Sunteți pe pagina 1din 15

Experimental and Numerical Simulation

Activities for the Assessment of Explosion


Effects in a Train Station
George Solomos, Folco Casadei, Georgios Giannopoulos
Joint Research Centre of the European Commission
Martin Larcher
Universität der Bundeswehr München, Germany

Abstract. Recent terrorist attacks in biggest cities of the world revealed the vulnerability of transport network as a whole and rail transport
in particular. A particular role in improving the safety situation may play architecture alleviating the consequences of such attacks.
The behaviour of structural materials and components at higher strain-rates, and experimental techniques and facilities for relevant
studies has been considered. Numerical simulation tools, suitable for fast dynamic phenomena, have also in parallel been developed.
Using these tools the authors attempted to assess the physical vulnerabilities of rail transport infrastructures to explosion loads, mainly
due to terrorist bombing attacks. The source term, the air medium and the structure are properly modeled using the computational
framework of the explicit finite element code Europlexus. Particular attention has been focused on modeling of glass failure, which
may be the cause of extended injuries. A case study of a rail station has been conducted, whose geometry has been reconstructed using
a laser scanning technique. For several explosion scenarios the structural response is reliably determined, and through appropriate probit
functions, the associated risk of human injuries has also been calculated. The potential of the technique to serve as a scene reconstruction
tool for forensic purposes is outlined.

Keywords: blast loading, counter terrorism, explosions, forensic science, security by design

Introduction
The recent terrorist attacks in Madrid, London, and Moscow have shown that rail
transport constitutes a high-impact target. They have also exposed the vulnerabilities
of the system, which are mainly due to its open architecture and its widely dispersed
assets. While intelligence and  other security measures will always play the  major
role in preventing and foiling such attacks, improved architectural design may also
significantly contribute towards mitigating the effects of explosions. Thus reliable
simulation tools for assessing structural vulnerabilities are indispensable. To this end
both the proper numerical modelling and the knowledge of the structural materials
behaviour under such loading conditions are important.
Concerning human casualties, the number of deaths and injuries of the 2004 ter-
rorist attack in Madrid is reported in one of the issues of Critical Care.1 A comparison
of the injuries from detonations in confined structures and in open‐air is presented
in J Trauma Inj Infect Crit Care,2 where it shown that the mortality rate in confined spaces
1
Peral-Gutierrez de Ceballos J., Turégano-Fuentes F., Pérez-Díaz D., Sanz-Sánchez M.,
Martín-Llorente C., Guerrero-Sanz J.E., 11 March 2004: The terrorist bomb explosions in Madrid,
Spain — an  analysis of  the  logistics, injuries sustained and  clinical management of  casualties
treated at the closest hospital. Critical Care 2004, Vol. 9(1), pp. 104–111.
2
Leibovici D., Gofrit O.N., Stein M., Shapira S.C., Noga Y., Heruti R.J., Shemer J., Blast injuries:
bus versus open-air bombings — a comparative study of injuries in survivors of open-air versus
confined-space explosions. J Trauma Inj Infect Crit Care 1996, Vol. 41(6), pp. 1030–1035.

Internal Security, January–June 2011 49


Copyright © 2011 Police Academy in Szczytno. All Rights Reserved
George Solomos, Martin Larcher, Folco Casadei, Georgios Giannopoulos

is higher. The reason is that the pressure of an air blast wave decays much slower inside
a long tube-like train or metro tunnel than in open spaces. This implies that both
the response of the air medium inside the structure and the failure of the structure
itself have to be considered for the determination of the risk in case of a blast.
Experimental data for explosions between scaled-down buildings are provided
in Int J Impact Eng,3 and it has been shown that shielding and channelling phenomena
have to be considered in city streets.4 Some researchers5 have investigated the damage
of complex buildings using a pressure‐time function to describe the air blast wave. More
sophisticated Computational Fluid Dynamics (CFD) codes have also been employed for
the determination of the behaviour of air inside complex configurations, like large city
centres.6 Similar are also the methodologies in other works devoted to the subject,7
where the interaction of air blast waves with buildings in combination with a damage
criterion is shown. For a different urban setting, blast phenomena in underground road
tunnels with roof openings as venting areas are investigated in J Loss Prev Process Ind.8
Concerning studies for assessing in  particular the  security in  railway systems,
in a review paper Loukaitou‐Sideris et al.,9 compare the measures taken by four cities
to reduce the risk in their metro lines and stations. As reported, increased surveil-
lance and patrolling are mainly pursued, while strategies, such as the Crime Prevention
Through Environmental Design (CPTED), start gaining acceptance and are incorporated
into new station planning and design. Finally, several explosion scenarios and their
effects in railway stations and carriages have been analysed and presented by authors
in J Hazard Mater10 and J Transp Secur.11
Along the same lines, the blast analysis inside partially confined, large spaces,
like those of a train station, is attempted in this work. Properties of materials under

