been identified on
ground surface, but others are buried under the ground. An active fault in a tectonic plate may
fracture once in a few thousand years.
It is not possible at this stage to accurately predict the time, location and magnitude of an
earthquake occurrence.
Earthquake Ground Motions: A seismometer to measure ground displacement during an
earthquake was developed in late nineteenth century. The seismometer has been used by
seismologists to understand the source mechanism of earthquakes, but it does not provide
acceleration records necessary for engineering purpose. The seismologist believed that the
acceleration signal was affected by accidental phenomena such as local geology.
A strong motion accelerograph to record ground acceleration was developed in the early 1930s.
The characteristics of earthquake ground motions were studied through the observed records; i.e.,
common features of acceleration records were abstracted and general shapes of response spectra
were established for design purpose taking the local effect of soil into account.
Historical records about earthquake occurrences are studied to estimate the probability of the
maximum earthquake intensity in a region. A large uncertainty exists in the estimated maximum
ground acceleration attributable to the inaccuracy and the limited period of the historical
documentation.
Engineering Seismology: Earthquake ground motions specific at a construction site is influenced
by the geometry of active faults, dynamic rupture process of earthquake sources, and the
transmission of earthquake motions from the earthquake source to the construction site. There have
been efforts by engineering seismologists to estimate the characteristics of future earthquake
motions.
The global parameters (fault length, width and seismic moment) of future earthquakes can be
estimated by the seismic history, geological investigation and source modeling of active faults near
the construction site. The local source parameters (slip heterogeneity on fault plane) are important to
characterize the fault movement along the slip plane, especially the slip and slip velocity.
The local parameters, such as the transmission characteristics of earthquake motion from
earthquake source to the construction site, cannot be evaluated theoretically, but must be
determined by the source inversion of past major events with the use of statistical analysis or by the
observation of minor earthquakes. The transfer function from the source to the construction site may
be estimated by an empirical Green’s function; i.e., the transfer function of past small earthquakes in
the region. The intensity of earthquake wave decays with distance.
The earthquake wave is generally transmitted from the fracture fault through hard rock layers and
then to the ground surface through relatively soft surface soil. The characteristics of ground motion
are significantly modified by the properties of surface soil layers, such as the properties and
16
geometry of the subsurface soil layers, surface topography, and depth and properties of the
underlying bedrock. Soft soil layers consisting of river deposits tend to amplify long period
components of an earthquake motion, causing serious damage to houses and buildings.
It is important to recognize that seismic design of a structure is based on a large uncertainty
about the characteristics, especially intensity, of a design earthquake motion.
Reference:
Architectural Institute of Japan: Seismic Loading  State of the Art and Future Developments (in
Japanese), November 1988, 438 pp.
1
Chapter 6. Numerical Integration of Equation of Motion
6.1 Introduction
The equation of motion of a structure (dynamic equilibrium of horizontal forces) under horizontal
earthquake motion may be expressed;
i i i i
i i i i
y e M R D x M
t y
dt
d
e M t R t D t x
dt
d
M
& & & & } ]{ [ } { } { } ]{ [
) ( } ]{ [ )} ( { )} ( { )} ( { ] [
2
2
2
2
− = + +
− = + +
in which [M]: mass matrix,
i i
x t x
dt
d
} { )} ( {
2
2
& & = : acceleration vector relative to the structure’s base at
time t
i
,
i i
D t D } { ) ( { = : damping force vector at time t
i
,
i
R t R
i
} { )} ( { = : resistance vector at time t
i
,
i i
y t y
dt
d
& & = ) (
2
2
: horizontal ground acceleration at time t
i
, } {e : vector with elements equal to 1.0 for
horizontal degrees of freedom and zero for the rest of degrees of freedom.
Assuming that resistance and damping force vary linearly with displacement and velocity relative
to the base, respectively, over a short time increment t ∆
from
i
t to
1 + i
t , the equation of motion at
the new time step
1 + i
t may be written;
1 1 1 1
1 1 1 1
[ ]{ } { } { } [ ]{ }{ }
[ ]{ } { } { } { } { } [ ]{ }{ }
i i i i
i i i i i i
M x D R M e y
M x D D R R M e y
+ + + +
+ + + +
+ + = −
+ + ∆ + + ∆ = −
&& &&
&& &&
or expressing incremental resistance
1
{ }
i
R
+
∆ as the product of instantaneous stiffness [ ]
i
K at
time t
i
and incremental displacement
1
{ }
i
x
+
∆ , and incremental damping force
1
{ }
i
D
+
∆ as the
product of instantaneous damping matrix [ ]
i
C at time t
i
and incremental velocity
1
{ }
i
x
+
∆& ;
1 1 1 1
1 1 1 1
[ ]{ } { } [ ]{ } { } [ ]{ } [ ]{ }{ }
[ ]{ } [ ]{ } [ ]{ } [ ]{ }{ } { } { }
i
i i i i i i
i i i i i
M x D C x R K x M e y
M x C x K x M e y R D
+ + + +
+ + + +
+ + ∆ + + ∆ = −
+ ∆ + ∆ = − − −
&& & &&
&& & &&
1
} {
+ i
x& & ,
1
} {
+
∆
i
x& and
1
} {
+
∆
i
x are solved by a numerical integration procedure and numerical
solution of a set of linear algebraic equations.
It is important to evaluate
1
} {
+ i
R and
1
} {
+ i
D on the basis of calculated displacement
1
} {
+ i
x and velocity
1
} {
+ i
x& at time
1 + i
t
considering nonlinear stiffness and damping
characteristics. The damping force and
resistance should not be calculated as
1 1
1 1
} }{ [ } { } {
} ]{ [ } { } {
+ +
+ +
∆ + =
∆ + =
i i i
i i i
x C D D
x K R R
&
because the stiffness and damping may not be
proportional to incremental velocity and
displacement during the time increment
1 +
∆
i
t
due to nonlinearity.
x
i
x
i+1
R
i
R
i+1
1 i
R
+
∆
1 i
K x
+
∆
1 i
x
+
∆
2
6.2 NigamJennings' Direct Integration Method
Consider a linearly elastic singledegreeoffreedom (SDF) system, with mass m, damping
coefficient c and stiffness k, subjected to linearly varying load p(t):
b at p kx x c x m
t p kx
dt
dx
c
dt
x d
m
+ = = + +
= + +
& & &
) (
2
2
The solution of a differential equation is expressed as
the sum of a particular solution x
p
(t), which satisfies
the original differential equation, and a complementary
solution x
c
(t), which is a general solution of the
differential equation for right hand function to be zero.
There is no systematic way to search for a particular
solution; the complementary solution is given as the
solution for free vibration.
0
p p p
c c c
mx cx kx at b
mx cx kx
+ + = +
+ + =
&& &
&& &
A particular solution
p
x for this differential equation is assumed to be of the following form;
2 1
c t c x
p
+ =
then, differentiating the relation with time,
0
1
=
=
p
p
x
c x
& &
&
The above relations are substituted into the equation of motion to determine the constants c
1
and c
2
of the assumed particular solution:
b at c t c k cc + = + + ) (
2 1 1
or
0 ) ( ) (
2 1 1
= − + + − b kc cc t a kc
The particular solution must satisfy the differential equation for any time t; hence
0 ) (
0 ) (
2 1
1
= − +
= −
b kc cc
a kc
Solving for c
1
and c
2
, we obtain
k
k
a
c b c
k
a
c
/ ) (
2
1
− =
=
or
2
k
ac bk
t
k
a
x
p
−
+ =
The complementary solution is given as
( cos sin )
n
h t
c d d
x e A t B t
ω
ω ω
−
= +
where
mk
c
h
m
k
h
n
n d
2
1
2
=
=
− =
ω
ω ω
t
p(t)
a
b
3
d
ω : damped circular frequency,
n
ω : undamped circular frequency, h: damping factor. A and B are
integration constants dependent on the initial condition.
Therefore, the complete solution x(t) for the loading is
) sin cos (
2
t B t A e
k
ac bk
t
k
a
x x x
d d
t h
c p
n
ω ω
ω
+ +
−
+ =
+ =
−
The velocity is given by differentiating the displacement with time;
)} cos sin ( ) sin cos ( { t B t A t B t A h e
k
a
x
d d d d d n
t h
n
ω ω ω ω ω ω
ω
+ − + + − + =
−
&
For given initial displacement x
o
and velocity v
o
at t = 0,
B A h
k
a
v
k
ac bk
A x
d n o
o
ω ω + − =
−
+ =
2
The integration constants A and B are solved from the initial conditions;
) (
1
2
2
k
a
k
ac bk
h v x h B
k
ac bk
x A
n o o n
d
o
−
−
− + =
−
− =
ω ω
ω
Namely, the response at any time t (> 0) is expressed as a function of the system constants h, k,
n
ω
and
d
ω , the loading function p(t)=a t + b, and the initial conditions x
o
and v
o
at time t = 0.
In general, if a linearly varying loading function is given from t=t
i
to t=t
i+1
as
1 1
1
1
1
1
) (
) ( ) (
+ +
+
+
+
+
+ =
−
∆
∆
+ =
−
−
− + =
i i
i
i
i
i
i i
i
i i i
b a
t t
t
p
p
t t
t t
p p p t p
τ
where
i
i i
i
i
i
t t
p b
t
p
a
− =
=
∆
∆
=
+
+
+
+
τ
1
1
1
1
The response at any time t (
1 +
< <
i i
t t t ,
1
0
+
∆ < <
i
t τ ) can be expressed by known initial conditions
x
i
and
i
x& at time
i
t t = (τ = 0.0). Therefore, the response velocity
1 + i
x& and displacement x
i+1
at
time t
i+1
are determined.
4
)} cos sin (
) sin cos ( {
1
) sin cos ( ) (
1
) sin cos ( ) (
1
1 1
1 1
1
1
1
1 1
1
1
1
1 1
1
1
2
1
1
1
1
1
1
1
+ +
+ +
∆ −
+
+
+
+ +
∆ −
+
+
+
+ +
∆ −
+
+ +
+
+
+
∆ + ∆ − +
∆ + ∆ − +
∆
∆
=
∆ + ∆ +
∆
∆
− =
∆ + ∆ +
∆
∆
− +
∆
∆
∆
=
+
+
+
i d i d d
i d i d n
t h
i
i
i
i d i d
t h
i
i
i
i d i d
t h
i
i
i
i
i
i
i
t B t A
t B t A h e
t
p
k
x
t B t A e
k
c
t
p
p
k
t B t A e c
t
p
k p
k k
t
t
p
x
i n
i n
i n
ω ω ω
ω ω ω
ω ω
ω ω
ω
ω
ω
&
}
1
) ( {
1
) (
1
1
1
1
1
2
1
1
2
+
+
+
+
+
+
∆
∆
−
∆
∆
− − + =
∆
∆
− − =
i
i
i
i
i
n
i i n
d
i
i
i i
t
p
k t
p
c k p
k
h
x x h B
c
t
p
k p
k
x A
ω
ω
ω
&
The procedure may be successively applied as long as the loading function is given in a piecewise
linear form over a short time increment.
The response can be calculated in the following form:
m
kx x c p
x
p B p B x B x B x
p A p A x A x A x
i i i
i
i i i i i
i i i i i
1 1 1
1
4 1 3 2 1 1
4 1 3 2 1 1
+ + +
+
+ +
+ +
− −
=
+ + + =
+ + + =
&
& &
& &
&
where,
t
h
t
t
h
t
t
h
h
h
e A
t
h
t
h
t
h
t
e A
t e A
t t
h
h
e A
n
d
n
d
d n
t h
n n
d
n
d
d n
t h
d
d
t h
d d
t h
n
n
n
n
∆
+ ∆
∆
+ + ∆
∆
−
+
−
=
∆
− + ∆ + ∆
−
∆
=
∆ =
∆ + ∆
−
=
∆ −
∆ −
∆ −
∆ −
3
2
2
2 4
2
2
2 3
2
2
1
2
} cos )
2
1 ( sin )
1 2
1
{(
1
}
2
1 {
1
} cos
2
sin
1 2
{
1
sin
1
} cos sin
1
{
ω
ω
ω
ω
ω ω
ω ω
ω
ω
ω
ω ω
ω
ω
ω ω
ω
ω
ω
ω
t
t h t
t
h
t
h
h
t
t
h
e B
t
h
h
t e B
t
h
e B
n
d n d d
n
d d
n
t h
d d
t h
d
n
t h
n
n
n
∆
+ ∆ + ∆
∆
−
∆
−
− ∆
∆
−
=
∆
−
− ∆ =
∆
−
− =
∆ −
∆ −
∆ −
2 3
2
2
2
3
2
2
2
1
1
)} cos sin (
2
) sin
1
(cos
1 2
{
) sin
1
(cos
sin
1
ω
ω ω ω ω
ω
ω ω
ω
ω ω
ω
ω
ω
ω
ω
t
t
h
h
t
t
h h
t h t
t
h
e B
n
d d
n n
d n d d
n n
t h
n
∆
− ∆
−
− ∆
∆
−
+ −
∆ + ∆
∆
+ =
∆ −
2
2
2
2
3 2 4
1
)} sin
1
)(cos
1 2
(
) cos sin )(
2 1
{(
ω
ω ω
ω ω
ω ω ω ω
ω ω
ω
A1 = E (H S + C)
5
A2 = E S /
d
ω
A3 = E (1 / k)(H1 S + H2 C) + (1  H2) / k
A4 =  E [(H + H1) S + (1 + H2) C] / k + H2 / k
B1 =  E
n
ω S /
2
1 h −
B2 = E (C  H S)
B3 = E {H1(
d
ω C  h
n
ω S)  H2(
d
ω S + h
n
ω C)} / k + 1 / (k t ∆ )
B4 = E [(1 + H2)(
d
ω S + h
n
ω C)  (H + H1)(
d
ω C  h
n
ω S)} / k  1 / (k t ∆ )
where
C = cos
d
ω t ∆
S = sin
d
ω t ∆
E = e
 h
n
ω t ∆
H = h /
2
1 h −
H1 = (2h
2
1) / (
d
ω t ∆ )
H2 = 2h / (
n
ω t ∆ )
The coefficients A
i
and B
i
(i= 1, 2, 3, and 4) can be made constant if time increment t ∆ of
numerical integration is fixed and system properties m, c, and k do not change with time. In such a
case, the computing time may be significantly reduced.
Not that this procedure is exact if the exciting function is given as a series of piecewise linear
functions.
Therefore, this procedure is most desirable for the response analysis of a linearly elastic
singledegreeoffreedom system. However, in the case of a nonlinear problem, the coefficients As
and Bs must be evaluated whenever the tangent stiffness k and damping coefficient c are altered.
Reference:
Nigam, N.C., and P. C. Jennings, "Calculation of Response Spectra from StrongMotion Earthquake
Records," Bulletin, Seismological Society of America, Vol. 59, No. 2, April 1969, pp. 909  922.
6
6.3 Linear Acceleration Method
The Taylor series expansion of a vector function {x} with respect to time t about time t
o
is
expressed as
o
o o o
t t
k
k
k
k
o
t t
o
t t
o
t t o o
dt
x d
k
t t
dt
x d t t
dt
x d t t
dt
dx
t t t x t x
=
=
= = =
∑
−
=
+
−
+
−
+ − + =
} {
!
) (
} {
! 3
) (
} {
! 2
) (
} ){ ( )} ( { )} ( {
0
3
3 3
2
2 2
L
The accuracy of the expression will be improved by considering more terms and by evaluating the
function near t
o
.
The displacement and velocity are expanded by the Taylor series at time t
i
to evaluate the
functions at time t
i+1
;
L & & & & & & &
L & & & & & &
+
∆
+ ∆ + =
+
∆
+
∆
+ ∆ + =
+
+
i i i i
i i i i i
x
t
x t x x
x
t
x
t
x t x x
} {
2
} { } { } {
} {
6
} {
2
} { } { } {
2
1
3 2
1
where
i i
t t t − = ∆
+1
.
If the acceleration is assumed to vary linearly over a short time increment t ∆ ;
1
{ } { }
{ }
i i
i
x x
x
t
+
−
=
∆
&& &&
&&&
The following relations are obtained for displacement and velocity increments;
1 1 1
1
2 2
1 1
} {
2
} {
2
} { } { } {
} {
6
} {
3
} { } { } { } {
+ + +
+ + +
∆
+
∆
= − = ∆
∆
+
∆
+ ∆ = − = ∆
i i i i i
i i i i i i
x
t
x
t
x x x
x
t
x
t
x t x x x
& & & & & & &
& & & & &
The equation of motion at time t
i+1
may be solved for
1
} {
+ i
x& & ,
1
} {
+
∆
i
x& and
1
} {
+
∆
i
x ;
i i i i i i
R D y e M x K x C x M } { } { } ]{ [ } ]{ [ } ]{ [ } ]{ [
1 1 1 1
− − − = ∆ + ∆ +
+ + + +
& & & & &
Using the integration relation and the equation of motion, acceleration
1
{ }
i
x
+
&& at new time step t
i+1
can be solved
2 2
1 1
([ ] [ ] [ ]){ } { } [ ]{ } ([ ] [ ]){ } ( [ ] [ ]){ }
2 6 2 3
i i i i i
t t t t
m c k x p k x c t k x c k x
+ +
∆ ∆ ∆ ∆
+ + = − − + ∆ − + && & &&
i i i i i
x A x A x A p A x } ]{ [ } ]{ [ } ]{ [ } ]{ [ } {
4 3 2 1 1 1
+ + + =
+ +
& & & & &
where,
] ][ [ ] [
]) [ ] ]([ [ ] [
]) [
3
] [
2
]( [ ] [
]) [
6
] [
2
] ([ ] [
1 4
1 3
2
1 2
1
2
1
k A A
k t c A A
k
t
c
t
A A
k
t
c
t
m A
− =
∆ + − =
∆
+
∆
− =
∆
+
∆
+ =
−
The displacement and velocity are calculated using the linear acceleration procedure;
7
1 1
1
2 2
1
} {
2
1
} {
2
1
} { } {
} {
6
1
} {
3
1
} { } { } {
+ +
+ +
∆ + ∆ + =
∆ + ∆ + ∆ + =
i i i i
i i i i i
x t x t x x
x t x t x t x x
& & & & & &
& & & & &
The time increment t ∆ for the response analysis must be chosen to satisfy the following
conditions;
(a) Excitation function, hysteresis relations, and response waveforms can be expressed
with a satisfactory accuracy,
(b) The accuracy of response results can be attained, and
(c) The numerical integration gives stable results.
The solution of the linear acceleration method diverges if the time increment is selected larger
than 1/3 of the shortest modal period of oscillation of the system.
8
6.4 Newmark Beta Method
For a set of differential equation of motion given below,
i i i i i i
R D y e M x K x C x M } { } { } ]{ [ } ]{ [ } ]{ [ } ]{ [
1 1 1 1
− − − = ∆ + ∆ +
+ + + +
& & & & &
Newmark (1959) suggested the following relations for the numerical integration;
1 1 1
1
2 2
1 1
} { } { ) 1 ( } { } { } {
} { } { )
2
1
( } { } { } { } {
+ + +
+ + +
∆ + ∆ − = − = ∆
∆ + ∆ − + ∆ = − = ∆
i i i i i
i i i i i i
x t x t x x x
x t x t x t x x x
& & & & & & &
& & & & &
γ γ
β β
where β and γ are constants of the Newmark Beta method.
The value of γ must be 1/2, and 1/ 4 β ≤ . If the value of γ is selected to be greater than 1/2, the
response amplitude becomes greater than the true value; if the value is smaller than 1/2, then the
amplitude becomes smaller.
For the numerical integration to give stable results, the time increment t ∆ must be less than
onesixth of the shortest modal period of the system for β = 1/6. The scheme is known to be
unconditionally stable for β = 1/4.
The equation motion can be solved for the acceleration
1
} {
+ i
x& & at the new time step;
i i i i i i
i
x K t x K t x C t R D p
x K t C t M
} ]{ [ )
2
1
( } ]{ [ } ]{ [ ) 1 ( } { } { } {
} ]){ [ ] [ ] {[
2
1
1
2
& & & & &
& &
∆ − − ∆ − ∆ − − − − =
∆ + ∆ +
+
+
β γ
β γ
Newmark suggested to solve the set of equations by an iteration method;
(a) Assume acceleration vector
*
1
} {
+ i
x& & at time t
i+1
,
(b) Evaluate the displacement increment
*
1
} {
+
∆
i
x and velocity increment
*
1
} {
+
∆
i
x& by the
following relations;
*
1
*
1
*
1
*
1
*
1
*
1
*
1
2 2 *
1
} { } { } {
} { } { } {
} { } { ) 1 ( } {
} { } { )
2
1
( } { } {
+ +
+ +
+ +
+ +
∆ + =
∆ + =
∆ + ∆ − = ∆
∆ + ∆ − + ∆ = ∆
i i i
i i i
i i i
i i i i
x x x
x x x
x t x t x
x t x t x t x
& & &
& & & & &
& & & & &
γ γ
β β
(c) Evaluate damping force {D}*
i+1
and resistance {R}*
i+1
for the calculated velocity
*
1
} {
+ i
x&
and displacement
*
1
} {
+ i
x on the basis of damping and hysteresis models;
(d) Reevaluate the acceleration
1
} {
+ i
x& & by the equation of motion at time t
i+1
;
1 1 1 1
} { } { } ]{ [ } ]{ [
+ + + +
− − − =
i i i i
R D y e M x M & & & &
(e) If the reevaluated acceleration
1
} {
+ i
x& & differs from the assumed acceleration
*
1
} {
+ i
x& & by
more than a specified tolerance } {ε , the assumed acceleration
*
1
} {
+ i
x& & is replaced by the
reevaluated acceleration
1
} {
+ i
x& & in step (a), and the procedure is iterated until a satisfactory
conversion is achieved.
Sharpe and Carr (1974) studied a condition for numerical stability as follows;
1.0
derived acceleration true acceleration
assumed acceleration true acceleration
−
≤
−
at the end of each time step using Newmark Beta scheme. They derived a convergence criteria as
9
follows;
2
1 1
[ ( ) ]
2 2 2
t h h
T π β β β
∆
< − + +
Newmark (1959) showed the convergence criteria for an undamped linear system as follows;
1 1
2
t
T π β
∆
≤
For
1
6
β = equivalent to the linear acceleration scheme,
0.39
t
T
∆
≤
Sharpe and Carr (1974) extended Newmark’s derivation of convergence for undamped system to
damped system. The criterion is shown below;
2
1
1 4
t h
T π β
∆ −
<
−
References
Newmark, N. M., "A Method of Computation for Structural Dynamics," Journal, Engineering
Mechanics Division, ASCE, July 1959, pp. 67  94.
Sharpe, R. D., and A. J. Carr, "The Seismic Response of Inelastic Structures," Research Report
7413, Department of Civil Engineering, University of Canterbury, New Zealand, November
1974.
10
6.5 Wilson Theta Method
When the finite element method was applied in a
structural dynamic problem, a major problem was the
computation time because the finite element analysis deals
a system having a large number of degrees of freedom. In
order to reduce the computation time, a stable numerical
method allowing the use of a large time increment t ∆ is
desired. Wilson (Wilson, 1968, Bathe and Wilson, 1973)
proposed a stable numerical integration scheme.
A unique feature of this method is to satisfy the
equation of motion at an imaginary time
*
1 i i
t t t θ
+
= + ∆
(θ > 1.0) for an extrapolated imaginary force {p}*
i+1
θ ] } { } [{ } { } {
1
*
1 i i i i
p p p p − + =
+ +
Using the Newmark Beta method, the equation of motion at time t=t*
i+1
is given as
2
1
2
1
{[ ] [ ] ( ) [ ]){ }
1
{ } { } { } (1 ) [ ]{ } [ ]{ } ( )( ) [ ]{ }
2
i
i i i i i i
M t C t K x
p D R t C x t K x t K x
γ θ β θ
γ θ θ β θ
+
+
+ ∆ + ∆
= − − − − ∆ − ∆ − − ∆
&&
&& & &&
The relationship may be solved for
*
1
} {
+ i
x& & , and the acceleration at time t
i+1
is determined by the
interpolation (assuming linear variation of acceleration over time increment);
i i i
x x x } ){
1
1 ( } {
1
} {
*
1 1
& & & & & &
θ θ
− + =
+ +
Velocity
1
} {
+ i
x& and displacement
1
} {
+ i
x at time t t + ∆ are evaluated by the Newmark Beta method;
1 1
1
2 2
1
} { } { ) 1 ( } { } {
} { } { )
2
1
( } { } { } {
+ +
+ +
∆ + ∆ − + =
∆ + ∆ − + ∆ + =
i i i i
i i i i i
x t x t x x
x t x t x t x x
& & & & & &
& & & & &
γ γ
β β
The value of θ was suggested to be greater than 1.37 for numerical stability; θ = 1.37 to 1.40
is often used.
The Wilson's Theta method is known to introduce numerical damping, a fictitious damping caused
by the numerical integration scheme especially for higher mode (short period) oscillation; the period
of oscillation is elongated and response amplitude decays with time (Bathe and Wilson, 1973). It is
believed that some error may be tolerated in higher mode oscillation because the higher mode
response may not govern the total response in a normal case.
Stability and Accuracy: It is important in selecting numerical integration scheme to examine the
stability of the procedure. Some scheme is stable when the time increment is selected less than a
certain fraction of the period of the highest mode; otherwise, error is continuously amplified during
the numerical integration and the response diverges with time. The stability of a numerical
integration scheme may be examined for a linearly elastic singledegreeoffreedom system without
any damping under free vibration;
2
0.0
n
x x ω + = &&
The error can be expressed in terms of period elongation and amplitude decay as a function of time
increment t ∆ over natural period T .
The accuracy of numerical integration is another problem. Although a numerical integration
scheme may be unconditionally stable for the choice of time increment, the result may not be
accurate if the time increment is large with respect to the natural period. Such inaccuracy may be,
i
t
1 i
t
+
t θ ∆
1 i
t
+
∆ *
1 i
t
+
t
p(t)
11
sometimes, tolerated for higher mode response because the higher mode response is often
negligibly small compared with the dominant response.
12
It is often said that “numerical damping can be considered a good feature in numerical schemes
because it may be used to damp out and practically suppress the response of those modes for
which the response cannot be calculated accurately.” However, the range of errors should be
estimated.
A sixstory twobay linearly elastic frame was analyzed using the Wilson's Theta method and the
Newmark Beta method (Sharpe and Carr, 1974). Vertical and horizontal masses were considered at
each node. The stability criterion for an analysis using the linear acceleration technique requires a
time step of approximately 1/400 sec for this structure. Beta value of the Newmark Beta method was
varied from 1/12 ( t ∆ = 1/400 sec) to 1/4 ( t ∆ = 1/100 sec), and theta value of the Wilson's Theta
method was selected to be 1.5 and 2.0 with a time increment of t ∆ =1/100 sec.
The top story displacement waveforms are compared. The response waveforms calculated by the
Newmark Beta method were almost identical, while the waveforms calculated by the Wilson Theta
method showed a difference from the response waveforms calculated by the Newmark Beta
methods. This particular response was dominated by the first mode component, but the appreciable
difference can be observed. If the oscillation is governed by higher frequency components such as
the acceleration waveform, then the effect of numerical (artificial) damping would appear more in the
calculated waveform.
It should be noted that the role of the analytical tool is to give results as close to the exact solution
as possible. In this standpoint, the Wilson Theta method does not satisfy the criteria.
References:
Bathe, K.J., and E. L. Wilson, "Stability and Accuracy Analysis of Direct Integration Methods,"
International Journal of Earthquake Engineering and Structural Dynamics, Vol. 1, 1973, pp. 283
 291.
Bathe, K.J., and E. L. Wilson, "Linear and Nonlinear Earthquake Analysis of Complex Structures,"
Proceedings, World Conference on Earthquake Engineering, Rome, Italy, 1973, Paper No. 224.
13
Hilber, H.M., et al., “Improved Numerical Dissipation for Time Integration Algorithms in Structural
Dynamics,” Earthquake Engineering & Structural Dynamics, No. 5, 1977, pp. 283  292.
Sharpe, R. D., and A. J. Carr, "The Seismic Response of Inelastic Structures," Research Report
7413, Department of Civil Engineering, University of Canterbury, New Zealand, November
1974.
Wilson, E. L., "A Computer Program for the Dynamic Stress Analysis of Underground Structures,"
SESM Report 681, Department of Civil Engineering, University of California, Berkeley, 1968.
14
6.6 RungeKuttaGill Method (Fourth Order)
A procedure was developed by Gill for the numerical integration. For a differential equation of the
first order,
) , ( t x F
dt
dx
=
the initial value problem is solved in a form
6
2 2
3 2 1 0
1
f f f f
x x
i i
+ + +
+ =
+
where,
) , (
)
2
,
2
(
)
2
,
2
(
) , (
2 3
1
2
0
1
0
t t f x tF f
t
t
f
x tF f
t
t
f
x tF f
t x tF f
i i
i i
i i
i i
∆ + + ∆ =
∆
+ + ∆ =
∆
+ + ∆ =
∆ =
This integration scheme is extended for the ordinary differential equation of the second order;
}) ]{ [ } ]{ [ )} ( ({ ] [ )} , , ( {
} { )} , , ( {
1
x K x C t p M t x x G
x t x x F
− − =
=
−
& &
& &
then,
6
} { } { 2 } { 2 } {
} { } {
6
} { } { 2 } { 2 } {
} { } {
3 2 1 0
1
3 2 1 0
1
g g g g
x x
f f f f
x x
i i
i i
+ + +
+ =
+ + +
+ =
+
+
& &
where,
}) { } ({ )} , , ( { } {
)
2
} {
} ({ )}
2
,
2
,
2
( { } {
)
2
} {
} ({ )}
2
,
2
,
2
( { } {
} { )} , , ( { } {
2 2 2 3
1 1 1
2
0 0 0
1
0
g x t t t g x f x F t f
g
x t
t
t
g
x
f
x F t f
g
x t
t
t
g
x
f
x F t f
x t t x x F t f
i i i i
i i i i
i i i i
i i i i
+ ∆ = ∆ + + + ∆ =
+ ∆ =
∆
+ + + ∆ =
+ ∆ =
∆
+ + + ∆ =
∆ = ∆ =
& &
& &
& &
& &
) } ]{ [ } ]{ [ } ({ ] [
)} , , ( { ) {
) }
2
]{ [ }
2
]{ [ }
2
( ({ ] [
)}
2
,
2
,
2
( { } {
) }
2
]{ [ }
2
]{ [ }
2
( ({ ] [
)}
2
,
2
,
2
( { } {
) } ]{ [ } ]{ [ } ({ ] [
)} , , ( { } {
2 2 1
1
2 2 3
1 1 1
1 1
2
0 0 1
0 0
1
1
0
i i i
i i i
i i i
i i i
i i i
i i i
i i i
i i i
f x K g x C p M t
t t g x f x G t g
f
x K
g
x C
t
t p M t
t
t
g
x
f
x G t g
f
x K
g
x C
t
t p M t
t
t
g
x
f
x G t g
x K x C p M t
t x x G t g
+ − + − ∆ =
∆ + + + ∆ =
+ − + −
∆
+ ∆ =
∆
+ + + ∆ =
+ − + −
∆
+ ∆ =
∆
+ + + ∆ =
− − ∆ =
∆ =
+
−
−
−
−
&
&
&
&
&
&
&
&
15
Reference:
Gill, S., "A Procedure for the Stepbystep Integration of Differential Equations in an Automatic
Computing Machine," Proceedings, Cambridge Philosophical Society, 49:96, 1951.
16
Appendix: Linearly elastic response of SDF System under earthquake motion (FORTRAN）
C
C RESPONSE OF A SINGLEDEGREEOFFREEDOM LINEARLY ELASTIC SYSTEM
C TO EARTHQUAKE MOTION
C
C PROGRAMMED BY OTANI, S.
C ON OCTOBER 28, 1975
C AT UNIVERSITY OF TORONTO
C MODIFIED BY OTANI, S.
C ON APRIL 18, 1993
C AT UNIVERSITY OF TOKYO
C MODIFIED BY OTANI, S.
C ON MAY 26, 2000
C AT UNIVERSITY OF TOKYO
C
C THE EQUATION OF MOTION OF SDF SYSTEM UNDER GROUND MOTION
C
C DDX + 2.0 * BT * WN * DX + WN * WN * X =  DDY
C
C INPUT DATA
C 1. TITL
C TITL: TITLE OF STUDY
C 2. LUNT,TUNT
C LUNT: UNIT OF LENGTH
C TUNT: UNIT OF TIME
C 3. TN,BT,DT,ASCL,V0,D0
C TN: PERIOD
C BT: DAMPING RATIO
C DT: TIME INTERVAL OF NUMERICAL INTEGRATION
C ASC: SCALE FACTOR FOR GROUND ACCELERATION
C V0: INITIAL VELOCITY
C D0: INITIAL DISPLACEMENT
C 4. NSTT,NSTP,MTHD
C NSTT: STARTING POINT OF EARTHQUAKE DATA
C NSTP: TERMINATING POINT OF EARTHQUAKE DATA
C MTHD: POINTER FOR NUMBERICAL INTEGRATION METHOD
C 5. EQNM
C EQNM: NAME OF EARTHQUAKE RECORD
C 6. TIME,GACC
C TIME: TIME COORDINATE OF EARTHQUAKE DATA
C GACC: GROUND ACCELERATION OF EARTHQUAKE DATA
C
CHARACTER*4 TITLE(18)*4,EQNM(18)*4,LUNT*4,TUNT*4
DIMENSION TIME(2),GACC(2)
C
READ (5,500) TITL
READ (5,500) LUNT,TUNT
C
READ (5,*) TN,BT,DT,ASCL,V0,D0
C TN UNDAMPED NATURAL PERIOD
IF (TN.LE.0.0) GO TO 60
C BT DAMPING FACTOR
IF (BT.LT.0.0) BT=0.0
C DT TIME INCREMENT OF NUMERICAL INTEGRATION
IF (DT.LE.0.0) DT=0.05*TN
C ASCL SCALE FACTOR FOR GROUND ACCELERATION DATA
IF (ASCL.LE.0.0) ASCL=1.0
C FN UNDAMPED NATURAL FREQUENCY
FN=1.0/TN
WN=6.2830*FN
BW=2.0*BT*WN
W2=WN*WN
C PRINT INPUT INFORMATION
WRITE (6,600) TITL,LUNT,TUNT
WRITE (6,602) TN,TUNT,FN,TUNT,WN,TUNT,BT,DT,TUNT,ASCL,
1 V0,LUNT,TUNT,D0,LUNT
17
C
READ (5,*) NSTT,NSTP,MTHD
IF (NSTT.LE.0) NSTT=1
IF (NSTP.LE.NSTT) NSTP=NSTT
IF (MTHD.LE.0) MTHD=1
IF (MTHD.GT.4) MTHD=1
WRITE (6,604) NSTT,NSTP
C
C CALCULATION OF NUMERICAL INTEGRATION CONSTANTS
C
GO TO (1,2,3,4),MTHD
1 CALL JNNGS1 (DT,WN,BT,A1,A2,A3,A4,B1,B2,B3,B4,&9)
2 CALL LNACC1 (DT,WN,BT,A1,A2,A3,A4,B1,B2,B3,&9)
3 CALL NWMRK1 (DT,WN,BT,A1,A2,A3,A4,&9)
4 CALL RKGIL1 (DT,A1,&9)
C
C EARTHQUAKE RECORD
C
9 READ (5,500) EQNM
WRITE (6,606) EQNM
C
C INITIAL CONDITIONS
C
ISTP=0
IST=0
NST=0
GACC(2)=0.0
TIME(2)=0.0
TT=0.0
C
10 TIME(1)=TIME(2)
GACC(1)=GACC(2)
READ (5,*) TIME(2),GACC(2)
NST=NST+1
IF (NST.GT.NSTT) GO TO 20
IF (NST.LT.NSTT) GO TO 10
GA=GACC(2)*ASCL
A0=(GA+BT*V0+W2*D0)
TIM=TIME(2)
CALL PRINT (IST,TIM,DT,A0,V0,D0,GA,LUNT,TUNT)
GO TO 10
C
20 IF (NST.GT.NSTP) GO TO 60
DLT=TIME(2)TIME(1)
IF (DLT) 60,10,30
C
30 GR=GACC(1)*ASCL
SLP=(GACC(2)GACC(1))*ASCL/DLT
TI=TT
TT=TT+DLT
C
38 IF (DT.GT.TT) GO TO 10
TT=TTDT
TI=TI+DT
G0=GA
GA=GR+SLP*TI
C
C NUMERICAL INTEGRATION
C PREVIOUS RESPONSE VALUE (A0,V0,D0) MUST BE REPLACED BY
C NEW RESPONSE VALUES (A0,V0,D0)
C A0: ACCELERATION RELATIVE TO GROUND
C V0: VELOCITY RELATIVE TO GROUND
C D0: DISPLACEMENT RELATIVE TO GROUND
C GA: GROUND ACCELERATION
C
40 GO TO (41,42,43,44),MTHD
41 CALL JNNGS2 (A0,V0,D0,G0,GA,W2,BW,A1,A2,A3,A4,B1,B2,B3,B4,&49)
18
42 CALL LNACC2 (A0,V0,D0,GA,DT,A1,A2,A3,A4,B1,B2,B3,&49)
43 CALL NWMRK2 (A0,V0,D0,GA,DT,W2,BW,A1,A2,A3,A4,&49)
44 CALL RKGIL2 (A0,V0,D0,G0,GA,DT,W2,BW,A1,&49)
C
49 IST=IST+1
CALL PRINT (IST,TIM,DT,A0,V0,D0,GA,LUNT,TUNT)
GO TO 38
C END OF RESPONSE CALCULATION
60 WRITE (6,610)
STOP
500 FORMAT (18A4)
600 FORMAT (18A4,/,
1 "UNITS USED IN CALCULATION",/,
2 5X,"LENGTH = ",A4,/,
4 5X,"TIME = ",A4,/)
602 FORMAT ("SYSTEM PROPERTIES",/,
1 5X,"UNDAMPED PERIOD = ",F10.3," (",A4,")",/,
2 5X,"UNDAMPED FREQENCY = ",F10.3," (1/",A4,")",/,
3 5X,"UNDAMPED CIRCULAR FREQ. = ",F10.3," (rad/",A4,")",/,
4 5X,"DAMPING FACTOR = ",F10.3,/,
5 "CONSTANTS FOR NUMERICAL INTEGRATION",/,
6 5X,"TIME INTERVAL = ",F10.3," (",A4,")",/,
7 5X,"ACCELERATION SCALE FACTOR = ",F10.3,/,
8 "INITIAL CONDITIONS",/,
9 5X,"VELOCITY = ",1PE10.3," (",A4,"/",A4,")",/,
1 5X,"DISPLACEMENT = ",E10.3," (",A4,")",/)
604 FORMAT ("START COMPUTATION = ",I5," POINTS",/,
1 "END COMPUTATION = ",I5," POINTS",/)
606 FORMAT ("EARTHQUAKE RECORD ",/,
1 5X,"NAME: ",18A4,/)
610 FORMAT ("ALL DATA PROCESSED")
END
SUBROUTINE PRINT (IST,TIM,DT,A0,V0,D0,GA,LUNT,TUNT)
C
C PRINT RESPONSE VALUES
C
C IST: STEP NUMBER
C TIM: TIME AT FIRST RESPONSE CALCULATION
C DT: TIME STEP
C A0: RESPONSE ACCELERATION RELATIVE TO GROUND
C V0: RESPONSE VELOCITY RELATIVE TO GROUND
C D0: RESPONSE DISPLACEMENT RELATIVE TO GROUND
C GA: GROUND ACCELERATION
C
IF (IST.GT.0) GO TO 10
C FORMAT FOR PRINTING RESPONSE
WRITE (6,600) TUNT,LUNT,TUNT,LUNT,TUNT,
1 LUNT,TUNT,LUNT,LUNT,TUNT
10 T=TIM+DT*FLOAT(IST)
C ABSOLUTE ACCELERATION RESPONSE
A=A0+GA
WRITE (6,602) IST,T,GA,A0,V0,D0,A
600 FORMAT (" STEP"," TIME "," GR. ACC. "," REL. ACC. ",
1 " VELOCITY ","DISPLACEMENT"," ABS. ACC. ",/,
2 3X," (",A4,") ","(",A4,"/",A4,"**2)",
3 "(",A4,"/",A4,"**2)","(",A4,"/",A4,")",
4 " (",A4,") ","(",A4,"/",A4,"**2)",/)
602 FORMAT (I5,F8.3,5(2X,1PE10.3))
RETURN
END
SUBROUTINE JNNGS1 (DT,WN,BT,A1,A2,A3,A4,B1,B2,B3,B4,*)
C
C CALCULATION OF NUMERICAL CONSTANTS IN JENNINGSNIGAM METHOD
C
C INPUT
19
C DT TIME INTERVAL OF NUMERICAL INTEGRATION
C WN UNDAMPED CIRCULAR FREQUENCY
C BT DAMPING FACTOR
C
C FUNCTIONS USED IN THE SUBROUTINE
C SQRT
C EXP
C SIN
C COS
C
DW=WN*DT
EX=EXP(BT*DW)
SB=SQRT(1.0BT*BT)
BS=BT/SB
WD=WN*SB
SW=SIN(WD*DT)
CW=COS(WD*DT)
B2=2.0*BT*BT1.0
W2=1.0/WN**2
TB=B2/SB
WT=W2/DW
A1=EX*(BS*SW+CW)
A2=EX*SW/WD
A3=(EX*(TB*SW+2.0*BT*CW)BTBT)*WT+W2
A4=EX*((BS+TB/DW)*SW+(1.0+2.0*BT/DW)*CW)*W2+2.0*BT*WT
B1=EX*WN*SW/SB
B2= EX*(CWBS*SW)
B3=(EX*(CW+BS*SW)+1.0)*W2/DT
B4= EX*SW/WD+(EX*(BS*SW+CW)1.0)*W2/DT
WRITE (6,600)
RETURN 1
600 FORMAT ("JENNINGSNIGAM METHOD",/)
END
SUBROUTINE LNACC1 (DT,WN,BT,A1,A2,A3,A4,B1,B2,B3,*)
C
C CALCULATION OF NUMERICAL CONSTANTS FOR LINEAR ACCELERATION METHOD.
C
C INPUT
C DT TIME INTEVAL OF NUMERICAL INTEGRATION
C WN UNDAMPED CIRCULAR FREQUENCY
C BT DAMPING FACTOR
C
B1=0.5*DT
B2=DT*DT/3.0
B3=B2*0.5
A=1.0+BT*WN*DT+WN*WN*B3
A1=1.0/A
A2=WN*DT*(BT+WN*DT/3.0)/A
A3=WN*(BT+BT+WN*DT)/A
A4=WN*WN/A
WRITE (6,600)
RETURN 1
600 FORMAT ("LINEAR ACCELERATION METHOD",/)
END
SUBROUTINE NWMRK1 (DT,WN,BT,A1,A2,A3,A4,*)
C
C CALCULATION OF NUMERICAL CONSTANTS FOR NEWMARK BETA METHOD
C
C NIPUT
C DT TIME INTERVAL OF NUMERICAL INTEGRATION
C WN UNDAMPED CIRCULAR FREQUENCY
C BT DAMPING FACTOR
C
BETA=1.0/6.0
C
C BETA CONSTANT FOR NEWMARK BETA METHOD
C
20
A1=0.5*DT
A2=(0.5BETA)*DT**2
A3=BETA*DT**2
A4=0.0005
WRITE (6,600) BETA, A4
RETURN 1
600 FORMAT ("NEWMARK BETA METHOD", /,
1 5X,"BETA = ",F10.3, /,
2 5X,"ERROR LIMIT = ",1PE10.3,/)
END
SUBROUTINE RKGIL1 (DT,A1,*)
C
C CALCULATION OF A NUMERICAL CONSTANT FOR RUNGEKUTTAGILL METHOD.
C
C INPUT
C DT TIME INTERVAL OF NUMERICAL INTEGRATION
C
A1=0.5*DT
WRITE (6,600)
RETURN 1
600 FORMAT ("RUNGEKUTTAGILL MEHOTD",/)
END
SUBROUTINE JNNGS2 (A0,V0,D0,G0,GA,W2,BW,A1,A2,A3,A4,B1,B2,B3,B4,*)
C
C JENNNINGSNIGAM METHOD OF NUMERICAL INTEGRATION
C
C INPUT
C A0 RESPONSE ACCLERATION AT PREVIOUS TIME STEP
C V0 RESPONSE VELOCITY AT PREVIOUS TIME STEP
C D0 RESPONSE DISPLACEMENT AT PREVIOUS TIME STEP
C G0 GROUND ACCELERATION AT PREVIOUS TIME STEP
C GA GROUND ACCELERATION AT CURRENT TIME STEP
C W2 = WN * WN
C BW = 2.0 * BT * WN
C A1  A4 CONSTANTS FOR NUMERICAL INTEGRATION
C B1  B4 CONSTANTS FOR NUMERICAL INTEGRATION
C
C OUTPUT
C A0 RESPONSE ACCELERATION AT CURRENT TIME STEP
C V0 RESPONSE VELOCITY AT CURRENT TIME STEP
C D0 RESPONSE DISPLACEMENT AT CURRENT TIME STEP
C
D=A1*D0+A2*V0A3*GAA4*G0
V0=B1*D0+B2*V0B3*GAB4*G0
A0=GAW2*DBW*V0
D0=D
RETURN 1
END
SUBROUTINE LNACC2 (A0,V0,D0,GA,DT,A1,A2,A3,A4,B1,B2,B3,*)
C
C LINEAR ACCELERATION METHOD OF NUMERICAL INTEGRATION
C
C INPPUT
C A0 RESPONSE ACCLERATION AT PREVIOUS TIME STEP
C V0 RESPONSE VELOCITY AT PREVIOUS TIME STEP
C D0 RESPONSE DISPLACEMENT AT PREVIOUS TIME STEP
C GA GROUND ACCEERATION AT CURRENT TIME STEP
C DT TIME INTERVAL OF NUMERICAL INTEGRATION
C A1  A4 CONSTANTS FOR NUMERICAL INTEGRATION
C B1  B3 CONSTANTS FOR NUMERICAL INTEGRATION
C
C OUTPUT
C A0 RESPONSE ACCELERATION AT CURRENT TIME STEP
C V0 RESPONSE VELOCITY AT CURRENT TIME STEP
C D0 RESPONSE DISPLACEMENT AT CURRENT TIME STEP
C
A=A1*GA+A2*A0+A3*V0+A4*D0
21
D0=D0+V0*DT+A0*B2+A*B3
V0=V0+(A0+A)*B1
A0=A
RETURN 1
END
SUBROUTINE NWMRK2 (A0,V0,D0,GA,DT,W2,BW,A1,A2,A3,A4,*)
C
C NEWMARK BETA METHOD OF NUMERICAL INTEGRATION
C
C INPPUT
C A0 RESPONSE ACCLERATION AT PREVIOUS TIME STEP
C V0 RESPONSE VELOCITY AT PREVIOUS TIME STEP
C D0 RESPONSE DISPLACEMENT AT PREVIOUS TIME STEP
C GA GROUND ACCELERATION AT CURRENT TIME STEP
C DT TIME INTERVAL OF NUMERICAL INTEGRATION
C W2 = WN * WN
C BW = 2.0 * BT * WN
C A1 = 0.5 * DT
C A2 = (0.5  BETA) *DT **2
C A3 = BETA * DT **2
C
C OUTPUT
C A0 RESPONSE ACCELERATION AT CURRENT TIME STEP
C V0 RESPONSE VELOCITY AT CURRENT TIME STEP
C D0 RESPONSE DISPLACEMENT AT CURRENT TIME STEP
C
B=V0+A1*A0
C=D0+V0*DT+A0*A2
10 A=A0
V0=B+A1*A
D0=C+A3*A
A0=GABW*V0W2*D0
ER=ABS(A0A)/(ABS(A0)+ABS(A))
IF (ER.GT.A4) GO TO 10
RETURN 1
END
SUBROUTINE RKGIL2 (A0,V0,D0,G0,GA,DT,W2,BW,A1,*)
C
C RUNGEKUTTAGILL METHOD OF NUMERICAL INTEGRATION
C
C INPPUT
C A0 RESPONSE ACCLERATION AT PREVIOUS TIME STEP
C V0 RESPONSE VELOCITY AT PREVIOUS TIME STEP
C D0 RESPONSE DISPLACEMENT AT PREVIOUS TIME STEP
C G0 GROUND ACCELERATION AT PREVIOUS TIME STEP
C GA GROUND ACCELERATION AT CURRENT TIME STEP
C DT TIME INTERVAL OF NUMERICAL INTEGRATION
C W2 = WN * WN
C BW = 2.0 * BT * WN
C A1 CONSTANT FOR NUMERICAL INTEGRATION
C
C OUTPUT
C A0 RESPONSE ACCELERATION AT CURRENT TIME STEP
C V0 RESPONSE VELOCITY AT CURRENT TIME STEP
C D0 RESPONSE DISPLACEMENT AT CURRENT TIME STEP
C
F0=DT*V0
G0=DT*A0
F1=F0+A1*G0
G1=G0A1*(G0+GA+BW*G0+W2*F0)
F2=F0+A1*G1
G2=G0A1*(G0+GA+BW*G1+W2*F1)
F3=F0+DT*G2
G3=G0DT*(G0+GA+BW*G2+W2*F2)
D0=D0+(F0+F1+F1+F2+F2+F3)/6.0
V0=V0+(G0+G1+G1+G2+G2+G3)/6.0
A0=GABW*V0W2*D0
22
RETURN 1
END
1
Chapter 7. Matrix Analysis of Linearly Elastic Plane Frame
A structural analysis method for plane (twodimensional) frames is presented in this chapter. The
efficiency of the analysis is not intended here, but the basic principle of the analysis should be
clearly understood. The procedure can be easily extended to a threedimensional frame analysis or
a finite element analysis as long as the degree of freedom is increased for nodal displacements and
forces of each structural unit. The elastic behavior of members is considered, as an example, in the
formulation of member stiffness matrix. The procedure can be also used in the nonlinear earthquake
response analysis of a structure because the stiffness of members is assumed to be linearly elastic
over a short time increment in the analysis.
7.1 Assumptions
The following assumptions are made in the structural analysis of a momentresisting plane frame
structure;
(a) All members and loads lie in the vertical plane of a frame (Plane Frame),
(b) Each member is prismatic and straight,
(c) Small displacement,
(d) Each element is rigidly connected at a joint. Member end displacement is equal to the
displacement at the connecting joint (Continuity Condition).
(e) Members behave within linearly elastic region,
(f) Axial and flexural deformation is considered for each member,
(g) External forces act at joints,
(h) External forces acting at a joint are equal to the sum of member end forces of members
connected to the joint (Equilibrium of Forces).
Each member is represented by a straightline element passing through the geometrical centroid
of the section of a member. Axial and flexural elastic deformations are considered for each member.
Shear deformation is neglected for simplicity, but shear deformation can be included in the analysis
by introducing shear deformation in the member stiffness matrix. Flexible connection can be
considered by introducing rotational or shear springs at member ends.
Under a small displacement, the equilibrium of forces can be expressed using the coordinate
system for predeformed configuration.
Internal bending moment should be calculated at the geometrical centroid of the section. The
axial deformation is zero under bending at the geometrical centroid because the neutral axis under
bending passes through the geometrical centroid when the member is in the linearly elastic state.
Therefore, the axial deformation at the geometrical centroid is null under bending.
However, the neutral axis shifts from the
geometric centroid after the stress in section exceeds
the proportional limit of materials. Axial strain at the
geometric centroid develops elongation of the
member even under pure bending. The interaction of
axial and bending deformation should be considered
in the analysis.
The forces acting within a member should be
replaced by the equivalent member end actions
(forces equal to fixed end forces, but having opposite
sign). The effect of forces acting in the member will
be considered after the member end forces are
determined in the structural analysis.
2
1
m
z1 m
z2
p
y1 p
y2
Fixed End Forces
m
z2
p
y1
p
y2
m
z1 Equivalent Joint Loads
2
7.2 Member Stiffness Matrix in Local Coordinates
A set of local (member) coordinates of a member are selected in the plane of a frame. The xaxis
is defined in the direction of the member, and yaxis perpendicular to the member following the
righthand rule; i.e., rotation is positive in the counterclockwise direction.
Coordinate System: Member direction is defined for each member from start end 1 to terminal end
2. Member end actions { } p
i 1
and { } p
i 2
at start end 1 and terminal end 2 of member i are
{ } p
p
p
m
i
x
y
z
i
1
1
1
1
=
¦
´
¦
¹
¦
¹
`
¦
)
¦
{ } p
p
p
m
i
x
y
z
i
2
2
2
2
=
¦
´
¦
¹
¦
¹
`
¦
)
¦
For each member, member end displacements { } d
i 1
and { } d
i 2
at start end 1 and terminal end
2 of member i, are
{ } d
d
d
i
x
y
z
i
1
1
1
1
=
¦
´
¦
¹
¦
¹
`
¦
)
¦
θ
{ } d
d
d
i
x
y
z
i
2
2
2
2
=
¦
´
¦
¹
¦
¹
`
¦
)
¦
θ
Member stiffness relation: The relation of member end actions and member end displacement
through member stiffness submatrices.
{ } [ ] { } [ ] { } p k d k d
i i i i i 1 11 1 12 2
= +
{ } [ ] { } [ ] { } p k d k d
i i i i i 2 21 1 22 2
= +
For a linearly elastic member, stiffness submatrices are given below:
y
x p
x1
p
x2
p
y1
p
y2
m
z1
m
z2
1
2
d
y1
d
x1
d
x2
p
y2
θ
z1
θ
z2 y
x
1
2
3
11 3 2
2
0 0
12 6
[ ] 0
6 4
0
i
i
EA
L
EI EI
k
L L
EI EI
L L
=
[ ] [ ] k k
EA
L
EI
L
EI
L
EI
L
EI
L
i i
T
i
12 21
3 2
2
0 0
0
12 6
0
6 2
= =
−
−
−
[ ] k
EA
L
EI
L
EI
L
EI
L
EI
L
i
i
22
3 2
2
0 0
0
12 6
0
6 4
= −
−
in which E : Young's modulus of material, A: cross sectional area, I : moment of inertia of section,
and L : member length.
1
1.0
x
d =
EA
L
EA
L
−
1
1.0
y
d =
3
12EI
L
3
12EI
L
−
2
6EI
L
2
6EI
L
1
1.0
z
θ =
1 2
, EA EI
L
2
6EI
L
2
6EI
L
−
4EI
L
2EI
L
4
7.3 Coordinate Transformation
Global coordinates ( , , X Y Z ) are defined in the plane of a frame. The Xaxis is defined in the
horizontal direction, the Yaxis in the upward vertical direction, and the rotation Θ in the
counterclockwise using the righthand rule.
Suppose the member axis x of
member i is inclined from the horizontal
global axis X by angle α
counterclockwise. Forces { } p
i
in a
member coordinate system may be
transformed into forces { } P
i
in the
global coordinate system considering the
equilibrium of forces in each direction.
cos sin
sin cos
X x y
Y x y
Z z
P p p
P p p
M m
α α
α α
= −
= +
=
or
cos sin
sin cos
x X Y
y X Y
z Z
p P P
p P P
m M
α α
α α
= +
= − +
=
These relations can be expressed in the matrix form,
i
z
y
x
i i
Z
Y
X
m
p
p
M
P
P
¦
)
¦
`
¹
¦
¹
¦
´
¦
−
=
¦
)
¦
`
¹
¦
¹
¦
´
¦
1 0 0
0 cos sin
0 sin cos
α α
α α
cos sin 0
sin cos 0
0 0 1
x X
y Y
z Z
i i i
p P
p P
m M
α α
α α
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
= −
´ ` ´ `
¦ ¦ ¦ ¦
¹ ) ¹ )
The relations of forces in the global and local coordinates can be expressed in a symbolic matrix
form using a transformation matrix  
i
T ;
{ } [ ] { } P T p
i i i
=
i i i
P T p } { ] [ } {
1 −
=
Note the following property of coordinate transformation matrix [ ] T
i
of member i:
1
[ ] [ ]
T
i i
T T
−
=
Similarly, displacement
i
D} { in the global coordinate system may be transformed into
displacement
i
d} { in a member coordinate system;
i
z
y
x
i i
Z
Y
X
d
d
D
D
¦
)
¦
`
¹
¦
¹
¦
´
¦
−
=
¦
)
¦
`
¹
¦
¹
¦
´
¦
Θ θ
α α
α α
1 0 0
0 cos sin
0 sin cos
X
Y
x
y
α
θ
α α
α α
= Θ
+ =
− =
cos sin
sin cos
y x Y
y x X
5
i
Z
Y
X
i i
z
y
x
D
D
d
d
¦
)
¦
`
¹
¦
¹
¦
´
¦
Θ
− =
¦
)
¦
`
¹
¦
¹
¦
´
¦
1 0 0
0 cos sin
0 sin cos
α α
α α
θ
or
i i i
d T D } { ] [ } { =
{ } [ ] { } d T D
i i
T
i
=
where, matrix [ ]
T
i
T is the transpose of matrix [ ]
i
T .
6
7.4 Member Stiffness Matrix in Global Coordinates
Member end actions { } P
i 1
at start end of
member i in the global coordinate system are
expressed by those in the local coordinate system
by the coordinate transformation;
{ } [ ] { } P T p
i i i 1 1
=
Using the member stiffness relation,
i i i i i
d k d k p } { ] [ } { ] [ } {
2 12 1 11 1
+ =
The member end actions { }
i
P in the global
coordinate system can be expressed by member
end displacements { }
i
d in the member
coordinate system:
) } { ] [ } { ] ([ ] [ } { ] [ } {
) } { ] [ } { ] ([ ] [ } { ] [ } {
2 22 1 21 2 2
2 12 1 11 1 1
i i i i i i i i
i i i i i i i i
d k d k T p T P
d k d k T p T P
+ = =
+ = =
Transformation of displacement coordinates from the global coordinate system to the local
coordinate system;
{ } [ ] { } d T D
i i
T
i 1 1
=
{ } [ ] { } d T D
i i
T
i 2 2
=
The member end actions { } P
i 1
at the starting end of member i are expressed in terms of
member end displacements { } D
i 1
and { } D
i 2
both in the global coordinate system:
i i i i
i
T
i i i i
T
i i i
i
T
i i i
T
i i i
i i i i i i
D K D K
D T k T D T k T
D T k D T k T
d k d k T P
} { ] [ } { ] [
} { ] [ ] [ ] [ } { ] [ ] [ ] [
) } { ] [ ] [ } { ] [ ] ([ ] [
) } { ] [ } { ] ([ ] [ } {
2 12 1 11
2 12 1 11
2 12 1 11
2 12 1 11 1
+ =
+ =
+ =
+ =
where,
T
i i i i
T
i i i i
T k T K
T k T K
] [ ] [ ] [ ] [
] [ ] [ ] [ ] [
12 12
11 11
=
=
Similarly, the member end action {P
2
}
i
at the terminal end can be expressed in the global
coordinate system;
i i i i
i
T
i i i i
T
i i i
i
T
i i i
T
i i i
i i i i i i
D K D K
D T k T D T k T
D T k D T k T
d k d k T P
} { ] [ } { ] [
} { ] [ ] [ ] [ } { ] [ ] [ ] [
) } { ] [ ] [ } { ] [ ] ([ ] [
) } { ] [ } { ] ([ ] [ } {
2 22 1 21
2 22 1 21
2 22 1 21
2 22 1 21 2
+ =
+ =
+ =
+ =
where,
T
i i i i
T
i i i i
T k T K
T k T K
] [ ] [ ] [ ] [
] [ ] [ ] [ ] [
22 22
21 21
=
=
Therefore, member end actions
i
P} { , in
the global coordinates, of member i can be
expressed by member end displacements
X
Y
x
y
{p
1
}
i
{P
1
}
i
1
2
D
y1
D
x1
D
x2
D
y2
Θ
z1
Θ
z2
Y
X
1
2
Member End Displacements in Global Coordinates
Joint J Joint K Starting End 1
Terminal End 2
Member i
7
i
D} { using member stiffness matrix
i
K] [ in the global coordinates;
i i i
i i i
D
D
K K
K K
P
P
D K P
)
`
¹
¹
´
¦
)
`
¹
¹
´
¦
=
)
`
¹
¹
´
¦
=
2
1
22 21
12 11
2
1
} { ] [ } {
where,
2 , 1 , ] ][ [ ] [ ] [ = = k j for T k T K
T
i jk i jk
8
7.5 Continuity of Displacement at Joint
Suppose member i is rigidly
connected to joints J and K of the
structure; starting end 1 of member i is
connected to joint J and terminal end 2
of member i to joint K.
By the continuity conditions of
displacement at joints, member end
displacements and joint displacements
are equal as long as they are expressed
in the common coordinate system:
K i
J i
D D
D D
} { } {
} { } {
2
1
=
=
where
i
D } {
1
: displacement at start end 1 of member i, connected to joint J of the structure,
expressed in the global coordinate system,
i
D } {
2
: displacement at terminal end 2 of member i,
connected to joint K of the structure, expressed in the global coordinate system,
J
D} { :
displacement at joint J of the structure expressed in the global coordinate system,
K
D} { :
displacement at joint K of the structure expressed in the global coordinate system.
Joint J Joint K Starting End 1
Terminal End 2
Member i
9
7.6 Equilibrium of Forces at Joint
The sum of member end forces of members connected at a joint is equal to the external forces
acting at the joint, both expressed in the global coordinate system.
∑
=
=
n
i
i J J
P P
1
} { } {
where,
J
P} { : external forces acting at joint J of the structure expressed in the global coordinate
system,
i J
P } { : member end forces of member i at a member end connected to joint J of the
structure expressed in the global coordinate system, n: number of members connected to joint J of
the structure. If start end 1 of member i is connected to joint J of the structure,
i i J
P P } { } {
1
= ; if
terminal end 2 is connected to joint J,
i i J
P P } { } {
2
= .
Member 1
Member 2
Member i
Member n
Joint J
J
1
J
2
J
i
J
n
J
i
: End of member i connected to
Joint J of the structure.
10
7.7 Formulation of Structural Stiffness Matrix
Suppose start end 1 of member i is connected to joint J of the structure and terminal end 2 to joint
K
i
of the structure, member end forces
i J
P } { of member i at joint J are expressed in the global
coordinate system, as
i i i i
i i J
D K D K
P P
} { ] [ } { ] [
} { } {
2 12 1 11
1
+ =
=
By the continuity condition of member end displacements and structural joint displacements at a
joint in the global coordinate system,
i
K i
J i
D D
D D
} { } {
} { } {
2
1
=
=
Therefore,
i
K i J i i J
D K D K P } { ] [ } { ] [ } {
12 11
+ =
Similarly, if terminal end 2 of member i is connected to joint J and starting end 1 to joint K
i
,
member end forces
i J
P } { of member i at Joint J are expressed as
2
21 1 22 2
21 22
{ } { }
[ ] { } [ ] { }
[ ] { } [ ] { }
i
J i i
i i i i
i K i J
P P
K D K D
K D K D
=
= +
= +
By the equilibrium condition of member end forces and external forces at joint J, where n
members are connected,
∑
=
=
n
i
i J J
P P
1
} { } {
Suppose m members of start end 1 are connected to joint J of the structure and (nm) members of
terminal end 2 are connected to the same joint,
Member i 1
2
Joint J
Joint K
i
{P
J
}
i
={P
1
}
i
X
Y
Member i 1
2
Joint J
Joint K
i
{P
J
}
i
={P
2
}
i
X
Y
11
∑ ∑ ∑ ∑
∑ ∑
+ = + = = =
+ = =
+ + + =
+ + + =
n
m i
J i K
n
m i
i
m
i
m
i
K i J i
J i K
n
m i
i
m
i
K i J i J
D K D K D K D K
D K D K D K D K P
i i
i i
1
22
1
21
1 1
12 11
22
1
21
1
12 11
} { ] [ } { ] [ } { ] [ } { ] [
} { ] [ } { ] [ } { ] [ } { ] [ } {
If the equilibrium of forces is written for every joint including the support joint, the external forces
} {P at joints are expressed as a linear function of joint displacement } {D ;
} ]{ [ } { D K P =
in which [K] is called a structural stiffness matrix.
Normally, the structural stiffness matrix [K] is formulated member by member rather than joint by
joint. Suppose starting end of member i is connected to joint J and terminal end to joint K, then
member stiffness submatrix
i
K ] [
11
is added at (J, J) location of the structural stiffness matrix [K],
i
K ] [
12
at (J, K) location,
i
K ] [
21
at (K, J) location, and
i
K ] [
22
at (K, K) location. This process of
adding member submatrices to the structural stiffness matrix is repeated for all members.
J K
l l
l l
l l
J (J,J)(J,K)
l l
l l
l l
l l
[K] = l l
l l
l l
l l
l l
l l
K (K,J)(K,K)
l l
l l
l l
i
T
i
i
i
K K K K K K K
K J K K J K K
K K J K K J K
K J J K J J K
] [ )] , ( [ )] , ( [
] [ )] , ( [ )] , ( [
] [ )] , ( [ )] , ( [
] [ )] , ( [ )] , ( [
22
12
12
11
+ =
+ =
+ =
+ =
If the symmetric properties of the structural stiffness matrix are recognized, then only upper
triangle part of the structural stiffness matrix need be formulated.
¦
¦
)
¦
¦
`
¹
¦
¦
¹
¦
¦
´
¦
=
¦
¦
)
¦
¦
`
¹
¦
¦
¹
¦
¦
´
¦
M
M
O M M M
L L
L L
L M M O
M
M
J
I
JJ JI
IJ II
J
I
D
D
K K
K K
P
P
12
7.8 Free Joint Displacements and Support Reactions
Joints are classified into support joints and free joints. Rigidly supported joints cannot displace in
any direction, hence the displacement } {
s
D at the support joint is zero; however, force (reaction)
} {
s
P at a support joint is not known. On the other hand, at free end, displacement } {
f
D is not
known, but force (external load) } {
f
P is given.
known: } {
f
P (given as joint loads)and } {
s
D (=0 for fixed supports)
unknown: } {
f
D and } {
s
P
Therefore, it is desirable to reorganize the structural stiffness relation and separate displacement
and force at free joints and support joints;
)
`
¹
¹
´
¦
=
)
`
¹
¹
´
¦
s
f
ss sf
fs ff
s
f
D
D
K K
K K
P
P
where } 0 { } { =
s
D . The free joint displacement } {
f
D can be solved from the first equation:
} { ] [ } {
1
f ff f
P K D
−
=
and then, the support reaction {P
s
} is solved.
} ]{ [ } {
f sf s
D K P =
Support Joint
Free Joint
13
7.9 Member End Actions
Once joint displacements are all calculated, then member end actions in the local coordinate
system can be calculated using joint displacements and member stiffness matrix.
The member end actions and displacements
relation for member i is
i i i i i
i i i i i
d k d k p
d k d k p
} { ] [ } { ] [ } {
} { ] [ } { ] [ } {
2 22 1 21 2
2 12 1 11 1
+ =
+ =
If starting end 1 of the member is connected
to joint J and terminal end 2 to joint K, using the
continuity condition of displacement at joint an transformation of coordinates, joint end
displacements in local coordinates are expressed
K
T
i i
J
T
i i
D T d
D T d
} { ] [ } {
} { ] [ } {
2
1
=
=
Combining the above two relations;
K
T
i i J
T
i i i
K
T
i i J
T
i i i
D T k D T k p
D T k D T k p
} { ] [ ] [ } { ] [ ] [ } {
} { ] [ ] [ } { ] [ ] [ } {
22 21 2
12 11 1
+ =
+ =
The bending moment distributes linearly along the member because no intermediate loads act in
the member.
If intermediate loads act in the member, the above member end actions should be added to the
member actions of fixedfixed member under the given intermediate loads.
Joint J Joint K Starting End 1
Terminal End 2
Member i
p
y2
p
x2
m
z2
p
y1
p
x1
Member i
Structural Analysis
FEM
1
FEM
2
Stress due to member loading
m
z1
Final Member Stresses
14
7.10 Example: Onestory Onebay Frame
Structure: A onestory onebay frame is analyzed. The stiffness matrix of the structure is formulated
in this example.
The story height is h m and the span is l
m. The base of the columns is fixed. Only
flexural deformation is considered in the beam
and axial and flexural deformations are
considered in the columns. The flexural rigidity
of the beam and the columns are
b
EI and
c
EI , and the axial rigidity of the columns is
c
EA .
Joints are numbered from 1 to 4 starting
from the left beamcolumn joint to the right
column base. Members are numbered from 1 to 3 starting from the beam.
Global Coordinate System: The global coordinate
system is taken as Xaxis in the horizontal direction
and Yaxis in the vertical direction. Displacement
{ , , }
X Y Z
D D Θ and force { , , }
X Y Z
P P M are
defined at each joint in the direction of the global
coordinates. Rotation and moment are positive
counterclockwise.
Member Coordinate System: The xaxis of a member is taken in the direction of the member, and
the yaxis normal to the member axis. The member direction of a beam is from left to right, and that
of a column is downward. Displacement { , , }
x y z
d d θ and force { , , }
x y z
p p m are defined at starting
and terminal ends in the direction of the member coordinates. Rotation and moment are positive
counterclockwise.
The transformation matrices for a column (α =90.0 deg) and a beam (α =0.0 deg) are given as
follows;
EI
b
EI
c
,EA
c
EI
c
,EA
c
h
l
1
2
3
4
X
Y
1
2
3
1
2
3 4
1 X
D
1 Y
D
1 Z
Θ
2 X
D
2 Y
D
2 Z
Θ
3 X
D
3 Y
D
3 Z
Θ
4 X
D
4 Y
D
4 Z
Θ
1 1
,
x x
d p
1 1
,
y y
d p
1 1
,
z z
m θ
2 2
,
x x
d p
2 2
,
y y
d p
2 2
,
z z
m θ
1
2 x
y
1 1
,
y y
d p
1 1
,
x x
d p
1 1
,
z z
m θ
2 2
,
y y
d p
2 2
,
x x
d p
2 2
,
z z
m θ
1
2
y
x
15
0.0 1.0 0.0 1.0 0.0 0.0
[ ] 1.0 0.0 0.0 [ ] 0.0 1.0 0.0
0.0 0.0 1.0 0.0 0.0 1.0
column beam
T T
= − =
Beam Stiffness Matrix in Member Coordinates: For a beam 12, no axial deformation is
considered. Therefore, the stiffness matrix of a beam in the member coordinates is given as below;
1 1 2
1 1 2 3 2 3 2
1 1 2
2 2
2
2
2
0 0 0 0
12 6 12 6
0 0
6 4 6 2
0 0
b b
x x x
b b b b
y y y
z z z
beam beam beam
b b b b
beam beam
x
y
z
EA EA
p d d
EI EI EI EI
p d d
m
EI EI EI EI
p
p
m
θ θ
−
¦ ¹ ¦ ¹ ¦ ¹
¦ ¦ ¦ ¦ ¦ ¦
= + −
´ ` ´ ` ´ `
¦ ¦ ¦ ¦ ¦ ¦
¹ ) ¹ ) ¹ )
−
¦ ¹
¦ ¦
´ `
¦
¹ )
l l
l l l l
l l l l
1 2
1 2 3 2 3 2
1 2
2 2
0 0 0 0
12 6 12 6
0 0
6 2 6 4
0 0
b b
x x
b b b b
y y
z z
beam beam beam
b b b b
beam beam
EA EA
d d
EI EI EI EI
d d
EI EI EI EI
θ θ
−
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
= − − + −
´ ` ´ `
¦ ¦ ¦ ¦ ¦
¹ ) ¹ )
−
l l
l l l l
l l l l
but
1 2 x x
d d = and
1 2 x x
p p = − .
1
2
1 x
p
1 y
p
1 z
m
2 z
θ
1 z
θ
2 x
p
2 z
m
2 y
p
1 y
d
2 y
d
Beam Stiffness Matrix in Global Coordinates: The member coordinates of the beam coincide with
the global coordinates; hence the member stiffness matrix of the beam is expressed in the same
form;
1 1 2
1 1 2 3 2 3 2
1 1 2
2 2
2
2
2
0 0 0 0
12 6 12 6
0 0
6 4 6 2
0 0
b b
X X X
b b b b
Y Y Y
Z Z Z
beam beam beam
b b b b
beam beam
X
Y
Z
EA EA
P D D
EI EI EI EI
P D D
M
EI EI EI EI
P
P
M
−
¦ ¹ ¦ ¹ ¦ ¹
¦ ¦ ¦ ¦ ¦ ¦
= + −
´ ` ´ ` ´ `
¦ ¦ ¦ ¦ ¦ ¦
Θ Θ
¹ ) ¹ ) ¹ )
−
¦ ¹
¦ ¦
´ `
¦
¹ )
l l
l l l l
l l l l
1 2
1 2 3 2 3 2
1 2
2 2
0 0 0 0
12 6 12 6
0 0
6 2 6 4
0 0
b b
X X
b b b b
Y Y
Z Z
beam beam beam
b b b b
beam beam
EA EA
D D
EI EI EI EI
D D
EI EI EI EI
−
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
= − − + −
´ ` ´ `
¦ ¦ ¦ ¦ ¦
Θ Θ
¹ ) ¹ )
−
l l
l l l l
l l l l
but
1 2 X X
D D = and P
X1
=P
X2
.
16
Column Stiffness Matrix in Member Coordinates: The stiffness matrix of a column is given in the
member coordinates as follows;
1 1 2
1 1 2
3 2 3 2
1 1 2
2 2
2
2
2
0 0 0 0
12 6 12 6
0 0
6 4 6 2
0 0
c c
x x x
c c c c
y y y
z z z
column column column
c c c c
x
y
z
c
EA EA
h h
p d d
EI EI EI EI
p d d
h h h h
m
EI EI EI EI
h h h h
p
p
m
θ θ
−
¦ ¹ ¦ ¹ ¦ ¹
¦ ¦ ¦ ¦ ¦ ¦
= + −
´ ` ´ ` ´ `
¦ ¦ ¦ ¦ ¦ ¦
¹ ) ¹ ) ¹ )
−
¦ ¹
¦ ¦
´ `
¦ ¦
¹ )
1 2
1 2
3 2 3 2
1 2
2 2
0 0 0 0
12 6 12 6
0 0
6 2 6 4
0 0
c c
x x
c c c c
y y
z z
olumn column column
c c c c
EA EA
h h
d d
EI EI EI EI
d d
h h h h
EI EI EI EI
h h h h
θ θ
−
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
= − − + −
´ ` ´ `
¦ ¦ ¦ ¦
¹ ) ¹ )
−
1 z
θ
2 z
θ
1 y
u
2 y
u
1
A
2
x
y
1 x
u
2 x
u
Column Stiffness Matrix in Global Coordinates: The member coordinates of a column rotate 90
degrees clockwise. The member stiffness matrix of the beam is expressed as follows;
3 2 3 2
1 1 2
1 1 2
1 1 2
2 2
2
2
2
12 6 12 6
0 0
0 0 0 0
6 4 6 2
0 0
c c c c
X X X
c c
Y Y Y
Z Z Z
column column column
c c c c
X
Y
Z
c
EI EI EI EI
h h h h
P D D
EA EA
P D D
h h
M
EI EI EI EI
h h h h
P
P
M
−
¦ ¹ ¦ ¹ ¦ ¹
¦ ¦ ¦ ¦ ¦ ¦
= + −
´ ` ´ ` ´ `
¦ ¦ ¦ ¦ ¦ ¦
Θ Θ
¹ ) ¹ ) ¹ )
−
¦ ¹
¦ ¦
´ `
¦ ¦
¹ )
3 2 3 2
1 2
1 2
1 2
2 2
12 6 12 6
0 0
0 0 0 0
6 2 6 4
0 0
c c c c
X X
c c
Y Y
Z Z
olumn column column
c c c c
EI EI EI EI
h h h h
D D
EA EA
D D
h h
EI EI EI EI
h h h h
− − −
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
= − +
´ ` ´ `
¦ ¦ ¦ ¦
Θ Θ
¹ ) ¹ )
−
17
Continuity of Displacement at Joint: The continuity of displacement is considered at each joint
using the global coordinates;
Joint 1:
1 1
1 1
1 1
int 1 1 2
X X X
Y Y Y
Z Z Z
Jo Member Member
D D D
D D D
¦ ¹ ¦ ¹ ¦ ¹
¦ ¦ ¦ ¦ ¦ ¦
= =
´ ` ´ ` ´ `
¦ ¦ ¦ ¦ ¦ ¦
Θ Θ Θ
¹ ) ¹ ) ¹ )
Joint 2:
2 1
2 1
2 1
int 2 1 3
X X X
Y Y Y
Z Z Z
Jo Member Member
D D D
D D D
¦ ¹ ¦ ¹ ¦ ¹
¦ ¦ ¦ ¦ ¦ ¦
= =
´ ` ´ ` ´ `
¦ ¦ ¦ ¦ ¦ ¦
Θ Θ Θ
¹ ) ¹ ) ¹ )
Joint 3:
2
2
2
int 3 2
X X
Y Y
Z Z
Jo Member
D D
D D
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
=
´ ` ´ `
¦ ¦ ¦ ¦
Θ Θ
¹ ) ¹ )
Joint 4:
2
2
2
int 4 3
X X
Y Y
Z Z
Jo Member
D D
D D
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
=
´ ` ´ `
¦ ¦ ¦ ¦
Θ Θ
¹ ) ¹ )
Equilibrium of Forces at Joint: The equilibrium of forces is considered at each joint;
Joint 1:
1 1
1 1
1 1
int 1 1 2
X X X
Y Y Y
Z Z Z
Jo Member Member
P P P
P P P
M M M
¦ ¹ ¦ ¹ ¦ ¹
¦ ¦ ¦ ¦ ¦ ¦
= +
´ ` ´ ` ´ `
¦ ¦ ¦ ¦ ¦ ¦
¹ ) ¹ ) ¹ )
Joint 2:
2 1
2 1
2 1
int 2 1 3
X X X
Y Y Y
Z Z Z
Jo Member Member
P P P
P P P
M M M
¦ ¹ ¦ ¹ ¦ ¹
¦ ¦ ¦ ¦ ¦ ¦
= +
´ ` ´ ` ´ `
¦ ¦ ¦ ¦ ¦ ¦
¹ ) ¹ ) ¹ )
Joint 3:
2
2
2
int 3 2
X X
Y Y
Z Z
Jo Member
P P
P P
M M
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
=
´ ` ´ `
¦ ¦ ¦ ¦
¹ ) ¹ )
Joint 4:
2
2
2
int 3 2
X X
Y Y
Z Z
Jo Member
P P
P P
M M
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
=
´ ` ´ `
¦ ¦ ¦ ¦
¹ ) ¹ )
1 2
3 4
1 2
1
2
1
2
1
2
3
1 2
3 4
1 2
1
2
1
2
1
2
3
18
Structural Stiffness: The stiffness of the structure is formulated by using the equilibrium and
continuity conditions at each joint;
Joint 1:
1 1
1 1
1 1
int 1 1 2
3 2 3 2
1
2 2
1 1
0 0 0 0
12 6 12 6
0 0
6 4 6 2
0 0
X X X
Y Y Y
Z Z Z
Jo Member Member
b b
X
b b b b
Y
Z
b b b b
P P P
P P P
M M M
EA EA
D
EI EI EI EI
D
EI EI EI EI
¦ ¹ ¦ ¹ ¦ ¹
¦ ¦ ¦ ¦ ¦ ¦
= +
´ ` ´ ` ´ `
¦ ¦ ¦ ¦ ¦ ¦
¹ ) ¹ ) ¹ )
−
¦ ¹
¦ ¦
= + −
´ `
¦ ¦
Θ
¹ )
−
l l
l l l l
l l l l
2
3 2 3 2
1 3
2 2
2 2
12 6 12 6
0 0
0 0 0 0
6 4 6 2
0 0
X
Y
Z
c c c c
X X
c c
Y Y
Z Z
c c c c
D
D
EI EI EI EI
h h h h
D D
EA EA
D D
h h
EI EI EI EI
h h h h
¦ ¹
¦ ¦
´ `
¦ ¦
Θ
¹ )
−
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
+ + −
´ ` ´ `
¦ ¦ ¦ ¦
Θ Θ
¹ ) ¹ )
−
Joint 2:
2 1
2 1
2 1
int 2 1 3
3 2 3 2
1
2 2
1
0 0 0 0
12 6 12 6
0 0
6 2 6 4
0 0
X X X
Y Y Y
Z Z Z
Jo Member Member
b b
X
b b b b
Y
Z
b b b b
P P P
P P P
M M M
EA EA
D
EI EI EI EI
D
EI EI EI EI
¦ ¹ ¦ ¹ ¦ ¹
¦ ¦ ¦ ¦ ¦ ¦
= +
´ ` ´ ` ´ `
¦ ¦ ¦ ¦ ¦ ¦
¹ ) ¹ ) ¹ )
−
¦ ¹
¦ ¦
= − − + −
´ `
¦ ¦
Θ
¹ )
−
l l
l l l l
l l l l
2
1
3 2 3 2
2 4
2 2
3 3
12 6 12 6
0 0
0 0 0 0
6 4 6 2
0 0
X
Y
Z
c c c c
X X
c c
Y Y
Z Z
c c c c
D
D
EI EI EI EI
h h h h
D D
EA EA
D D
h h
EI EI EI EI
h h h h
¦ ¹
¦ ¦
´ `
¦ ¦
Θ
¹ )
−
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
+ + −
´ ` ´ `
¦ ¦ ¦ ¦
Θ Θ
¹ ) ¹ )
−
Joint 3:
19
2
2
2
int 3 2
3 2 3 2
1 2
1 2
1 2
2 2
2 2
12 6 12 6
0 0
0 0 0 0
6 2 6 4
0 0
12
X X
Y Y
Z Z
Jo Member
c c c c
X X
c c
Y Y
Z Z
c c c c
P P
P P
M M
EI EI EI EI
h h h h
D D
EA EA
D D
h h
EI EI EI EI
h h h h
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
=
´ ` ´ `
¦ ¦ ¦ ¦
¹ ) ¹ )
− − −
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
= − +
´ ` ´ `
¦ ¦ ¦ ¦
Θ Θ
¹ ) ¹ )
−
−
=
3 2 3 2
1 3
2 2
6 12 6
0 0
0 0 0 0
6 2 6 4
0 0
c c c c
X X
c c
Y Y
Z Z
c c c c
EI EI EI EI
h h h h
D D
EA EA
D D
h h
EI EI EI EI
h h h h
− −
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
− +
´ ` ´ `
¦ ¦ ¦ ¦
Θ Θ
¹ ) ¹ )
−
Joint 4:
2
2
2
int 4 3
3 2 3 2
1 2
1 2
1 2
3 3
2 2
12 6 12 6
0 0
0 0 0 0
6 2 6 4
0 0
12
X X
Y Y
Z Z
Jo Member
c c c c
X X
c c
Y Y
Z Z
c c c c
P P
P P
M M
EI EI EI EI
h h h h
D D
EA EA
D D
h h
EI EI EI EI
h h h h
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
=
´ ` ´ `
¦ ¦ ¦ ¦
¹ ) ¹ )
− − −
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
= − +
´ ` ´ `
¦ ¦ ¦ ¦
Θ Θ
¹ ) ¹ )
−
−
=
3 2 3 2
2 4
2 2
6 12 6
0 0
0 0 0 0
6 2 6 4
0 0
c c c c
X X
c c
Y Y
Z Z
c c c c
EI EI EI EI
h h h h
D D
EA EA
D D
h h
EI EI EI EI
h h h h
− −
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
− +
´ ` ´ `
¦ ¦ ¦ ¦
Θ Θ
¹ ) ¹ )
−
From the boundary condition of the support;
3
4
0.0
0.0
0.0
0.0
0.0
0.0
X
Y
Z
X
Y
Z
D
D
D
D
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
=
´ ` ´ `
¦ ¦ ¦ ¦
Θ
¹ ) ¹ )
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
=
´ ` ´ `
¦ ¦ ¦ ¦
Θ
¹ ) ¹ )
Therefore, the stiffness matrix of the structure is written as
20
3 2
3 2
int 1 1 1
2 2
1 2
3 2
2
12 6
0 0 0
12 6
0 0 0
6 4 6 4
0 0
0 0
12 6
0
6 2
0
c c b
X X X
b b c
Y Y Y
Z Z Z
Jo
b b c c
b
b b
b b
EI EI EA
h h
P D D
EI EI EA
P D D
h
M
EI EI EI EI
h h
EA
EI EI
EI EI
¦ ¹ ¦ ¹ ¦ ¹
¦ ¦ ¦ ¦ ¦ ¦
= +
´ ` ´ ` ´ `
¦ ¦ ¦ ¦ ¦ ¦
Θ Θ
¹ ) ¹ ) ¹ )
−
+ −
−
l
l l
l l
l
l l
l l
2
1
X
Y
Z
D
D
¦ ¹
¦ ¦
´ `
¦ ¦
Θ
¹ )
Free joint 1
3 2 3 2
int 2 1 2
2 2
1 1
3 2
2
0 0 0 0
12 6 12 6
0 0
6 2 6 4
0 0
12 6
0
0 0
6 4
0
b b
X X X
b b b b
Y Y Y
Z Z Z
Jo
b b b b
c c
c
c c
EA EA
P D D
EI EI EI EI
P D D
M
EI EI EI EI
EI EI
h h
EA
h
EI EI
h h
−
¦ ¹ ¦ ¹ ¦ ¹
¦ ¦ ¦ ¦ ¦ ¦
= − − + −
´ ` ´ ` ´ `
¦ ¦ ¦ ¦ ¦ ¦
Θ Θ
¹ ) ¹ ) ¹ )
−
+
l l
l l l l
l l l l
2
3
X
Y
Z
D
D
¦ ¹
¦ ¦
´ `
¦ ¦
Θ
¹ )
Free joint 2
3 2
int 3 1
2
12 6
0
0 0
6 2
0
c c
X X
c
Y Y
Z Z
Jo
c c
EI EI
h h
P D
EA
P D
h
M
EI EI
h h
− −
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
= −
´ ` ´ `
¦ ¦ ¦ ¦
Θ
¹ ) ¹ )
Support joint 3
3 2
int 4 2
2
12 6
0
0 0
6 2
0
c c
X X
c
Y Y
Z Z
Jo
c c
EI EI
h h
P D
EA
P D
h
M
EI EI
h h
− −
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
= −
´ ` ´ `
¦ ¦ ¦ ¦
Θ
¹ ) ¹ )
Support joint 4
but,
@ int1 @ int 2 X jo X jo
D D = and
2 1 X X
P P − = .
The horizontal joint displacement at joints 1 and 2 are the same because no axial deformation is
considered in the beam;
1 2 1 X atjnt X atjnt X
D D D = = . As we consider the horizontal displacement at
21
joints 1 and 2 simultaneously,
2 1 1 jnt at X jnt at X X
P P P + = .
1 X
P
3
24
h
EI
0
2
6
h
EI
c
0
2
6
h
EI
c
1 X
D
1 jnt at Y
P
0
h
EA EI
c b
+
3
12
l
2
6
l
b
EI
3
12
l
b
EI
−
2
6
l
b
EI
1 jnt at Y
D
1 jnt at Z
M
=
2
6
h
EI
c
2
6
l
b
EI
h
EI EI
c b
4 4
+
l
2
6
l
b
EI
−
l
b
EI 2
1 jnt at Z
M
2 jnt at Y
P
0
3
12
l
b
EI
−
2
6
l
b
EI
−
h
EA EI
c b
+
3
12
l
2
6
l
b
EI
−
2 jnt at Y
D
2 jnt at Z
M
2
6
h
EI
c
2
6
l
b
EI
l
b
EI 2
2
6
l
b
EI
−
h
EI EI
c b
4 4
+
l
2 jnt at Z
M
22
Home Assignment No. 3
20020304
S. Otani
Analyze a twomember structure. Define the member stiffness matrices in local coordinate
system, coordinate transformation matrices, member stiffness matrix in global coordinate system,
and stiffness matrix of the structure. Determine bending moment, shear and axial forces of the two
members. The length, cross sectional area and moment of inertia are common in the two members.
4 10
2 5
2 3
10
10
/ 10 200
00 . 3
mm I
mm A
mm N E
m
=
=
× =
= l
A
B
C
M= 100 kN m
I A E , , , l
I A E , , , l
30 θ =
o
23
Joint displacement (unit: m)
4
0.0124
{ } 0.0214 10
0.1937
B
D
−
¦ ¹
¦ ¦
= ×
´ `
¦ ¦
−
¹ )
Joint reactions (unit: m, N)
5
5
0.2473
{ } 0.1428 10
0.2418
0.2473
{ } 0.1428 10
0.2418
C
A
P
P
¦ ¹
¦ ¦
= − ×
´ `
¦ ¦
−
¹ )
− ¦ ¹
¦ ¦
= ×
´ `
¦ ¦
−
¹ )
1
Chapter 8. Numerical Solution of Linear Equations
The equilibrium of a structure at time t is expressed as the equilibrium of internal force (structural
resistance) { ( )} R t and external force { ( )} F t at structural nodes;
{ ( )} { ( )} F t R t =
In a dynamic problem, the external force may be the sum of externally applied dynamic forces, and
negative values of inertia and damping forces. The internal force is the restoring force of the
structure.
{ ( )} { ( )} [ ]{ ( )} { ( )} F t P t M z t D t = − − &&
where { ( )} P t : external force at time t, [ ] M : mass matrix, normally defined as lumped mass at
each floor corresponding to the horizontal degree of freedom, { ( )} z t && : absolute acceleration at mass
point, { ( )} D t : damping resistance, normally defined at each floor for the horizontal degree of
freedom. Damping force may be assumed to be proportional to velocity { ( )} x t & relative to the base
of the structure, but the damping matrix may be made proportional to instantaneous stiffness in the
horizontal degree of freedom.
8.1 Incremental Formulation
A nonlinear problem is normally solved by reducing the problem into a linear problem over a short
time increment in an incremental approach.
For example, the equilibrium of dynamic forces at time
1 i
t
+
can be written as
 { } { } { } { } 0 ) ( ) ( ) (
1 1 1
= + +
+ + + i i i
t R t D t z M & &
The structural resistance and damping resistance may be written in incremental form;
  { } { } { } { } { } { } { } 0 ) ( ) ( ) ( ) ( ) ) ( ) ( (
1 1 1 1
= ∆ + + ∆ + + +
+ + + + i i i i i i
t R t R t D t D e t y t x M & & & &
where
1
( )
i
y t
+
&& : ground acceleration (given value as an input earthquake motion) at
1 i
t
+
, and { } e :
vector having unit value for all elements.
Separating the known quantities and unknown quantities, the equilibrium of dynamic forces at
time
1 + i
t can be written as
 { } { } { }   { } { } { } ) ( ) ( ) ( ) ( ) ( ) ( ) (
1 1 1 1 1 i i i i i i i
t R t D e t y M t y t R t D t x M − − − = ∆ + ∆ +
+ + + + +
& & & & & &
The increment of internal force
1
{ ( )}
i
R t
+
∆ may be approximated by product of tangent stiffness
[ ]
i
t
K and incremental displacement { ( )} x t ∆ relative to the base of the structure, and the
increment of damping force
1
{ ( )}
i
D t
+
∆ by the product of tangent damping matrix [ ]
i
t
C and
incremental velocity { ( )} x t ∆& .
 { }   { }   { }   { } { } { } ) ( ) ( ) ( ) ( ) ( ) ( ) (
1 1 1 1 1 i i i i i t i t i
t R t D e t y M t y t x K t x C t x M
i i
− − − = ∆ + ∆ +
+ + + + +
& & & & & & &
The incremental acceleration, velocity and displacement are related through a numerical
integration method assumed in the analysis. For example, using the Newmark β method scheme,
2 2
1 1
1 1
1
{ ( )} { ( )} ( ) { ( )} { ( )}
2
1 1
{ ( )} { ( )} { ( )}
2 2
i i i i
i i i
x t t x t t x t t x t
x t t x t t x t
β β
+ +
+ +
∆ = ∆ + − ∆ + ∆
∆ = ∆ + ∆
& && &&
& && &&
where β is a constant of the Newmark Beta method, { ( )}
i
x t & and { ( )}
i
x t && are known velocity and
acceleration vectors at previous time step t
i
.
2
The three linear equations may be solved to determine incremental displacement and velocity
1
{ ( )}
i
x t
+
∆ ,
1
{ ( )}
i
x t
+
∆& and acceleration
1
{ ( )}
i
x t
+
&& at the new time step.
The stiffness of constituent members is nonlinear and the incremental internal force
1
{ ( )}
i
R t
+
∆
may not be expressed as the product of tangent stiffness [ ]
i
t
K and incremental displacement
1
{ ( )}
i
x t
+
∆ . Nor may the incremental damping force
1
{ ( )}
i
D t
+
∆ at new time step
1 i
t
+
not be equal
to the product of tangent damping matrix [ ]
i
t
C and incremental velocity
1
{ ( )}
i
x t
+
∆& . Therefore, the
equilibrium force may not be satisfied at new time step
1 i
t
+
. In other words, the internal force
{ } ) (
1 +
∆ +
i i
t t R may not be calculated as the sum of internal force { } ) (
i
t R at previous time step
i
t
and the product of tangent stiffness matrix [ ]
t
K and incremental displacement { } ) (
1 +
∆
i
t x .
i
x
1 i
x
+
i
F
1 i
F
+
1 i
F
+
∆
1 i i
K x
+
∆
1 i
x
+
∆
i
x&
1 i
x
+
&
i
D
1 i
D
+
1 i
D
+
∆
1 i i
C x
+
∆&
1 i
x
+
∆&
x
x&
i
K
i
C
Structural resistance
1
{ ( )}
i
R t
+
and damping resistance
1
{ ( )}
i
D t
+
at new time step must be
reevaluated for calculated new displacement
1
{ ( )}
i
x t
+
and velocity
1
{ ( )}
i
x t
+
& to satisfy the
hysteresis and damping relations. It should be noted, therefore, that the equilibrium of dynamic force
at new time increment is not satisfied;
1 1 1 1 1
{ ( )} [ ]{ ( )} { ( )} { ( )} { ( )}
i i i i i
R t M z t D t P t Error t
+ + + + +
+ + − = &&
It is generally time consuming to correct this error within the current time step because stiffness
and damping may be nonlinear in the dynamic problem. The unbalanced force
1
{ ( )}
i
Error t
+
must
be corrected in the equilibrium of dynamic force in the next time interval.
If the correction of error is desired within the same time step, NewtonRaphson iteration method
may be used. The incremental linearization of the equilibrium equation leads to;
[ ]{ } { } K u S ∆ =
The right hand side of the equation
( )
{ }
i
S may be updated for residue vector
1
{ ( )}
i
Error t
+
, and
new solution may be sought;
( ) ( )
[ ]{ } { }
i i
K u S ∆ =
The solution may be updated at iteration i;
( 1) ( ) ( )
{ } { } { }
i i i
u u u
+
= + ∆
In this solution process, the stiffness matrix should be reformulated. However, the reformulation of
stiffness matrix and its factorization is computationally expensive. Therefore, the initial stiffness
0
[ ] K is normally maintained during iteration steps. The convergence may be accelerated by the use
of a scale factor δ in updating the solution;
( 1) ( ) ( )
{ } { } { }
i i i
u u u δ
+
= + ∆
3
8.2 Modified Cholesky Matrix Decomposition
A set of n simultaneous linear algebraic equations are expressed in a matrix form:
} { } ]{ [ b X A =
where, ] [ A : coefficient matrix of size n x n, } {X : column vector of n unknowns, } {b : column
vector of n constants. The coefficient matrix ] [ A is symmetric and positive definite in a normal
structural analysis.
A symmetric positive definite matrix ] [ A can be decomposed into the product of three matrices as
follows:
] ][ [ ] [
] ][ ][ [ ] [
U D U
U D L A
T
=
=
in which ] [L : lower unit triangular matrix, ] [D : diagonal matrix, and ] [U : upper unit triangular
matrix (=
T
L] [ ). The diagonal elements of lower and upper unit triangular matrices are equal to
unity.
=
=
1 0 0 0
1 0 0
1 0
1
0 0 0
0 0 0
0 0 0
0 0 0
1
0 1
0 0 1
0 0 0 1
1 0 0 0
1 0 0
1 0
1
0 0 0
0 0 0
0 0 0
0 0 0
1
0 1
0 0 1
0 0 0 1
] [
3
2 23
1 13 12
33
22
11
3 2 1
23 13
12
3
2 23
1 13 12
33
22
11
3 2 1
32 31
21
L
M O M M M
L
L
L
L
M O M M M
L
L
L
L l l
M O M M M
L
L
L
L
M O M M M
L
L
L
L
M O M M M
L
L
L
L l l l
M O M M M
L l l
L l
L
n
n
n
nn n n n
n
n
n
nn n n n
u
u u
u u u
d
d
d
d
u u u
u u
u
u
u u
u u u
d
d
d
d
A
where,
ji ij
u = l .
+ + + + + + +
+ + + + +
+ + +
=
=
+ nn n n n n n n n n
n n n
n n
n
n
n
n
nn n n n
d u d u d u d u d u u d u d u d u u d u d
u d u u d u u d d u d u d u d u u d u d
u d u u d u d u u d d u d u d
u d u d u d d
u
u u
u u u
d u d u d u d
d u d u d
d u d
d
A
L L
M O M M M
L
L
L
L
M O M M M
L
L
L
L
M O M M M
L
L
L
2
3 33
2
2 22
2
1 11 3 33 2 23 22 13 11 2 22 1 12 11 1 11
3 33 2 23 22 1 13 11 33
2
23 22
2
13 11 23 22 13 12 11 13 11
2 22 1 12 11 23 22 13 12 11 22
2
12 11 12 11
1 11 13 11 12 11 11
3
2 23
1 13 12
3 33 2 22 1 11
33 23 22 13 11
22 12 11
11
1 0 0 0
1 0 0
1 0
1
0
0 0
0 0 0
] [
Expanding the product,
For i = 1; j = i:
11 11
d a =
j > i:
j j
u d a
1 11 1
=
4
For i > 1; j = i:
2
1
1
2
3 33
2
2 22
2
1 11 ki
i
k
kk ii ii i i i ii
u d d d u d u d u d a
∑
−
=
+ = + + + + = L
j > i:
∑
−
=
+ = + + + + =
1
1
3 3 33 2 2 22 1 1 11
i
k
kj ki kk ij ii ij ii j i j i j i ij
u u d u d u d u u d u u d u u d a L
The relations may be solved for
ij
u (for i<j) and
ii
d . Note that ] [
ij
u is an upper triangular
matrix with null lower triangular elements ( 0 =
ij
u for i > j) and with unit diagonal element ( 1 =
ii
u ).
Therefore, only upper triangular elements need be stored; unit diagonal elements may not have to
be stored. Instead diagonal element
ii
d of the diagonal matrix ] [D can be stored at the diagonals
ii
u of upper triangle matrix ] [U .
For i = 1; j = i:
11 11 11
) ( a d u = =
j > i:
11 1 1
/ d a u
j j
=
For i > 1; j = i:
∑
−
=
− = =
1
1
2
) (
i
k
ki kk ii jj ii
u d a d u
j > i:
ii kj ki
i
k
kk ij ij
d u u d a u / ) (
1
1
∑
−
=
− =
It should be noted that matrix elements
ij
a and
ij
u appear in the same expression; in other words,
ij
a may be replaced by
ij
u in the operation to save computer memory in programming.
Banded Coefficient Matrix: The coefficient matrix ] [ A of linear algebraic equations is normally
banded. Remember the formulation of structural stiffness matrix;
} ]{ [ } { D K P =
The structural stiffness matrix is formulated member by member rather than joint by joint. For
member i with start end connected to joint J and terminal end to joint K, member stiffness submatrix
[K
11
]
i
is added at (J, J) location of the structural stiffness matrix ] [K ,
i
K ] [
12
at (J, K) location,
i
K ] [
21
at (K, J) location, and
i
K ] [
22
at (K, K) location.
J K
l l
l l
l l
J (J,J)(J,K)
l l
l l
l l
l l
[K] = l l
l l
l l
l l
l l
l l
K (K,J)(K,K)
l l
l l
l l
J K
Member i
5
11 12
21 22
[ ]
i
i
K K
K
K K
=
Unless a member connects joints L and M, a structural stiffness submatrix at (L,M) is null; in
other words, (L,M) submatrix is not null if joints L and M are connected by a member. By proper
choice of joint numbering, the structural stiffness can be made narrowly banded.
To conserve computer time and storage, only the upper band of the stiffness matrix [K] is
normally constructed in a rectangular array of size N x NB, where N: the number of degrees of
freedom and NB: semibandwidth. The diagonal elements are stored in the first column of the
rectangular array.
The band width of the upper unit triangular matrix [U] is the same as that of the coefficient matrix
[A]. An element at ID(J) line and ID(K) column of the original square matrix can be stored at IR line
and IC column in a compact storage format.
IR=ID(J)
IC=ID(K)(IR1)
6
SUBROUTINE BANFAC (N, NB, A, *)
A flow chart to decompose a banded rectangular coefficient matrix ] [ A to the product of lower
triangular matrix, diagonal matrix and upper triangular matrix ] [U (Weaver and Gere, 1980). The
decomposed banded upper triangular matrix ] [U is stored in matrix ] [ A . The diagonal elements of
the diagonal matrix ] [D are stored in the first column of the array ] [ A .
N: number of unknowns in coefficient matrix ] [A ,
NB: semiband width of rectangular coefficient matrix ] [ A ,
] [ A : rectangular coefficient matrix of array size N x NB,
*: statement number in a calling program to which a nonstandard RETURN is directed.
References:
L. Fox: An Introduction to Numerical Linear Algebra, Oxford University Press, New York, 1965.
W. Weaver, Jr. and J. M. Gere: Matrix Structural Analysis of Framed Structures, Second Edition,
Van Nostrand Reinhold, New York, 1980.
7
8.3 Solution of Linear Algebraic Equations
A set of linear algebraic equations are given in a matrix form
} { } ]{ [ B X A =
The coefficient matrix ] [ A can be decomposed into the products of lower triangular matrix,
diagonal matrix and upper triangular matrix:
} { } ]{ ][ [ ] [ B X U D U
T
=
Let
} { } ]{ [ Y X U =
and
} { } ]{ [ Z Y D =
then
} { } { ] [ B Z U
T
=
The third relation can be expressed in a matrix form:
¦
¦
¦
)
¦
¦
¦
`
¹
¦
¦
¦
¹
¦
¦
¦
´
¦
=
¦
¦
¦
)
¦
¦
¦
`
¹
¦
¦
¦
¹
¦
¦
¦
´
¦
n n n n n
b
b
b
b
z
z
z
z
u u u
u u
u
M M
L
M O M M M
L
L
L
3
2
1
3
2
1
3 2 1
23 13
12
1
0 1
0 0 1
0 0 0 1
This relation can be solved for z
i
's by "forward substitution procedure",
∑
∑
−
=
−
=
− =
− =
− − =
− =
=
1
1
1
1
2 23 1 13 3 3
1 12 2 2
1 1
n
k
k kn n n
i
k
k ki i i
z u b z
z u b z
z u z u b z
z u b z
b z
M
M
The z’s are stored in vector {b}.
The equation } { } ]{ [ Z Y D = can be solved for y
i
's as follows;
¦
¦
¦
)
¦
¦
¦
`
¹
¦
¦
¦
¹
¦
¦
¦
´
¦
=
¦
¦
¦
)
¦
¦
¦
`
¹
¦
¦
¦
¹
¦
¦
¦
´
¦
n n nn
z
z
z
z
y
y
y
y
d
d
d
d
M M
L
M O M M M
L
L
L
3
2
1
3
2
1
33
22
11
0 0 0
0 0 0
0 0 0
0 0 0
ii i i
d z y / =
The y’s are stored in vector {z}.
8
The equation } { } ]{ [ Y X U = is expressed as follows:
¦
¦
¦
)
¦
¦
¦
`
¹
¦
¦
¦
¹
¦
¦
¦
´
¦
=
¦
¦
¦
)
¦
¦
¦
`
¹
¦
¦
¦
¹
¦
¦
¦
´
¦
n n
n
n
n
y
y
y
y
x
x
x
x
u
u u
u u u
M M
L
M O M M M
L
L
L
3
2
1
3
2
1
3
2 23
1 13 12
1 0 0 0
1 0 0
1 0
1
The equation can be solved for } {X by "backward substitution procedure":
∑
∑
=
+ =
− − − − − −
− − −
− =
− =
− − =
− =
=
n
k
k k
n
i k
k ik i i
n n n n n n n n
n n n n n
n n
x u y x
x u y x
x u x u y x
x u y x
y x
2
1 1 1
1
, 2 1 1 , 2 2 2
, 1 1 1
M
M
The final solution x’s are stored in vector {y}.
9
SUBROUTINE BANSOL (N, NB, U, B, X)
The solution of a set of linear equation in which the coefficient matrix ] [ A has been decomposed
into the product of unit lower triangular matrix
T
U] [ , diagonal matrix ] [D and unit upper triangular
matrix ] [U .
N: number of unknowns (number of columns of coefficient matrix ] [ A ),
NB: semiband width of coefficient matrix ] [ A ,
] [U : decomposed unit upper triangular matrix, stored in banded rectangular array, with
diagonal elements of diagonal matrix ] [D stored in the first column, array size of N x NB,
} {B : column vector of n constants,
} {X : column vector of n unknowns, used as working area for } {Y and } {Z .
10
8.4 Static Condensation
Frame structures are often analyzed by assuming floor slabs to be rigid in their own plane,
leading the same horizontal displacement at each floor. The mass at a floor is assumed to
concentrate at a floor level; the mass associated with vertical and rotational inertia is ignored.
The displacement degrees of freedom are selected to be
one horizontal displacement for a floor level, and vertical and
rotational displacements at each node.
Floor horizontal displacements
1
2
{ }
n
X
X
X
X
¦ ¹
¦ ¦
¦ ¦
=
´ `
¦ ¦
¦ ¦
¹ )
M
Node displacements
1
1
{ }
m
m
Y
Y
Y
¦ ¹
¦ ¦
Θ
¦ ¦
=
´ `
¦ ¦
¦ ¦
Θ
¹ )
where X
i
: horizontal displacement at floor i, Y
j
: vertical
displacement at node j, and
j
θ : rotation at node j.
If lumped translational masses are assumed at each floor level, the mass matrix becomes
diagonal with nonzero elements associated with the floor horizontal degrees of freedom and zero
elements associated with the nodal displacements.
] [M = diagonal [M
1
,M
2
,...,M
n
,0,0,....,0]
=
0 0
0
XX
M
The stiffness matrix is normally banded, but can be expressed in a partitioned form;
=
YY YX
XY XX
K K
K K
K] [
The equation of motion of an undamped system under horizontal base motion y
dt
y d
& & =
2
2
may be
expressed as
i
Y
X
i
XX
i
YY YX
XY XX
i
XX
R
R
y
e M
Y
X
K K
K K
Y
X M
)
`
¹
¹
´
¦
−
)
`
¹
¹
´
¦
− =
)
`
¹
¹
´
¦
∆
∆
+
)
`
¹
¹
´
¦
+
+ +
1
1 1
0 0 0
0
0 0
0
& &
& &
& &
where, } {X : horizontal floor displacement, } {Y : nodal vertical displacement and rotation, } {e :
vector of unit elements.
From the second equation;
i Y i YY i YX
R Y K X K } { } ]{ [ } ]{ [
1 1
− = ∆ + ∆
+ +
and
) } { } ]{ ([ ] [ } {
1
1
1 i Y i YX YY i
R X K K Y + ∆ − = ∆
+
−
+
Therefore, the first equation can be written as
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Degrees of freedom
11
i X i Y YY KY XX
i YX YY XY XX i XX
R R K K y e M
X K K K K X M
} { } { ] ][ [ } ]{ [
} ]){ [ ] ][ [ ] ([ } ]{ [
1
1
1
1
− + − =
∆ − +
−
+
−
+
& &
& &
or
i X i Y YY XY
i i
R R K K y e M
X K X M
} { } { ] ][ [ } ]{ [
} ]{ [ } ]{ [
1 *
1
*
1
*
− + − =
∆ +
−
+ +
& &
& &
where, ] [
*
M : story mass matrix, ] [
*
K : story stiffness matrix (= ] [ ] ][ [ ] [
1
YX YY XY XX
K K K K
−
− ).
The equation of motion can be solved for incremental story displacement
1
} {
+
∆
i
X and then
incremental nodal displacement and rotation
1
} {
+
∆
i
Y can be determined by
) } { } ]{ ([ ] [ } {
1
1
1 i Y i YX YY i
R X K K Y + ∆ − = ∆
+
−
+
It should be noted that the inverse matrix of ] [
YY
K is not formed in formulating the story matrix;
but a set of linear algebraic equations of the following form are solved;
} { } ]{ [
] [ ] ][ [
Y YY
YX YY
R b K
K a K
=
=
or
] [ ] ][ [
Y YX YY
R K b a K M M =
by decomposing the matrix ] [
YY
K into ] ][ ][ [ U D L .
} { } ]{ [ } {
1 1
b X a Y
i i
− ∆ − = ∆
+ +
A series of plane frames were analyzed by (a) direct ] ][ ][ [ U D L decomposition of the entire
matrix and (b) the static decomposition. The number of stories was varied from 3 to 30. The number
of bays in a frame was 2, 4 or 9. The computation time required to solve a set of equations was
compared.
12
Number of Stories
C
o
m
p
u
t
a
t
i
o
n
t
i
m
e
,
m
s
9bay frame
4bay frame
2bay frame
L
T
DL Decomposition
Static condensation
The computation time is reduced to onehalf to onethird by the direct [L][D][U] decomposition
method compared to the static condensation method for a frame of more than 10story high.
However, the number of operation is much less for the static condensation if the stiffness does
not change during a time increment. The computation time by the static condensation may not be
reduced even if the number of parallel frames increased.
13
Home Assignment No. 4
20020304
Otani, S.
For a fourstory onebay with rigid beams, formulate the stiffness matrix of a structure. The
material properties are constant throughout the structure. The moment of inertia of columns is 3I
in the first and second stories and 2I in the third and fourth stories.
h
h
h
h
3EI
3EI
2EI
2EI
If shear and axial deformation are ignored, each floor joint can move only in the horizontal
direction. Neither vertical nor rotation can take place at the joint.
Slope deflection equation of column AB with flexural rigidity EI ;
2
(2 3 )
2
( 2 3 )
AB A B AB
BA A B AB
EI
M R
h
EI
M R
h
θ θ
θ θ
= + −
= + −
Because the beam is rigid;
0.0
A B
θ θ = =
Story shear and member rotation angle relation
2
1 3 3
24
2
24 24
( )
AB BA
i AB
AB i i
M M EI
V R
h h
EI EI
x x
h h
−
+
= − =
= ∆ = −
The stiffness matrix of the structure is
 
