Sei sulla pagina 1di 7

Structural Response of Offshore Blast Walls under

Accidental Explosion
Analysis of Offshore Blast Wall
Shaikh Atikur Rahma
Student ID-G02726
Department of Civil Engineering
Universiti Teknologi Petronas
Bander Seri Iskandar, Tronoh, Perak, Malaysia
atik.ce.uap@gmail.com
Abstract Adequate blast resistant barriers are requisite to
protect personnel and critical systems from the consequences of
an accidental explosion and subsequent fire. Many of the blast
walls currently installed in offshore structures were designed
using simplified calculation approaches like Single Degree of
Freedom models (SDOF) as recommended in many design
guidelines. Over simplified and idealized explosion load used for
response calculation and design of blast wall can lead to
inadequate or overdesign of offshore blast walls. Due to lack of
presence of a well-accepted design guidelines supported by
extensive study, the protection provided by the conventional blast
walls for offshore structures can be inadequate. In-depth
understanding of structural response of blast walls under
different blast loading can provide better design practice of blast
walls for adequate protection. In this study, structural responses
of conventional offshore blast walls were investigated. A
computation fluid dynamics (CFD) approach was used to predict
effect of different explosions on the barrier walls and non-linear
finite elements analyses were performed to study the behavior of
the blast-loaded walls under different explosions. Effect of
different parameters related to blast wall and accidental
explosions were investigated to gain detail understanding of
structural behavior of typical steel blast wall
Keywords Offshore blast wall, non-linear finite element
analysis, blast loading,computational fluid dynamics (CFD).

INTRODUCTION
Offshore platforms are usually exposed to potential
accidental explosions while drilling and conveying
combustible crude oil. In the event of an accidental explosion
on offshore structure, the blast wall contains blast and
prevents fire from spreading to cause further escalation. Blast
wall are normally non-load bearing members connected to
primary framework . A number of research and investigations
were carried out after the Piper Alpha oil rig (1988) accidental
explosion on different aspects related to the protection of
offshore structures under accidental explosion. But, in-depth
research on the structural response of conventional blast walls
under different explosion loading has not been performed.
Realistic modeling of blast wave propagation was often
ignored in the past studies. Simplified blast load on structural

Dr. Teo Wee


Department of Civil Engineering
Universiti Teknologi Petronas
Bander Seri Iskandar ,Tronoh,Perak ,Malaysia
teo_wee@petronas.com.my

analysis fails to capture the actual and realistic response of the


structures in analysis. This paper presents an overview of the
structural response to resist blast and explosion, focusing on
the use of non-linear finite element analysis (NLFEA). In this
research realistic explosion loads were simulated using
computational fluid mechanics technique and the blast
pressures obtained were coupled with NLFEA to achieve
better response calculation. This approach overcomes the
limitations of commonly used analytical methods which are
predominantly developed for elastic response or limited plastic
response and do not allow for large deflection and unstable
responses. In offshore structures high safety measure and
precaution need to be established not only to protect the
dormitory area of the members on board, but also to ensure
that whole structure does not collapse. Therefore, vulnerability
assessment of offshore blast wall, by investigating in-depth
structural response under realistic loading is highly desirable.

ACCIDENTAL EXPLOSION
A . Hydrocarbon Explosion
Hydrocarbon can be explored by ignition with the mixture
of an oxidizer (usually air).Thus when the temperature is to
the point at which molecules of hydrocarbons react
spontaneously with the oxidizer, combustion take place. This
hydrocarbon explosion causes a blast or a rapid increase in
pressure. There are various types of disaster in offshore
industries such as explosion, fire, ship collision, extreme wave
load and dropped object, etc. Especially, explosion in offshore
structure is extremely hazardous. This kind of accident causes
serious casualties, property losses and marine pollution.
Topside of offshore platforms are most likely to be expose to
such hazards as hydrocarbon explosion and a number of major
accidents involving them already have been reported.
Therefore, considerable interests in explosion and blast wall
have been increased on both industries and researcher since
the piper Alpha explosion (6 July 1988 Figure 1). Since the
piper Alpha accident took place, a substantial amount of effort
has been directed towards the management of explosion and

fire in offshore industries. In spite of these efforts, accidents


occurred continuously as evidence by the recent Deepwater
Horizon incident which occurred on 20 April 2010 in the Gulf
of Mexico as shown Figure 2.

bottom connections usually consist of two angles welded


together and to the structural steelwork of the platform.
Existing blast wall of an offshore platform has been in Figure
3.

