P C M RC-M (C & B) S B L: Katchalla Bala Kishore
P C M RC-M (C & B) S B L: Katchalla Bala Kishore
P C M RC-M (C & B) S B L: Katchalla Bala Kishore
11
PROBABILISTIC CAPACITY MODEL FOR RC-MEMBERS
(COLUMNS & BEAMS) SUBJECT TO BLAST LOADING
Submitted
By
CERTIFICATE
It is certified that the work contained in the project report entitled Probabilistic Capacity
Model for RC-Members (Columns & Beams) Subject to Blast Loading, by Katchalla
Bala Kishore (134104016) has been carried out under my supervision and that this work has
not been submitted elsewhere for the award of a degree or diploma.
ACKNOWLEDGEMENTS
ABSTRACT
The failure of Reinforced Concrete (RC) structures is increasing by the day, all over
the world due to natural as well as man-made events, blast loading being one of the prime
causes in case of man-made disasters. The number of occurrences of blast loading on the
structures has noticeably increased all over as a part of anti-social/terror activities, accidental
References 4
44
explosions and similar causes due to negligence. As a result, in addition to injuries to the
mankind and habitations, the structural components like column, beam and slab sustain
damage and eventual failure of the system they are part of. So examining the response of RC
components becomes an important issue at present time.
This research aims to develop the probabilistic model for estimating the deflection
capacity of RC-Column and RC-Beam which are integral and vital component of structural
system subject to blast loading. A parametric analysis of different dimensioned (ranges in
practice) RC columns and RC beams subject to blast loading are carried out. The deflection
capacity of the RC column and RC beam are estimated under ultimate limit state, subject to
blast loading. Detailed Finite Element (FE) models of RC column and RC beam subject to
different blast loading scenarios are simulated and a statistical analysis of FE simulated data
has revealed a high level of variability of deflection capacity; where variability would be
expected to be minimum. Hence a probabilistic capacity model for RC column and RC beam
subject to blast loading is then developed for estimating the deflection capacity that considers
variability and/or model error.
The developed probabilistic capacity models account the realistic aspects such as
multi-modal response of the structure, the interaction among the different components and the
inherent uncertainty associated with the modeling, configuration and material properties of
the concrete structure. Hence the current research will facilitate the provisions for
improvement and modifications in existing codal provisions for blast resistant design of
structures.
TABLE OF CONTENTS
Abstract .i
Table of Contents.....ii
List of Figures......iv
List of Tables....vi
Chapter 1. Introduction......................................................................................................1
1.1. Intoduction.................................................................................................................1
References 5
55
1.2. Motivation of Research..............................................................................................2
1.3. Objective....................................................................................................................2
1.4. Methodology..............................................................................................................3
4.4.4. Assessment.......................................................................................................31
6.2. Conclusions..............................................................................................................46
References ..49
LIST OF FIGURES
References 7
77
Fig. 1.1 Collapse of Yichang Bridge in China when a truck carrying fireworks exploded (left)
and Building damage in India due to blast (right). (BBC News, February 1, 2013 and
Fig. 2.1 Decay of Pressure with distance and time [Ngo et al. (2007)].....................................5
Fig. 2.4 Strain rates exhibited by different types of loading [Ngo et al. (2007)].....................10
Fig. 3.3 Schematic view of (a) SDOF system and (b) Blast loading.......................................17
LIST OF TABLES
CHAPTER 1. INTRODUCTION
1.1. INTODUCTION
As a part of the urban environment or infrastructure and as a part of different types of
civilian and military facilities, a large number of Reinforced concrete (RC) structures exist. It
is known that tremendous strides have been made on technologies meant to improve the state
of human life. However, often such technologies have come with new threat for damage or
collapse of structures in the form of Explosions and Blast detonations. Explosions due to
different civilian accidents or intentional events, detonations of high explosives or weapons
result in extreme loading conditions on all objects nearby. It may cause a non-recoverable
References 9
99
damage to the infrastructure and also loss of life within its proximity, thereby leading to great
consequences on regional, national, economic, socioeconomic and security. Thus, the safety
of life and infrastructure against the blast effects got an increased attention.
Fig. 1.1 Collapse of Yichang Bridge in China when a truck carrying fireworks exploded (left)
and Building damage in India due to blast (right). (BBC News, February 1, 2013 and The
Hindu News, March 12, 1993)
In this sense, Structural Engineers are receiving the renewed attention and challenge
for designing a structure resisting the blast load. In the design of RC Building to resist the
blast load or any severe dynamic load, it is not economical to only consider the elastic
response of the structural element. Therefore, the structural elements should be allowed for
certain plastic deformations, which better utilizes their energy absorbing capabilities if
exposed to any dynamic loads. It is also important to design for ductile response of the
member, so that the partial or total collapse of the structure can be prevented due to local
failure of the structural elements.
1.3. OBJECTIVE
This research is mainly aimed at developing probabilistic models for estimating the
deflection capacity of RC members subject to blast loading. This research also aimed at to
make the design philosophy handy with the developed model. The predicted values of the
model will account the following:
Multi-modal response of the structure
Interaction among the different components and inherent uncertainty associated with
the modeling, configuration and material properties of the concrete structure.
