Research ArticleAccess Control Points Reducing a Possible BlastImpact by Meandering

Martin Larcher Georgios Valsamos and Vasilis Karlos

European Commission Joint Research Centre Ispra Italy

Correspondence should be addressed to Martin Larcher martinlarchereceuropaeu

Received 24 August 2017 Accepted 7 December 2017 Published 11 February 2018

Academic Editor Filipe Santos

Copyright copy 2018 Martin Larcher et al (is is an open access article distributed under the Creative Commons AttributionLicense which permits unrestricted use distribution and reproduction in any medium provided the original work isproperly cited

In response to the heightened terror threat in recent years there is an increasing interest in the introduction of access controlzones at sites that are characterized by an increased likelihood of being the target of a terrorist attack as latest data reveal thatunprotected areas of mass congregation of people have become attractive to terrorist groups Such control zones could be locatedwithin the building that has to be protected or attached to it (e elevated security needs for these areas call for a design that willconsider the risk of internal explosive events(e purpose of this article is to outline a strategy for limiting the consequences of aninternal blast while guaranteeing that the produced blast wave does not propagate into vulnerable areas In particular the focus ison the introduction of a protective wall system in the form of a meander that allows unobstructed access of the public and at thesame time reduces the possible blast inflow to the buildingrsquos interior (e performed numerical simulations show that theproposed strategy yields much smaller injury risk areas compared to a design without the addition of protective walls and isrecommended for upgrading the security of buildings

1 Introduction

11 Motivation Several trends in modern terrorism need tobe considered for developing an effective protection strategyagainst a possible attack In the past the focus was on theprotection of critical infrastructures However recent ter-rorist attacks shifted towards soft targets due to the increasedlikelihood of success without the need for careful planningbig resources and special training of the perpetrators Softtargets are in general defined as unprotected areas with highconcentration of people like shopping malls stadia hotelsuniversities urban city centres festivals fairs places ofworship train and subway stations or Christmas marketsSuch targets might consist only of the attending public(eg markets open-air fairs parks festivals and parades) orcould also entail structural components (eg concert hallstheatres museums churches and transportation terminals)In the recent past Jihadist perpetrators have chosen this typeof targets with the intent to cause mass casualties and due tothe high social political andor economic impact of a suc-cessful attack As can be seen in Figure 1 all major attacks

(involving fatalities) in Europe over the last years targetedsoft targets which is also attributed to their increasedvulnerability

It was noticed that these attacks were often performed byutilizing unconventional andor ldquolow-techrdquo weapons suchas improvised explosive devices (IEDs) knives axes carsand trucks In particular for the attacks in Brussels (032016)and Paris (112015) IEDs were used as modus operandi Forinstance physical protection against vehicle-ramming at-tacks can be achieved by a properly designed barrier systemHowever the detection of smaller weapons and explosivesby specialized devices [1] is fundamental for reducing thepossibility of an attack (e adoption of detection measuresmight be advisable for places of mass congregation of peoplebut is crucial for controlling the access at high-risk sitesMoreover the location of this detection zone is of greatimportance since it is an area characterized by an increasedlikelihood of terrorist attacks (by using explosives andorfirearms) (e present article concentrates on the design ofsuch access control points and in particular their design inrelation to possible attacks by the use of explosives

HindawiAdvances in Civil EngineeringVolume 2018 Article ID 3506892 11 pageshttpsdoiorg10115520183506892

Concerning the physical resistance or structural robust-ness of a building a progressive collapse mechanism [2] mustbe avoided and glass surfaces like windows or faccedilades [3]should be protected against the impact of blast waves orprojectiles e physical protection of crowded places againstexplosive events can be performed by proper shadowingtechniques for example by using concrete blocks [4]

12 RiskMitigation e creation of an ecient plan for theprotection of critical infrastructures and soft targets startswith identifying the potential risk To ascertain the risk of anattack the analyst has to weigh its probable consequencesthe most suitable tactic and the weaknesses that may beexploited by the aggressors in their eshyort to impose damageand casualties is means that for calculating the potentialrisk the combination of threat level vulnerability level andtarget value is requirede threat level can be dened as theprobability of occurrence of an attack in a specic period oftime causing abnormal loading to the structure e vul-nerability level is linked to the identication of securityweaknesses and the potential damage to the system resultingfrom an event with a given intensity e target value can bequantied with respect to the damage in the targetrsquos assetssuch as injuries of the occupants human life loss capitalvalue reconstruction cost and disruption of services andfunctions

A carefully designed protection strategy and crisismanagement plan can eshyectively reduce the impact of anattack to the targetrsquos asset value e implementation ofa detection and investigation scheme can deter possibleaggressors and decrease the threat level while the assessedvulnerability can be reduced by strengthening the examinedstructure to limit the consequences of an attack

Many buildings can be identied as potential targets ofa terrorist attack Accessibility to these buildings is oftenrestricted and people entering the premises (in particularexternal visitors) are subjected to security checks to reducethe likelihood of an attack at the buildingrsquos interior Eventhough an access control strategy does not guarantee de-tection of all possible threats it can certainly be assumed that

perpetrators cannot enter carrying big amounts of explosivesor rearms Practically this means that both the risk ofpossible attacks inside the building and their consequencesare reduced

e location of a security control zone where people arechecked for the presence of explosive materials and weaponsprior entering to the building should be carefully selectedPrior experience from military applications shows thata dedicated building (guardhouse) detached from anexisting structure reduces possible consequences in case ofan attack at its interior Such a design approach is feasible ifspace is available but often in the densely built urban en-vironment buildings occupying nearly all the available lotand construction of a detached structure is not an option Inthe current article measures for the protection of accesscontrol areas in closed environment against explosions arepresented In particular the proposed strategy is based onthe employment of a meandering wall to reduce the possibleinow of the produced blast wave to a buildingrsquos interior

13 Policy Context Security is one of the recent priorities ofthe European Commission e European Agenda on Securityclearly advocates that the protection of critical infrastructuresand soft targets presents a real challenge e main objective isto gather best practise principles and produce guidance formitigating the risk of terrorism In particular it is highlightedthat the protection of soft targets and human life is of utmostimportance and additional eshyorts are required from both stateand private security stakeholders

2 Explosion Effects

21 Blast Loading Blast waves are typically characterized bya compression phase (positive phase) with a very high peakoverpressure followed by an underpressure phase (negativephase) Figure 2 shows an idealized prole of the pressure inrelation to time for the case of a blast wave and presentssome relevant parameters e compression phase includesan almost instantaneous increase from the ambient pressure (p0)to a peak pressure (p0 +pmax) at the arrival time ta when the

0

20

40

60

80

100

120

0114

BrusselsParis

Paris

Brussels

Berlin

Nice

KnifeExplosiveRamming

Human forceGun

0115 0116 0117Month

Fata

litie

s

Figure 1 Fatalities due to terrorist attacks in the European Union 2014ndash2017 modus operandi

2 Advances in Civil Engineering

shock front reaches the point of interest e time required forthe pressure to reach its peak value is very small and for designpurposes it is usually considered as equal to zero After its peakvalue the pressure decreases at an exponential rate until itreaches the ambient value at ta+ td td being called the positivephase duration After the positive phase the pressure dropsbelow the ambient value attaining its minimum value pminbefore nally returning to it after tn (negative phase duration)e positive impulse i is dened as the integral of the over-pressure curve over time and is distinguished in positive andnegative according to the relevant phase of the blast wave timehistory More detailed information on the blast wave param-eters is provided in [5ndash8]

e loading conditions of an externally loaded structurecan be distinguished into three types impulsive dynamic andquasi-static loading (Figure 3) Loads with very short duration(relative to the structurersquos or componentrsquos fundamental naturalperiod) are known as impulsive loads for example near-eldexplosions of smaller charges (rucksack bombs)e structuralconsequences of such explosions range from shear-type failures(eg of fragile parts like window panes) to bending-typefailures (eg stishyer concrete structures) Explosions withvery small distances or even contact detonations could causeeven spalling and scabbing [9] If the duration of the excitationis longer than the natural period of the structure or of thecomponent (dynamic loading) the predominant structuralfailuremode is bending as in the case of far-eld explosions (carbombs) and gas explosions A static load simulation approachcan be used only if the examined pressure increases at a verylow rate (quasi-static loading) For the other two loadingconditions sophisticated numerical models should be utilizedfor example an explicit nite element analysis For a particularstructure and load type (eg pressure) these loading regimescan be schematically presented by an iso-response curve ina so-called PI (pressure-impulse) diagram (Figure 3) Ifa probabilistic approach is of interest the probability of damagecould also be inserted in the diagram (eg Figure 4)

22VentingPressureRelease e diagram shown in Figure 2is only valid for spherical or hemispherical free air or

surface bursts respectively Reections shadowing orchannelling phenomena can signicantly alter the behaviourof a blast wave In such cases the calculation of the blastwave parameters with the empirical and semiempiricalformulas proposed in [5] could result in signicant errorsus the use of numerical uid-structure interaction cal-culations can provide a detailed identication of the loadingtime history of a structure or a component

Internal explosions in conned rooms or bunkers wherethe pressure wave cannot escape result in a static remainingpressure after the explosion Certain concepts for designingventing surfaces for gas explosions (eg [10 11]) are publicallyavailable However blasts from solid explosives in connedspaces are in general faster and are characterized by muchhigher peak pressure and often smaller impulse values incomparison to gas explosions e remaining static pressurefrom such explosions is signicant and must be consideredduring design Release surfaces might prove eshyective [12] asthey are able to reduce the remaining pressure after an ex-plosion even though their eshyects are not immediately evidentNevertheless venting surfaces cannot reduce the peak pres-sure associated with the rst arrival of the blast wave

23 Afterburning e afterburning phenomenon mayprove signicant for internal explosions depending on thetype of explosive burnable products that remain after thedetonationese hot products mix with the surrounding airand burn if the conditions (in particular oxygen availability)allow combustion Also the presence of aluminium powderin the explosive contributes to the combustion processwhich leads to an additional pressure after the rst peakpressure e eshyect of afterburning on the resulting pres-sures is inuenced by many factors In numerical simula-tions the approach of Miller [13] is commonly used whoproposed the addition of a certain amount of energy to themodel after a specic time Nevertheless afterburning isnot considered in the present article since injuries are mostlikely associated with the initial blast wave and not theafterburning

Impu

lsive

load

ing

Dynamic loading

Quasi-static loading

Unsafe

Safe

Impulse (Pamiddots)

Pres

sure

(Pa)

Figure 3 Typical PI diagram impulsive dynamic and quasi-staticloading (logarithmic scale)

pmin

pmaxi

t

p

ta td

p0

tn

Figure 2 Pressure-time history for an ideal free-eld air-blastwave

Advances in Civil Engineering 3

24 Vulnerability of Humans e eshyects of an explosion onhuman body can be distinguished into three broad cate-gories (Figure 5) Concerning explosion-induced injuriesthe risk due to the primary blast eects on humans has al-ready been precisely described by several authors [14]However this is not yet the case for secondary blast eectsthat are related to the produced fragments either from theexplosive charge or from the failure of structures that arepropelled with high velocities striking the human body Astraightforward procedure for assessing the additional riskdue to the presence of splinters as part of the secondary blasteshyects is currently under development [15]

In this work the formulation of the human injury riskmodel is based on the work of Gonzalez Ferradas et al [16]andMannan [17] It utilizes the peak overpressure pmax (in Pa)and the positive impulse I (in Pamiddots) to determine the proba-bility of eardrum rupture and the probability of death

241 Causes of Death Regarding possible causes of deaththree dishyerent cases are considered head impact wholebody impact and lung haemorrhage Using a PI diagram[14] it has been proven that the head impact is the dominantcause of death and its probit function can be dened as

0

2000

4000

6000

8000

10000

Impu

lse (P

amiddots)

00E

+0

10E

+5

20E

+5

30E

+5

40E

+5

50E

+5

60E

+5

70E

+5

80E

+5

90E

+5

10E

+6

Overpressure (Pa)

5 vulnerability50 vulnerability

95 vulnerability1000 Pamiddots limit

Figure 4 Vulnerability of humans (head impact)

Primary effects Blast

Secondary effects Fragments

Tertiary effects Impact on structures

Figure 5 Blast eshyects of an explosive charge (yellow) on humans (green) due to the created blast wave (red) and fragmentation (brown)

4 Advances in Civil Engineering

Y1 5minus 849ln2430pmax

+4 middot 108

pmax middot I1113888 1113889 (1)

242 Eardrum Rupture (e probit function Y4 of eardrumrupture is described according to the following equation

Y4 minus126 + 1524ln pmax( 1113857 (2)

(e probability of occurrence R (or the percentage ofthe affected population) of the corresponding injury is nextdetermined for each death-related function using thefollowing equation from Gonzalez Ferradas et al [16]which is a good approximation of the relevant cumulativenormal distribution

Ri minus325Y3i + 4876Y

2i minus 2066Yi + 27035 (3)

Human risk is presented in Figure 4 as a PI diagram fora 5 50 and 95 probability of death Clearly for calculatingthe risk of life loss at every point both the peak pressure andthe impulse values are required meaning that a full fluid-structure interaction (FSI) analysis is required It is notedthat the influence of flying debris (secondary blast effects) isnot included in these formulas

(ese formulations are developed for the positive phaseof spherical or hemispherical blast waves from free air andsurface bursts respectively and can be used with both in-cident and reflected blast parameters In case of confinedexplosions the remaining static pressure results in an in-creased impulse a phenomenon that is not considered fromthe available formulas (e duration of the explosionssimulated in [16] is in general shorter than the time requiredfor the effects of venting to be evident and therefore theabove formulas can be used In order to reduce the com-plexity as an additional criterion an impulse limit of1000 Pamiddots is used to identify critical zones at the buildingrsquosinterior UFC 3-340-02 [18] proposes a limit of 1380ndash1725 kPafor lethal damage as a result of lung haemorrhage (is valueignores the head impact a fact that is also recognized in UFC3-340-02 where a much lower threshold is referenced (16 kPa)In the present study the impulse value of 1000Pamiddots is used asan indicative limit motivated by Figure 4 and taking intoaccount aminimum peak overpressure value of about 250 kPa(e impulse parameter is used since it is more appropriatefor confined explosions (e objective of this approach isto acquire qualitative results to investigate the effects ofmeandering For quantitative results the abovementionedrisk formulas should be used

3 Blast Protective Design

Protective design of critical infrastructures against terroristattacks is usually performed in several steps

(i) (e first step is establishing plausible attack sce-narios which is a joint decision of the owner orstakeholder and the designer Assessing the prob-ability of occurrence of a terrorist attack throughestimating sources of threat could be an appropriate

tool for defining and choosing possible scenariosAn indication is given for example by the NorthAtlantic Treaty Organizationrsquos (NATO) Standard-ization Agreement (STANAG) 2280 [19]

(ii) (e next step is to calculate the corresponding loadson the structure by identifying the worst case sce-nario In some cases this can be performed by usingsimple formulas for the blast wave propagation asthose proposed by Kingery and Bulmash [5] andKinney and Graham [20] or by using a series ofdiagrams included in Unified Facilities Criteria suchas UFC 3-340-2 [18] Bogosian et al [21] comparedthese equations with experimental data and showedthat in urban environments the propagation of theblast wave is not spherical and differences appear atthe calculated blast parameters For such compli-cated geometries numerical simulations may benecessary to assess with greater precision theloading history of a structure as described in [22]