Figure 1: The piper Alpha accidental explosion (1988).

C. Blast loading
The magnitude of the peak overpressure and the variation
of the incident overpressure with time depend on the type and
mass of the explosive, the location of the explosive centre to
the ground, and the observed point. From the process of
formation of an air-blast wave, it is evident that apart from
incident pressure, dynamic pressure that is caused by the
moving air stream behind the shock front is also formed. This
moving air stream, sometimes referred to as a blast wind,
induces a pressure on any obstacle. The variation of dynamic
pressure with time at a given point is presented in Figure 4.
Arrival time, ta, is defined as a finite time that the blast wave
requires to reach the observed point. This time depends on the
air density, the type and mass of explosive, and the distance
from the explosion point .

Figure 2: Deepwater Horizon accident (2010)

Figure 4: Variation of overpressure and dynamic pressure with time

BLAST SIMULATION

Figure 3: Offshore Blast wall under construction

B. Blast wall
Blast walls in offshore structure are often made of
stainless steel for its light-weight, corrosion preventing ability
and better performance in resisting fire. Typically stainlesssteel blast walls are made of about 12m wide and 4m high,
with corrugations running top to bottom. In modern days blast
walls are welded top and bottom and down both sides to the
primary steelwork through angle connections. The top and

A . Interaction of Air Blast Waves and Structures


A blast wave from an explosion propagates outwards in
space in all directions until it is distorted and interfered with by
obstacles along the way. The interaction of blast waves with
structures is a complex process that is caused by a combination
of reflection, diffraction, and refraction. When the blast wave
strikes the front of a surface, reflection occurs. This leads to an
increase in pressure, density and temperature of the air in the
infinitesimal area. As a result, the peak overpressure is
magnified and the resulting pressure is termed peak reflected
pressure. In addition, depending on the geometric properties of
the structure, diffraction may occur at its corners, which
reduces the effects of reflection on the surrounding areas.
Consequently, the pressure on the front face, the reflected
pressure, is much higher in magnitude whereas the side walls,
back walls and roof encounter a lower magnitude pressure
referred to as incident overpressure (Figure 5).
The drag loading component is caused by the dynamic
overpressure and the corresponding drag coefficients, and is
dependent on the characteristics of the fluid flow. The drag
pressure on a surface is the product of dynamic pressure and
the appropriate drag coefficient of the surface.

subsequent failure modes of a blast resisting structure. It is


always recommended to use actual stress strain curve material
data in the model if available, or otherwise an appropriate
idealized stress strain curve. Traditionally, for analysis
purposes steels have been assumed to possess behaviour of
rigid plastic, elastic perfectly plastic or linear hardening post
yield response up to ultimate stress. However, a more realistic
stress strain curve for stainless steel can be obtained by the
modified Ramberg and Osgood formulation given below
For
Figure 5: Blast wave interaction on structure

Blast simulation using CFD techniques


CFD was a traditional tool in the aerospace industry
focusing on the problems of supersonic and hypersonic flight,
but it has the same basic equations of motion and is applicable
to simulate blast modeling problems. However, only recently,
with the enhancement of CFD techniques, of numerical
methods and faster solver algorithms; and the tremendous
development of computer technology, CFD was applied into
blast simulation to investigate the problems of blast wave
propagation and its interaction with structures.

(1)

B.

MODELLING OF BLAST WALL


A . Finite Element Modeling
Blast walls were modeled by using finite element
package both LS-DYNA and ANSYS version 14.5. The
complexity of blast loading and the material non-linearity it
produced in the target structure necessitate the use of the finite
element techniques for realistic response calculation. NLFEA
also allows accurate modeling of the supports which is vital in
a blast response assessment. For efficiency and adequate
accuracy, it has been found that the shell elements are
appropriate for general assessment purpose for blast barriers.
Shell elements that are capable for both thick and thin shell
applications involving large displacements and finite
membrane strains (change of thickness) are particularly
suitable for wall type structure. Hence shell elements were used
to model typical blast wall. For improved modeling,
appropriate selection of mesh or element size is very important.
Mesh sensitivity analysis was performed to select the suitable
mesh size for the models. The blast wall properties suggested
by Schleyer (2007) were used for this study. The geometric and
material properties of the stainless steel blast wall used in this
project are presented in Table 2. Additionally mild steel was
used for modeling of blast wall to verify the model and
analysis results. Configuration and the meshing of the
conventional blast wall modeled in this study are presented in
Figure 7.
B. Material Modeling.
The material nonlinearity and strain rate effects have a
direct and significant influence on the response and

For
.. (2)
Where

and n, m =5, 2, 5 respectively.