3
Smith P.D., Mays G.C., Rose T.A., Teo K.G., Roberts B.J., Small scale models of  complex
geometry for blast overpressure assessment. Int J Impact Eng 1992, Vol. 12(3), pp. 345–360.
4
Smith P.D., Rose T.A., Blast wave propagation in city streets — an overview. Prog Struct Eng
Mater 2006, Vol. 8(1), pp. 16–28.
5
Bogosian D.D., Dunn B.W., Chrostowski J.D., Blast analysis of  complex structures using
physics-based fast-running models. Comput Struct 1999, Vol. 72(1), pp. 81–92.
6
Remennikov A.M., Rose T.A., Modelling blast loads on buildings in complex city geometries.
Comput Struct 2005, Vol. 83(27), pp. 2197–2205.
7
Langdon G., Schleyer G., Inelastic deformation and failure of profiled stainless steel blast
wall panels. Part II: analytical modelling considerations. Int J Impact Eng 2005, Vol. 31, pp. 371–399;
Löhner  R., Cebral J., Yang C., Baum J.D., Mestreau E., Charman C., Pelessone D., Large-scale
fluid-structure interaction simulations. Comput Sci Eng 2004, Vol. 6(3), pp. 27–37; Luccioni B.M,
Ambrosini D., Danesi  R. Blast load assessment using hydrocodes. Eng Struct 2006, Vol. 28(12),
pp. 1736–1744; Luccioni, B.M., Ambrosini, R.D., Danesi, R.F., Analysis of building collapse under
blast loads. Eng Struct 2004, Vol. 26, pp. 63–71.
8
Van den Berg A.C., Weerheijm J., Blast phenomena in  urban tunnel systems, J Loss Prev
Process Ind 2006, Vol. 19, pp. 598–603.
9
Loukaitou-Sideris A., Taylor B.D., Fink C.N.Y., Rail transit security in an international context:
lessons from four cities. Urban Aff Rev 2006, Vol. 41(6), pp. 727–748.
10
Giannopoulos G., Larcher M., Casadei F., Solomos G., Risk assessment of the fatality due
to explosion in land mass transport infrastructure by fast transient dynamic analysis, J Hazard
Mater 2010, Vol. 173, pp. 401–408.
11
Larcher M, Casadei F., Solomos G., Risk analysis of explosions in trains by fluid–structure
calculations. J Transp Secur 2010, Vol. 3, pp. 57–71.

50 Internal Security, January–June 2011

Copyright © 2011 Police Academy in Szczytno. All Rights Reserved


Experimental and Numerical Simulation Activities...

fast-transient dynamic conditions and proper modelling constitute the basis of this


investigation. The finite element code Europlexus is used, which has the feature of effi-
ciently treating fast fluid-structure interaction phenomena. Thus estimates of the risk
to both the structures and to humans are calculated.

Explosion Loads
Air blast waves are considered to result from the detonation of a solid high-explosive
charge, which is expressed as a TNT equivalent.12 The magnitude of the pressure of an air
blast wave that arrives at a certain point depends on the distance and on the size
of the charge. An idealised form of a pressure-time function at a certain distance from
the explosive is shown in Figure 1.

Figure 1. Idealised form of a pressure-time function in open-air.

Shown in this figure are the main characteristics of a free-field air blast: the arrival
time ta (it includes the detonation time itself), the peak overpressure pmax over the ref-
erence pressure p 0 (atmospheric pressure), the  duration of  the  positive phase td,
the minimum pressure, and the duration of the negative phase tn. The positive impulse
I is the integral of the overpressure curve over the positive phase, where the pressure p
at time t is usually described by the modified Friedlander equation:13

Equation 1.

12
Baker W.E., Cox P.A., Westine P.S., Kulesz J.J., Strehlow R.A., Explosion Hazards and Evalua-
tion. Amsterdam: Elsevier 1983.
13
Baker W.E., Explosions in Air, Austin TX: University of Texas Press 1973.