(iii) Subsequently the protection strategy against thedetermined blast load according to the individualstructural characteristics of the examined build-ing (eg windows or facades) should be definedSeveral options validated by numerical or ex-perimental methods can be adopted as de-scribed in [23]

Among the various structural and nonstructural elementswindows and facades are considered critical when facing anexplosion due to their extremely fragile nature In additiona progressive collapse mechanismmust always be prevented asit could lead to a large death toll In general the explosive energydecreases rapidly by increasing the standoff distance Henceemploying perimeter protection measures and restricting theaccess to a site constitute one of the most effective methods fordecreasing its vulnerability For example protection againstvehicle access can be achieved by the application of variousbarrier systems as shown in [24]

31 Access Point Design for Building Structures Dependingon the security level of a site access control measures may alsobe employed for the entrance of the public (e control zoneshould be placed in an area which should be resistant to bothexterior and interior explosive events Long queuing shouldbe avoided to minimize the likelihood of the access pointbecoming a target of a terrorist attack In case of a blast flyingdebris from the failed parts of the structure (in particularfrom glass failure) may cause extensive injuries or fatalities(erefore incorporating venting in the form of release sur-faces is important for mitigating the effects of a blast inparticular if the access zone is inside the main building orattached to it (e release surfaces should be light to be easilypushed aside by the propagating blast wave or fragile toinstantly fail under the blast pressure To prevent peoplersquosinjury from the potential fragments of the failed opening therelease surfaces are often located at the roof of the structureNevertheless the effects of venting should not be over-estimated (e release surfaces (whose size is up to 10 of the

Advances in Civil Engineering 5

roomrsquos size) have a very small impact on the peak pressure butare eshyective for reducing the remaining static pressure afterthe explosion Clearly the proper design of release surfaces isa challenging task since no clear indications exist in the openliterature Usually it is advised to follow a design approachthat depends on the geometry of the control zone while thesize of the openings ranges from 3 to 8 of the roomrsquos(control arearsquos) size us these values serve only as an in-dication since the actual release surface size depends also onits cross section and the determined blast scenario

If the access control zone is located within or is con-nected to the main building the propagation of the pressurewave into the adjacent structure should be avoided toconne the explosive eshyects only in one area Constructionshaving very stishy or blast resistant doors are complicatedimpractical and space demanding since these doors must beduplicated As shown below a so-calledmeandering wall canachieve a similar protection level maintaining a design thatfavors the unobstructed ow of people

4 Numerical Simulations

e current section is dedicated to the analysis of a casestudy simulating an explosion within an access control zonee objective of the case study is to investigate the waya rigid (meandering) wall ashyects the propagation of the blastwave into an area of the building that needs to be protectedMoreover a parametric study on the inuence of the basicdimensions of the meandering wall on the attenuation of theproduced detonation wave is carried out e numericalsimulations are performed with the explicit FE codeEUROPLEXUS (EPX) [25] which is jointly developed by theFrench Commissariat a lrsquo Energie Atomique et aux EnergiesAlternatives (CEA) and the Joint Research Centre of the

European Commission (JRC Ispra) EPX employs sophis-ticated uid-structure interaction techniques and is suitablefor simulating fast and severe loadings on structures such asimpact and blast e CAST3M code [26] also developed byCEA was used to generate the mesh for the numericalmodels while ParaView software [27] was used for post-processing the results

41 Case Study To reduce complexity a simple geometryhas been selected for the case study A rectangular area isseparated by a rigid wall into two partse rst compartmentis considered as the entrance zone where a possible intrudercan enter carrying an improvised explosive device (IED) andwalk till the opening of the rigid wall that leads to the secondcompartment Just before the opening that connects the twoparts a thorough control of all persons entering the secondpart is considered to be implemented e second com-partment is regarded a vulnerable zone that needs to beprotected as a possible blast wave can become critical for thepersonnel or the structural integrity of the building

Figure 6 shows the geometrical details of the investigatedarea e entrance zone located on the left is connected tothe protected zone on the right through a 3m wideopening e openingrsquos width guarantees a relatively highow of the people to avoid queuinge height of both zonesis 3 meters while the detonation centre is located 15 metersfrom the ground level (average value for the location ofa backpack in the vertical axis) and is centred to the openingbased on a worst case scenario is is assumed to be theclosest point to the protected zone that a perpetrator canreach prior to being intercepted e mass of the IED hasbeen selected to be equal to 25 kilograms of TNTequivalentas an upper limit for the amount of explosives a pedestriancould carry in a backpack or small suitcase

10 mEntrance zone

Explosive material

Reflecting boundaries

Absorbing boundaries

Protected zone

Blocking wall (meandering)

2 m

3 m

E

D

10 m 20 mRigid wall

Figure 6 Geometry of the case study model

6 Advances in Civil Engineering

As shown in Figure 6 the distance D between theopening and the introduced meandering wall and the di-mension E that refers to the overlapping between the ad-ditional blocking wall and the preexisting rigid wall are notexplicitly determined (ese two parameters are selected asvariables in the performed parametric study in order toquantify their influence on the blast wave propagationD ranges from 1 to 3 meters corresponding to constructionspecifications (eg wheelchair access and personnel flow)Similarly E ranges from 05 to 2 meters to allow a minimumdistance of 15 meters between each extremity of themeandering wall and the lateral walls of the building and toguarantee comfortable access to the attached protected zone

42NumericalModel (e structural parts of the building inthe case study are assumed rigid meaning that the walls theceiling and the floor of the structure are nondeformableA postanalysis investigation of the recorded maximum pres-sures indicated that thick stiff structural elements like concreteor steel walls respect this basic assumption (is hypothesisresults in decreased computational power demand sincethere is no need to include deformable structural parts orperform cumbersome fluid-structure interaction (FSI)calculations Moreover the absence of structural de-formations allows an accurate study of the influence of thetwo parameters (D and E) on the blast wave propagation

As already mentioned for the numerical calculationsa purely Eulerian approach is utilized where only the fluidpart of the model is considered (e Eulerian formulationconsiders the computational mesh fixed while the fluid(particles) moves relative to these fixed grid points (econservation laws of mass momentum and energy areexpressed in a spatial framework (e discretization is basedon cell-centred finite volume (FV) formulations in which thegoverning equations are solved in integral form (ismethod is conservative since the formulation ensures thatthe flux entering a given volume is identical to that leaving

the adjacent volume (e numerical fluxes between adjacentFVs were calculated using the Harten-Lax-van Leer-contact(HLLC) Reimann solver [28]

(e fluid domain was modelled with 3D cubic finitevolumes (CUVF) of 01 meter edge size (e size of the cubicvolumes is selected after a convergence parametric study Onthe extremity surfaces of the two different zones (yellow lineson Figure 6) absorbing boundaries have been introduced inorder to simulate nonreflecting atmospheric boundariesSuitable shell elements (CL3D) that allow the introduction ofan absorbing medium have been employed and froma mathematical point of view represent a supersonic outletAs already mentioned the rigid parts of the model aresimulated as reflecting boundaries For a 3D cubic FV a freesurface is considered by definition as a reflecting boundaryA simple duplication of the fluid nodes at the area where therigid structure is located results into free FV surfaces thatbehave as reflecting boundaries

(e representation of the blast load is performed viathe 1D to 3D mapping technique [29] during whicha 1D numerical calculation that simulates the blast load isperformed with a very fine mesh(e data obtained from the1D calculation are mapped into a spherical shape andinserted in the coarser 3D mesh An important detail thatshould be taken into account is that when transferring theresults of the 1D analysis into the 3D mesh the blast load isconsidered spherical so the transformation should beperformed before the wave comes to contact with any ob-stacles (is mapping technique can significantly acceleratethe calculation time without sacrificing the accuracy of theresults For the representation of the detonation load inthe 1D calculation the Jones-Wilkins-Lee EOS law is used[30](e material parameters for the TNTare proposed by[31] and are included in Table 1 (e perfect gas law isemployed for the representation of the atmospheric air inthe fluid simulation (e material parameters for theatmospheric air are shown in Table 2 Finally one

Table 1 Parameters for the JWL EOS for TNT (Dobratz and Crawford [31])

Parameters Description Parameters for explosive Parameters for airA (Pa) Material constant (experimental) 37121e11 37121e11B (Pa) Material constant (experimental) 323e9 323e9R1 Material constant (experimental) 415 415R2 Material constant (experimental) 095 095Beta Material constant (experimental) 025 025RHO (kgm3) Density 1630 13EINT (Jkg) Current internal energyPer unit mass 429e6 021978e6OMEGA Specific heat ratio 03 03D (ms) Detonation velocity 6930 mdash

Table 2 Material properties for the ideal gas law for atmospheric air

Density (kgm3) Specific heat at constant volume (JkgmiddotK) Gamma ratio Reference pressure (Pa)13 71675 14 1e5

Advances in Civil Engineering 7

symmetry plane was utilized by modelling only one half ofthe uid mesh

43 Numerical Results e objective of the numerical in-vestigation is to determine the eshyectiveness of the in-troduced meandering wall Figure 7 displays the behaviourof the propagated blast wave for the ldquoopen doorrdquo case anda ldquomeanderingrdquo case (D 1m E 1m) for a plane view atthe height of 15 meters A comparison of the two cases isperformed at several time frames in order to provide a clearrepresentation on the diversion of the blast wave from theadditional wall As already mentioned half of the geometrywas modelled in the numerical calculations so in order to

present the full model in Figure 7 special mirroring tech-niques are used in the postprocessing phase e rstsubgure presents the pressure eld distribution at 3msafter the detonation At that time frame the wave front startsbeing reected on the meandering wall as it is clear from thedishyerences between the two models For the ldquomeanderingrdquocase the blast wave starts to enter the protected zone at 4msand is split into two smaller waves propagating from bothsides of themeandering wall At 8ms it is evident that in theldquoopen doorrdquo case there is one strong wave front propagatinginto the protected zone while for the ldquomeanderingrdquo casethere are two weaker wave fronts e wave propagationcontinues in a similar way for both cases until 30ms after thedetonation where the two wave fronts of the ldquomeanderingrdquo

3 ms

10 ms 15 ms

Pressure

30 ms

4 ms 8 ms

10e+05 15e+05 20e+05 25e+05 30e+05

Figure 7 Pressure distribution of the blast wave for the ldquoopen doorrdquo model and a ldquomeanderingrdquo model (D 1m E 1m) for several timeframes (Pa)

500 10 20 30 40 50 60 70

100

150

200

250

300

Time (ms)

Pres

sure

(kPa

)

Open doorMeandering

(a)

0 10 20 30 40 50 60 70Time (ms)

Impu

lse (P

amiddots)

Open doorMeandering

0

500

1000

1500

(b)

Figure 8 Comparison of (a) pressure and (b) impulse values for a point located in the protected zone for an ldquoopen doorrdquo model anda ldquomeandering wallrdquo model (D 1m E 1m)

8 Advances in Civil Engineering

case are united into one But as can be noticed even in thatphase the wave front of the ldquomeanderingrdquo case is weaker thanthe one of the ldquoopen doorrdquo case

e meandering wall reects part of the initial blast waveand guides it back to the entrance area resulting in a slightlystronger blast wave into the entrance zone Since this zone hasalready been signicantly ashyected from the initial blast wave theadditional energy due to the presence of the ldquomeanderingrdquo walldoes not ashyect the injury risk probability Figure 8 presents thepressure and the impulse history for the abovementioned casesat a point that is located at the middle of the protected zone7 meters from the opening and 15 meters from the ground

level From the pressure-time history it is clear that in theldquomeanderingrdquo case the rst peak is eliminated while the secondpeak that results from the reection of the wave onto the lateralwalls is lower Similar results can be derived from comparingthe impulse-time histories as there is a signicant reduction ofthe total impulse when the meandering wall is present

e next gures present the inuence of the dimensionand location of the meandering wall on the blast wavepropagation As has already been described in previoussections areas with an impulse value larger than 1000 Pamiddotsare considered as high injury risk zones Figure 9 presentsthe injury risk areas for the protected zone of the examined

E 05

D1

D2Impulse

30e+03

20e+03

10e+03

00e+00

D3

E 1 E 2

Figure 9 Zone of impulse higher than 1000 Pamiddots for the ldquomeanderingrdquo studied cases and the ldquoopen doorrdquo model (parameters D and E in mand impulse in pamiddots)

105 151

E (m)

D (m

)

()

2

15

2

25

3

60

40

20

60

40

20

0

Figure 10 Percentage of ashyected area spectrum versus the investigated parameters for the ldquomeanderingrdquo model

Advances in Civil Engineering 9

building for several studied geometries (e results refer tothe same horizontal cut that is presented in Figure 7 but theoutput quantity is the calculated impulse Areas charac-terized by impulse values higher than 1000 Pamiddots are high-lighted D and E determine the location and the size of themeandering wall as described earlier

In the same figure the ldquoopen doorrdquo case is also includedfor comparison purposes In the ldquoopen doorrdquo case themajority of the protected zone results in a high injury riskarea In the ldquomeanderingrdquo case the size of the high injuryrisk area varies but is always smaller than the ldquoopen doorrdquocase It is clear that the area before the additional wall isalways unprotected irrespective of the presence of themeandering wall Moreover in some cases high impulsevalues are also recorded at the sides of the meandering walldue to the relatively strong wave fronts that are present atboth those areas Figure 10 summarizes the output of theparametric study by presenting the percentage of the injuryrisk area in the protected zone (above 1000 Pamiddots) versus thetwo examined parameters From comparing the severalldquomeanderingrdquo cases it can be concluded that if D is small(distance between the door and the added wall) the influ-ence of E (dimension of overlapping length) is of minorimportance However if D is large the influence of E issignificant (if E decreases the blast wave that enters into theprotected zone is stronger and the high injury risk areas arebecoming larger)

5 Conclusions

Access control points that are located inside or attached tobuildings are often adopted due to local constraints (eirdesign must consider possible terrorist attack scenarios(eg explosions) at the interior and exterior of these accesscontrol points (e mitigation of the consequences shouldsuch a terrorist attack materialize inside an access point andin the attached building can be achieved by using venting(release surfaces) and in particular meandering walls (einvestigation shows that the reduction of the pressure issignificant if meandering walls are used and that these wallscan therefore play an important role in separating accesscontrol points from the attached buildings Nevertheless it isobvious that meandering walls can be considered an effectiverisk reduction measure and not as a full protection solution(e risk reduction concerns both blast wave effects and forobvious reasons fragments resulting from explosions andshooting attacks Herein the idealized situation that is in-vestigated can be extended to include additional parametersAs has been presented in the current article the presence ofa rigid wall standing between an explosionrsquos detonationcentre and a target can effectively reduce the severity of theimpacting blast wave It is therefore recommended theaddition of such elements (varying in size and geometry) inorder to mitigate the risk of terrorist attacks

Conflicts of Interest

(e authors declare that there are no conflicts of interestregarding the publication of this paper

References

[1] A M Lewis ldquoState of the art report on European legislationrelating to explosives and explosive detection system for non-aviation configurationsrdquo Tech Rep JRC84076 Joint ResearchCentre Ispra Italy 2013

[2] B M Luccioni R D Ambrosini and R F Danesi ldquoAnalysisof building collapse under blast loadsrdquo Engineering Structuresvol 26 no 1 pp 63ndash71 2004