1.4. METHODOLOGY
A total of 50 columns and 50 beams of different sizes and reinforcement ratios in
realistic ranges are selected and subjected to blast loading of varying charge and stand-off
distance. Then the deflection capacity of these columns and beams at ultimate performance
level will be estimated. Also, their numerical deflection capacities are to be obtained and
depending the level of damage, the FE data should be segregated into equality and lower
bound data. Then, for all the members in the equality data, corresponding analytical results
are integrated with the FE results to develop the model.
References 11
1111
1.5. SCOPE OF WORK
The analysis assumes air-blast on the structural components. Fragments and their
effects are neglected.
The blast loading is chosen as uniformly distributed load on member for mechanical
model calculations.
No other material than reinforced concrete (RC) was used. Hence the methodology is
valid only for RC elements.
Research is limited to beams and columns in flexure with propped-cantilever and
fixed boundary conditions.
This research is limited to ultimate performance level only.
Review of Literature
For the past few decades, considerable researches are going on to understand the
explosion science and emphasis has been given to study its effect on the strength and stability
of structural members and to design for its safety. Due to the nature of blast and its effect on
human mankind, most of the available literature on blast characteristics and its experimental
response are restricted. This led to a confinement of literature in this review. However, the
information about the development in this field is made available mostly through the
publication of the US Army Corps of Engineers, Departments of The Army, The Navy and
References 12
1212
The Air Force and other government offices and public institutes. The following reviews
some of the work done in this area.
An explosion or blast is the sudden release of energy within a limited volume, leading
to an increase of light and temperature; but above all an enormous increase in pressure. In
general, an explosion in air generates a pressure bulb that grows in size at supersonic
velocity, where some part of energy is released as flash (thermal radiation), some part into air
and some part into ground, as radially expanding shock waves.
An external blast wave is created when the atmosphere surrounding the explosion is
pushed back due to a compressive pulse travelling outward from the center of the explosion
(Kinney et. al., 1985).
Fig. 2.2 Decay of Pressure with distance and time [Ngo et al. (2007)]
This development of both positive phase and negative phase can be described with a
pressure-time relationship, as schematically shown in Figure 2.2. The overpressure in the
positive phase is considerably higher than in the negative phase. Also, the duration in the
positive phase is shorter than in the negative phase, resulting in an impulse of the negative
phase that is somewhat larger than in the positive phase.
The most commonly used blast wave scaling is the cube root scaling law otherwise
known as Hopkinsons Law (Baker, 1973). The blast wave scaling law defined by Hopkins
(1915) states two different weights of the same explosive have same blast characteristics at
some scaled distances in similar atmospheric conditions. The Hopkins scaling distance is:
R
Scaled distance (or) Proximity factor ( Z )= (m/kg1/3)
Q1 /3
R is the Stand-off distance (m)-(distance between center of explosion and the
structure)
For example, to produce a given blast overpressure at twice given distance, it requires
eight times the explosive energy release.
Another scaling law known as Sachs scaling law (1944), attempts to correct for
changes in atmospheric pressure (Baker, 1973).
References 15
1515
1/ 3
R Po
Scaled distance ( R )= 1 /3
E
E is the energy of the charge and Sachs law assumes air behaves as perfect gas.
Air blasts can be further subdivided into air bursts or surface bursts, depending on the
height above ground of the detonation. Free-air bursts are a special case of air bursts where
the detonation occurs directly above the structure such that no interference takes place with
the blast wave before reaching the structure (TM 5-1300, 1990).
Air burst environment is produced by the detonations that occur above the ground
surface and at a distance away from the protective structure, so that the initial shock wave,
propagating away from the center of explosion, impinges on the ground surface prior to
arrival at the structure, however depends on the stand-off. Similar to that, the same happens
with the surface burst, but the detonations that occur will be very close to the ground surface
or all most on the ground surface. But when the charge is close to the ground, the blast wave
will be reflected (mirrored) and reinforced by the ground. In this case, incident and reflected
wave merge to form a single wave and travel hemi-spherically, whereas the airburst waves
initially have a spherical wave front. Theoretically, the merged blast wave is more powerful
and its effect is equivalent to an explosive of twice the charge in free air. However in real,
due to some ground movement and cratering, equivalent increase in the explosive charge is
given a factor 1.8 (Goel et al., 2012).
After the detonation of charge, enormous energy will be released in the form of
waves, which has a pressure on very higher side of ambient air pressure, which is referred to
as positive phase; followed by negative phase. However, considering the impulse of both the
phases, the load is normalized to a triangular load accounting only positive phase of an
explosion.
References 16
1616
Several authors have recommended their formulations for predicting the peak
overpressure from conventional explosives. Following are the most widely used formulations.