−
− −
− −
−
=
2 2 0 0
2 4 2 0
0 2 5 3
0 0 3 6
24
3
h
EI
K
Decompose the stiffness matrix in [ ] [ ][ ][ ] K L D U = form.
14
[Solution]
The stiffness matrix of the structure is
3
3
6 3 0 0
3 5 2 0
24
[ ]
0 2 4 2
0 0 2 2
1 0 0 0 1
144 0 0 0 1 0 0
2
1
0 84 0 0 1 0 0
4
2
0 1 0
480
7
4 0 0 0
0 1 0
7
7
7
0 0 1
72
10
7 0 0 0
0 0 1
5
0 0 0 1
10
EI
K
h
EI
h
−
− −
=
− −
−
−
−
−
=
−
−
−
1
Chapter 9. Formulation of Member Stiffness Matrix
9.1 Introduction
Nonlinear analysis of a building considers the nonlinear behavior of
materials. The nonlinear geometrical properties associated with large
deformation are normally ignored in the analysis. However, the ∆ − P
effect, which is defined as the overturning effect of horizontally
displaced mass, may be included in the analysis using an approximate
formulation. The ∆ − P effect is known to increase the structural
displacement response when a story drift angle (interstory
deformation divided by story height) exceeds approximately 0.01.
The equilibrium of external forces { }
i
P at time step
i
t and
internal resistances { }
i
R at all joints are expressed as;
{ } { }
i i
R P =
Assuming that structural members behave linearly elastic between time steps
i
t and
1 i
t
+
,
incremental resistance
1
{ }
i
R
+
∆ between the two adjacent time steps may be written as the product
of instantaneous (tangent) stiffness
1
[ ]
i
K
+
and incremental displacement
1
{ }
i
D
+
∆ at joints. Hence,
1 1 1 1
[ ] { } { } { }
i i i i
K D R P
+ + + +
∆ = ∆ = ∆
in which
1
{ }
i
P
+
∆ : incremental external forces at joint during the time increment.
Thus, the linearly analysis of a structure reflecting the damage state is necessary in the nonlinear
response analysis.
The structural stiffness matrix of a frame structure can be formulated for free joints from
(a) member stiffness matrices in the local coordinate system,
(b) transformation of coordinates,
(c) continuity conditions of displacement at joint,
(d) equilibrium condition of forces at joint,
(e) separating the free joint and the support joint displacements and forces,
(f) solution of a set of linear equations for free joint displacement, and then
(g) determination of member end actions using the member stiffness matrix.
Steps (b) to (f) can be processed automatically by considering the geometrical conditions and the
coordinate transformation. Therefore, it is more important to formulate stiffness matrices of structural
members considering their damaged state and the formulation of a member stiffness matrix needs a
special attention.
P
∆
P
M P = ×∆
2
9.2 Formulation of Member Stiffness Matrix
A general procedure to formulate a member stiffness matrix is outlined in this section. The
stiffness matrix of a member can be formulated from a stiffness relation of a statically determinate
member. First, a method is introduced to formulate the member stiffness from a stiffness relation of
a cantilever member. The equilibrium matrix is used in the derivation.
Flexibility of Arbitrary Cantilever: For a member of given stiffness distribution, fixed at the start
end and free at the terminal end, free end displacement { } d
i 2
can be calculated for any free end
actions { } p
i 2
, for example, by Castigliano's theorem or the unit load method.
Displacement
2 jk
f at the free end in jdirection due to a unit load in kdirection applied at the
free end is calculated by the unit load method, by
evaluating axial forces ( )
j
n x and ( )
k
n x , bending
moments ( )
j
m x and ( )
k
m x , and shear forces
( )
j
v x and ( )
k
v x due to unit load in j and
kdirections, separately;
2
0
( ) ( ) ( ) ( ) ( ) ( )
{ }
( ) ( ) ( )
L
j k j k j k
jk
s
n x n k v x v k m x m k
f dx
EA x GA x EI x
⋅ ⋅ ⋅
= + +
∫
where, E : Young’s modulus, ( ) A x : cross sectional area, ( )
s
A x : effective shear area, and ( ) I x :
moment of inertia of section at coordinate x.
Castigliano's Theorem is expressed in the following form to calculate displacement
i
d :
2 2 2
0
( ) ( ) ( )
( ) ( )
2 ( ) 2 ( ) 2 ( )
i
i
L
i
U
d
p
m x v x n x
dx
p EI x GA x EA x
κ
∂
=
∂
∂
= + +
∂
∫
where U: strain energy stored in the structure expressed in terms of external forces
2 2 2
{ , , }
x y z
p p m
at the free end, ( ) m x : bending moment, ( ) v x : shear, ( ) n x : axial force, κ : shape factor for shear
deformation, E : Young’s modulus of material, ( ) I x : moment of inertia of section, ( ) A x : cross
sectional area.
The flexibility matrix of a cantilever can be expressed as
{ } [ ] { } d f p
i i i 2 22 2
=
where, each element
ij
f of flexibility matrix
22
[ ]
i
f represents the displacement in jdirection at the
free end due to unit load applied in the kdirection at the free
end.
Stiffness of Elastic Prismatic Cantilever: For a linearly
elastic prismatic cantilever, neglecting shear deformation,
the flexibility relation is expressed as follows;
1 2
{p
2
}
i
{d
2
}
i
EI, EA
member i
L
2 2
,
x x
p d
2 2
,
y y
p d
2 2
,
z z
m θ
member i
( ), ( ), ( )
s
EA x EI x GA x
1 2
3
d
d
L
EA
L
EI
L
EI
L
EI
L
EI
p
p
m
x
y
z
i
i
x
y
z
i
2
2
2
3 2
2
2
2
2
0 0
0
3 2
0
2
θ
¦
´
¦
¹
¦
¹
`
¦
)
¦
=
¦
´
¦
¹
¦
¹
`
¦
)
¦
Solving this relationship for free end action
i
p } {
2
, we obtain the stiffness relation for the
cantilever;
{ } [ ] { }
[ ] { }
p f d
k d
i i i
i i
2 22
1
2
22 2
=
=
−
[ ] [ ] k f
i i 22 22
1
=
−
in which
22
[ ]
i
k is the stiffness matrix of the cantilever member i .
For a linearly elastic prismatic member; the stiffness matrix is expressed
p
p
m
EA
L
EI
L
EI
L
EI
L
EI
L
d
d
x
y
z
i
i
x
y
z
i
2
2
2
3 2
2
2
2
2
0 0
0
12 6
0
6 4
¦
´
¦
¹
¦
¹
`
¦
)
¦
= −
−
¦
´
¦
¹
¦
¹
`
¦
)
¦
θ
Stiffness Matrix of Cantilever with Shear Deformation: The shear deformation increases lateral
deformation
2 y
d at the free end due to lateral force
2 y
p . The flexibility relation is written as
2 2
3 2
2 2
2 2 2
0 0
0
3 2
0
2
x x
y y
z z
i i
i
L
EA
d p
L L L
d p
GA EI EI
m
L L
EI EI
κ
θ
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
= +
´ ` ´ `
¦ ¦ ¦ ¦
¹ ) ¹ )
By inverting the flexibility relation, the stiffness matrix of a prismatic cantilever can be expressed as
2 2
2 2 3 2
2 2
2
0 0
1 12 1 6
0
1 2 1 2
1
1 6 4
2
0
1 2 1 2
x x
y y
z z
i i
i
EA
L
p d
EI EI
p d
L L
m
EI EI
L L
γ γ
θ
γ
γ γ
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
= −
´ ` ´ `
+ +
¦ ¦ ¦ ¦
¹ ) ¹ )
+
−
+ +
where
2
6EI
GAL
κ
γ = , κ : shape factor for shear deformation.
4
Equilibrium Matrix [ ] H
jk i
: For a straight member i of
length L, without any intermediate loads, member end
actions
i
p } {
1
and
i
p } {
2
in the member coordinates
must satisfy the equilibrium conditions:
1 2
1 2
@1 1 2 2
0 : 0
0 : 0
0 : 0
x x x
y y y
z z z y
p p p
p p p
m m m Lp
= + =
= + =
= + + =
∑
∑
∑
in which, xaxis of the member coordinate system is taken in the direction of the straight member.
Writing in a matrix form; the member end forces { } p
i 1
and { } p
i 2
of member i are related
through an equilibrium matrix [ ] H
i 12
{ } [ ] { } { } p H p
i i i 1 12 2
0 + =
where,
{ }
{ }
i
z
y
x
i
i
z
y
x
i
m
p
p
p
m
p
p
p
¦
)
¦
`
¹
¦
¹
¦
´
¦
=
¦
)
¦
`
¹
¦
¹
¦
´
¦
=
2
2
2
2
1
1
1
1
[ ] H
L
i
i
12
1 0 0
0 1 0
0 1
=
Force
12 2
[ ] { }
i i
H p is a force developed at end 1 due to force
2
{ }
i
p acting at end 2.
By the same token,
[ ] { } { } { } H p p
i i i 21 1 2
0 + =
where the member length L in matrix [ ] H
i 21
is measured from the terminal point.
[ ] H
L
i
i
21
1 0 0
0 1 0
0 1
=
−
Note that
{ } [ ] { }
[ ] [ ] { }
p H p
H H p
i i i
i i i
1 12 2
12 21 1
= −
=
hence,
[ ] [ ] H H
12 21
1
=
−
The inverse matrix of an equilibrium matrix is obtained by changing the sign of offdiagonal terms.
p
y1
p
y2
p
x1
p
x2
m
z1
m
z2
L
1 2
member i
5
Rigid Body Displacement: For small displacement
( sin , cos 1.0 θ θ θ ≈ ≈ ), member end displacements
1
{ }
i
d and
2
{ }
i
d through rigid body displacement are
also related by the equilibrium matrix;
1 2
1 1 2
1 2
z z
z y y
x x
L d d
d d
θ θ
θ
=
+ =
=
This relation can be written in a matrix form:
d
d L
d
d
x
y
z
i i
x
y
z
i
2
2
2
1
1
1
1 0 0
0 1
0 0 1 θ θ
¦
´
¦
¹
¦
¹
`
¦
)
¦
=
¦
´
¦
¹
¦
¹
`
¦
)
¦
or symbolically,
2 12 1
{ } [ ] { }
T
i i i
d H d =
Note that
2
{ }
i
d is a displacement at the member end 2 caused by a rigid body displacement at
member end 1. By the same token, we obtain
1 21 2
{ } [ ] { }
T
i i i
d H d =
Member Stiffness Matrix: For a free body of member i, the member end displacement and force
relation is formulated. If a starting end 1 of member i is allowed to displace, additional displacement
takes place at the terminal end 2 by a rigid
body movement; i.e., the deformation { } e
i
of the member is the difference of terminal
end displacement { } d
i 2
and rigid body
displacement [ ] { } H d
i
T
i 12 1
at the terminal
end 2 caused by displacement { } d
i 1
at the
starting end 1:
{ } { } [ ] { } e d H d
i i i
T
i
= −
2 12 1
The deformation { } e
i
is related to the
deformation at the free end of a cantilever
member due to forces
2
{ }
i
p applied at the
free end. Therefore,
2 22
{ } [ ] { }
i i i
p k e =
Expressing the deformation by the member end displacements
1
{ }
i
d and
2
{ }
i
d , the member
end force
2
{ }
i
p is related to the member end displacements;
{ } [ ] { }
[ ] ({ } [ ] { } )
[ ] [ ] { } [ ] { }
[ ] { } [ ] { }
p k e
k d H d
k H d k d
k d k d
i i i
i i i
T
i
i i
T
i i i
i i i i
2 22
22 2 12 1
22 12 1 22 2
21 1 22 2
=
= −
= − +
= +
where
d
y1
d
y2
d
x1
d
x
θ
1z
θ
2z
1
2
1 2
{d
1
}
i
{e}
i
2’
1’
{d
2
}
i
2”
[H
12
]
T
i
{d
1
}
i
member i
6
[ ] [ ] [ ] k k H
i i i
T
21 22 12
= −
For a prismatic member,
22 3 2
2
0 0
12 6
[ ] 0
6 4
0
i
i
EA
L
EI EI
k
L L
EI EI
L L
= −
−
21 3 2
2
0 0
12 6
[ ] 0
6 2
0
i
i
EA
L
EI EI
k
L L
EI EI
L L
−
= − −
From the equilibrium of member end forces,
i i i i
i i i
T
i i
i i
T
i i
i i i
d k d k
d k H d H k H
d d H k H
p H p
} { ] [ } { ] [
} ]{ [ ] [ } { ] ][ [ ] [
) } { } { ] [ ]( [ ] [
} { ] [ } {
2 12 1 11
2 22 12 1 12 22 12
2 1 12 22 12
2 12 1
+ =
− =
+ − − =
− =
where,
T
i
i i i
T
i i i i
k
k H k
H k H k
] [
] [ ] [ ] [
] [ ] [ ] [ ] [
21
22 12 12
12 22 12 11
=
− =
=
For a prismatic member,
11 3 2
2
0 0
12 6
[ ] 0
6 4
0
i
i
EA
L
EI EI
k
L L
EI EI
L L
=
12 3 2
2
0 0
12 6
[ ] 0
6 2
0
i
i
EA
L
EI EI
k
L L
EI EI
L L
−
= −
−
1 11 12 1
2 21 22 2
i i i
p k k d
p k k d
¦ ¹ ¦ ¹
=
´ ` ´ `
¹ ) ¹ )
7