[3] M Larcher M Arrigoni C Bedon et al ldquoDesign of blast-loaded glazing windows and facades a review of essentialrequirements towards standardizationrdquo Advances in CivilEngineering vol 2016 Article ID 2604232 14 pages 2016

[4] R Hajek and M Foglar ldquoNumerical and experimentalanalysis of the effect of rigid barriers on blast wave propa-gationrdquo Journal of Structural Engineering vol 141 no 12p 04015061 2015

[5] C N Kingery and G Bulmash ldquoAirblast parameters fromTNT spherical air burst and hemispherical surface burstrdquoTech Rep ARBRL-TR-02555 Defence Technical InformationCenter Ballistic Research Laboratory Aberdeen ProvingGround MD USA Tech Rep ARBRL-TR-02555 1984

[6] M Larcher ldquoPressure-time functions for the description of airblast wavesrdquo Tech Rep JRC46829 Joint Research CentreIspra Italy 2008

[7] V Karlos and G Solomos ldquoCalculation of blast loads forapplication to structural componentsrdquo Tech Rep EUR26456EN Joint Research Centre Publications Office of theEuropean Union Luxembourg 2013

[8] V Karlos G Solomos and M Larcher ldquoAnalysis of the blastwave decay coefficient using the Kingery-Bulmash datardquoInternational Journal of Protective Structures vol 7 no 3pp 409ndash429 2016

[9] M Larcher ldquoDevelopment of discrete cracks in concreteloaded by shock wavesrdquo International Journal for ImpactEngineering vol 36 pp 700ndash710 2009

[10] J Li and H Hao ldquoInternal and external pressure prediction ofvented gas explosion in large rooms by using analytical andCFD methodsrdquo Journal of Loss Prevention in the ProcessIndustries vol 49 pp 367ndash381 2017

[11] EN14494 ldquoGas explosion venting protective systemsrdquo 2007[12] M Larcher F Casadei and G Solomos ldquoInfluence of venting

areas on the air blast pressure inside tubular structures likerailway carriagesrdquo Journal of Hazardous Materials vol 183no 1-3 pp 839ndash846 2010

[13] P J Miller ldquoA reactive flow model with coupled reactionkinetics for detonation and combustion in non-ideal explo-sivesrdquo MRS Proceedings vol 418 1995

[14] M Larcher R Forsberg U Bjoernstig A Holgersson andG Solomos ldquoEffectiveness of finite element modelling ofdamage and injuries for explosions inside trainsrdquo Journal ofTransportation Safety and Security vol 8 pp 83ndash100 2015

[15] G Valsamos F Casadei M Larcher and G SolomosldquoImplementation of flying debris fatal risk calculation inEUROPLEXUSrdquo Tech Rep JRC94805 Joint Research CentrePublications Office of the European Union Luxembourg 2015

[16] E Gonzalez Ferradas F Dıaz Alonso M Doval MintildearroA Mintildeana Aznar J Ruiz Gimeno and J Sanchez PerezldquoConsequence analysis by means of characteristic curves todetermine the damage to humans from bursting sphericalvesselsrdquo Process Safety and Environmental Protection vol 86no 2 pp 121ndash129 2008

10 Advances in Civil Engineering

[17] S Mannan Leesrsquo Loss Prevention in the Process IndustriesHazard Identification Assessment and Control Vol 2 ElsevierAmsterdam 2005

[18] Department of Defence USA Unified Facilities Criteria UFC3-340ndash2 2008

[19] ldquoNATO STANAG 2280 design threat levels and handoverprocedure for temporary protective structuresrdquo 2008

[20] G F Kinney and K J Graham Explosive Shocks in AirSpringer Berlin Germany 1985

[21] D Bogosian J Ferritto and Y Shi ldquoMeasuring uncertaintyand conservatism in simplified blast modelsrdquo in Proceedingsof 30th Explosive Safety Seminar Atlanta Georgia August 2002

[22] M Larcher and F Casadei ldquoExplosions in complexgeometriesndasha comparison of several approachesrdquo InternationalJournal of Protective Structures vol 1 no 2 pp 169ndash195 2010

[23] M Larcher A Stolz L Rudiger and M Steyerer ldquoStatischeErsatzlasten und ldquoIngenieurverfahrenrdquordquo in Berichte ausdem Konstruktiven Ingenieurbau Workshop Bau-ProtectN Gebbeken K (oma M Rudiger and M Klaus Edspp 117ndash130 Universitat der Bundeswehr MunchenNeubiberg Germany 2012

[24] FEMA ldquoSite and urban design for securityndashguidance againstpotential terrorist attacksrdquo Tech Rep FEMA 430 FEMAWashington DC USA 2007

[25] Joint Research Centre and Commissariat a lrsquoenergie atomiqueet aux energies alternatives EUROPLEXUS httpwww-epxceafr

[26] Commissariat a lrsquoenergie atomique et aux energies alterna-tives CAST3M httpwww-cast3mceafr

[27] ParaView httpwwwparavieworg[28] F Daude and P Galon ldquoOn the computation of the baerndash

nunziato model using ALE formulation with HLL-and HLLC-type solvers towards fluidndashstructure interactionsrdquo Journal ofComputational Physics vol 324 pp 189ndash230 2016

[29] M Larcher F Casadei and G Solomos ldquoSimulation of blastwaves by usingmapping technology in EUROPLEXUSrdquo TechRep JRC91102 Joint Research Centre Publications Office ofthe European Union Luxembourg 2014

[30] E Lee J Horning and J Kury ldquoAdiabatic expansion of highexplosives detonation productsrdquo University of CaliforniaLawrence Livermore National Laboratory Livermore CAUSA 1968

[31] B Dobratz and P Crawford LLNL Explosives HandbookProperties of Chemical Explosives and Explosive SimulantsUniversity of California Lawrence Livermore NationalLaboratory Livermore CA USA 1985

Advances in Civil Engineering 11

shock front reaches the point of interest e time required forthe pressure to reach its peak value is very small and for designpurposes it is usually considered as equal to zero After its peakvalue the pressure decreases at an exponential rate until itreaches the ambient value at ta+ td td being called the positivephase duration After the positive phase the pressure dropsbelow the ambient value attaining its minimum value pminbefore nally returning to it after tn (negative phase duration)e positive impulse i is dened as the integral of the over-pressure curve over time and is distinguished in positive andnegative according to the relevant phase of the blast wave timehistory More detailed information on the blast wave param-eters is provided in [5ndash8]

e loading conditions of an externally loaded structurecan be distinguished into three types impulsive dynamic andquasi-static loading (Figure 3) Loads with very short duration(relative to the structurersquos or componentrsquos fundamental naturalperiod) are known as impulsive loads for example near-eldexplosions of smaller charges (rucksack bombs)e structuralconsequences of such explosions range from shear-type failures(eg of fragile parts like window panes) to bending-typefailures (eg stishyer concrete structures) Explosions withvery small distances or even contact detonations could causeeven spalling and scabbing [9] If the duration of the excitationis longer than the natural period of the structure or of thecomponent (dynamic loading) the predominant structuralfailuremode is bending as in the case of far-eld explosions (carbombs) and gas explosions A static load simulation approachcan be used only if the examined pressure increases at a verylow rate (quasi-static loading) For the other two loadingconditions sophisticated numerical models should be utilizedfor example an explicit nite element analysis For a particularstructure and load type (eg pressure) these loading regimescan be schematically presented by an iso-response curve ina so-called PI (pressure-impulse) diagram (Figure 3) Ifa probabilistic approach is of interest the probability of damagecould also be inserted in the diagram (eg Figure 4)

22VentingPressureRelease e diagram shown in Figure 2is only valid for spherical or hemispherical free air or

surface bursts respectively Reections shadowing orchannelling phenomena can signicantly alter the behaviourof a blast wave In such cases the calculation of the blastwave parameters with the empirical and semiempiricalformulas proposed in [5] could result in signicant errorsus the use of numerical uid-structure interaction cal-culations can provide a detailed identication of the loadingtime history of a structure or a component

Internal explosions in conned rooms or bunkers wherethe pressure wave cannot escape result in a static remainingpressure after the explosion Certain concepts for designingventing surfaces for gas explosions (eg [10 11]) are publicallyavailable However blasts from solid explosives in connedspaces are in general faster and are characterized by muchhigher peak pressure and often smaller impulse values incomparison to gas explosions e remaining static pressurefrom such explosions is signicant and must be consideredduring design Release surfaces might prove eshyective [12] asthey are able to reduce the remaining pressure after an ex-plosion even though their eshyects are not immediately evidentNevertheless venting surfaces cannot reduce the peak pres-sure associated with the rst arrival of the blast wave

23 Afterburning e afterburning phenomenon mayprove signicant for internal explosions depending on thetype of explosive burnable products that remain after thedetonationese hot products mix with the surrounding airand burn if the conditions (in particular oxygen availability)allow combustion Also the presence of aluminium powderin the explosive contributes to the combustion processwhich leads to an additional pressure after the rst peakpressure e eshyect of afterburning on the resulting pres-sures is inuenced by many factors In numerical simula-tions the approach of Miller [13] is commonly used whoproposed the addition of a certain amount of energy to themodel after a specic time Nevertheless afterburning isnot considered in the present article since injuries are mostlikely associated with the initial blast wave and not theafterburning

Impu

lsive

load

ing

Dynamic loading

Quasi-static loading

Unsafe

Safe

Impulse (Pamiddots)

Pres

sure

(Pa)

Figure 3 Typical PI diagram impulsive dynamic and quasi-staticloading (logarithmic scale)

pmin

pmaxi

t

p

ta td

p0

tn

Figure 2 Pressure-time history for an ideal free-eld air-blastwave

Advances in Civil Engineering 3

24 Vulnerability of Humans e eshyects of an explosion onhuman body can be distinguished into three broad cate-gories (Figure 5) Concerning explosion-induced injuriesthe risk due to the primary blast eects on humans has al-ready been precisely described by several authors [14]However this is not yet the case for secondary blast eectsthat are related to the produced fragments either from theexplosive charge or from the failure of structures that arepropelled with high velocities striking the human body Astraightforward procedure for assessing the additional riskdue to the presence of splinters as part of the secondary blasteshyects is currently under development [15]

In this work the formulation of the human injury riskmodel is based on the work of Gonzalez Ferradas et al [16]andMannan [17] It utilizes the peak overpressure pmax (in Pa)and the positive impulse I (in Pamiddots) to determine the proba-bility of eardrum rupture and the probability of death

241 Causes of Death Regarding possible causes of deaththree dishyerent cases are considered head impact wholebody impact and lung haemorrhage Using a PI diagram[14] it has been proven that the head impact is the dominantcause of death and its probit function can be dened as

0

2000

4000

6000

8000

10000

Impu

lse (P

amiddots)

00E

+0

10E

+5

20E

+5

30E

+5

40E

+5

50E

+5

60E

+5

70E

+5

80E

+5

90E

+5

10E

+6

Overpressure (Pa)

5 vulnerability50 vulnerability

95 vulnerability1000 Pamiddots limit

Figure 4 Vulnerability of humans (head impact)

Primary effects Blast

Secondary effects Fragments

Tertiary effects Impact on structures

Figure 5 Blast eshyects of an explosive charge (yellow) on humans (green) due to the created blast wave (red) and fragmentation (brown)

4 Advances in Civil Engineering

Y1 5minus 849ln2430pmax

+4 middot 108

pmax middot I1113888 1113889 (1)

242 Eardrum Rupture (e probit function Y4 of eardrumrupture is described according to the following equation

Y4 minus126 + 1524ln pmax( 1113857 (2)

(e probability of occurrence R (or the percentage ofthe affected population) of the corresponding injury is nextdetermined for each death-related function using thefollowing equation from Gonzalez Ferradas et al [16]which is a good approximation of the relevant cumulativenormal distribution

Ri minus325Y3i + 4876Y

2i minus 2066Yi + 27035 (3)

Human risk is presented in Figure 4 as a PI diagram fora 5 50 and 95 probability of death Clearly for calculatingthe risk of life loss at every point both the peak pressure andthe impulse values are required meaning that a full fluid-structure interaction (FSI) analysis is required It is notedthat the influence of flying debris (secondary blast effects) isnot included in these formulas

(ese formulations are developed for the positive phaseof spherical or hemispherical blast waves from free air andsurface bursts respectively and can be used with both in-cident and reflected blast parameters In case of confinedexplosions the remaining static pressure results in an in-creased impulse a phenomenon that is not considered fromthe available formulas (e duration of the explosionssimulated in [16] is in general shorter than the time requiredfor the effects of venting to be evident and therefore theabove formulas can be used In order to reduce the com-plexity as an additional criterion an impulse limit of1000 Pamiddots is used to identify critical zones at the buildingrsquosinterior UFC 3-340-02 [18] proposes a limit of 1380ndash1725 kPafor lethal damage as a result of lung haemorrhage (is valueignores the head impact a fact that is also recognized in UFC3-340-02 where a much lower threshold is referenced (16 kPa)In the present study the impulse value of 1000Pamiddots is used asan indicative limit motivated by Figure 4 and taking intoaccount aminimum peak overpressure value of about 250 kPa(e impulse parameter is used since it is more appropriatefor confined explosions (e objective of this approach isto acquire qualitative results to investigate the effects ofmeandering For quantitative results the abovementionedrisk formulas should be used

3 Blast Protective Design

Protective design of critical infrastructures against terroristattacks is usually performed in several steps

(i) (e first step is establishing plausible attack sce-narios which is a joint decision of the owner orstakeholder and the designer Assessing the prob-ability of occurrence of a terrorist attack throughestimating sources of threat could be an appropriate

tool for defining and choosing possible scenariosAn indication is given for example by the NorthAtlantic Treaty Organizationrsquos (NATO) Standard-ization Agreement (STANAG) 2280 [19]

(ii) (e next step is to calculate the corresponding loadson the structure by identifying the worst case sce-nario In some cases this can be performed by usingsimple formulas for the blast wave propagation asthose proposed by Kingery and Bulmash [5] andKinney and Graham [20] or by using a series ofdiagrams included in Unified Facilities Criteria suchas UFC 3-340-2 [18] Bogosian et al [21] comparedthese equations with experimental data and showedthat in urban environments the propagation of theblast wave is not spherical and differences appear atthe calculated blast parameters For such compli-cated geometries numerical simulations may benecessary to assess with greater precision theloading history of a structure as described in [22]

(iii) Subsequently the protection strategy against thedetermined blast load according to the individualstructural characteristics of the examined build-ing (eg windows or facades) should be definedSeveral options validated by numerical or ex-perimental methods can be adopted as de-scribed in [23]

Among the various structural and nonstructural elementswindows and facades are considered critical when facing anexplosion due to their extremely fragile nature In additiona progressive collapse mechanismmust always be prevented asit could lead to a large death toll In general the explosive energydecreases rapidly by increasing the standoff distance Henceemploying perimeter protection measures and restricting theaccess to a site constitute one of the most effective methods fordecreasing its vulnerability For example protection againstvehicle access can be achieved by the application of variousbarrier systems as shown in [24]