Brode (1955), analyzed the differential equation of gas motion in Lagrangian form
and presented the analytical solution for the peak positive over-pressure (P m) in near-field and
medium to far-field conditions as:
6.7
Pm = +1 bar (Pm > 10 bar)
Z3
Q Q 0.5
Pm = 6784 + 93 ( 3 ) (bar)
R3 R
Kinney et al., (1985), based on the experimental analysis of large explosion data;
presented the following equation to compute the peak positive over pressure:
[ ( )]
2
Z
808 1+
4.5
Pm = (bar)
[ ( ) ][ ( ) ][ ( ) ]
2 2 2
Z Z Z
1+ 1+ 1+
0.048 0.32 1.35
Held (1983), based on the experimental analysis of explosion data; presented the
following equation to compute the peak positive over pressure:
2
Pm = (MPa)
Z2
Sadovskiy (2004), based on explosion data analysis; presented the following equation
to compute peak positive over pressure:
Pm
log 10
log 10
For analyzing the loading, the explosion is taken into ideal form having only the
overpressure phase (positive phase) and several authors have recommended their
formulations for predicting is the time between decay of the shock front from peak pressure
to zero over pressure of the positive phase of explosion, i.e., the positive overpressure
duration (td).
Henrych (1979), presented the following equation to compute the positive over
pressure duration:
1/ 3
Q
2.75+ 0.27 log Z +log
td = (millisecond)
e
Kinney et al., (1985), presented the following equation to compute the positive over
pressure duration:
10
Z
( )]
980[1+
0.54
td = Q1/3 (millisecond)
Sadovskiy (2004), presented the following equation to compute the positive over
pressure duration:
td = 1.2 6 Q 2 R (millisecond)
References 18
1818
Izadifard et al., (2010), presented the following equation to compute the positive over
pressure duration:
During an explosion, a pressure bulb is formed and pressure waves gets expanded
with a spherical wave front. These waves strike with the ground surface and reflects back
reinforcing with the incident waves and then travels parallel to earths surface with an
enhanced pressure, velocity and duration. This shock front generated with the interaction of
incident and reflected wave is known as Mach front or Mach stem and the pressure field
under it is very critical with cylindrical shock front. From the analytical studies, tall structures
should be designed for critical ground-zero distance (the horizontal distance between the
location of ground-zero point and Mach stem origin, where the point on ground vertically
below explosion is ground-zero) in the case of surface blast (Dharaneepathy et al., 1995).
For the cases of blasts under bridge deck, severity of blast effect depends on the
bridge geometry and the clearance under the deck. In many cases, for the bridges with deep
girders, due to confinement effect the blast pressure got enhanced and resulted in more
damage than predicted. Therefore, higher clearances should be provided for reducing the
confinement effect and average loads on piers can be expected when the detonation takes
place at increased average stand-off distance to the pier (Winget et al., 2005).
Byfield in 2006, from a case study on the behavior of non-military building that are
damaged by vehicle bombs during the World War II, most of the buildings exhibited an
impressive resistance against the damage from the blast effects due to their closely spaced
column standards. Whereas many modern commercial buildings are more susceptible to
progressive collapse due to the wider column centerlines. The progressive collapses are due
to poor connection of beam to column and he conventional design standards creates over-
References 19
1919
strong beams followed by weak and brittle connections which are incapable of carrying
higher rate loading of vehicle bombs.
Fig. 2.5 Strain rates exhibited by different types of loading [Ngo et al. (2007)]
Based on the work performed by Curione (1958) and ASCE (1985) referenced in
Bischoff et al., (1991), the design compressive strength of concrete can increase about 25 to
30% percent during dynamic loading of concrete.
In analysis and design of blast resistant structures, materials enhanced strength due to
high rate of loading is taken into account by using Dynamic Increase factor (DIF), which is
the ratio of dynamic to static strength.
fy is the yield stress of steel in MPa and s is the strain rate of steel.
fc is the yield stress of steel in MPa and s is the strain rate of concrete.
TM 5-1300 (1990) and UFC 3-340-02 (2008), introduced DIF values for steel and
concrete which can be directly used in the design of blast resisting structures.
Dynamic nonlinear Finite Element analysis is the most accurate method available for
studying the response of structure under blast loading (Krauthammer, 1998; and Conrath et
al., 1999). The finite element (FE) method can capture local effects, and can allow the analyst
to model a component response to match response observed through experiments. The FE
method requires more time and expertise, and is often reserved for research, forensics, and
for complex configurations that cant be modeled as SDOF systems.
Considering the problem into account that cost of actual RC building and performing
realistic blast is almost impracticable; Luccioni et al. (2003), analyzed the failure of an RC
building caused by a blast load using the hydrocodes that are integrated in software
AUTODYN. The numerical results made an accurate match with the photographs of original
damaged building by analyzing for the same magnitude and location of the explosion which
took place earlier and so, the numerical models are reliable to represent the behavior of a
structure when subject to blast, by some simplified assumptions made for structure and
material.
Based on the work done by Yaramada (2010), three material models are selected from
the library of LS-DYNA (Finite Element Modeling Software), in order to study the response
of concrete to high strain rate loading. They are material Type 159 (Continuous Surface Cap
Model, CSCM), material Type 84 (Winfrith Concrete) and material Type 72R3 (Concrete
Damage REL3).
2.3.