Note:
T T T
A B B A ] [ ] [ ]) ][ ([ =
8
9.3 Member with Rigid Ends
It is often necessary to consider rigid zones at member ends. For example, a beamcolumn
connection may be assumed to be rigid or a structural wall is represented by a column with rigid
beams at each floor. Let us consider the treatment of a member having rigid zones at the ends.
Beamcolumn Connection
Structural Wall
Equilibrium Matrix: The equilibrium matrix ] [
AB
H for member AB, not lying in xdirection, can be
defined in a more general form if the lengths of the member are
AB
x and
AB
y in x and ydirection,
respectively;
@
0: 0
0: 0
0: 0
x xA xB
y yA yB
z A zA zB xB AB yB AB
p p p
p p p
m m m p y p x
= + =
= + =
= + − + =
∑
∑
∑
{ } [ ]{ } {0}
A AB B
p H p + =
The equilibrium matrix for this member is given as
−
=
1
0 1 0
0 0 1
] [
AB AB
AB
x y
H
It should be noted that force { }
A
p at starting end A is necessary to satisfy the equilibrium when
force { }
B
p acts at the terminal end B.
Product of Equilibrium Matrices: Consider regions AB and BC connected at B. No external force
is assumed to act at joint B.
If force { }
C
p acts at member end C of region BC, force { '}
B
p at B end is necessary to satisfy
the equilibrium in the region;
} ]{ [ } ' {
} 0 { } ]{ [ } ' {
C BC B
C BC B
p H p
p H p
− =
= +
The equilibrium of forces at end B is
} ]{ [ } ' { } {
} 0 { } ' { } {
C BC B B
B B
p H p p
p p
= − =
= +
A
B
y
AB
x
AB
x
y
9
In other words, force { }
B
p is a force developed at
joint B due to force { }
C
p acting at joint C.
By the same token, force { '}
A
p is developed by
member end force { }
B
p acting at member end B to
satisfy the equilibrium of forces in member AB.
} ]{ [
} ]{ ][ [
} ]{ [ } ' {
C AC
C BC AB
B AB A
p H
p H H
p H p
− =
− =
− =
Therefore,
[ ] [ ][ ]
1 0 0 1 0 0
0 1 0 0 1 0
1 1
1 0 0
0 1 0
( ) ( ) 1
1 0 0
0 1 0
1
AC AB BC
AB AB BC BC
AB BC AB BC
AC AC
H H H
y x y x
y y x x
y x
=
=
− −
=
− + +
=
−
The equilibrium matrix [ ]
AC
H of a region ABC
combing regions AB and BC can be expressed as the
product of corresponding equilibrium matrices [ ]
AB
H and [ ]
BC
H . The product matrix can be
formulated by adding the corresponding offdiagonal elements of the element matrices.
Rigid Zones and Equilibrium Matrices: Consider a member consisting of an elastic middle part BC
and rigid zones AB and CD at the two ends. The two rigid zones may not be in line with the elastic
part. The local coordinate system is defined for the middle elastic part. No external loads acts within
compound region of ABCD.
The equilibrium matrix [ ]
AB
H for rigid zone AB can be defined if the lengths of the rigid zones
are
AB
x and
AB
y in x and ydirections:
A
B
C
} {
C
p
} ' {
A
p
A
} ' {
A
p
} {
C
p
} ' {
B
p
} {
B
p
B
B
C
AB
x
BC
x
BC
y
AB
y
AC
x
AC
y
x
y
A
B
C
A
B
C
D
x
AB
x
BC x
CD
y
AB
y
CD
y
x
y
BC
10
−
=
1
0 1 0
0 0 1
] [
AB AB
AB
x y
H
Similarly, for the middle elastic part BC and rigid zone CD, equilibrium matrices [ ]
BC
H and
[ ]
CD
H can be formulated:
−
=
−
=
1
0 1 0
0 0 1
] [
1
0 1 0
0 0 1
] [
CD CD
CD
BC BC
BC
x y
H
x y
H
Member Stiffness Matrix of Elastic Part: Consider a
cantilever member; fixed at start end B and free at
terminal end C. A flexibility matrix
22
[ ]
BC
f of the middle
elastic part can be formulated by calculating deformation
2
{ }
BC
d at free end 2 caused by free end force
2
{ }
BC
p .
BC BC BC
p f d } { ] [ } {
2 22 2
=
The relation can be inverted to express force
2
{ }
BC
p
required to deform the terminal end by
2
{ }
BC
d ;
BC BC BC
d k p } [ ] [ } {
2 22 2
=
For a prismatic member,
−
− =
BC BC
BC BC
BC
BC
x
EI
x
EI
x
EI
x
EI
x
EA
k
4 6
0
6 12
0
0 0
] [
2
2 3 22
As discussed in Section 6.2 "Formulation of Member Stiffness Matrix, member stiffness
submatrices of the middle elastic part BC are formulated;
T
BC BC BC
T
BC BC BC BC
H k H
H k H k
] [ ] ][ [
] [ ] [ ] [ ] [
22
12 22 12 11
=
=
BC BC
BC BC
T
BC BC
k H
k H
k k
] ][ [
] [ ] [
] [ ] [
22
22 12
21 12
− =
− =
=
BC BC BC
d
d
k k
k k
p
p
)
`
¹
¹
´
¦
=
)
`
¹
¹
´
¦
2
1
22 21
12 11
2
1
1 2
{p
2
}
BC
{d
2
}
BC
BC
p
y1 p
y2
p
x1
p
x2
m
z1
m
z2
L
1 2
BC
11
Member Stiffness Matrix of Compound Member: Consider a cantilever member ABCD (rigid zone
AB, elastic zone BC and rigid zone CD) with fixed end A and free end D. Let force { }
D
p be
member end force at end D of rigid zone CD,
and
2
{ }
BC
p be member end force at end C of
elastic part BC. These two forces are related by
the equilibrium matrix [ ]
CD
H as follows;
} ]{ [ } {
2 D CD BC
p H p =
The rigid body displacement { }
D
d at D
caused by displacement { }
C
d at end C (or terminal end of elastic part BC) is expressed using
equilibrium matrix [ ]
CD
H ;
BC
T
CD
C
T
CD D
d H
d H d
} { ] [
} { ] [ } {
2
=
=
The region AB is rigid and does not deform. Therefore, terminal member end displacement
2
{ }
BC
d of part BC is expressed by the flexibility
22
[ ]
BC
f of part BC and terminal end force
2
{ }
BC
p :
BC BC BC
p f d } { ] [ } {
2 22 2
=
Using the relation of displacements at nodes C and D, the flexibility [ ]
DD
f of ABCD as a
cantilever fixed at A can be expressed as follows;
} ]{ [
} ]{ [ ] [ ] [
} { ] [ ] [
} { ] [ } {
22
2 22
2
D DD
D CD BC
T
CD
BC BC
T
CD
BC
T
CD D
p f
p H f H
p f H
d H d
=
=
=
=
where
] [ ] [ ] [ ] [
22 CD BC
T
CD DD
H f H f =
The stiffness submatrix [ ]
DD
k of a cantilever ABCD is obtained by inverting the flexibility [ ]
DD
f
1
22
1
1 1
22
1
1
22
1
) ] ([ ] [ ] [
) ] ([ ] [ ] [
]) [ ] [ ] ([
] [ ] [
− −
− − −
−
−
=
=
=
=
T
CD BC CD
T
CD BC CD
CD BC
T
CD
DD DD
H k H
H f H
H f H
f k
Note that the inverse of an equilibrium matrix can be obtained by changing the sign of offdiagonal
elements; i.e.,
−
=
−
1
0 1 0
0 0 1
] [
1
CD CD
CD
x y
H
and also the inverse of a product of matrices is the product of inverse matrices reversing the order of
product;
1 1 1
([ ][ ]) [ ] [ ] A B B A
− − −
=
provided each element matrix has an inverse matrix.
Other member stiffness submatrices are obtained as follows by noting:
A
B C
D
{p
D
}
{d
D
}
{p
2
}
BC
={p
C
}
{p
C
}
{p
B
}
{p
B
}
12
T
AB
T
BC
T
CD
T
CD BC AB
T
AD
T
CD BC CD DD
CD BC AB AD
H H H H H H H
H k H k
H H H H
] [ ] [ ] [ ]) ][ ][ ([ ] [
) ] ([ ] [ ] [ ] [
] ][ ][ [ ] [
1
22
1
= =
=
=
− −
and
T
BC BC BC BC
H k H k ] [ ] ][ [ ] [
22 11
=
then
T
AB BC AB
T
AB
T
BC BC BC AB
T
AB
T
BC
T
CD
T
CD BC CD CD BC AB
T
AD DD AD AA
H k H
H H k H H
H H H H k H H H H
H k H k
] [ ] ][ [
] [ ] [ ] ][ ][ [
] [ ] [ ] [ ) ] ([ ] [ ] ][ ][[ ][ [
] ][ ][ [ ] [
11
22
1
22
1
=
=
=
=
− −
1
12
1
22
) ] ([ ] ][ [
) ] ([ ] ][ ][ [
] ][ [
] [ ] [
−
−
=
− =
− =
=
T
CD BC AB
T
CD BC BC AB
DD AD
T
DA AD
H k H
H k H H
k H
k k
or compound member AD consisting of
rigid zones AB and CD at ends and
elastic part BC, member stiffness matrix
relations in local coordinate system are
expressed as follows;
} ]{ [ } ]{ [ } {
} ]{ [ } ]{ [ } {
D DD A DA D
D AD A AA A
d k d k p
d k d k p
+ =
+ =
It should be noted that the submatrices [ ]
IJ
k of the compound member could be easily derived
from the member stiffness submatrices [ ] k
ij b
of the elastic part by transformation operation.
A
B C
D
} {
A
P
{ }
D
P
{ }
D
d
{ }
A
d
13
9.4 Member with Flexible Ends
A member AB has springs AA’ and BB’ at both ends.
Each spring has stiffness for axial deformation, lateral
deformation and rotation with spring constants
,
x y
k k and k
θ
, respectively. The length of the spring is
none.
The flexibility matrix of spring at A’ or B’ is expressed as
'
'
'
'
1
0 0
1
[ ] 0 0
1
0 0
1
0 0
1
[ ] 0 0
1
0 0
x
AA
y
AA
x
B B
y
B B
k
f
k
k
k
f
k
k
θ
θ
=
=
The equilibrium matrices of parts AB and A’B are equal because no length is given at the springs.
'
[ ] [ ]
1 0 0
0 1 0
0 1
AB A B
AB
H H
L
=
=
and the equilibrium matrix of part B’B is unit
matrix;
'
1 0 0
[ ] 0 1 0
0 0 1
B B
H
=
For a cantilever member AB fixed at end A, internal forces at end A’ induced by external force
{ }
B
p is given as
' '
{ } [ ]{ }
A A B B
p H p =
The deformation of spring
'
{ }
A
d at AA’ is given as
' ' '
' '
{ } [ ]{ }
[ ][ ]{ }
A AA A
AA A B B
d f p
f H p
=
=
The rigid body displacement at end B due to the deformation of the spring at A end is
' ' ' ' '
[ ] { } [ ] [ ][ ]{ }
T T
A B A A B AA A B B
H d H f H p =
k
θ
y
k
x
k
x
y
A A’
Spring at AA’
A B
A’
B’
, EI EA
L
B
A’
B’
L
{ }
B
p
{ }
B
d
A
14
The displacement at end B’ due to the deformation of middle region A’B’ caused by external force
{ }
B
p is given as
' 22 ' ' ' 22 ' '
[ ] [ ] [ ]{ } [ ] { }
T
B B A B B B B A B B
H f H p f p =
The displacement at B end due to the deformation of spring B’B caused by external force { }
B
p
is given by
' ' ' '
[ ] [ ][ ]{ } [ ]{ }
T
B B B B B B B B B B
H f H p f p =
The total displacement at end B caused by the external force { }
B
p is the sum of the
contributions of springs AA’ and B’B and middle elastic region A’B’;
' ' ' 22 ' ' '
' ' ' 22 ' ' '
{ } [ ] [ ][ ]{ } [ ] { } [ ]{ }
([ ] [ ][ ] [ ] [ ]){ }
T
B A B AA A B B A B B B B B
T
A B AA A B A B B B B
d H f H p f p f p
H f H f f p
= + +
= + +
The flexibility matrix [ ]
BB
f is given as
' ' ' ' ' '
[ ] [ ] [ ][ ] [ ] [ ]
T
BB A B AA A B A B B B
f H f H f f = + +
The stiffness matrix [ ]
BB
k of the cantilever can be obtained by inverting the flexibility matrix
[ ]
BB
f ;
1
[ ] [ ]
BB BB
k f
−
=
For the member AB, the stiffness submatrices
are obtained as
[ ] [ ][ ][ ]
[ ] [ ][ ]
T
AA AB BB AB
AB AB BB
k H k H
k H k
=
= −
{ } [ ]{ } [ ]{ }
{ } [ ]{ } [ ]{ }
A AA A AB B
B BA A BB B
p k d k d
p k d k d
= +
= +
For a special case where no resistance is given in the spring, the spring constant can be set to
be zero in the member stiffness relation.
B A’ B’
{ }
B
p
{ }
B
d
A
{ }
A
d
{ }
A
p
15
9.5 Member Stiffness based on Stiffness of Simply Supported Member
A simply supported member is often used as a basic statically determinate system in formulating
the member stiffness matrix. Deformation of a member can be expressed by the axial deformation e,
member end rotations
A
θ and
B
θ , and member deformation angle
AB
R .
Member end displacements of a simply supported
member AB are extension e and member end
rotations
A
θ and
B
θ . These displacements are
related with displacements
1
{ } d and
2
{ } d at the
member ends expressed in the member coordinates:
1 2
1 1 2
2 1 2
( ) /
( ) /
AB x x
A z y y
B z y y
e d d
d d L
d d L
θ θ
θ θ
= − +
= + −
= + −
In a matrix form;
1 2
1 2
1 2
1 0 0 1 0 0
1 1
0 1 0 0
1 1
0 0 0 1
AB x x
A y y
B z z
i i
i i
e d d
d d
L L
L L
θ
θ θ θ
−
¦ ¹ ¦ ¹ ¦ ¹
¦ ¦ ¦ ¦ ¦ ¦
= + −
´ ` ´ ` ´ `
¦ ¦ ¦ ¦ ¦ ¦
¹ ) ¹ ) ¹ )
−
The relation can be written in a matrix form:
1 1 2 2
{ } [ ] { } [ ] { }
AB i i i i i
e B d B d = +
Member end forces in the member coordinates
can be expressed by member end forces of the
simply supported member AB;
A z
B A y
AB x
m m
L m m p
p p
=
+ =
− =
1
1
1
/ ) (
B z
B A y
AB x
m m
L m m p
p p
=
+ − =
=
2
2
2
/ ) (
or in a matrix form:
i
B
A
AB
i i
z
y
x
i
B
A
AB
i i
z
y
x
m
m
p
L L
m
p
p
m
m
p
L L
m
p
p
¦
)
¦
`
¹
¦
¹
¦
´
¦
− − =
¦
)
¦
`
¹
¦
¹
¦
´
¦
¦
)
¦
`
¹
¦
¹
¦
´
¦
−
=
¦
)
¦
`
¹
¦
¹
¦
´
¦
1 0 0
/ 1 / 1 0
0 0 1
0 1 0
/ 1 / 1 0
0 0 1
2
2
2
1
1
1
or
m
A
m
B
p
AB
p
x1
p
x2
p
y1
p
y2
m
z1
m
z2
1 2
2 z
θ
2 x
d
2 y
d
AB
R
1 z
θ
1 x
d
1 y
d
B
θ
A
θ
y
x
member i
16
{ }   { }
{ }   { }
i AB
T
i i
i AB
T
i i
p B p
p B p
2 2
1 1
=
=
Member Stiffness Matrix in Local Coordinates: The incremental member end forcedeformation
of a simply supported member may be expressed in a stiffness matrix form;
{ }  { }
AB AB AB
e k p =
¦
)
¦
`
¹
¦
¹
¦
´
¦
=
¦
)
¦
`
¹
¦
¹
¦
´
¦
=
B
A
AB
AB
B
A
AB
AB
e
e
m
m
p
p
θ
θ } {
} {
For a linearly elastic prismatic member,
=
L
EI
L
EI
L
EI
L
EI
L
EA
k
AB
4 2
0
2 4
0
0 0
] [
The member stiffness matrix in the local coordinate system is obtained by the transformation of
forces and displacements;
     
i i
i
i i AB
T
i
i
d
d
k k
k k
d
d
B B k B B
p
p
)
`
¹
¹
´
¦
=
)
`
¹
¹
´
¦
=
)
`
¹
¹
´
¦
2
1
22 21
12 11
2
1
2 1 2 1
2
1
, ,
in which
       
         
       
i i AB
T
i i
i i AB
T
i
T
i i
i i AB
T
i i
B k B k
B k B k k
B k B k
2 2 22
2 1 21 12
1 1 11
=
= =
=
17
Member with Rigid Zones: If rigid zones of length
λ
A AB
L and λ
B AB
L are attached at member ends, the
transformation of deformation and force must be
considered.
For a simply supported flexible part A'B', member
end moments m
A'
and m
B'
and member end rotations
θ
A'
and θ
B'
are defined. For a simply supported total
member AB, member end moments m
A
and m
B
and
member end rotation θ
A
and θ
B
are defined. Let the
length of the total member be L
AB
, rigid zones λ
A AB
L
and λ
B AB
L . The inflection point is denoted by C.
Looking at a triangle AA'C of the deformed shape,
we have the following relation among angles,
θ θ
A A B A
R
' ' '
= +
in which
R
L L
L
A B
A A AB B B AB
A B AB
' '
( )
=
+
− −
θ λ θ λ
λ λ 1
Therefore,
'
(1 )
1
B A B B
A
A B
λ θ λ θ
θ
λ λ
− +
=
− −
Similarly,
'
(1 )
1
A A A B
B
A B
λ θ λ θ
θ
λ λ
+ −
=
− −
In a matrix form,
'
'
(1 )
1
(1 ) 1
B B A A
A A B A B B
λ λ θ θ
λ λ θ λ λ θ
− ¦ ¹ ¦ ¹
=
´ ` ´ `
− − −
¹ ) ¹ )
or
{ }   { } θ θ A = '
For the moments, member end moments
A
m and
B
m are expressed by moments
' A
m and
' B
m at the ends of the middle elastic region, by using constant shear acting throughout the member;
' '
'
' '
' '
'
' '
( )
( )
(1 )
(1 )
(1 ) (1 )
( )
( )
(1 )
(1 )
(1 ) (1 )
A B
A A A AB
A B AB
B A A B
A B A B
A B
B B B AB
A B AB
B A A B
A B A B
m m
m m L
L
m m
m m
m m L
L
m m
λ
λ λ
λ λ
λ λ λ λ
λ
λ λ
λ λ
λ λ λ λ
+
= +
− −
−
= +
− − − −
+
= +
− −
−
= +
− − − −
In a matrix form,
A' B'
M
B’
M
A’
A
A'
B'
B
θ
A
θ
B
θ
B’
θ
A’
R
A’B’
M
A
M
B
M
A’
M
B’
C
C
D
A AB
L λ (1 )
A B AB
L λ λ − −
B AB
L λ
18
(1 ) '
1
(1 ) ' 1
A B A A
B B A B A B
m m
m m
λ λ
λ λ λ λ
− ¦ ¹ ¦ ¹
=
´ ` ´ `
− − −
¹ ) ¹ )
{ }   { } ' m A m
T
=
The member end rotation and moment stiffness relation of a simple member is expressed as
' '
' '
4 2
2 4
A A
B B
i
i
EI EI
m
L L
m EI EI
L L
θ
θ
¦ ¹ ¦ ¹
=
´ ` ´ `
¹ ) ¹ )
{ }  { } ' ' ' θ k m =
The stiffness matrix of a simple member having rigid ends can be assembled by using the
following relations:
'
'
(1 )
1
(1 ) 1
B B A A
A A B A B B
λ λ θ θ
λ λ θ λ λ θ
− ¦ ¹ ¦ ¹
=
´ ` ´ `
− − −
¹ ) ¹ )
(1 ) '
1
(1 ) ' 1
A B A A
B B A B A B
m m
m m
λ λ
λ λ λ λ
− ¦ ¹ ¦ ¹
=
´ ` ´ `
− − −
¹ ) ¹ )
' '
' '
4 2
2 4
A A
B B
i
i
EI EI
m
L L
m EI EI
L L
θ
θ
¦ ¹ ¦ ¹
=
´ ` ´ `
¹ ) ¹ )
In a symbolic format,
{ }   { }
   { }
    { } θ
θ
A k A
k A
m A m
T
T
T
'
' '
'
=
=
=
19
Home Assignment No. 5
20020306
Otani, S.
Formulate a member stiffness submatrices [k
11
], [k
12
], [k
21
] and [k
22
] for a beam having rigid ends
and rotational springs as shown below. Ignore the axial deformation and shear deformation, and
consider only flexural deformation of the elastic member and rotation of rotational springs. Use
member end forces and displacements in the local coordinate system.
Use the following symbols; E: Young's modulus, I: moment of inertia of prismatic section, L: total
length of member, and L
1
λ : length of rigid zone at starting end, and L
2
λ : length of rigid zone at
terminal end,
1 θ
k : spring constant of the rotational spring at the starting end 1,
2 θ
k : spring constant
of the rotational spring at the terminal end 2.
)
`
¹
¹
´
¦
=
)
`
¹
¹
´
¦
)
`
¹
¹
´
¦
=
)
`
¹
¹
´
¦
=
)
`
¹
¹
´
¦
=
)
`
¹
¹
´
¦
=
2
1
22 21
12 11
2
1
2
2
2
1
1
1
2
2
2
1
1
1
} {
} {
} {
} {
d
d
k k
k k
p
p
d
d
d
d
m
p
p
m
p
p
z
y
z
y
z
y
z
y
θ
θ
Each matrix must be defined.
(a) Calculate
BC
k ] [
22
without the rotational springs at the two ends.
(b) Calculate
BC
k ] [
22
including the rotational springs at the two ends,
(c) Express
AD
k ] [
22
and define all related matrices necessary to calculate
AD
k ] [
22
. Matrix
operation is not required.
(d) Express
AD
k ] [
11
and
AD
k ] [
12
. Define all related matrices necessary to calculate
AD
k ] [
11
and
AD
k ] [
12
. Matrix operation is not required.
1
y
x
2
θ 1
k
θ 2
k
EI
L
1
λ L
2
λ L ) 1 (
2 1
λ λ − −
A B
C
D
20
(1) Stiffness
22
[ ]
BC
k without any rotational spring
3 2
22
2
12 6
[ ]
6 4
BC
BC
BC
BC BC
EI EI
L L
k
EI EI
L L
−
=
−
(2) Calculation of stiffness matrix of the part consisting of two rotational spring and mid elastic
element.
The flexibility matrix
22
[ ]
BC
F of the mid elastic member BC
3 2
1
22 22
2
3 2
[ ] [ ]
2
BC BC
BC BC
BC BC
L L
EI EI
f k
L L
EI EI
−
= =
The flexibility matrix
22 ' '
[ ]
B C
f of the part consisting of the two springs and mid elastic element;
22 ' ' ' 22 ' ' ' 22 ' 22 '
3 2
2
1 2
3 2 2
1 1
2
[ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ]
0 0 0 0
1 0 1
3 2
1 1
0 0 1 0 1
2
3 2
2
T T
B C BC B B BC CC BC CC CC
BC BC
BC
BC
BC BC
BC BC BC BC
BC B
f H f H H f H f
L L
L
EI EI
L
L L
k k
EI EI
L L L L
EI k EI k
L L
EI
θ θ
θ θ
= + +
= + +
+ +
=
+
1 1 2
1 1
C BC
L
k EI k k
θ θ θ
+ +
Inverting the flexibility matrix
22 ' '
[ ]
B C
f , the stiffness matrix of the part consisting of the two springs
and mid elastic element;
2
1 2 1
22 4 3 3 2
2 3 2
2 2
1 2 1 2
1 1
1 1
2
1
[ ]
12 3 3
2
BC BC BC
BC
BC BC BC BC
BC BC BC BC
L L L
EI k k EI k
k
L L L L
L L L L
E I EIk EIk k k
EI k EI k
θ θ θ
θ θ θ θ
θ θ
+ + − −
=
+ + +
− − +
(3) Calculation of stiffness matrix
22
[ ]
AD
k of the entire member with rigid zones at the ends.
1 1
22 22 ' '
[ ] [ ] [ ] ([ ] )
T
AD CD B C CD
k H k H
− −
=
and
1
1
1 0
[ ]
1
1
([ ] )
0 1
CD
CD
CD T
CD
H
L
L
H
−
−
=
−
−
=
Therefore,
21
22 4 3 3 2
2 2
1 2 1 2
1 2 1 2
2 2 2
1 2 1 2
1
[ ]
12 3 3
(2 ) 1 1
2
(2 ) {3 ( ) } ( )
2 3
AD
BC BC BC BC
BC CD BC BC CD BC CD
CD BC BC CD BC CD CD CD BC BC BC CD BC CD
k
L L L L
E I EIk EIk k k
L L L L L L L
EI k k EI k k
L L L L L L L L L L L L L L
EI k k EI k k
θ θ θ θ
θ θ θ θ
θ θ θ θ
=
+ + +
+ +
+ + − − −
×
+ + + + +
− − − + +
(3) Other stiffness submatrices
11
[ ]
AD
k and
12
[ ]
AD
k
11 22
4 3 3 2
2 2
1 2 1 2
1 2 1 2
2
1 2
[ ] [ ][ ] [ ]
1
12 3 3
(2 ) 1 1
2
(2 ) {3 ( 2 ) 6( ) 2 }
2
T
AD AD AD AD
BC BC BC BC
BC CD BC BC CD BC CD
CD BC BC CD BC CD CD CD BC BC
k H k H
L L L L
E I EIk EIk k k
L L L L L L L L L L
EI k k EI k k
L L L L L L L L L L L L L L L L L
EI k k
θ θ θ θ
θ θ θ θ
θ θ
=
=
+ + +
+ − + − −
+ + − − −
×
+ − + − − − + − +
− − −
2 2
1 2
( ) ( )
6
BC CD BC CD
L L L L L
EI k k
θ θ
+ − −
+ +
12 22
4 3 3 2
2 2
1 2 1 2
1 2 1 2
2
1 2
[ ] [ ][ ]
1
12 3 3
(2 ) 1 1
2
(2 ) {6 ( ) 3(2 ) 2 } (
2 6
AD AD AD
BC BC BC BC
BC CD BC BC CD BC CD
CD BC BC CD BC CD CD CD CD BC BC BC CD
k H k
L L L L
E I EIk EIk k k
L L L L L L L L
EI k k EI k k
L L L L L L L L L L L L L L L L
EI k k EI
θ θ θ θ
θ θ θ θ
θ θ
= −
=
+ + +
+ + −
+ + − − −
×
+ + − − + − + +
− − − +
1 2
)( ) ( )
BC CD BC CD CD
L L L L L L L
k k
θ θ
+ − −
+
1
Chapter 10. Member Stiffness Models
10.1 Member Stiffness Model
The modeling of a reinforced concrete structure at the material level, such as the finite element
method analysis, allows the representation of details of the structural geometry and material
properties. In the past decade, a great progress has been made in the field of constitutive modeling
of plain and reinforced concrete under multiaxial loading. Despite this progress, the computational
and memory requirements of such a modeling have restricted its application to the analysis of
individual members or their subassemblage. The efforts to extend the application to the dynamic
analysis of a small reinforced concrete structure have not been successful. Therefore, the
memberbymember modeling is normally used in the nonlinear response analysis of a reinforced
concrete structure.
In the response of a structure in the nonlinear range, piecewise linear response is assumed and
the member stiffness changes with the development of damage along the member. Therefore,
member stiffness matrix needs to be reevaluated and reformulated with a development of new
damage.
The damage distribution within a member
is affected by the distribution of stress within
the member; namely, the stress caused by
gravity loads affects the damage distribution
along the member. This gravity load effect is
important in a building designed with
relatively small earthquake forces; the
yielding may take place in the middle region
of a member. Unfortunately, the effect of
gravity loading cannot be treated in a simple
model at the moment. The effect of gravity
loading is normally ignored in the laboratory
testing of structural members as well as in
the nonlinear response analysis.
Consequently, moment is normally assumed
to distribute linearly along a member.
Although the moment is assumed to
distribute linearly within a member, where
the largest moment occurs at member ends
under earthquakeinduced forces, the
damage (inelastic deformation) does not concentrate at the critical section, but rather spreads along
the member. Many cracks develop in the middle part of a member. Therefore, the distribution of
stiffness within a member needs to be
modeled in the nonlinear analysis. Many
member models have been proposed in the
past to represent the distribution of stiffness
along the member reflecting the
development of damage.
For the formulation of a member stiffness
matrix, let us consider a simply supported
member AB of any stiffness distribution within the member. The relation between incremental
member end forces and deformations needs to be formulated; i.e., the relation between incremental
member end moments,
A
m ∆ and
B
m ∆ , axial force,
AB
p ∆ , and incremental member end rotations,
A
θ ∆ and
B
θ ∆ , axial deformation,
AB
e ∆ , at member ends. In a matrix form, the elements of the
instantaneous member stiffness matrix must be evaluated:
Bending Moment
under Gravity
Loads
Bending Moment
under
Earthquake
Bending Moment
under combined
Gravity and
Earthquake Loads
2
i
B
A
AB
i i
B
A
AB
e
k k k
k k k
k k k
m
m
p
¦
)
¦
`
¹
¦
¹
¦
´
¦
∆
∆
∆
=
¦
)
¦
`
¹
¦
¹
¦
´
¦
∆
∆
∆
θ
θ
33 32 31
23 22 21
13 12 11
In a linearly elastic structural analysis, a
prismatic member, such as a beam or column, is
represented by a straight element passing
through "the geometrical centroid" of the section
because the longitudinal strain under bending is zero at the geometrical centroid. Therefore, a
member does not develop axial deformation under bending moment acting along the member, nor
an axial force causes any member end rotation. In other words, there is no interaction between axial
and rotational response in a linearly elastic stage.
With flexural cracks forming along a reinforced concrete member due to bending, the neutral axis
shifts from the geometrical centroid into the compression side. The tensile strain is developed at the
geometric centroid of previously uncracked section, causing an elongation of the member measured
at the centroid of the section. Such elongation of a member is also measured in member tests under
bending in the laboratory. This phenomenon should be recognized. Very few model recognizes the
interaction of axial and bending response.
Strain before
Cracking
Strain after
Cracking
Section
In a nonlinear analysis of a frame structure, however, this axial and rotational interaction is
normally ignored. Therefore, member end momentrotation relation and member end axial
forceelongation relation are treated separately; i.e., it is assumed that k k
12 21
0 = = and
k k
13 31
0 = = .
Once the momentaxial force interaction is ignored, stiffness element k
11
represents the axial
forcedeformation relation of the reinforced concrete
member under uniaxial loading and is determined
from the axial forceaxial deformation relationship
observed in uniaxial tests. The relationship in tension
is linear up to initial tensile cracking, and then the
stiffness gradually deteriorates with additional
cracking, followed by yielding of all longitudinal
reinforcement. The relationship in compression is
similar to that of concrete in compression. The axial
stressstrain relation is normally considered to be
linear for a practical range of analysis, especially in
the analysis of lowrise buildings, where axial stress
due to the overturning effect of earthquake forces is relatively small.
If the interaction of bending moment and axial force interaction, the member end moment and
A
m ∆
B
m ∆
AB
p ∆
1 x
p ∆
2 x
p ∆
1 y
p ∆
2 y
p ∆
1 z
m ∆
2 z
m ∆
Yielding
Strain
Stress
Ccompresson
Tension
Yielding
3
rotation relation should be defined; i.e., stiffness elements
22
k ,
23
k ,
32
k and
33
k must be defined.
The symmetry of a stiffness matrix (the reciprocal theorem) gives
32 23
k k = , and three independent
stiffness elements
22
k ,
23
k and
33
k must be determined. The test of reinforced concrete members
is carried out under a prescribed loading history. It is not possible to vary the ratio of member end
moments because the combination is infinite. Therefore, the member end moments are chosen to be
same in the test, and the antisymmetric bending moment distribution is developed in a specimen
with the inflection point at the midspan. The member end moment and rotation relation may be
determined to define diagonal elements
22
k and
33
k , but the offdiagonal element
23
k may not be
defined. A member model is necessary to define the member stiffness matrix.
Rotation
M
o
m
e
n
t
Cracking
Yielding
m
A
m
B
=m
A
A
θ
B
θ
AB
e
p
AB
Member end momentrotation relation from tests
4
10.2 Fiber Model
A member may be divided into short segments, and boundary section between adjacent short
segments may be divided into small element, where the Bernoulli’s assumption of plane section
remaining plane after deformation may be used as kinematic constraints. Rotations about two axis
and elongation at the centroid may express the degrees of freedom at the section. Such a model is
generally called a fiber model and useful to represent flexural behavior of a member.
The fiber model may be considered as a simple finite element method applied to the
onedimensional continuum. A member coordinate system may consist of xaxis in the direction of
the member, yaxis in the vertical direction, and zaxis in the horizontal direction. A section is divided
into small elements (filaments or fibers) by lines parallel to the zaxis and yaxis. Each fiber
represents either concrete or steel reinforcement. Nonlinear uniaxial stressstrain relation of the
material is assigned to the center of each fiber. Shear deformation is normally ignored in this
formulation.
Stress and Strain in Section: Plane section before bending
is assumed to remain plane after bending. The normal tensile
strain ( , ) y z ε at ( , ) y z in the section is expressed by the
tensile strain
0
ε at the geometric centroid ( 0, 0) y z = =
and curvatures
y
φ and
z
φ (positive counterclockwise
rotation about the coordinate axis) about the centroidal axes y
and z;
0
( , ) ( ) ( ) ( )
x z y
y z x y x z x ε ε φ φ = − +
The incremental tensile stress ( , )
x
y z σ ∆ in the fiber is
evaluated by using tangent modulus of elasticity ( , )
t
x
E y z of
the material at coordinate ( , y z ) at section x ;
( , ) ( , ) ( , )
t
x x x
y z E y z y z σ ε ∆ = ⋅ ∆
Stiffness of Section: The incremental tensile force
( ) n x ∆ of section at x is evaluated by summing the
incremental tensile stress ( , ) y z σ ∆ over the section;
x
y
z
y
z
Coordinate System of Fiber Model
y
z
y
z
0
Coordinate System
σ
ε
E’
ε ∆
σ ∆
5
0
0
0
( ) ( , )
( , ) ( , )
( , ) { ( ) ( ) ( )}
( , ) ( ) ( , ) ( ) ( , ) ( )
( ) ( , )
x
Section
t
x x
Section
t
x z y
Section
t t t
x x z x y
Section Section Section
t
x
Section
n x y z dA
E y z y z dA
E y z x y x z x dA
E y z x dA E y z y x dA E y z z x dA
x E y z dA
σ
ε
ε φ φ
ε φ φ
ε
∆ = ∆
= ⋅ ∆
= ⋅ ∆ − ∆ + ∆
= ⋅ ∆ − ∆ + ∆
= ∆ −
∫
∫
∫
∫ ∫ ∫
∫
( ) ( , ) ( ) ( , )
t t
z x y x
Section Section
x E y z ydA x E y z z dA φ φ ∆ + ∆
∫ ∫
Similarly, the incremental bending moment about yaxis is calculated by summing up the
contribution of fiber stresses to the moment;
0
2
0
0
( ) ( , )
( , ) ( , )
( , ) { ( ) ( ) ( )}
( , ) ( ) ( , ) ( ) ( , ) ( )
( ) ( , )
y x
Section
t
x
Section
t
x z y
Section
t t t
x x z x y
Section Section Section
t
x
m x y z zdA
E y z y z zdA
E y z x y x z x zdA
E y z z x dA E y z yz x dA E y z z x dA
x E y z z
σ
ε
ε φ φ
ε φ φ
ε
∆ = ∆
= ⋅ ∆ ⋅
= ⋅ ∆ − ∆ + ∆ ⋅
= ⋅ ∆ − ⋅ ∆ + ∆
= ∆ ⋅
∫
∫
∫
∫ ∫ ∫
2
( ) ( , ) ( ) ( , )
t t
z x y x
Section Section Section
dA x E y z yzdA x E y z z dA φ φ − ∆ ⋅ + ∆
∫ ∫ ∫
The incremental moment about zaxis is evaluated by summing up the moment contributions of
stresses in fiber elements;
0
2
0
0
( ) ( , )
( , ) ( , )
( , ) { ( ) ( ) ( )}
( , ) ( ) ( , ) ( ) ( , ) ( )
( ) (
z x
Section
t
x
Section
t
x z y
Section
t t t
x x z x y
Section Section Section
t
x
m x y z ydA
E y z y z ydA
E y z x y x z x ydA
E y z y x dA E y z y x dA E y z yz x dA
x E
σ
ε
ε φ φ
ε φ φ
ε
∆ = − ∆
= − ⋅ ∆ ⋅
= − ⋅ ∆ − ∆ + ∆ ⋅
= − ⋅ ∆ + ⋅ ∆ − ⋅ ∆
= −∆
∫
∫
∫
∫ ∫ ∫
2
, ) ( ) ( , ) ( ) ( , )
t t
z x y x
Section Section Section
y z ydA x E y z y dA x E y z yz dA φ φ ⋅ + ∆ ⋅ − ∆
∫ ∫ ∫
The instantaneous stiffness matrix [ ( )]
s
k x relates the incremental internal forces { ( )}
s
s x ∆ and
incremental strains { ( )} x ε ∆ ;
{ ( )} { ( ), ( ), ( )}
T
s y z
s x m x m x n x ∆ = ∆ ∆ ∆
0
{ ( )} { ( ), ( ), ( )}
T
s y z
x x x x ε φ φ ε ∆ = ∆ ∆ ∆ ;
{ ( )} [ ( )]{ ( )}
s s
s x k x x ε ∆ = ∆
6
2
2
[ ( )]
t t t
x x x
t t t
s x x x
t t t
x x x
E z dA E yzdA E zdA
k x E yzdA E y dA E ydA
E zdA E ydA E dA
⋅ − ⋅ ⋅
= − ⋅ ⋅ − ⋅
⋅ − ⋅
∫ ∫ ∫
∫ ∫ ∫
∫ ∫ ∫
For uniaxial bending with ( ) 0
y
m x = , the stiffness relation is reduced to 2x2.
2
0
( ) ( )
( ) ( )
t t
x x
z z
t t
x x
E y dA E ydA
x m x
x n x
E ydA E dA
φ
ε
⋅ − ⋅
∆ ∆ ¦ ¹ ¦ ¹
=
´ ` ´ `
∆ ∆
¹ ) − ⋅ ¹ )
∫ ∫
∫ ∫
Actual evaluation of section stiffness is carried out not by integration but by summing up the
contribution from small fiber segments. Zeris and Mahin (1988) pointed out that this formulation
sometimes causes a numerical problem once the maximum section capacity is reached. An iterative
approach is suggested.
The section stiffness may be defined by a hysteresis model once skeleton momentcurvature
relation is estimated under monotonically increasing curvature.
Member Stiffness Matrix: The
principle can be discussed,
without loosing generality, using
a twodimensional plane frame
member under uniaxial bending
and deformation. A simple
beam is considered as a basic
statically determinate system.
The tangent stiffness matrix of the member relates the member end forces and displacements
without intermediate loading.
¦
)
¦
`
¹
¦
¹
¦
´
¦
∆
∆
∆
=
¦
)
¦
`
¹
¦
¹
¦
´
¦
∆
∆
∆
AB
B
A
t
m
AB
B
A
e
k
p
m
m
θ
θ
] [
(1) Stiffness Approach
Member stiffness matrix [ ]
t
m
k is evaluated by integrating the tangent section stiffness matrix
[ ]
t
s
k along the length of the member;
0
[ ] [ ( )] [ ][ ( )]
L
t T t
m s
k B x k B x dx =
∫
where [ ( )] B x : matrix relating the section generalized strain increment vector to the member
deformation increment vector;
{ ( )} [ ( )]{ } x B x ε θ ∆ = ∆
For a linearly elastic prismatic member, the transverse incremental displacement is cubic
polynomials of the distance along the member axis, and matrix [ ( )] B x may be evaluated as
0 2(3 2) 2(3 1) 1
[ ( )]
0
0 0
x x
B x
− −
=
l l
l
For nonuniform distribution of stiffness, the evaluation of matrix [ ( )] B x is the critical problem.
A B
,
A A
m θ ∆ ∆
,
B B
m θ ∆ ∆
x
y
AB AB
e p ∆ ∆ ,
7
The main shortcoming of the stiffnessbased elements is their inability to represent member
behavior near the peak resistance since there exists numerical instability problem. Therefore, the
member may be subdivided into short segments, and simple flexibility distribution should be
assumed for each short segments. Mahasuverachai and Powell (1982) suggested the use of
flexibilitydependent shape functions that are continuously updated during the analysis.
(2) Flexibility Approach
For a statically determinate member such as a cantilever or a simple beam, section forces (e.g.,
axial force and bending moment) { ( )} s x along a member are precisely defined as a function of
member end forces { } Q .
{ ( )} [ ( )]{ }
Q
s x N x Q =
where [ ( )]
Q
N x : force interpolation functions which define section forces as a function of member
end forces. For a simply supported plane frame member, matrix [ ( )]
Q
N x is expressed as
0 1
[ ]
1
0 0
Q
x x
N
−
=
l l
A section constitutive law is written in the incremental form at each section;
{ ( )} [ ( )]{ ( )} x f x s x ε ∆ = ∆
where [ ( )] f x : section flexibility matrix and { ( )} x ε : section deformations.
The compatibility equation is expressed in the following form;
[ ]{ } { } F Q U ∆ = ∆
where { } U : nodal displacement of the statically determinate member, and flexibility matrix [ ] F is
defined as
0
[ ] [ ( )] { ( )}[ ( )]
L
T
F N x f x N x dx =
∫
The flexibility matrix of a member is evaluated in discretized form. Therefore, the compatibility
between section { ( )} e x and member end displacement { } U is maintained in an integral sense
(Coleman and Spacone, 2001);
0
{ } [ ( )]{ ( )}
L
Q
U N x e x dx =
∫
The draw back of the flexibilitybased formulation is the implementation the flexibility relation in
existing stiffnessbased analysis procedure.
The drawback of the fiber model is the difficulty in evaluating member tangent stiffness matrix.
The memory requirement is also significant to keep track of stress and strain levels in each fiber in
the analysis.
References:
Menegotto, M., and P. E. Pinto, “Method of Analysis for Cyclically Loaded RC Plane Frames
Including Changes in Geometry and Nonelastic Behaviour of Elements under Combined Normal
Force and Bending,” Preliminary Report, IABSSE, 1973, Vol. 13, pp. 15  22.
Aktan, A. E., et al., “R/C Column Earthquake Response in Two Dimensions,” Journal, Structures
Division, ASCE, Vol. 100, No. ST10, October 1974, 1999  2015.
8
Aziz, T. S., “Inelastic Dynamic Analysis of Building Frames,” Research Report R7637, Department
of Civil Engineering, Massachusetts Institute of Technology, Cambridge, MA, 1976.
Coleman, J., and E. Spacone, “Localization Issues in Forcebased Frame Elements,” Journal,
Structural Engineering, ASCE, Vol. 127, No. 11, November 2001, pp. 1257  1265.
Kaba, S. and S. A. Mahin, “Refined Modeling of Reinforced Concrete Columns for Seismic Analysis,”
EERC Report 84/03, Earthquake Engineering Research Center, University of California, Berkeley,
1984.
Mahasuverachai, M., and G. H. Powell, “Inelastic Analysis of Piping and Tubular Structures,” EERC
Report 8227, Earthquake Engineering Research Center, University of California, Berkeley, 1982.
Mark, K., “Nonlinear Dynamic Response of Reinforced Concrete Frames,” Research Report R7638,
Department of Civil Engineering, Massachusetts Institute of Technology, Cambridge, MA, 1976.
Spacone, E., F. C. Filippou and F. F Taucer, “Fiber Beamcolumn Model for Nonlinear Analysis of
R/C Frames: Part I. Formulation,” Earthquake Engineering and Structural Dynamics, Vol. 25,
1996, pp. 711  725.
Spacone, E., F. C. Filippou and F. F Taucer, “Fiber Beamcolumn Model for Nonlinear Analysis of
R/C Frames: Part II. Applications,” Earthquake Engineering and Structural Dynamics, Vol. 25,
1996, pp. 727  742.
Zeris, C. and S. A. Mahin, “Analysis of Reinforced Concrete Beamcolumns under uniaxial excitation,
Journal, Structures Division, ASCE, Vol. 114, No. ST4, April 1988, pp. 804  820.
Zeris, C., and S. A. Mahin, “Behavior of Reinforced Concrete Structures subjected to Biaxial
Excitation,” Journal, Structures Division, ASCE, Vol. 117, No. ST9, September 1991, pp. 2657 
2673.
9
10.3 Discrete Element Models
In order to overcome difficult problems of variable stiffness along a member, the member can be
subdivided into short line segments along the length, with each short segment assigned nonlinear
hysteretic characteristics. The nonlinear stiffness can be assigned within a segment, or to flexible
springs at the connection of two adjacent segments.
Discrete spring model Discrete segment model
Discrete Segment Model: A member is divided into short segments, each segment i with uniform
flexural rigidity
i
EI that varies with a stress history of the segment. Variation of stiffness along the
member can be easily handled by this model if the flexural rigidity can be estimated for each short
segment. Structural walls are often idealized by this model although the interaction of axial
deformation and flexure cannot be considered in this model; i.e., the elongation due to the shift of
neutral axis in the section cannot be considered. More computation effort is required in this model
due to the increased number of degrees of freedom.
A member is subdivided into (n+1)
elements; element i has constant flexural
rigidity
i
EI and length
i
x ∆ . A common
local coordinate system may be chosen for
the entire member. The stiffness matrix for
each element using the common local
coordinate system is expressed as
i i i
d
d
k k
k k
p
p
)
`
¹
¹
´
¦
∆
∆
=
)
`
¹
¹
´
¦
∆
∆
2
1
22 21
12 11
2
1
where
 
∆ ∆
∆ ∆
∆
=
i
i
i
i
i
i
i
i
i
i
i
x
EI
x
EI
x
EI
x
EI
x
EA
k
4
) (
6
0
) (
6
) (
12
0
0 0
2
2 3 11
   
∆ ∆
−
∆ ∆
−
∆
−
= =
i
i
i
i
i
i
i
i
i
i
T
i i
x
EI
x
EI
x
EI
x
EI
x
EA
k k
2
) (
6
0
) (
6
) (
12
0
0 0
2
2 3 21 12
A B
1 2
i
n
EI
i
Discrete Segment Model
Flexural Rigidity EI
Rotational Spring
Rigid Element
x
y
1
2 i
2 z
m ∆
2 y
p ∆
2 x
p ∆
EI
i
, EA
i
1 x
p ∆
1 y
p ∆
1 z
m ∆
i
x ∆
10
 