31 Access Point Design for Building Structures Dependingon the security level of a site access control measures may alsobe employed for the entrance of the public (e control zoneshould be placed in an area which should be resistant to bothexterior and interior explosive events Long queuing shouldbe avoided to minimize the likelihood of the access pointbecoming a target of a terrorist attack In case of a blast flyingdebris from the failed parts of the structure (in particularfrom glass failure) may cause extensive injuries or fatalities(erefore incorporating venting in the form of release sur-faces is important for mitigating the effects of a blast inparticular if the access zone is inside the main building orattached to it (e release surfaces should be light to be easilypushed aside by the propagating blast wave or fragile toinstantly fail under the blast pressure To prevent peoplersquosinjury from the potential fragments of the failed opening therelease surfaces are often located at the roof of the structureNevertheless the effects of venting should not be over-estimated (e release surfaces (whose size is up to 10 of the

Advances in Civil Engineering 5

roomrsquos size) have a very small impact on the peak pressure butare eshyective for reducing the remaining static pressure afterthe explosion Clearly the proper design of release surfaces isa challenging task since no clear indications exist in the openliterature Usually it is advised to follow a design approachthat depends on the geometry of the control zone while thesize of the openings ranges from 3 to 8 of the roomrsquos(control arearsquos) size us these values serve only as an in-dication since the actual release surface size depends also onits cross section and the determined blast scenario

If the access control zone is located within or is con-nected to the main building the propagation of the pressurewave into the adjacent structure should be avoided toconne the explosive eshyects only in one area Constructionshaving very stishy or blast resistant doors are complicatedimpractical and space demanding since these doors must beduplicated As shown below a so-calledmeandering wall canachieve a similar protection level maintaining a design thatfavors the unobstructed ow of people

4 Numerical Simulations

e current section is dedicated to the analysis of a casestudy simulating an explosion within an access control zonee objective of the case study is to investigate the waya rigid (meandering) wall ashyects the propagation of the blastwave into an area of the building that needs to be protectedMoreover a parametric study on the inuence of the basicdimensions of the meandering wall on the attenuation of theproduced detonation wave is carried out e numericalsimulations are performed with the explicit FE codeEUROPLEXUS (EPX) [25] which is jointly developed by theFrench Commissariat a lrsquo Energie Atomique et aux EnergiesAlternatives (CEA) and the Joint Research Centre of the

European Commission (JRC Ispra) EPX employs sophis-ticated uid-structure interaction techniques and is suitablefor simulating fast and severe loadings on structures such asimpact and blast e CAST3M code [26] also developed byCEA was used to generate the mesh for the numericalmodels while ParaView software [27] was used for post-processing the results

41 Case Study To reduce complexity a simple geometryhas been selected for the case study A rectangular area isseparated by a rigid wall into two partse rst compartmentis considered as the entrance zone where a possible intrudercan enter carrying an improvised explosive device (IED) andwalk till the opening of the rigid wall that leads to the secondcompartment Just before the opening that connects the twoparts a thorough control of all persons entering the secondpart is considered to be implemented e second com-partment is regarded a vulnerable zone that needs to beprotected as a possible blast wave can become critical for thepersonnel or the structural integrity of the building

Figure 6 shows the geometrical details of the investigatedarea e entrance zone located on the left is connected tothe protected zone on the right through a 3m wideopening e openingrsquos width guarantees a relatively highow of the people to avoid queuinge height of both zonesis 3 meters while the detonation centre is located 15 metersfrom the ground level (average value for the location ofa backpack in the vertical axis) and is centred to the openingbased on a worst case scenario is is assumed to be theclosest point to the protected zone that a perpetrator canreach prior to being intercepted e mass of the IED hasbeen selected to be equal to 25 kilograms of TNTequivalentas an upper limit for the amount of explosives a pedestriancould carry in a backpack or small suitcase

10 mEntrance zone

Explosive material

Reflecting boundaries

Absorbing boundaries

Protected zone

Blocking wall (meandering)

2 m

3 m

E

D

10 m 20 mRigid wall

Figure 6 Geometry of the case study model

6 Advances in Civil Engineering

As shown in Figure 6 the distance D between theopening and the introduced meandering wall and the di-mension E that refers to the overlapping between the ad-ditional blocking wall and the preexisting rigid wall are notexplicitly determined (ese two parameters are selected asvariables in the performed parametric study in order toquantify their influence on the blast wave propagationD ranges from 1 to 3 meters corresponding to constructionspecifications (eg wheelchair access and personnel flow)Similarly E ranges from 05 to 2 meters to allow a minimumdistance of 15 meters between each extremity of themeandering wall and the lateral walls of the building and toguarantee comfortable access to the attached protected zone

42NumericalModel (e structural parts of the building inthe case study are assumed rigid meaning that the walls theceiling and the floor of the structure are nondeformableA postanalysis investigation of the recorded maximum pres-sures indicated that thick stiff structural elements like concreteor steel walls respect this basic assumption (is hypothesisresults in decreased computational power demand sincethere is no need to include deformable structural parts orperform cumbersome fluid-structure interaction (FSI)calculations Moreover the absence of structural de-formations allows an accurate study of the influence of thetwo parameters (D and E) on the blast wave propagation

As already mentioned for the numerical calculationsa purely Eulerian approach is utilized where only the fluidpart of the model is considered (e Eulerian formulationconsiders the computational mesh fixed while the fluid(particles) moves relative to these fixed grid points (econservation laws of mass momentum and energy areexpressed in a spatial framework (e discretization is basedon cell-centred finite volume (FV) formulations in which thegoverning equations are solved in integral form (ismethod is conservative since the formulation ensures thatthe flux entering a given volume is identical to that leaving

the adjacent volume (e numerical fluxes between adjacentFVs were calculated using the Harten-Lax-van Leer-contact(HLLC) Reimann solver [28]

(e fluid domain was modelled with 3D cubic finitevolumes (CUVF) of 01 meter edge size (e size of the cubicvolumes is selected after a convergence parametric study Onthe extremity surfaces of the two different zones (yellow lineson Figure 6) absorbing boundaries have been introduced inorder to simulate nonreflecting atmospheric boundariesSuitable shell elements (CL3D) that allow the introduction ofan absorbing medium have been employed and froma mathematical point of view represent a supersonic outletAs already mentioned the rigid parts of the model aresimulated as reflecting boundaries For a 3D cubic FV a freesurface is considered by definition as a reflecting boundaryA simple duplication of the fluid nodes at the area where therigid structure is located results into free FV surfaces thatbehave as reflecting boundaries

(e representation of the blast load is performed viathe 1D to 3D mapping technique [29] during whicha 1D numerical calculation that simulates the blast load isperformed with a very fine mesh(e data obtained from the1D calculation are mapped into a spherical shape andinserted in the coarser 3D mesh An important detail thatshould be taken into account is that when transferring theresults of the 1D analysis into the 3D mesh the blast load isconsidered spherical so the transformation should beperformed before the wave comes to contact with any ob-stacles (is mapping technique can significantly acceleratethe calculation time without sacrificing the accuracy of theresults For the representation of the detonation load inthe 1D calculation the Jones-Wilkins-Lee EOS law is used[30](e material parameters for the TNTare proposed by[31] and are included in Table 1 (e perfect gas law isemployed for the representation of the atmospheric air inthe fluid simulation (e material parameters for theatmospheric air are shown in Table 2 Finally one

Table 1 Parameters for the JWL EOS for TNT (Dobratz and Crawford [31])

Parameters Description Parameters for explosive Parameters for airA (Pa) Material constant (experimental) 37121e11 37121e11B (Pa) Material constant (experimental) 323e9 323e9R1 Material constant (experimental) 415 415R2 Material constant (experimental) 095 095Beta Material constant (experimental) 025 025RHO (kgm3) Density 1630 13EINT (Jkg) Current internal energyPer unit mass 429e6 021978e6OMEGA Specific heat ratio 03 03D (ms) Detonation velocity 6930 mdash

Table 2 Material properties for the ideal gas law for atmospheric air

Density (kgm3) Specific heat at constant volume (JkgmiddotK) Gamma ratio Reference pressure (Pa)13 71675 14 1e5

Advances in Civil Engineering 7

symmetry plane was utilized by modelling only one half ofthe uid mesh

43 Numerical Results e objective of the numerical in-vestigation is to determine the eshyectiveness of the in-troduced meandering wall Figure 7 displays the behaviourof the propagated blast wave for the ldquoopen doorrdquo case anda ldquomeanderingrdquo case (D 1m E 1m) for a plane view atthe height of 15 meters A comparison of the two cases isperformed at several time frames in order to provide a clearrepresentation on the diversion of the blast wave from theadditional wall As already mentioned half of the geometrywas modelled in the numerical calculations so in order to

present the full model in Figure 7 special mirroring tech-niques are used in the postprocessing phase e rstsubgure presents the pressure eld distribution at 3msafter the detonation At that time frame the wave front startsbeing reected on the meandering wall as it is clear from thedishyerences between the two models For the ldquomeanderingrdquocase the blast wave starts to enter the protected zone at 4msand is split into two smaller waves propagating from bothsides of themeandering wall At 8ms it is evident that in theldquoopen doorrdquo case there is one strong wave front propagatinginto the protected zone while for the ldquomeanderingrdquo casethere are two weaker wave fronts e wave propagationcontinues in a similar way for both cases until 30ms after thedetonation where the two wave fronts of the ldquomeanderingrdquo

3 ms

10 ms 15 ms

Pressure

30 ms

4 ms 8 ms

10e+05 15e+05 20e+05 25e+05 30e+05

Figure 7 Pressure distribution of the blast wave for the ldquoopen doorrdquo model and a ldquomeanderingrdquo model (D 1m E 1m) for several timeframes (Pa)

500 10 20 30 40 50 60 70

100

150

200

250

300

Time (ms)

Pres

sure

(kPa

)

Open doorMeandering

(a)

0 10 20 30 40 50 60 70Time (ms)

Impu

lse (P

amiddots)

Open doorMeandering

0

500

1000

1500

(b)

Figure 8 Comparison of (a) pressure and (b) impulse values for a point located in the protected zone for an ldquoopen doorrdquo model anda ldquomeandering wallrdquo model (D 1m E 1m)

8 Advances in Civil Engineering

case are united into one But as can be noticed even in thatphase the wave front of the ldquomeanderingrdquo case is weaker thanthe one of the ldquoopen doorrdquo case

e meandering wall reects part of the initial blast waveand guides it back to the entrance area resulting in a slightlystronger blast wave into the entrance zone Since this zone hasalready been signicantly ashyected from the initial blast wave theadditional energy due to the presence of the ldquomeanderingrdquo walldoes not ashyect the injury risk probability Figure 8 presents thepressure and the impulse history for the abovementioned casesat a point that is located at the middle of the protected zone7 meters from the opening and 15 meters from the ground

level From the pressure-time history it is clear that in theldquomeanderingrdquo case the rst peak is eliminated while the secondpeak that results from the reection of the wave onto the lateralwalls is lower Similar results can be derived from comparingthe impulse-time histories as there is a signicant reduction ofthe total impulse when the meandering wall is present

e next gures present the inuence of the dimensionand location of the meandering wall on the blast wavepropagation As has already been described in previoussections areas with an impulse value larger than 1000 Pamiddotsare considered as high injury risk zones Figure 9 presentsthe injury risk areas for the protected zone of the examined

E 05

D1

D2Impulse

30e+03

20e+03

10e+03

00e+00

D3

E 1 E 2

Figure 9 Zone of impulse higher than 1000 Pamiddots for the ldquomeanderingrdquo studied cases and the ldquoopen doorrdquo model (parameters D and E in mand impulse in pamiddots)

105 151

E (m)

D (m

)

()

2

15

2

25

3

60

40

20

60

40

20

0

Figure 10 Percentage of ashyected area spectrum versus the investigated parameters for the ldquomeanderingrdquo model

Advances in Civil Engineering 9

building for several studied geometries (e results refer tothe same horizontal cut that is presented in Figure 7 but theoutput quantity is the calculated impulse Areas charac-terized by impulse values higher than 1000 Pamiddots are high-lighted D and E determine the location and the size of themeandering wall as described earlier

In the same figure the ldquoopen doorrdquo case is also includedfor comparison purposes In the ldquoopen doorrdquo case themajority of the protected zone results in a high injury riskarea In the ldquomeanderingrdquo case the size of the high injuryrisk area varies but is always smaller than the ldquoopen doorrdquocase It is clear that the area before the additional wall isalways unprotected irrespective of the presence of themeandering wall Moreover in some cases high impulsevalues are also recorded at the sides of the meandering walldue to the relatively strong wave fronts that are present atboth those areas Figure 10 summarizes the output of theparametric study by presenting the percentage of the injuryrisk area in the protected zone (above 1000 Pamiddots) versus thetwo examined parameters From comparing the severalldquomeanderingrdquo cases it can be concluded that if D is small(distance between the door and the added wall) the influ-ence of E (dimension of overlapping length) is of minorimportance However if D is large the influence of E issignificant (if E decreases the blast wave that enters into theprotected zone is stronger and the high injury risk areas arebecoming larger)

5 Conclusions

Access control points that are located inside or attached tobuildings are often adopted due to local constraints (eirdesign must consider possible terrorist attack scenarios(eg explosions) at the interior and exterior of these accesscontrol points (e mitigation of the consequences shouldsuch a terrorist attack materialize inside an access point andin the attached building can be achieved by using venting(release surfaces) and in particular meandering walls (einvestigation shows that the reduction of the pressure issignificant if meandering walls are used and that these wallscan therefore play an important role in separating accesscontrol points from the attached buildings Nevertheless it isobvious that meandering walls can be considered an effectiverisk reduction measure and not as a full protection solution(e risk reduction concerns both blast wave effects and forobvious reasons fragments resulting from explosions andshooting attacks Herein the idealized situation that is in-vestigated can be extended to include additional parametersAs has been presented in the current article the presence ofa rigid wall standing between an explosionrsquos detonationcentre and a target can effectively reduce the severity of theimpacting blast wave It is therefore recommended theaddition of such elements (varying in size and geometry) inorder to mitigate the risk of terrorist attacks

Conflicts of Interest

(e authors declare that there are no conflicts of interestregarding the publication of this paper

References

[1] A M Lewis ldquoState of the art report on European legislationrelating to explosives and explosive detection system for non-aviation configurationsrdquo Tech Rep JRC84076 Joint ResearchCentre Ispra Italy 2013

[2] B M Luccioni R D Ambrosini and R F Danesi ldquoAnalysisof building collapse under blast loadsrdquo Engineering Structuresvol 26 no 1 pp 63ndash71 2004

[3] M Larcher M Arrigoni C Bedon et al ldquoDesign of blast-loaded glazing windows and facades a review of essentialrequirements towards standardizationrdquo Advances in CivilEngineering vol 2016 Article ID 2604232 14 pages 2016

[4] R Hajek and M Foglar ldquoNumerical and experimentalanalysis of the effect of rigid barriers on blast wave propa-gationrdquo Journal of Structural Engineering vol 141 no 12p 04015061 2015

[5] C N Kingery and G Bulmash ldquoAirblast parameters fromTNT spherical air burst and hemispherical surface burstrdquoTech Rep ARBRL-TR-02555 Defence Technical InformationCenter Ballistic Research Laboratory Aberdeen ProvingGround MD USA Tech Rep ARBRL-TR-02555 1984

[6] M Larcher ldquoPressure-time functions for the description of airblast wavesrdquo Tech Rep JRC46829 Joint Research CentreIspra Italy 2008

[7] V Karlos and G Solomos ldquoCalculation of blast loads forapplication to structural componentsrdquo Tech Rep EUR26456EN Joint Research Centre Publications Office of theEuropean Union Luxembourg 2013