Internal Security, January–June 2011 51


Copyright © 2011 Police Academy in Szczytno. All Rights Reserved
George Solomos, Martin Larcher, Folco Casadei, Georgios Giannopoulos

The constant b influences the decay of the overpressure curve, and all parameters


of Eq. 1 can be readily found in appropriate diagrams and equations,14 for free field
conditions and spherical and hemi-spherical charge shapes.
The form of the pressure wave of Figure 1 can be greatly altered by the morphology
of the medium encountered along its propagation. For instance, incident peak pressure
can be substantially magnified if the wave is reflected by a rigid obstacle, while these
reflection effects depend on the geometry, the intensity of the wave and its angle
of incidence. The situation is much more complicated if there are several reflection
boundaries, as occurs between or inside buildings. Obviously these arguments make
clear that in confined environments simplified models that relate the peak overpressure
to the distance from the explosion can lead to grossly erroneous estimates.
There are available various ways of modelling explosions and blast loading, which,
in turn, define also the fineness of the finite element discretization. Three of them have
been considered in this study.
— The solid TNT model. It describes the mechanical behaviour of the explosive with
a material law, e.g. the Jones-Wilkins-Lee (JWL) equation. A fine mesh is essential
to obtain realistic results. The calculation is therefore very expensive in terms
of computation time.
— The  bursting balloon model.15 The  pressure-time function resulting from
the release of a compressed balloon can match the air blast history. The amount
of initial compression can be calibrated with the impulse. The computational
time is smaller compared to that of the solid TNT model.
— Load-time function. Only the structure is modelled and loaded by an appro-
priately scaled pressure-time function, as presented previously in Figure 1.
The  calculation is  relatively inexpensive. Clearly, since the  air is  neglected,
the method cannot represent reflections, shadowing and channelling.
In the present work the bursting balloon technique has been preferably used.
The advantages of this approach lie in the fact that for big structures and spaces
(e.g. a rail station) larger dimension finite elements can be used, and thus the computa-
tion time becomes reasonable, while at the same time both the structure and the fluid
are modelled.

Materials Data
Under the dynamic conditions pertaining to blast or impact, materials may exhibit
a different behaviour than that under static conditions.16 This behaviour under higher
strain-rates must be investigated and be known in order to formulate the materi-
als’ constitutive equations at that loading regimes. These material models will be

14
Ibid.; Kingery C.N., Bulmash G., Airblast parameters from TNT spherical air burst and hemi-
spherical surface burst, Tech. rep., Defence Technical Information Center, Ballistic Research Labo-
ratory, Aberdeen Proving Ground Maryland 1984; Hyde D.W., CONWEP: Conventional Weapons
Effects Program, US Army Waterways Experimental Station, Vicksburg 1991; US Army TM 5–855–1:
Fundamentals of  Protective Design for Conventional Weapons, Headquarters Department
of the Army, Washington, D.C. 1986.
15
Larcher M., op. cit.
16
Krauthammer T., Modern Protective Structures, CRC Press 2008.

52 Internal Security, January–June 2011

Copyright © 2011 Police Academy in Szczytno. All Rights Reserved


Experimental and Numerical Simulation Activities...

subsequently inserted in the numerical simulation codes in order to produce realistic


response estimates.
This type of experimental studies have been extensively conducted at the Joint
Research Centre using Hopkinson bar techniques. An example of a relevant stress-strain
diagram, concerning concrete in compression,17 is shown in Figure 2. As is evident,
the dynamic strength of concrete is substantially higher than its static one (more than
twice), this being coined as the “dynamic increase factor”, DIF.

Figure 2. Stress-strain curves of concrete under compression for several strain-rates


(10 –6 /s static up to 20/s dynamic).

Another aspect of crucial importance in the case of a terrorist bombing attack


is the behaviour of non-structural elements, like glass panes. Glass may fracture to small
pieces, which, propelled by the blast wave, act as sharp shards causing extensive
injuries. For this reason laminated glasses (which limit these phenomena through
the action of the PVB interlayer) are used in safety and security applications. Their
dynamic behaviour has been considered herein, and it has been attempted to simulate
their response. Two examples of detailed analyses of laminated glass subjected to blast
loading are shown in Figure 3. The experimental result comes from the shock-tube
testing18 of a laminated panel 7.5mm thick (3mm floatglas/ 1.5mm PVB/ 3mm floatglass),
of exposed area of 1x0.8=0.8m2. It is seen that the actual crack pattern is adequately
reproduced in the simulations, where a layered-shell model and a 3D-solid model have
been employed, respectively.19

17
Kranzer C., Gürke G., Mayrhofer C., Testing of bomb resistant glazing systems. Experimental
investigation of  the  time dependent deflection of  blast loaded 7.5mm laminated glass. Glass
Processing Days 2005, Electronic source: http://www.gpd.fi, accessed: 2005.
18
Ibid.
19
Larcher M., Simulation of laminated glass loaded by air blast waves. DYMAT 2009. Vol. 1.
Les Ulis (France): EDP SCIENCES S A, pp. 1553–1559.

Internal Security, January–June 2011 53


Copyright © 2011 Police Academy in Szczytno. All Rights Reserved
George Solomos, Martin Larcher, Folco Casadei, Georgios Giannopoulos

(a) (b) (c)


Figure 3. Crack pattern of  laminated glass: (a) experiment, (b) simulation with
layered-shell model and mesh 12.5mm, (c) simulation with 3D-solid model
and mesh size 2.5mm.