∆ ∆
−
∆
−
∆
∆
=
i
i
i
i
i
i
i
i
i
i
i
x
EI
x
EI
x
EI
x
EI
x
EA
k
4
) (
6
0
) (
6
) (
12
0
0 0
2
2 3 22
At each node, two elements are connected. Using the
continuity condition of element end displacement at the
connection and the equilibrium condition of member end
forces at each node, a stiffness matrix of the member can be
formulated;
} ]{ [ } { d k p ∆ = ∆
¦
¦
¦
¦
)
¦
¦
¦
¦
`
¹
¦
¦
¦
¦
¹
¦
¦
¦
¦
´
¦
∆
∆
⋅
∆
⋅
∆
∆
+ ⋅ ⋅
⋅ ⋅ ⋅ ⋅ ⋅ ⋅
⋅ ⋅ ⋅ + ⋅
⋅ ⋅ ⋅
⋅ +
⋅ ⋅
=
¦
¦
¦
¦
)
¦
¦
¦
¦
`
¹
¦
¦
¦
¦
¹
¦
¦
¦
¦
´
¦
∆
∆
⋅
∆
⋅
∆
∆
−
+ −
B
n
i
A
nB nB
nB nB n n
i i i i
A A
A A
B
n
i
A
d
d
d
d
d
k k
k k k
k k
k
k k k k
k k
p
p
p
p
p
1
22 21
12 11
, 1
22
1 ,
11
, 1
22
12
21
12
12
12
11
1
22
1
21
1
12
1
11
1
0 0 0 0 0
0 0 0
0
0 0 0
0 0 0
0 0 0
0 0 0
The linear equation can be rearranged in the form
)
`
¹
¹
´
¦
∆
∆
=
)
`
¹
¹
´
¦
∆
∆
i
e
ii ie
ei ee
i
e
d
d
k k
k k
p
p
in which, { }
e
p ∆ and { }
e
d ∆ are forces and displacements at member ends A and B, { }
i
p ∆ and
{ }
i
d ∆ are forces and displacements at intermediate nodes from 1 to n.
¦
)
¦
`
¹
¦
¹
¦
´
¦
∆
∆
= ∆
¦
)
¦
`
¹
¦
¹
¦
´
¦
∆
∆
= ∆
)
`
¹
¹
´
¦
∆
∆
= ∆
)
`
¹
¹
´
¦
∆
∆
= ∆
n
i
n
i
B
A
e
B
A
e
d
d
d
p
p
p
d
d
d
p
p
p
M M
1 1
} { } {
} { } {
If intermediate loads are not considered along a member, the sum of internal forces should be
zero at intermediate joints;
{ } { } 0 = ∆
i
p
hence,
{ }    { }
e ie ii i
d k k d ∆ − = ∆
−1
and
{ }        { }
 { }
e
e ie ii ei ee e
d k
d k k k k p
∆ =
∆ − = ∆
−
) (
1
2 2
{ } ,{ }
i i
p d ∆ ∆
Member i Member i+1
1 1 1 1
{ } ,{ }
i i
p d
+ +
∆ ∆
11
This procedure is called "static condensation." A reduced stiffness matrix of 3 x 3 is obtained.
Incremental displacement { }
i
d ∆ at internal joints is calculated after the end incremental
displacement { }
e
d ∆ is obtained:
{ }    { }
e ie ii i
d k k d ∆ − = ∆
−1
Correction of Unbalanced Forces: The
structural analysis is based on (a) constitutive
relation of members, (b) equilibrium of forces
at joints, and (c) compatibility of displacement
at joints. It is generally assumed that the
stiffness does not change during a small load
(displacement) increment. However, this
assumption is often violated and the
forcedeformation relationship deviates from
the linear instantaneous stiffness. If the
constitutive relation is to be satisfied, either
equilibrium or compatibility must be violated
at a joint at the next load (displacement)
increment.
It is normally advisable to correct the
element force in accordance with the
constitutive relation because the violation of
equilibrium can be corrected easily by applying an imaginary external force at the joint.
Due to the unbalanced force { }
*
i
p ∆ at intermediate joints within a member and { }
*
e
p ∆ at the
ends, the static condensation must be performed in a more complicated manner;
)
`
¹
¹
´
¦
∆
∆
−
)
`
¹
¹
´
¦
∆
∆
=
)
`
¹
¹
´
¦
∆
∆
*
*
i
e
i
e
ii ie
ei ee
i
e
p
p
d
d
k k
k k
p
p
and { } { } 0 = ∆
i
p because there acts no external load at the internal nodes. Hence,
{ }  { }  { }
{ }   { }  { } ( )
e ie i ii i
i ii e ie i
d k p k d
d k d k p
∆ − ∆ = ∆
∆ + ∆ = ∆
− * 1
*
and,
{ }  { }  { } { }
       { } { }    { }
* 1 * 1
*
) (
i ii ei e e ie ii ei ee
e i ei e ee e
p k k p d k k k k
p d k d k p
∆ + ∆ − ∆ − =
∆ − ∆ + ∆ = ∆
− −
Application of Unit Load Method: If there are no external loads acting at internal nodes, the
stiffness matrix of a member can be formulated in a simpler manner using unit load method of
calculating displacement under a given loading set.
Suppose a member AB of length L is simply supported at the two ends subjected to incremental
member end moments
A
m ∆ and
B
m ∆ . The member is divided into n segments, not necessarily of
equal length. Length of segment i is
i
x ∆ , flexural rigidity
i
EI , and distance from A end to the
center of segment i is x
i
.
Incremental bending moment
x
m ∆ (positive for tension at bottom) at distance x is expressed as
D
F
D
j
D
j+1
F
j
F*
j+1
F
j+1
k
j
Correction of overshooting
Correction of Overshooting in Hysteresis Relation
1 +
∆
j
D
1 +
∆
j
F
*
1 +
∆
j
p
*
1 1 + +
∆ − ∆ = ∆
j j j j
p D k F
12
) ( ) 1 (
L
x
m
L
x
m m
B A x
∆ + − ∆ − = ∆
and incremental curvature
x
φ ∆ at the
center of segment i is
)
2
( )
2
(
i
i
i
i
i
x
x
x
x x
x
x for
EI
m
∆
+ ≤ ≤
∆
−
∆
= ∆φ
For a unit moment applied at A end,
bending moment
ux
m at distance x is
given as
L
x
m
ux
+ − = 1
Incremental member end rotation
A
θ ∆ at
A end is calculated by unit load method:
dx m
dx m
ux
i
x x
x x
x
ux
L
x A
i i
i i
∑
∫
∫
∆ +
∆ −
∆ =
∆ = ∆
2 /
2 /
0
φ
φ θ
B ABi A AAi
x x
x x
i
B
x x
x x
i
A
x x
x x
i B A ux
x x
x x
x
m f m f
dx
L
x
L
x
EI
m
dx
L
x
EI
m
dx EI
L
x
m
L
x
m
L
x
dx m
i i
i i
i i
i i
i i
i i
i i
i i
∆ + ∆ =
−
∆
− −
∆
=
∆ + − ∆ + − = ∆
∫ ∫
∫ ∫
∆ +
∆ −
∆ +
∆ −
∆ +
∆ −
∆ +
∆ −
2 /
2 /
2 /
2 /
2
2 /
2 /
2 /
2 /
) 1 )( ( ) 1 (
/ )} ( ) 1 ( ){ 1 ( φ
1
A
m =
A
B
i
x
x
i
m
EI
φ
∆
∆ =
i
x
i
x ∆
ux
m
Calculation of member end rotation
Flexural Rigidity EI
B
m ∆
A
m ∆
x
x
m ∆
L
13
} ) (
12
1
) {( ) (
1
) 1 )( (
1
} ) (
12
1
) 1 {( ) 1 (
1
2 2
2 /
2 /
2
2 /
2 /
2 2
2 /
2 /
2
L
x
L
x
EI
x
dx
L
x
EI
f
L
x
EI
x
f dx
L
x
L
x
EI
f
L
x
L
x
EI
x
dx
L
x
EI
f
i i
i
i
x x
x x
i
BBi
i
i
i
BBi
x x
x x
i
ABi
i i
i
i
x x
x x
i
AAi
i i
i i
i i
i i
i i
i i
∆
+
∆
= =
∆
+ = − − =
∆
+ −
∆
= − =
∫
∫
∫
∆ +
∆ −
∆ +
∆ −
∆ +
∆ −
B AB A AA
B
i
ABi A
i
AAi A
m f m f
m f m f
∆ + ∆ =
∆ + ∆ = ∆
∑ ∑
) ( ) ( θ
Similarly, applying unit moment at end B, bending moment
ux
m at distance x from end A;
L
x
m
ux
=
and incremental rotation
B
θ ∆ at member end B is calculated using the unit load method:
B BB A BA
i
B BBi A
i
BAi
ux
i
x x
x x
x
ux
L
o
x B
m f m f
m f m f
dx m
dx m
i i
i i
∆ + ∆ =
+ =
∆ =
∆ = ∆
∑ ∑
∑
∫
∫
∆ +
∆ −
) ( ) (
2 /
2 /
φ
φ θ
where,
ABi BAi
f f =
)
`
¹
¹
´
¦
∆
∆
=
)
`
¹
¹
´
¦
∆
∆
B
A
BB BA
AB AA
B
A
m
m
f f
f f
θ
θ
The stiffness relation at the member end can be expressed by inverting the flexibility relation:
)
`
¹
¹
´
¦
∆
∆
=
)
`
¹
¹
´
¦
∆
∆
B
A
BB BA
AB AA
B
A
k k
k k
m
m
θ
θ
in which
2
2
2
AB BB AA
AA
BB
AB BB AA
AB
BA AB
AB BB AA
BB
AA
f f f
f
k
f f f
f
k k
f f f
f
k
−
=
−
− = =
−
=
The number of degrees of freedom of a member is normally reduce to three (two member end
rotations and an extension) using the static condensation technique so that the size of a structural
stiffness matrix is kept small.
Discrete Spring Model: Wen and Janssen (1965) presented a method of analysis for a plane frame
14
consisting of elastoplastic segments. The mass and flexibility of a member were lumped at the
connecting points on a tributary basis. Powell (1975) suggested the use of a degrading stiffness
hysteresis model for rigid inelastic connecting springs. Shorter segments were recommended in a
region of high bending moment, and longer segments in a low bending moment region.
Flexibility relation of a simply supported
member consisting of short rigid segments
and rotational springs at internal joints is
expressed as;
)
`
¹
¹
´
¦
∆
∆
=
)
`
¹
¹
´
¦
∆
∆
B
A
BB BA
AB AA
B
A
m
m
f f
f f
θ
θ
in which
i
i
i
BB
i
i
i i
BA AB
i
i
i
AA
f
L
x
f
f
L
x
L
x
f f
f
L
x
f
2
2
) (
) 1 )( (
) 1 (
∑
∑
∑
=
− = =
− =
where,
i
x : distance of spring i from A end,
L: length of member,
i
f : rotational flexibility of spring i. The stiffness relation can be obtained by
inverting the flexibility relation.
References:
Kaba, S. and S. A. Mahin, “Refined modeling of reinforced concrete columns for seismic analysis,”
Earthquake Engineering Research Center, University of California at Berkeley, EERC Report
843, 1984.
Powell, G. H., "Supplement to Computer Program DRAIN2D, Supplement to Report, DRAIN2D
User's Guide, University of California, Berkeley, 1975.
Wen, R. K., and J. G. Janssen, "Dynamic Analysis of ElastoInelastic Frames," Proceedings, Third
World Conference on Earthquake Engineering, Wellington, New Zealand, January 1965, Vol. II,
pp. 713729.
Zeris, C. and S.A. Mahin, “Analysis of Reinforced Concrete Beamcolumns under Uniaxial
Excitation,” Journal of Structural Engineering, ASCE, Vol. 114, No. 4, April 1988, pp. 804  820.
Zeris, C., and S.A. Mahin, “Behavior of Reinforced Concrete Structures subjected to Biaxial
Excitation,” Journal of Structural Engineering, ASCE, Vol. 117, No. 9, September 1991, pp.
2657  2673.
Rotational Spring
Rigid Element
x
y
L
x
i
spring i
m
A
m(x
i
)
m
B
15
10.4 Onecomponent Model
An elastoplastic frame structure was often analyzed by introducing a plastic hinge at the location
where the bending moment reached the plastic moment (Berg and DaDeppo, 1960). All plastic
deformation was assumed to occur in the hinge and no strain hardening was considered.
This model, called onecomponent model, was extended for the general use in a nonlinear frame
analysis under earthquake excitation by Giberson (1967); he used nonlinear rotational springs at two
member ends instead of rigidplastic hinges. All the inelastic deformation of a member was assumed
to concentrate at the member ends, and the middle part was assumed to remain elastic.
If the stiffness properties of each spring can be
evaluated from the material properties and
geometry of the member, the flexibility matrix of a
simply supported member can be formulated by
considering the member end rotation as the sum of
an elastic element rotation and spring rotation at
each end:
Bp Be B
Ap Ae A
θ θ θ
θ θ θ
∆ + ∆ = ∆
∆ + ∆ = ∆
For a simply supported elastic element:
)
`
¹
¹
´
¦
∆
∆
−
−
=
)
`
¹
¹
´
¦
∆
∆
B
A
Be
Ae
m
m
EI
L
EI
L
EI
L
EI
L
3 6
6 3
θ
θ
in which EI: flexural rigidity of section in the elastic region, L: member length. For each inelastic
rotational spring:
A
B C
D
Elastic Response
P
A
B
C
D
Inelastic Response
P
One Component Model
m
A
m
B
A
θ
A
B
k
A
k
B
Elastic Element
B
θ
L
θ
m
Elastic
Deformation
Spring Rotation
Member End MomentRotation Relation
16
B
B
Bp
A
A
Ap
k
m
k
m
∆
= ∆
∆
= ∆
θ
θ
where k
A
and k
B
: instantaneous (tangent) spring constant of rotational springs at A end and B end,
respectively. Therefore, the flexibility relation is written in the form;
)
`
¹
¹
´
¦
∆
∆
+ −
− +
=
)
`
¹
¹
´
¦
∆
∆
B
A
B
A
B
A
m
m
k EI
L
EI
L
EI
L
k EI
L
1
3 6
6
1
3
θ
θ
Let
A A
k
L
EI
s /
6
= , and
B B
k
L
EI
s /
6
= ,
)
`
¹
¹
´
¦
∆
∆
+ −
− +
=
)
`
¹
¹
´
¦
∆
∆
b
A
B
A
B
A
m
m
s
s
EI
L
2 1
1 2
6 θ
θ
The flexibility relation can be inverted to the stiffness relation; i.e.,
)
`
¹
¹
´
¦
∆
∆
+
+
− + +
=
)
`
¹
¹
´
¦
∆
∆
B
A
A
B
B A B
A
s
s
s s
L EI
m
m
θ
θ
2 1
1 2
1 ) 2 )( 2 (
) / 6 (
The spring properties may be determined by assuming the length of a yield hinge zone and a
uniform curvature distribution over the hinge region.
In many cases, however, it becomes more desirable to
faithfully simulate a member end momentrotation relation
of a member observed in the laboratory under a
prescribed loading condition. Such a relation is idealized
and given by a hysteresis model. Therefore, it is more
convenient to use an instantaneous stiffness
'
A
k and
'
B
k
of the member end momentrotation relation, in which
member end rotation includes elastic rotation (=
L EI m
A
/ 6 ∆ ) at the member end under antisymmetric
moment distribution; consider
B A
m m = in the flexibility
relation above,
1
( ) ( )
3 6
1
( )
6
A A B A
A
A
A
L L
m m m
EI k EI
L
m
EI k
θ ∆ = + ∆ − ∆ = ∆
= + ∆
This relation is now expressed as
A
A
A
m
k
∆ = ∆ )
1
(
'
θ
Therefore,
A A
k EI
L
k
1
6
1
'
+ =
'
A
A
A
m
k
θ
∆
∆ =
A
m ∆
A
m
EI
L
∆
6
Member End MomentRotation Relation
17
or
A A
s s + =1 ' , where,
A A
k
L
EI
s ' /
6
' = and
A A
k
L
EI
s ' /
6
' = .
This relation can be substituted in the stiffness relation above:
)
`
¹
¹
´
¦
∆
∆
+
+
− + +
=
)
`
¹
¹
´
¦
∆
∆
B
A
A
B
B A B
A
s
s
s s
L
EI
m
m
θ
θ
'
'
' '
1 1
1 1
1 ) 1 )( 1 (
6
A major advantage of the model is that inelastic memberend deformation depends solely on the
moment acting at the end so that any momentrotation hysteresis relation can be assigned to the
spring. The stiffness of an inelastic spring is normally defined by assuming an antisymmetric
moment distribution along a member with the inflection point at midspan.
This fact is also a weakness of the model because the memberend rotation should be dependent
on the curvature distribution along the member, hence dependent on moments at both member ends.
Consider two cases of moment distribution along a member AB with corresponding to a curvature
distribution shown below;
The inelastic rotations at the A end are given by
shaded areas. For the same moments at A end,
Case II causes larger inelastic rotation at A end.
Consequently, this simple model does not simulate
actual member behavior if the member moment
distribution changes significantly during an
earthquake. Furthermore, it is not rational to lump all
inelastic deformation of a reinforced concrete
member at member ends.
Therefore, the momentrotation relation at a
member end using this member model tends to
deviate from the actual relation if the stress
distribution becomes different from that assumed in
determining the spring properties.
Suko and Adams (1971) suggested the use of the
initial location of the inflection point in evaluating spring properties, assuming the inflection point
during an earthquake does not shift much from the initial elastic location. However, once yielding is
developed at one member end, the moment at the other end must increase to resist a higher stress,
moving the inflection point toward the member center. At the same time, a large concentrated
rotation starts to occur near the critical section.
Despite rational criticisms against this simple model, the performance of the onecomponent
model is expected to be reasonably good for a relatively lowrise frame structure, in which the
inflection point of a column locates reasonable close to midheight.
A finite size of the plastic regions may be considered in the analysis (Roufaiel and Meyer, 1987).
References:
Berg, G. V., and D. A. DaDeppo, "Dynamic Analysis of ElastoPlastic Structures," Proceedings,
Engineering Mechanics Division, ASCE, Vol. 86, EM2, April 1960, pp. 3558.
Giberson, M. F., "The Response of Nonlinear Multistory Structures subjected to Earthquake
Excitation," EERL Report, Earthquake Engineering Research Laboratory, California Institute of
Technology, Pasadena, 1967.
Roufaiel, M.S.L., and C. Meyer, “Analytical Modeling of Hysteretic Behavior of R/C Frames,” Journal
Inelastic memberend rotation
for different moment distributions
18
of Structural Engineering, ASCE, Vol. 113, No. 3, March 1987, pp. 429444.
Suko, M., and P. F. Adams, "Dynamic Analysis of Multibay Multistory Frames," Journal, Structural
Division, ASCE, Vol. 97, No. ST10, October 1971, pp. 25192533.
19
10.5 Multicomponent Model
In an effort to analyze frame structures well
into the inelastic range under earthquake
excitation, Clough et al. (1965) proposed a
twocomponent model: a frame member was
divided into two imaginary parallel elements: an
elastoplastic element to represent a yielding
phenomenon, and a fully elastic element to
representing strainhardening behavior. The sum
of the flexural rigidity of the two elements was
equal to the initial elastic flexural rigidity, EI. The
two elements were rigidly connected at the two
member ends, and the member endmoment
was equal to the sum of the moments at the
element ends. When a memberend moment
reached the yield level, a plastic hinge was
inserted at the end of the elastoplastic element.
Upon unloading from a peak, the plastic hinge
was removed at the yielding end. A memberend
rotation depended on both memberend
moments.
This model was unique at the time when a frame was
analyzed only for an elasticperfectly plastic condition. Bilinear
forcedeformation relation was made possible to use at
member ends by this model.
Aoyama and Sugano (1968) adapted the two component
model, creating the multicomponent model using four parallel
elements to account for flexural cracking, yield levels different
at two member ends, and strainhardening. The deformation
compatibility of the imaginary components is satisfied only at
their ends.
The stiffness matrix of a simply supported
member having four parallel elements can be
formulated by recognizing (a) the rotation at each
member end is common among four elements, and
(b) the moment at each end is the sum of moment
resisted by four elements. In other words, the
member stiffness matrix is the sum of stiffness
matrices of four parallel elements.
For a simply supported elastic member with
flexural rigidity
1
p EI ,
)
`
¹
¹
´
¦
∆
∆
=
)
`
¹
¹
´
¦
∆
∆
B
A
B
A
L
EI p
m
m
θ
θ
4 2
2 4
1
1
1
For an element with flexural rigidity
2
p EI and a plastic hinge at A end, the moment at A end is
known to be zero;
)
`
¹
¹
´
¦
∆
∆
=
)
`
¹
¹
´
¦
∆
∆
B
A
B
A
L
EI p
m
m
θ
θ
3 0
0 0
2
2
2
Member Rotation
(a) Member
(b) Member division
(c) Left end yielding
Element I
Element II
Elements I and II
Member rotation
M
e
m
b
e
r
e
n
d
m
o
m
e
n
t
(a) Member
(b) Stiffness of divided members
20
Similarly, for an element with flexural rigidity
3
p EI and a plastic hinge at B end;
)
`
¹
¹
´
¦
∆
∆
=
)
`
¹
¹
´
¦
∆
∆
B
A
B
A
L
EI p
m
m
θ
θ
0 0
0 3
3
3
3
For an element with plastic hinges at the two ends, the stiffness matrix is zero;
)
`
¹
¹
´
¦
∆
∆
=
)
`
¹
¹
´
¦
∆
∆
B
A
B
A
m
m
θ
θ
0 0
0 0
4
4
From the equilibrium of forces;
4 3 2 1
4 3 2 1
B B B B B
A A A A A
m m m m m
m m m m m
∆ + ∆ + ∆ + ∆ = ∆
∆ + ∆ + ∆ + ∆ = ∆
Therefore, the stiffness matrix of a simply supported member is written as follows;
)
`
¹
¹
´
¦
∆
∆
+
+
=
)
`
¹
¹
´
¦
∆
∆
B
A
B
A
p p p
p p p
L
EI
m
m
θ
θ
) 3 4 ( 2
2 ) 3 4 (
2 1 1
1 3 1
Note that 0
3 2
= ⋅ p p ; an element having a plastic hinge at A end and another element having a
plastic hinge at B end cannot exist simultaneously; i.e., either
2
p or
3
p or both should be zero at
a stage.
The multicomponent model appears to have a merit; rotation at one end of a member depends
on both memberend moments. In other words, the moment distribution along a member can be
approximately reflected in the analysis. However, the stiffness of the multiparallel components must
be evaluated under a certain assumed moment distribution. Therefore, the stiffness parameters are
valid only under such a moment distribution, and are bound to be approximate when the moment
distribution becomes drastically different in the analysis.
Giberson (1967) discussed the advantage and disadvantage of the onecomponent and the
twocomponent models, and concluded that the onecomponent model was more versatile than the
twocomponent model because the twocomponent model was restricted to the bilinear hysteresis
characteristics.
The ratios p's of element stiffness may be varied with damage if a more general hysteresis
relation is desired for a member. Takizawa (1976) suggested the ratios p's be determined as a
function of member end stiffness under an antisymmetric moment distribution. The flexibility matrix
including shear deformation is given below:
)
`
¹
¹
´
¦
∆
∆
=
)
`
¹
¹
´
¦
∆
∆
B
A
B
A
m
m
f f
f f
22 21
12 11
θ
θ
where, for
2
( 0)
B A
f f p < = :
B
B
B A
f f
f f f
f f f
γ
γ
γ
γ
γ γ
γ
γ
2 1
2
2 1
) 1 (
) 2 )( 2 1 (
) 1 (
2
3
22
21 12
2
11
+
+
=
+
−
− = =
+ +
−
+
+
=
and for
3
( 0)
B A
f f p > = ,
21
B A
A
A
f f f
f f f
f f
γ γ γ
γ
γ
γ
γ
γ
+
+
+ +
−
=
+
−
− = =
+
+
=
2
3
) 2 )( 2 1 (
) 1 (
2 1
) 1 (
2 1
2
2
22
21 12
11
γ : parameter describing the relative contribution of shear deformation (=
2
6 / EI GAL κ ). Flexibility
parameters
A
f and
B
f are given by a hysteresis model at member ends A and B;
B
B
B
A
A
A
m
f
m
f
∆
∆
=
∆
∆
=
θ
θ
under antisymmetric bending moment distribution.
The axial loadbending moment interaction may be approximated by changing the yield moment
level of elements as a function of existing axial load.
References:
Aoyama, H., and T. Sugano, "A Generalized Inelastic Analysis of Reinforced Concrete Structures
based on the Tests of Members," Recent Researches of Structural Mechanics, Contribution in
Honor of the 60th Birthday of Professor Y. Tsuboi, Uno Shoten, Tokyo, 1968, pp. 1530.
Clough, R. W., K. L. Benuska and E. L. Wilson, "Inelastic Earthquake response of tall buildings,"
Proceedings, Third World Conference on Earthquake Engineering, New Zealand, Vol. II,
Session II, 1965, pp. 6889.
Giberson, M. F., "The Response of Nonlinear Multistory Structures subjected to Earthquake
Excitation," EERL Report, Earthquake Engineering Research Laboratory, California Institute of
Technology, 1967.
Takizawa, T., "Notes on Some Basic Problems in Inelastic Analysis of Planar R/C Structures (Part
1)," Transactions, Architectural Institute of Japan, No. 240, February 1976, pp. 5162.
22
10.6 Distributed Flexibility Model
Once cracks develop in a member, the stiffness becomes nonuniform along the member length.
Instead of dividing a member into short segments, Takizawa (1973,1976) developed a model that
assumed a prescribed distribution pattern of crosssectional flexural flexibility (reciprocal of flexural
rigidity EI) along the member length. The flexibility matrix of a simply supported member can be
formulated by considering a member with varying section properties EI(x) and (GA/ κ )(x) with
distance x from A end.
)
`
¹
¹
´
¦
∆
∆
=
)
`
¹
¹
´
¦
∆
∆
B
A
BB BA
AB AA
B
A
m
m
f f
f f
θ
θ
in which,
dx
x GA L
dx
x EI
L
x
f
dx
x GA L
dx
x EI
L
x
L
x
f f
dx
x GA L
dx
x EI
L
x
f
L L
BB
L L
BA AB
L L
AA
∫ ∫
∫ ∫
∫ ∫
+ =
+
− −
= =
+
−
=
0
2
0
2
0
2
0
0
2
0
2
) ( /
1 1
) (
) (
) ( /
1 1
) (
) 1 )( (
) ( /
1 1
) (
) 1 (
κ
κ
κ
EI(x): tangent flexural rigidity, GA/ κ (x): tangent shear rigidity.
A parabolic distribution of flexural flexibility (1/EI(x)) was assumed; the function may be defined if
two end values (1/EI(0) and 1/EI(L)) and the minimum value (min {1/EI(x)}) are given; an inflection
point can stay within a member or outside the member. The smallest flexural flexibility may be
assumed to be the same as the initial elastic value. For member end moments, the member end
flexural flexibility may be evaluated on the basis of a momentcurvature relation.
Takizawa (1973, 1976) assumed a momentrotation relation at the member end under the
antisymmetric moment distribution to determine the flexibility coefficients
AA
f ,
AB
f and
BB
f ,
rather than flexural flexibility.
Flexibility relation of a member for incremental member end rotations
A
θ ∆ and
B
θ ∆ and
moments
A
m ∆ and
B
m ∆ is given in the following form by ignoring shear deformation;
23
)
`
¹
¹
´
¦
∆
∆
−
−
+ +
+
−
+
+
− −
−
+
=
)
`
¹
¹
´
¦
∆
∆
B
A
AB
o A
B
AB B A
AB B A
AB
o B
A
B
A
m
m
f
f f
f
f f f
f f f
f
f f
f
3
2
3
2
2
) (
3
2
2
) (
3
2
θ
θ
in which
o
f : member end flexibility at the initial elastic stage (= L / 6 EI).
Member end flexibility
A
f or
B
f was determined as an incremental member end rotation to
moment ratio under the antisymmetric bending moment distribution with the same member end
moments of amplitude
A
m or
B
m acting at the two ends. The curvature was assumed to distribute
along the member length in a parabolic form with the initial elastic value at the midspan. Normally,
this flexibility is given by a hysteresis model.
The interacting element
AB
f is defined as
) sgn( ) )( (
B A o B o A AB
m m f f f f f − − =
This is an interesting concept in analyzing an inelastic member. However, the parabolic flexibility
distribution may not describe the actual concentration of deformation at critical section (normally at
member ends) due to flexural yielding and deformation attributable to slippage of longitudinal
reinforcement within a beamcolumn connection. The usage of inelastic springs at locations of
concentrated deformation in conjunction with this model may be a useful solution.
References:
Takizawa, H., "Analysis of Reinforced Concrete Buildings under Strong Earthquake Motion (in
Japanese)," Concrete Journal, Japan Concrete Institute, Vol. 11, No. 2, February 1973, pp.
1021.
Takizawa, H., "Notes on Some Basic Problems in Inelastic Analysis of Planar R/C Structures (Part
1)," Transactions, Architectural Institute of Japan, No. 240, February 1976, pp. 5162.
24
10.7 Multispring Model
The momentcurvature analysis of a reinforced concrete section under load reversal is normally
based on (a) an assumption, known as Bernoulli's hypothesis, that a plane section remains plane
after bending, (b) uniaxial stressstrain relations of materials and (c) equilibrium of forces. The
"lamina (fiber) model" may be used for a sectional momentcurvature analysis. However, the
deformation of a member calculated by the integration of curvature is known to underestimate the
deformation observed in a test. Therefore, a member model, based on member end momentrotation
relation, is favored in a nonlinear analysis of a reinforced concrete structure.
The flexural behavior of a reinforced concrete column section is influenced by existing axial load;
the phenomenon normally called as "the interaction of axial force and bending moment." The
interaction of axial deformation and curvature also exists. For example, in a reinforced concrete
member under pure bending, the neutral axis of section shifts to compression side after flexural
cracking, which accompanies the elongation at the centroid of the section although no tensile force
is applied; i. e., pure bending causes the elongation of a member after cracking. Member stiffness
models such as onecomponent and multicomponent models cannot model such interaction.
Furthermore, a column during a real earthquake motion is subjected to bidirectional shear and
bending. The bending capacity in a principal direction is influenced by an existing bending moment
in the orthogonal direction, the phenomenon called "the bidirectional bending interaction." Proper
analytical model is necessary to analyze a threedimensional structure under bidirectional horizontal
and vertical ground motions.
Lai Model: Lai et al. (1984) proposed a multispring (MS) model to simulate the flexural behavior of
reinforced concrete columns under varying axial load and bidirectional lateral load reversals. A
column member was idealized to be a linear element with its length equal to the column clear height
and two multispring elements with zero length at the top and bottom. The multispring element
model was similar to the "lamina (fiber) model," but the section was not divided into many elements;
cross sectional properties at each member end were represented by four steel springs and five
concrete springs. All inelastic flexural deformation was assumed to concentrate in the multispring
elements at the member ends (onecomponent model).
A column was idealized by an elastic line element b with two multispring elements a and c;
member end nodes are numbered A and B. The node between the top multispring element and the
elastic element is C and the node between the bottom multispring element and the elastic element
is D. For each element, the start and terminal ends 1 and 2 are assigned in the direction of A to B.
A
B
C
D
a
b
c
y
z
x
z
25
The following assumptions were
used in formulating stiffness of a
multispring element:
(1) The plane section remains plane
after bending in the multispring
element;
(2) The forcedeformation relation of
each spring under monotonically
increasing load is elasticperfectly
plastic;
(3) Concrete springs resist only
compressive stresses;
(4) Deformation of steel springs is
caused by the pullout deformation of the
longitudinal reinforcement from
anchorage zone and the yielding of the
reinforcement;
(5) Concrete and steel springs yield
at the same deformation.
The elastic element in the middle
part is already cracked and deforms in
flexure only. The moment of inertia of
section is evaluated for fully cracked
section.
The steel spring simulates the elastoplastic behavior and the Bauschinger effect under load
reversals. The area of a steel spring is calculated by summing the areas of reinforcing bars in the
corner quadrant of the section. The yield force
sy
P of a steel spring is determined as the steel area
(= A
s
/ 4) in the tributary area and the yield
strength
sy
σ of the steel.
4
sy g
sy
A
P
σ
=
The deformation due to the pullout of
longitudinal reinforcement from the
anchorage zone is assigned to the steel
spring. The steel stress was assumed to
distribute linearly with constant bond stress
u along the development length
d
l
within the joint;
u d
f A
b
y b
d
π
= l
where
b
d : diameter of bar,
b
A : cross sectional area of bar,
y
f : yield stress of bar. The pullout
deformation
sy
d of the longitudinal reinforcement at yielding is estimated by assuming a linear
distribution of steel strain over the development length
d
l ;
1
2
sy d y
d ε = l
where,
y
ε : yield strain of steel.
The elastic stiffness
se
k of steel spring is the ratio of yield force to the yield displacement;
26
s y
se
sy
A f
k
d
=
Postyield stiffness of a steel spring
was assumed to be zero.
The concrete spring resisted only
compression stress. The initial elastic
deformation of the spring simulates the
initial elastic depression at the joint
core. The plastic deformation of a
concrete spring represents the
accumulated crushing behavior of the
concrete over the plastic hinge length.
The yielding force level of the effective
concrete spring is expressed by
ci B cy
A P σ 85 . 0 =
where
cy
P : yielding force of the
concrete spring,
ci
A : tributary area for a
concrete spring.
The tributary area
ci
A of a corner
concrete spring was evaluated by
considering the equilibrium of axial force
at the balanced point, when the tensile
steel springs yield simultaneously with
the crushing (compression yielding) of
the concrete springs. As the yield
deformation of the steel and concrete
springs was assumed to be the same,
the steel spring elements in tension and
compression both yield at the balanced
point. The neutral axis lies at the center
of the section and the central concrete
element does not carry any force.
Therefore, the axial force at the
balanced point must be resisted by the
two concrete spring elements;
2(0.85 )
b
ci
B
P
A
σ
=
The area
0 c
A of the central concrete
spring was calculated as the remaining
area after steel area and tributary areas
of the four corner concrete springs were
removed from the gross sectional area.
The distance d between the four
corner springs was determined by the
equilibrium of bending moment
b
M at
the balanced point. As the yield
deformation of the steel and concrete
springs was assumed to be identical, the
neutral axis is at the centroid of the
Cross Section
Deformation and resistance in springs
at the balanced point
P
b
M
b
27
section. Therefore, the distance between the springs is determined as
∑ ∑
+
=
ci B y si
b
A f A
M
d
σ 85 . 0 ) (
2
The axial forcebending moment interaction diagram may be represented by four zones. The
equations for the four zonse may be expressed as follows;
(a) Zone 1: Tension failure zone (TAB)
d
M
P P P
y
y s y s
2
) ( 2
3 1
− + =
(b) Zone 2: Compression failure zone (BD)
0
2 2 4 4 )
1 3
0 0
1 3
[2 ]( )
2( )
2
[1 ]
2 2( )
c y
c y s y c y s y
s y s y
c y c y y
s y s y
P
P P P P P
P P
P P M
P P d
= − + + + +
+
− + +
+
(c) Zone 3 (CD):
0
2 2 4 4
1 1 3 3
0 0
1 1 3 3
[2 ]( )
2( )
2
[1 ]
2 2( )
c y
c y s y c y s y
s y c y s y c y
c y c y y
s y c y s y c y
P
P P P P P
P P P P
P P M
P P P P d
= − + + + +
+ + +
− + +
+ + +
The interaction curve of the model deviated from the interaction curve due to the use of fewer
number of springs. With an increase in the number of springs in a multispring element, the
simulation of the interaction behavior is improved, and the determination of the stiffness properties of
each spring is simplified.
Jiang and Saiidi (1990) proposed to combine the
hysteresis properties of the concrete and steel
springs in each corner to simplify the model .
because the yield deformations of the concrete
and steel springs located in the same quadrant is
identical.
Forcedeformation relation
(Jiang and Saiidi, 1990)
d
F
F
sy
+F
cy
F
sy
d
sy
d
sy k
1
k
2
2
k
s
β
2
k
s
β
28
Li Model: Li et al. (1988, 1990) simplified the method to determine spring properties and modified
the hysteretic properties of the concrete and steel springs; they demonstrated the reliability of the
model with respect to column test results using simple five spring models.
Li (1991) suggested the use of fiber models at the
member end. A member is represented by plastic
zones (multiaxial spring elements) at the member
ends and an elastic zone in the middle part.
A multiaxial spring element consists of uniaxial
springs in the direction of member axis. The number
of springs may be chosen considering the material
properties, section size and reinforcement
arrangement. Each single steel bar may be represented by a steel spring at the bar center. Concrete
section may be divided into several subareas; a concrete spring is placed at the geometrical
centroid of a concrete subarea. The multispring element is rigid against shearing force. The length
of a multiaxial spring element is given as p
z
, which is suggested as the smaller of onehalf of total
depth of end section or onetenth of the clear length of the member.
L
o
p
z
Multispring element
Elastic element
29
The axial force f
i
and deformation d
i
of each axial spring are given as follows;
z i i
i i i
p d
A f
ε
σ
=
=
where
i
σ and
i
ε : stress and strain at spring point i, and
i
A . area of subarea i. Multilinear
stressstrain relation may be used for concrete and steel springs. Some adjustment is required for
the stressstrain relation of a steel spring to take into account the stiffness degradation, for example,
due to the bond slip along the longitudinal reinforcement or the pullout deformation of the
longitudinal reinforcement from the anchorage zone. The yield deformation may be increased by a
factor κ ;
0 . 1 0 . 1
0 . 1
0 . 1
0 . 1
≤ =
>
−
+ =
D
h
for
D
h
for
D
h
D
h
o
o
o
o
κ
κ
where h
o
: shear span, and D: overall depth of section.
30
The flexural and axial deformations may be considered in the middle elastic part.
A common member coordinate system is used to define forces and displacements, with xaxis in
the direction of the member and y and zaxes in the principal directions of the section. Forces and
displacements at node J are denoted by
J
P} { and
J
D} { , whereas the start and terminal end
forces and displacements of an element i are denoted by
i
p } {
1
and
i
p } {
2
,
i
d } {
1
and
i
d } {
2
.
The incremental stiffness relation of a multispring model i (i = a or c) may be expressed;
) (
2
2
2
1
1
1
2
2
2
1
1
1
i
z
y
x
i
z
y
x
i
zz zy zx
yz yy yx
xz xy xx
i
z
y
x
i
z
y
x
d d
k k k
k k k
k k k
m
m
p
m
m
p
¦
)
¦
`
¹
¦
¹
¦
´
¦
∆
∆
∆
−
¦
)
¦
`
¹
¦
¹
¦
´
¦
∆
∆
∆
=
¦
)
¦
`
¹
¦
¹
¦
´
¦
∆
∆
∆
− =
¦
)
¦
`
¹
¦
¹
¦
´
¦
∆
∆
∆
θ
θ
θ
θ
in which,
∑
∑
∑
∑
∑
∑
=
=
=
=
=
=
i
i i zz
i i
i
i yz
i
i
i yy
i
i
i xz
i
i
i xy
i
i xx
y k k
z y k k
z k k
y k k
z k k
k k
2
2
and
i
k : tangent stiffness of spring i, (y
i
, z
i
): coordinates of spring i with respect to the centroid of
section,
x
p : axial force,
y
m : bending moment about yaxis,
z
m : bending moment about zaxis,
x
d : axial deformation at the centroid,
y
θ : rotation about yaxis,
z
θ : rotation about zaxis.
It should be noted that the following relations hold because no length is considered in multispring
elements a and c;
a
z
y
a
z
y
p
p
p
p
)
`
¹
¹
´
¦
∆
∆
− =
)
`
¹
¹
´
¦
∆
∆
2
2
1
1
c
z
y
c
z
y
p
p
p
p
)
`
¹
¹
´
¦
∆
∆
− =
)
`
¹
¹
´
¦
∆
∆
2
2
1
1
and
a
z
y
a
z
y
d
d
d
d
)
`
¹
¹
´
¦
∆
∆
− =
)
`
¹
¹
´
¦
∆
∆
2
2
1
1
c
z
y
c
z
y
d
d
d
d
)
`
¹
¹
´
¦
∆
∆
− =
)
`
¹
¹
´
¦
∆
∆
2
2
1
1
In a symbolic expression, for multispring a;
a a a a a
a a a a a
d k d k p
d k d k p
} { ] [ } { ] [ } {
} { ] [ } { ] [ } {
2 1 2
2 1 1
∆ + ∆ − = ∆
∆ − ∆ = ∆
and for multispring element c;
c c c c c
c c c c c
d k d k p
d k d k p
} { ] [ } { ] [ } {
} { ] [ } { ] [ } {
2 1 2
2 1 1
∆ + ∆ − = ∆
∆ − ∆ = ∆
A
C
D
a
b
c
B
31
For an elastic element of length L, flexural rigidity EI, axial rigidity EA, and shear rigidity GA/κ, a
stiffness relation can be formulated in the form:
b b b b b
b b b b b
d k d k p
d k d k p
} { ] [ } { ] [ } {
} { ] [ } { ] [ } {
2 22 1 21 2
2 12 1 11 1
∆ + ∆ = ∆
∆ + ∆ = ∆
where,
b k
p } { and
b k
d } { at kend are
T
zk zk yk yk xk
m p m p p } , , , , { and
T
zk zk yk yk xk
d d d } , , , , { θ θ .
The compatibility of displacements at nodes:
c b D
b a C
c B
a A
d d D
d d D
d D
d D
} { } { } {
} { } { } {
} { } {
} { } {
1 2
1 2
2
1
∆ = ∆ = ∆
∆ = ∆ = ∆
∆ = ∆
∆ = ∆
Equilibrium of external forces and the sum of internal element end forces at a node;
c b D
b a C
c B
a A
p p P
p p P
p P
p P
} { } { } {
} { } { } {
} { } {
} { } {
1 2
1 2
2
1
∆ + ∆ = ∆
∆ + ∆ = ∆
∆ = ∆
∆ = ∆
Special care must be exercised in formulating a member stiffness matrix with four nodes (A, B, C
and D) to include the following conditions;
(a) for displacements:
c
z
y
c
z
y
B
z
y
D
z
y
a
z
y
a
z
y
C
z
y
A
z
y
d
d
d
d
D
D
D
D
d
d
d
d
D
D
D
D
)
`
¹
¹
´
¦
∆
∆
=
)
`
¹
¹
´
¦
∆
∆
=
)
`
¹
¹
´
¦
∆
∆
=
)
`
¹
¹
´
¦
∆
∆
)
`
¹
¹
´
¦
∆
∆
=
)
`
¹
¹
´
¦
∆
∆
=
)
`
¹
¹
´
¦
∆
∆
=
)
`
¹
¹
´
¦
∆
∆
2
2
1
1
2
2
1
1
and (b) for forces:
a
z
y
a
z
y
p
p
p
p
)
`
¹
¹
´
¦
∆
∆
− =
)
`
¹
¹
´
¦
∆
∆
2
2
1
1
c
z
y
c
z
y
p
p
p
p
)
`
¹
¹
´
¦
∆
∆
− =
)
`
¹
¹
´
¦
∆
∆
2
2
1
1
References:
Lai, S.S., G. T. Will and S. Otani, "Model for Inelastic Biaxial Bending of Concrete Member,"
Journal, Structural Division, ASCE, Vol. 110, No. ST11, November 1984, pp. 25632584.
Jiang, Y., and S. M. Saiidi, "FourSpring Element for Cyclic Response of R/C Columns," Journal,
Structural Engineering Division, ASCE, Vol. 116, No. ST4, April 1990, pp. 10181029.
Li, K.N., S. Otani and H. Aoyama, "Reinforced Concrete Columns under Varying Axial Load and
Bdirectional Lateral Load Reversals," Proceedings, Ninth World Conference on Earthquake
Engineering, TokyoKyoto, August 1988, Vol. VIII, pp. 537542.
Li, K.N., S. Otani and H. Aoyama, "Study on Reinforced Concrete Columns subjected to Varying
Axial Load and Bidirectional Horizontal Earthquake Loads (in Japanese)," Report, Aoyama
Laboratory, Department of Architecture, Faculty of Engineering, University of Tokyo, March
1990.
Li, KangNing, S. Otani and H. Aoyama, “R/C Columns under Axial and Bidirectional Lateral Loads,”
Proceedings, Mechanics Computing in 1990’s and Beyond, ASCE, Vol. 2, Structural and
Material Mechanics, May 1991, pp. 681  685.
32
10.10 Wall Models
Column Model: A structural wall is often represented by a
column (lineal) model at the center of the wall section. The
boundary girders at the top and bottom of a wall panel are
normally assumed to be rigid. The bending and shear
deformations are considered for the wall.
Onecomponent model is often used in an inelastic analysis,
in which a yield rotation is estimated for uniform or
antisymmetric moment distribution along the wall height. The
uniform bending moment distribution may be realistic for a
structural wall at the lower level. The flexibility relation of a wall
including a shear spring is
¦
)
¦
`
¹
¦
¹
¦
´
¦
∆
∆
∆
+ + + −
+ − + + =
¦
)
¦
`
¹
¦
¹
¦
´
¦
∆
∆
∆
B
A
AB
B
A
B
A
AB
m
m
p
f f f
f f f
EA
h
e
γ γ
γ γ
θ
θ
2 0
2 0
0 0
in which
EI
h
f
6
= ,
A
A
k
f
1
= ,
B
B
k
f
1
= ,
h GA
w
κ
γ = , EI : flexural rigidity of the middle elastic
region, h : clear height of the wall panel,
κ
w
GA
: shear rigidity of wall section including shape factor
κ for shear. Shear rigidity may be reduced with inelastic deformation.
A rigid zone, length equal to the onehalf width of a wall, must be considered at the end of a
girder connected to the structural wall.
Distributed flexibility model may be used to represent a distribution of damage for a wall. Shear
deformation of a wall needs be considered. A wall may be subdivided into a short segment (the
discrete element model) to reflect the distribution of the damage.
The problem of representing a wall by a single line member at the center is that the “three
dimensional effect” of a flexural wall cannot be represented; i.e., the axial elongation of a wall at the
centroid due to the shift of the neutral axis after flexural cracking cannot be modeled. The boundary
girders connected on both sides of a wall displace the same amount in the vertical direction at the
wall faces.
Brace Model: A structural wall is sometimes idealized by a
braced frame, in which shear deformation is represented by
the deformation of diagonal braces and flexural deformation by
the deformation of vertical elements. This model is useful
when the shear deformation is dominant in a structural wall.
The axial stiffness of tensile bracing and vertical elements
may be reduced to take into account the degradation of
stiffness due to cracking.
The flexibility of the wall element is given for the coordinate
system shown in the figure,
k
A
k
B
Axial Spring
Shear Spring
Flexibility of springs
h
Rigid Beam
Rigid Beam
L
p
x
p
y
m
z
EA
1
EA
2
EA
2
33
3 3 2
2 2
1 2 2
3
1
3 3
2
1 2
2
2 2
2 2
1 ' 1
( ) 0
2
'
0 0
' 2
1 2
0
x x
y y
z z
L h h
L EA EA EA L
L
d p
h EA
d p
L h EA
m
EA EA
h h
EA L EA L
θ
+ −
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
=
´ ` ´ `
¦ ¦ ¦ ¦
+
¹ ) ¹ )
−
where
1
2 2
2
' ( ) L L h = + : length of diagonal braces.
Boundary Column Model: A structural wall, especially behaving dominantly in flexure, is modeled
by three springs at the boundary columns and at the wall center (Otani et al., 1985). The two outside
springs are provided with the axial stiffness of the boundary columns. The central element
represents the vertical uniaxial, lateral shear and flexural rotational characteristics of the wall panel;
a rotational spring is placed only at the bottom of the central element. The girder within a wall is
considered to be rigid.
The axial stiffness in compression is assumed to be
elastic, and the tensile stiffness is reduced in comparison
with the compression stiffness ignoring "the tension
stiffening effect," and the stiffness is reduced to a small
value after tensile yielding of the longitudinal
reinforcement.
The flexibility relation of the wall panel as a simply
supported member is
¦
)
¦
`
¹
¦
¹
¦
´
¦
∆
∆
∆
+ + + −
+ − + =
¦
)
¦
`
¹
¦
¹
¦
´
¦
∆
∆
∆
B
A
AB
B B
A
AB
m
m
p
f f f
f f
a e
γ γ
γ γ
θ
θ
2 0
2 0
0 0
in which
w
a a a + + =
2 1
α ,
w
w
h
a
EA
= ,
w
EI
h
f
6
= ,
B
f :
flexibility of rotational spring,
h GA
w
κ
γ = ,
κ
w
GA
: shear
rigidity of the wall section, taking stiffness degradation with shear cracking into account,
w
EI :
elastic flexural rigidity of wall panel section. Forces and displacements are defined at the midpoint
of the rigid girders.
The stiffness of the wall model is formulated by including the stiffness contribution from the
boundary columns.
Reference:
Kabeyasawa, Toshimi, Chapter 7 Earthquake Response Analysis, Design of Modern Highrise
Reinforced Concrete Structures, Series on Innovation in Structures and Construction vol. 3,
Imperial College Press, 2001, pp. 315  344.
Otani, S., T. Kabeyasawa, H. Shiohara and H. Aoyama, "Analysis of the Fullscale Sevenstory
Reinforced Concrete Test Structure," ACI SP84, Earthquake Effects on Reinforced Concrete
k
B
a
1 a
2
a
w
γ
Rigid Girder
Boundary Column Model
(Kabeyasawa Model)
34
Structure, USJapan Research, American Concrete Institute, Detroit, 1985, pp. 203239.
Vulcano, A., and V. V. Bertero, “Analytical Models for Predicting the Lateral Response of RC Shear
Walls,” University of California at Berkeley, Earthquake Engineering Research Center, Report
No. EERC 8719, 1987.
35
Home Assignment No. 6
(Formulation of Member Stiffness Matrix)
20020308
Otani, S.
Consider a multicomponent model, consisting
of two parallel elements I and II, supported at A
end by a pin and at B end by a roller. Element I is
connected at A end by a plastic hinge and is
rigidly connected at B end. Element II is rigidly
connected at the two ends. The elastic flexural
rigidity of elements I and II is p
1
EI and p
2
EI. The
stiffness relation is expressed as follows:
)
`
¹
¹
´
¦
∆
∆
+
=
)
`
¹
¹
´
¦
∆
∆
B
A
B
A
p p p
p p
L
EI
m
m
θ
θ
1 2 2
2 2
3 4 2
2 4
(1) Express member end rotations
A
θ ∆ and
B
θ ∆ in terms of member end moments
A
m ∆
and
B
m ∆ in a matrix form.
(2) The incremental momentand rotation relations are given for antisymmetric bending moment
distribution below. Namely, the relations were obtained by applying equal incremental moments
A
m ∆ at the two ends of a simply supported beam and measuring resultant incremental member
end rotations
A
θ ∆ at A end. Similarly, equal incremental member end moments
B
m ∆ were
applied to a simple beam and incremental member end rotation
B
θ ∆ was measured at B end. The
flexibility at the member ends is defined from the incremental relation as follows;
1
1
A
A
A A
B
B
B B
f
k m
f
k m
θ
θ
∆
= =
∆
∆
= =
∆
A
θ ∆
A
m ∆
A
θ
A
m
A
k
B
θ ∆
B
m ∆
B
θ
B
m
B
k
Determine the stiffness ratios
1
p and
2
p which satisfy the flexibility
A
f and
B
f at the two
member ends as defined above.
A B
Plastic hinge
1
p EI
2
p EI
B
m
A
m
36
[Solution]
For a twocomponent model with a plastic hinge at left end, the tangent stiffness relation is given as
)
`
¹
¹
´
¦
∆
∆
+
=
)
`
¹
¹
´
¦
∆
∆
B
A
B
A
p p p
p p
L
EI
m
m
θ
θ
1 2 2
2 2
3 4 2
2 4
(1) The stiffness relation is solved for the flexibility relation,
2 1 2
2
2 2 2 1 2 2
4 3 2
1
2 4 (4 3 )4 4
A A
B B
p p p m
L
p p m EI p p p p
θ
θ
∆ + − ∆ ¦ ¹ ¦ ¹
=
´ ` ´ `
∆ − ∆ + −
¹ ) ¹ )
(2) Under antisymmetric bending,
A B
m m ∆ = ∆ ;
2 1 2
2
2 1 2 2
2 1
2
2 1 2 2
2 1
1 2 2
2 2 2
2 1 2 2
2
2
2 1 2 2
1 2
1
(4 3 2 )
(4 3 )4 4
1
(2 3 )
(4 3 )4 4
2 3
12( )
1
( 2 4 )
(4 3 )4 4
1
(2 )
(4 3 )4 4
1
6( )
A A
A
A
A
A
B B
B
B
B
B
L
p p p m
EI p p p p
L
p p m
EI p p p p
L p p
m
EI p p p
m
k
L
p p m
EI p p p p
L
p m
EI p p p p
L
m
EI p p
m
k
θ
θ
∆ = + − ∆
+ −
= + ∆
+ −
+
= ∆
+
∆
=
∆ = − + ∆
+ −
= ∆
+ −
= ∆
+
∆
=
Solving for
1
p and
1
p ,
2
1
2 2
3 2
A B
A B
B A
B
A B
L k k
p
EI k k
L k k
p k
EI k k
=
+
−
=
+
37
Home Assignment No. 6
20011205
Otani, S.
Consider a system consisting of diagonal braces with
axial rigidity EA
1
and vertical elements with axial rigidity
EA
2
. The braces and vertical elements are connected to
rigid girders (girder depth is zero) at the top and bottom by
hinges. The distance between the vertical elements is L
and clear height between the top and bottom girder is h.
Formulate a flexibility relation of the system at the
center of the rigid girder at the top, considering the bottom
girder to be fixed. In other words, find horizontal, vertical
and rotational displacements under each of unit horizontal
force, unit vertical force and unit moment applied at the
midspan of the top girder.
h
Rigid Beam
Rigid Beam
L
p
x
p
y
m
z
EA
1
EA
2
EA
2
38
[Solution]
This structure is statically indeterminate by one degree.
The reaction at the support at lower right is selected as
an indeterminate force and is released to make a
statically determinate structure.
A
B
C
D
Basic Determinate Structure
α
Axial forces in the members are calculated for unit force
applied separately in the horizontal, vertical and
rotational directions at the center of the top rigid beam, and also for a unit applied in the horizontal
direction at the released support. Axial force is positive when in tension.
A
B
C
D
h
L
−
1
cosα
h
L
−
1
cosα
N=1
A
B
C
D A
B
C
D
P
y
=1.0
M
z
=1.0
1
2
1
2
A
B
C
D
P
x
=1.0
h
L
−
1
cosα
0
0 0
0
1
L
−
1
L
0
0
N
H0
N
V0
N
M0
N
u
Horizontal displacement is calculated in the direction of the indeterminate force under the unit force
using the unit load method.
Member
L
EA
0 H
N
0 V
N
0 M
N
u
N
0 H u
L
N N
EA
0 V u
L
N N
EA
0 M u
L
N N
EA
2
u
L
N
EA
AB
2
h
EA
0
1
2
1
L
−
h
L
−
0
2
2
1
2
h
EA L
−
2
2
2
1 h
EA L
−
3
2
2
1 h
EA L
AC
1
' L
EA
1
cosα
0 0
1
cosα
2
1
' 1
cos
L
EA α
0 0
2
1
' 1
cos
L
EA α
h
Rigid Beam
Rigid Beam
L
p
x
p
y
m
z
EA
1
EA
2
EA
2
39
BD
1
' L
EA
0 0 0
1
cosα
0 0 0
2
1
' 1
cos
L
EA α
CD
2
h
EA
h
L
−
1
2
1
L
h
L
−
3
2
2
1 h
EA L
2
2
1
2
h
EA L
−
2
2
2
1 h
EA L
3
2
2
1 h
EA L
Total     
3 3
2
1 2
1 '
( )
L h
L EA EA
+
2
2
1 h
EA L
−
0
3 3
2
1 2
2 '
( )
L h
L EA EA
+
where
2 2
' L L h = +
From the displacement boundary condition at support D, the horizontal reaction
1
N at the support
due to unit horizontal load acting at the top beam is calculated;
3 3 3 3
1
2 2
1 2 1 2
1 ' 2 '
( ) ( ) 0.0
L h L h
N
L EA EA L EA EA
+ + + =
1
1
2
N = −
The horizontal reaction
1
N due to unit vertical force is calculated;
2 3 3
1
2
2 1 2
2
2
1 3 3
1 2
1 2 '
( ) 0.0
' 2
h L h
N
EA L L EA EA
h
L EA
N
L h
EA EA
− + + =
=
+
The horizontal reaction
1
N due to unit moment is calculated;
3 3
1
2
1 2
1
2 '
0 ( ) 0.0
0.0
L h
N
L EA EA
N
+ + =
=
Therefore, the axial forces in he brace members due to unit load applied at the top beam are
expressed as;
A
B
C
D A
B
C
D
P
y
=1.0
M
z
=1.0
1
1
2
h
N
L
−
1
1
2
h
N
L
−
A
B
C
D
P
x
=1.0
2
h
L
−
1
2cosα
2
h
L
1
2cosα
−
1
cos
N
α
1
cos
N
α
1
L
−
1
L
0
0
N
H
N
V
N
M
0.0
0.5
2
2
1 3 3
1 2
' 2
h
L EA
N
L h
EA EA
=
+
Calculation of displacement due to unit force applied at the center of top beam. Note that a unit force
in the direction of desired displacement may be applied to the original structure, but unit forces in the
direction of applied forces can be applied to the statically determinate structure.
40
A
B
C
D A
B
C
D
P
y
=1.0
M
z
=1.0
1
1
2
h
N
L
−
1
1
2
h
N
L
−
A
B
C
D
P
x
=1.0
2
h
L
−
1
2cosα
2
h
L
1
2cosα
−
1
cos
N
α
1
cos
N
α
1
L
−
1
L
0
0
N
H0
N
V0
N
M0
0.0
0.5
2
2
1 3 3
1 2
' 2
h
L EA
N
L h
EA EA
=
+
A
B
C
D A
B
C
D
P
y
=1.0
M
z
=1.0
1
2
1
2
A
B
C
D
P
x
=1.0
h
L
−
1
cosα
0
0 0
0
1
L
−
1
L
0
0
N
Hu
N
Vu
N
Mu
(1) Horizontal displacement
x
d at the top beam due to horizontal force
x
p , vertical force
y
p and
moment
z
m acting at the top beam.
Member
L
EA
0 H
N
0 V
N
0 M
N
u
N
0 H u
L
N N
EA
0 V u
L
N N
EA
0 M u
L
N N
EA
AB
2
h
EA
0
1
2
1
L
−
2
h
L
0
2
2
1
4
h
EA L
2
2
2
1
2
h
EA L
−
AC
1
' L
EA
1
cosα
0 0
1
2cosα
2
1
' 1
2cos
L
EA α
0 0
BD
1
' L
EA
0 0 0
1
2cosα
−
0 0 0
CD
2
h
EA
h
L
−
1
2
1
L
2
h
L
−
3
2
2
1
2
h
EA L
2
2
1
4
h
EA L
−
2
2
2
1
2
h
EA L
−
Total     
3 3
2
1 2
1 '
( )
2
L h
L EA EA
+
0
2
2
2
1 h
EA L
−
3 3 2
2 2
1 2 2
1 ' 1
( )
2
x x z
L h h
d p m
L EA EA EA L
= + −
41
(2) Vertical displacement
y
d at the top beam due to horizontal force
x
p , vertical force
y
p and
moment
z
m acting at the top beam.
Member
L
EA
0 H
N
0 V
N
0 M
N
u
N
0 H u
L
N N
EA
0 V u
L
N N
EA
0 M u
L
N N
EA
AB
2
h
EA
0
1
2
1
L
−
1
1
2
h
N
L
−
0
1
2
1
( )
2 2
h h
N
EA L
−
1 2
2
1
( )
2
h h
N
EA L L
− −
AC
1
' L
EA
1
cosα
0 0
1
cos
N
α
1
2
1
'
cos
L N
EA α
0 0
BD
1
' L
EA
0 0 0
1
cos
N
α
0 0 0
CD
2
h
EA
h
L
−
1
2
1
L
1
1
2
h
N
L
−
2
1 2
2
( )
2
h h h
N
EA L L
−
1
2
1
( )
2 2
h h
N
EA L
−
1 2
2
1
( )
2
h h
N
EA L L
−
Total     
1
2
1
2
1
2
2
'
cos
( )
2
L N
EA
h h h
N
EA L L
α
+
−
1
2
1
( )
2
h h
N
EA L
−
0
2
2
1 3 3
1 2
' 2
h
L EA
N
L h
EA EA
=
+
2
1
1 1
2 2
1 2 2
' 1
{ ( )} ( )
cos 2 2
y x y
L N h h h h h
d N p N p
EA EA L L EA L α
= + − + −
2
1
1
2 2
1 2
3 3
1
2
1 2 2
2
3 3 2
2
3 3 2
1 2 2
1 2
2 2
2 2
'
( )
cos 2
'
( )
2
1 ' 1
( )
' 2 2
1
0
2 2
L N h h h
N
EA EA L L
N L h h h
L EA EA EA L
h
L L h h EA
L h L EA EA EA L
EA EA
L h h
EA EA L
α
+ −
= + −
= + −
+
= − =
Therefore,
42
1
2
2
2
3 3
2
1 2
3
2
3 3
2
1 2
3
1
3 3
2
1 2
1
( )
2
1
( )
' 2 2
(1 )
' 2
'
' 2
y y
y
y
y
h h
d N p
EA L
h
h h L EA
p
L h EA L
EA EA
h
h EA
p
L h EA
EA EA
L
h EA
p
L h EA
EA EA
= −
= −
+
= −
+
=
+
(3) Rotation
z
θ at the top beam due to horizontal force
x
p , vertical force
y
p and moment
z
m
acting at the top beam.
Member
L
EA
0 H
N
0 V
N
0 M
N
u
N
0 H u
L
N N
EA
0 V u
L
N N
EA
0 M u
L
N N
EA
AB
2
h
EA
0
1
2
1
L
−
1
L
−
0
2
2
h
EA L
−
2
2
h
EA L
AC
1
' L
EA
1
cosα
0 0 0 0 0 0
BD
1
' L
EA
0 0 0 0 0 0 0
CD
2
h
EA
h
L
−
1
2
1
L
1
L
2
2
2
h
EA L
−
2
2
h
EA L
2
2
h
EA L
Total     
2
2
2
1 h
EA L
−
0
2
2
2h
EA L
2
2 2
2 2
1 2
z x z
h h
p m
EA L EA L
θ = − +
Collecting the information,
43
3 3 2
2 2
1 2 2
3
1
3 3
2
1 2
2
2 2
2 2
1 ' 1
( ) 0
2
'
0 0
' 2
1 2
0
x x
y y
z z
L h h
L EA EA EA L
L
d p
h EA
d p
L h EA
m
EA EA
h h
EA L EA L
θ
+ −
¦ ¹ ¦ ¹
¦ ¦ ¦ ¦
=
´ ` ´ `
¦ ¦ ¦ ¦
+
¹ ) ¹ )
−
where
1
2 2
2
' ( ) L L h = +
44
10.11 Modeling of Foundation
A building structure is often analyzed with rigid foundation. A structure stands on flexible ground,
and it sometimes becomes necessary to consider the effect of soil flexibility. The finite element
model is sometimes used for the ground. A simple model is to consider a sway spring and a rocking
spring under a rigid foundation.
Free surface
Sway
Rocking
Pile foundation supporting a structural wall is
idealized by an elastic member supported on a fixed
base or by a line member supported by vertical springs
at various levels. Vertical elastic stiffness k
v
of a pile
may be evaluated by using a reduced length taking into
account friction along the depth of a pile (Japan Road
Association, 1990);
l
p p
p
E A
k α =
where, α: effective length factor,
p
A : cross sectional
area at the bottom end of a pile,
p
E : Young's modulus
of pile material, l : length of the pile.
The value of a was studied for loading test data (L/D > 10, where L/D: length to diameter ratio of
a pile). The vertical stiffness was determined as the secant stiffness at yielding point on log Plog S
relation, where P: load applied at the pile top, and S: vertical deformation at the pile top. The
following expressions are suggested (Japan Road Association, 1990);
(a) driving steel tube pile:
78 . 0 014 . 0 + =
D
l
α
(b) driving PC (precast prestressed concrete) pile or HPC (high strength precast prestressed
concrete) pile:
61 . 0 013 . 0 + =
D
l
α
(c) castinsitu reinforced concrete pile:
15 . 0 031 . 0 − =
D
l
α
(d) drilled steel tube pile:
Free surface
No mass
45
39 . 0 009 . 0 + =
D
l
α
(e) drilled PC and HPC pile:
36 . 0 011 . 0 + =
D
l
α
The above expressions tend to give small vertical stiffness of a pile foundation.
The vertical stiffness of a pile may be calculated by considering a pile being modeled by a single
line member with a series of vertical friction springs attached along the depth and a vertical spring at
the bottom of the pile. The friction spring properties may be determined for the displacement.
Swayrocking Model: A swayrocking model of a large structure may be formulated by the dynamic
ground compliance proposed by T. Kobori for square foundation.
The effective shear modulus
0
G is estimated from shear wave velocity
s
V and unit weight per
volume ρ ;
2
0 s
G V
g
ρ
=
where, g : gravity acceleration.
The dynamic shear modulus may be estimated to be onehalf of the static shear modulus;
0
/ 2
e
G G =
The first mode undimensional frequency
0
a is defined as
0
e
a d
G
ρ
ω =
where, ω : circular frequency of structure on rigid foundation, d : dimension of the square
foundation.
The equivalent spring constants
eS
K and
eR
K for sway and rocking are evaluated from charts
Equivalent constants for rocking
Equivalent constants for sway
Nondimensional frequency a
0
Nondimensional frequency a
0
E
q
u
i
v
a
l
e
n
t
s
p
r
i
n
g
c
o
n
s
t
a
n
t
s
K
e
S
a
n
d
E
q
u
i
v
a
l
e
n
t
d
a
m
p
i
n
g
c
o
e
f
f
i
c
i
e
n
t
C
e
S
E
q
u
i
v
a
l
e
n
t
s
p
r
i
n
g
c
o
n
s
t
a
n
t
K
e
R
a
n
d
E
q
u
i
v
a
l
e
n
t
d
a
m
p
i
n
g
c
o
e
f
f
i
c
i
e
n
t
C
e
R
46
for longitudinal dimension 2c and transverse dimension 2b. Spring constants
S
K and
R
K for
rocking and sway springs and associated damping factors
S
h and
R
h are evaluated as;
3
0
0
3
1
2
1
2
S eS e
e
R eR
eS
S
eS
eR
R
eR
K K d G
G
K K d
a C
h
K
a C
h
K
=
=
=
=
Reference:
Japan Road Association, "Standard Specifications for Design of Road Bridge and Commentary, Part
4: Underground Structures (in Japanese)," revised in February 1990.
1
Chapter 11. Member Hysteresis Models
11.1 Introduction
An inelastic earthquake response analysis of structures requires realistic hysteresis models,
which can represent resistancedeformation relationship of a structural member model.
The resistancedeformation relations are different for constitutive materials of a section, for a
section, for a member, for a story and for an entire structure. The resistancedeformation relation of
a structural analysis unit observed in a laboratory test must be idealized into a
resistancedeformation hysteresis model. Different levels of resistancedeformation models must be
used for structural elements considered in an analysis; e.g., a constitutive model of materials in a
finite element method analysis, a hysteresis model for a rotational spring in a onecomponent
member model, a story sheardrift hysteresis model for a massspring model.
A hysteresis model is derived by extracting common features of resistancedeformation relations
observed in laboratory tests of members of similar properties. The hysteresis model of a member
must be able to express resistancedeformation relations under any loading history, including load
reversals.
Resistancedeformation relationship under monotonically increasing loading is called the primary
curve, skeleton curve or backbone curve. The skeleton curve provides an envelope of the hysteresis
resistancedeformation relationship if the behavior is governed by stable flexure. The skeleton curve
for reinforced concrete member is normally represented by a trilinear relation with stiffness changes
at flexural cracking and tensile yielding of longitudinal reinforcement. The skeleton curve of a
member must be defined on the basis of mechanical properties of constitutive materials and
geometry of the member. Some researchers suggest the use of a bilinear relation with a stiffness
change at yielding, ignoring the initial uncracked stage, because a reinforced concrete member
subjected to light axial force can be easily cracked by shrinkage or accidental and gravity loading.
The stateoftheart does not provide a reliable method to estimate the initial stiffness, yield
deformation and ultimate deformation. The stiffness degrades from the initial elastic stiffness with
increased inelastic deformation and the number of cycles under reversed loading. The elastic
modulus of concrete varies significantly with concrete strength and mix; initial cracks cause decay in
the stiffness. The estimate of yield deformation is more complicated by the interaction of bending
and shear deformation and additional deformation due to pullout of longitudinal reinforcement from
the anchorage zone and due to bar slip of longitudinal reinforcement along the longitudinal
reinforcement within the member. Empirical expressions are necessary for the estimate of yield and
ultimate deformation.
The coordinates of a response point on a deformationresistance plane are given by (D, F), in
which, D: deformation, F: resistance. The skeleton curve is represented by either "bilinear" or
"trilinear" lines for a reinforced concrete member, with stiffness changes at "cracking (C)" and
"yielding (Y)" points.
The following terms are defined to clarify the hysteresis
description;
Loading: a case where the absolute value of resistance (or
deformation) increases on the skeleton curve;
Unloading: a case where the absolute value of resistance
(or deformation) decreases after loading or reloading; and
Reloading; a case where the absolute value of resistance
(or deformation) increases after unloading before the
response point reaching the skeleton curve.
The hysteresis model is formulated on the basis of resistancedeformation relations observed in
the laboratory tests. The loading program for a test should include the followings;
(1) At least two cycles of load reversals at an amplitude to study the decay in resistance at the
Loading
Unloading
Reloading
D
F
2
amplitude,
(2) Small deflection amplitude
excursion must be placed after a large
amplitude excursion to study the
sliptype behavior
A lateral loaddeflection relation of a
reinforced concrete member was
obtained from the test of a slender
column (Otani and Cheung, 1981). The
behavior was dominantly by flexure
although flexural cracks started to
incline due to the presence of high
shear stresses before flexural yielding.
The yielding of the longitudinal
reinforcement was observed in cycle 3.
The general hysteretic characteristics
can be summarized as follows:
(a) Stiffness changed due to the flexural cracking of concrete and the tensile yielding of the
longitudinal reinforcement (cycle 1);
(b) When a deflection reversal was repeated at the same newly attained maximum deformation
amplitude, the loading stiffness in the second cycle was noticeably lower than that in the first cycle,
although the resistance at the peak displacement was almost identical (cycles 3 and 4). This
reduction in stiffness is attributable to the formation of new cracks during loading cycle 3, and also to
a reduced stiffness of the longitudinal reinforcement in cycle 4 due to the Bauschinger effect.
(c) Average peaktopeak stiffness of a complete cycle decreases with previous maximum
displacement. Note that the peaktopeak stiffness of cycle 5 is significantly smaller than that of cycle
2, although the displacement amplitudes of the two cycles are comparable. The peaktopeak
stiffness of cycle 5 is closer to that of cycles 3 and 4;
(d) The hysteresis characteristics of reinforced concrete are dependent on the loading history,
and
(e) The resistance at the peak deflection is almost the same for the two successive cycles in the
member dominated by flexural behavior.
A hysteresis model of a reinforced concrete "flexural" member must be able to represent the
above characteristics. The skeleton curve is similar to an "envelope curve" of a forcedeformation
relation under load reversals. The state of the art is not sufficient to determine the ultimate point, at
the deformation of which the resistance of a member starts to decay. The forcedeformation relation
after the onset of strength decay is normally not modeled because the behavior is strongly
dependent on a particular local deterioration of materials.
If the reinforced concrete is subjected to
high shear stress reversals, or if the
slippage of the reinforcement from concrete
within the anchorage area occurs, the
forcedeflection curve exhibits a pronounced
"pinching". The pinching behavior is also
observed;
(a) in a "flexural" member when the
amount of longitudinal reinforcement differs
significantly for the tension and compression
sides at the critical sections, typically in a
girder with monolithically cast slabs,
(b) at a member end where additional
deformation may be caused by anchorage
slip of longitudinal reinforcement within the
adjacent member or connection, and
Hysteresis of slip type (Bertero and Popov, 1977)
3
(c) in a member where bond splitting cracks develop along the longitudinal reinforcement.
Because such hysteresis relationship is highly dependent on loading history and structural
properties of the member, a general hysteresis model is difficult to formulate; or the parameters of
hysteresis models cannot be analytically determined by the properties of the member. In the design
of earthquake resistant structures, the pinching type behavior is generally thought to be undesirable
because small hysteresis energy can be dissipated by the behavior. Therefore, a proper design care
must be exercised to reduce such pinching behavior due to shear and bond deterioration.
Many hysteresis models have been developed in the past. Some hysteresis models are elaborate,
and include many hysteresis rules; others are simple. The complicatedness of a hysteresis model
indicates a large memory to store the hysteresis rule program in a computer. It does not lead to a
longer computation time because the complicatedness of a hysteresis model requires simply many
branches in a computer program, and only a few branches are referred to for a step of response
computation.
A class of hysteresis models, in which the unloading and reloading relation is defined by
enlarging the skeleton curve by a factor of two, are called "Masing type." Some examples of Masing
type models are shown below:
A hysteresis energy dissipation index (E
h
) is
used to express the amount of hysteresis energy
dissipation W ∆ per cycle during displacement
reversals of equal amplitudes in the positive and
negative directions;
m m
h
D F
W
E
π 2
∆
=
in which F
m
: resistance at peak displacement D
m
.
The value of the index was derived by equating
the area of hysteresis and the energy W ∆
dissipated by an equivalent viscous damper of a
linearly elastic system in one cycle under the
"resonant" "steadystate" oscillation.
The steady state response amplitude
m
D
under sinusoidal excitation with amplitude
o
p
and circular frequency ω , is given by
Hysteresis energy dissipation index
4
) sin( ) (
) ( 4 } ) ( 1 {
1
2 2 2
φ ω
ω
ω
ω
ω
+ =
+ −
=
t D t x
h
k
p
D
m
n n
o
m
The energy dissipated W ∆ by viscous damper per cycle is
2
2
2 2
0 0
2
2
) ( cos ) )( (
m
m
T T
m
D mk h
D c
dt t D c dt
dt
dx
dt
dx
c W
n n
ω π
ω π
φ ω ω
=
=
+ = = ∆
∫ ∫
where k c m , , ,: mass, damping coefficient and stiffness of an SDF system, h : damping factor
(
k m
c
2
= ),
n
T : natural period of the system (
k
m
π 2 = ),
n
ω : circular frequency of the system
(
m
k
= ).
At the resonant condition (
n
ω ω = ), the energy dissipated per cycle can be expressed
2
2
m
D k h W π = ∆
Therefore, the damping factor corresponding to the hysteresis energy dissipation W ∆ is
m
m
m m
m
D
F
k
D F
W
D k
W
h
=
∆
=
∆
=
π
π
2
2
2
The equivalent damping factor should not be confused with a damping factor of a viscously
damped system because the equivalent damping factor is not relevant in random oscillation.
References:
Bertero, V. V., and E. P. Popov, "Seismic Behavior of Ductile Moment Resisting Reinforced
Concrete Frames," ACI SP53, American Concrete Institute, Detroit, 1977, pp. 247291.
Comite EuroInternational du Beton: RC Frames under Earthquake Loading, State of the Art Report,
Thomas Telford, 1996.
Otani, S, "Hysteresis Models of Reinforced Concrete for Earthquake Response Analysis," Journal,
Faculty of Engineering, University of Tokyo, Vol. XXXVI, No. 2, 1981, pp. 125156.
Otani, S., and V. W.T. Cheung, "Behavior of Reinforced Concrete Columns Under Biaxial Lateral
Load Reversals  (II) Test Without Axial Load," Publication 8102, Department of Civil
Engineering, University of Toronto, 1981.
Saatcioglu, M., "Modeling Hysteretic ForceDeformation Relationships for Reinforced Concrete
Elements," ACISP127, American Concrete Institute, Detroit, 1991, pp. 153198.
5
11.2 Bilinear Model
At the initial development stage of nonlinear dynamic analysis, the elasticperfectly plastic
hysteretic model ("elastoplastic model") was used by many investigators. The response point
moves on the elastic stiffness line before the yield stress is reached. After yielding, the response
point moves on the perfectly plastic line until unloading takes place. Upon unloading, the response
point moves on the line parallel to the initial elastic line.
This model does not consider degradation of stiffness under cyclic loading. Energy dissipation
during a small excursion is not included.
A finite positive slope was assigned to
the stiffness after yielding to simulate the
strain hardening characteristics of the
steel and the reinforced concrete
("bilinear model"). Unloading stiffness
after yielding is equal to the initial elastic
stiffness. The stiffness degradation with
inelastic deformation and energy
dissipation during small amplitude
oscillation are not considered in the
model.
Neither the elastoplastic model nor
the bilinear model represents the
behavior of reinforced concrete and steel
members. The steel member softens
during reloading after plastic deformation
by the "Bauschinger effect." The
response of the elastoplastic model is
compared with a test result of a
reinforced concrete column above.
When the degradation in stiffness was recognized in the behavior of the reinforced concrete, the
loading and unloading stiffness K
r
was proposed to degrade with the previous maximum
displacement (Nielsen and Imbeault, 1970) in a form:
α −
= ) (
y
m
y r
D
D
K K
in which, α : unloading stiffness degradation parameters (0 < α <1); K
y
: initial elastic stiffness, and
D
m
: previously attained maximum displacement in any direction. The unloading stiffness remains
Bilinear Model
Specimen SP5
Column top displacement, cm
C
o
l
u
m
n
t
o
p
f
o
r
c
e
,
k
N
Response of Bilinear model and RC column
6
constant until the response displacement amplitude exceeds the previous maximum displacement in
either direction. The model is called a "degrading" bilinear hysteresis model." If the value of a is
chosen to be zero, the unloading stiffness does not degrade with yielding. A smaller value of a tends
to yield a larger residual displacement. The degrading bilinear model does not dissipate hysteretic
energy until the yield is developed. For a reinforced concrete member, the value of α is normally
selected to be around 0.4.
The hysteretic energy dissipation index E
h
of
the degrading bilinear model is given by
) 1 )( 1 (
)} 1 ( ){ 1 ( 2
α
α
βµ µβ β µ π
µβ β µ µ β
− + −
+ − − −
=
h
E
in which β : ratio of the postyielding stiffness
to the initial elastic stiffness; and µ : "ductility
factor" (ratio of the maximum displacement to
the initial yield displacement).
The equation is valid for a ductility factor
greater than 1.0. The hysteresis energy index of
a regular bilinear model (α = 0) reaches as
high as 0.33 at a ductility factor of 4.0. However,
such large amplitude oscillations do not
continue during an earthquake; no hysteresis
energy is dissipated by the model during small
amplitude oscillations. The total energy
dissipation of the bilinear model over the
duration of an earthquake is much smaller than
that expected from the hysteretic energy
dissipation index.
Reference:
Nielsen, N. N., and F. A. Imbeault, "Validity of
Various Hysteretic Systems," Proceedings,
Third Japan National Conference on
Earthquake Engineering, 1971, pp. 707714.
7
Appendix FORTRAN PROGRAM LISTING OF BILINEAR HYSTERESIS MODEL
SUBROUTINE HYSTR1 (LL,SS,DD,DS,FF,FS)
C
C BILINEAR HYSTERESIS RULES FOR A GIVEN DISPLACEMENT INCREMENT.
C
C PROGRAMMED BY OTANI, S.
C ON NOVEMBER 11,1978
C AT THE UNIVERSITY OF TORONTO.
C
C INPUT DATA
C FY INITIAL YIELD FORCE LEVEL.
C DY INITIAL YIELD DISPLACEMENT.
C SY INITIAL ELASTIC STIFFNESS.
C SU POST YIELDING STIFFNESS.
C B0 STIFFNESS DEGRADATION FACTOR.
C LL HYSTERESIS RULE POINTER AT PREVIOUS STEP.
C DD DISPLACEMENT AT CURRENT STEP.
C DS DISPLACEMENT AT PREVIOUS STEP.
C
C OUTPUT DATA
C LL HYSTERESIS RULE POINTER AT CURRENT STEP.
C SS STIFFNESS AT CURRENT STEP.
C FF FORCE AT CURRENT STEP.
C
C VARIABLES
C ES VARIABLE ELASTIC STIFFNESS AFTER FIRST YIELDING
C DU UPPER LIMIT DISPLACEMENT FOR POSTYIELDING ELASTIC STAGE.
C DL LOWER LIMIT DISPLACEMENT FOR POSTYIELDING ELASTIC STAGE.
C FL LOWER LIMIT FORCE FOR POSTYIELDING ELASTIC STAGE.
C DMX ABSOLUTE MAXIMUM DISPLACEMENT.
C
C
C BILINEAR HYSTERESIS RULES
C
C RULE 1 INITIAL ELASTIC STAGE.
C FF=SY*DD
C RULE 2 POSTYIELDING STAGE IN POSITIVE DIRECTION.
C FF= FY+(DDDY)*SU
C RULE 3 POSTYIELDING STAGE IN NEGATIVE DIRECTION.
C FF=FY+(DD+DY)*SU
C RULE 4 POSTYIELDING ELASTIC STAGE.
C FF=FL+(DDDL)*SS
C
COMMON /STFF/MD,DC,DY,FC,FY,SC,SY,SU,B0,B1
GO TO (1,2,3,4),LL
C INITIAL ELASTIC STAGE.
1 IF (DYABS(DD)) 110,110,100
100 FF=SY*DD
GO TO 1000
110 DMX=DY
IF (DD) 300,200,200
C POSTYIELDING STAGE IN POSITIVE DIRECTION.
2 IF (DDDS) 220,220,210
200 LL=2
SS=SU
210 FF= FY+(DDDY)*SU
GO TO 1000
220 IF (DMX.LT.ABS(DS)) DMX=ABS(DS)
ES=SY*(DY/DMX)**B0
DU=DS
DL=(FS+FYDY*SUDU*ES)/(SUES)
FL=FY+(DL+DY)*SU
IF (DDDL) 300,300,400
C POSTYIELDING STAGE IN NEGATIVE DIRECTION.
3 IF (DDDS) 310,320,320
300 LL=3
SS=SU
8
310 FF=FY+(DD+DY)*SU
GO TO 1000
320 IF (DMX.LT.ABS(DS)) DMX=ABS(DS)
ES=SY*(DY/DMX)**B0
DL=DS
FL=FY+(DL+DY)*SU
DU=(FYFL+DL*ESDY*SU)/(ESSU)
IF (DDDU) 400,200,200
C POSTYIELDING ELASTIC STAGE.
4 IF (DDDU) 420,200,200
400 LL=4
SS=ES
410 FF=FL+(DDDL)*ES
GO TO 1000
420 IF (DDDL) 300,300,410
C
1000 RETURN
END
9
11.3 RambergOsgood Model
A stressstrain relation of the metal was expressed using three parameters
y
D ,
y
F and γ by
Ramberg and Osgood (1943), where
y
D : yield displacement,
y
F : yield resistance and γ : a
parameter of the model. Jennings (1963) introduced the fourth parameter η to the model. The
initial loading curve of the model under monotonically increasing deformation, as modified by
Jennings, is expressed by
) 1 (
1 −
+ =
γ
η
y y y
F
F
F
F
D
D
in which, γ : exponent of the RambergOsgood model; and η : parameter introduced by Jennings
(1963).
The initial tangent modulus is equal to (F
y
/D
y
), and the initial loading curve passes a point (F
y
,
(1+η )D
y
) for any value of γ . The shape of the primary curve can be controlled by the exponent γ
from linearly elastic (γ = 1.0) to elastoplastic (γ = infinity). For a larger value of γ , the behavior
becomes similar to that of the bilinear model.
Upon unloading from a peak response point (D
o
, F
o
), the unloading, load reversal and reloading
branches of the relationship is given by
)
2
1 (
2 2
1 −
−
+
−
=
−
γ
η
y
o
y
o
y
o
F
F F
F
F F
D
D D
until the response point reaches the peak point of one outer hysteresis loop.
The resistance F is not explicitly expressed by a given displacement D in this model. The
resistance F at a given displacement D must be computed numerically, for example, using the
NewtonRapson's iterative procedure.
The RambergOsgood model is often used for stressstrain relation of the steel in the finite
element analysis or in the lamina model, and for resistancedeformation relation of steel members in
a frame analysis.
The hysteresis energy dissipation index of the RambergOsgood model is expressed as
) 1 )(
1
2
1 (
2
m
m
y
y
h
D
F
F
D
E −
+
− =
γ
η
π
The model can dissipate some hysteresis energy even if the ductility factor is less than unity. The
10
index is sensitive to the exponent γ
of the model, and the hysteresis
energy dissipation capacity increases
with increasing value of the exponent.
References:
Jennings, P. C., "Response of Simple
Yielding Structures to Earthquake
Excitation," Ph.D. Thesis,
California Institute of Technology,
Pasadena, 1963.
Ramberg, W., and W. R. Osgood,
"Description of StressStrain
Curves by Three Parameters,"
National Advisory Committee on
Aeronautics, Technical Note 902,
1943.
11
SUBROUTINE HYSTR2 (LL,SS,DD,DS,FF,FS)
C
C RAMBERGOSGOOD HYSTERESIS MODEL
C
C PROGRAMMED BY OTANI, S.
C ON FEBRUARY 9, 1979
C AT UNIVERSITY OF TORONTO
C
C
C (DD/DY)=(FF/FY)*(1.0+ABS(FF/FY)**(B01.0))
C
C THIS PROGRAM FINDS THE RESISTANCE AT A GIVEN DISPLACEMENT.
C THE NEWTONRAPSON'S METHOD IS USED TO SOLVE THE NONLINEAR PROBLEM.
C
C
C INPUT DATA
C LL HYSTERESIS RULE POINTER
C DD CURRENT DISPLACEMENT
C DS PREVIOUS DISPLACEMENT
C FS PREVIOUS RESISTANCE
C FY YIELD RESISTANCE
C DY YIELD DISPLACEMENT
C B0 RAMBERGOSGOOD PARAMETER
C B1 CONVERGENCE LIMIT
C
C OUTPUT DATA
C FF CURRENT RESISTANCE
C SS CURRENT TANGENT STIFFNESS
C
COMMON /STFF/MD,DC,DY,FC,FY,SC,SY,SU,B0,B1
C
GO TO (1,2,3,4,5,6,7,8,9,10), LL
C RULE 1 LOADING ON PRIMARY CURVE
1 CONTINUE
IF ((DDDS)*DS) 120,110,110
100 CONTINUE
LL=1
110 CONTINUE
XX=DD/DY
QQ=FS/FY
CALL RAMOSG (XX,QQ,B0,B1)
FF=QQ*FY
SS=FY/DY/XX
GO TO 10000
120 CONTINUE
D0=DS
F0=FS
130 CONTINUE
IF (ABS(DD)ABS(D0)) 200,200,100
C RULE 2 UNLOADING FROM POINT (D0,F0) ON PRIMARY CURVE
2 CONTINUE
IF ((DSD0)*(DDDS)) 230,220,220
200 CONTINUE
LL=2
210 CONTINUE
XX=(DDD0)/(DY+DY)
QQ=(FSF0)/(FY+FY)
CALL RAMOSG (XX,QQ,B0,B1)
FF=QQ*(FY+FY)+F0
SS=FY/DY/XX
GO TO 10000
220 CONTINUE
IF (ABS(DD)ABS(D0)) 210,210,100
230 CONTINUE
D1=DS
F1=FS
240 IF (ABS(DD)ABS(D0)) 300,300,100
C RULE 3 UNLOADING FROM POINT (D1,F1) ON FIRST INNER LOOP
12
3 CONTINUE
IF ((DSD1)*(DDDS)) 330,320,320
300 CONTINUE
LL=3
310 CONTINUE
XX=(DDD1)/(DY+DY)
QQ=(FSF1)/(FY+FY)
CALL RAMOSG (XX,QQ,B0,B1)
FF=QQ*(FY+FY)+F1
SS=FY/DY/XX
GO TO 10000
320 CONTINUE
IF (ABS(DD)ABS(D0)) 310,310,100
330 CONTINUE
D2=DS
F2=FS
340 CONTINUE
IF ((D1D2)*(DDD1)) 400,130,130
C RULE 4 UNLOADING FROM POINT (D2,F2) ON FIRST INNER LOOP
4 CONTINUE
IF ((D1D2)*(DDDS)) 430,420,420
400 CONTINUE
LL=4
410 CONTINUE
XX=(DDD2)/(DY+DY)
QQ=(FSF2)/(FY+FY)
CALL RAMOSG (XX,QQ,B0,B1)
FF=QQ*(FY+FY)+F2
SS=FY/DY/XX
GO TO 10000
420 CONTINUE
IF ((D1D2)*(DDD1)) 410,130,130
430 CONTINUE
D3=DS
F3=FS
440 CONTINUE
IF ((D1D2)*(DDD2)) 240,240,500
C RULE 5 UNLOADING FROM POINT (D3,F3) ON SECOND INNER LOOP
5 CONTINUE
IF ((D2D3)*(DDDS)) 530,520,520
500 CONTINUE
LL=5
510 CONTINUE
XX=(DDD3)/(DY+DY)
QQ=(FSF3)/(FY+FY)
CALL RAMOSG (XX,QQ,B0,B1)
FF=QQ*(FY+FY)+F3
SS=FY/DY/XX
GO TO 10000
520 CONTINUE
IF ((D2D3)*(DDD2)) 510,240,240
530 CONTINUE
D4=DS
F4=FS
540 CONTINUE
IF ((D2D3)*(DDD3)) 340,340,600
C RULE 6 UNLOADING FROM POINT (D4,F4) ON SECOND INNER LOOP
6 CONTINUE
IF ((D3D4)*(DDDS)) 630,620,620
600 CONTINUE
LL=6
610 CONTINUE
XX=(DDD4)/(DY+DY)
QQ=(FSF4)/(FY+FY)
CALL RAMOSG (XX,QQ,B0,B1)
FF=QQ*(FY+FY)+F4
SS=FY/DY/XX
GO TO 10000
13
620 CONTINUE
IF ((D3D4)*(DDD3)) 610,340,340
630 CONTINUE
D5=DS
F5=FS
640 CONTINUE
IF ((D3D4)*(DDD4)) 440,440,700
C RULE 7 UNLOADING FROM POINT (D5,F5) ON THIRD INNER LOOP
7 CONTINUE
IF ((D4D5)*(DDDS)) 730,720,720
700 CONTINUE
LL=7
710 CONTINUE
XX=(DDD5)/(DY+DY)
QQ=(FSF5)/(FY+FY)
CALL RAMOSG (XX,QQ,B0,B1)
FF=QQ*(FY+FY)+F5
SS=FY/DY/XX
GO TO 10000
720 CONTINUE
IF ((D4D5)*(DDD4)) 710,440,440
730 CONTINUE
D6=DS
F6=FS
740 CONTINUE
IF ((D4D5)*(DDD5)) 540,540,800
C RULE 8 UNLOADING FROM POINT (D6,F6) ON THIRD INNER LOOP
8 CONTINUE
IF ((D5D6)*(DDDS)) 830,820,820
800 CONTINUE
LL=8
810 CONTINUE
XX=(DDD6)/(DY+DY)
QQ=(FSF6)/(FY+FY)
CALL RAMOSG (XX,QQ,B0,B1)
FF=QQ*(FY+FY)+F6
SS=FY/DY/XX
GO TO 10000
820 CONTINUE
IF ((D5D6)*(DDD5)) 810,540,540
830 CONTINUE
D7=DS
F7=FS
840 CONTINUE
IF ((D5D6)*(DDD6)) 640,640,900
C RULE 9 UNLOADING FROM POINT (D7,F7) ON FOURTH INNER LOOP
9 CONTINUE
IF ((D6D7)*(DDDS)) 930,920,920
900 CONTINUE
LL=9
910 CONTINUE
XX=(DDD7)/(DY+DY)
QQ=(FSF7)/(FY+FY)
CALL RAMOSG (XX,QQ,B0,B1)
FF=QQ*(FY+FY)+F7
SS=FY/DY/XX
GO TO 10000
920 CONTINUE
IF ((D6D7)*(DDD6)) 910,640,640
930 CONTINUE
D8=DS
F8=FS
940 CONTINUE
IF ((D6D7)*(DDD7)) 740,740,1000
C RULE 10 LINEAR RELATION BETWEEN POINTS (D7,F7) AND (D8,F8)
10 CONTINUE
IF ((D7D8)*(DDD8)) 840,1020,1020
1000 CONTINUE
14
LL=10
1010 CONTINUE
SS=(F7F8)/(D7D8)
FF=F8+(DDD8)*SS
GO TO 10000
1020 CONTINUE
IF ((D7D8)*(DDD7)) 1010,740,740
C
10000 CONTINUE
RETURN
END
SUBROUTINE RAMOSG (XX,QQ,B0,ERR)
C
C DETERMINATION OF FORCE LEVEL QQ AT A GIVEN DISPLACEMENT XX OF A
C RAMBERGOSGOOD HYSTERESIS MODEL BY THE NEWTONRAPSON'S ITERATIVE
C PROCEDURE.
C
C XX=QQ*(1.0+ABS(QQ)**(B01.0))
C
C
C INPUT DATA
C XX CURRENT DISPLACEMENT
C QQ INITIAL VALUE OF RESISTANCE
C B0 RAMBERGOSGOOD PARAMETER
C ERR CONVERGENCE LIMIT
C
C OUTPUT
C QQ CURRENT RESISTANCE OF THE HYSTERESIS MODEL
C XX CURRENT STIFFNESS
C =1.0+B0*ABS(QQ)**(B01.0)
C
10 CONTINUE
Q0=QQ
EX=ABS(Q0)**(B01.0)
GQ=Q0*(1.0+EX)XX
TQ=1.0+B0*EX
QQ=Q0GQ/TQ
IF (ABS(QQQ0)/(ABS(Q0)+ABS(QQ)).LT.ERR) GO TO 20
GO TO 10
20 CONTINUE
XX=1.0+B0*ABS(QQ)**(B01.0)
RETURN
END
15
11.4 Degrading Trilinear Model
A model that simulates dominantly flexural stiffness characteristics of the reinforced concrete was
used extensively in Japan (Fukada, 1969). The primary curve is of trilinear shape with stiffness
changes at flexural cracking and yielding. Up to yielding, the model behaves in a manner the same
as the bilinear model. When the response exceeds a yield point, response point follows the
strainhardening part of the
primary curve. Once
unloading takes place from
a point on the primary curve,
the unloading point is
considered to be a new
"yield point" in the direction.
The model behaves in a
bilinear manner between
the positive and negative
"yield points" with stiffness
degraded proportional to
the ratio of the slopes
connecting "current yield
points" and "the initial yield
points."
The ratio of the first and second stiffness is kept constant even after yielding.
This model has the following properties:
(a) the stiffness continuously degrades
with increasing maximum amplitude beyond
yielding,
(b) the hysteretic energy dissipation is
large in the first load reversal cycle after
yielding, and becomes steady in the following
cycles, and
(c) the steady hysteretic energy dissipation
is proportional to the displacement amplitude.
The hysteretic energy dissipation index of
the degrading trilinear model is expressed as
y
c
c
y
h
F
F
K
K
E ) 1 (
2
− =
π
in which K
y
: secant stiffness at yielding (=
y y
D F / ), and
c
K : initial elastic stiffness (=
c c
D F / ). The index is independent of the
displacement amplitude, but dependent on
the stiffness and resistance ratios at cracking
and yielding. Cracking point of this model
controls the fatness of a hysteresis loop.
Therefore, it is important to choose the
cracking point taking into account the degree
of a hysteresis loop.
Nomura (1976) used an arbitrary skeleton
curve; when the response point reached the previous maximum response point, it moves on the
skeleton curve. Upon unloading, the newly attained maximum response point was considered as the
yield point in the direction, similar to the degrading trilinear model.
Degrading trilinear model
16
References:
Fukada, Y., "Study on the Restoring Force
Characteristics of Reinforced Concrete Buildings
(in Japanese)," Proceedings, Kanto Branch
Symposium, Architectural Institute of Japan, No.
40, 1969, pp. 121124.
Nomura, S., "Restoring Characteristics and their
Modeling," Data for Earthquake Resistant
Design for Buildings, No. 65, Magazine of
Architectural Institute of Japan, June 1976.
Nomura model (1976)
17
SUBROUTINE HYST6 (LL,SS,DD,DS,FF,FS)
C
C DEGRADING TRILINEAR MODEL
C
C PROGRAMMED BY OTANI, S.
C ON FEBRUARY 9, 1979
C AT THE UNIVERSITY OF TORONTO.
C
C THE MODEL BEHAVES AS A BILINEAR MODEL BETWEEN POSITIVE (DX,FX) AND
C NEGATIVE (DN,FN) YIELD POINTS WITH A CHANGE OF STIFFNESS AT A
C CRACKING POINT (DU,FU) OR (DL,FL). ONCE THE RESPONSE EXCEEDS A
C YIELD DISPLACEMENT (DX OR DN), THE RESPONSE FOLLOWS THE PRIMARY
C CURVE. THE MAXIMUM RESPONSE POINT (DX,FX) OR (DN,FN) IS NOW
C TREATED AS A NEW YIELD POINT IN THAT DIRECTION. THE STIFFNESSES
C ARE MODIFIED ACCORDINGLY.
C
C REFERENCE
C FUKADA, Y.,'STUDY ON THE RESTORING FORCE CHARACTERISTICS OF
C REINFORCED CONCRETE BUILDINGS (IN JAPANESE)', PROCEEDINGS,KANTO
C DISTRICT SYMPOSIUM, AIJ, NO.40, NOV. 1969, PP.1214.
C
C THE PROGRAM COMPUTES THE RESISTANCE OF THE MODEL AT A GIVEN
C DISPLACEMENT.
C
C
C INPUT DATA
C FC INITIAL CRACKING FORCE LEVEL
C DC INITIAL CRACKING DISPLACEMENT
C FY INITIAL YIELD FORCE LEVEL.
C DY INITIAL YIELD DISPLACEMENT.
C SC INITIAL ELASTIC STIFFNESS