[8] V Karlos G Solomos and M Larcher ldquoAnalysis of the blastwave decay coefficient using the Kingery-Bulmash datardquoInternational Journal of Protective Structures vol 7 no 3pp 409ndash429 2016

[9] M Larcher ldquoDevelopment of discrete cracks in concreteloaded by shock wavesrdquo International Journal for ImpactEngineering vol 36 pp 700ndash710 2009

[10] J Li and H Hao ldquoInternal and external pressure prediction ofvented gas explosion in large rooms by using analytical andCFD methodsrdquo Journal of Loss Prevention in the ProcessIndustries vol 49 pp 367ndash381 2017

[11] EN14494 ldquoGas explosion venting protective systemsrdquo 2007[12] M Larcher F Casadei and G Solomos ldquoInfluence of venting

areas on the air blast pressure inside tubular structures likerailway carriagesrdquo Journal of Hazardous Materials vol 183no 1-3 pp 839ndash846 2010

[13] P J Miller ldquoA reactive flow model with coupled reactionkinetics for detonation and combustion in non-ideal explo-sivesrdquo MRS Proceedings vol 418 1995

[14] M Larcher R Forsberg U Bjoernstig A Holgersson andG Solomos ldquoEffectiveness of finite element modelling ofdamage and injuries for explosions inside trainsrdquo Journal ofTransportation Safety and Security vol 8 pp 83ndash100 2015

[15] G Valsamos F Casadei M Larcher and G SolomosldquoImplementation of flying debris fatal risk calculation inEUROPLEXUSrdquo Tech Rep JRC94805 Joint Research CentrePublications Office of the European Union Luxembourg 2015

[16] E Gonzalez Ferradas F Dıaz Alonso M Doval MintildearroA Mintildeana Aznar J Ruiz Gimeno and J Sanchez PerezldquoConsequence analysis by means of characteristic curves todetermine the damage to humans from bursting sphericalvesselsrdquo Process Safety and Environmental Protection vol 86no 2 pp 121ndash129 2008

10 Advances in Civil Engineering

[17] S Mannan Leesrsquo Loss Prevention in the Process IndustriesHazard Identification Assessment and Control Vol 2 ElsevierAmsterdam 2005

[18] Department of Defence USA Unified Facilities Criteria UFC3-340ndash2 2008

[19] ldquoNATO STANAG 2280 design threat levels and handoverprocedure for temporary protective structuresrdquo 2008

[20] G F Kinney and K J Graham Explosive Shocks in AirSpringer Berlin Germany 1985

[21] D Bogosian J Ferritto and Y Shi ldquoMeasuring uncertaintyand conservatism in simplified blast modelsrdquo in Proceedingsof 30th Explosive Safety Seminar Atlanta Georgia August 2002

[22] M Larcher and F Casadei ldquoExplosions in complexgeometriesndasha comparison of several approachesrdquo InternationalJournal of Protective Structures vol 1 no 2 pp 169ndash195 2010

[23] M Larcher A Stolz L Rudiger and M Steyerer ldquoStatischeErsatzlasten und ldquoIngenieurverfahrenrdquordquo in Berichte ausdem Konstruktiven Ingenieurbau Workshop Bau-ProtectN Gebbeken K (oma M Rudiger and M Klaus Edspp 117ndash130 Universitat der Bundeswehr MunchenNeubiberg Germany 2012

[24] FEMA ldquoSite and urban design for securityndashguidance againstpotential terrorist attacksrdquo Tech Rep FEMA 430 FEMAWashington DC USA 2007

[25] Joint Research Centre and Commissariat a lrsquoenergie atomiqueet aux energies alternatives EUROPLEXUS httpwww-epxceafr

[26] Commissariat a lrsquoenergie atomique et aux energies alterna-tives CAST3M httpwww-cast3mceafr

[27] ParaView httpwwwparavieworg[28] F Daude and P Galon ldquoOn the computation of the baerndash

nunziato model using ALE formulation with HLL-and HLLC-type solvers towards fluidndashstructure interactionsrdquo Journal ofComputational Physics vol 324 pp 189ndash230 2016

[29] M Larcher F Casadei and G Solomos ldquoSimulation of blastwaves by usingmapping technology in EUROPLEXUSrdquo TechRep JRC91102 Joint Research Centre Publications Office ofthe European Union Luxembourg 2014

[30] E Lee J Horning and J Kury ldquoAdiabatic expansion of highexplosives detonation productsrdquo University of CaliforniaLawrence Livermore National Laboratory Livermore CAUSA 1968

[31] B Dobratz and P Crawford LLNL Explosives HandbookProperties of Chemical Explosives and Explosive SimulantsUniversity of California Lawrence Livermore NationalLaboratory Livermore CA USA 1985

Advances in Civil Engineering 11

24 Vulnerability of Humans e eshyects of an explosion onhuman body can be distinguished into three broad cate-gories (Figure 5) Concerning explosion-induced injuriesthe risk due to the primary blast eects on humans has al-ready been precisely described by several authors [14]However this is not yet the case for secondary blast eectsthat are related to the produced fragments either from theexplosive charge or from the failure of structures that arepropelled with high velocities striking the human body Astraightforward procedure for assessing the additional riskdue to the presence of splinters as part of the secondary blasteshyects is currently under development [15]

In this work the formulation of the human injury riskmodel is based on the work of Gonzalez Ferradas et al [16]andMannan [17] It utilizes the peak overpressure pmax (in Pa)and the positive impulse I (in Pamiddots) to determine the proba-bility of eardrum rupture and the probability of death

241 Causes of Death Regarding possible causes of deaththree dishyerent cases are considered head impact wholebody impact and lung haemorrhage Using a PI diagram[14] it has been proven that the head impact is the dominantcause of death and its probit function can be dened as

0

2000

4000

6000

8000

10000

Impu

lse (P

amiddots)

00E

+0

10E

+5

20E

+5

30E

+5

40E

+5

50E

+5

60E

+5

70E

+5

80E

+5

90E

+5

10E

+6

Overpressure (Pa)

5 vulnerability50 vulnerability

95 vulnerability1000 Pamiddots limit

Figure 4 Vulnerability of humans (head impact)

Primary effects Blast

Secondary effects Fragments

Tertiary effects Impact on structures

Figure 5 Blast eshyects of an explosive charge (yellow) on humans (green) due to the created blast wave (red) and fragmentation (brown)

4 Advances in Civil Engineering

Y1 5minus 849ln2430pmax

+4 middot 108

pmax middot I1113888 1113889 (1)

242 Eardrum Rupture (e probit function Y4 of eardrumrupture is described according to the following equation

Y4 minus126 + 1524ln pmax( 1113857 (2)

(e probability of occurrence R (or the percentage ofthe affected population) of the corresponding injury is nextdetermined for each death-related function using thefollowing equation from Gonzalez Ferradas et al [16]which is a good approximation of the relevant cumulativenormal distribution

Ri minus325Y3i + 4876Y

2i minus 2066Yi + 27035 (3)

Human risk is presented in Figure 4 as a PI diagram fora 5 50 and 95 probability of death Clearly for calculatingthe risk of life loss at every point both the peak pressure andthe impulse values are required meaning that a full fluid-structure interaction (FSI) analysis is required It is notedthat the influence of flying debris (secondary blast effects) isnot included in these formulas

(ese formulations are developed for the positive phaseof spherical or hemispherical blast waves from free air andsurface bursts respectively and can be used with both in-cident and reflected blast parameters In case of confinedexplosions the remaining static pressure results in an in-creased impulse a phenomenon that is not considered fromthe available formulas (e duration of the explosionssimulated in [16] is in general shorter than the time requiredfor the effects of venting to be evident and therefore theabove formulas can be used In order to reduce the com-plexity as an additional criterion an impulse limit of1000 Pamiddots is used to identify critical zones at the buildingrsquosinterior UFC 3-340-02 [18] proposes a limit of 1380ndash1725 kPafor lethal damage as a result of lung haemorrhage (is valueignores the head impact a fact that is also recognized in UFC3-340-02 where a much lower threshold is referenced (16 kPa)In the present study the impulse value of 1000Pamiddots is used asan indicative limit motivated by Figure 4 and taking intoaccount aminimum peak overpressure value of about 250 kPa(e impulse parameter is used since it is more appropriatefor confined explosions (e objective of this approach isto acquire qualitative results to investigate the effects ofmeandering For quantitative results the abovementionedrisk formulas should be used

3 Blast Protective Design

Protective design of critical infrastructures against terroristattacks is usually performed in several steps

(i) (e first step is establishing plausible attack sce-narios which is a joint decision of the owner orstakeholder and the designer Assessing the prob-ability of occurrence of a terrorist attack throughestimating sources of threat could be an appropriate

tool for defining and choosing possible scenariosAn indication is given for example by the NorthAtlantic Treaty Organizationrsquos (NATO) Standard-ization Agreement (STANAG) 2280 [19]

(ii) (e next step is to calculate the corresponding loadson the structure by identifying the worst case sce-nario In some cases this can be performed by usingsimple formulas for the blast wave propagation asthose proposed by Kingery and Bulmash [5] andKinney and Graham [20] or by using a series ofdiagrams included in Unified Facilities Criteria suchas UFC 3-340-2 [18] Bogosian et al [21] comparedthese equations with experimental data and showedthat in urban environments the propagation of theblast wave is not spherical and differences appear atthe calculated blast parameters For such compli-cated geometries numerical simulations may benecessary to assess with greater precision theloading history of a structure as described in [22]

(iii) Subsequently the protection strategy against thedetermined blast load according to the individualstructural characteristics of the examined build-ing (eg windows or facades) should be definedSeveral options validated by numerical or ex-perimental methods can be adopted as de-scribed in [23]

Among the various structural and nonstructural elementswindows and facades are considered critical when facing anexplosion due to their extremely fragile nature In additiona progressive collapse mechanismmust always be prevented asit could lead to a large death toll In general the explosive energydecreases rapidly by increasing the standoff distance Henceemploying perimeter protection measures and restricting theaccess to a site constitute one of the most effective methods fordecreasing its vulnerability For example protection againstvehicle access can be achieved by the application of variousbarrier systems as shown in [24]

31 Access Point Design for Building Structures Dependingon the security level of a site access control measures may alsobe employed for the entrance of the public (e control zoneshould be placed in an area which should be resistant to bothexterior and interior explosive events Long queuing shouldbe avoided to minimize the likelihood of the access pointbecoming a target of a terrorist attack In case of a blast flyingdebris from the failed parts of the structure (in particularfrom glass failure) may cause extensive injuries or fatalities(erefore incorporating venting in the form of release sur-faces is important for mitigating the effects of a blast inparticular if the access zone is inside the main building orattached to it (e release surfaces should be light to be easilypushed aside by the propagating blast wave or fragile toinstantly fail under the blast pressure To prevent peoplersquosinjury from the potential fragments of the failed opening therelease surfaces are often located at the roof of the structureNevertheless the effects of venting should not be over-estimated (e release surfaces (whose size is up to 10 of the

Advances in Civil Engineering 5

roomrsquos size) have a very small impact on the peak pressure butare eshyective for reducing the remaining static pressure afterthe explosion Clearly the proper design of release surfaces isa challenging task since no clear indications exist in the openliterature Usually it is advised to follow a design approachthat depends on the geometry of the control zone while thesize of the openings ranges from 3 to 8 of the roomrsquos(control arearsquos) size us these values serve only as an in-dication since the actual release surface size depends also onits cross section and the determined blast scenario

If the access control zone is located within or is con-nected to the main building the propagation of the pressurewave into the adjacent structure should be avoided toconne the explosive eshyects only in one area Constructionshaving very stishy or blast resistant doors are complicatedimpractical and space demanding since these doors must beduplicated As shown below a so-calledmeandering wall canachieve a similar protection level maintaining a design thatfavors the unobstructed ow of people

4 Numerical Simulations

e current section is dedicated to the analysis of a casestudy simulating an explosion within an access control zonee objective of the case study is to investigate the waya rigid (meandering) wall ashyects the propagation of the blastwave into an area of the building that needs to be protectedMoreover a parametric study on the inuence of the basicdimensions of the meandering wall on the attenuation of theproduced detonation wave is carried out e numericalsimulations are performed with the explicit FE codeEUROPLEXUS (EPX) [25] which is jointly developed by theFrench Commissariat a lrsquo Energie Atomique et aux EnergiesAlternatives (CEA) and the Joint Research Centre of the

European Commission (JRC Ispra) EPX employs sophis-ticated uid-structure interaction techniques and is suitablefor simulating fast and severe loadings on structures such asimpact and blast e CAST3M code [26] also developed byCEA was used to generate the mesh for the numericalmodels while ParaView software [27] was used for post-processing the results

41 Case Study To reduce complexity a simple geometryhas been selected for the case study A rectangular area isseparated by a rigid wall into two partse rst compartmentis considered as the entrance zone where a possible intrudercan enter carrying an improvised explosive device (IED) andwalk till the opening of the rigid wall that leads to the secondcompartment Just before the opening that connects the twoparts a thorough control of all persons entering the secondpart is considered to be implemented e second com-partment is regarded a vulnerable zone that needs to beprotected as a possible blast wave can become critical for thepersonnel or the structural integrity of the building

Figure 6 shows the geometrical details of the investigatedarea e entrance zone located on the left is connected tothe protected zone on the right through a 3m wideopening e openingrsquos width guarantees a relatively highow of the people to avoid queuinge height of both zonesis 3 meters while the detonation centre is located 15 metersfrom the ground level (average value for the location ofa backpack in the vertical axis) and is centred to the openingbased on a worst case scenario is is assumed to be theclosest point to the protected zone that a perpetrator canreach prior to being intercepted e mass of the IED hasbeen selected to be equal to 25 kilograms of TNTequivalentas an upper limit for the amount of explosives a pedestriancould carry in a backpack or small suitcase

10 mEntrance zone

Explosive material

Reflecting boundaries

Absorbing boundaries

Protected zone

Blocking wall (meandering)

2 m

3 m

E

D

10 m 20 mRigid wall

Figure 6 Geometry of the case study model

6 Advances in Civil Engineering

As shown in Figure 6 the distance D between theopening and the introduced meandering wall and the di-mension E that refers to the overlapping between the ad-ditional blocking wall and the preexisting rigid wall are notexplicitly determined (ese two parameters are selected asvariables in the performed parametric study in order toquantify their influence on the blast wave propagationD ranges from 1 to 3 meters corresponding to constructionspecifications (eg wheelchair access and personnel flow)Similarly E ranges from 05 to 2 meters to allow a minimumdistance of 15 meters between each extremity of themeandering wall and the lateral walls of the building and toguarantee comfortable access to the attached protected zone

42NumericalModel (e structural parts of the building inthe case study are assumed rigid meaning that the walls theceiling and the floor of the structure are nondeformableA postanalysis investigation of the recorded maximum pres-sures indicated that thick stiff structural elements like concreteor steel walls respect this basic assumption (is hypothesisresults in decreased computational power demand sincethere is no need to include deformable structural parts orperform cumbersome fluid-structure interaction (FSI)calculations Moreover the absence of structural de-formations allows an accurate study of the influence of thetwo parameters (D and E) on the blast wave propagation