Numerical Simulations

Numerical simulations are performed within Europlexus,20 an explicit finite element


code for non-linear dynamic analysis. This finite element tool has been jointly devel-
oped by the French Commissariat a l’Energie Atomique (CEA) and the Joint Research
Centre (JRC). Among the main advantages of Europlexus over similar software is its
ability to handle complex fluid-structure interaction problems, which include the phe-
nomena encountered when a blast wave impinges onto a deformable structure.

Geometrical modelling and discretization


A relatively old typical train station has been selected for evaluation. The geometry
of its structures has been acquired using a 3D laser scanning technique. A laser scanner
was placed at various positions inside and outside these buildings and the whole geom-
etry was stored as a cloud of points. These points were next transformed to continuous
geometrical primitives, like surfaces and volumes, using the JRC-RECONSTRUCTOR [25],
an in-house developed tool, and a final elaboration of these data yielded the finite
element model for the structure, depicted in Figure 4. It is composed of two princi-
pal domains which are connected through a short, narrow passage. The dimensions
of the waiting hall area are about (LxWxH) 50×30×12 m (some advertisement posters
are hanging from the roof) and those of the long corridor about 100×10×8 m.
As mentioned above, for representing a certain TNT quantity the bursting balloon
model has mainly been used. The air inside this balloon behaves as a perfect gas,
and the diameter and internal energy (or initial pressure) of the balloon are properly
chosen so as to generate a far-field blast impulse equivalent to that produced by
the TNT charge under consideration. The full simulation of the explosion is performed
using an Eulerian formulation for the explosive and for the fluid representing the air.
Apart from the nature of the problem, this choice is justified by the fact that the sub-
sequent risk analysis requires the calculation of pressure and impulse of the air inside
the volume of the structure.

20
Commissariat a l’´Energie Atomique (CEA), Joint Research Centre (JRC) 2010. EUROPLEXUS
User’s manual. Electronic source: http://europlexus.jrc.ec.europa.eu/.

54 Internal Security, January–June 2011

Copyright © 2011 Police Academy in Szczytno. All Rights Reserved


Experimental and Numerical Simulation Activities...

Figure 4. Geometrical finite element model of  structure consisting of  a  main hall
and a long corridor, connected by a narrower passage.

A variety of material models exist in Europlexus for large strain analysis of metals up
to failure. The structural elements of the station (columns, beams and roof trusses) are
modeled as metallic (steel) elements with rather ductile characteristics. The material
model used is based on isotropic hardening formulation in order to describe the elas-
toplastic behaviour, however failure is also added. Strain rate effects are not considered,
and the floor and some walls are taken as rigid.
A brittle material model and element erosion techniques are used for the glass tiles
of the roof. The laminated glass used in the metro station is modelled by using layered
elements with a special failure criterion. After the failure of the glass, the stresses are set
to zero if the strains are positive (traction), but the material can still react to compres-
sion. The failure behaviour of the interlayer of the laminated glass requires a finer
meshing. Since a rather coarse element mesh is implemented (due to the large dimen-
sions of the complete numerical model), a displacement criterion is used instead for
the failure of the interlayer.

Scenario cases
Several scenarios with different quantities of explosive have been run, whereas
the explosive has always been placed on the floor in a booth at the center of the main
hall. Representative results of three charge levels (16kg, 100kg and 275kg) are shown.

Figure 5. Structural response of main hall at three time instants (1ms, 25ms and 50ms)
after detonation of the 275kg-charge.

The case of the explosive charge of approximately 275 kg TNT equivalent is pre-


sented in Figure 5. The results are based on the more sophisticated bursting balloon
approach, where both the fluid and the structure are included. It is observed that

Internal Security, January–June 2011 55


Copyright © 2011 Police Academy in Szczytno. All Rights Reserved
George Solomos, Martin Larcher, Folco Casadei, Georgios Giannopoulos

the methodology manages to reproduce the salient characteristics of the evolution