References 23
2323
In case of dynamic loading, the structural response is also time-dependent and hence
varies with time. Dynamic response of the structure is usually measured in terms of
deformations (displacements or rotations), velocity and acceleration.
In the investigation of the dynamic response of a structure subject to a blast load, the
following approach is to be made:
(b) The natural period of response of the structure/structural element must be determined.
(c) The positive phase duration of the blast wave is then compared with the natural period
of response of the structure.
Based on the above comparisons (c), the response of the structure can be described as
follows:
(a) If the positive phase duration of the blast pressure is shorter than the natural period of
vibration of the structure, the response is described as impulsive. In this case, most of
the deformation of the structure will occur after the blast loading has diminished.
References 25
2525
(b) If the positive phase duration of the blast pressure is longer than the natural period of
vibration of the structure, the response is described as quasi-static. In this case, the
structural deformation occurs whilst the blast loading is still being applied
(c) If the positive phase duration of the blast pressure is close to the natural period of
vibration of the structure, then the response is described as dynamic. In this case, the
deformation of the structure is a function of time and the response is determined by
solving the equation of motion of the structural system
In order to evaluate the dynamic structural response subject to blast load, the
following methods are in state of practice:
However, structural response can also be analysed by Multi Degree Freedom (MDOF)
method and Continuous System method; but due to their complexity in calculations and
consumption in time, they are of not in the research interest.
In this method, the approach is based on a statement that, the maximum potential
energy from the static load should be equal to the maximum potential energy from the
dynamic load. Since the total energy of an un-damped system is constant over time, it is fair
to assume that the maximum kinetic energy is reached when the potential energy is zero.
Therefore, potential energy in static case is equal to maximum kinetic energy in dynamic
case.
For the elastic case, the equivalent static load (F) can be expressed as:
F=k x e
k x 2e F2 I2
E= = =
2 2k 2m
References 26
2626
t1
here (t0 , t1) = time-interval and F(t) is forcing function; and m = mass.
Hence, F = I
k
m
For the plastic case, the maximum static load (F) is determined by the maximum
resistance, Rm.
I2
Hence, F =
2mxp
Fig. 3.8 Schematic view of (a) SDOF system and (b) Blast loading
References 28
2828
From the idealised triangular pulse having a peak force Fm and positive phase duration
td as shown in Figure 3.3 (b), the forcing function can be given as:
( tt )
F ( t )=F m 1
d
The blast impulse can be approximated as the area under the force-time curve, and is
given as:
1
I = Fm t d
2
m x +c x + kx=F (t)
where m is the structural mass; c is the structural damping; and k is the structural
stiffness.
As the structural damping has a very little effect on the structures undergoing larger
deflections, the effect of damping can be neglected. Hence, the equation of motion can be
reduced to:
m x +kx =F (t )
The values of mass, stiffness and external force values used in the equation of motion
are actual values only if all the elements of the mass of the structure experience the same
force and consequently move as a single unit, in which the entire mas may be assumed to be
concentrated at its centre of gravity. In the case of blast loading, the motion of particles of
mass varies along the length of the member; hence, the assumption of uniform motion of
entire mass cannot be made and transformation factors are required for the structure to be
represented as an equivalent SDOF system.
8 MN 4 (M N +2 M P)
re= r u=
L 2
L2
12 M N 8( M N + M P )
re= r u=
L2 L2
384 EI 384 EI
307 E I
L4 5 L4
L4
I is Moment of Inertia;
1 4
I* (average cracked Moment of Inertia) = Ig+ Ic (Magnusson, 2007);
5 5
The modified equation of motion for equivalent SDOF system for a time ranging from
0 to the positive phase duration, td is given as:
t = time of loading
Fm F m sin ( t )
x=
k (1cos ( t ))+
k td (
t )
Where, x is the displacement of the structure; is the natural frequency, given as
=
k
m
and Natural Time period, T N =
2
On solving the displacement equation for maximum response by equation the velocity
to zero, i.e., dx/dt = 0, the maximum elastic displacement (xe) is obtained as (Biggs, 1964):
References 31
3131
Fm F
( 1cos ( 2 tan t d ) ) ) m ( sin ( 2 tan1 ( t d ) )2 tan 1 ( t d ) )
1
xe= (
k k td
The maximum elastic response in an elastic state, can also be obtained from DLF
plots that are presented in TM5-1300 manual or in UFC 3-340-02 manual. Where DLF is the
Dynamic Load Factor and is defined as the ratio of the maximum dynamic deflection to the
deflection which would have resulted from the static application of peak load.
xm
Thus the Dynamic load factor (DLF) is given by: DLF =
xs
where, xs = static deflection or, the displacement produced in the system (elastic or
Under high velocity impact or blast loading, structural elements are expected to
undergo large deflections; and hence these elements are also analysed beyond the elastic limit
i.e., elasto-plastic state, after which structure completely transforms into plastic state, forming
a condition of mechanism.