As already mentioned for the numerical calculationsa purely Eulerian approach is utilized where only the fluidpart of the model is considered (e Eulerian formulationconsiders the computational mesh fixed while the fluid(particles) moves relative to these fixed grid points (econservation laws of mass momentum and energy areexpressed in a spatial framework (e discretization is basedon cell-centred finite volume (FV) formulations in which thegoverning equations are solved in integral form (ismethod is conservative since the formulation ensures thatthe flux entering a given volume is identical to that leaving

the adjacent volume (e numerical fluxes between adjacentFVs were calculated using the Harten-Lax-van Leer-contact(HLLC) Reimann solver [28]

(e fluid domain was modelled with 3D cubic finitevolumes (CUVF) of 01 meter edge size (e size of the cubicvolumes is selected after a convergence parametric study Onthe extremity surfaces of the two different zones (yellow lineson Figure 6) absorbing boundaries have been introduced inorder to simulate nonreflecting atmospheric boundariesSuitable shell elements (CL3D) that allow the introduction ofan absorbing medium have been employed and froma mathematical point of view represent a supersonic outletAs already mentioned the rigid parts of the model aresimulated as reflecting boundaries For a 3D cubic FV a freesurface is considered by definition as a reflecting boundaryA simple duplication of the fluid nodes at the area where therigid structure is located results into free FV surfaces thatbehave as reflecting boundaries

(e representation of the blast load is performed viathe 1D to 3D mapping technique [29] during whicha 1D numerical calculation that simulates the blast load isperformed with a very fine mesh(e data obtained from the1D calculation are mapped into a spherical shape andinserted in the coarser 3D mesh An important detail thatshould be taken into account is that when transferring theresults of the 1D analysis into the 3D mesh the blast load isconsidered spherical so the transformation should beperformed before the wave comes to contact with any ob-stacles (is mapping technique can significantly acceleratethe calculation time without sacrificing the accuracy of theresults For the representation of the detonation load inthe 1D calculation the Jones-Wilkins-Lee EOS law is used[30](e material parameters for the TNTare proposed by[31] and are included in Table 1 (e perfect gas law isemployed for the representation of the atmospheric air inthe fluid simulation (e material parameters for theatmospheric air are shown in Table 2 Finally one

Table 1 Parameters for the JWL EOS for TNT (Dobratz and Crawford [31])

Parameters Description Parameters for explosive Parameters for airA (Pa) Material constant (experimental) 37121e11 37121e11B (Pa) Material constant (experimental) 323e9 323e9R1 Material constant (experimental) 415 415R2 Material constant (experimental) 095 095Beta Material constant (experimental) 025 025RHO (kgm3) Density 1630 13EINT (Jkg) Current internal energyPer unit mass 429e6 021978e6OMEGA Specific heat ratio 03 03D (ms) Detonation velocity 6930 mdash

Table 2 Material properties for the ideal gas law for atmospheric air

Density (kgm3) Specific heat at constant volume (JkgmiddotK) Gamma ratio Reference pressure (Pa)13 71675 14 1e5

Advances in Civil Engineering 7

symmetry plane was utilized by modelling only one half ofthe uid mesh

43 Numerical Results e objective of the numerical in-vestigation is to determine the eshyectiveness of the in-troduced meandering wall Figure 7 displays the behaviourof the propagated blast wave for the ldquoopen doorrdquo case anda ldquomeanderingrdquo case (D 1m E 1m) for a plane view atthe height of 15 meters A comparison of the two cases isperformed at several time frames in order to provide a clearrepresentation on the diversion of the blast wave from theadditional wall As already mentioned half of the geometrywas modelled in the numerical calculations so in order to

present the full model in Figure 7 special mirroring tech-niques are used in the postprocessing phase e rstsubgure presents the pressure eld distribution at 3msafter the detonation At that time frame the wave front startsbeing reected on the meandering wall as it is clear from thedishyerences between the two models For the ldquomeanderingrdquocase the blast wave starts to enter the protected zone at 4msand is split into two smaller waves propagating from bothsides of themeandering wall At 8ms it is evident that in theldquoopen doorrdquo case there is one strong wave front propagatinginto the protected zone while for the ldquomeanderingrdquo casethere are two weaker wave fronts e wave propagationcontinues in a similar way for both cases until 30ms after thedetonation where the two wave fronts of the ldquomeanderingrdquo

3 ms

10 ms 15 ms

Pressure

30 ms

4 ms 8 ms

10e+05 15e+05 20e+05 25e+05 30e+05

Figure 7 Pressure distribution of the blast wave for the ldquoopen doorrdquo model and a ldquomeanderingrdquo model (D 1m E 1m) for several timeframes (Pa)

500 10 20 30 40 50 60 70

100

150

200

250

300

Time (ms)

Pres

sure

(kPa

)

Open doorMeandering

(a)

0 10 20 30 40 50 60 70Time (ms)

Impu

lse (P

amiddots)

Open doorMeandering

0

500

1000

1500

(b)

Figure 8 Comparison of (a) pressure and (b) impulse values for a point located in the protected zone for an ldquoopen doorrdquo model anda ldquomeandering wallrdquo model (D 1m E 1m)

8 Advances in Civil Engineering

case are united into one But as can be noticed even in thatphase the wave front of the ldquomeanderingrdquo case is weaker thanthe one of the ldquoopen doorrdquo case

e meandering wall reects part of the initial blast waveand guides it back to the entrance area resulting in a slightlystronger blast wave into the entrance zone Since this zone hasalready been signicantly ashyected from the initial blast wave theadditional energy due to the presence of the ldquomeanderingrdquo walldoes not ashyect the injury risk probability Figure 8 presents thepressure and the impulse history for the abovementioned casesat a point that is located at the middle of the protected zone7 meters from the opening and 15 meters from the ground

level From the pressure-time history it is clear that in theldquomeanderingrdquo case the rst peak is eliminated while the secondpeak that results from the reection of the wave onto the lateralwalls is lower Similar results can be derived from comparingthe impulse-time histories as there is a signicant reduction ofthe total impulse when the meandering wall is present

e next gures present the inuence of the dimensionand location of the meandering wall on the blast wavepropagation As has already been described in previoussections areas with an impulse value larger than 1000 Pamiddotsare considered as high injury risk zones Figure 9 presentsthe injury risk areas for the protected zone of the examined

E 05

D1

D2Impulse

30e+03

20e+03

10e+03

00e+00

D3

E 1 E 2

Figure 9 Zone of impulse higher than 1000 Pamiddots for the ldquomeanderingrdquo studied cases and the ldquoopen doorrdquo model (parameters D and E in mand impulse in pamiddots)

105 151

E (m)

D (m

)

()

2

15

2

25

3

60

40

20

60

40

20

0

Figure 10 Percentage of ashyected area spectrum versus the investigated parameters for the ldquomeanderingrdquo model

Advances in Civil Engineering 9

building for several studied geometries (e results refer tothe same horizontal cut that is presented in Figure 7 but theoutput quantity is the calculated impulse Areas charac-terized by impulse values higher than 1000 Pamiddots are high-lighted D and E determine the location and the size of themeandering wall as described earlier

In the same figure the ldquoopen doorrdquo case is also includedfor comparison purposes In the ldquoopen doorrdquo case themajority of the protected zone results in a high injury riskarea In the ldquomeanderingrdquo case the size of the high injuryrisk area varies but is always smaller than the ldquoopen doorrdquocase It is clear that the area before the additional wall isalways unprotected irrespective of the presence of themeandering wall Moreover in some cases high impulsevalues are also recorded at the sides of the meandering walldue to the relatively strong wave fronts that are present atboth those areas Figure 10 summarizes the output of theparametric study by presenting the percentage of the injuryrisk area in the protected zone (above 1000 Pamiddots) versus thetwo examined parameters From comparing the severalldquomeanderingrdquo cases it can be concluded that if D is small(distance between the door and the added wall) the influ-ence of E (dimension of overlapping length) is of minorimportance However if D is large the influence of E issignificant (if E decreases the blast wave that enters into theprotected zone is stronger and the high injury risk areas arebecoming larger)

5 Conclusions

Access control points that are located inside or attached tobuildings are often adopted due to local constraints (eirdesign must consider possible terrorist attack scenarios(eg explosions) at the interior and exterior of these accesscontrol points (e mitigation of the consequences shouldsuch a terrorist attack materialize inside an access point andin the attached building can be achieved by using venting(release surfaces) and in particular meandering walls (einvestigation shows that the reduction of the pressure issignificant if meandering walls are used and that these wallscan therefore play an important role in separating accesscontrol points from the attached buildings Nevertheless it isobvious that meandering walls can be considered an effectiverisk reduction measure and not as a full protection solution(e risk reduction concerns both blast wave effects and forobvious reasons fragments resulting from explosions andshooting attacks Herein the idealized situation that is in-vestigated can be extended to include additional parametersAs has been presented in the current article the presence ofa rigid wall standing between an explosionrsquos detonationcentre and a target can effectively reduce the severity of theimpacting blast wave It is therefore recommended theaddition of such elements (varying in size and geometry) inorder to mitigate the risk of terrorist attacks

Conflicts of Interest

(e authors declare that there are no conflicts of interestregarding the publication of this paper

References

[1] A M Lewis ldquoState of the art report on European legislationrelating to explosives and explosive detection system for non-aviation configurationsrdquo Tech Rep JRC84076 Joint ResearchCentre Ispra Italy 2013

[2] B M Luccioni R D Ambrosini and R F Danesi ldquoAnalysisof building collapse under blast loadsrdquo Engineering Structuresvol 26 no 1 pp 63ndash71 2004

[3] M Larcher M Arrigoni C Bedon et al ldquoDesign of blast-loaded glazing windows and facades a review of essentialrequirements towards standardizationrdquo Advances in CivilEngineering vol 2016 Article ID 2604232 14 pages 2016

[4] R Hajek and M Foglar ldquoNumerical and experimentalanalysis of the effect of rigid barriers on blast wave propa-gationrdquo Journal of Structural Engineering vol 141 no 12p 04015061 2015

[5] C N Kingery and G Bulmash ldquoAirblast parameters fromTNT spherical air burst and hemispherical surface burstrdquoTech Rep ARBRL-TR-02555 Defence Technical InformationCenter Ballistic Research Laboratory Aberdeen ProvingGround MD USA Tech Rep ARBRL-TR-02555 1984

[6] M Larcher ldquoPressure-time functions for the description of airblast wavesrdquo Tech Rep JRC46829 Joint Research CentreIspra Italy 2008

[7] V Karlos and G Solomos ldquoCalculation of blast loads forapplication to structural componentsrdquo Tech Rep EUR26456EN Joint Research Centre Publications Office of theEuropean Union Luxembourg 2013

[8] V Karlos G Solomos and M Larcher ldquoAnalysis of the blastwave decay coefficient using the Kingery-Bulmash datardquoInternational Journal of Protective Structures vol 7 no 3pp 409ndash429 2016

[9] M Larcher ldquoDevelopment of discrete cracks in concreteloaded by shock wavesrdquo International Journal for ImpactEngineering vol 36 pp 700ndash710 2009

[10] J Li and H Hao ldquoInternal and external pressure prediction ofvented gas explosion in large rooms by using analytical andCFD methodsrdquo Journal of Loss Prevention in the ProcessIndustries vol 49 pp 367ndash381 2017

[11] EN14494 ldquoGas explosion venting protective systemsrdquo 2007[12] M Larcher F Casadei and G Solomos ldquoInfluence of venting

areas on the air blast pressure inside tubular structures likerailway carriagesrdquo Journal of Hazardous Materials vol 183no 1-3 pp 839ndash846 2010

[13] P J Miller ldquoA reactive flow model with coupled reactionkinetics for detonation and combustion in non-ideal explo-sivesrdquo MRS Proceedings vol 418 1995

[14] M Larcher R Forsberg U Bjoernstig A Holgersson andG Solomos ldquoEffectiveness of finite element modelling ofdamage and injuries for explosions inside trainsrdquo Journal ofTransportation Safety and Security vol 8 pp 83ndash100 2015

[15] G Valsamos F Casadei M Larcher and G SolomosldquoImplementation of flying debris fatal risk calculation inEUROPLEXUSrdquo Tech Rep JRC94805 Joint Research CentrePublications Office of the European Union Luxembourg 2015

[16] E Gonzalez Ferradas F Dıaz Alonso M Doval MintildearroA Mintildeana Aznar J Ruiz Gimeno and J Sanchez PerezldquoConsequence analysis by means of characteristic curves todetermine the damage to humans from bursting sphericalvesselsrdquo Process Safety and Environmental Protection vol 86no 2 pp 121ndash129 2008

10 Advances in Civil Engineering

[17] S Mannan Leesrsquo Loss Prevention in the Process IndustriesHazard Identification Assessment and Control Vol 2 ElsevierAmsterdam 2005

[18] Department of Defence USA Unified Facilities Criteria UFC3-340ndash2 2008

[19] ldquoNATO STANAG 2280 design threat levels and handoverprocedure for temporary protective structuresrdquo 2008

[20] G F Kinney and K J Graham Explosive Shocks in AirSpringer Berlin Germany 1985

[21] D Bogosian J Ferritto and Y Shi ldquoMeasuring uncertaintyand conservatism in simplified blast modelsrdquo in Proceedingsof 30th Explosive Safety Seminar Atlanta Georgia August 2002

[22] M Larcher and F Casadei ldquoExplosions in complexgeometriesndasha comparison of several approachesrdquo InternationalJournal of Protective Structures vol 1 no 2 pp 169ndash195 2010

[23] M Larcher A Stolz L Rudiger and M Steyerer ldquoStatischeErsatzlasten und ldquoIngenieurverfahrenrdquordquo in Berichte ausdem Konstruktiven Ingenieurbau Workshop Bau-ProtectN Gebbeken K (oma M Rudiger and M Klaus Edspp 117ndash130 Universitat der Bundeswehr MunchenNeubiberg Germany 2012

[24] FEMA ldquoSite and urban design for securityndashguidance againstpotential terrorist attacksrdquo Tech Rep FEMA 430 FEMAWashington DC USA 2007

[25] Joint Research Centre and Commissariat a lrsquoenergie atomiqueet aux energies alternatives EUROPLEXUS httpwww-epxceafr

[26] Commissariat a lrsquoenergie atomique et aux energies alterna-tives CAST3M httpwww-cast3mceafr

[27] ParaView httpwwwparavieworg[28] F Daude and P Galon ldquoOn the computation of the baerndash

nunziato model using ALE formulation with HLL-and HLLC-type solvers towards fluidndashstructure interactionsrdquo Journal ofComputational Physics vol 324 pp 189ndash230 2016

[29] M Larcher F Casadei and G Solomos ldquoSimulation of blastwaves by usingmapping technology in EUROPLEXUSrdquo TechRep JRC91102 Joint Research Centre Publications Office ofthe European Union Luxembourg 2014

[30] E Lee J Horning and J Kury ldquoAdiabatic expansion of highexplosives detonation productsrdquo University of CaliforniaLawrence Livermore National Laboratory Livermore CAUSA 1968

[31] B Dobratz and P Crawford LLNL Explosives HandbookProperties of Chemical Explosives and Explosive SimulantsUniversity of California Lawrence Livermore NationalLaboratory Livermore CA USA 1985

Advances in Civil Engineering 11

Y1 5minus 849ln2430pmax

+4 middot 108

pmax middot I1113888 1113889 (1)

242 Eardrum Rupture (e probit function Y4 of eardrumrupture is described according to the following equation

Y4 minus126 + 1524ln pmax( 1113857 (2)