of  the  structural behaviour, such as  large deformations, failure of  the  roof, frag-
mentation, projectile formation and motion. The deformation and damage pattern
of the structure at the selected three time instants is as intuitively expected, and this
is also true for the whole response period. The same scenario has been carried out
using the pressure-time function approach (where only the structure is considered),
and the results were almost identical. Thus it can be concluded that for the structures
themselves exposed to a direct air blast, the pressure-time functions can provide a good
and inexpensive way for calculating their behaviour. However, this approach is not
sufficient if the blast waves arrive indirectly to the structure (reflections, channeling),
or if information about the air-pressure field is required.
Figure 6 depicts the “final” damage state corresponding to each of the above three
charges; more precisely, it is the situation 250 milliseconds after detonation. It is noted
that the 16kg-charge causes moderate structural damage; some columns are overly
deformed and  the  light-weight central booth is  disintegrated. The  100kg-charge
causes extensive structural damage, while the 275kg-charge causes total destruction.
The fact that the floor and some side-walls appear to remain intact is only because
(for simplicity) they have been modelled as absolutely rigid. It is also worth noticing
that the analysis can follow the trajectories of the fragmented debris and determine
their eventual landing position. These distinct differences in the final damage pattern
is the type of information that can be incorporated as scientific evidence for the recon-
struction of the event, especially as regards the place and quantity of the explosive.
With appropriate refinements and calibration, use of this methodology for forensic
investigations can thus be envisaged.

Figure 6. Structural damage after 250 ms from detonation for three explosive charge
quantities.

For the computational effort, as described above, in the comprehensive simula-


tion with the bursting balloon approach an Eulerian formulation has been used for
the fluid mesh. The mesh is first conforming to the external envelope mesh on which
the appropriate boundary conditions are set. Areas of the station that are commu-
nicating with the external environment are modeled using an absorbing boundary
condition. The complete model consists of about 4000000 fluid brick elements, 175000
absorbing boundary elements and 55000 elements for the structure.
Appropriate modelling of fluid-structure interaction (FSI) phenomena has been
a central issue in the current numerical simulations, as Europlexus contains quite power-
ful automatic FSI algorithms. However, the simulation of terrorist attacks up to possible
complete failure and fragmentation of some structural components introduces a new

56 Internal Security, January–June 2011

Copyright © 2011 Police Academy in Szczytno. All Rights Reserved


Experimental and Numerical Simulation Activities...

issue, for which a novel dedicated FSI model has been developed.21 The fluid and struc-
tural sub-domains are topologically uncoupled (independent). Each sub-domain
is discretized separately and the two meshes are simply superposed. At each time
instant of the computation, a topological search is performed (by suitable optimized
algorithms) of the fluid nodes which are reasonably close to the structure, and appropri-
ate FSI coupling conditions are imposed.

Human risk analysis


The formulation of the risk of human injuries is based on existing work, described
in,22 and mainly referring to explosion accidents in chemical process plants. It utilizes
the peak overpressure pmax (Pa) and the positive impulse I (Pa·s) calculated inside
each fluid finite element in order to determine the probability of eardrum rupture
and the probability of death.
Three different causes of death are considered through the following probit func-
tions:

Equation 2.

Equation 3.

Equation 4.

Y1 is the death probit function due to head impact, Y2 is the one for whole body
impact and Y3 is the one for lung haemorrhage. A probit function for body impact by
flying debris has not yet been implemented. The probit function of eardrum rupture Y4
is described through the equation:

Equation 5.

The probability of occurrence R (or the percentage of the affected population)


of the corresponding injury type is next determined for each of the above probit
21
Giannopoulos G., et all., op. cit.; Casadei F., Fast Transient Fluid-Structure Interaction with
Failure and Fragmentation, WCCM8 Congress, Venice (Italy), June 30–July 5, 2008.
22
González Ferradás E., Díaz Alonso F., Doval Miñarro M., Miñana Aznar A., Ruiz Gimeno J.,
Sánchez Pérez J., Consequence analysis by means of characteristic curves to determine the dam-
age to  humans from bursting spherical vessels. Process Safety Environ Protect 2008, Vol. 86,
pp. 121–129; Yet-Pole I., Cheng Te-Lung, The Development of a 3D Risk Analysis Method, J Hazard
Mater 2008, Vol. 153(1–2), pp. 600–608; Mannan S., Lee’s loss prevention in the process industries;
Volume 2: Hazard Identification, Assessment and Control. Amsterdam: Elsevier 2005.

Internal Security, January–June 2011 57


Copyright © 2011 Police Academy in Szczytno. All Rights Reserved
George Solomos, Martin Larcher, Folco Casadei, Georgios Giannopoulos

functions using the underlying specific cumulative normal distribution function. Plau-