C SY STIFFNESS AFTER INITIAL CRACKING, BUT BEFORE
C INITIAL YIELDING.
C SU POST YIELDING STIFFNESS.
C LL HYSTERESIS RULE POINTER AT PREVIOUS STEP.
C SS STIFFNESS AT PREVIOUS STEP.
C DD DISPLACEMENT AT CURRENT STEP.
C DS DISPLACEMENT AT PREVIOUS STEP.
C FS RESISTANCE AT PREVIOUS STEP
C
C OUTPUT DATA
C LL HYSTERESIS RULE POINTER AT CURRENT STEP.
C SS STIFFNESS AT CURRENT STEP.
C FF FORCE AT CURRENT STEP.
C
C VARIABLES
C DU UPPER LIMIT DISPLACEMENT FOR POSTCRACKING ELASTIC
C STAGE.
C DL LOWER LIMIT DISPLACEMENT FOR POSTCRACKING ELASTIC
C STAGE.
C FL LOWER LIMIT FORCE FOR POSTCRACKING ELASTIC STAGE.
C FU UPPER LIMIT FORCE FOR POSTCRACKING ELASTIC STAGE.
C DX POSITIVE MAXIMUM DISPLACEMENT
C DN NEGATIVE MAXIMUM DISPLACEMENT
C FX POSITIVE MAXIMUM RESISTANCE
C FN NEGATIVE MAXIMUM RESISTANCE
C B0 STIFFNESS DEGRADATION FACTOR.
C =(DY*(FXFN))/(FY*(DXDN))
C
C
C DEGRADING TRILINEAR HYSTERESIS RULES
C
C RULE 1 INITIAL ELASTIC STAGE.
C FF=SC*DD
C RULE 2 LOADING IN POSTCRACKING STAGE IN POSITIVE DIRECTION
C S1=SY*B0
C FF=FU+(DDDU)*S1
C RULE 3 LOADING IN POSTCRACKING STAGE IN NEGATIVE DIRECTION
18
C S1=SY*B0
C FF=FL+(DDDL)*S1
C RULE 4 POSTCRACKING ELASTIC STAGE
C S0=SC*B0
C FF=FL+(DDDL)*S0
C RULE 5 LOADING IN POSTYIELDING STAGE IN POSITIVE DIRECTION
C FF= FY+(DDDY)*SU
C RULE 6 LOADING IN POSTYIELDING STAGE IN NEGATIVE DIRECTION
C FF=FY+(DD+DY)*SU
C
COMMON /STFF/MD,DC,DY,FC,FY,SC,SY,SU,B0,B1
C
GO TO (1,2,3,4,5,6), LL
C RULE 1 INITIAL ELASTIC STGE BEFORE CRACKING AT (DC,FC)
1 IF (DCABS(DD)) 110,110,100
100 FF=SC*DD
GO TO 1000
110 DX= DY
DN=DY
FX= FY
FN=FY
S0=SC
S1=SY
IF (DD.GT.DY) GO TO 500
IF (DD.LT.DY) GO TO 600
IF (DD) 300,200,200
C RULE 2 LOADING IN POSTCRACKING STAGE IN POSITIVE DIRECTION
2 IF (DDDS) 230,230,220
200 LL=2
SS=S1
210 FF=FX+(DDDX)*S1
GO TO 1000
220 IF (DXDD) 500,500,210
230 DU=DS
FU=FS
DL=(FUFN+DN*S1DU*S0)/(S1S0)
FL=FU+(DLDU)*S0
IF (DLDD) 400,240,240
240 IF (DNDD) 300,600,600
C RULE 3 LOADING IN POSTCRACKING STAGE IN NEGATIVE DIRECTION
3 IF (DDDS) 320,330,330
300 LL=3
SS=S1
310 FF=FN+(DDDN)*S1
GO TO 1000
320 IF (DNDD) 310,600,600
330 DL=DS
FL=FS
DU=(FXFLDX*S1+DL*S0)/(S0S1)
FU=FL+(DUDL)*S0
IF (DUDD) 340,340,400
340 IF (DXDD) 500,500,200
C RULE 4 POSTCRACKING ELASTIC STAGE
4 IF (DUDD) 340,340,420
400 LL=4
SS=S0
410 FF=FL+(DDDL)*S0
GO TO 1000
420 IF (DLDD) 410,240,240
C RULE 5 LOADING IN POSTYIELDING STAGE IN POSITIVE DIRECTION
5 IF (DDDS) 520,520,510
500 LL=5
SS=SU
510 FF= FY+(DDDY)*SU
GO TO 1000
520 DX=DS
FX=FS
B0=DY*(FXFN)/(FY*(DXDN))
19
S0=SC*B0
S1=SY*B0
GO TO 230
C RULE 6 LOADING IN POSTYIELDING STAGE IN NEGATIVE DIRECTION
6 IF (DDDS) 610,620,620
600 LL=6
SS=SU
610 FF=FY+(DD+DY)*SU
GO TO 1000
620 DN=DS
FN=FS
B0=DY*(FXFN)/(FY*(DXDN))
S0=SC*B0
S1=SY*B0
GO TO 330
C
1000 RETURN
END
20
11.5 Clough Degrading Model
A hysteretic model with an elastoplastic
skeleton curve was proposed by Clough
and Johnston (1966) to represent the
hysteretic behavior of a reinforced concrete
beamcolumn subassemblage.
During loading, the response point
follows the elastoplastic skeleton curve.
The unloading stiffness after yielding was
kept equal to the initial elastic stiffness. The
response point during reloading moves
toward the previous maximum response
point in the direction of reloading, simulating
the stiffness degradation. If yielding has not
taken place in the direction of reloading, the
response point moves toward the yield point
in the reloading direction.
A minor deficiency of the Clough model
was pointed out by Mahin and Bertero
(1976). After unloading from point A,
consider a situation in which reloading takes
place from point B. The original Clough
model assumed that the response point
should move toward the previous maximum
response point C. This is not realistic.
Therefore, a minor modification was added
so that the response point should move
toward an immediately preceding unloading
point A during reloading. When the
response point reaches the point A, the
response point moves toward the previous maximum point C.
The model was made more versatile by incorporating the reduction in unloading stiffness K
r
with
a maximum displacement in a form:
α −
= ) (
y
m
y r
D
D
K K
in which, α : unloading stiffness
degradation parameter;
y
K : initial elastic
stiffness; and
m
D : previous maximum
displacement. The different unloading
stiffness may be assigned taking
m
D to be
a maximum deformation in the direction
unloading takes place.
If the value of a is chosen to be zero, the
unloading stiffness of the model remains
equal to the initial elastic stiffness.
The response of the Clough model is
shown to compare well with the response of
a reinforced concrete column tested in the
structures laboratory.
Clough Model
RC Column
Column Top Displacement, mm
C
o
l
u
m
n
R
e
s
i
s
t
a
n
c
e
,
k
N
100
50
0
50 100
100
200
0
100
200
D
F
B
C
Y
Y
K
r
=K
y
K
y
A
Clough Model
D
F
B
C
Y
Y
K
r
K
y
D
m
D
y
F
y
A
Modified Clough Model
21
Saiidi and Sozen (1979) and Riddell and Newmark (1979) used models similar to the modified
Clough model.
Wang and Shah (1987) introduced the strength and stiffness degradation effect of cumulative
damage. The strength and stiffness degrade in proportion to (1D
ws
), where D
ws
is the Wang and
Shah damage index. The ordinates of the bilinear skeleton curve in monotonic loading is multiplied
by the current value of (1D
ws
). Unloading and reloading stiffness is reduced by the same amount, as
they are defined on the basis of the location of the point of reversal and of the maximum previous
deformation in the direction of loading, on the degraded skeleton curve. The Wang and Shah
damage index is defined separately for each direction of loading as
1
1
n
ws
n
e
D
e
δ
−
=
−
where the damage prameter δ is expressed in terms of chord rotation,
i
i
u
c
θ
δ
θ
=
∑
The hysteretic energy dissipation index
of the modified Clough model is expressed
as
}
) 1 (
1 {
1
µ
µ µβ β
π
α
+ −
− =
h
E
where β : ratio of postyielding stiffness to
the initial elastic stiffness, and µ : ductility
factor.
The equation is valid for ductility factor
greater than unity. The Clough model can
continuously dissipate hysteretic energy
even at a small amplitude oscillation after
yielding.
References:
Clough, R. W., and S. B. Johnston, "Effect
of Stiffness Degradation on
Earthquake Ductility Requirements,"
Proceedings, Second Japan National
Conference on Earthquake
Engineering, 1966, pp. 227232.
Mahin, S. A., and V. V. Bertero, "Rate of
Loading Effect on Uncracked and Repaired Reinforced Concrete Members," EERC No. 736,
Earthquake Engineering Research Center, University of California, Berkeley, 1972.
Riddell, R., and N. M. Newmark, "Statistical Analysis of the Response of Nonlinear Systems
subjected to Earthquakes," Structural Research Series No. 468, Civil Engineering Studies,
University of Illinois at UrbanaChampaign, Illinois, 1979.
Saiidi, M., and M. A. Sozen, "Simple and Complex Models for Nonlinear Seismic Response of
Reinforced Concrete Structures," Structural Research Series No. 465, Civil Engineering Studies,
University of Illinois at UrbanaChampaign, Illinois, 1979.
Wang, M.L., and S. P. Shah, “Reinforced Concrete Hysteresis Model based on the Damage
Concept,” Earthquake Engineering and Structural Dynamics, John Wiley & Sons, Chichester,
Sussex, Vol. 15, 1987, pp. 993 1003.
22
SUBROUTINE HYSTR3 (LL,SS,DD,DS,FF,FS)
C
C CLOUGH'S DEGRADING HYSTERESIS MODEL WITH TWO MODIFICATIONS.
C (1) UNLOADING STIFFNESS DEGRADES WITH MAXIMUM DISPLACEMENT ON
C THE SIDE OF UNLOADING POINT,
C (2) RELOADING STIFFNESS, AFTER UNLOADING IN THE SAME STRESS
C REGION, IS THE SAME AS THE UNLOADING STIFFNESS UP TO THE
C INNER UNLOADING POINT.
C
C PROGRAMMED BY OTANI, S.
C ON FEBRUARY 9, 1979
C AT UNIVERSITY OF TORONTO
C
C
C THE CLOUGH MODEL ASSUMES THE YIELD POINT TO BE THE PREVIOUS MAXIMUM
C DISPLACEMENT POINT IN THE DIRECTION OF LOADING.
C THE RESISTANCE INCREASES TOWARD THIS VARIABLE YIELD POINT IN THE
C DIRECTION OF LOADING. THE UNLOADING STIFFNESS IS MODIFIED IN THIS
C PROGRAM TO DEGRADE WITH THE VALUE OF MAXIMUM DISPLACEMENT ON THE
C SIDE OF UNLOADING POINT. THIS PROGRAM RETURNS TO THE CALLING
C PROGRAM WITH NEW HYSTERESIS POINT VALUES.
C
C
C INPUT INFORMATION
C DY INITIAL YIELD DISPLACEMENT
C FY INITIAL YIELD RESISTANCE
C SY ELASTIC STIFFNESS
C SU POSTYIELDING STIFFNESS
C B0 CONSTANT DEFINING DEGRADATION OF STIFFNESS
C LL HYSTERESIS CASE NUMBER AT PREVIOUS STEP
C DD DISPLACEMENT AT CURRENT STEP
C FS RESISTANCE AT PREVIOUS STEP
C DS DISPLACEMENT AT PREVIOUS STEP
C
C OUTPUT INFORMATION
C LL HYSTERESIS CASE NUMBER AT CURRENT STEP
C SS STIFFNESS AT CURRENT STEP
C FF RESISTANCE AT CURRENT STEP
C
C VARIABLES
C FX(1) NEGATIVE MAXIMUM RESISTANCE REACHED
C FX(2) POSITIVE MAXIMUM RESISTANCE REACHED
C DX(1) NEGATIVE MAXIMUM DISPLACEMENT REACHED
C DX(2) POSITIVE MAXIMUM DISPLACEMENT REACHED
C ES(1) UNLOADING STIFFNESS AFTER YIELDING IN NEGATIVE SIDE.
C ES(2) UNLOADING STIFFNESS AFTER YIELDING IN POSITIVE SIDE.
C XD DISPLACEMENT AT CROSSING OF ZERORESISTANCE AXIS IN
C HYSTERESIS RULE 3.
C F1 RESISTANCE AT INNER UNLOADING POINT
C D1 DISPLACEMENT AT INNER UNLOADING POINT
C X1 DISPLACEMENT AT CROSSING OF ZERORESISTANCE AXIS IN
C HYSTERESIS RULE 5.
C DC DUMMY VARIABLE
C FC DUMMY VARIABLE
C SC DUMMY VARIABLE
C B1 DUMMY VARIABLE
C
C
C CLOUGH HYSTERESIS RULES
C
C RULE 1 INITIAL ELASTIC STAGE
C FF=SY*DD
C RULE 2 POSTYIELDING STAGE
C FF=SN*FY+(DDSN*DY)*SU
C RULE 3 UNLOADING FROM POSTYIELDING POINT (DMX(IS),FMX(IS))
C FF=FX(IS)+(DDDX(IS))*ES(IS)
C RULE 4 LOADING TOWARD PREVIOUS MAXIMUM POINT (DMX(IS),FMX(IS))
C FF=FX(IS)*(DDXD)/(DX(IS)XD)
23
C RULE 5 UNLOADING FROM INNER PEAK POINT (D1,F1)
C FF=F1+(DDD1)*ES(IS)
C
COMMON /STFF/MD,DC,DY,FC,FY,SC,SY,SU,B0,B1
DIMENSION DX(2),FX(2),ES(2)
C
C IS=1 NEGATIVE RESISTANCE AT PREVIOUS STEP
C IS=2 POSITIVE RESISTANCE AT PREVIOUS STEP
IS=2
IF (FS.LT.0.0) IS=1
SN=FLOAT(IS+IS3)
C GO TO PREVIOUS HYSTERESIS RULE
GO TO (1,2,3,4,5), LL
C RULE 1 = INITIAL ELASTIC STAGE
1 IF (DYABS(DD)) 110,110,100
100 FF=SY*DD
GO TO 1000
110 DX(1)=DY
DX(2)= DY
FX(1)=FY
FX(2)= FY
ES(1)=SY
ES(2)=SY
GO TO 200
C RULE 2 = LOADING BEYOND YIELD POINT (DX,FX)
2 IF ((DDDS)*SN) 220,220,210
200 LL=2
SS=SU
210 FF=SN*FY+(DDSN*DY)*SU
GO TO 1000
220 DX(IS)=DS
FX(IS)=FS
ES(IS)=SY*(DY/ABS(DS))**B0
XD=DX(IS)FX(IS)/ES(IS)
IF ((XDDD)*SN) 300,230,230
230 IS=3IS
SN=FLOAT(IS+IS3)
240 IF ((DX(IS)DD)*SN) 200,200,400
C RULE 3 = UNLOADING FROM YIELD POINT (DMX,FMX)
3 IF ((XDDD)*SN) 320,230,230
300 LL=3
SS=ES(IS)
310 FF=FX(IS)+(DDDX(IS))*ES(IS)
GO TO 1000
320 IF ((DX(IS)DD)*SN) 200,200,310
C RULE 4 = LOADING TOWARD YIELD POINT (DMX,FMX)
4 IF ((DDDS)*SN) 430,430,420
400 LL=4
SS=FX(IS)/(DX(IS)XD)
410 FF=SS*(DDXD)
GO TO 1000
420 IF ((DX(IS)DD)*SN) 200,200,410
430 D1=DS
F1=FS
X1=D1F1/ES(IS)
IF ((X1DD)*SN) 500,440,440
440 XD=X1
GO TO 230
C RULE 5 = UNLOADING FROM INNER PEAK (D1,F1)
5 IF ((X1DD)*SN) 520,440,440
500 LL=5
SS=ES(IS)
510 FF=F1+(DDD1)*ES(IS)
GO TO 1000
520 IF ((D1DD)*SN) 240,240,510
C
1000 RETURN
END
24
11.6 Takeda Degrading Model
Based on the experimental observation on the behavior of a number of mediumsize reinforced
concrete members tested under lateral load reversals with light to medium amount of axial load, a
hysteresis model was developed by Takeda, Sozen and Nielsen (1970). The model has been widely
used in the nonlinear earthquake response analysis of reinforced concrete structures.
Takeda Hysteresis Model:
1. Condition: The cracking load, P
cr
, has not been exceeded in one direction. The load is reversed
from a load P in the other direction. The load P is smaller than the yield load P
y
.
Rule: Unloading follows a straight line from the position at load P to the point representing the
cracking load in the other direction.
2. Condition: A load P
1
is reached in one direction on the primary curve such that P
1
is larger than
P
cr
but smaller the yield load P
y
. The load is then reversed to P
2
such that P
2
<P
1
.
Rule: Unload parallel to loading curve for that half cycle.
3. Condition: A load P
1
is reached in one direction such that P
1
is larger than P
cr
but not larger
than the yield load P
y
. The load is then reversed to P
3
such that P
3
>P
1
.
Rule: Unloading follows a straight line joining the point of return and the point representing
cracking in the other direction.
4. Condition: One or more loading cycles have occurred. The load is zero.
Rule: To construct the loading curve, connect the point at zero load to the point reached in the
previous cycle, if that point lies on the primary curve or on a line aimed at a point on the primary
curve. If the previous loading cycle contains no such point, go to the preceding cycle and continue
the process until such a point is found. Then connect that point to the point at zero load.
Exception: If the yield point has not been exceeded and if the point at zero load is not located
within the horizontal projection of the primary curve for that direction of loading, connect the point at
zero load to the yield point to obtain the loading slope.
5. Condition: The yield load P
y
is exceeded in one direction.
Rule: Unloading curve follows the slope given by the following equation:
4 . 0
) (
D
D
k k
y
y r
=
in which
r
k : slope of unloading curve,
y
k : slope of a line joining the yield point in one direction to
the cracking point in the other direction, D: maximum deflection attained in the direction of the
loading, and
y
D : deflection at yield.
6. Condition: The yield load is exceeded in one direction but the cracking load is not exceeded in
the opposite direction.
Rule: Unloading follows Rule 5. Loading in the other direction continues as an extension of the
unloading line up to the cracking load. Then, the loading curve is aimed at the yield point.
7. Condition: One or more loading cycles have occurred.
Rule: If the immediately preceding quartercycle remained on one side of the zeroload axis,
unload at the rate based on rule 2, 3 and 5 whichever governed in the previous loading history. If the
immediately preceding quartercycle crossed the zeroload axis, unload at 70% of the rate based on
rule 2, 3, or 5, whichever governed in the previous loading history, but not at a slope flatter than the
immediately preceding loading slope.
Takeda model included (a) stiffness changes at flexural cracking and yielding, (b) hysteresis rules
for inner hysteresis loops inside the outer loop, and (c) unloading stiffness degradation with
deformation. The response point moves toward a peak of the one outer hysteresis loop. The
unloading stiffness K
r
after yielding is given by
25
α −
+
+
=
y
m
y c
y c
r
D
D
D D
F F
K
in which, α : unloading stiffness degradation parameter; and
m
D : previous maximum displacement
beyond yielding in the direction concerned. The hysteresis rules are extensive and comprehensive.
The hysteretic energy dissipation index of
the Takeda model is expressed as
}
) 1 (
1
1
1 {
1
µ
µβ β µ
π
α
+ −
+
+
− =
y
c
y
c
h
F
F
D
D
E
The expression is valid for a ductility factor
greater than unity.
It should be noted that the Takeda
hysteresis rule was originally developed to
simulate the behavior of reinforced concrete
members. If this model is used to simulate the
behavior of a story or a simplified structure,
some rules need to be simplified.
For example, hysteresis rules prior to
yielding may be simplified such that unloading
takes place toward the origin of the relation
(Muto Model). This model is often used in a
storybased (massspring) earthquake
response analysis.
26
Bilinear Takeda Model: The primary curve of the Takeda model can be made bilinear simply
choosing the cracking point to be the origin of the hysteretic plane. Such a model is called the
"bilinear Takeda" model, similar to the Clough model except that the bilinear Takeda model has
more hysteresis rules for inner hysteresis loops (Otani and Sozen, 1972); i.e., the response point
moves toward an unloading point on the immediately outer hysteresis loop.
The behavior before yielding is
sometimes made simple by letting the
response point moves toward the
origin during unloading, and toward
the maximum response point in the
opposite side upon reloading. The
Takeda hysteresis rules are applied
after the yielding.
This model is similar to Clough
Degrading Model, but is more
complicated having rules for inner
hysteresis loops.
Additional modifications of the Takeda model with bilinear backbone curve may be found in
literature (Powell, 1975, Riddle and Newmark, 1979, Saiidi and Sozen, 1979, Saiidi, 1982). Riddle
and Newmark (1979) used a bilinear skeleton curve and unloading stiffness equal to the initial elastic
stiffness; loading occurs either on the strain hardening branch or towards the furthest point attained
in the previous cycle. Saiidi and Sozen (1979) claimed to simplify the Takeda model using a bilinear
skeleton curve; the model, however, is identical to the modified Clough model with reduced
unloading stiffness with maximum deformation, and reloading to the immediate prior unloading point
if reloading occurs during unloading and then to the unloading point on the skeleton curve.
References:
Takeda, T., M. A. Sozen and N. N. Nielsen, "Reinforced Concrete Response to Simulated
Earthquakes," Journal, Structural Division, ASCE, Vol. 96, No. ST12, 1970, pp. 25572573.
Otani, S., and M. A. Sozen, "Behavior of Multistory Reinforced Concrete Frames During
Earthquakes," Structural Research Series No. 392, Civil Engineering Studies, University of
Illinois, Urbana, 1972.
Powell, G. H., “Supplement to Computer Program DRAIN2D,” Supplement to Report, DRAIN2D
User’s Guide, University of California, Berkeley, August 1975.
Riddle, R., and N. M. Newmark, “Statistical Analysis of the Response of Nonlinear Systems
subjected to Earthquakes,” Structural Research Series No. 468, Civil Engineering Studies,
University of Illinois, Urbana, 1979.
Saiidi, M., “Hysteresis Models for Reinforced Concrete,” Journal, Structural Division, ASCE, Vol. 108,
No. ST5, May 1982, pp. 1077  1087.
Saiidi, M., and M. A. Sozen, “Simple and Complex Models for Nonlinear Seismic Response of
Reinforced Concrete Structures,” Structural Research Series No. 465, Civil Engineering Studies,
University of Illinois, Urbana, 1979.
D
F
D
m
D’
m
X
0
(D
0
,F
0
)
X
1
(D
1
,F
1
)
X
3
(D
2
,F
2
)
(D
3
,F
3
)
27
SUBROUTINE HYSTR4 (LL,SS,DD,DS,FF,FS)
C
C BILINEAR TAKEDA MODEL
C
C PROGRAMMED BY OTANI, S.
C ON FEBRUARY 9, 1979
C AT UNIVERSITY OF TORONTO
C
C
C THE MODEL WAS SLIGHTLY MODIFIED FROM THE ORIGINAL TAKEDA MODEL.
C THE PRIMARY CURVE IS CHANGED TO A BILINEAR TYPE.
C
C MODIFICATION IS MADE SUCH THAT UNLOADING STIFFNESS IS DEGRADED
C WITH THE MAXIMUM DISPLACEMENT AMPLITUDE IN THE DIRECTION, RATHER
C THAN WITH THE ABSOLUTE MAXIMUM DISPLACEMENT AMPLITUDE.
C
C THIS PROGRAM FINDS A RESISTANCE OF THE TAKEDA HYSTERESIS MODEL AT
C A GIVEN DISPLACEMENT.
C
C
C INPUT DATA
C LL HYSTERESIS RULE POINTER AT PREVIOUS STEP.
C SS STIFFNESS AT PREVIOUS TIME STEP
C DD DISPLACEMENT AT PRESENT STEP
C DS DISPLACEMENT PREVIOUS TIME STEP
C FS FORCE AT PREVIOUS TIME STEP
C FY YIELDING FORCE
C DY YIELDING DISPLACEMENT
C SY DISPLACEMENT STIFFNESS BEFORE YIELDING
C SU DISPLACEMENT STIFFNESS AFTER YIELDING
C B0 STIFFNESS DEGRADATION FACTOR
C B1 STIFFNESS DEGRADATION FACTOR
C
C OUTPUT DATA
C LL HYSTERESIS RULE POINTER AT PRESENT STEP.
C SS STIFFNESS AT PRESENT STEP.
C FF FORCE AT PRESENT TIME STEP
C
C DESCRIPTION OF VARIABLES
C S1(IS) UNLOADING STIFFNESS IN OUTER HYSTERESIS LOOP
C S1(IS)=(DY/DM(IS))**B0*(FY/DY)
C S2(IS) UNLOADING STIFFNESS IN INNER HYSTERESIS LOOPS
C S2(IS)=S1(IS)*B1
C F0 UNLOADING FORCE LEVEL FROM STEP 6
C F1 UNLOADING FORCE LEVEL FROM STEP 8
C F2 UNLOADING FORCE LEVEL FROM STEP 10
C F3 UNLOADING FORCE LEVEL FROM STEP 12
C D0 UNLOADING DISPLACEMENT FROM STEP 6
C D1 UNLOADING DISPLACEMENT FROM STEP 8
C D2 UNLOADING DISPLACEMENT FROM STEP 10
C D3 UNLOADING DISPLACEMENT FROM STEP 12
C X0 INTERCEPT OF DISPLACEMENT AXIS WITH STEP 6
C X1 INTERCEPT OF DISPLACEMENT AXIS WITH STEP 8
C X2 INTERCEPT OF DISPLACEMENT AXIS WITH STEP 10
C X3 INTERCEPT OF DISPLACEMENT AXIS WITH STEP 12
C FM(*) UNLOADING FORCE LEVEL ON PRIMARY CURVE
C DM(*) UNLOADING DISPLACEMENT ON PRIMARY CURVE
C
COMMON /STFF/MD,DC,DY,FC,FY,SC,SY,SU,B0,B1
DIMENSION FM(2),DM(2),S1(2),S2(2)
C
C
IS=2
IF (FS.LT.0.0) IS=1
SN=FLOAT(IS+IS3)
C
GO TO (1,2,3,4,5,6,7,8,9,10,11), LL
C RULE 1 LINEARLY ELASTIC STAGE BEFORE YIELDING AT (DY,F!)
28
1 IF (DYABS(DD)) 110,110,100
100 FF=SY*DD
GO TO 10000
110 FM(1)=FY
FM(2)= FY
DM(1)=DY
DM(2)= DY
S1(1)= SY
S1(2)= SY
GO TO 200
C RULE 2 LOADING ON PRIMARY CURVE AFTER YIELDING AT (DY,FY),
C OR BEYOND PREVIOUS MAXIMUM POINT (DM,FM).
2 IF ((DDDS)*SN) 220,220,210
200 LL=2
SS=SU
210 FF=(FY+(ABS(DD)DY)*SU)*SN
GO TO 10000
220 FM(IS)=FS
DM(IS)=DS
S1(IS)=SY*(DY/ABS(DM(IS)))**B0
X0=DM(IS)FM(IS)/S1(IS)
IF ((X0DD)*SN) 300,230,230
230 IS=3IS
SN=FLOAT(IS+IS3)
240 IF ((DM(IS)DD)*SN) 200,200,400
C RULE 3 UNLOADING WITH STIFFNESS S1 FROM PREVIOUS MAXIMUM
C POINT (DM,FM) AFTER YIELDING AT (DY,FY).
3 IF ((X0DD)*SN) 320,230,230
300 LL=3
SS=S1(IS)
310 FF=FM(IS)+(DDDM(IS))*S1(IS)
GO TO 10000
320 IF ((DM(IS)DD)*SN) 200,200,310
C RULE 4 LOADING TOWARD PREVIOUS MAXIMUM POINT (DM,FM) ON
C PRIMARY CURVE FROM ZERO CROSSING POINT (X0,0.0).
4 IF ((DDDS)*SN) 430,430,420
400 LL=4
SS=FM(IS)/(DM(IS)X0)
410 FF=(DDX0)*SS
GO TO 10000
420 IF ((DM(IS)DD)*SN) 200,200,410
430 F0=FS
D0=DS
S2(IS)=S1(IS)*B1
IF (SS.LT.S2(IS)) SS=S2(IS)
X1=D0F0/SS
IF ((X1DD)*SN) 500,440,440
440 IS=3IS
SN=FLOAT(IS+IS3)
450 IF ((DM(IS)DD)*SN) 200,200,600
C RULE 5 UNLOADING FROM INNER PEAK POINT (D0,F0) WITH
C STIFFNESS S2.
5 IF ((X1DD)*SN) 520,440,440
500 LL=5
IF (SS.LT.S2(IS)) SS=S2(IS)
510 FF=F0+(DDD0)*SS
GO TO 10000
520 IF ((D0DD)*SN) 240,240,510
C RULE 6 LOADING TOWARD PREVIOUS MAXIMUM POINT (DM,FM) ON
C PRIMARY CURVE FROM ZERO CROSSING POINT (X1,0.0).
6 IF ((DDDS)*SN) 630,630,620
600 LL=6
SS=FM(IS)/(DM(IS)X1)
610 FF=(DDX1)*SS
GO TO 10000
620 IF ((DM(IS)DD)*SN) 200,200,610
630 F1=FS
D1=DS
29
S2(IS)=S1(IS)*B1
IF (SS.LT.S2(IS)) SS=S2(IS)
X2=D1F1/SS
IF ((X2DD)*SN) 700,640,640
640 IS=3IS
SN=FLOAT(IS+IS3)
650 IF ((D0DD)*SN) 240,240,800
C RULE 7 UNLOADING FROM INNER PEAK POINT (D1,F1) WITH
C STIFFNESS S2.
7 IF ((X2DD)*SN) 720,640,640
700 LL=7
IF (SS.LT.S2(IS)) SS=S2(IS)
710 FF=F1+(DDD1)*SS
GO TO 10000
720 IF ((D1DD)*SN) 450,450,710
C RULE 8 LOADING TOWARD INNER PEAK POINT (D0,F0) FROM ZERO
C CROSSING POINT (X2,0.0).
8 IF ((DDDS)*SN) 830,830,820
800 LL=8
SS=F0/(D0X2)
810 FF=(DDX2)*SS
GO TO 10000
820 IF ((D0DD)*SN) 240,240,810
830 F2=FS
D2=DS
IF (SS.LT.S2(IS)) SS=S2(IS)
X3=D2F2/SS
IF ((X3DD)*SN) 900,840,840
840 IS=3IS
SN=FLOAT(IS+IS3)
850 IF ((D1DD)*SN) 450,450,1000
C RULE 9 UNLOADING FROM INNER PEAK POINT (D2,F2) WITH
C STIFFNESS S2.
9 IF ((X3DD)*SN) 920,840,840
900 LL=9
IF (SS.LT.S2(IS)) SS=S2(IS)
910 FF=F2+(DDD2)*SS
GO TO 10000
920 IF ((D2DD)*SN) 650,650,910
C RULE 10 LOADING TOWARD INNER PEAK POINT (D1,F1) FROM ZERO
C CROSSING POINT (X3,0.0).
10 IF ((DDDS)*SN) 1030,1030,1020
1000 LL=10
SS=F1/(D1X3)
1010 FF=(DDX3)*SS
GO TO 10000
1020 IF ((D1DD)*SN) 450,450,1010
1030 F3=FS
D3=DS
IF (SS.LT.S2(IS)) SS=S2(IS)
X2=D3F3/SS
IF ((X2DD)*SN) 1100,640,640
C RULE 11 UNLOADING FROM INNER PEAK POINT (D3,F3) WITH
C STIFFNESS S2.
11 IF ((X2DD)*SN) 1120,640,640
1100 LL=11
IF (SS.LT.S2(IS)) SS=S2(IS)
1110 FF=F3+(DDD3)*SS
GO TO 10000
1120 IF ((D3DD)*SN) 850,850,1110
10000 RETURN
END
30
SUBROUTINE HYST5 (LL,SS,DD,DS,FF,FS)
C
C TAKEDA HYSTERESIS MODEL
C
C PROGRAMMED BY OTANI, S.
C ON FEBRUARY 9, 1979
C AT UNIVERSITY OF TORONTO
C
C
C INPUT DATA
C LL HYSTERESIS RULE POINTER AT PREVIOUS TIME STEP
C DS DISPLACEMENT AT PREVIOUS TIME STEP
C DD DISPLACEMENT AT PRESENT TIME STEP
C FS FORCE AT PREVIOUS TIME STEP
C SS STIFFNESS AT PREVIOUS TIME STEP
C FC CRACKING FORCE
C FY YIELDING FORCE
C DC CRACKING DISPLACEMENT
C DY YIELDING DISPLACEMENT
C SC DISPLACEMENT STIFFNESS BEFORE CRACKING
C SY DISPLACEMENT STIFFNESS BEFORE YIELDING
C SU DISPLACEMENT STIFFNESS AFTER YIELDING
C B0 STIFFNESS DEGRADATION FACTOR
C B1 STIFFNESS DEGRADATION FACTOR
C OUTPUT DATA
C LL HYSTERESIS RULE POINTER AT PRESENT TIME STEP
C SS STIFFNESS AT PRESENT TIME STEP
C FF FORCE AT PRESENT TIME STEP
C DESCRIPTION OF VARIABLES
C S1(*) UNLOADING STIFFNESS FROM A PEAK ON PRIMARY CURVE
C S2(*) UNLOADING STIFFNESS ON INNER HYSTERESIS LOOPS
C F0 UNLOADING FORCE LEVEL FROM STEP 6
C F1 UNLOADING FORCE LEVEL FROM STEP 8
C F2 UNLOADING FORCE LEVEL FROM STEP 10
C F3 UNLOADING FORCE LEVEL FROM STEP 12
C D0 UNLOADING DISPLACEMENT FROM STEP 6
C D1 UNLOADING DISPLACEMENT FROM STEP 8
C D2 UNLOADING DISPLACEMENT FROM STEP 10
C D3 UNLOADING DISPLACEMENT FROM STEP 12
C X0 ZEROCROSSING POINT FROM STEP 4 TO STEP 5
C ZEROCROSSING POINT FROM STEP 4 TO STEP 6
C X1 ZEROCROSSING POINT FROM STEP 7 TO STEP 8
C X2 ZEROCROSSING POINT FROM STEP 9 TO STEP 10
C ZEROCROSSING POINT FROM STEP 13 TO STEP 10
C X3 ZEROCROSSING POINT FROM STEP 11 TO STEP 12
C FM(*) UNLOADING FORCE LEVEL ON PRIMARY CURVE
C DM(*) UNLOADING DISPLACEMENT ON PRIMARY CURVE
C
C
COMMON /STFF/MD,DC,DY,FC,FY,SC,SY,SU,B0,B1
DIMENSION FM(2),DM(2),S1(2),S2(2)
C
IS=2
IF (FS.LT.0.0) IS=1
SN=FLOAT(IS+IS3)
GO TO (1,2,3,4,5,6,7,8,9,10,11,12,13), LL
C RULE 1 ELASTIC STAGE UP TO CRACKING
1 IF (DCABS(DD)) 110,110,100
100 FF=SC*DD
GO TO 10000
110 FM(1)=FC
FM(2)= FC
DM(1)=DC
DM(2)= DC
S1(1)= SC
S1(2)= SC
120 IF (DYABS(DD)) 300,300,200
C RULE 2 LOADING ON THE POSTCRACKING FRIMARY CURVE UP TO
31
C YIELDING.
2 IF ((DDDS)*SN) 230,230,220
200 LL=2
SS=SY
210 FF=(FC+(ABS(DD)DC)*SY)*SN
GO TO 10000
220 IF (DYABS(DD)) 300,300,210
230 S1(IS)=(ABS(FS)+FC)/(ABS(DS)+DC)
GO TO 330
C RULE 3 LOADING ON THE POSTYIELDING PRIMARY CURVE
3 IF ((DDDS)*SN) 320,320,310
300 LL=3
SS=SU
310 FF=(FY+(ABS(DD)DY)*SU)*SN
GO TO 10000
320 S1(IS)=(DY/ABS(DS))**B0*(FC+FY)/(DC+DY)
330 FM(IS)=FS
DM(IS)=DS
X0=DSFS/S1(IS)
IF ((DDX0)*SN) 430,430,400
C RULE 4 UNLOADING FROM A PEAK (DM,FM) ON THE PRIMARY CURVE.
4 IF ((DM(IS)DD)*SN) 120,120,420
400 LL=4
SS=S1(IS)
410 FF=FM(IS)+(DDDM(IS))*S1(IS)
GO TO 10000
420 IF ((DDX0)*SN) 430,430,410
430 IS=3IS
SN=FLOAT(IS+IS3)
IF (ABS(DM(IS)).GE.DY) GO TO 740
X=FM(IS)/(DM(IS)X0)
Y=FY*SN/(DY*SNX0)
IF (XY) 440,440,450
440 FM(IS)=FY*SN
DM(IS)=DY*SN
450 IF (FCABS(FM(IS))) 740,460,460
460 X=FM(IS)/(DM(IS)X0)
IF (XSS) 470,470,740
470 D0=X0+FC/SS*SN
F0=FC*SN
IF ((D0DD)*SN) 530,530,500
C RULE 5 LOAD REVERSED AT A ZEROCROSSING POINT (X0,0) WITHOUT
C PREVIOUS CRACKING IN NEW LOADING DIRECTION.
5 IF ((DDX0)*SN) 540,540,520
500 LL=5
510 FF=(DDX0)*SS
GO TO 10000
520 IF ((D0DD)*SN) 530,530,510
530 FM(IS)=FY*SN
DM(IS)=DY*SN
X0=SN*(DYFY*(DY*SND0)/(FY*SNF0))
GO TO 740
540 IS=3IS
SN=FLOAT(IS+IS3)
IF ((DM(IS)DD)*SN) 120,120,400
C RULE 6 LOAD REVERSED AT ZEROCROSSING POINT (X0,0), AND
C LOADING TOWARD A PEAK (DM,FM) ON THE PRIMARY CURVE.
6 IF ((DDDS)*SN) 630,630,620
600 LL=6
SS=FM(IS)/(DM(IS)X0)
610 FF=(DDX0)*SS
GO TO 10000
620 IF ((DM(IS)DD)*SN) 120,120,610
630 F0=FS
D0=DS
S2(IS)=S1(IS)*B1
IF (SS.LT.S2(IS)) SS=S2(IS)
X1=D0F0/SS
32
640 IF ((DDX1)*SN) 730,730,700
C RULE 7 UNLOADING FROM A PEAK (D0,F0) TOWARD A ZEROCROSSING
C POINT (X1,0).
7 IF ((D0DD)*SN) 740,740,720
700 LL=7
IF (SS.LT.S2(IS)) SS=S2(IS)
710 FF=F0+(DDD0)*SS
GO TO 10000
720 IF ((DDX1)*SN) 730,730,710
730 IS=3IS
SN=FLOAT(IS+IS3)
GO TO 940
740 IF ((DM(IS)DD)*SN) 120,120,600
C RULE 8 LOAD REVERSED AT ZERO CROSSING POINT (X1,0), AND
C LOADING TOWARD A PEAK (DM,FM) ON THE PRIMARY CURVE.
8 IF ((DDDS)*SN) 830,830,820
800 LL=8
SS=FM(IS)/(DM(IS)X1)
810 FF=(DDX1)*SS
GO TO 10000
820 IF ((DM(IS)DD)*SN) 120,120,810
830 F1=FS
D1=DS
S2(IS)=S1(IS)*B1
IF (SS.LT.S2(IS)) SS=S2(IS)
X2=D1F1/SS
IF ((DDX2)*SN) 930,930,900
C RULE 9 UNLOADING FROM A PEAK (D1,F1) TOWARD A ZEROCROSSING
C POINT (X2,0).
9 IF ((D1DD)*SN) 940,940,920
900 LL=9
IF (SS.LT.S2(IS)) SS=S2(IS)
910 FF=F1+(DDD1)*SS
GO TO 10000
920 IF ((DDX2)*SN) 930,930,910
930 IS=3IS
SN=FLOAT(IS+IS3)
GO TO 1140
940 IF ((DM(IS)DD)*SN) 120,120,800
C RULE 10 LOAD REVERSED AT A ZEROCROSSING POINT (X2,0), AND
C LOADING TOWARD A PEAK (D0,F0).
10 IF ((DDDS)*SN) 1030,1030,1020
1000 LL=10
SS=F0/(D0X2)
1010 FF=(DDX2)*SS
GO TO 10000
1020 IF ((D0DD)*SN) 740,740,1010
1030 F2=FS
D2=DS
IF (SS.LT.S2(IS)) SS=S2(IS)
X3=D2F2/SS
IF ((DDX3)*SN) 1130,1130,1100
C RULE 11 UNLOADING FROM A PEAK (D2,F2) TOWARD A ZEROCROSSING
C POINT (X3,0).
11 IF ((D2DD)*SN) 1140,1140,1120
1100 LL=11
IF (SS.LT.S2(IS)) SS=S2(IS)
1110 FF=F2+(DDD2)*SS
GO TO 10000
1120 IF ((DDX3)*SN) 1130,1130,1110
1130 IS=3IS
SN=FLOAT (IS+IS3)
GO TO 1330
1140 IF ((D0DD)*SN) 740,740,1000
C RULE 12 LOAD REVERSED AT A ZEROCROSSING POINT (X3,0), AND
C LOADING TOWARD A PEAK (D1,F1).
12 IF ((DDDS)*SN) 1230,1230,1220
1200 LL=12
33
SS=F1/(D1X3)
1210 FF=(DDX3)*SS
GO TO 10000
1220 IF ((D1DD)*SN) 940,940,1210
1230 F3=FS
D3=DS
IF (SS.LT.S2(IS)) SS=S2(IS)
X2=D3F3/SS
IF ((DDX2)*SN) 930,930,1300
C RULE 13 UNLOADING FROM A PEAK (D3,F3) TOWARD A ZEROCROSSING
C POINT (X2,0).
13 IF ((D3DD)*SN) 1330,1330,1320
1300 LL=13
IF (SS.LT.S2(IS)) SS=S2(IS)
1310 FF=F3+(DDD3)*SS
GO TO 10000
1320 IF ((DDX2)*SN) 930,930,1310
1330 IF ((D1DD)*SN) 940,940,1200
10000 RETURN
END
34
11.7 Pivot Model
Major features of the forcedeflection hysteresis results of largescale reinforced concrete
members are;
(1) Unloading stiffness decreases as displacement
ductility increases,
(2) Following a nonlinear excursion in one direction,
upon load reversal, the forcedeflection path crosses the
idealized initial stiffness line prior to reaching the idealized
yield force, and
(3) The effect of precracked stiffness may be ignored.
The use of the pivot point in defining degraded unloading
stiffness was first proposed by Kunnath et al. (1990).
Four quadrants are defined by the horizontal axis and
the elastic loading lines (positive and negative). Primary
Pivot points P
1
through P
4
on the elastic loading lines
control the amount of softening in each quadrant. Pinching
Pivot points PP
2
and PP
4
fix the degree of pinching
following load reversal in each quadrant.
(1) The response follows the strength envelope so
long as no displacement reversal occurs. The initial
elastic stiffness, yield resistance and maximum
resistance can be different in each direction.
(2) Once the yield deformation has been exceeded
in either direction, a subsequent strength envelope
is developed requiring the introduction of points S
1
and S
2
which move along the strength envelope and
defined by the previous maximum displacements.
The initial points of S
1
and S
2
are yield point Y
1
and
Y
2
in each direction.
(3) The modified strength envelope (acting as the
upper bound for future cyclic loading) is defined by
lines joining the pinching pivot point PP
4
(PP
2
) to
maximum response point S
1
(S
2
) until the response
point reaches the strength envelope.
(4) The pinching pivot points PP
4
and PP
2
are
initially fixed, but they move toward the
forcedeflection origin with the strength degradation. The
resistance at a pinching pivot point is given by
*
i yi
F β
where
) (
) (
*
*
ti iMAX
ti
iMAX
i i
ti iMAX i i
d d
F
F
d d
> =
≤ =
β β
β β
where
i
β defines the degree of pinching for a ductile
flexural response prior to strength degradation.
ti iMAX
d d , : maximum displacement and strength
degradation displacement (displacement at the highest
resistance) in the ith direction of loading (i=1 or 2).
(5) The unlading stiffness of the maximum displacement
Q
1
Q
2
Q
3
Q
4
4 2 2
( )
y
P F α
2 1 1
( )
y
P F α
Y
2
(D
y2
,F
y2
)
Y
1
(D
y1
,F
y1
)
* *
4 1 1 1 1
( , )
y y
PP D F β β
* *
2 2 2 2 2
( , )
y y
PP D F β β
Q
1
Q
2
Q
4
Q
3
D
F
Y
1
Y
2
P
1
P
4
P
3
P
2
d
y
d
t1
d
d1
d
f1
F
y1
F
y2
d
y2
2 2 y
F α
1 1 y
F α
d
t2
d
d2 d
f2
PP4
PP
2
F
t1
F
t2
S1(D1max,F1max)
S2(D2max,F2max)
Q
2
Q
3
Q
4
Q
1
S
1
d
1max
D
PP
4
PP
4
*
1 1 y
F α
1 1
(1 )
y
F η α +
P
1
P
1
*
F
P
4
P
4
*
35
excursion in Quadrant Q
1
is guided toward point P
1
until point P
1
* is reached at force ( η + 1 ) times
larger than the force at point P
1
. A line extending from point P
1
* through origin defines the new
softened elastic loading line K*. Point P
4
* is on the new elastic loading line at the same force level
as point P
4
. Point PP
4
* is also on the new elastic loading line but at a force defined by the
intersection of the modified strength envelope (line between points PP
4
and S
1
) and K*.
Hysteresis Rules:
(1) Loading and unloading in Quadrants Q
n
is directed away from or toward point P
n
, respectively.
(2) Loading in Quadrant Q
n
is directed toward point PP
n,
then to maximum response point Si, followed
by the strength envelope.
(3) Unloading in Quadrant Q
n
is directed away from point P
n
.
Reference:
Dowell, R. K., F. Seible and E. L. Wilson, “Pivot
Hysteresis Model for Reinforced Concrete
Members,” ACI Structural Journal, Title No.
95S55, Vol. 95, No. 5, SeptemberOctober
1998, pp. 607  617.
Kunnath, S. K., A. M. Reinhorn, and Y. J. Park,
“Analytical Modeling of Inelastic Seismic
Response of RC Structures,” Journal,
Structural Engineering Division, ASCE, Vol.
116, No. 4, April 1990, pp. 996  1017,”
Q
1
Q
2
Q
4
Q
3
D
F
Y
1
Y
2
P
1
P
4 P
3
P
2
PP
4
PP2
S
1
S
2
36
11.8 Stable Hysteresis Models with Pinching
The forcedeformation relation of a reinforced concrete member is highly dependent on a loading
history, characterized by strength decay with load reversals and pinching behavior at a low stress
level during reloading, when the behavior of the member is dominated by sliding along inclined shear
cracks or slippage of longitudinal reinforcement.
A flexuredominated reinforced concrete
girder sometimes exhibits a pinching
characteristic when the amount of
longitudinal reinforcement (or bending
resistance) is significantly different at the top
and bottom of section. This is attributable to
the fact that a wide crack in weak side
cannot close due to large residual strain in
tensile reinforcement after load reversal; the
compressive stress must be resisted by the
longitudinal reinforcement before concrete
faces make contact at cracks.
Many hysteresis models have been
developed on the basis of test results of a
particular set of specimens under a specific
loading history. However, the parameters of most models may not be analytically defined by the
member properties (material properties and member geometry).
Takedaslip Model: Eto and Takeda (1973) modified the Takeda model to incorporate a sliptype
behavior at low stress level due to pullout of longitudinal reinforcement from the anchorage zone.
The skeleton curve is trilinear with stiffness changes at cracking and yielding where the cracking
and yielding levels can be different in positive and negative directions. The performance of the
model is identical to the Takeda model before yielding.
Pinching takes place only when the
yielding has occurred in the direction of
reloading. The reloading (pinching)
stiffness K
s
is defined as
m m
s
m o y
F D
K
D D D
γ −
=
−
where
o
D : displacement at the end of
unloading (resistance equal to zero),
m
D and
m
F : maximum deformation and
resistance in the direction of reloading,
y
D : yield deformation in the direction of
reloading, γ : slip stiffness degradation
index (slip stiffness degradation index γ
is suggested to be 0.5). The pinching
stiffness is revised only when the
maximum response point is exceeded in
the direction of reloading.
When the response point crosses a line connecting the origin and the maximum response point in
the direction of reloading, the response point moved toward the previous maximum response point
and then on the skeleton curve. The unloading stiffness is defined in the same manner as the
D
F
Y
(D
m
,F
m
)
K
d
o
D
C
C
Y
K
s
K
s
’
(D
m
’,F
m
’)
'
o
D
Takedaslip model
Hysteresis Relation of Beams with
Unbalanced Amount of Reinforcement
37
Takeda model.
The same pinching and unloading stiffness is used during reloading and unloading in an inner
loop.
'
'
c y
m
d
c y y
F F
D
K
D D D
α −
+
=
+
where, '
c
F and '
c
D : resistance and deformation at cracking on the opposite side,
y
F and
y
D :
resistance and deformation at yielding on the unloading side,
m
D : maximum deformation on the
unloading side, α : unloading degradation index.
KabeyasawaShiohara Model: Kabeyasawa et al. (1983) modified the TakedaEto slip model to
represent the behavior of a girder with the amount of longitudinal reinforcement significantly different
at the top and bottom;
38
(1) the pinching occurs only in one direction where the yield resistance is higher than the other
direction,
(2) the pinching occurs only after the initial yielding in the direction of reloading, and
(3) the stiffness K
s
during slipping is a function of the maximum response point (D
m
, F
m
) and the
point of load reversal (D
o
, F
o
=0.0) in the resistancedeformation plane.
The reloading (slip) stiffness K
s
, after unloading in the direction of the smaller yield resistance,
was determined as
γ
o m
m
o m
m
s
D D
D
D D
F
K
− −
=
where (
m m
F D , ): deformation and resistance at the previous maximum response point,
o
D :
displacement at the end of unloading on the zeroload axis, γ : slip stiffness degradation index. No
slip behavior will be generated for γ = 0; the degree of slip behavior increases with γ > 1.0. γ =
1.2 was suggested.
The slip stiffness is used until the response point crosses a line with slope K
p
through the
previous maximum response point (D
m
, F
m
); the stiffness is reduced from the slope connecting the
origin and the maximum response point by reloading stiffness index η ,
) (
m
m
p
D
F
K η =
The values of unloading stiffness degradation index α of Takeda model, slipping stiffness
degradation index γ , and reloading stiffness index η were chosen to be 0.4, 1.0 and 1.0,
respectively by Kabeyasawa et al. (1983).
Costa and Costa model: Costa and Costa (1987) proposed a
trilinear model for the forcedisplacement response of a
singledegreeoffreedom oscillator, including pinching and
strength degradation.
Unloadingreloading loops prior to yielding in either direction
are bilinear, with slopes equal to those of the precracking and
postcracking branches in the virgin loading. After the initial
yielding, the reloading stiffness
s
K is reduced from the stiffness
toward the previous extreme point by factor ( / )
y m
D D
γ
; i.e.,
( )
y
m
s
m o m
D
F
K
D D D
γ
=
−
where,
m
F and
m
D : resistance and deformation at the
previous maximum response point, and
o
D : deformation at
load reversal point. Once the response point crosses the line
connecting the origin and the maximum response point, then
response point moves toward the maximum response point.
The unloading stiffness after yielding is reduced from the
elastic stiffness by factor ( / )
y r
D D
α
.
Postyield strength and stiffness degradation with cycling is
modeled by directing the reloading branch, after modification for
pinching, toward a point at a displacement equal to (1 )
m
D λ +
D
F
C
Y
K
s
O
Dc
Dy
F
c
F
y
F’c
F’
y
D’c D’y
39
and at a moment (1 )
m
F λ − , where
m
F is the resistance at the extreme point if the previous
excursion. After reaching this terminal point of the reloading branch, further loading takes place
parallel to the postyielding stiffness of the virgin loading curve.
References:
Costa, A. C., and A. G. Costa, “Hysteretic Model of ForceDisplacement Relationships for Seismic
Analysis of Structures,” National Laboratory for Civil Engineering, Lisbon, 1987.
Eto, H, and T. Takeda, "Elasto Plastic Earthquake Response Analysis of Reinforced Concrete
Frame Structure (in Japanese)," Report, Annual Meeting, Architectural Institute of Japan, 1973,
pp. 12611262.
Kabeyasawa, T., H. Shiohara, S. Otani and H. Aoyama, "Analysis of the Fullscale Sevenstory
Reinforced Concrete Test Structure," Journal of the Faculty of Engineering, the University of
Tokyo, (B), Vol. XXXVII, No. 2, 1983, pp. 431478.
F’c
F’
y
D’
c
D’
y
D
c
D
y
F
c
F
y
40
SUBROUTINE TAKEDSLP (FC,RC,FY,RY,SU,B0,B1,B2,
+ FM,RM,F0,R0,F1,R1,X0,X1,XM,S1,
+ FF,RR,DR,SS,IL)
C
C TAKEDASLIP HYSTERESIS MODEL
C THE PRIMARY CURVE IS TRILINEAR WITH STIFFNESS CHANGES AT
C CRACKING AND YIELDING. THE PINCHING BEHAVIOUR DUE TO THE
C DIFFERENT AMOUNT OF TOP AND BOTTOM LONGITUDINAL REINFORCEMENT
C WAS MODELED BY MODIFYING THE TAKEDA MODEL.
C
C REFERENCE:
C OTANI, S., T. KABEYASAWA, H. SHIOHARA, AND H. AOYAMA,
C "ANALYSIS OF THE FULLSCALE SEVENSTORY REINFORCED CONCRETE
C TEST STRUCTURE", EARTHQUAKE EFFECTS ON REINFORCED CONCRETE
C STRUCTURES, U.S.JAPAN RESEARCH, PUBLICATION SP84, ACI, 1985,
C PP. 203239.
C
C PROGRAMMED BY SHIOHARA, H.
C ON DECEMBER 15, 1982
C AT THE UNIVERSITY OF TOKYO
C MODIFIED BY OTANI, S.
C ON APRIL 21, 1994
C AT UNIVERSITY OF CANTERBURY
C
C THE ORIGINAL PROGRAM BY SHIOHARA WAS WRITTEN TO DEFINE THE FORCE
C AND DEFORMATION FOR A GIVEN FORCE INCREMENT.
C THE PROGRAM WAS MODIFIED BY OTANI TO DEFINE THE HYSTERESIS RELATION
C FOR A GIVEN DISPLACEMENT INCREMENT. MANY COMMENTS WERE ADDED TO
C DESCRIBE THE HYSTERESIS RULES IN DETAILS.
C
C TAKEDA HYSTERESIS MODEL WAS MODIFIED AS FOLLOWS :
C (1) THE YIELD RESISTANCES IN POSITIVE AND NEGATIVE DIRECTIONS
C ARE MADE DIFFERENT ALTHOUGH CRACKING FORCE AND DEFORMATION
C ARE KEPT THE SAME. POSTYIELDING STIFFNESS IS ALSO THE SAME
C IN POSITIVE AND NEGATIVE DIRECTIONS.
C (2) THE PINCHING TAKES PLACE ONLY IN THE DIRECTION OF THE
C HIGHER YIELD RESISTANCE, AND ONLY AFTER THE INITIAL YIELDING.
C (3) THE HYSTERESIS RULES WERE SIMPLIFIED FROM THE ORIGINAL
C TAKEDA MODEL.
C (4) STIFFNESS DURING UNLOADING IS MADE EQUAL FOR UNLOADING
C FROM MAXIMUM RESPONSE POINT AND UNLADING FROM INNER PEAK.
C
C INPUT DATA:
C IL HYSTERESIS RULE POINTER AT PREVIOUS STEP.
C INITIAL VALUE MUST BE 1.
C RR DISPLACEMENT AT PREVIOUS STEP.
C INITIAL VALUE MUST BE LESS THAN RC.
C FF FORCE AT PREVIOUS STEP.
C INITIAL VALUE MUST BE LESS THAN FC.
C SS STIFFNESS AT PREVIOUS STEP.
C INITIAL VALUE MUST BE SC.
C DR DISPLACEMENT INCREMENT.
C PROPERTIES OF HYSTERESIS SYSTEM (INPUT DATA):
C FC CRACKING FORCE.
C RC CRACKING DISPLACEMENT.
C FY(*) YIELDING FORCE (POSITIVE AND NEGATIVE).
C RY(*) YIELDING DISPLACEMENT (POSITIVE AND NEGATIVE).
C SU STIFFNESS AFTER YIELDING.
C B0 UNLOADING STIFFNESS DEGRADATION INDEX.
C STIFFNESS S1(*) DURING UNLOADING IS DEGRADED AFTER
C YIELDING BY FACTOR (RY(IS)/RM(IS))**B(1) FROM THE
C STIFFNESS CONNECTING PEAK POINT (DM(*),FM(*)) AND
C CRACKING POINT IN THE OTHER DIRECTION.
C B0=0.0 TO 0.5 IS COMMONLY USED FOR REINFORCED CONCRETE.
C B1 SLIP STIFFNESS DEGRADATION PARAMETER.
C SLIP STIFFNESS ET IS DEFINED BY
C ET=ABS(FM(IS)/(XRM(IS)))*ABS(RM(IS)/(XRM(IS)))**B1
C B2 RELOADING STIFFNESS PARAMETER.
41
C STIFFNESS EU AFTER SLIPPING IS DEFINED BY
C EU=B2*FM(IS)/RM(IS)
C
C OUTPUT DATA:
C IL HYSTERESIS RULE POINTER AT PRESENT STEP
C RR DISPLACEMENT AT PRESENT STEP
C FF FORCE AT PRESENT STEP
C SS STIFFNESS AT PRESENT STEP
C
C DESCRIPTION OF VARIABLES
C RZ DISPLACEMENT AT PREVIOUS STEPS (TEMPORARY)
C FZ FORCE AT PREVIOUS STEP (TEMPORARY)
C FM(*) FORCE AT UNLOADING POINT FROM PRIMARY CURVE
C RM(*) DISPLACEMENT AT UNLOADING POINT FROM PRIMARY CURVE
C F0 FORCE AT UNLOADING POINT TO RULE 5
C R0 DISPLACEMENT AT UNLOADING POINT TO RULE 5
C F1 FORCE AT UNLOADING POINT TO RULE 8
C R1 DISPLACEMENT AT UNLOADING POINT TO RULE 8
C XM DISPLACEMENT AT STIFFNESS HARDENING POINT AFTER SLIPPING
C TOWARD PREVIOUS MAXIMUM RESPONSE POINT (RM(*),FM(*)),
C (FROM RULE 6 TO RULE 7).
C X0 DISPLACEMENT AT ZEROCROSSING POINT AFTER UNLOADING FROM
C PEAK (RM(*),FM(*)) ON PRIMARY CURVE.
C X1 DISPLACEMENT AT ZEROCROSSING POINT AFTER UNLOADING FROM
C PEAK.
C S1(*) STIFFNESS DURING UNLOADING; THE STIFFNESS IS COMMON FOR
C UNLOADING FROM PEAK (RM(*),FM(*)) ON PRIMARY AND FROM
C PEAK OF INNER LOOP.
C EU STIFFNESS AFTER SLIPPING TOWARD PREVIOUS MAXIMUM POINT.
C EU=B2*FM(IS)/RM(IS)
C ET STIFFNESS DURING SLIPPING AFTER RELOADING IN THE DIRECTION
C OF HIGHER YIELD RESISTANCE.
C ET=ABS(FM(IS)/(XRM(IS)))*ABS(RM(IS)/(XRM(IS)))**B1
C IS SIGN POINTER FOR FORCE AT PREVIOUS STEP.
C (=1 FOR NEGATIVE FORCE, AND =2 FOR POISTIVE FORCE).
C SN SIGN OF FORCE AT PREVIOUS STEP (=1.0 OR 1.0).
C ERR SMALL VALUE TO CHECK THE DIFFERENCE IN STIFFNESS.
C
PARAMETER (ERR=0.0001)
C
DIMENSION RY(2),FY(2),S1(2),FM(2),RM(2)
C
C DR INCREMENTAL DISPLACEMENT
C RR DISPLACEMENT AT PREVIIOUS STEP, REPLACED BY
C DISPLACEMENT AT PRESENT STEP
C SS STIFFNESS AT PREVIOUS STEP
C FF FORCE AT PREVIOUS STEP, REPLACED BY FORCE AT PRESENT
C STEP
C RZ DISPLACEMENT AT PREVIOUS STEP (TEMPORARY USE)
C FZ FORCE AT PREVIOUS STEP (TEMPORARY USE)
C
RZ=RR
FZ=FF
RR=RR+DR
FF=FF+SS*DR
C IS: SIGN POINTER FOR FORCE FZ AT PREVIOUS STEP
C (=1 FOR POISTIVE FORCE, AND =2 FOR NEGATIVE FORCE)
C SN: SIGN OF FORCE FZ AT PREVIOUS STEP (=1.0 OR 1.0)
IS=1
IF (FZ.LT.0.0) IS=2
SN=FLOAT(3ISIS)
C IL: HYSTERESIS RULE POINTER AT PREVIOUS TIME STEP
GO TO (1, 2, 3, 4, 5, 6, 7, 8, 9), IL
C
C RULE 1: ELASTIC STAGE UP TO CRACKING
C
C TEST IF DISPLACEMENT RR EXCEEDS CRACKING DISPLACEMENT RC.
C
42
1 IF (RCABS(RR)) 110, 110, 2000
C THE FOLLOWING PARAMETERS ARE SET
C FM(*) FORCE AT PREVIOUS MAXIMUM RESPONSE ON PRIMARY CURVE
C RM(*) DISPLACEMENT AT PREVIOUS MAXIMUM RESPONSE ON PRIMARY
C CURVE
C S1(*) STIFFNESS DURING UNLOADING FROM (RM,FM) ON PRIMARY
C CURVE
C WHERE *=1 FOR POSITIVE AND *=2 FOR NEGATIVE DIRECTION.
110 FM(1)= FC
FM(2)=FC
RM(1)= RC
RM(2)=RC
S1(1)= FC/RC
S1(2)= FC/RC
C CHECK IF DISPLACEMENT RR EXCEEDS YIELDING DISPLACEMENT AFTER CRACKING
120 IF (ABS(RY(IS))ABS(RR)) 300, 300, 200
C
C RULE 2: LOADING ON POSTCRACKING PRIMARY CURVE TO YIELD POINT
C
C TEST IF LOADING CONTINUES OR UNLOADING TAKES PLACE
C DR*SN>0: LOADING CONTINUES
C DR*SN<0: LOADING DIRECTION CHANGES, AND UNLOADING TAKES PLACE
C
2 IF (DR*SN) 220, 220, 210
C RESPONSE POINT MOVES ON PRIMARY CURVE AFTER CRACKING TOWARD YIELD POINT
C STIFFNESS SS AND FORCE FF ARE REVISED WITH STIFFNESS CHANGE
200 IL=2
SS=(FY(IS)SN*FC)/(RY(IS)SN*RC)
FF=SN*FC+(RRRC*SN)*SS
GO TO 2000
C TEST IF YIELDING TAKES PLACE DURING LOADING ON POSTCRACKING PRIMARY CURVE
210 IF ((RY(IS)RR)*SN) 300, 300, 2000
C UNLOADING TAKES PLACE FROM PEAK (RM(*),FM(*)) ON PRIMARY CURVE
C BETWEEN CRACKING AND YIELDING POINTS TOWARD CRACKING POINT
C IN THE OTHER DIRECTION.
C FORCE (FZ) AND DISPLACEMENT (RZ) AT PREVIOUS STEP ARE ASSIGNED
C TO FORCE FM(*) AND DISPLACEMENT RM(*) AT MAXIMUM RESPONSE POINT.
C STIFFNESS S1(*) DURING UNLOADING IS DEFINED AS SLOPE CONNECTING
C PREVIOUS MAXIMUM RESPONSE POINT (FM(*),RM(*)) AND CRACKING POINT
C IN THE OTHER DIRECTION.
220 RM(IS)=RZ
FM(IS)=FZ
S1(IS)=(FM(IS)+SN*FC)/(RM(IS)+SN*RC)
GO TO 320
C
C RULE 3: LOADING ON THE POSTYIELDING PRIMARY CURVE
C
C TEST IF LOADING CONTINUES OR UNLOADING TAKES PLACE
C DR*SN>0: LOADING CONTINUES
C DR*SN<0: LOADING DIRECTION CHANGES, AND UNLOADING TAKES PLACE
C
3 IF (DR*SN) 310, 310, 2000
C RESPONSE POINT MOVES ON POSTYIEDING PRIMARY CURVE.
C FORCE FF AND STIFFNESS SS ARE REVISED FOR STIFFNESS CHANGE.
300 IL=3
FF=FY(IS)+(RRRY(IS))*SU
SS=SU
GO TO 2000
C UNLOADING TAKES PLACE FROM PEAK (RM(*),FM(*)) ON PRIMARY
C CURVE AFTER YIELDING USING UNLOADING STIFFNESS S1(*).
C FORCE (FZ) AND DISPLACEMENT (RZ) AT PREVIOUS STEP ARE ASSIGNED
C TO FORCE FM(*) AND DISPLACEMENT RM(*) AT MAXIMUM RESPONSE POINT.
C STIFFNESS S1(*) DURING UNLOADING IS DEGRADED BY FACTOR
C (RY(IS)/RM(IS))**B0 FROM THE STIFFNESS CONNECTING THE PREVIOUS
C MAXIMUM RESPONSE POINT (DM(*),FM(*)) AND THE CRACKING POINT IN
C THE OTHER DIRECTION.
310 RM(IS)=RZ
FM(IS)=FZ
43
S1(IS)=(SN*FC+FY(IS))/(SN*RC+RY(IS))*(RY(IS)/RM(IS))**B0
C X0: DISPLACEMENT AT ZERO CROSSING POINT DURING UNLOADING
C FROM PEAK (RM(*),FM(*)) ON PRIMARY CURVE.
320 X0=RM(IS)FM(IS)/S1(IS)
C TEST IF THE LOADING DIRECTION CHANGES AFTER UNLOADING.
C (RRX0)*SN>0 UNLOADING CONTINUES.
C (RRX0)*SN<0 UNLOADING FINISHES, AND RELOADING TAKES PLACE IN
C THE OPPOSITE DIRECTION.
IF ((RRX0)*SN) 420, 400, 400
C
C RULE 4: UNLOADING FROM PEAK (FM,RM) ON THE PRIMARY CURVE
C
C TEST IF PREVIOUS RESPPONSE DISPLACEMENT IS EXCEEDED DURING
C REOADING AFTER UNLOADING FROM PEAK (RM(*),FM(*)).
C
4 IF ((RM(IS)RR)*SN) 120, 120, 410
C UNLOADING FROM POINT (RM,FM) ON PRIMARY CURVE.
C FORCE AND STIFFNESS ARE REVISED WITH STIFFNESS CHANGE.
400 IL=4
FF=FM(IS)+(RRRM(IS))*S1(IS)
SS=S1(IS)
GO TO 2000
C TEST IF UNLOADING IS COMPLETED, AND RELOADING TAKES PLACE IN
C THE OTHER DIRECTION
C (RRX0)*SN>0 UNLOADING CONTINUES.
C (RRX0)*SN<0 UNLOADING IS COMPLETED AND RELOADING TAKES PLACE IN
C THE OPPOSITE DIRECTION.
410 IF ((RRX0)*SN) 420, 420, 2000
C UNLOADING IS COMPLETED AND RELOADING TAKES PLACE IN THE OPPOSITE
C DIRECTION PASSING ZERO CROSSING POINT (X0,0).
C SIGN SN OF FORCE FF IS CHANGED FROM FORCE FZ OF PREVIOUS STEP.
420 IS=3IS
SN=FLOAT(3ISIS)
C IF CRACKING HAS NOT OCCURRED IN THE RELOADING DIRECTION, RELAODING
C ASSUMES THE SAME STIFFNESS AS THE UNLOADING STIFFNESS UNTIL FORCE
C EXCEEDS THE CRACKING FORCE AT POINT (R0,F0).
IF (ABS(RM(IS)).LE.RC) THEN R0=X0+SN*FC/S1(3IS)
F0=FC*SN
IF ((R0RR)*SN) 520, 520, 500
C IF YIELDING HAS NOT TAKEN PLACE IN THE RELOADING DIRECTION, AND IF
C STIFFNESS TOWARD PREVIOUS MAXIMUM RESPONSE POINT (RM(*),FM(*))
C IS SMALLER THAN STIFFNESS TOWARD YIELDING POINT, PREVIOUS MAXIMUM
C RESPONSE POINT (RM(*),FM(*)) IS REPLACED BY THE YIELD POINT.
ELSE IF (ABS(RM(IS)).LE.ABS(RY(IS))) THEN
X=FM(IS)/(RM(IS)X0)
Y=FY(IS)/(RY(IS)X0)
IF (X.LT.Y) THEN RM(IS)=RY(IS)
FM(IS)=FY(IS)
END IF
GO TO 530
C TEST IF YIELD FORCE IN RELOADING DIRECTION IS LARGER THAN YIELD
C FORCE IN THE OPPOSITE DIRECTION
ELSE IF (ABS(FY(IS)).GT.ABS(FY(3IS))) THEN GO TO 430
ELSE GO TO 530
END IF
C AFTER COMPLETION OF UNLOADING FROM MAXIMUM RESPONSE POINT
C (RM(*),FM(*)) AT ZERO CROSSING POINT (X,0), SLIP TAKES
C PLACE DURING RELOADING IN OPPOSITE DIRECTION AT SLIP STIFFNESS ET.
430 X=RM(3IS)FM(3IS)/S1(3IS)
ET=FM(IS)/(RM(IS)SN*X)*(RM(IS)/(RM(IS)SN*X))**B1
EU=FM(IS)/RM(IS)*B2
C DISPLACEMENT X AT START OF STRAIN HARDENING AFTER SLIP IS
C SOLVED AS THE INTERSECTION OF LINE PASSING ZERO CROSSING POINT
C (X0,0) WITH SLOPE ET AND LINE PASSING PREVIOUS MAXIMUM POINT
C (RM(*),FM(*)) WITH SLOP EU.
IF (ABS(EUET).LT.ERR) THEN XM=X0
ELSE XM=(EU*RM(IS)ET*X0FM(IS))/(EUET)
END IF
44
C TEST IF RESPONSE PASSES STRAINHARDENING POINT
IF ((RRXM)*SN.GT.0.0) THEN
IF ((RRRM(IS))*SN) 700, 120, 120
ELSE GO TO 600
END IF
C
C RULE 5: LOAD REVERSED AT ZERO CROSSING PONT (X0,0) WITHOUT
C PREVIOUS CRACKING IN RELOADING DIRECTION.
C
5 IF ((RRX0)*SN) 540, 540, 510
C RELAODING TOWARD POINT (R0,F0) AFTER UNLODING FROM PEAK
C (RM(*),FM(*)) ON PRIMARY CURVE USING THE UNLOADING STIFFNESS
C S1(*). NO CRACKING HAS TAKEN PLACE IN THE RELOADING DIRECTION.
500 IL=5
SS=S1(3IS)
FF=(RRX0)*SS
GO TO 2000
C RELOADING TOWARD POINT (R0,F0).
C TEST IF RESPONSE POINT EXCEEDS POINT (R0,F0).
510 IF ((R0RR)*SN) 520, 520, 2000
C RELOADING BEYOND POINT (R0,F0) AND RELOADING TOWARD
C PREVIOUS MAXIMUM RESPONSE POINT (RM(*),FM(*)).
520 IF (ABS(RM(IS)).LE.RC
+ .AND.ABS(RM(3IS)).LE.ABS(RY(3IS))) GO TO 120
C IF YIELDED IN THE OPPOSITE DIRECTION, MAXIMUM RESPONSE POINT
C (RM(*),FM(*)) IS REPLACED BY THE YIELD POINT ALTHOUGH CRACKING
C HAS NOT TAKEN PLACE IN RELOADING DIRECTION.
FM(IS)=FY(IS)
RM(IS)=RY(IS)
X0=RY(IS)FY(IS)*(RY(IS)R0)/(FY(IS)F0)
530 IF ((RRRM(IS))*SN) 900, 120, 120
C DURING RELOADING, UNLOADING TAKES PLACE INTO OPPOSITE
C DIRECTION PASSING ZERO CROSSING POINT (X0,0), AND SIGN
C OF FORCE SN HAS BEEN CHANGED.
540 IS=3IS
SN=FLOAT(3ISIS)
C TEST IF RESPONSE POINT PASSES PREVIOUS MAXIMUM RESPONSE POINT
C (RM(*),FM(*)) DURING RELOADING TOWARD PREVIOUS MAXIMUM RESPONSE
C POINT (RM(*),FM(*))
IF ((RM(IS)RR)*SN) 120, 120, 400
C
C RULE 6: RELOADING WITH SLIP STIFFNESS ET TOWARD STRAIN HARDENING
C POINT AT DISPLACEMENT XM
C
C TEST IF RELOADING CONTINUES
C
6 IF (DR*SN) 630, 630, 610
C STIFFNESS SS AND FORCE FF ARE REVISED FOR STIFFNESS CHANGE
600 IL=6
X=RM(3IS)FM(3IS)/S1(3IS)
ET=FM(IS)/(RM(IS)SN*X)*(RM(IS)/(RM(IS)SN*X))**B1
SS=ET
FF=ET*(RRX0)
GO TO 2000
C RELOADING CONTINUES WITH SLIP STIFFNESS ET
C TEST IF RESPONSE POINT PASSES STRAIN HARDENING POINT AT
C DISPLACEMENT AT XM.
610 IF ((XMRR)*SN) 620, 620, 2000
C RESPONSE POINT PASSES THE STRAIN HARDENING POINT AT
C DISPLACEMENT XM. TEST IF RESPONSE PINT PASSES PREVIOUS
C MAXIMUM RESPONSE POINT (RM(*),FM(*)).
620 IF ((RRRM(IS))*SN) 700, 700, 300
C UNLOADING TAKES PLACE FROM THE PREVIOUS RESPONSE POINT
C AT PEAK (R1,F1) OF INNER LOOP.
630 R1=RZ
F1=FZ
C X1: DISPLACEMENT AT ZERO CROSSING POINT DURING UNLOADING
C FROM PEAK (R1,F1) OF INNER LOOP WITH UNLOADING STIFFNESS S1(*).
45
X1=R1F1/S1(IS)
GO TO 8
C
C RULE 7: LOADING WITH STRAIN HARDENING AFTER SLIPPING TOWARD
C PREVIOUS PEAK (RM(*),FM(*))
C
C TEST IF LOADING CONTINUES.
C
7 IF (DR*SN) 630, 630, 710
700 IL=7
SS=FM(IS)/RM(IS)*B2
FF=FM(IS)SS*(RM(IS)RR)
GO TO 2000
C LOADING CONTINUES TOWARD PREVIOUS MAXIMUM RESPONSE POINT
C (RM(*),FM(*)).
C TEST IF RESPONSE POINT EXCEEDS THE PREVIOUS MAXIMUM POINT
C (RM(*),FM(*)).
710 IF ((RRRM(IS))*SN) 2000, 300, 300
C
C RULE 8: UNLOADING FROM INNER PEAK POINT (R1,F1).
C
C TEST IF UNLOADING IS COMPLETED AND RELOADING STARTS IN THE
C OPPOSITE DIRECTION.
C
8 IF ((X1RR)*SN) 810, 840, 840
C UNLOADING FROM INNER PEAK (R1,F1) WITH STIFFNESS CHANGE.
800 IL=8
SS=S1(IS)
FF=F1(R1RR)*SS
GO TO 2000
C UNLOADING OR RELOADING IN THE SAME DIRECTION CONTINUES.
C TEST IF RELOADING RESPONSE POINT EXCEEDS THE INNER PEAK
C POINT (R1,F1).
810 IF ((R1RR)*SN) 820, 820, 2000
C RESPONSE POINT DURING RELOADING EXCEEDS THE INNER PEAK
C POINT (R1,F1), AND RESPONSE POINT MOVES TOWARD PREVIOUS
C MAXIMUM POINT (RM(*),FM(*)).
820 IF (ABS(RM(IS)).GT.ABS(RY(IS))
: .AND.ABS(FY(IS)).GT.ABS(FY(3IS))) THEN
C YIELDING HAS TAKEN PLACE IN BOTH DIRECTIONS.
IF ((RRXM)*SN.LE.0.0) GO TO 600
C TEST IF THE PREVIOUS MAXIMUM POINT (RM(*),FM(*)) IS EXCEEDED.
830 IF ((RRRM(IS))*SN) 900, 120, 120
C YIELDING HAS NOT TAKEN PLACE IN ONE DIRECTION.
ELSE GO TO 830
840 X0=X1
IS=3IS
SN=FLOAT(3ISIS)
IF (ABS(RM(IS)).GT.ABS(RY(IS))
+ .AND.ABS(FY(IS)).GT.ABS(FY(3IS))) THEN
EU=FM(IS)/RM(IS)*B2
XF=RM(IS)FM(IS)/EU
XM=X0
IF ((XFX0)*SN.LE.0.0) THEN
IF ((RRRM(IS))*SN) 900, 120, 120
ELSE
GO TO 430
END IF
ELSE
IF ((RRRM(IS))*SN) 900, 120, 120
END IF
C
C RULE 9: RELOADING TOWARD PEAK (RM(*),FM(*)) ON PRIMARY CURVE
C WITHOUT PINCHING
C
C TEST IF LOADING CONTINUES.
C
9 IF (DR*SN) 920, 920, 910
46
C RELOADING TOWARD PREVIOUS MAXIMUM POINT (RM(*),FM(*)) FROM
C ZERO CROSSING POINT (X0,0).
900 IL=9
SS=FM(IS)/(RM(IS)X0)
FF=SS*(RRX0)
GO TO 2000
C TEST IF RESPONSE POINT EXCEEDS PREVIOUS MAXIMUM RESPONSE
C POINT (RM(*),FM(*)) ON PRIMARY CURVE DURING RELOADING.
910 IF ((RRRM(IS))*SN) 2000, 2000, 120
C UNLOADING TAKES PLACE. PEAK (R1,F1) IS DEFINED AS RESPONSE
C POINT AT PREVIOUS STEP.
920 R1=RZ
F1=FZ
C X1 ZERO CROSSIG POINT DURING UNLOADING FROM INNER PEAK
C (R1,F1) USING COMMON UNLOADING STIFFNESS.
X1=R1F1/S1(IS)
C TEST IF UNLOADING IS COMPLETED.
IF ((X1RR)*SN) 800, 840, 840
C RETURN TO CALLING PROGRAM
2000 RETURN
END
47
11.9 Sheartype Hysteresis Models
Reinforced concrete members exhibit progressive loss of strength under reversed cycles of
inelastic deformation due to lack of shear capacity of member or bond resistance along longitudinal
reinforcement; the monotonic strength of such members cannot be attained.
The response of a reinforced concrete member, exhibiting early strength decay, is difficult to
model because such behavior is sensitive to loading history. General features can be summarized
as the decay in resistance with cyclic loading and pinching response during reloading followed by
hardening.
The undesirable features can be avoided or reduced by following design requirements and
detailing of reinforcement. Therefore, hysteresis models for shearfailing performance may not be
necessary for the response analysis of new construction, but may be necessary for the seismic
evaluation of existing construction.
TakayanagiSchnobrich Model: Takayanagi and Schnobrich (1976) modified the Takeda model to
incorporate pinching and strength decay features caused by high shear acting in short coupling
beams connecting parallel structural walls. The skeleton curve is trilinear.
The reloading (loading in the opposite direction after unloading) is made smaller than the stiffness
toward the previous maximum response point in the direction of reloading; the response point moves
toward the previous maximum response point after the response deformation changes its sign.
The resistance at a target point for reloading in the hardening range is reduced from the
resistance at the previous maximum response point; e.g., the resistance at the target point is
selected on a strength decay guideline which descends from the yield point. After the response
reaches the target point, the response point moves along a line parallel to the post yielding line.
The pinching stiffness is based on the reinforcement resistance for bending. The rate of strength
decay is assumed to proportionally increase with the rotation.
RoufaielMeyer Model: Roufaiel and Meyer (1987) used a hysteresis model that includes strength
decay, stiffness degradation and pinching effect.
Pinching
Decay Guideline
M
Mc
Mc
My
My
θ
D
m
D
m
’
Y’
Y
TakayanagiSchnobrich Model of
Pinching and Strength Decay
48
The moment resistance of a bilinear
momentcurvature relation was assumed to decay
when a given strain is reached at the extreme
compression fiber. The curvature at the
commencement of strength decay is called the
critical curvature. The degradation in resistance was
assumed to be proportional to the amount by which
the critical curvature was exceeded.
An auxiliary unloading branch AB is drawn
parallel to the elastic branch of the bilinear skeleton
curve until it intersects a line OB through the origin O
parallel to the strainhardening branch YA of the
skeleton curve. The line connecting this latter point B
of intersection to the point of previous extreme
deformation in the opposite direction defines the end
C of the unloading branch on the horizontal axis. If
yielding has not taken place in the direction of
loading, the yield point is used as the previous
maximum response point.
From that point on reloading is not always directed straight to the point of the previous extreme
postyield excursion in the direction of reloading, but it may include pinching, depending on the shear
ration, M/Vh. Pinching is accomplished by directing the reloading branch first towards a point on the
elastic branch of the skeleton curve at an ordinate equal to that of the intersection of this branch with
the line of straight reloading to the previous extreme deformation point, times m<1. The second part
of the reloading branch heads towards this latter extreme deformation point. Parameter m assumes
the following values;
for M/Vh<1.5 m=0
for 1.5<M/Vh<4 m=0.4(M/Vh)0.6
for 4<M/Vh m=1
The slope of slipping stiffness is
'
'
m
s
m o
F
K m
D D
=
−
Chung et al. (1987) extended the
Roufaiel and Meyer model to include
strength and stiffness degradation at
constant amplitude cycling. The
degradation model requires two
additional parameters: the value of
curvature
f
φ and the moment
f
m at
failure in monotonic loading. The failure
is defined as rupture or buckling of
longitudinal reinforcement, concrete
crushing, or the reduction of resistance
to 75 %. If the bilinear approximation to
the momentcurvature curve under
monotonic loading is denoted by ( )
p
m φ ,
a single halfcycle of loading up to
monotonic failure causes a drop in the moment at failure is ( )
p f f
m m φ − . By extension, during a
halfcycle to a curvature φ the drop in moment relative to the bilinear monotonic envelope ( )
p
m φ
is given by
Modification of Roufaiel and Meyer model
by Chung (1987)
Y
F A
B
C
D
O
s
K
o
D
(D
m
,F
m
)
Roufaiel and Meyer Model
49
3
2
( ) { ( ) }
y
p f f
f y
m half cycle at m m
φ φ
φ φ
φ φ
−
∆ = −
−
Accordingly, a branch of reloading in the direction where the previous maximum curvature is equal to
φ , moves toward a point at ( ( ) ,
p
m m φ φ − ∆ ), rather than at ( ( ),
p
m φ φ ) as in the original Roufaiel
and Meyer model.
BanonBiggsIrvine Model: Banon, Biggs and Max
Irvine (1981) modified Takeda hysteresis model by (a)
using a bilinear skeleton curve, (b) incorporating
pinching and stiffness degradation. The pinching
hysteresis was adopted to simulate the propagation of
inclined cracks due to high shear and slippage of
longitudinal reinforcement.
Hysteresis rules are summarized below;
(a) Momentrotation relationship is elastic up to the
yield point,
(b) Once the yield point is exceeded, loading
proceeds on the second slope of the bilinear envelope,
(c) Unloading is parallel to the elastic stiffness,
(d) The stiffness during reloading immediately after
unloading is reduced to 50 % of the second slope of the
bilinear envelope,
2
2
s
K
K =
(e) When the direction of loading changes during unloading and resistance (or deformation)
starts to increase again, the reloading stiffness is parallel to the elastic stiffness before the response
point reaches a point where the last unloading started,
(f) When the sign of deformation changes during reloading, the response point moves toward
previous maximum response point in the direction of reloading.
If the strengthdegrading feature is introduced, the response point after the pinching does not
move toward the previous maximum point, but a point on the skeleton curve at deformation greater
than the previous maximum deformation.
*
m
m
D
D
α
=
and 0.8 α = is suggested in the study.
The skeleton curve may be different in positive and negative directions.
Kato Shear Model: Kato et al. (1983) used a
hysteresis model to represent the behavior of a
reinforced concrete member failing in shear, in
which strength decay and stiffness reduction due to
load reversals were incorporated. A trilinear
skeleton curve was used with stiffness changes at A
and B. By choosing the skeleton curve without
descending stiffness, the stable flexural behavior
may be represented by this model. The trilinear
skeleton curve may include descending slopes. The
following example shows a skeleton curve with two
descending slopes.
F
D
Y
Y’
D
m
D’
m
s
K
1
K
2
K
BanonBiggsIrvine Model (1981)
D
F
A
B
Skeleton Curves of Kato Model (1983)
50
The response is linearly elastic before the response point reaches point A. The response point
follows the skeleton curve if the slope of the skeleton curve is positive; if the slope of the skeleton
curve is negative, the response point increases its deformation without the change in resistance
(plastic behavior).
If a response point crosses the descending branches during loading or reloading, the deformation
increases without change in resistance (perfectly plastic stiffness). Upon unloading from a maximum
response point on the perfectly plastic branch, the response point moves on a line parallel to the
initial elastic stiffness
e
K until the response point crosses the descending skeleton curve; the point
is termed as the maximum response point (D
max
, F
max
). Then the response point follows a line with
reduced stiffness
u
K ;
max
( )
u e
y
D
K K
D
α −
=
where α : unloading stiffness degradation index,
y
D : yield deformation.
Upon reloading after crossing zero resistance line, the response point moves on a line with
reloading (slip) stiffness
s
K ;
max min
min
( )
s
o y
D F
K
D D D
β −
=
−
where (
min min
, F D ): previous maximum response point on the skeleton curve in the direction of
reloading,
o
D : deformation at the completion of unloading,
y
D : yield deformation in the opposite
direction.
This slip stiffness is used for deformation l
s
(= γ l), where l: length from the unloading point to
the intersection of slip line and the line connecting the origin and the negative maximum response
point (
min min
, F D ). The response point during strain softening moves toward the previous maximum
point (
min min
, F D ) or the yield point if no yielding was experienced in the reloading direction.
If unloading takes place during reloading toward previous maximum response point, the
unloading stiffness from the previous maximum response point is used. If the response point crosses
s
l
l
min
F
min
D
max
D
xo
D
yp
D
e
K
s
K
u
K
51
the zero resistance axis,
the response point
follows the same slip
stiffness previously
defined in the reloading
direction. The length of
slip deformation is
defined for l: length from
the new unloading point
to the intersection of slip
line and the line
connecting the origin and
the maximum response
point (
max max
, D F ).
Values for the
parameters of this model
recommended for shear
failing reinforced
concrete members are
α =0.4, β =0.6 and
γ =0.95.
Values for flexure
dominated members are
α =0.2, β =γ = 0.0.
Park et al. model: The
model developed by Park
et al. (1987) includes (a)
stiffness degradation, (b)
pinching and (c) strength
degradation with cycling.
The skeleton curve is a
trilinear relation. The
extension of unloading from the postcracking branch of the virgin loading curve intersects the
precracking branch of the trilinear virgin loading in the direction of unloading at an ordinate equal to
approximately two times the corresponding yield moment. The reloading branch is initially directed
towards a point on the previous extreme unloading branch, at a moment ordinate equal to a
userspecified percentage γ (approximately 0.5) of the yield moment. Before reaching this point
and upon exceedance of the previous maximum permanent deformation (curvature at the
intersection of the previous extreme unloading branch and the horizontal axis), the reloading branch
stiffens and moves toward the point of maximum deformation in the direction of reloading. The
strength degrades in proportion to the amount of energy dissipated up to the current point. The
proportionality constant depends on the amount of longitudinal reinforcement and confining
reinforcement.
Origin Oriented Model: Shiga (1976) suggested
a simple hysteresis model, in which the response
moves on the line connecting the previous
absolute maximum response point and the origin.
If the response point reaches the maximum
response point, it moves on the skeleton curve.
When unloading takes place from a point on the
skeleton curve, the response point moves on the
line connecting the newly attained maximum
Hysteresis model by Park et al. (1987)
D
F
C
Y
C
Y
52
response point and the origin.
The model was obtained from the observation on steadystate response of reinforced concrete
structural model which oscillated about the origin of the forcedeformation relation. No hysteresis
energy is dissipated during the oscillation within the
previous maximum response amplitude. Therefore,
viscous damping proportional to the initial stiffness is
suggested as a mechanism to dissipate energy with
degradation of stiffness in a system.
Any shape may be used for the skeleton curve of
this model. This model is sometimes used to represent
a feature of sheardominated member, which
dissipates small hysteresis energy and degrades its
stiffness with plastic deformation. The model, however,
does not give residual displacement when the load
was removed. Therefore, the model may not be suited
for the simulation analysis of response waveform.
Similar to the origin oriented model, the response
point may directed toward the previous maximum
response point on the opposite direction. Such model may be called a peak oriented model.
Matsushima Strength Reduction Model: Short reinforced
concrete columns, failing in shear, exhibit strength decay
with load reversals and associated stiffness degradation.
Matsushima (1969) used a model to explain the damage of a
structure after shear failure in columns. The characteristics
of the model are basically of bilinear type, but the elastic
stiffness K
n
and the yield resistance F
n
were degraded
whenever unloading takes place from a point on the
postyielding line in a form;
y
n
n
y
n
n
F F
k K
β
α
=
=
where K
y
: initial elastic stiffness, F
y
: initial yield resistance, n:
number of unloading from the postyield stiffness line, α
and β are constants to decay rate.
References:
Banon, H., J. M. Biggs and H. Max Irvine, "Seismic Damage in Reinforced Concrete Frames,"
Journal of Structural Division, ASCE, Vol. 107, No. ST9, September 1981, pp. 17131729.
Chung, Y. S., et al., “Seismic Damage Assessment of Reinforced Concrete Members,” National
Center for Earthquake Engineering Research, State University of New York, Buffalo, Technical
Report NCEER870022, 1987.
Kato, D., S. Otani, H. Katsumata and H. Aoyama, "Effect of Wall Base Rotation Behavior of
Reinforced Concrete FrameWall Building," Proceedings, Third South Pacific Regional
Conference on Earthquake Engineering, Victoria University of Wellington, New Zealand, May
1983.
Matsushima, Y., "Discussion of Restoring Force Characteristics of Buildings, the Damage from
Tokachioki Earthquake (in Japanese)," Report, Annual Meeting, Architectural Institute of Japan,
August 1969, pp. 587588.
Park, Y. J., et al., “IDARC: Inelastic Damage Analysis of Reinforced Concrete FrameShear Wall
Structures,” National Center for Earthquake Engineering Research, State University of New
D
F
C
Y
C
Y
F
D
0
k
0
0
N
y y
N
F a F
k b k
=
=
0 y
F
k
53
York at Buffalo, Technical Report NCEER870008, 1987.
Roufaiel, M. S. L., and C. Meyer, "Analytical Modeling of Hysteretic Behavior of R/C Frames,"
Journal of Structural Division, ASCE, Vol. 113, No. 3, March 1987, pp. 429444.
Shiga, T., Vibration of Structures (in Japanese), Structural Series, Vol. 2, Kyoritsu Shuppan, 1976.
Takayanagi, T., and W. C. Schnobrich, "Computed Behavior of Reinforced concrete Coupled Shear
Walls," Structural Research Series No. 434, Civil Engineering Studies, University of Illinois at
UrbanaChampaign, 1976.
54
11.10 Axial ForceBending Moment Interaction
It is known that bending resistance varies with
existing axial force in a reinforced concrete section.
The effect of axial load on flexural yield level was
considered by Mahin and Bertero (1976), in which
the yield moment of the multicomponent model
was varied with the amount of axial load.
Takayanagi and Schnobrich (1976) modified
the Takeda model to include the effect of axial
forcebending resistance interaction in the analysis
of a coupled structural wall. The skeleton curve is
trilinear. A set of trilinear skeleton curves were
prepared for different level of axial force, and the
change in bending resistance with unit axial load
was evaluated. The moment m is assumed to
vary with curvature φ and axial force n , while
the axial force n is assumed to vary with
curvature φ and axial strain ε ;
( , )
( , )
m m n
n n
φ
φ ε
=
=
The assumption leads to an unsymmetric relation in an incremental form;
( )
m m m m n m n
m n
n n n
n n
n
φ φ ε
φ φ φ ε
φ ε
φ ε
∂ ∂ ∂ ∂ ∂ ∂ ∂
∆ = ∆ + ∆ = + ∆ + ∆
∂ ∂ ∂ ∂ ∂ ∂ ∂
∂ ∂
∆ = ∆ + ∆
∂ ∂
The above relation for incremental curvature φ ∆ and strain ε ∆ , and then modification factor was
developed to regain the symmetry;
1
( ) *
1
1
{ } *
1 ( / )( )
m
m EI
m n
n m
n
n EA
n m m m
n n
φ φ
φ
ε ε
ε
φ φ
∂
∆ = ∆ = ∆
∂ ∆
∂
−
∂ ∆
∂
∆ = ∆ = ∆
∂ ∂ ∆ ∂
∂
− −
∂ ∂ ∆ ∂
where * EI : instantaneous flexural rigidity, and * EA : instantaneous axial rigidity. The ratio
n
m
∆
∆
is assumed to remain constant during a small load increment.
The stiffness is updated for the subsequent load increment considering the existing axial force
level. For an increase in axial force, the momentrotation hysteresis relation is directed to the
corresponding loop with increased yield moment.
The axial forcemoment interaction effect can be easily handled by "fiber" model. Curvature may
be assumed to distribute uniformly over a specified hinge region, for which a momentrotation
relation can be evaluated on the basis of the momentcurvature relation at the critical section.
References:
Mahin, S. A., and V. V. Bertero, "Nonlinear Seismic Response of a Coupled Wall System," Journal
TakayanagiSchnobrich model for
axial loadmoment interaction
55
of Structural Division, ASCE, Vol. 102, 1976, pp. 17591780.
Takayanagi, T., and W. C. Schnobrich, "Computed Behavior of Reinforced Concrete Coupled Shear
Walls," Structural Research Series No. 434, Civil Engineering Studies, University of Illinois at
UrbanaChampaign, 1976.
56
11.11 Special Purpose Models
Hysteresis Model for MS Model: A steel spring and a
concrete spring in the corner of section are located in the
same point, and are subjected to identical displacement
history. Therefore, the two springs may be combined into
a single composite spring. The skeleton curve is
expressed by a bilinear relation; the compressive yield
resistance is determined as the sum of the compressive
strengths of the concrete and the steel springs, and the
tensile yield resistance is equal to the yield resistance
of the steel spring.
Hysteresis relation is of the Takeda model type with
the bilinear skeleton curve; unloading stiffness in a
compression zone and in a tension zone was made
different:
In a compression zone:
sy m
y
m
ce SE
sy m ce se
D D for
D
D
K K S
D D for K K S K
> + =
≤ + =
−λ
) (
) (
1
1 1
In a tensile zone:
sy m
sy
m
se
sy m se
D D for
D
D
K S
D D for K S K
− < =
− ≥ =
−
'
'
'
2
2 2
λ
where K
se
and K
ce
: initial elastic stiffness of the steel spring and the concrete spring, D
sy
: yield
deformation of the concrete and steel springs, D
m
: previous maximum response deformation in
compression, D
m
': previous maximum deformation in tension, S
1
= S
2
=2.0 and k = 0.4.
Post yielding stiffness was chosen to be 0.02 times the initial elastic stiffness of the direction of
loading. Upon reloading in compression, the response point moves on the slip stiffness line toward a
point (Dm, Fm"), where
m m
F F θ = " and θ = 0.4. When the sign of deformation changes, the
response point moves toward the previous maximum point in compression. Similar to the Takeda
model, the response point moves toward a peak of immediately outer loop.
Axial Forcedeformation Model: Kabeyasawa and Shiohara et al. (1983) used a hysteresis model
for an axial forcedeformation relation of a boundary column in the analysis of a structural wall. The
model was developed on the basis of the observed axial deformation behavior of the boundary
column in the test of the fullscale sevenstory structure tested as a part of U.S.Japan Cooperative
Program (Yoshimura and Kurose, 1985).
The tension stiffening was ignored; concrete was assumed to resist no tensile stress. The axial
stiffness in tension was made equal to the stiffness of the reinforcing steel in the boundary column,
and the stiffness in compression was assumed to be linearly elastic including the stiffness of the
concrete. The stiffness in tension changed at the tensile yielding of the longitudinal reinforcement.
The gravity loads was considered as the initial stress.
Composite spring hysteresis model
for MS model
Tension
Compression
57
A response point followed bilinear
hysteresis rules between the maximum
response point (
max max
, F D ) in the tension
side after yielding and a reference point Y'
(
y yc
F D − , ) on the skeleton curve in the
compression zone. The resistance
y
F − at
the reference point was determined at the
compressive yielding of the longitudinal
reinforcement.
The unloading stiffness K
r
was degraded
with plastic deformation;
α −
= ) (
max
yt
c r
D
D
K K
where,
yt
D : tensile yielding deformation,
max
D : maximum deformation greater than
y
D , α : unloading stiffness degradation
parameter (= 0.9).
When the response point reached the previous maximum point (
max max
, F D ) in tension, the
response point moved on the second slope of the skeleton curve, renewing the maximum response
point.
When the response point approached the compressive characteristic point Y' (D
yc
, F
y
) in
compression, the response point was directed to move toward a point Y" (2D
yc
, 2F
y
) from a point P
(D
p
, F
p
) on the bilinear relation:
) (
yc x yc p
D D D D − + = β
where, β : parameter for stiffness hardening point (=0.2), and
x
D : deformation at unloading
stiffness changing point. This rule was introduced to reduce an unbalanced force at the compressive
characteristic point Y' due to a large stiffness change. The compressive characteristic point Y' did
not change under any loading history.
This axialstiffness hysteresis model was used for the axial deformation of an independent
column as well as boundary columns of a wall.
Slip Model: Reinforced concrete
members exhibit sliptype (pinching)
behavior before a wide crack closes or
when longitudinal reinforcing bars slip
after bond deterioration. The sliptype
behavior is characterized by a small
stiffness during reloading at low
resistance level after a large amplitude
deformation in the opposite direction
and by the gradual increase in
stiffness with deformation.
Tanabashi and Kaneta (1962) used
a slip model with elastoplastic
skeleton curve and zero slip stiffness in their nonlinear response analysis. No hysteresis energy was
dissipated until the response point exceeded the previous maximum response point.
Axial forcedeformation model for
wall boundary element (Kabeyasawa et al., 1983)
Tension
Initial Load
Elongation
Compression
58
A finite stiffness may be assigned to the slip stiffness and a stress hardening may start to occur
before the initiation of slip at preceding unloading.
Bond Slip Model: Morita and Kaku
(1984) proposed a hysteresis model to
represent the bond stressbar slip relation
on the basis of their observation of the
test results. The model is prepared for
assuming various loading situations and
may be useful in a finite element analysis
of a reinforced concrete member.
References:
Fillipou, F. C., E. P. Popov and V. V. Bertero, “Effect of Bond Deterioration on Hysteretic Behavior of
Reinforced Concrete Joints,” Report No. EERC 8319, University of California, Berkeley, August
1983, 184 pp.
Fillipou, F. C., E. P. Popov and V. V. Bertero, “Modeling of Reinforced Concrete Joints under Cyclic
Excitations,” Journal, Structural Engineering, ASCE, Vol. 109, No. 11, November 1983, pp.
2666  2684.
Fillipou, F. C., “A Simplified Model for Reinforcing Bar Anchorages under Cyclic Excitations,” Report
No. EERC 8505, University of California, Berkeley, March 1985, 61 pp.
Kabeyasawa, T., H. Shiohara, S. Otani and H. Aoyama, "Analysis of the Fullscale Sevenstory
Reinforced Concrete Test Structure," Journal, Faculty of Engineering, University of Tokyo (B),
Vol. XXXVII, No. 2, 1983, pp. 432478.
Li, K.N., S. Otani and H. Aoyama, "Study on the Elasticplastic Behavior of Reinforced Concrete
Columns subjected to Bidirectional Horizontal Earthquake Forces and Varying Axial Load (in
Japanese)," Report, Aoyama Laboratory, Department of Architecture, Faculty of Engineering,
(a) Reloading relation (S > SA)
Bond Stress
Bond Stress
Slip Slip
Bond Stress
Slip
59
University of Tokyo, March 1990.
Morita, S., and T. Kaku, "Slippage of Reinforcement in Beamcolumn Joint of Reinforced Concrete
Frames," Proceedings, Eighth World Conference on Earthquake Engineering, San Francisco, U.
S. A., Vol. 6, 1984, pp. 477484.
Tanabashi, R., and K. Kaneta, "On the Relation between the Restoring Force Characteristics of
Structures and the Pattern of Earthquake Ground Motion," Proceedings, Japan National
Conference on Earthquake Engineering, November 1962, pp. 5762.
Yoshimura, M., and Y. Kurose, "Inelastic Behavior of the Building," ACI SP84, Earthquake Effects
on Reinforced Concrete Structures, U.S.Japan Research, American Concrete Institute, Detroit,
1985, pp. 163202.
60
11.12 Hysteresis Model for Prestressed Concrete Members
This note summarizes a hysteresis model for prestressed concrete members. The model was
proposed by M. Hayashi et al. (1995).
Member end moment M and rotation
θ of a prestressed concrete member under
antisymmetric bending moment distribution
with the inflection point at mid span is
considered. The skeleton curve of the
momentcurvature ( , M θ ) relation is
represented by a trilinear relation with
stiffness changes at flexural cracking of
concrete and tensile yielding of longitudinal
reinforcement.
Trilinear Skeleton Relation: The two points to define a trilinear skeleton curve may be estimated as
follows.
(1) Initial elastic stiffness
1
K is calculated for a prismatic line member considering flexural and
shear deformation:
1 2
3
c c c c
L
K
L
E I G A
κ
=
+
where L : member length from the face of the orthogonal member to the inflection point (0nehalf of
clear span or height),
c
E and
c
G : elastic and shear moduli of concrete,
e
I : moment of inertia of
the transformed concrete section,
c
A : cross sectional area of the transformed concrete section, κ :
shape factor for shear deformation (=1.2 for a rectangular section).
(2) Cracking moment
c
M and rotation
c
θ are calculated for a condition that the tensile stress at
the extreme tensile fiber reaches the tensile strength
t
σ of concrete:
1
( )
e
c t e
c
c
c
P
M Z
A
M
K
σ
θ
= +
=
where,
e
P : axial force acting on the section including effective prestressing force,
c
A : cross
sectional area of concrete,
e
Z : section modulus of the transformed section. Tensile strength
t
σ of
concrete may be assumed to be equal to 1.8
B
σ in kgf/cm
2
where compressive strength
B
σ of
concrete is expressed in kgf/cm
2
.
(3) Yielding moment
y
M should be calculated for a given axial force and effective prestressing
force assuming (a) plain section to remain plain after bending, (b) nonlinear axial stressstrain
relation of concrete and reinforcement, and (c) equilibrium of internal and external forces. A parabola
and straight descending line may be used to represent stressstrain relation of concrete in
compression; an elastoplastic stressstrain relation may be used for steel reinforcement ignoring
strain hardening.
Rotation
y
θ at yielding may be evaluated by integrating the curvature along the member, but
this often underestimates the deformation. Sugano (1970) proposed an empirical expression for the
A
m
B A
m m =
δ
A
θ
B A
θ θ =
61
ratio of secant stiffness at yielding to the initial elastic stiffness as follows;
2
1
{0.43 1.64 0.043 0.33 }( )
y
y
y t
B
M
a N d
n p
K D b D D
θ
α
σ
= = + + +
where n : modular ratio of steel to concrete,
t
p :
tensile reinforcement ratio including prestressing
reinforcement area as increased by the yield stress
ratio of prestressing reinforcement to ordinary
reinforcement, a : shear span, b and D: width and
depth of member section, N : axial force of section
including effective prestressing force, d : effective