C.

Cowper & Symonds strain rate


Typical strain rates encountered in the offshore industry
are in the range of 103 to10-2 s-1 . There are several
approaches that can be adopted to deal with strain rate in the
design or analysis of blast resistant structures. One of the most
commonly constitutive equations adopted by the offshore
industry is the CowperSymonds equation.
(3)
Where

and

are the dynamic and static yield stress,

respectively, is the strain rate and, D and q are the curvefitting material constants. Eq. (3) may be recast into a form of
dimensionless overstress power law:
. (4)
Eq. (4) has the form very similar to the models proposed by
Perzyna and Malvern . In this equation,
is the equivalent
plastic strain rate, is the dynamic flow stress and
is the
static flow stress which is a function of the equivalent plastic
strain. The material constants can be evaluated from a least
mean square fit to the experimental data by plotting the
logarithmic curve derived from Eq. (3) or (4). A series of test
carried out to determine the dynamic properties of stainless
steel has shown that experimental data tends towards a nonlinear curve in small strain rate regions as opposed to the
requirement of the straight line relation indicated in Eq. (5),.

(5)

Cowper Symonds stain rate p

10

The Coefficient for Cowper Symonds equation for different


standards is presented at Table 1. The numerical values of D
and q are already verified by series of experiments and widely
recognized for finite element modeling.
Table 1 Coefficient for the Cowper Symonds equation .

Material
Mild steel
Stainless steel
High tensile steel

D(1/s)
40.4
100
3200

q
5
10
5

Figure 6: Material property of Bilinear Isotropic Hardening

For the realistic simulation material properties of mild steel


and stainless steel was defined as bilinear isotropic hardening
and plastic kinematic hardening. Both models are suited to
model isotropic and kinematic hardening plasticity with the
option of including rate effect, those material modeling are
very cost effective and available for shell, solid and beam
elements. Only detail material properties of stainless steel
blast wall were presented in Table 2.
Table 2: Material and Geometric properties of Blast wall

Profile width (mm)

220.0mm

Profile depth (mm)

40.5mm
3

Profile material density (kg/m )

7750 kg/m3

Blast wall height (m)

4m.

Blast wall width (m)

3m

Youngs Modulus (pa)

1.930x1011 pa

Poissons ratio

0.3100

Yield stress (pa)

2.1x108 pa

Tangent modulus (pa)

1.8x109 pa

Cowper Symonds stain rate c

100

Figure 7: Finite element modeling of blast wall

NUMERICAL RESULTS AND DISCUSSION


Numerical results for offshore blast wall are presented in this
section, together with blast simulation using CFD approach.
Modeled blast walls were subjected to different blast loading
to obtain the structural responses for those loading conditions.
Blast pressures were obtained from different equivalent TNT
explosions to represent effect of different amount of
explosions on offshore structures. According to previous
investigations on Piper Alpha platform by Pask et al. (2000)
and Schleyer (2007), it has found that typical blast pressure for
hydrocarbon explosions are in the range of 250 kPa to 400
kPa. From the blast simulation performed, it was observed that
1kg and 2kg TNT equivalent weight at a scale distance of 2m
produce blast pressure profile which can represent the typical
hydrocarbon explosion pressure range. To identify extensive
response of the blast wall, more detail studies were perform
by using various amount of TNT charge on different location.
However both Stainless steel and Mild steel were used for
finite element analysis of blast wall. Hence in the finite
element modeling of Mild steel blast wall were studied only
using the pressure obtained from 1 and 2 kg TNT. The peak
pressure (Reflected on incident) value obtained for 1kg and
2kg TNT were 238 kPa and 431 kPa as shown in Figure 8.
Much more localized failure was identified by evaluating the

each element deformation for 238 kPa and 431kPa blast

loading. Maximum element deformations for those loading are


presented in Figure 10.