References 33
3333
The equation of motion for equivalent SDOF system for a time ranging from 0 to the
positive phase duration, td in elastic stage is:
m x +
k x = Fm 1 ( tt ) d
and the solution for response in elastic stage is: (Biggs, 1964)
Fm F m sin ( t )
x=
k
(1cos ( t ))+
k td ( t )
The equation of motion in plastic stage is:
t
m x + r u = Fm 1 ( ) td
and the solution for response in plastic stage is: (Biggs, 1964)
t d t m
()
( 1 )
cos
r F
x p=x e + u m
k
where xp is the maximum plastic displacement; t m being the time at maximum elastic
displacement; and xe is the maximum elastic displacement, given as:
Fm F
( 1cos ( 2 tan t d ) ) ) m ( sin ( 2 tan1 ( t d ) )2 tan 1 ( t d ) )
1
xe= (
k k td
re
X E=x e +x p 1( ) ru
The maximum elasto-plastic response can also be obtained from the plots that are
presented in TM5-1300 manual or in UFC 3-340-02 manual. Figure 3.5 shows the Maximum
Response of Elastic-Plastic, One-Degree-of-Freedom system for Triangular Load.
References 34
3434
However, a detailed review of Finite Element (FE) simulations are provided in the
fore coming chapter.
Blast Analysis 36
4.1. INTRODUCTION
Estimating the response or maximum deflection of an RC member through a simple
SDOF analysis is erroneous, as they doesnt include other modes and it is also practical that
any member can fail in any higher mode. In this sense, the present research aims to develop
probabilistic models for estimating the capacity of Reinforced Concrete column which are
integral and vital component of structural system subject to blast loading. A parametric
analysis of RC columns and RC beams subject to blast loading are carried out. The ultimate
deflection capacity of the RC column is estimated due to blast loading. Detailed Finite
Element (FE) models of RC column subject to different blast loading scenarios are simulated.
Using the results of the FE models, probabilistic models for estimating the deflection capacity
are developed for Ultimate/Plastic Performance level.
low to medium level blast protection and so, a blast threat level from a bomb or any explosive
which can be carried by luggage, automobile and van (FEMA 428) are accounted.
The experimental design is split into design of column/beam and design of load cases.
The experimental design considers 50 RC columns and 50 RC beams subject to various blast
pressures by varying the blast charge and its stand-off distance. The load cases are separately
designed to create cases of all combinations during a Blast detonation. In case the
experimental design considered the column variables and load cases together, there could
have been cases when some columns do not experience the blast events to impractical
scenarios. This is avoided by splitting the experimental design into basic and dependent
variables and then performing a realistic combination of the RC column parameters as well as
blast loading parameters.
The D-optimal point selection scheme (Myers et. al., 1995) is used for the selection of
the best set of cases from a given range. The D-optimal scheme is chosen because it has the
flexibility of allowing any number of designs to be placed appropriately in a design space
with an irregular boundary. The D-optimal scheme is also the recommended point selection
scheme for polynomial response surfaces (Livermore Software Technology Corporation
2006). Overall 50 cases of Columns and 50 cases of beams are simulated to access the
probabilistic deflection capacity under blast loading at ultimate performance level. Table 4.1
and Table 4.3 show the range of basic variables considered in this research; and Table 4.2 and
Table 4.4 show the range of dependent variables. FE models of the 50 combinations of RC
columns and RC beams with the blast loading scenarios are made in the next section.
Blast Analysis 38
2
Spacing of Transverse Bar s (100 d 2 ( B+ D ) )/(BD s)
4 l
Clear cover (mm) c ( 0.050.1 ) D
Blast Analysis 39
Diameter of transverse bar (mm) ds 9.525, 12.7, 15.875, 19.05, 22.225, 25.4
2
Spacing of Transverse Bar s (100 d 2 ( B+ D ) )/(BD s)
4 l
A three dimensional solid model is used for modeling the RC column/beam subject to
blast loading. The RC column/beam is modeled by a fully integrated quadratic eight node
element with nodal rotations. A rate dependent material model is used for all the materials
due to the sensitivity of material properties. Concrete is modeled as 3D-solid elements with
the formulation of continuous surface cap model (CSCM) available in the software LS-
DYNA. This is a cap model with a smooth intersection between the shear yield surface and
hardening cap; and this model contributes for erosion in concrete and also takes into account
the strain rate dependency of the concrete strength. The reinforcement bars are modeled
explicitly as one dimensional element with the formulation of an elasto-plastic material as it
accounts for strain rate dependency and also failure based on plastic strain. The contact
between the concrete and reinforcement is modeled using the Lagrangian coupling method.
This method provides the coupling mechanism for steel and concrete interaction by allowing
the coupling between edges of each part and saves the effort of matching the nodes of the
reinforcement and the concrete which might be very difficult in some cases. Mesh refinement
is done and a convergence is achieved at 25 mm mesh size and minimization of the hourglass
energy is ensured.