sible influence and interaction of the different causes of death is treated in a simplified
manner by considering the maximum of the relevant three probabilities.
This analysis performed for the above-mentioned scenarios reveals the areas for
which the human injury risk is high. The results obtained at the end of the simulation
time (250ms) are shown for the three-dimensional space in Figure 7 for the death risk,
and for the charges 16kg and 275kg, respectively. From these figures it is evident that
for a large part of the main hall the death risk is very high; in particular for the charge
of 275kg it is 100% everywhere at floor level, where the passengers are found. However,
the abrupt change of geometry between the main hall of the station and the corridor
appears to play an important role in reducing the devastating effects of the explosion.
At this section of the structure the death risk is nil for the 16kg charge and it is substan-
tially reduced for the 275kg charge, thus providing a relatively safe area for the occupants
of the station under this explosion scenario. The eardrum rupture risk (not shown here)
is high for both charges, and remains high even along a large part of the corridor. For
the main hall it is also worth noticing spots at corners, where enhanced values of death
risk estimates are encountered due to wave reflection phenomena. For the same reason
the region on the roof appears deceivingly to be of higher risk for the 16kg charge than
for 275kg one. As shown in Figure 6, in the first case the roof does not fail and thus blast
pressure accumulates there, whereas in the latter, with the incipient disintegration
of the roof, the overpressure there is effectively released. As previously outlined, this
type of information on potential human injuries and casualties can also be important
for complementing forensic investigations, where however judicious interpretation
of the results is essential.

(a) (b)
Figure 7. Top view of final death risk contours for the main hall and corridor for (a) 16 kg
charge and (b) 275 kg charge; red colour corresponds to 100% probability.

Conclusions

A methodology for assessing the vulnerability of a train station to explosion effects


has been presented. After reviewing the behaviour of materials at higher strain-rates,
a finite element modelling of a large structure representing a typical train station has
been elaborated and the numerical simulation of a terrorist bombing attack scenario
has been conducted. Even though several simplifying assumptions have been made

58 Internal Security, January–June 2011

Copyright © 2011 Police Academy in Szczytno. All Rights Reserved


Experimental and Numerical Simulation Activities...

with regards to  material and  structural properties, the  dominant characteristics
of the response clearly emerge.
The Europlexus code used has proven to be an efficient and robust computational
platform. Several modelling approaches have been presented and  implemented
in the code. It has been found that the bursting balloon model demonstrates definite
advantages when explosion effects in such large spaces are simulated. At the same
time it allows a comprehensive study to be carried out by including both the fluid
and the structure.
Among the additional simulation features introduced, is the human injuries risk
analysis module. It is thought that such risk analysis capabilities can be a useful tool for
decision makers, transport operators and stakeholders etc. in order for them to assess
the risks of certain events and perform cost-benefit analyses of contemplated interven-
tions Finally these simulation techniques can be of aid in forensic investigations towards
the reconstruction of the most probable scenario of a bomb explosion attack or acci-
dent, based on the spatial extent and distribution of the final damage and injuries.

Theses of the article presented above were set out during the European Materials Research
Society (E-MRS) 2010 Fall Meeting which was held in Warsaw on September 13–17. Founded
in  1983, the  E-MRS has more than 3,000 members from industry, government, academic
and research laboratories, who meet regularly to debate about recent technological develop-
ments of functional materials. Presentation of the problems described in the article was a part
of the Symposium on New Materials and Homeland Security, one of the 15 symposia which were
included to the E-MRS 2010 Fall Meeting.

References:

1. Baker W.E., Cox P.A., Westine P.S., Kulesz J.J., Strehlow R.A., Explosion Hazards
and Evaluation. Amsterdam: Elsevier 1983.
2. Baker W.E., Explosions in Air, Austin TX: University of Texas Press 1973.
3. Bogosian D.D., Dunn B.W., Chrostowski J.D., Blast analysis of complex structures
using physics-based fast-running models. Comput Struct 1999, Vol. 72(1).
4. Boström G., Fiocco M., Gonçalves J., Puig Alcoriza D., Sequeira V., Chartier B., Kiesser
R., Mariotte F., Richard M., Zamora P., 3D reconstruction in nuclear security. ESARDA
Bull 2008, Vol. 38.
5. Casadei F., Fast Transient Fluid-Structure Interaction with Failure and Fragmentation,
WCCM8 Congress, Venice (Italy), June 30–July 5 2008.
6. Commissariat a l’´Energie Atomique (CEA), Joint Research Centre (JRC) 2010. EURO-
PLEXUS User’s manual. Electronic source: http://europlexus.jrc.ec.europa.eu/.
7. Giannopoulos G., Larcher M., Casadei F., Solomos G., Risk assessment of the fatality
due to explosion in land mass transport infrastructure by fast transient dynamic
analysis, J Hazard Mater 2010, Vol. 173.
8. González Ferradás E., Díaz Alonso F., Doval Miñarro M., Miñana Aznar A., Ruiz
Gimeno J., Sánchez Pérez J., Consequence analysis by means of characteristic curves
to determine the damage to humans from bursting spherical vessels. Process Safety
Environ Protect 2008, Vol. 86.