Figure 8: pressure vs. time history of TNT Explosive

Figure 10 Total deformation for different loading

Figure 9: pressure vs. time history of TNT Explosive

Figure 11: Stress vs. Strain diagram for different loading

Figure 12: Maximum Element displacement for 320 kPa blast pressure

One of the comparisons of element displacement in various


location of blast wall under 320 kPa blast pressures has
presented in . Elements were chosen from the first point of
blast hit on wall and nearby area of it. To identify a detail
scenario of blast response on structure a parametric study was
performed for various pressure curves loading. It has been

summarized on . Pressure curve history is represented by its


peak pressure and maximum numerical values of evaluated
results are only shown on the table. Element displacement for
particular point, where the first blast wave was impacted is
higher in comparison with total body displacement of the wall.
From these analysis results it can be stated that localized

failures are more vulnerable than others. Finally total rigid

body displacement and stress vs. strain diagram are displayed


at & respectively.

Table 3: Finite element results for different blast pressure

Blast pressure (peak)


Principal stress
(mPa)
(V-M) stress (mPa)
-1

Plastic strain (s )
Deformation(mm)

320 kPa 290kPa 281kPa 250kPa

223kPa

407

402

395

390

385

368

367

366

365

361

0.034

0.027

0.026

0.025

0.024

150

141

135

125

120

This non-linear approach in understanding the behaviour and


subsequent design of blast wall is expected to improve the
performance of future offshore blast walls.

ACKNOWLEDGEMENT
The authors would like to express their highest gratitude to
Universiti Teknologi PETRONAS (UTP), the Offshore
Engineering Centre of UTP and also PETRONAS for their
utmost cooperation in this research.
[1]

Figure 13: Rigid body displacement for 320 kPa blast pressure

[2]

[3]
[4]

[5]

[6]

[7]
Figure 14: Stress vs. plastic strain diagram for stainless steel .

CONCLUSION
This paper presents preliminary results on the structural
response analysis of typical blast walls under different blast
explosions. Initial investigation shows the possible response of
offshore blast walls under highly impulsive loading. More
detailed study, which is currently undergoing, is required to
have a comprehensive understanding of the response of blast
walls exposed to different explosions.

[8]

[9]

J. M. Sohn, S. J. Kim, B. H. Kim, and J. K. Paik, "Nonlinear


structural consequence analysis of FPSO topside blastwalls,"
Ocean Engineering, vol. 60, pp. 149-162, 3/1/ 2013.
L. A. Louca, M. Punjani, and J. E. Harding, "Non-linear analysis of
blast walls and stiffened panels subjected to hydrocarbon
explosions," Journal of Constructional Steel Research, vol. 37, pp.
93-113, 4// 1996.
L. A. Louca, J. W. Boh, and Y. S. Choo, "Design and analysis of
stainless steel profiled blast barriers," Journal of Constructional
Steel Research, vol. 60, pp. 1699-1723, 12// 2004.
A. S. Fallah and L. A. Louca, "Pressureimpulse diagrams for
elastic-plastic-hardening and softening single-degree-of-freedom
models subjected to blast loading," International Journal of
Impact Engineering, vol. 34, pp. 823-842, 4// 2007.
G. K. Schleyer, M. J. Lowak, M. A. Polcyn, and G. S. Langdon,
"Experimental investigation of blast wall panels under shock
pressure loading," International Journal of Impact Engineering,
vol. 34, pp. 1095-1118, 6// 2007.
G. S. Langdon and G. K. Schleyer, "Inelastic deformation and
failure of profiled stainless steel blast wall panels. Part I:
experimental investigations," International Journal of Impact
Engineering, vol. 31, pp. 341-369, 4// 2005.
G. S. Langdon and G. K. Schleyer, "Inelastic deformation and
failure of profiled stainless steel blast wall panels. Part II:
analytical modelling considerations," International Journal of
Impact Engineering, vol. 31, pp. 371-399, 4// 2005.
G. S. Langdon and G. K. Schleyer, "Deformation and failure of
profiled stainless steel blast wall panels. Part III: finite element
simulations and overall summary," International Journal of Impact
Engineering, vol. 32, pp. 988-1012, 6// 2006.
E. Nwankwo, A. Soleiman Fallah, G. S. Langdon, and L. A. Louca,
"Inelastic deformation and failure of partially strengthened profiled
blast walls," Engineering Structures, vol. 46, pp. 671-686, 1//
2013.

Potrebbero piacerti anche