The end restraints selected for these FE simulations of columns are those of a
propped-cantilever. These boundary conditions are a fixed condition at bottom of the column
and a pinned condition (simply supported) at top of the column. Here, the pinned condition at
top of the column models the superstructure, which is assumed to be axially stiff along its
primary axis, though capable of allowing rotations to occur because of its limited flexural
rigidity. The fixed condition at the base of the column models the column foundation. It is
assumed that the axial loads experienced by the any column in its service typically do not
Blast Analysis 42
exceed the balance point of the column, and any applied axial load only improves the shear
and flexural capacity of the members. Thus, simulating the columns as propped-cantilevers
with no applied axial load is expected to be conservative.The end restraints selected for these
FE simulations of beams are those of a fixed boundary conditioned, where both left and right
sides of the beam are restricted from any sort of translations and rotations.
A spherical-free air burst at the mid height of the column/beam is simulated with
variation of blast charge (Q) and its stand-off distance (R) as obtained from the experimental
design done in the previous section. To overcome the current limitations of conventional
approach in blast load simulations, an approach similar to Slavik (2009) and Agrawal et al.
(2008) is adopted. In this approach, an Arbitrary Lagrangian Eulerian (ALE) formulation is
used. A single element layer of ambient formulation makes up the exterior surface of the air
domain which faces the blast. Its function is to receive information from the blast loading as
provided by ConWep formulation (LSTC, 2006), which is then converted to thermodynamic
state data and subsequently applied as a source to adjoining ALE air elements. The air mesh
interacts with Lagrangian structural elements to apply the load on the elements. Figure 4.1
shows a schematic view of blast loading on FE model of RC Column/Beam.
Blast Analysis 43
4.4.4. ASSESSMENT
The following sections illustrate the numerical capacity assessment for RC column
and RC beam subject to blast loading. Figure 4.2 shows an example FE model of RC Column
at different stages of evolution of blast loading.
Blast Analysis 44
For obtaining the deflection capacity at desired performance level, time at first crack
(tc) and time at ultimate capacity (tu) are to be known to find crack deflection and ultimate
deflection. Time at first crack is the time at which plots of internal energy and total internal
energy gets separated. The time at which the internal energy reaches its peak, is the time (t u)
that accounts for ultimate capacity. As the current research interest is for ultimate limit state,
time corresponding to peak of internal energy plot (t u) is noted and for that tu corresponding
value of displacement can be obtained, i.e., its ultimate deflection capacity.
Figure 4.3 and Figure 4.4 shows the variation of Energy and Deflection with time, at
mid height of the RC column (example).
Blast Analysis 45
Figure 4.5 shows the comparison between analytical deflection capacity and Numerical
deflection capacity of RC Columns subject to Blast loading.
For obtaining the deflection capacity at desired performance level, time at first crack
(tc) and time at ultimate capacity (tu) are to be known to find crack deflection and ultimate
deflection. Time at first crack is the time at which plots of internal energy and total internal
energy gets separated. The time at which the internal energy reaches its peak, is the time (t u)
that accounts for ultimate capacity. As the current research interest is for ultimate limit state,
time corresponding to peak of internal energy plot (t u) is noted and for that tu corresponding
value of displacement can be obtained, i.e., its ultimate deflection capacity.
Blast Analysis 47
Figure 4.6 and Figure 4.7 shows the variation of Energy and Deflection with time, at
mid height of the RC beam (example).
CHAPTER 5. PROBABILISTIC
CAPACITY MODEL AND VALIDATION
Using the results in FE simulations, the probabilistic models are developed for
estimating the probabilistic deflection capacity of a RC column and RC beam subject to blast
loading under ultimate performance level. The calculated 50 values of the ultimate deflection
from the FE models are used to formulate the probabilistic models for RC column and RC
beam. The estimated probabilistic capacity models takes into account the multi-modal
response of the structure, the interaction among the different components and the inherent
uncertainty associated with the modeling, configuration and material properties of the
concrete structure. The equality data as well as lower bound data are used to construct the
probabilistic models.
ln [ Pi ( x , P ) ]=ln [ pi ( x ) ]+ P ( x , P ) + P e P
i i i i i
P e P = model error,
i i
e P =Gaussian error and
i
P = P = set of unknowns model
i i
For a RC column under ultimate performance level, analytical model for deflection
developed by Newmark is used in this research (Brooks et al., 1953).
Xm td
XE
1+ ( )( 12 13 )
ru
is explained as the ratio of yield strength to average applied load, =
P
ru
X E=
KE
Xm td
XE
1+ ( )
6
The model correction terms are used to capture the physical phenomena that are not
accounted for in the mechanical model. The first explanatory function, h 1 accounts for the
contribution of neutral axis (NA) depth and limiting NA depth ( x x ) . The second
limit
explanatory function, h2 accounts for the contribution of the longitudinal reinforcement ratio
( bDA ) . The third explanatory function, h accounts for the contribution of the slenderness
s
3
Blast Analysis 52
ratio ( Hr ) ; where r is the radius of gyration. The fourth explanatory function, h 4 accounts
for the contribution of the cross section ( Db ) . The fifth explanatory function, h5 accounts
td
for the contribution of positive blast duration and natural period of structure ( ) TN
. The
sixth explanatory function, h6 account for the contribution of dynamic increase factors of
DI F steel
steel and concrete ( DI F concrete ) .