Internal Security, January–June 2011 59


Copyright © 2011 Police Academy in Szczytno. All Rights Reserved
George Solomos, Martin Larcher, Folco Casadei, Georgios Giannopoulos

9. Hyde D.W., CONWEP: Conventional Weapons Effects Program, US Army Waterways


Experimental Station. Vicksburg 1991.
10. Kingery C.N., Bulmash G., Airblast parameters from TNT spherical air burst and hemi-
spherical surface burst, Tech. rep., Defence Technical Information Center, Ballistic
Research Laboratory, Aberdeen Proving Ground Maryland 1984.
11. Kranzer C., Gürke G., Mayrhofer C., Testing of bomb resistant glazing systems.
Experimental investigation of the time dependent deflection of blast loaded 7.5mm
laminated glass. Glass Processing Days 2005. Electronic source: http://www.gpd.fi,
accessed: 2005.
12. Krauthammer T., Modern Protective Structures. CRC Press 2008.
13. Langdon G., Schleyer G., Inelastic deformation and failure of profiled stainless steel
blast wall panels. Part II: analytical modelling considerations. Int J Impact Eng 2005,
Vol. 31.
14. Larcher M, Casadei F., Solomos G., Risk analysis of explosions in trains by fluid–struc-
ture calculations. J Transp Secur 2010, Vol. 3.
15. Larcher M., Simulation of laminated glass loaded by air blast waves. DYMAT 2009.
Vol. 1. Les Ulis (France): EDP SCIENCES S A.
16. Leibovici D., Gofrit O.N., Stein M., Shapira S.C., Noga Y., Heruti R.J., Shemer J., Blast
injuries: bus versus open-air bombings — a comparative study of injuries in survi-
vors of open-air versus confined-space explosions. J Trauma Inj Infect Crit Care 1996,
Vol. 41(6).
17. Löhner R., Cebral J., Yang C., Baum J.D., Mestreau E., Charman C., Pelessone D.,
Large-scale fluid-structure interaction simulations. Comput Sci Eng 2004, Vol. 6(3).
18. Loukaitou-Sideris A., Taylor B.D., Fink C.N.Y., Rail transit security in an international
context: lessons from four cities. Urban Aff Rev 2006, Vol. 41(6).
19. Luccioni B.M., Ambrosini R.D., Danesi R., Blast load assessment using hydrocodes.
Eng Struct 2006, Vol. 28(12).
20. Luccioni, B.M., Ambrosini, R.D., Danesi, R.F., Analysis of building collapse under blast
loads. Eng Struct 2004, Vol. 26.
21. Mannan S., Lee’s loss prevention in the process industries; Volume 2: Hazard Identi-
fication, Assessment and Control. Amsterdam: Elsevier 2005.
22. Peral-Gutierrez de Ceballos J., Turégano-Fuentes F., Pérez-Díaz D., Sanz-Sánchez M.,
Martín-Llorente C., Guerrero-Sanz J.E., 11 March 2004: The terrorist bomb explo-
sions in Madrid, Spain — an analysis of the logistics, injuries sustained and clinical
management of casualties treated at the closest hospital. Critical Care 2004, Vol. 9(1).
23. Remennikov A.M., Rose T.A., Modelling blast loads on buildings in complex city
geometries. Comput Struct 2005, Vol. 83(27).
24. Smith P.D., Mays G.C., Rose T.A., Teo K.G., Roberts B.J., Small scale models of complex
geometry for blast overpressure assessment. Int J Impact Eng 1992, Vol. 12(3).
25. Smith P.D., Rose T.A., Blast wave propagation in city streets — an overview. Prog
Struct Eng Mater 2006, Vol. 8(1).
26. Solomos, G., Berra, M., Compressive behaviour of plain concrete at higher strain
rates. Journal of the Mechanical Behaviour of Materials 2005, Vol. 16, Nos. 1–2.
27. US Army TM 5–855–1: Fundamentals of Protective Design for Conventional Weap-
ons, Headquarters Department of the Army, Washington, D.C. 1986.
28. Van den Berg A.C., Weerheijm J., Blast phenomena in urban tunnel systems, J Loss
Prev Process Ind 2006, Vol. 19.

60 Internal Security, January–June 2011

Copyright © 2011 Police Academy in Szczytno. All Rights Reserved


Experimental and Numerical Simulation Activities...