depth of section,
B
σ : compressive strength of
concrete.
Characteristic Points on Hysteresis Relations:
The following points and stiffness are used in this
model.
(1) Characteristic point A ( ,
A A
M θ ) is defined on the
initial elastic stiffness line with stiffness
1
K . This
point is used for the hysteresis relation of a prestressed concrete member. Moment resistance
A
M
of the characteristic point is defined as decompression moment; i.e., for the effective prestressing
force P
e
,
e
A e
c
P
M Z
A
=
The moment is zero for a reinforced concrete member without prestressing force. Rotation
A
θ is
calculated for moment
A
M and initial stiffness
1
K ;
1
A
A
M
K
θ =
(2) Characteristic point B ( ,
B B
M θ ) is defined for Takeda hysteresis model (Takeda, Sozen and
Nielsen, 1970) as the terminal point (zero moment resistance) of unloading from the maximum
response point M ( ,
m m
M θ ). The unloading stiffness
B
K is defined as follows;
(a) unloading before yielding:
'
'
m c
B
m c
M M
K
θ θ
−
=
−
(b) unloading after yielding:
'
( )
'
y c
m
B
y c y
M M
K
γ
θ
θ θ θ
−
−
=
−
where
y
M and
y
θ : yield moment and rotation on the side of the unloading point, '
c
M and '
c
θ :
cracking moment and rotation on the opposite side, and γ : unloading stiffness degradation index of
the Takeda model (=0.5 for normal reinforced concrete members). The rotation
B
θ is calculated as
m
B m
B
M
K
θ θ = −
C
Y
y
θ
c
θ
c
M
y
M
Rotation
M
o
m
e
n
t
1
K
1 y
K α
A
A
θ
A
M
62
C
Y
Rotation
M
o
m
e
n
t
A
C’
'
c
M
'
c
θ
C
Y
y
θ
y
M
Rotation
M
o
m
e
n
t
A
C’
'
c
M
'
c
θ
M
m
M
m
θ
B
B
θ
B
K
B
B
K
m
θ
B
θ
M
(3) Unloading stiffness
A
K of fully prestressed concrete members from the maximum response
point M ( ,
m m
M θ ) on the skeleton curve is defined as follows;
C
Y
Rotation
M
o
m
e
n
t
A
A
θ
A
M
M
m
M
A
K
C
Y
y
θ
y
M
Rotation
M
o
m
e
n
t
A
A
θ
A
M
M
m
M
m
θ
A
K
m
θ
(a) unloading before yielding:
m A
A
m A
M M
K
θ θ
−
=
−
(b) unloading after yielding:
( )
y A
m
A
y A y
M M
K
γ
θ
θ θ θ
−
−
=
−
(4) Unloading stiffness
D
K of this model from the maximum response point M ( ,
m m
M θ ) on the
skeleton curve is defined by index ' λ taking into consideration the characteristics of both reinforced
concrete and fully prestressed concrete members.
' (1 ')
D A B
K K K λ λ = + −
(5) Characteristic point D ( ,
D D
M θ ) is defined as an intersection of line AB and the unloading line
MD of this model with unloading stiffness
D
K from the maximum response point M( ,
m m
M θ ) on the
63
skeleton curve.
Coordinates of the intersection D ( ,
D D
M θ ) must satisfy the two equations;
:
:
A D
A B D B
m D
D
m D
M M
line AB
M M
K line MD
θ θ θ θ
θ θ
=
− −
−
=
−
The moment resistance
D
M is thus solved from the two simultaneous equations as
m
m B
D
D A
A
A B
D
M
K
M M
M
K
θ θ
θ θ
− −
=
− −
(6) Unloading stiffness
E
K after reaching characteristic point D
(61) No yielding has taken place on the unloading side:
1 E
K K = (initial elastic stiffness)
(a1) The terminal point of this unloading stiffness is point E’ at moment level equal to moment
'
A
M of characteristic point A’ on the opposite side if no cracking has taken place on the
opposite side.
(a2) The terminal point of this unloading stiffness is point E’ at moment level equal to moment
'
D
M of characteristic point D’, which was defined during unloading after cracking on the
other side.
(62) Yielding has already occurred on the unloading side,
1 1
1
' ( )
'
( )
'
' ' (1 ')
m
y
y c
m
B
y c y
E B
K K
M M
K
K K K
γ
γ
θ
θ
θ
θ θ θ
λ λ
−
−
=
−
=
−
= + −
where rotation
m
θ is the maximum response rotation where the unloading initiated. The terminal
point of this unloading stiffness is point E’ at moment level equal to moment '
D
M of
characteristic point D’ in the opposite direction.
D
K
Rotation
Moment
Y
M
C
A
D
B
m
M
A
M
m
θ
B
θ
A
θ
Rotation
Moment
C
Y
M
D
K
A
D
B
m
M
A
M
A
θ
y
θ
m
θ
64
Characteristics of Prestressed Concrete:
(1) Unlading stiffness degradation index γ
Index γ was initially used in
the Takeda hysteresis model
(Takeda, Sozen and Nielsen,
1970) to control the unloading
stiffness from the maximum
deformation on the postyielding
skeleton curve and also to control
the hysteresis area per cycle. The
unloading stiffness was degraded
from the reference stiffness,
defined as a slope connecting the
yield point of unloading side and
the cracking point on the opposite
side.
'
( )
'
y c
m
B
y c y
M M
K
γ
θ
θ θ θ
−
−
=
−
The index value from 0.4 to 0.5 is normally used for reinforced concrete members. The hysteresis
energy dissipation decreases with increasing value of the index.
The unloading stiffness in momentrotation relations of reinforced concrete and prestressed
concrete member tests were examined. The following graph was suggested by Hayashi et al. (1995)
to express the unloading stiffness degradation index γ as a function of the ratio λ of ultimate
moment resistance attributable to prestressing reinforcement to the total ultimate moment. The ratio
λ is zero for a reinforced concrete member, unity for a fully prestressed concrete member, and
between zero and unity for a partially prestressed concrete member. The unloading stiffness
degradation index γ for partially prestressed to fully prestressed concrete is 0.7 to 0.8. The
hysteresis energy dissipation per cycle is less compared with a reinforced concrete member.
(2) Index ' λ for unloading stiffness
Index ' λ controls unloading
stiffness of reinforced concrete,
partially prestressed reinforced
concrete and fully prestressed
concrete members. The unloading
stiffness of the momentrotation
relations of member tests was
examined. Hayashi et al. (1995)
suggested expressing index ' λ
for unloading stiffness as a function
of the ratio λ of ultimate moment
resistance attributable to
prestressing reinforcement to the
total ultimate moment. The index
' λ increases with increasing
ultimate moment ratio λ .
The effect of prestressing on unloading stiffness is negligible when the ultimate moment ratio was
less than 0.3.
0.0 0.2 0.4 0.6 0.8 1.0
1.0
0.0
0.2
0.4
0.6
0.8
U
n
l
o
a
d
i
n
g
s
t
i
f
f
n
e
s
s
d
e
g
r
a
d
a
t
i
o
n
i
n
d
e
x
Ultimate moment ratio of prestressing reinforcement λ
I
n
d
e
x
f
o
r
u
n
l
o
a
d
i
n
g
s
t
i
f
f
n
e
s
s
0.0 0.2
0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
Ultimate moment ratio
λ
of prestressing reinforcement
65
Hysteresis Rules:
Rule 1: Before flexural cracking at C ( ,
c c
M θ ), the
relation is linearly elastic with stiffness
1
K .
If the response point reaches flexural cracking point C
( ,
c c
M θ ), the response point follows Rule 2.
Rule 2: The response point ( , M θ ) moves on the
second skeleton line CY toward yield point Y ( ,
y y
M θ )
with stiffness
2
K .
If the response point reaches the yielding point Y
( ,
y y
M θ ), the response point follows Rule 3.
If the unloading takes place from the maximum
response point M ( ,
m m
M θ ), the response point follows Subrule 21.
Subrule 21: The response point moves on line MD
with unloading stiffness
D
K , where M ( ,
m m
M θ ) is the
previous maximum point on the second skeleton line
CY. The unloading stiffness
D
K is defined as
'
'
' (1 ')
m A
A
m A
m c
B
m c
D A B
M M
K
M M
K
K K K
θ θ
θ θ
λ λ
−
=
−
−
=
−
= + −
Characteristic point D is defined as the intersection of
the unloading line MD and line AB connecting two
characteristic points A and B. The moment
D
M at
point D is given by
m
m B
D
D A
A
A B
D
M
K
M M
M
K
θ θ
θ θ
− −
=
− −
The response point moves on line MD during unloading and reloading.
If the response point reaches the previous maximum point M ( ,
m m
M θ ) during reloading, the
response point follows Rule 2.
If the response point reaches point D during unloading, the response point follows Subrule
22.
Subrule 22: The response point follows Subrule 221 if no cracking has taken place in the
opposite direction, or Subrule 222 if cracking has taken place in the opposite direction.
Subrule 221: The response point moves elastically from the characteristic point D
( ,
D D
M θ ) of Subrule 21 to point E’ whose moment level is equal to moment '
A
M of
C
Y
y
θ
c
θ
c
M
y
M
Rotation
M
o
m
e
n
t
1
K
2
K
Rule 1
Rule 2
Rule 3
Momentrotation relation during loading
C
Rotation
A
A
M
M
m
θ
B
D
K
Y
D
M
o
m
e
n
t
B
θ
m
M
A
θ
66
characteristic point A’ in the direction of reloading.
The slope
E
K of line DE’ is equal to the initial
stiffness
1
K . Between characteristic point D and
cracking point E’, the response point moves on line
DE’ during reloading and unloading.
The previous response point M’ on the side of
point E’ is defined as the cracking point C’. The
unloading stiffness '
D
K from the previous
maximum M’ (cracking point C’) is defined as
1
K .
Characteristic point D’ is defined as the
characteristic point A’, and moment '
D
M of the
characteristic point D’ is equal to moment '
A
M .
If the response point reaches point E’, the
response point follows Rule 4.
Subrule 222: The response point moves
elastically from the characteristic point D ( ,
D D
M θ )
of Subrule 21 to point E’ whose moment level is
equal to moment level '
D
M of characteristic point
D’ in the direction of reloading. The unloading
stiffness
E
K is equal to the initial elastic stiffness
1
K . Point D’ and its moment '
D
M have
been defined by Subrule 21 or
Subrule 31 upon previous unloading
from point M’ ( ' , '
m m
M θ ) on the second
skeleton line.
Between characteristic points D and
E’, the response point moves on the
same line DE’ during reloading and
unloading.
If the response point reaches
characteristic point D during reloading,
the response point moves toward
previous maximum response point M
( ,
m m
M θ ) in the direction of reloading
following Subrule 21.
If the response point reaches point E’
after crossing zero moment axis
(moment reversal), the response point
follows Subrule 4.
Rule 3: The response point follows the third
skeleton line with stiffness
3
K .
If the unloading takes place at maximum
response point M ( ,
m m
M θ ), the response point
follows Subrule 31.
1
K
A
B
C
D
E’
A’
C’
1 E
K K =
m
θ
m
M
No cracking in
reloading
direction
M
D
K
Subrule 21
Subrule 221
Rule 4
1
K
D
K
A
B
C
D
E’
A’
B’
C’
M’
D’
Cracked in
reloading
direction
D
M
m
M
m
θ
M
Subrule 21
Subrule 222
Rule 4
67
Subrule 31: The response point follows the unloading stiffness
D
K ,
( )
'
( )
'
' (1 ')
y A
m
A
y A y
y c
m
B
y c y
D A B
M M
K
M M
K
K K K
γ
γ
θ
θ θ θ
θ
θ θ θ
λ λ
−
−
−
=
−
−
=
−
= + −
where yielding point Y ( ,
y y
M θ ) is on the side of
maximum response point M ( ,
m m
M θ ) and
cracking point C’ ( ' , '
c c
M θ ) is on the other side.
Characteristic point D is defined as the
intersection of the unloading line MD and line AB
connecting characteristic points A and B.
Moment
D
M at characteristic point D is
determined by
m
m B
D
D A
A
A B
D
M
K
M M
M
K
θ θ
θ θ
− −
=
− −
The response is elastic between unloading point M and characteristic point D.
If the response point reaches the unloading point M, the response point follows Rule 3 for
loading on the third skeleton line.
If the response point reaches the
characteristic point D ( ,
D D
M θ ) of unloading,
the response point follows Subrule 32.
Subrule 32: The response point follows
Subrule 321 if no cracking has taken place in
the direction of reloading, Subrule 322 if
cracking has taken place in the direction of
reloading.
Subrule 321: The response point
moves elastically on line DE’ with
unloading stiffness
E
K , where unloading
stiffness is defined as
1 1
1
' ( )
'
( )
'
' ' (1 ')
m
y
y c
m
B
y c y
E B
K K
M M
K
K K K
γ
γ
θ
θ
θ
θ θ θ
λ λ
−
−
=
−
=
−
= + −
Point E’ is defined on the unloading line at
moment level equal to moment '
A
M of
characteristic point A’ on the initial
stiffness.
A
C
Y
M
B
'
A
M
Y’
C’
D
D
M
D
K
E’
E
K
A’
No cracking on
opposite side
Subrule 31
Subrule 321
Rule 4
C
Y
M
A
D
B
D
K
m
θ y
θ
y
M
u
M
Subrule 31
Rule 3
68
The previous response point M’ on the side of point E’ is defined as the yielding point Y’.
The unloading stiffness '
D
K from the yield point is defined as
' '
' '
'
'
' (1 ')
y A
A
y A
y c
B
y c
D A B
M M
K
M M
K
K K K
θ θ
θ θ
λ λ
−
=
−
−
=
−
= + −
Characteristic point D is defined as the intersection of the unloading line MD and line AB
connecting two characteristic points A and B. The moment
D
M at point D is given by
y
y B
D
D A
A
A B
D
M
K
M M
M
K
θ θ
θ θ
− −
=
− −
If the response point reaches point D, the response point follows Subrule 31.
If the response point reaches point E’, the response point follows Rule 4.
Subrule 322: The response
point moves elastically on line DE’
with unloading stiffness
E
K . The
unloading stiffness
E
K is defined
by
1 1
1
' ( )
'
( )
'
' ' (1 ')
m
y
y c
m
B
y c y
E B
K K
M M
K
K K K
γ
γ
θ
θ
θ
θ θ θ
λ λ
−
−
=
−
=
−
= + −
Point E’ is defined on the
unloading line at moment level
equal to moment level '
D
M of
the characteristic point D’. The
characteristic point D’ and its
moment '
D
M were defined
during previous unloading from
point M’ on the skeleton curve in
the opposite direction under
Subrule 21 or 31.
If the response point reaches point D, the response point follows Subrule 31.
If the response point reaches point E’, then the response point follows Rule 4.
Rule 4: The response point moves on line E’M’ toward the previous maximum response point M’
( ' , '
m m
M θ ) in the direction of reloading. The characteristic point E’ is defined either in Subrule 22
or 32.
A
B
C
D
Y
M
C’
A’
Y’
M’
B’
D’
E’
E
K
Cracking in reloading
direction
Subrule 31
Subrule 322
Rule 4
69
When the response point reaches the previous maximum response point M’ in the direction of
reloading, the response point follows either Rule 2 if no yielding has taken place or Rule 3 if yielding
has taken place in the direction of reloading.
If unloading takes place at point N’ before reaching the previous maximum response point, the
response point follows Subrule 41.
1 E
K K =
D
K
A
B
C
D
E’
A’
B’
C’
M’
D’
D
M
F’
G
N’
'
D
K
'
m
M
'
m
θ
A
B
C
D
Y
M
C’
A’
B’
D’
Y’
M’
G’
E
N
F
N’
F’
G
Subrule 41: The response point moves on line N’F’ with unloading stiffness '
D
K where
point N’ is an unloading point before the response point reaches the previous maximum
response point M’ in Subrule 22 or
32. Point F’ is defined on the
unloading line N’F’ at moment level
equal to moment '
D
M of
characteristic point D’. The
unloading stiffness '
D
K ,
characteristic point D’ and its
moment '
D
M were defined during
previous unloading from point M’
( ' , '
m m
M θ ) on the skeleton curve in
Subrule 21 or 31.
The response point follows line
N’F’ during reloading and
unloading.
If the response point reaches
point N’ during reloading, it follows
Rule 4.
If the response point reached
point F’ during unloading, the
response point follows Subrule 42.
A
C
Y
M
B
'
A
M
'
D
M
B’
Y’
C’
D
D
M
D
K
E’
E
K
N’
A’
'
D
K
D’
D
K
F’
G
70
Subrule 42: The response point follows line F’G with unloading (reloading) stiffness '
E
K
where characteristic point F’ is defined in Subrule 41. The moment level of point G is equal
to moment
D
M of characteristic point D. The characteristic point D’ and its moment '
D
M
were defined during previous unloading from point M’ ( ' , '
m m
M θ ) on the skeleton curve in
Subrule 21 or 31. The unloading stiffness '
E
K depends on the previous maximum
response on the unloading side;
(a) If no yielding has taken place on the unloading side, the unloading stiffness '
E
K is
equal to the initial elastic stiffness
1
K .
(b) If yielding has taken place on the unloading side, the unloading stiffness is given
below;
1 1
1
' ( )
'
'
( )
' '
' ' ' (1 ')
m
y
y c
m
B
y c y
E B
K K
M M
K
K K K
γ
γ
θ
θ
θ
θ θ θ
λ λ
−
−
=
−
=
−
= + −
The response point follows line F’G during unloading and reloading.
If the response point reaches point G, the response point follows Rule 4.
If the response point reaches F’ during reloading, the response point moves toward point
N’ following Subrule 41.
References:
Hayashi, M., S. Okamoto, S. Otani, H. Kato, and J. Fu, “Hysteresis Model for Prestressed Concrete
Members and its Effect on Earthquake Response (in Japanese),” Journal, Prestressed Concrete,
Japan Prestressed Concrete Engineering Association, Vol. 37, No. 4, July 1995, pp. 5767.
Sugano, S, "Experimental Study on Restoring Force Characteristics of Reinforced Concrete
Members (in Japanese)," Thesis submitted to fulfill the requirements of Doctor of Philosophy,
University of Tokyo, March 1970.
Takeda, T., M. A. Sozen and N. N. Nielsen, "Reinforced Concrete Response to Simulated
Earthquakes," Journal, Structural Division, ASCE, Vol. 96, No. ST12, 1970, pp. 25572573.
71
Home Assignment No. 6
20020311
Otani, S.
Displacement history is given below;
(1) D= 0.0
(2) D= 2.0 Dy
(3) D=2.0 Dy
(4) D= Dy
(5) D=0.5 Dy
(6) D= 4.0 Dy
(7) D= 2.0 Dy
(8) D= 4.0 Dy
where Fy=1.0, Dy=1.0. The skeleton curve is elastoplastic.
Draw resistancedeformation (FD) relation for the two hysteresis models;
(1) Clough Model (no unloading stiffness degradation)
(2) Pivot hysteresis model ( 2.0 α = , 0.5 β = )
1
Chapter 12 Response of Different Models
12.1 Effect of Member Modeling
Various member models have been proposed to represent the distribution of stiffness within a
reinforced concrete member. The effect of member models on the nonlinear response is studied by
Shiohara et al. (1983).
Member Models: Four member models are studied; (a) onecomponent model (Giberson, 1967), (b)
multicomponent model (Clough et al., 1965), (c) distributed flexibility model (Takizawa, 1976) and (d)
discrete spring model.
Onecomponent model consists of a linearly
elastic element and two rotational springs at the
ends of the elastic element. All inelastic
deformation is assumed to concentrate at the
member ends.
The multicomponent model assumes that a
member is divided to four imaginary parallel
elements; an element rigidly connected at both
ends, two elements with a hinge at one end and
rigidly connected at the other end and a truss
element with hinges at both ends. The last
element does not contribute to the resistance
and stiffness of the member. The dividing ratios are varied with a strain history to accommodate an
arbitrary member end hysteresis relation as proposed by Takizawa (1976).
The discrete spring model utilized 20
equallength rigid segments and 19 rotational
springs inbetween; the number of segments was
increased in a preliminary study, but the
response was not influenced by the number
greater than 20.
The distributed flexibility model assumes a
parabolic distribution of flexural flexibility
(reciprocal of flexural rigidity EI) along the member
with an elastic flexibility at the lowest flexibility
point. The flexural flexibility at member ends was
given by a hysteresis model. The model is useful
to represent spread of damage (cracks) along the
member, but not suitable to represent the
concentration of inelastic deformation at the
member end.
Stiffness Properties: Momentcurvature relation of section
was assumed to be bilinear with a postyielding stiffness of 2
percent of the initial elastic stiffness. The response of the discrete
element model was believed to give the most realistic response of
the four models if the behavior of the member was dominated by
flexure; however, the yielding was found to occur at slightly later
stage than the other models because rotational springs were not
placed at the member ends. The Clough hysteresis model
(Clough, 1966) was used for the momentcurvature relation with a
bilinear skeleton relation and a degrading unloading stiffness k
u
:
k
y
k
p
=0.02k
y
k
u
D
y
D
m
F
y
F
D
Rotational Spring
Rigid Element
One Component Model
m
A
m
B
A
θ
A
B
k
A
k
B
Elastic Element
B
θ
L
2
k k
p y
= 0 02 .
k k D D
u y m y
=
−
( / )
α
in which, k
p
: postyielding stiffness, k
u
: unloading stiffness, k
y
: initial elastic stiffness, D
m
:
maximum deformation beyond the yield displacement D
y
, α : unloading stiffness degradation
coefficient (= 0.4).
A member end momentrotation relation was calculated for the antisymmetric moment distribution
using the discrete element model. The calculated member end momentrotation relation was
idealized into a bilinear relationship connecting the yield point and a point corresponding to rotational
ductility factor (a ratio of a rotation divided by the yield rotation) of 9.0. The Clough hysteresis model
was also used for a member end momentrotation relation on the basis of the skeleton relation
calculated by the discrete spring model. The unloading degradation factor for a member end
momentrotation relation was estimated to be 0.43 for the discrete spring model under load reversal
at a rotational ductility factor of 9.0.
The stiffness properties of the onecomponent mode, multicomponent model and distributed
flexibility modes are determined on the basis of a hysteresis model provided for the member end
momentrotation relation under an imaginary antisymmetric moment distribution along the member.
Note that the actual member end rotation is not necessarily identical to the rotation given by the
hysteresis model, or the member stiffness identical to the stiffness given by the corresponding
hysteresis model.
Response under Uniform Bending: A hysteresis
model is derived for a member end momentrotation
relation under the antisymmetric moment distribution.
The use of such a hysteresis relation for a loading
situation drastically different from the antisymmetric
moment distribution is expected to clarify the
difference in the performance of the member models.
Therefore, the member models are subjected to a
stress history of uniform bending moment distribution along the member, although such a uniform
bending moment distribution is not expected to develop in a member during an earthquake.
A simply supported member was
subjected to external moments,
A
m and
B
m , of equal amplitude but of the opposite
sign at the two ends. The amplitude of the
moments was monotonically increased, and
the response of the member was calculated
for the four member models. Naturally, the
four models exhibited the identical elastic
stiffness; the yield point was also identical
for the three models except for the discrete
spring model which showed a slightly higher
yield resistance and deformation because
rotational springs were not placed at the
ends.
The difference among the four models was observed in the post yield stiffness. The post yield
stiffness of a member end momentrotation relation must be identical to the that of the
momentcurvature relation because a uniform bending moment was applied; the member end
rotation should be equal to φ ( / ) L 4 ; i.e. the post yielding stiffness should be 0.02 of the initial
stiffness.
A
B
m
A
m
B
A
θ
B
θ
Member under uniform bending
0 5 10 15
20
Stiffness reduction factor after yielding, %
One component model
Multicomponent model
Distributed flexibility model
Discrete element model
3
The one component model exhibited the highest postyield stiffness, more than 13 percent of the
initial stiffness because the model does not consider the distribution of curvature along the member.
The multicomponent model and the distributed flexibility model developed 5 percent of the initial
stiffness.
Response under Predetermined
Displacement History: A simply supported
member is subjected to member end rotations θ
A
and θ
B
of a predetermined history. A member
end rotation θ
B
was delayed from the other end
rotation θ
A
by a quarter of a cycle. Large
inelastic deformation is intended at the member
ends; the rotation amplitudes in the first and
second loading cycles are 6 and 12 times the yield
rotation.
A similar tendency is observed in the first cycle
of loading. The three models exhibit comparable
hysteresis shapes. The discrete spring model dissipates the largest hysteresis energy per cycle,
followed by the one component model, and the multicomponent model. The multicomponent model
dissipates less hysteresis energy at A end and more at B end.
One component model
Discrete spring model
Multicomponent model
(a) A end (b) B end
The momentrotation relation in the second cycle (load stages 5 to 8 in the loading history) after
yielding in the first cycle is compared for the four models. The hysteresis area is much thinner at A
end than at B end; the hysteresis loops were generally thin at A end. Lager hysteresis energy is
dissipated at Aend by the onecomponent model compared to the other two models. The post
yielding stiffness is highest for the onecomponent model as observed in the member end
momentrotation relation at B end. The response of the multicomponent model is closer to that of the
discrete element model.
Earthquake Response of Twostory Onebay Frame: A twostory onebay frame is analyzed under
El Centro (NS) 1940 earthquake motion, the acceleration amplitude is scaled to the maximum
acceleration of 3.90 m/sec
2
. The one component model and the multicomponent model are used in
the analysis.
A end
B end
D
u
c
t
i
l
i
t
y
f
a
c
t
o
r
4
Span width is 6.0 m; the firststory height is
3.75 m, and the secondstory height 3.00 m. Floor
weights at the two levels are the same and 50 tonf.
The base of the firststory columns is fixed to the
rigid ground. Young's modulus of the concrete is
2.37 x 10
5
kgf/cm
2
. No damping is assumed in the
analysis.
The beams have the following stiffness
properties; initial elastic moment of inertia Ib = 5.9
x 10
3
m
4
, cracking moment M
c
= 5.2 tonfm, M
y
+ =
21.7 tonfm for positive bending, and M
y
 = 8.5
tonfm for negative bending. The ratio β of
tangent stiffness after cracking to the initial
stiffness is 0.273 for positive bending and 0.080
for negative bending; post yield stiffness is
arbitrarily assumed to be 0.02 of the initial stiffness.
The Takedaslip model (Kabeyasawa, 1982) is
used for the hysteresis relation.
The columns have the following stiffness
properties both in positive and negative directions;
initial elastic moment of inertia I
c
= 5.0 x 10
3
m
4
,
cracking moment M
c
= 5.2 tonfm, and yield
moment M
y
= 19.3 tonfm. The ratio β of tangent
stiffness after cracking to the initial stiffness is
0.154; post yield stiffness is arbitrarily assumed to
be 0.02 of the initial stiffness. The Takeda model
(Takeda et al., 1970) is used for the hysteresis
relation.
The response waveform of the second floor
level (roof level) displacement, base shear,
member end rotations is compared for the two
models.
Although the secondfloor displacement
waveform is similar for the two model, the
multicomponent model exhibited a slightly larger
maximum response and also residual
displacement. The residual displacement must be
associated with the large plastic deformation at beam ends of the multicomponent model.
Member end deformation of the second floor beam, using the multicomponent model, shows the
drift corresponding to the deformation under uniform bending. The multicomponent model developed
a maximum deformation larger than the onecomponent model.
The deformation waveform at the ends of the first story column is similar, but the onecomponent
model calculates larger amplitudes Some residual deformation was calculated by the
multicomponent model at the top of the first story column.
Response of Fourstory and Sevenstory Buildings: In the nonlinear earthquake response
analysis of a fourstory fivebay frame and a sevenstory threebay frame, the response of three
member models was compared (Shiohara et al., 1983); i.e., (a) the onecomponent model, (b)
multicomponent model and (c) distributed flexibility model. No damping was assumed. The 1940 El
Centro (NS) motion was scaled to yield the maximum acceleration of 4.37 m/sec
2
. Displacement
5
response at the roof level and base shear response were compared.
The response waveforms were comparable for the three models in the analysis of the fourstory
frame. The multicomponent model and distributed flexibility model developed slightly larger residual
displacement. Multicomponent model developed slightly smaller base shear response. Beam end
rotation response of the three models was comparable, but the distributed flexibility model showed
accumulation of residual displacement with a number of oscillations.
The three models responded almost the same way up to 6 sec in the analysis of the sevenstory
frame, but the multispring model and the distributed flexibility model lengthened the period of
oscillation after a large amplitude response at 6 sec; the distributed flexibility model behaved quite
different manner than the other two models. The difference may be attributable to the contribution of
higher modes in the response; i.e., the inflection point of a column may locate near the midheight
under the oscillation in the fundamental mode. Therefore, the difference in the model characteristics
may not appear in the response of a lowrise structure. However, the shift of an inflection point due to
the higher mode oscillation tends to reveal the difference in model characteristics.
References:
Clough, R. W., K. L. Benuska and E. L. Wilson, "Inelastic Earthquake response of tall buildings,"
Proceedings, Third World Conference on Earthquake Engineering, New Zealand, Vol. II, Session
II, 1965, pp. 68  89.
Roof level displacement, cm
Base shear, x10
3
tonf
Response of sevenstory building
15
15
8.0
8.0
Ground motion acceleration, m/sec
2
Rooflevel Displacement, cm
Reponse of fourstory building
Base shear, x 10
3
tonf
4.0
4.0
15
15
2.0
2.0
Onecomponent model
Multicomponent model
Distributed flexibility model
6
Clough, R. W., and S. B. Johnston, "Effect of Stiffness Degradation on Earthquake Ductility
Requirements," Proceedings, Second Japan Conference of Earthquake Engineering, October
1966, pp. 227  232.
Giberson, M. F., "The Response of Nonlinear Multistory Structures subjected to Earthquake
Excitation," EERL Report, Earthquake Engineering Research Laboratory, California Institute of
Technology, Pasadena, 1967.
Kabeyasawa, T., "U.S.Japan Cooperative Research,  Study on the Earthquake Resistance of
Sevenstory Reinforced Concrete Structure (Part 3: PseudoDynamic Analysis) (in Japanese),"
Proceedings, Sixth Japan Conference of Earthquake Engineering, December 1982, pp. 1161 
1168.
Shiohara, H., S. Otani and H. Aoyama, "Comparison of Various Member Models for Reinforced
Concrete Earthquake Response Analysis," Transactions, Japan Concrete Institute, Vol. 5, 1983,
pp. 269  276.
Shiohara, H., S. Otani and H. Aoyama, "Comparison of Various Member Models for Reinforced
Concrete Earthquake Response Analysis," Transactions, Japan Concrete Institute, Vol. 5, 1983,
pp. 269  276.
Takeda, T., M. A. Sozen and N. N. Nielsen, "Reinforced Concrete Response to Simulated
Earthquakes," Journal, Structural Division, ASCE, Vol. 96, No. ST12. December 1970, pp. 2557 
2573.
Takizawa, H., "Notes on Some Basic Problems in Inelastic Analysis of Planar R/C Structures (Part
1)," Transactions, Architectural Institute of Japan, No. 240, February 1976, pp. 51  62.
7
12.2 Effect of Damping Modeling
The effect of viscous damping on the earthquake response of a singledegreeoffreedom system
is studied using an instantaneous damping coefficient * c proportional to constant mass m and
varying instantaneous stiffness
*
k :
* *
k c m c c
k m
+ =
An instantaneous damping factor
*
h may be defined as
m
k c
k
m c
mk
c
h
k m
*
*
*
*
*
2 2
2
+ =
=
Note that, with the degradation of stiffness, the instantaneous damping factor * h associated with
the mass increases and that associated with the instantaneous stiffness tends to decrease.
A hysteresis model includes the hysteresis energy dissipation, and it may not be reasonable to
expect additional energy dissipation by the damping during inelastic oscillation. A constant
massproportional damping tends to exaggerate the damping effect.
For a given earthquake motion, the degree of damping effect on the response may depend on (a)
type of damping, (b) period of vibration, (c) capacity of hysteresis energy dissipation, and (d) level of
ductility demand. The effect of damping is studied from these view points.
Type of Damping: Massproportional
damping is expected to be more effective
reducing the response amplitude where
many cycles of oscillation occurs with
highly degraded stiffness. On the contrary,
instantaneous stiffnessproportional
damping is effective during oscillation in a
small ductility range.
Attained ductility of the Takeda model
with an unloading stiffness degradation
parameter α of 0.0 is compared. The
"yielding period" of
singledegreeoffreedom systems was
varied from 0.14 sec to 1.13 sec. Taft
(N21E) record was used as an excitation
function.
The system with massproportional
damping produced small displacement
response than that with
stiffnessproportional damping having the
same initial damping factor. With an
increase in the value of initial damping
factor, the mass proportional damping is
more effective in reducing the response
amplitude. This tendency is larger for a shorter period system. The response amplitudes of systems
with stiffnessproportional damping are not so sensitive to the increase in the value of initial damping
factor partially attributable to the fact that the hysteretic energy dissipation is appreciable when the
unloading stiffness degradation parameter α was 0.0. When the initial damping factor is made of
equal contributions for the massproportional damping and the stiffnessproportional damping, the
massproportional damping tends to have a dominant influence on the maximum response.
Effect of types of damping
8
Displacement response waveforms are compared for two types of damping.
Singledegreeoffreedom systems with yielding period of 0.4 sec were subjected to the El Centro
(NS) 1940 earthquake motion. The Takeda hysteresis model was used with unloading stiffness
degradation parameters α of 0.0 and 0.5.
The waveforms are generally similar. The effect of damping is large when an unloading stiffness
degradation parameter α is 0.5; i.e., the hysteresis energy dissipation is small. The system with the
stiffnessproportional damping produces a larger response.
Period and Ductility Range: The effect of massproportional damping on maximum response is
pronounced, and the massproportional damping was found useful to exaggerate the damping effect
and to clarify a general trend of the damping effect of maximum response. Therefore, the
massproportional damping is used to study the variation of maximum response with the amount of
damping.
Singledegreeoffreedom (SDF)
systems with the Takeda hysteresis
model (unloading stiffness degrading
parameter α = 0.0) were subjected
to Taft (N21E) 1952 earthquake
motion. The yielding period was
selected to be 0.14 sec and 1.13 sec.
The yield resistance level was varied
to control the maximum ductility
response and to study the effect of
damping at different ductility ranges.
The initial damping factor was varied
from zero to 20 percent of the
critical.
The maximum response is
reduced significantly with increasing
damping amplitude in shortperiod
systems (T
y
= 0.14 sec), but is not so
much affected in longperiod systems
(T
y
= 1.13 sec). The general trend of
decreasing response amplitude with logarithmically increasing damping amplitude is observed for
both high and low ductility ranges.
Hysteretic Energy Dissipation Capacity: Some models have large hysteretic energy dissipation
Effect of damping with yield level and period
9
capacity, and others have small capacity. The latter model can dissipate kinetic energy only through
viscous damping, hence its response amplitude is likely affected by the amount of viscous damping.
The effect of damping on the response amplitude of singledegreeoffreedom systems with a
Takeda hysteresis model (unloading stiffness degrading parameter α = 0.0) and a Peakoriented
hysteresis model is studied under the Taft (N21E) 1952 earthquake motion. The yield period of the
systems was selected to be 0.14 sec and 1.13 sec.
The response point of the Peakoriented hysteresis model moves toward a previous maximum
response point in the loading direction, and the model behavior is linearly elastic between the positive
and negative maximum response points without any hysteretic energy dissipation. Once the
response point reaches the previous maximum response point, it moves on the primary curve.
The maximum response of the Peakoriented model at a yielding period of 0.14 sec was too large,
and the yield resistance was increased by 50 percent from the standard value to reduce the response
amplitude.
Note that the amount of damping has a larger influence on the response ductility of the
Peakoriented models, especially in the shortperiod systems. The difference in ductility of the two
models was relatively small in the longperiod systems.
It should be noted that the viscous damping dissipated energy even at a small amplitude
oscillation as long as there exists velocity. Hysteresis energy is dissipated at a large amplitude
oscillation beyond yielding. Therefore, the viscous damper is more effective in reducing response
amplitudes.
1
Chapter 13. Response of Different Hysteresis Models
The effect of different stiffness parameters on earthquake response of singledegreeoffreedom
systems is studied. By choosing stiffness properties and hysteretic energy dissipation capacity as
similar as possible, the effect of different hysteresis models on earthquake response waveforms and
amplitudes was investigated. Hysteresis models simulating flexural behavior of the reinforced
concrete were used.
13.1 Analysis Method
Common forcedeflection properties were specified to all the models so that the influence of
different model hysteretic behavior would be clarified. However, some models have a trilinear
skeleton curve under monotonically increasing load, and others have a bilinear skeleton curve.
Consequently, the yield point and the postyielding stiffness were chosen common among the
models, and the cracking point was added to the models with a trilinear skeleton curve. Hysteresis
models used are Degrading bilinear model, RambergOsgood model, Clough model, Bilinear Takeda
model, Takeda model, and Degrading trilinear model.
The mass of an SDF system was arbitrarily chosen to be 1.0 ton since the overturning effect
( ∆ − P effect) due to the mass' side sway was not included in the analysis. A series of hysteretic
models were designed with "yield period" (periods related to the secant stiffness at the yield point) of
0.10, 0.14, 0.20, 0.28, 0.40, 0.57, 0.80, 1.13 and 1.60 sec.
Earthquake response amplitudes are known to
vary with the system's period and the yield level.
To make the comparative study easy, SDF
systems with different periods were desired to
produce comparable ductility ratios (attained
maximum displacement divided by the yield
value). Hence, the Newmark's design criteria
(Veletsos and Newmark, 1960) were adopted.
Namely, the yield resistance of an SDF system
was determined by:
(a) dividing the maximum elastic inertia force
by the allowable ductility factor, µ , of the system
if the system's period is greater than 0.5 sec; and
(b) dividing the maximum elastic inertia force
by a factor 1 2 − µ if the system's period less
than 0.5 sec.
The stiffness of an elastic system was made
equal to the "yielding stiffness," K
y
(the secant
stiffness at the yield point). The allowable ductility
factor of a nonlinear system was arbitrarily
assumed to be 4.0. In this manner, the yield resistance was different for different periods and
different earthquake motions.
The postyielding stiffness, K
u
, was assumed to be 10 percent of the yielding stiffness; the
uncracked stiffness, K
c
, to be 2.0 times the yield stiffness; and the cracking resistance to be
onethird the yield resistance. These assumptions were used to approximate the stiffness properties
of a reinforced concrete structure.
The parameters of the RambergOsgood model were chosen so that the resistance at the
allowable ductility should be the same as the other models; i.e., η =1.0 and γ =3.79.
Hysteresis energy dissipation indices of different models are calculated at a ductility factor of 4.0
Determination of stiffness properties
2
and compared below:
Clough Model (α = 0.5): E
h
= 0.11
Degrading Trilinear Model: E
h
= 0.11
Takeda Model (α = 0.5): E
h
= 0.14
Degrading Bilinear Model (α = 0.5): E
h
= 0.19
Clough Model (α = 0.0): E
h
= 0.21
Takeda Model (α = 0.0): E
h
= 0.23
RambergOsgood Model(η =1.0; γ = 3.79): E
h
= 0.28
Bilinear Model (α = 0.0): E
h
= 0.33
where, α : unloading stiffness degradation index. Note that a large discrepancy exists among the
models in the capacity to dissipate hysteretic energy under a steadystate condition. the unloading
stiffness degradation prameter α has an appreciable effect on the value of hysteretic energy
dissipation index.
Damping: Viscous damping was assumed; the damping coefficient was assumed to be proportional
to varying instantaneous stiffness. The damping factor was 5 percent of the critical at the initial
elastic stage.
The damping is assumed to be proportional to instantaneous stiffness, where the maximum
response of an SDF system is not so sensitive to the amplitude of initial damping factor. Therefore,
an initial elastic damping factor of 0.05 is used for a system with bilinear primary curve, and 0.0707
for a system with trilinear primary curve. In this manner, the damping factor of all the system is
made identical at the yielding period because the precracking stiffness of a trilinear primary curve
is chosen twice the yielding stiffness.
Earthquake Motions: Four
earthquake accelerograms from two
California earthquakes were used in
this study: the NS and EW
components of the 1940 El Centro
record and the N21E and S69E
components of the 1952 Taft record,
digitized at the University of Illinois
at UrbanaChampaign (Amin and
Ang, 1966).
Linearly elastic response spectra
of these four records were studied
using the entire duration and the
first 15 sec part of the records. The
damped spectra were almost
identical for a period range less
than 2.0 sec using either the entire
duration or the first 15 sec part
except for the El Centro (EW)
record. Consequently, the response
computation was terminated
approximately at 15 sec when the
Taft (N21E and S69E) and El
Centro (NS) records were used. On
the other hand, the maximum
response of some linearly elastic
systems under the El Centro (EW)
motion occurred after 15 sec. Therefore, the entire 30 sec record of the El Centro (EW) motion was
Earthquake accelerograms
3
used.
The response spectra of the four motions are shown below:
Numerical Method: The equation of motion was solved numerically using the Newmark β method
(Newmark, 1959) with β =1/6 and γ =1/2. Both the equation of motion and the displacement 
velocity  acceleration relations were satisfied only at the discrete time step using an iterative
procedure. In other words, the "overshooting" of the hysteresis curve was adjusted within the time
step.
A constant time increment of the numerical integration was taken either as onetwentieth the
initial elastic period or 0.02 sec, whichever was shorter. The former was necessary to faithfully trace
the hysteresis curve rather than numerical stability requirements. The latter criterion became
necessary because the earthquake accelerograms were given at a 0.02 sec interval.
4
13.2 Effect of Initial Stiffness (Takeda Model)
The initial stiffness was arbitrarily chosen in this
paper to be 2.0 times the yielding stiffness. For a
normal reinforced concrete member, the ratio of initial
to the yielding stiffness may vary from 1.5 to 4.0.
F
y
F
c
k
y
D
F
Maximum response of the Takeda models is compared by varying the stiffness ratio, keeping a
crackingtoyielding resistance ratio to be onethird and secant yield stiffness the same. When an
attained ductility is greater than 4.0, the effect of initial stiffness is minimal for both shortperiod and
intermediateperiod systems. It is expected, however, that the initial stiffness should influence
maximum response amplitude if an attained ductility is less than or around unity.
13.3 Effect of Cracking Force Level (Takeda Model)
The effect of cracking force levels on
maximum response amplitudes of the Takeda
model is studied, keeping the initial stiffness to
be 2.0 times the yielding stiffness. Little effect is
observed when an attained ductility is greater
than 4.0 even for a shortperiod system.
F
y
F
c
k
y
D
F
13.4 Effect of Yield Resistance Level (Takeda Model)
The level of yield resistance is expected to be one major factor to influence maximum response
amplitudes. The yield resistance was varied from the standard value. As yield strength increases, an
attained ductility factor is significantly reduced, especially for a system with a short yielding period;
the required ductility was reduced to one half due to a 30 percent increase in the yield strength.
5
It is important to note that the value of
yield displacement increased proportional to
the level of yield resistance. Consequently,
the maximum response amplitude did not
decrease with the level of yield resistance so
much as the attained ductility did, although
the 0.14 sec period system showed a rapid
increase in the displacement amplitude with
the reduction in yield strength. The 1.13 sec
system also showed an increase in the
maximum displacement with decreasing
yield strength, but reached a peak at the
standard resistance, and then gradually
decreased its maximum displacement
amplitude.
The level of yield resistance has a
significant effect on maximum response
amplitude, especially in a shortperiod
range.
Response waveforms of Takeda models
( α =0.0) with different yield resistances are
compared. The yield resistances are 0.70, 0.60,
0.47 times the standard yield strength at a 0.4
sec yield period under El Centro (NS) 1940
motion. The maximum amplitude is largest for
the weakest system. However, the peak
amplitudes in the positive direction were largest
for the strongest system since the lower
strength system produced a large residual
displacement in the negative direction. This
observation is not necessarily true for a general
weakstrength system, but it is related to the
unloading stiffness degradation parameter. A low value of the parameter, for example α =0.0, tends
to cause a large residual displacement with little elastic recovery during unloading.
F
y
F
c
D
F
D
y
D
m
Maximum displacement and ductility factor
6
13.5 Effect of PostYielding Stiffness (Takeda Model)
The strain hardening of reinforcing
bars will give a finite positive stiffness
after the flexural yielding. Very small
postyielding stiffness has been routinely
used in Japan. The standard model in
this paper assumes a 10 percent of the
yielding stiffness as the postyielding
stiffness.
Maximum response of Takeda models
is compared varying postyield stiffness.
Maximum response decreases with an
increasing postyielding stiffness,
remarkably in a shortperiod system, and
insignificantly in a longperiod system.
The response amplitude changes more
with postyielding stiffness when the
postyielding stiffness is 0.05 to 0.20
times the yield stiffness. For a high
postyielding stiffness, less inelastic
displacement is required to store a given
magnitude of strain energy.
Effect of postyielding stiffness
7
13.6 Effect of Unloading Stiffness Degradation Parameters (Takeda Model)
Some models use an unloading stiffness degradation parameter, which controls the fatness of a
hysteresis loop and also the plastic residual deformation. It is not possible to determine the value of
this parameter from the material and geometrical properties of a reinforced concrete structure.
Normal range of this parameter is 0.0 to 0.5, and a value of 0.4 has been often used for the
reinforced concrete.
The effect of the value of the
unloading stiffness degradation
parameter on maximum response
of Takeda models is studied.
Maximum response increases with
an increasing value of the
parameter, and this tendency is
remarkable for shorter period
systems. The system's capacity
(either through damping or through
hysteresis) to dissipate kinetic
energy has a conspicuous
influence on the maximum
response of a short period
structure. The same tendency is
observed when the yielding period
of systems was varied from 0.1 to
1.6 sec; the effect becomes small
for a system of yielding period
greater than 0.4 sec.
Response waveforms of Takeda
models under El Centro (NS) 1940
motion are compared for the
yielding period of 0.4 sec. The yield level was chosen to be 0.6 times that of the standard model. For
8
a large value of the parameter, peak amplitudes are larger both in the positive and negative
directions, having comparable amplitudes in the two directions. For smaller values of the parameter,
the system tends to produce large amplitude only in one direction. This is clearly observed in the
hysteretic curve. Peaktopeak stiffness in a low amplitude oscillation is lower for a system with a
larger parameter, causing a long period of oscillation from approximately 6.0 sec.
The effect of the unloading stiffness degradation parameter is significant on response amplitude,
response waveform, residual displacement and hysteresis shape.
9
13.7 Effect of Hysteresis Energy Dissipation
The hysteretic energy dissipation capacity of a Degrading
Trilinear model is known to be sensitive to the choice of a
cracking point relative to the yielding point. The effect of
cracking force level on maximum response amplitude of
Degrading Trilinear models is studied. An attained ductility
factor decreases with an increasing cracking force level,
especially in a shortperiod system. This is another example to
show that maximum response amplitude of a shortperiod
system is significantly influenced by the capacity to dissipate
kinetic energy through either damping or hysteresis.
The response of two degrading trilinear models is shown above. The yielding period is 0.4 sec. The
two systems were subjected to the 1940 El Centro (NS) motion. A significant difference in hysteresis
shape can be observed. The response up to the first large oscillation is similar, but difference started
to be apparent in subsequent response cycles in the response waveform.
Effect of energy dissipation
10
13.8 Effect of Parameter of RambergOsgood Model
The parameter γ of the RambergOsgood model influences the hysteresis shape. When the
parameter is small, the hysteresis area becomes small, but
postyielding stiffness is high. When the parameter increases,
the hysteresis shape becomes similar to that of the
elastoplastic model.
The response of short and longperiod systems is compared
with the parameter γ . The shortperiod system increased the
ductility demand for increasing value of the parameter.
The response waveforms and hysteresis relations of systems
are compared using different parameters γ . The yielding
period of the systems was 0.4 sec, and the 1940 El Centro (NS)
motion was used. The response waveforms are similar up to 2.0
sec from the beginning of the motion. The system with using the
smallest parameter exhibited the largest amplitude response.
The residual displacement increased with the value of the
parameter.
11
13.9 Response to Different Earthquake Motions
The each model was subjected to four different earthquake motions. The maximum response
ductility demand is compared with respect to yielding periods.
(a) Degrading Bilinear Model ( 5 . 0 , 0 . 0 = α )
12
13
13.10 Response of Different Models
Four earthquake records are used in this study; i.e., El Centro 1940 (NS) and (EW), and Taft 1952
(N21E) and (S69E). Maximum response of six different hysteresis models is compared using the
standard stiffness properties. Hysteresis models are (a) RambergOsgood model ( γ =3.79), (b)
Degrading Bilinear model (α = 0.0 and 0.5), (c) Clough model (α = 0.0 and 0.5), (d) Bilinear Takeda
model (α = 0.0 and 0.5), (e) Takeda model (α = 0.0 and 0.5), and (f) Degrading Trilinear model
(F
c
/F
y
= 1/3).
An attained ductility factor is defined as the ratio of the maximum displacement to the yield
displacement. For a design procedure to be conservative, the attained ductility factor should be less
than the allowable ductility factor of 4.0. Note that the Newmark's design criteria give a reasonable
ductility demand from all six hysteresis models for a wide range of yielding periods in the case of El
Centro (NS) 1940 motion. The undamped yield period is a period associated with secant stiffness at
the yield point. The initial uncracked period of an SDF system with a trilinear primary curve is
approximately 70 percent of the yield period.
Although the Newmark's design criteria appear to be acceptable for the El Centro (NS) 1940
motion, the other three earthquake motions caused attained ductility factors much greater than the
allowable value at the various periods. In general, the design criteria are not satisfactory in a very
shortperiod range, for example less than 0.15 sec. Distribution of maximum response with periods
is different from one earthquake motion to another, showing an irregular shape, although each
hysteresis models was designed on the basis of elastic response of individual earthquake motion.
On the other hand, distribution of maximum response with periods is similar from one hysteresis
model to another for a given earthquake motion, implying that maximum response amplitudes of
different hysteresis models can be made comparable if hysteresis parameters of each model are
properly adjusted.
14
For an unloading stiffness degradation parameter of 0.5, the Takeda, Clough and Degrading
Bilinear models developed comparable ductility factors. The Degrading Trilinear model also
developed ductility factors similar in magnitude to those three models at corresponding periods.
Therefore, maximum response amplitudes are not as sensitive to detail difference in hysteretic
rules of these models, but rather are influenced by more basic characteristics of hysteresis loops,
such as stiffness properties to define a primary curve and the fatness (hysteretic energy dissipating
capacity) of a hysteresis loop.
15
13.11 Response Waveforms and Hysteresis Relations
Resistance response normally oscillates about its neutral axis, and its amplitude is limited by the
yield resistance. On the other hand, displacement response does not necessarily oscillate about the
neutral axis, but the residual displacement amplitude is easily shifted by the properties of a
hysteresis model. Therefore, it is easy to study the effect of different hysteretic properties in a
displacement response waveform.
The El Centro (NS) 1940 motion was used for response computation. Five hysteretic models
were used for comparison; i.e., (a) Degrading Bilinear model ( α = 0.0 and 0.5), (b)
RambergOsgood model ( γ = 3.79), (c) Clough model (α = 0.0 and 0.5), (d) Takeda model (α = 0.0
and 0.5), and (e) Degrading Trilinear model (F
c
/F
y
= 1/3). The yielding period of these models was
arbitrarily chosen to be 0.4 sec, and the yield resistance level was taken to be 60 percent of the
standard model to allow a larger inelastic action.
The response amplitude was shown to be influenced by the fatness of a hysteresis loop even if
the stiffness properties of the primary curve are identical. Consequently, the response waveforms
are compared among systems having a relatively fat hysteresis loop and among those having a
relatively thin hysteresis loop.
In all fathysteresis systems, maximum displacement at 2.0 sec, and the second largest
amplitude at around 5.3 sec. The Takeda model shows a shortperiod oscillation at 1.0 sec, since
only the Takeda model has a trilinear primary curve among the models shown in this figure. The
Bilinear model oscillates in a period shorter than the other models, e.g., between 2.5 to 4.5 sec,
attributable to the nondegradation of stiffness with displacement amplitude. The Bilinear and
RambergOsgood models developed residual displacement in the negative direction at 7.0 sec,
whereas the Clough and Takeda models developed positive residual displacement. The former two
models behaved in a manner different from the Clough and Takeda models. The RambergOsgood,
Clough and Takeda models show similar hysteresis relations.
Thin hysteresis models show displacement response waveforms distinctly different from those of
fathysteresis models, oscillating regularly in larger amplitudes and in longer periods. The Degrading
Bilinear model exhibited a behavior different from the other models, especially in a waveform
between 6.5 and 8.0 sec. The Clough, Takeda and Degrading Trilinear models produced similar
displacement waveforms.
The Clough and Takeda models developed similar hysteretic relations although the Takeda
model had a trilinear primary curve. This may be attributable to the fact that a largeamplitude
oscillation occurred at an early stage of the earthquake motion. In other words, the behavior of
Takeda and Clough models can be different if a small oscillation continues for a long duration, or if
the yielding does not occur during an earthquake.
Therefore, the Takeda model is more preferable to the Clough model, although the former model
requires a larger memory in a computer to store the complicated hysteresis rules.
A hysteresis loop of the Degrading Trilinear model appears to be thinner than the Takeda model,
but the Degrading Trilinear model can dissipate larger hysteretic energy during mediumamplitude
oscillation.
The comparison of response waveforms of different hysteresis models points out the less
sensitive nature of response waveforms to a minor difference in hysteresis rules, as long as the
same primary curve is used in conjunction with a comparable capacity to dissipate hysteretic energy.
If maximum response amplitude is known, before analysis, to be much larger than the yield
displacement, the Clough model can produce a response waveform similar to that of the complicated
Takeda model. However, if that premise is not guaranteed, it is more conservative to use a
hysteresis model with a trilinear primary curve in the analysis of the reinforced concrete, recognizing
the stiffness changes at cracking and yielding; i.e., the Takeda model.
16
References:
Amin, M., and A. H.S. Ang, "A Nonstationary Model for Strong Motion Earthquakes," Structural
Research Series No. 306, Civil Engineering Study, University of Illinois, Urbana, 1966.
Otani, S, "Hysteresis Models of Reinforced Concrete for Earthquake Response Analysis," Journal,
Faculty of Engineering, University of Tokyo, Vol. XXXVI, No. 2, 1981, pp. 125156.
Newmark, N. M., "A Method of Computation for Structural Dynamics," Journal, Engineering
Mechanics Division, ASCE, Vol. 85, No. EM3, 1959, pp. 6794.
Veletsos, A. S., and N.M. Newmark, "Effect of Inelastic Behavior on the Response of Simple
Systems to Earthquake Motions," Proceedings, Second World Conference on Earthquake
Engineering, 1960, Vol. II, pp. 895912.
17
18
19
13.12 Effect of Hysteresis Shape on Frame Response
The bond deterioration along the beam longitudinal reinforcement within a beamcolumn joint is
normally thought to be undesirable because the energy dissipation at beam ends is reduced by
pinching in the hysteresis shape; the decay in energy dissipation capacity might increase the
response of a structure during an earthquake. However, it is not practical to expect perfect bond
along the beam longitudinal reinforcement.
The acceptable level of bond deterioration cannot be determined by the laboratory test of a
beamcolumn subassemblage if the problem is to evaluate how much response may be increased
by the decay in the energy dissipation characteristics. The influence of the energy dissipation on the
earthquake response well into an inelastic range is studied to reexamine the acceptable level of
bond deterioration (Kitayama, 1993).
Fourstory, sevenstory and sixteenstory reinforced concrete moment resisting frames were
analyzed with uniform 6.0 m bays and uniform story height of 3.5 m. In the analysis, a single
continuous column with connecting girders on the both sides was removed by cutting the connecting
girders at the midspan, and the girder ends were supported by pinhorizontal rollers. The yield was
allowed at the girder ends, but the columns were assumed to remain elastic during an earthquake.
The mass of each floor was estimated on the basis of member dimensions and floor slab thickness
(= 130 mm). The fundamental periods of oscillation were 0.36 sec, 0.62 sec and 0.98 sec for 4, 7
and 16story buildings.
Each member was represented by the onecomponent model, in which inelastic deformation
(rotation) was assumed to concentrate at member ends; a beamcolumn connection was assumed
to be rigid.
The Takedaslip hysteresis model (Kabeyasawa et al., 1983) at beam ends was selected to
simulate the pinching behavior caused by the bond deterioration along the beam reinforcement. As a
reference, the Takeda model (Takeda, Sozen and Nielsen, 1970) was used to simulate a good bond
situation with a spindleshape hysteresis. The primary curves of the both models were made
identical. Additional deformation caused by the pullout of reinforcement from the connection was not
considered. The forcedeformation relation of the hysteresis models are compared with the
hysteresis relations of interior beamcolumn subassemblages observed in the laboratory test.
Instantaneous viscous damping matrix was assumed to be proportional to instantaneous stiffness
matrix, and the initial elastic damping factor for the first mode was chosen to be 0.05 of the critical.
The 1940 El Centro (NS) and the 1952 Taft (S69E) records were used in the analysis. The
intensity of ground motion was adjusted to develop maximum member ductility factors of
approximately 4.0 at the beam ends using the Takeda hysteresis model.
20
Time histories of displacement response at the rooflevel are compared. Although the
displacement response waveforms of the Takedaslip model deviated from those of the Takeda
model, attained maximum response amplitudes are comparable. Large drift was noted more
frequently in the response waveforms of the Takedaslip models.
The attained story drift angles are compared. The maximum beam ductility factor of 4.0 was
adopted to determine the intensity of ground motion. The maximum drift angle was smaller than 1/50
rad in the three structures with the Takeda model. The story drift angle increased slightly with the
use of the Takedaslip model, by dissipating less kinetic energy; the story drift angle exceeded 1/50
rad in the 16story building using the Takedaslip model with small hysteresis energy dissipation
capacity.
The attained ductility factors at beam ends are compared. The distribution of beamend ductility
factors of a structure with the Takeda model is similar to that with the Takedaslip model (equivalent
damping factor h
e
= 0.15). The change in the energy dissipating capacity in terms of equivalent
viscous damping factors did not affect the ductility demand appreciably.
21
22
23
References:
Giberson, M. F., "The Response of Nonlinear Multistory Structures subjected to Earthquake
Excitation," EERL Report, Earthquake Engineering Research Laboratory, California Institute of
Technology, Pasadena, 1967.
Kabeyasawa, T., H. Shiohara, S. Otani and H. Aoyama, "Analysis of the Fullscale Sevenstory
Reinforced Concrete Test Structure," Journal, Faculty of Engineering, University of Tokyo (B), Vol.
XXXVII, No. 2, 1983, pp. 432478.
Kitayama, K., "Limitation of Beam Bar Bond Deterioration within Beamcolumn Joint," Earthquake
Resistance of Reinforced Concrete Structures, A Volume Honoring Hiroyuki Aoyama, November
1993, pp. 297  306.
Takeda, T., M. A. Sozen and N. N. Nielsen, "Reinforced Concrete Response to Simulated
Earthquakes," Journal, Structural Division, ASCE, Vol. 96, No. ST12, 1970, pp. 25572573.
1
Chapter 14. Reliability of Analysis Methods
14.1 Introduction
Various dynamic tests have been carried out to understand the dynamic behavior of a structure
and to test the reliability of analysis methods. The methods of dynamic testing may be classified by
(a) tests on a real structure or a model structure, (b) behavior within an elastic range or in an inelastic
range, (c) excitation by harmonic or random function, and (d) excitation at specific points within a
model or at the base.
A real structure is normally tested in an elastic range to avoid any damage to structural as well as
nonstructural elements; the structure is excited by a harmonic exciter to generate steady state
oscillation. The vibration characteristics such as natural periods, damping factors and mode shapes
are normally studied. The amplitudes of oscillation were measured at different point during the
steadystate response at resonance frequency. The stateoftheart in testing technique can define
the soilstructure interaction behavior or inplane and outofplane vibration mode shape of floor slabs
(Foutch, 1976).
An old building preceding demolition may be tested in an inelastic range under harmonic excitation.
The behavior of the structure to failure in such a test is significantly different from the behavior during
an earthquake because the structure may fail in lowcycle fatigue mode under steadystate excitation.
Smallscale model specimens are normally used to study the failing behavior due to the limitation
in the capacity of testing facilities or in the research funds. To study the behavior during an
earthquake, specimens may be tested on an earthquake simulator.
In testing a small scale test, the similitude laws must be carefully studied. The basic dimensions
for physical problem are force (mass), length and time. The relations between a prototype and model
may be expressed as
p m
p m
p m
T a T
L a L
F a F
3
2
1
=
=
=
Once the three scale factors a
1
, a
2
and a
3
are selected, then other properties must be determined;
e.g.,
Stress:
p
m
m
m
a
a
L
F
σ σ
2
2
1
2
= =
Young’s modulus:
p
m
m
m
E
a
a
L
F
E
2
2
1
2
= =
2
Velocity:
p
m
m
m
v
a
a
T
L
v
3
2
= =
Acceleration:
p
m
m
m
a
a
a
T
L
a
2
3
2
2
= =
Material properties (strength and stiffness) must satisfy the similitude laws although every property
cannot be satisfied by a scaled model. Note that the gravity acceleration must be the same in the
prototype and model environments.
When a smallscale test specimen is used in a dynamic test, the scale effect should be carefully
studied. The behavior of a specimen is sometimes influenced by its size; especially in shear failure
and bond failure in the reinforced concrete.
The pseudodynamic (online) testing method is also used in the laboratory, in which the response
of a specimen under given earthquake motion is calculated by a computer on the basis of the
observed resistance of the specimen; explicit numerical integration technique is used to determine
the displacement response at the next time step.
The test on an earthquake simulator may be classified into two types; (a) a proof test of a
particular structure against design earthquake excitation, in which the test specimen must represent
necessary stiffness and dynamic characteristics of the prototype structure and the excitation motion
must be carefully selected considering the soil properties of the construction site, and (b) a behavioral
test of a general structure, the test results of which may be used to understand the general behavior
and to test the reliability of an analytical method. The specimen for the behavioral test should
represent a mathematical model of a general structure rather than an actual structure. The difference
in the two types of testing lies in the design of a specimen and selection of loading function, but the
testing technique is the same.
Reference:
1. Foutch, D. A., "A Study of the Vibrational Characteristics of Two Multistory Buildings," EERL 7603,
Earthquake Engineering Research Laboratory, California Institute of Technology, September
1976.
3
14.2 Reinforced Concrete Column
As a simple model representing the
characteristics of the reinforced concrete
structure, a series of reinforced concrete
columns were tested under unidirectional base
excitation on the University of Illinois Earthquake
Simulator (Takeda, Sozen and Nielsen, 1970).
The dimensions of a specimen were 152 x
152 mm in section, reinforced by 4No.4 bars
(129 mm
2
/bar, yield stress of 351 MPa). Shear
reinforcement was diameter 4.8 mm plain wire
(yield stress of 276 MPa). Concrete strength was
30.3 MPa in compression and 2.52 MPa in
tension. A heavy steel weight (8.97 kN) was
attached at the top of the column through a
mechanical hinge so that rotational inertia of the
mass would not affect the response of the
specimen.
The lateral force resistancedeformation at the top of the specimen was calculated to define the
primary curve. Cracking moment was calculated for modulus of rupture concrete using the flexural
theory. The cracking deformation was calculated for the elastic stiffness and cracking moment. Yield
moment was calculated using the parabolic stressstrain relation of concrete and the elastoplastic
stressstrain of reinforcement. The yield deflection was calculated as the sum of (a) deflection caused
by curvature based on cracked section, (b) deflection caused by slip of the reinforcement (assuming
uniform anchorage bond over 20 bar diameter) and depression of semiinfinite plate under flexural
compression stress at the beamcolumn interface, (c) deflection caused by deformation of the test
platform (observed in a static test), and (d) the shearing deflection. The deflection was dominated by
parts (a) and (b) above.
Wire
Section
Unit
Cycle 1 Cycle 4
Displ., inch Displ., inch
Displ., inch
Displ., inch
Observed
Calculated
L
o
a
d
,
k
i
p
s
L
o
a
d
,
k
i
p
s
L
o
a
d
,
k
i
p
s
Cycle 6
Cycle 8
L
o
a
d
k
i
p
s
4
The forcedeflection relation obtained by a static test is compared with the hysteresis relation of
the Takeda model with the calculated primary curve. The model was subjected to the observed
deflection at the top. The overall behavior of the specimen and the model is generally similar.
In the earthquake simulator test, the time axis of earthquake motion was compressed to
oneeighth to satisfy the similitude relations (equal velocity, equal stress and scaled length) and the
capacity of the earthquake simulator. El Centro (NS) 1940 motion was simulated on the table. The
observed motion on the earthquake simulator table was used in the simulation analysis. Damping
factor was assumed to be zero and 2 percent of the critical; the damping coefficient was assumed to
remain constant (mass proportional damping).
The correlation was found more favorable with damping to 3.0 sec from the beginning of the
motion; however, the correlation was better without damping at around 3.5 sec probably because the
constant damping coefficient tends to dissipate more energy after the deterioration of stiffness with
damage.
The study showed that the analysis could simulate a complex behavior of the reinforced concrete
member in an inelastic range if the proper hysteresis model were to be used which was capable of
simulating static behavior under load reversals.
Reference
Takeda, T., M. A. Sozen, and N. N. Nielsen, "Reinforced Concrete Response to Simulated
Earthquakes," Journal, Structural Division, ASCE, Vol. 96, No. ST12, December 1970, pp. 2557 
2573.
(a) Base acceleration
(b) Observed acceleration
(c) Calculated acceleration (h=0.00)
(d) Calculated acceleration (h=0.02)
Time, sec
A
c
c
e
l
e
r
a
t
i
o
n
,
G
5
14.3 Frame Structures
Threestory Shear Model: A simple multistory
building model is a frame with rigid girders, in
which all the deformation takes place only in the
columns. The stiffness of each story can be
defined without coupling with stiffness of the other
stories.
Eto, Takeda and Omote (1973) tested a
onefifth scale threestory onebay frame model
with rigid girders on the Ohbayashigumi
earthquake simulator. Another specimen, designed
and constructed using identical specifications, was
tested under static lateral force reversals.
The resistancedeformation relation of each
column under monotonically increasing force was
evaluated taking into account flexural and shear
deformation and deformation due to pullout of
longitudinal reinforcement at column ends.
The Takeda model (Takeda, Sozen and Nielsen, 1970) was used to simulate the hysteresis
relation under lateral force reversals. The unloading stiffness degradation index of the model was
determined on the basis of the static test. The story shearinterstory deformation relations calculated
and observed in the static test agreed reasonably well.
The damping matrix was assumed to be proportional to initial stiffness, and the damping factor at
the initial elastic stage was determined to be 0.03 for the first mode on the basis of the free vibration
test conducted prior to the dynamic test.
The calculated and observed acceleration response waveforms are compared for test runs R2 and
R3. The columns in the three stories yielded in test run R3, but the analysis indicated no yielding in
the top story columns. The analysis could favorably simulate large amplitude oscillations.
Weight
Weight
Weight
Weight
Weight
Weight
Static test result
Calculated forcedeformation relation
6
(a) Third story acceleration
(b) Second story accleration
(c) First story accleration
(d) Ground acceleration
(a) Third story acceleration
(b) Second story acceleration
(c) First story acceleration
(d) Ground acceleration
Time, sec Time, sec
A
c
c
e
l
e
r
a
t
i
o
n
,
G
A
c
c
e
l
e
r
a
t
i
o
n
,
G
(a) Third story accelration
(b) Second story accelration
(c) First story acceleration
(d) Ground acceleration
(a) Third story acceleration
(b) Second story acceleration
(c) First story acceleration
(d) Ground acceleration
Time, sec Time, sec
A
c
c
e
l
e
r
a
t
i
o
n
,
G
A
c
c
e
l
e
r
a
t
i
o
n
,
G
7
Threestory Onebay Frame: A series of threestory onebay plane frame structures were tested
unidirectional base motion on the University of Illinois earthquake simulator (Otani and Sozen, 1972).
Specimens were approximately onesixth scale (=1/2.5
2
) of an imaginary prototype structure; two
parallel frames were tested to attain the stability in the orthogonal direction. Rigid steel weight was
attached at beam end through mechanical hinges, modeling rigid floor diaphragm.
The dimensions and reinforcement are shown below;
As the similitude relations, the followings were chosen as the basic relations;
(a) Linear scale: L
m
= (1/6.25) L
p
(b) Stress:
p m
σ σ =
(c) Acceleration: a
m
= a
p
where subscript m stands for model and p for prototype, and L: length, σ : stress, a: acceleration.
Linear scale was necessary to test a scaled model. Equal stress became necessary because the
materials used in the specimen were the same or similar to those used in a prototype structures;
strengths and elastic moduli of concrete and steel should be modeled. Equal acceleration was
adopted because the gravity acceleration could not be changed.
Three independent similitude relations can be selected in a dynamic model test. Some
researchers used equal velocity, instead of equal acceleration, as a criterion
The relations for forces and times can be derived from the chosen set of similitude relations; i.e.,
2 2
p
p
m
m
L
F
L
F
= : equal stress
2 2
p
p
m
m
T
L
T
L
= : equal acceleration
Wire
Wire
Wire
Wire
Beam Section
Column section Unit
8
and
p
p
m
p
p
m
m
T
L
L
T
F
L
L
F
=
=
2
Therefore, the time axis of an earthquake record is compressed by 2.5 in the test.
In an earthquake simulator test, a specimen is normally subjected to a series of excitations of
increasing magnitude. The excitation at a preceding test run may affect the response of the specimen
in the following test run.
The response waveforms of two specimens are compared under the same base motion; one
specimen (F1) was previously subjected to three test runs, doubling the magnitude after each test run,
and the other specimen (F2) was virgin at the test run.
The acceleration and displacement response waveforms were observed to be similar in the two
specimens. From this observation, the each test run of a reinforced concrete specimen may be
considered independent as long as (a) the behavior is governed by flexure without decay in
resistance and (b) large amplitude response occurs at the beginning of the motion.
The onecomponent model (Giberson, 1967) was used to represent the distribution of stiffness in
each member; the beamcolumn connection was assumed to be rigid. The momentrotation relation
under monotonically increasing force at a member end was calculated for the antisymmetric bending
moment distribution on the basis of observed material properties of the specimen. The Takeda
hysteresis model (Takeda, Sozen and Nielsen, 1970) was used to represent the forcedeformation
relation under load reversals.
The deformation at a member end due to the bar slip of longitudinal reinforcement within a
beamcolumn connection was considered in the analysis; i.e., the elastic deformation of a tensile
longitudinal bar under uniform bond stress was calculated as the pullout deformation, and the center
of rotation was assumed at the compressive reinforcement at the member end. The primary curve of
the momentbar slip rotation relation was simplified to a bilinear relation and the Bilinear Takeda
model (Otani and Sozen, 1972) was used for the hysteresis relationship. However, the slip type
behavior of bar slip was not represented by the model.
The damping properties could not be determined by the material and geometrical properties of a
specimen. Therefore, two types of damping matrix were studied (Otani and Sozen, 1972); i.e., (a) a
(a) Observed in Test F13
(b) Observed in Test F21
Time, sec
D
i
s
p
l
a
c
e
m
e
n
t
,
i
n
c
h
(a) Observed in Test F13
(b) Observed in Test F21
Time, sec
A
c
c
e
l
e
r
a
t
i
o
n
,
G
9
damping matrix proportional to the constant mass matrix, and (b) a damping matrix proportional to the
instantaneous stiffness. The damping factor for the first mode at the initial elastic stage was selected
to be 0.02 or 0.10.
The test specimen was subjected to an intense base motion causing a firststory drift angle of 1/20
rad. Two response waveforms were studied; (a) the topfloor (relative) displacement response
governed by the fundamental mode of oscillation, and (b) the secondfloor (absolute) acceleration
response with participation of higher modes.
The displacement response of the model using the instantaneous stiffness proportional damping
was not affected by the choice of the initial damping factor, but the higher frequency components in
the acceleration response was suppressed by the use of a larger initial damping factor; this
observation is consistent with the properties of elastic response using the stiffness proportional
damping.
(a) Observed displacement
(b) Calculated displacement
(stiffness proportional damping, h
1
=0.02)
(c) Calculated displacement,
(stiffness proportional damping, h
1
=0.10)
(d) Calculated displacement,
(mass proportional damping, h
1
=0.02)
(e) Calculated displacement
(mass proportional damping, h
1
=0.10)
Time, sec
T
h
i
r
d
s
t
o
r
y
d
i
s
p
l
a
c
e
m
e
n
t
,
i
n
c
h
10
The displacement amplitude was reduced by the use of large initial damping factor in the mass
proportional damping, but the value of initial damping factor did not influence the acceleration
response waveforms. The mass proportional damping has significant influence on the fundamental
mode response.
From the comparison of observed and calculated response waveforms, the mass proportional
damping must be small to better simulate the displacement response waveform, and must be large to
simulate the acceleration response waveform. The mass proportional damping may not be suited for
the simulation of the response of the particular specimen.
(a) Observed acceleration (Test F21)
(b) Calculated acceleration
(stiffness proportional damping, h
1
=0.02)
(c) Calculated acceleration
(stiffness proportional damping, h
1
=0.10)
(d) Calculated acceleration
(mass proportional damping, h
1
=0.02)
(e) Calculated acceleration
(mass proportional damping, h
1
=0.10)
Time, sec
F
i
r
s
t
s
t
o
r
y
a
c
c
e
l
e
r
a
t
i
o
n
,
G
11
If damping proportional to instantaneous stiffness was assumed, the calculated displacement
response waveforms for the two values of initial damping factors agreed reasonably well, while the
calculated acceleration amplitude was too large for the initial damping factor of 0.02, and too small for
the initial damping factor of 0.10. Therefore, the initial damping factor of approximately 0.05 might be
suitable for this particular specimen using the damping proportional to instantaneous stiffness.
If the damping proportional to instantaneous stiffness was used, the large amplitude oscillation of
displacement response was reasonably well simulated, but medium amplitude oscillations at 5 and 11
seconds from the beginning were not favorably reproduced. The calculated response showed
oscillation in a shorter period; the stiffness was estimated to be too large compared to the specimen.
This may be because the hysteresis model for bar slip deformation does not consider the pinching
characteristics.
The yielding was calculated at the ends of secondfloor beam, at top of the secondstory columns
and at the base of the firststory columns, where the yielding was observed in the test.
Pseudodynamic Test: In the nonlinear response of a building structure, the damping proportional to
the instantaneous stiffness gives better correlation with the observed response; the effect of the
damping proportional to the instantaneous stiffness is relatively small on the calculated response in
an inelastic range. Therefore, the equilibrium of inertia force and resistance should be considered.
Using an explicit integration scheme, for example the central difference method, it is possible to
determine the displacement response at the next time step:
2
1 1
1 1
1
1
1
} { } { 2 } {
} { } { } { } {
} { } {
} {
} { } {
} {
t
x x x
t
t
x x
t
x x
t
x x
x
t
x x
x
i i i
i i i i
i i
i
i i
i
∆
+ −
=
∆
∆
−
−
∆
−
=
∆
−
=
∆
−
=
− +
− +
+
+
+
& &
& &
&
t
i+1
t
i
t
i1