A step-wise deletion process is used to arrive at the posterior statistics for the capacity
model. The deletion process is stopped when the standard deviation of the model increases.
The following parameters correct for, the contribution of neutral axis (NA) depth and limiting
NA depth (1), the contribution of the longitudinal reinforcement ratio ( 2), the contribution
of the slenderness ratio (3), the contribution of the cross section ( 4), the contribution of
positive blast duration and natural period of structure ( 5), the contribution of dynamic
increase factors of steel and concrete ( 6) respectively. Table 5.1 shows the posterior statistics
of model parameters in estimating Probabilistic Deflection.
x lim
x
A t DIF steel
( )
0.601 ( 23.134 s 0.97
bD
H
r
2.646
D
b
( ) 1.462 d 1.987
Tn ( )
DIF concrete ( )
+ X e X ]
( )
X p =x me
Figure 5.1 shows the comparison between FE obtained results and predicted values of
Probabilistic deflection capacity model. The median value is plotted for the probabilistic
model.
Procedure for determining the Numerical deflection is explained in the section 4.4.4.
Figure 4.3 and figure 4.4 are the FE results of the above considered example. From those
plots the maximum numerical deflection is obtained as 44.50 mm
From the proposed model, the probabilistic deflection capacity is obtained as 39.39
mm.
On a summary,
From the above summary, the probabilistic deflection matched well with the
numerical deflection and this indicates that the proposed model predicted well. Also, figure
5.1 shows that the proposed model is capable of predicting the displacement value with a
greater accuracy as all the equality data lies within the one standard error ( X ) bin as
shown by the dotted lines. The lower bound data are above the 1:1 equality line.
Blast Analysis 56
For a RC beam under ultimate performance level, analytical model for deflection
developed by Newmark is used in this research (Brooks et al., 1953).
Xm td
XE
1+ ( )( 12 13 )
ru
is explained as the ratio of yield strength to average applied load, =
P
ru
X E=
KE
Xm td
XE
1+ ( )
6
The model correction terms are used to capture the physical phenomena that are not
accounted for in the mechanical model. The first explanatory function, h 1 accounts for the
contribution of neutral axis (NA) depth and limiting NA depth ( x x ) . The second
limit
explanatory function, h2 accounts for the contribution of the longitudinal reinforcement ratio
( bDA ) . The third explanatory function, h accounts for the contribution of the slenderness
s
3
Blast Analysis 57
ratio ( Hr ) ; where r is the radius of gyration. The fourth explanatory function, h 4 accounts
for the contribution of the cross section ( Db ) . The fifth explanatory function, h5 accounts
td
for the contribution of positive blast duration and natural period of structure ( )
TN
. The
sixth explanatory function, h6 account for the contribution of dynamic increase factors of
DI F steel
steel and concrete ( DI F concrete ) .
A step-wise deletion process is used to arrive at the posterior statistics for the capacity
model. The deletion process is stopped when the standard deviation of the model increases.
The following parameters correct for, the contribution of neutral axis (NA) depth and limiting
NA depth (1), the contribution of the longitudinal reinforcement ratio ( 2), the contribution
of the slenderness ratio (3), the contribution of the cross section ( 4), the contribution of
positive blast duration and natural period of structure ( 5), the contribution of dynamic
increase factors of steel and concrete ( 6) respectively. Table 5.2 shows the posterior statistics
of model parameters in estimating Probabilistic Deflection.
X =x e
p
[m
25.758 ( bDA )0.548( Hr )+1.624 ( Db )0.679 ( Tt )1.719( DIFDIF )+ e ]
s d
n
steel
concrete
X X
Figure 5.3 shows the comparison between FE obtained results and predicted values of
Probabilistic deflection capacity model. The median value is plotted for the probabilistic
model.
Procedure for determining the Numerical deflection is explained in the section 4.4.4.
Figure 4.6 and figure 4.7 are the FE results of the above considered example. From those
plots the maximum numerical deflection is obtained as 52 mm
Blast Analysis 60
From the proposed model, the probabilistic deflection capacity is obtained as 50.13
mm.
On a summary,
From the above summary, the probabilistic deflection matched well with the
numerical displacement and this indicates that the proposed model predicted well. Also,
figure 5.3 shows that the proposed model is capable of predicting the displacement value with
a greater accuracy as all the equality data lies within the one standard error ( X ) bin as
shown by the dotted lines. The lower bound data are above the 1:1 equality line.
CHAPTER 6. CONCLUDING
REMARKS
Interaction among the different components and inherent uncertainty associated with
the modeling, configuration and material properties of the concrete structure.
6.2. CONCLUSIONS
In this research, it is observed that predicting the actual response or realistic capacity
of structural elements is very important for the safety and economy concern. Therefore, this
research aimed to develop capacity model considering all the statistical uncertainties for
Blast Analysis 61
predicting the actual response. The developed probabilistic capacity model is realistic, easy to
use and also comforts in avoiding tedious FE simulations or experiments.