29. Yet-Pole I., Cheng Te-Lung, The Development of a 3D Risk Analysis Method, J Hazard
Mater 2008, Vol. 153(1-2).

About the authors

George Solomos, PhD, is a researcher of Italian Institute for the Protection and Security of the Citizen
in Ispra. The Institute is a structural unit of the Joint Research Centre of the European Commission. His
fields of scientific interest are structural vulnerability, fast dynamics of structures and materials under
impulsive loading.
E-mail: george.solomos@jrc.ec.europa.eu
Martin Larcher, DEng, is  a  researcher of  the  Institute of  Engineering Mechanics and  Structural
Mechanics of  the  Universität der Bundeswehr München. He is  also a  Head of  the  Laboratory for
computational engineering of that Institute. His fields of scientific interest are dynamics of structures,
high strain-rate behavior of materials and FEM numerical simulations.
E-mail: martin.larcher@unibw.de
Folco Casadei, PhD, is a researcher of Italian Institute for the Protection and Security of the Citizen
in  Ispra which is  a  part of  the  Joint Research Centre of  the  European Commission. His fields
of scientific interest are structural mechanics, fast dynamics of structures, FEM numerical simulations
and fluid-structure interaction.
E-mail: folco.casadei@jrc.ec.europa.eu
Georgios Giannopoulos, PhD, is  a  researcher of  Italian Institute for the  Protection and  Security
of the Citizen in Ispra which is a part of the Joint Research Centre of the European Commission. His
fields of  scientific interest are physical/chemical exposure, dynamics of  structures and  numerical
simulations.
E-mail: georgios.giannopoulos@jrc.ec.europa.eu

Abstracts
Niedawne akty terrorystyczne w metropoliach na całym świecie ujawniły zagrożenie sieci transportowych, głównie zaś transportu
kolejowego. Szczególną rolę w zwiększeniu poziomu bezpieczeństwa może odgrywać architektura, łagodząca skutki ataków
terrorystycznych. W artykule zbadano zachowania strukturalnych i budulcowych materiałów przy wysokim poziomie obciążenia,
a także eksperymentalne techniki i oprzyrządowanie do odpowiedniego rodzaju obróbki. Równocześnie opracowano także instrumenty
modelowania cyfrowego, przydatne w wypadku szybkich dynamicznych zmian. Przy użyciu tych narzędzi autorzy podjęli próbę oceny
fizycznego zagrożenia infrastruktury transportu kolejowego w wypadku obciążeń detonacyjnych, głównie przy zagrożeniu atakiem
terrorystycznym. Parametry czasowe, parametry powietrza i samego budynku zostały dokładnie zaprojektowane z wykorzystaniem
programów niezbędnego segmentalnego modelowania Europlexus. Szczególną uwagę skupiono na modelowaniu kruszenia szkła, które
może stać się przyczyną wzrostu liczby i stopnia urazów. Na potrzeby badania przestudiowano budowę dworca kolejowego, którego
geometria została odtworzona z zastosowaniem techniki skanowania laserowego. Opracowano reakcje strukturalne dla kilku scenariuszy
wybuchu, a także z wykorzystaniem odpowiednich probit funkcji obliczono ogólny poziom ryzyka urazów i strat ludności. Nakreślono
potencjał danej metody przy wykorzystaniu jej w charakterze narzędzia rekonstrukcji miejsca zdarzenia w celach sądowo-medycznych
i eksperckich.

Недавние террористические акты в мегаполисах по всему миру показали уязвимость транспортной сети вообще
и железнодорожного транспорта в частности. Особую роль в улучшении уровня безопасности может играть архитектура,
смягчающая последствия терактов. В статье исследуется поведение структурных материалов и составляющих при
высоком уровне нагрузки, а также экспериментальные техники и оборудование для соответствующего вида разработок.
Параллельно, были также разработаны инструменты численного моделирования, пригодные для быстрых динамических
явлений. С помощью этих инструментов авторы попытались оценить физическую уязвимость железнодорожной

Internal Security, January–June 2011 61


Copyright © 2011 Police Academy in Szczytno. All Rights Reserved
George Solomos, Martin Larcher, Folco Casadei, Georgios Giannopoulos

транспортной инфраструктуры при взрывных нагрузках, в основном из-за угрозы террористических взрывов. Временные
параметры, параметры воздушной среды и самого здания были точно спроектированы с использованием программного
обеспечения конечного элементного моделирования Europlexus. Особое внимание было сосредоточено при моделировании
крушения стекла, которое может быть причиной увеличения количества и степени травм. Для данного исследования
было проведено изучение железнодорожного вокзала, геометрия которого была восстановлена с использованием техники
лазерного сканирования. Были разработаны структурные реакции для нескольких сценариев взрыва, а также, используя
соответствующие функции пробит, были проведены вычисления общей степени риска человеческих травм и потерь.
Намечен потенциал данного метода для использования в качестве инструмента реконструкции места происшествия
для судебно-медицинских и экспертных целей.

62 Internal Security, January–June 2011

Copyright © 2011 Police Academy in Szczytno. All Rights Reserved


Reproduced with permission of the copyright owner. Further reproduction prohibited without
permission.