(e probability of occurrence R (or the percentage ofthe affected population) of the corresponding injury is nextdetermined for each death-related function using thefollowing equation from Gonzalez Ferradas et al [16]which is a good approximation of the relevant cumulativenormal distribution

Ri minus325Y3i + 4876Y

2i minus 2066Yi + 27035 (3)

Human risk is presented in Figure 4 as a PI diagram fora 5 50 and 95 probability of death Clearly for calculatingthe risk of life loss at every point both the peak pressure andthe impulse values are required meaning that a full fluid-structure interaction (FSI) analysis is required It is notedthat the influence of flying debris (secondary blast effects) isnot included in these formulas

(ese formulations are developed for the positive phaseof spherical or hemispherical blast waves from free air andsurface bursts respectively and can be used with both in-cident and reflected blast parameters In case of confinedexplosions the remaining static pressure results in an in-creased impulse a phenomenon that is not considered fromthe available formulas (e duration of the explosionssimulated in [16] is in general shorter than the time requiredfor the effects of venting to be evident and therefore theabove formulas can be used In order to reduce the com-plexity as an additional criterion an impulse limit of1000 Pamiddots is used to identify critical zones at the buildingrsquosinterior UFC 3-340-02 [18] proposes a limit of 1380ndash1725 kPafor lethal damage as a result of lung haemorrhage (is valueignores the head impact a fact that is also recognized in UFC3-340-02 where a much lower threshold is referenced (16 kPa)In the present study the impulse value of 1000Pamiddots is used asan indicative limit motivated by Figure 4 and taking intoaccount aminimum peak overpressure value of about 250 kPa(e impulse parameter is used since it is more appropriatefor confined explosions (e objective of this approach isto acquire qualitative results to investigate the effects ofmeandering For quantitative results the abovementionedrisk formulas should be used

3 Blast Protective Design

Protective design of critical infrastructures against terroristattacks is usually performed in several steps

(i) (e first step is establishing plausible attack sce-narios which is a joint decision of the owner orstakeholder and the designer Assessing the prob-ability of occurrence of a terrorist attack throughestimating sources of threat could be an appropriate

tool for defining and choosing possible scenariosAn indication is given for example by the NorthAtlantic Treaty Organizationrsquos (NATO) Standard-ization Agreement (STANAG) 2280 [19]

(ii) (e next step is to calculate the corresponding loadson the structure by identifying the worst case sce-nario In some cases this can be performed by usingsimple formulas for the blast wave propagation asthose proposed by Kingery and Bulmash [5] andKinney and Graham [20] or by using a series ofdiagrams included in Unified Facilities Criteria suchas UFC 3-340-2 [18] Bogosian et al [21] comparedthese equations with experimental data and showedthat in urban environments the propagation of theblast wave is not spherical and differences appear atthe calculated blast parameters For such compli-cated geometries numerical simulations may benecessary to assess with greater precision theloading history of a structure as described in [22]

(iii) Subsequently the protection strategy against thedetermined blast load according to the individualstructural characteristics of the examined build-ing (eg windows or facades) should be definedSeveral options validated by numerical or ex-perimental methods can be adopted as de-scribed in [23]

Among the various structural and nonstructural elementswindows and facades are considered critical when facing anexplosion due to their extremely fragile nature In additiona progressive collapse mechanismmust always be prevented asit could lead to a large death toll In general the explosive energydecreases rapidly by increasing the standoff distance Henceemploying perimeter protection measures and restricting theaccess to a site constitute one of the most effective methods fordecreasing its vulnerability For example protection againstvehicle access can be achieved by the application of variousbarrier systems as shown in [24]

31 Access Point Design for Building Structures Dependingon the security level of a site access control measures may alsobe employed for the entrance of the public (e control zoneshould be placed in an area which should be resistant to bothexterior and interior explosive events Long queuing shouldbe avoided to minimize the likelihood of the access pointbecoming a target of a terrorist attack In case of a blast flyingdebris from the failed parts of the structure (in particularfrom glass failure) may cause extensive injuries or fatalities(erefore incorporating venting in the form of release sur-faces is important for mitigating the effects of a blast inparticular if the access zone is inside the main building orattached to it (e release surfaces should be light to be easilypushed aside by the propagating blast wave or fragile toinstantly fail under the blast pressure To prevent peoplersquosinjury from the potential fragments of the failed opening therelease surfaces are often located at the roof of the structureNevertheless the effects of venting should not be over-estimated (e release surfaces (whose size is up to 10 of the

Advances in Civil Engineering 5

roomrsquos size) have a very small impact on the peak pressure butare eshyective for reducing the remaining static pressure afterthe explosion Clearly the proper design of release surfaces isa challenging task since no clear indications exist in the openliterature Usually it is advised to follow a design approachthat depends on the geometry of the control zone while thesize of the openings ranges from 3 to 8 of the roomrsquos(control arearsquos) size us these values serve only as an in-dication since the actual release surface size depends also onits cross section and the determined blast scenario

If the access control zone is located within or is con-nected to the main building the propagation of the pressurewave into the adjacent structure should be avoided toconne the explosive eshyects only in one area Constructionshaving very stishy or blast resistant doors are complicatedimpractical and space demanding since these doors must beduplicated As shown below a so-calledmeandering wall canachieve a similar protection level maintaining a design thatfavors the unobstructed ow of people

4 Numerical Simulations

e current section is dedicated to the analysis of a casestudy simulating an explosion within an access control zonee objective of the case study is to investigate the waya rigid (meandering) wall ashyects the propagation of the blastwave into an area of the building that needs to be protectedMoreover a parametric study on the inuence of the basicdimensions of the meandering wall on the attenuation of theproduced detonation wave is carried out e numericalsimulations are performed with the explicit FE codeEUROPLEXUS (EPX) [25] which is jointly developed by theFrench Commissariat a lrsquo Energie Atomique et aux EnergiesAlternatives (CEA) and the Joint Research Centre of the

European Commission (JRC Ispra) EPX employs sophis-ticated uid-structure interaction techniques and is suitablefor simulating fast and severe loadings on structures such asimpact and blast e CAST3M code [26] also developed byCEA was used to generate the mesh for the numericalmodels while ParaView software [27] was used for post-processing the results

41 Case Study To reduce complexity a simple geometryhas been selected for the case study A rectangular area isseparated by a rigid wall into two partse rst compartmentis considered as the entrance zone where a possible intrudercan enter carrying an improvised explosive device (IED) andwalk till the opening of the rigid wall that leads to the secondcompartment Just before the opening that connects the twoparts a thorough control of all persons entering the secondpart is considered to be implemented e second com-partment is regarded a vulnerable zone that needs to beprotected as a possible blast wave can become critical for thepersonnel or the structural integrity of the building

Figure 6 shows the geometrical details of the investigatedarea e entrance zone located on the left is connected tothe protected zone on the right through a 3m wideopening e openingrsquos width guarantees a relatively highow of the people to avoid queuinge height of both zonesis 3 meters while the detonation centre is located 15 metersfrom the ground level (average value for the location ofa backpack in the vertical axis) and is centred to the openingbased on a worst case scenario is is assumed to be theclosest point to the protected zone that a perpetrator canreach prior to being intercepted e mass of the IED hasbeen selected to be equal to 25 kilograms of TNTequivalentas an upper limit for the amount of explosives a pedestriancould carry in a backpack or small suitcase

10 mEntrance zone

Explosive material

Reflecting boundaries

Absorbing boundaries

Protected zone

Blocking wall (meandering)

2 m

3 m

E

D

10 m 20 mRigid wall

Figure 6 Geometry of the case study model

6 Advances in Civil Engineering

As shown in Figure 6 the distance D between theopening and the introduced meandering wall and the di-mension E that refers to the overlapping between the ad-ditional blocking wall and the preexisting rigid wall are notexplicitly determined (ese two parameters are selected asvariables in the performed parametric study in order toquantify their influence on the blast wave propagationD ranges from 1 to 3 meters corresponding to constructionspecifications (eg wheelchair access and personnel flow)Similarly E ranges from 05 to 2 meters to allow a minimumdistance of 15 meters between each extremity of themeandering wall and the lateral walls of the building and toguarantee comfortable access to the attached protected zone

42NumericalModel (e structural parts of the building inthe case study are assumed rigid meaning that the walls theceiling and the floor of the structure are nondeformableA postanalysis investigation of the recorded maximum pres-sures indicated that thick stiff structural elements like concreteor steel walls respect this basic assumption (is hypothesisresults in decreased computational power demand sincethere is no need to include deformable structural parts orperform cumbersome fluid-structure interaction (FSI)calculations Moreover the absence of structural de-formations allows an accurate study of the influence of thetwo parameters (D and E) on the blast wave propagation

As already mentioned for the numerical calculationsa purely Eulerian approach is utilized where only the fluidpart of the model is considered (e Eulerian formulationconsiders the computational mesh fixed while the fluid(particles) moves relative to these fixed grid points (econservation laws of mass momentum and energy areexpressed in a spatial framework (e discretization is basedon cell-centred finite volume (FV) formulations in which thegoverning equations are solved in integral form (ismethod is conservative since the formulation ensures thatthe flux entering a given volume is identical to that leaving

the adjacent volume (e numerical fluxes between adjacentFVs were calculated using the Harten-Lax-van Leer-contact(HLLC) Reimann solver [28]

(e fluid domain was modelled with 3D cubic finitevolumes (CUVF) of 01 meter edge size (e size of the cubicvolumes is selected after a convergence parametric study Onthe extremity surfaces of the two different zones (yellow lineson Figure 6) absorbing boundaries have been introduced inorder to simulate nonreflecting atmospheric boundariesSuitable shell elements (CL3D) that allow the introduction ofan absorbing medium have been employed and froma mathematical point of view represent a supersonic outletAs already mentioned the rigid parts of the model aresimulated as reflecting boundaries For a 3D cubic FV a freesurface is considered by definition as a reflecting boundaryA simple duplication of the fluid nodes at the area where therigid structure is located results into free FV surfaces thatbehave as reflecting boundaries

(e representation of the blast load is performed viathe 1D to 3D mapping technique [29] during whicha 1D numerical calculation that simulates the blast load isperformed with a very fine mesh(e data obtained from the1D calculation are mapped into a spherical shape andinserted in the coarser 3D mesh An important detail thatshould be taken into account is that when transferring theresults of the 1D analysis into the 3D mesh the blast load isconsidered spherical so the transformation should beperformed before the wave comes to contact with any ob-stacles (is mapping technique can significantly acceleratethe calculation time without sacrificing the accuracy of theresults For the representation of the detonation load inthe 1D calculation the Jones-Wilkins-Lee EOS law is used[30](e material parameters for the TNTare proposed by[31] and are included in Table 1 (e perfect gas law isemployed for the representation of the atmospheric air inthe fluid simulation (e material parameters for theatmospheric air are shown in Table 2 Finally one

Table 1 Parameters for the JWL EOS for TNT (Dobratz and Crawford [31])

Parameters Description Parameters for explosive Parameters for airA (Pa) Material constant (experimental) 37121e11 37121e11B (Pa) Material constant (experimental) 323e9 323e9R1 Material constant (experimental) 415 415R2 Material constant (experimental) 095 095Beta Material constant (experimental) 025 025RHO (kgm3) Density 1630 13EINT (Jkg) Current internal energyPer unit mass 429e6 021978e6OMEGA Specific heat ratio 03 03D (ms) Detonation velocity 6930 mdash

Table 2 Material properties for the ideal gas law for atmospheric air

Density (kgm3) Specific heat at constant volume (JkgmiddotK) Gamma ratio Reference pressure (Pa)13 71675 14 1e5

Advances in Civil Engineering 7

symmetry plane was utilized by modelling only one half ofthe uid mesh

43 Numerical Results e objective of the numerical in-vestigation is to determine the eshyectiveness of the in-troduced meandering wall Figure 7 displays the behaviourof the propagated blast wave for the ldquoopen doorrdquo case anda ldquomeanderingrdquo case (D 1m E 1m) for a plane view atthe height of 15 meters A comparison of the two cases isperformed at several time frames in order to provide a clearrepresentation on the diversion of the blast wave from theadditional wall As already mentioned half of the geometrywas modelled in the numerical calculations so in order to

present the full model in Figure 7 special mirroring tech-niques are used in the postprocessing phase e rstsubgure presents the pressure eld distribution at 3msafter the detonation At that time frame the wave front startsbeing reected on the meandering wall as it is clear from thedishyerences between the two models For the ldquomeanderingrdquocase the blast wave starts to enter the protected zone at 4msand is split into two smaller waves propagating from bothsides of themeandering wall At 8ms it is evident that in theldquoopen doorrdquo case there is one strong wave front propagatinginto the protected zone while for the ldquomeanderingrdquo casethere are two weaker wave fronts e wave propagationcontinues in a similar way for both cases until 30ms after thedetonation where the two wave fronts of the ldquomeanderingrdquo

3 ms

10 ms 15 ms

Pressure

30 ms

4 ms 8 ms

10e+05 15e+05 20e+05 25e+05 30e+05

Figure 7 Pressure distribution of the blast wave for the ldquoopen doorrdquo model and a ldquomeanderingrdquo model (D 1m E 1m) for several timeframes (Pa)

500 10 20 30 40 50 60 70

100

150

200

250

300

Time (ms)

Pres

sure

(kPa

)

Open doorMeandering

(a)

0 10 20 30 40 50 60 70Time (ms)

Impu

lse (P

amiddots)

Open doorMeandering

0

500

1000

1500

(b)

Figure 8 Comparison of (a) pressure and (b) impulse values for a point located in the protected zone for an ldquoopen doorrdquo model anda ldquomeandering wallrdquo model (D 1m E 1m)

8 Advances in Civil Engineering

case are united into one But as can be noticed even in thatphase the wave front of the ldquomeanderingrdquo case is weaker thanthe one of the ldquoopen doorrdquo case

e meandering wall reects part of the initial blast waveand guides it back to the entrance area resulting in a slightlystronger blast wave into the entrance zone Since this zone hasalready been signicantly ashyected from the initial blast wave theadditional energy due to the presence of the ldquomeanderingrdquo walldoes not ashyect the injury risk probability Figure 8 presents thepressure and the impulse history for the abovementioned casesat a point that is located at the middle of the protected zone7 meters from the opening and 15 meters from the ground

level From the pressure-time history it is clear that in theldquomeanderingrdquo case the rst peak is eliminated while the secondpeak that results from the reection of the wave onto the lateralwalls is lower Similar results can be derived from comparingthe impulse-time histories as there is a signicant reduction ofthe total impulse when the meandering wall is present

e next gures present the inuence of the dimensionand location of the meandering wall on the blast wavepropagation As has already been described in previoussections areas with an impulse value larger than 1000 Pamiddotsare considered as high injury risk zones Figure 9 presentsthe injury risk areas for the protected zone of the examined

E 05

D1

D2Impulse

30e+03

20e+03

10e+03

00e+00

D3

E 1 E 2

Figure 9 Zone of impulse higher than 1000 Pamiddots for the ldquomeanderingrdquo studied cases and the ldquoopen doorrdquo model (parameters D and E in mand impulse in pamiddots)

105 151

E (m)

D (m

)

()

2

15

2

25

3

60

40

20

60

40

20

0

Figure 10 Percentage of ashyected area spectrum versus the investigated parameters for the ldquomeanderingrdquo model