x
i+1
x
i
x
i1
1 + i
x&
i
x&
i
x& &
Solving for the displacement {x}
i+1
at time t
i+1
,
i i i i
x t x x x } { } { } { 2 } {
2
1 1
& & ∆ + − =
− +
The resistance vector {R}
i+1
at time t
i+1
can be determined in an experiment by applying statically
the forced displacement {x}
i+1
to the specimen in the laboratory.
The equation of motion without damping at time t
i+1
gives
1 1 1
} ]{ [ } { } ]{ [
+ + +
− = +
i i i
y e M R x M & & & &
The observed resistance {R}
i+1
and calculated displacement {x}
i+1
may be input to the above
relation to obtain the acceleration
1
} {
+ i
x& & .
12
This testing method was called "Online test" or "Pseudodynamic test," originally developed by Dr.
M. Hakuno (1969) using an analog computer.
The method has been extensively used in various earthquake response tests (e.g., Okada and
Seki, 1979). The major advantages of the test method are that (a) the development of damage in the
specimen can be observed and (b) the hysteresis relations need not be assumed in the analysis. The
difficulty in the method is to evaluate the numerical stability associated with the accuracy in the
instrumentation and the procedure and sequence to apply the forced deformation to a
multidegreeoffreedom system.
Teshigawara (1980) tested twostory onebay frames. One specimen (specimen FDR1) was
designed to form the beam yielding mechanism and the other specimen (specimen FDR2) was
designed to form the story sidesway mechanism. The beam section was 150x180 mm, and the
column section 150x150 mm. The horizontal forces were applied at the midspan of the beams.
In the analysis, one component model (Giberson, 1967) was used for beams and columns. The
momentrotation relation was determined by first analyzing the momentcurvature relation of section
by the lamina model, and then member end rotation was assumed proportional to the member end
curvature for the inflection point at the midpoint of the member. The degrading trilinear model
(Fukada, 1969) was used to define the hysteresis relation.
The observed (solid line) and calculated (dashed line) secondfloor displacement and firststory
shear waveforms are compared. The large amplitude oscillation was successfully reproduced by the
model, but calculated small to medium amplitude waveforms shows a shorter response period,
indicating the model stiffness was higher than the observed at a low stress level.
Loading Device
Loading device
Loading device
Loading device
Steel
Steel
Steel
Steel
(a) Specimen FDR1
(b) Specimen FDR2
Unit
13
Calculated
Observed
(a) First story displacement (Test FDR1)
(b) First story shear (Test FDR1)
Time, sec
Free vibration
Free vibration
Time, sec
Free vibration
Time, sec
(c) First story displacement (Test FDR2)
Free vibration
Time, sec
(d) First story shear (Test FDR2)
14
References:
Clough, R. W., and J. Gidwani, “Reinforced Concrete Frame 2; Seismic Testing and Analytical
Correlation,” EERC Report 7615, Earthquake Engineering Research Center, University of
California at Berkeley, 1976.
Eto, H., T. Takeda and Y. Omote, "Dynamic Destruction Test of Threestory Onespan Reinforced
Concrete Frames (Part 1: Test Results) (in Japanese)," Report, AIJ Annual Meeting, October 1972,
pp. 1119  1120, and "ditto, (Part 2: Discussion on Test Results) (in Japanese)," Report, AIJ Kanto
District Meeting, 1973, pp. 45  48.
Fukada, Y., "Study on the Restoring Force Characteristics of Reinforced Concrete Buildings (in
Japanese)," Proceedings, Kanto Branch Symposium, Architectural Institute of Japan, No. 40,
1969, pp. 121124.
Giberson, M. F., "The Response of Nonlinear Multistory Structures subjected to Earthquake
Excitation," EERL Report, Earthquake Engineering Research Laboratory, California Institute of
Technology, Pasadena, 1967.
Hakuno, M., Shidawara, and Hara, "Dynamic Failure Test of Beams Controlled by Computer,"
Transactions, Japan Society of Civil Engineers, No. 171, November 1969.
Healey, T. J. and M. A. Sozen, “Experimental Study of the Dynamic Response of a Tenstory
Reinforced Concrete Frame with a Tall Firststory,” Structural Research Series No. 450,
Department of Civil Engineering, University of Illinois at UrbanaChampaign, 1978.
Hidalgo, P., and R. W. Clough, “Earthquake Simulator Study of a Reinforced Concrete Frame,” EERC
Report 7413, Earthquake Engineering Research Center, University of California at Berkeley,
1974.
Kabeyasawa, T., H. Shiohara, S. Otani and H. Aoyama, "Analysis of the Fullscale Sevenstory
Reinforced Concrete Test Structure," Journal, Faculty of Engineering, University of Tokyo (B), Vol.
XXXVII, No. 2, 1983, pp. 432478.
Kunanath, S. K., and A. M. Reinhorn, “Inelastic Threedimensional Response Analysis of Reinforced
Concrete Building Structure (IDARC3D), Part 1 Modeling,” Technical Report NCEER890011,
State University of New York at Buffalo, New York, 1989.
Kunnath, S. K., A. M. Reinhorn, and Y. J. Park, “Analytical Modeling of Inelastic Seismic Response of
R/C Structures,“ Proceedings, Journal, Structural Engineering, ASCE, Vol. 116, No. 4, April 1990,
pp. 99610117.
Okada, T., and M. Seki, "Earthquake Response Testing of Reinforced Concrete Frames using
ComputerActuator Online System, Part 1: Objectives and Methodology (in Japanese),"
Transactions, Architectural Institute of Japan, No. 275, January 1979, pp. 25  31, "ditto, Part 2:
Online Test 1 (in Japanese)," Transactions, AIJ, No. 279, May 1979, pp. 77  84, "ditto, Part 3:
Online Test 2 (in Japanese)," Transactions, AIJ, No. 280, June 1979, pp. 79  89, "ditto, Part 4:
Discussion of Earthquake Response Characteristics (in Japanese)," Transactions, AIJ, No. 282,
August 1979, pp. 57  64.
Otani, S., "Earthquake Tests of Shear WallFrame Structures to Failure," Proceedings, ASCE/EMD
Specialty Conference, University of California at Los Angeles, March 1976, pp. 298  307.
Otani, S., and M. A. Sozen, "Behavior of Multistory Reinforced Concrete Frames during
Earthquakes," Civil Engineering Studies, SRS No. 392, University of Illinois at Urbana, November
1972.
Otani, S., "Failure Test of Reinforced Concrete Structure, Data for Dynamic Analysis (in Japanese),"
Report, AIJ Annual Meeting, September 1980, pp. 1555  1556.
Roufail, M. S. L., and C. Meyer, “Analytical Modeling of Hysteretic Behavior of R/C Frames,
“Proceedings, ASCE, Journal, Structural Engineering, Vol. 113, No. 3, March 1987, pp. 429444.
Takeda, T., M. A. Sozen and N. N. Nielsen, "Reinforced Concrete Response to Simulated
Earthquakes," Journal, Structural Division, ASCE, Vol. 96, No. ST12, 1970, pp. 25572573.
Teshigawara, M., "Simulation of Nonlinear Earthquake Response of Twostory Reinforced Concrete
Frames by ComputerActuator Online System," M. Eng. Thesis submitted to the University of
Tokyo, February 1980.
15
13.4 FrameWall Structures
Sevenstory Threebay Frame: A
sevenstory threebay frame with a
structural wall in the center bay was
tested on the Ohbayashigumi
earthquake simulator (Koike, Omote and
Takeda, 1980). The test specimen was
approximately onetenth scale of an
imaginary prototype structure.
Story height was uniform and 320 mm,
and bay width was 540 mm. Steel weights
were attached at each beamcolumn joint
through mechanical hinges.
The column section was 60 x 60 mm,
and beam section was 40 x 70 mm; wall
thickness was 20 mm. Four 4mm
diameter indented wires were used as
longitudinal reinforcement, 1.4mm
diameter plane wires as lateral
reinforcement in beams and columns,
2mm diameter wires as wall
reinforcement. Tensile reinforcement ratio
p
t
of the wall and beams was 0.52 % and
0.90 %, respectively. Gross reinforcement
ratio p
g
of columns was 1.40 %. The shear
reinforcement ratio p
w
was 0.21 % in
beams and columns.
In the analysis, the wall and columns
were divided into 13 short segments; the momentcurvature relation was varied in each segment. The
shear and axial deformations were considered in the column and wall. One component model was
used for a beam.
The Takeda hysteresis model (Takeda, Sozen and Nielsen, 1979) was used at beam ends
connected to the wall, and the Takedaslip hysteresis model, especially developed for this study, was
used at the beam ends connected to the exterior columns. The Takeda model was used for the
momentcurvature relation of the wall and column segments. The stiffness of the shear deformation
was reduced to onetenth after shear cracking, but the Clough model (Clough and Johnston, 1966)
was used as the hysteresis relation. The parameters of the hysteresis models were determined on
the basis of the observation in the static tests on individual components.
The specimen was tested using the Hachinohe Harbor (NS) record recorded during the 1968
Tokachioki earthquake. The observed and calculated response waveforms are compared for the
topfloor acceleration and displacement.
Reference:
Koike, Y. Omote and T. Takeda, "Reinforced Concrete Wallframe Structures subjected to Dynamic
Loading  Model Tests and the Simulations," Proceedings, Seventh World Conference on
Earthquake Engineering, Istanbul, Vol. 6, 1980, pp. 419  426.
Takeda, T., M. A. Sozen and N. N. Nielsen, "Reinforced Concrete Response to Simulated
Earthquakes," Journal, Structural Division, ASCE, Vol. 96, No. ST12, 1970, pp. 25572573.
Weight
Column
Beam
Wall
Reinf.
Beam
Column
Indented Wire Unit
16
M
e
m
b
e
r
m
o
d
e
l
B
e
a
m