From this laid foundation, in future the work can be focused on improvising the
current proposed model for higher accuracies by increasing the number of data sets. In this
research, probabilistic capacity models are developed only for predicting the deflection, so it
is suggested to focus further in developing models for the internal strain energy, which will
be helpful for energy based design. As these models are valid only for prescribed boundary
conditions, it is suggested to expand this work further for other boundary conditions that are
in current practice. The models developed in this research account for ultimate limit state
performance level only, so developing the models for different performance levels such as
crack limit state, post blast scenario etc., is recommended. Thus a performance based capacity
model of RC members can be developed; which leads to further improvement of current blast
codal provisions.
REFERENCES
AASHTO LRFD Bridge Design Specifications 2007, Loading and General information.
Agrawal, A. K. and Yi, Z. (2008), High Precision Analysis of Blast Events on Highway
Bridges, University Transportation Research Center, CCNY, New York, NY 10031.
Altair Computing (2003). HyperMesh Ver. 6.0 Basic Tutorial, Altair Engineering Inc., 1820
E. Big Beaver, Troy, MI 48083.
Baker, W. E. (1973), Explosions in Air. University of Texas Press, Austin, Texas.
Baker, W. E., Cox, P. A., Westine, P. S., Kulesz J. J. and Strehlow, R. A. (1983), Explosion
Hazards and Evaluation. Elsevier, Amsterdam.
Bangash, M. Y. H., and Bangash, T. (2006), Explosion-resistant buildings: Design, Analysis
and Case Studies, Springer Berlin Heidelberg, New York.
Beshara, F. B. A. (1994), Modelling of Blast Loading on Aboveground Structures-I. General
Phenomology and External Blast, Computers and Structures, Vo. 51. No. 5, pp. 585-
596.
Brooks, N. B. and Newmark, N. M. (1953), The Response of Simple Structures to Dynamic
Loads, Technical Report, Office of Naval Research.
Biggs, J. M. (1964), Introduction to Structural Dynamics, McGraw-Hill, New York.Brode,
H. L. (1955), Numerical Solutions of Spherical Blast Waves, Journal of Applied
Physics, 26(6), 766 75.
Bischoff, P. and Perry, S. (1991), "Compressive Behaviour of Concrete at High Strain Rates."
Materials and Structures 24(6): 425-450.
Blast Analysis 62
Mario Paz (2004), Structural Dynamics Theory and Computation, 5 th edition, Kluwer
Academic Publishers, London.
Myers, R. H., & Montgomery, D. C. (1995). Response Surface Methodology. Process and
Product Optimization Using Designed Experiments, Wiley, New York, USA.
Ngo, T., Mendis, P., Gupta, A. and Ramsay, J. (2007), Blast Loading and Blast Effects on
Structures An Overview, EJSE Special Issue: Loading on Structures.
Newmark, N. M. (1963), Design of Structures for Dynamic Loads Including the Effects of
Vibration and Ground Shock. In Symposium on Scientific Problems of Protective
Construction, pp. 148-248. Swiss Federal Institute of Technology, Zurich.
Padgett, J. E., Nielson, B. G. and DesRoches, R. (2008), Selection of optimal intensity
measures in probabilistic seismic demand models of highway bridge portfolios,
Earthquake Engineering and Structural Dynamics, 37(5), pp: 711-725.
Sadovskiy, M. A. (2004), Mechanical Effects of Air Shock Waves from Explosions
according to Experiments (1952), in: Sadovskiy, M. A. selected works: Geophysics and
Physics of Explosion, Nauka Press, Moscow.
Sharma, H., Hurlebaus, S. and Gardoni, P. (2012). Performance-based Response Evaluation
of RC Columns Subjected to Vehicle Impact, International Journal of Impact
Engineering, 43, 52-62.
Slavik, T. (2009), A Coupling of Empirical Explosive Blast loads to ALE Air Domains in
LS-DYNA, 7th European LS-DYNA Conferences Salzburg Austria.
Tai, Y.S., Chu, T.L., Hu, H.T. and Wu, J.Y. (2011), Dynamic Response of a Reinforced
Concrete Slab Subjected to Air Blast Load, Theoretical and Applied Fracture
Mechanics.
TM 5-1300 (1990). "Structures to Resist the Effects of Accidental Explosions", Departments
of the Army, The Navy and The Air Force.
U.S. Army Corps of Engineers, 2008, Users guide for the Single-Degree-of-Freedom Blast
Effects Design Spreadsheets (SBEDS), Report PDC TR-06-02 Rev 1.Luccioni, B.M.,
Ambrosini, R.D. and Danesi, R.F. (2003), Analysis of Building Collapse under Blast
Loads, Engineering Structures, Volume 26, Issue 1.
UFC 3-340-02 (2008), Structures to Resist the Effects of Accidental Explosions,
Departments of The Army, The Navy and The Air force.
Winget, D. G., Marchand, K. A. and Williamson, E. B. (2005), Analysis and Design of
Critical Bridges Subjected to Blast Loads, Journal of Structural Engineering, ASCE.
Yaramada, V. K. R. (2010), Numerical Response of Steel Reinforced Concrete Slab
Subjected to Blast and Pressure Loadings in LSDYNA, M.S, University of Missouri
Kansas City.