e
n
d

1
B
e
a
m
e
n
d

2
C
o
l
u
m
n
,
w
a
l
l
e
l
e
m
e
n
t
H
y
s
t
e
r
e
s
i
s
m
o
d
e
l
s
T
e
s
t
A
n
a
l
y
s
i
s
W
a
l
l
s
h
e
a
r
B
e
a
m
e
n
d

2
B
e
a
m
e
n
d

1
S
u
b

e
l
e
m
e
n
t
W
a
l
l
B
e
a
m

c
o
l
u
m
n
j
o
i
n
t
B
e
a
m
C
o
l
u
m
n
17
Seventh floor acceleration (observed)
Seventh floor acceleration (calculated)
Seventh floor displacement (observed)
Seventh floor displacement (calculated)
Base motion
Time, sec
A
c
c
e
l
e
r
a
t
i
o
n
,
c
m
/
s
e
c
2
D
i
s
p
l
a
c
e
m
e
n
t
,
m
m
A
c
c
e
l
e
r
a
t
i
o
n
,
c
m
/
s
s
e
c
2
18
13.5 Wall Structures
Tenstorey Coupled Shear Walls: Tenstorey coupled shear walls were tested on the University of
Illinois earthquake simulator (Aristizaba1Ochoa and Sozen 1976). Takayanagi and Schnobrich (l976)
divided a wall into short segments of uniform stiffness, and represented connecting beams by the
onecomponent model. The TakedaTakayanagi model with changing axial force was assigned to a
wall element, and the TakedaTakayanagi model with pinching action and strength decay was used in
a beam. It was judged that the usage of twodimensional plane stress elements for the walls was less
desirable because such an approach might cost more computational effort without any compensating
increase in accuracy.
The amplitude of the exterior column axial load varies greatly due to the earthquake overturning
moment, and changes its momentcarrying capacity. Takayanagi and Schnobrich (1976) incorporated
the effect of axial force variation in the Takeda model by preparing various backbone curves at
different axial load levels.
(a) axial force variation (b) pinching and strength decay
TakedaTakayanagi models (Takayanagi and Schnobrich 1976):
A pinching action and strength decay are inevitable in a short and deep member due to bar slip
and deterioration in shear resistance. Takayanagi and Schnobrich (1976) introduced a pinching
action and strength decay in the Takeda model. Whenever a response point was located in the
positive rotationnegative moment range or the negative rotationpositive moment range, the pinching
was introduced. After the moment exceeded the yield level, strength decay was incorporated. The
values of guideline for strength decay and pinching stiffness were not related to the member
geometry and material properties.
19
(a) displacement at level 10, in inches
(b)Acceleration at level 10, g.
Analysis of tenstory coupled shear wall (Takayanagi and Schnobrich 1976)
The comparison of the measured and
calculated displacement and acceleration is
excellent. It is necessary to include the
effects of inelastic axial rigidity of the wall
section and pinching action and strength
decay of the connecting beams to reproduce
the maximum displacement response and
the elongation of the perio