Advances in Civil Engineering 9

building for several studied geometries (e results refer tothe same horizontal cut that is presented in Figure 7 but theoutput quantity is the calculated impulse Areas charac-terized by impulse values higher than 1000 Pamiddots are high-lighted D and E determine the location and the size of themeandering wall as described earlier

In the same figure the ldquoopen doorrdquo case is also includedfor comparison purposes In the ldquoopen doorrdquo case themajority of the protected zone results in a high injury riskarea In the ldquomeanderingrdquo case the size of the high injuryrisk area varies but is always smaller than the ldquoopen doorrdquocase It is clear that the area before the additional wall isalways unprotected irrespective of the presence of themeandering wall Moreover in some cases high impulsevalues are also recorded at the sides of the meandering walldue to the relatively strong wave fronts that are present atboth those areas Figure 10 summarizes the output of theparametric study by presenting the percentage of the injuryrisk area in the protected zone (above 1000 Pamiddots) versus thetwo examined parameters From comparing the severalldquomeanderingrdquo cases it can be concluded that if D is small(distance between the door and the added wall) the influ-ence of E (dimension of overlapping length) is of minorimportance However if D is large the influence of E issignificant (if E decreases the blast wave that enters into theprotected zone is stronger and the high injury risk areas arebecoming larger)

5 Conclusions

Access control points that are located inside or attached tobuildings are often adopted due to local constraints (eirdesign must consider possible terrorist attack scenarios(eg explosions) at the interior and exterior of these accesscontrol points (e mitigation of the consequences shouldsuch a terrorist attack materialize inside an access point andin the attached building can be achieved by using venting(release surfaces) and in particular meandering walls (einvestigation shows that the reduction of the pressure issignificant if meandering walls are used and that these wallscan therefore play an important role in separating accesscontrol points from the attached buildings Nevertheless it isobvious that meandering walls can be considered an effectiverisk reduction measure and not as a full protection solution(e risk reduction concerns both blast wave effects and forobvious reasons fragments resulting from explosions andshooting attacks Herein the idealized situation that is in-vestigated can be extended to include additional parametersAs has been presented in the current article the presence ofa rigid wall standing between an explosionrsquos detonationcentre and a target can effectively reduce the severity of theimpacting blast wave It is therefore recommended theaddition of such elements (varying in size and geometry) inorder to mitigate the risk of terrorist attacks

Conflicts of Interest

(e authors declare that there are no conflicts of interestregarding the publication of this paper

References

[1] A M Lewis ldquoState of the art report on European legislationrelating to explosives and explosive detection system for non-aviation configurationsrdquo Tech Rep JRC84076 Joint ResearchCentre Ispra Italy 2013

[2] B M Luccioni R D Ambrosini and R F Danesi ldquoAnalysisof building collapse under blast loadsrdquo Engineering Structuresvol 26 no 1 pp 63ndash71 2004

[3] M Larcher M Arrigoni C Bedon et al ldquoDesign of blast-loaded glazing windows and facades a review of essentialrequirements towards standardizationrdquo Advances in CivilEngineering vol 2016 Article ID 2604232 14 pages 2016

[4] R Hajek and M Foglar ldquoNumerical and experimentalanalysis of the effect of rigid barriers on blast wave propa-gationrdquo Journal of Structural Engineering vol 141 no 12p 04015061 2015

[5] C N Kingery and G Bulmash ldquoAirblast parameters fromTNT spherical air burst and hemispherical surface burstrdquoTech Rep ARBRL-TR-02555 Defence Technical InformationCenter Ballistic Research Laboratory Aberdeen ProvingGround MD USA Tech Rep ARBRL-TR-02555 1984

[6] M Larcher ldquoPressure-time functions for the description of airblast wavesrdquo Tech Rep JRC46829 Joint Research CentreIspra Italy 2008

[7] V Karlos and G Solomos ldquoCalculation of blast loads forapplication to structural componentsrdquo Tech Rep EUR26456EN Joint Research Centre Publications Office of theEuropean Union Luxembourg 2013

[8] V Karlos G Solomos and M Larcher ldquoAnalysis of the blastwave decay coefficient using the Kingery-Bulmash datardquoInternational Journal of Protective Structures vol 7 no 3pp 409ndash429 2016

[9] M Larcher ldquoDevelopment of discrete cracks in concreteloaded by shock wavesrdquo International Journal for ImpactEngineering vol 36 pp 700ndash710 2009

[10] J Li and H Hao ldquoInternal and external pressure prediction ofvented gas explosion in large rooms by using analytical andCFD methodsrdquo Journal of Loss Prevention in the ProcessIndustries vol 49 pp 367ndash381 2017

[11] EN14494 ldquoGas explosion venting protective systemsrdquo 2007[12] M Larcher F Casadei and G Solomos ldquoInfluence of venting

areas on the air blast pressure inside tubular structures likerailway carriagesrdquo Journal of Hazardous Materials vol 183no 1-3 pp 839ndash846 2010

[13] P J Miller ldquoA reactive flow model with coupled reactionkinetics for detonation and combustion in non-ideal explo-sivesrdquo MRS Proceedings vol 418 1995

[14] M Larcher R Forsberg U Bjoernstig A Holgersson andG Solomos ldquoEffectiveness of finite element modelling ofdamage and injuries for explosions inside trainsrdquo Journal ofTransportation Safety and Security vol 8 pp 83ndash100 2015

[15] G Valsamos F Casadei M Larcher and G SolomosldquoImplementation of flying debris fatal risk calculation inEUROPLEXUSrdquo Tech Rep JRC94805 Joint Research CentrePublications Office of the European Union Luxembourg 2015

[16] E Gonzalez Ferradas F Dıaz Alonso M Doval MintildearroA Mintildeana Aznar J Ruiz Gimeno and J Sanchez PerezldquoConsequence analysis by means of characteristic curves todetermine the damage to humans from bursting sphericalvesselsrdquo Process Safety and Environmental Protection vol 86no 2 pp 121ndash129 2008

10 Advances in Civil Engineering

[17] S Mannan Leesrsquo Loss Prevention in the Process IndustriesHazard Identification Assessment and Control Vol 2 ElsevierAmsterdam 2005

[18] Department of Defence USA Unified Facilities Criteria UFC3-340ndash2 2008

[19] ldquoNATO STANAG 2280 design threat levels and handoverprocedure for temporary protective structuresrdquo 2008

[20] G F Kinney and K J Graham Explosive Shocks in AirSpringer Berlin Germany 1985

[21] D Bogosian J Ferritto and Y Shi ldquoMeasuring uncertaintyand conservatism in simplified blast modelsrdquo in Proceedingsof 30th Explosive Safety Seminar Atlanta Georgia August 2002

[22] M Larcher and F Casadei ldquoExplosions in complexgeometriesndasha comparison of several approachesrdquo InternationalJournal of Protective Structures vol 1 no 2 pp 169ndash195 2010

[23] M Larcher A Stolz L Rudiger and M Steyerer ldquoStatischeErsatzlasten und ldquoIngenieurverfahrenrdquordquo in Berichte ausdem Konstruktiven Ingenieurbau Workshop Bau-ProtectN Gebbeken K (oma M Rudiger and M Klaus Edspp 117ndash130 Universitat der Bundeswehr MunchenNeubiberg Germany 2012

[24] FEMA ldquoSite and urban design for securityndashguidance againstpotential terrorist attacksrdquo Tech Rep FEMA 430 FEMAWashington DC USA 2007

[25] Joint Research Centre and Commissariat a lrsquoenergie atomiqueet aux energies alternatives EUROPLEXUS httpwww-epxceafr

[26] Commissariat a lrsquoenergie atomique et aux energies alterna-tives CAST3M httpwww-cast3mceafr

[27] ParaView httpwwwparavieworg[28] F Daude and P Galon ldquoOn the computation of the baerndash

nunziato model using ALE formulation with HLL-and HLLC-type solvers towards fluidndashstructure interactionsrdquo Journal ofComputational Physics vol 324 pp 189ndash230 2016

[29] M Larcher F Casadei and G Solomos ldquoSimulation of blastwaves by usingmapping technology in EUROPLEXUSrdquo TechRep JRC91102 Joint Research Centre Publications Office ofthe European Union Luxembourg 2014

[30] E Lee J Horning and J Kury ldquoAdiabatic expansion of highexplosives detonation productsrdquo University of CaliforniaLawrence Livermore National Laboratory Livermore CAUSA 1968

[31] B Dobratz and P Crawford LLNL Explosives HandbookProperties of Chemical Explosives and Explosive SimulantsUniversity of California Lawrence Livermore NationalLaboratory Livermore CA USA 1985

Advances in Civil Engineering 11

roomrsquos size) have a very small impact on the peak pressure butare eshyective for reducing the remaining static pressure afterthe explosion Clearly the proper design of release surfaces isa challenging task since no clear indications exist in the openliterature Usually it is advised to follow a design approachthat depends on the geometry of the control zone while thesize of the openings ranges from 3 to 8 of the roomrsquos(control arearsquos) size us these values serve only as an in-dication since the actual release surface size depends also onits cross section and the determined blast scenario

If the access control zone is located within or is con-nected to the main building the propagation of the pressurewave into the adjacent structure should be avoided toconne the explosive eshyects only in one area Constructionshaving very stishy or blast resistant doors are complicatedimpractical and space demanding since these doors must beduplicated As shown below a so-calledmeandering wall canachieve a similar protection level maintaining a design thatfavors the unobstructed ow of people

4 Numerical Simulations

e current section is dedicated to the analysis of a casestudy simulating an explosion within an access control zonee objective of the case study is to investigate the waya rigid (meandering) wall ashyects the propagation of the blastwave into an area of the building that needs to be protectedMoreover a parametric study on the inuence of the basicdimensions of the meandering wall on the attenuation of theproduced detonation wave is carried out e numericalsimulations are performed with the explicit FE codeEUROPLEXUS (EPX) [25] which is jointly developed by theFrench Commissariat a lrsquo Energie Atomique et aux EnergiesAlternatives (CEA) and the Joint Research Centre of the

European Commission (JRC Ispra) EPX employs sophis-ticated uid-structure interaction techniques and is suitablefor simulating fast and severe loadings on structures such asimpact and blast e CAST3M code [26] also developed byCEA was used to generate the mesh for the numericalmodels while ParaView software [27] was used for post-processing the results

41 Case Study To reduce complexity a simple geometryhas been selected for the case study A rectangular area isseparated by a rigid wall into two partse rst compartmentis considered as the entrance zone where a possible intrudercan enter carrying an improvised explosive device (IED) andwalk till the opening of the rigid wall that leads to the secondcompartment Just before the opening that connects the twoparts a thorough control of all persons entering the secondpart is considered to be implemented e second com-partment is regarded a vulnerable zone that needs to beprotected as a possible blast wave can become critical for thepersonnel or the structural integrity of the building

Figure 6 shows the geometrical details of the investigatedarea e entrance zone located on the left is connected tothe protected zone on the right through a 3m wideopening e openingrsquos width guarantees a relatively highow of the people to avoid queuinge height of both zonesis 3 meters while the detonation centre is located 15 metersfrom the ground level (average value for the location ofa backpack in the vertical axis) and is centred to the openingbased on a worst case scenario is is assumed to be theclosest point to the protected zone that a perpetrator canreach prior to being intercepted e mass of the IED hasbeen selected to be equal to 25 kilograms of TNTequivalentas an upper limit for the amount of explosives a pedestriancould carry in a backpack or small suitcase

10 mEntrance zone

Explosive material

Reflecting boundaries

Absorbing boundaries

Protected zone

Blocking wall (meandering)

2 m

3 m

E

D

10 m 20 mRigid wall

Figure 6 Geometry of the case study model

6 Advances in Civil Engineering

As shown in Figure 6 the distance D between theopening and the introduced meandering wall and the di-mension E that refers to the overlapping between the ad-ditional blocking wall and the preexisting rigid wall are notexplicitly determined (ese two parameters are selected asvariables in the performed parametric study in order toquantify their influence on the blast wave propagationD ranges from 1 to 3 meters corresponding to constructionspecifications (eg wheelchair access and personnel flow)Similarly E ranges from 05 to 2 meters to allow a minimumdistance of 15 meters between each extremity of themeandering wall and the lateral walls of the building and toguarantee comfortable access to the attached protected zone

42NumericalModel (e structural parts of the building inthe case study are assumed rigid meaning that the walls theceiling and the floor of the structure are nondeformableA postanalysis investigation of the recorded maximum pres-sures indicated that thick stiff structural elements like concreteor steel walls respect this basic assumption (is hypothesisresults in decreased computational power demand sincethere is no need to include deformable structural parts orperform cumbersome fluid-structure interaction (FSI)calculations Moreover the absence of structural de-formations allows an accurate study of the influence of thetwo parameters (D and E) on the blast wave propagation

As already mentioned for the numerical calculationsa purely Eulerian approach is utilized where only the fluidpart of the model is considered (e Eulerian formulationconsiders the computational mesh fixed while the fluid(particles) moves relative to these fixed grid points (econservation laws of mass momentum and energy areexpressed in a spatial framework (e discretization is basedon cell-centred finite volume (FV) formulations in which thegoverning equations are solved in integral form (ismethod is conservative since the formulation ensures thatthe flux entering a given volume is identical to that leaving

the adjacent volume (e numerical fluxes between adjacentFVs were calculated using the Harten-Lax-van Leer-contact(HLLC) Reimann solver [28]

(e fluid domain was modelled with 3D cubic finitevolumes (CUVF) of 01 meter edge size (e size of the cubicvolumes is selected after a convergence parametric study Onthe extremity surfaces of the two different zones (yellow lineson Figure 6) absorbing boundaries have been introduced inorder to simulate nonreflecting atmospheric boundariesSuitable shell elements (CL3D) that allow the introduction ofan absorbing medium have been employed and froma mathematical point of view represent a supersonic outletAs already mentioned the rigid parts of the model aresimulated as reflecting boundaries For a 3D cubic FV a freesurface is considered by definition as a reflecting boundaryA simple duplication of the fluid nodes at the area where therigid structure is located results into free FV surfaces thatbehave as reflecting boundaries

(e representation of the blast load is performed viathe 1D to 3D mapping technique [29] during whicha 1D numerical calculation that simulates the blast load isperformed with a very fine mesh(e data obtained from the1D calculation are mapped into a spherical shape andinserted in the coarser 3D mesh An important detail thatshould be taken into account is that when transferring theresults of the 1D analysis into the 3D mesh the blast load isconsidered spherical so the transformation should beperformed before the wave comes to contact with any ob-stacles (is mapping technique can significantly acceleratethe calculation time without sacrificing the accuracy of theresults For the representation of the detonation load inthe 1D calculation the Jones-Wilkins-Lee EOS law is used[30](e material parameters for the TNTare proposed by[31] and are included in Table 1 (e perfect gas law isemployed for the representation of the atmospheric air inthe fluid simulation (e material parameters for theatmospheric air are shown in Table 2 Finally one

Table 1 Parameters for the JWL EOS for TNT (Dobratz and Crawford [31])

Parameters Description Parameters for explosive Parameters for airA (Pa) Material constant (experimental) 37121e11 37121e11B (Pa) Material constant (experimental) 323e9 323e9R1 Material constant (experimental) 415 415R2 Material constant (experimental) 095 095Beta Material constant (experimental) 025 025RHO (kgm3) Density 1630 13EINT (Jkg) Current internal energyPer unit mass 429e6 021978e6OMEGA Specific heat ratio 03 03D (ms) Detonation velocity 6930 mdash

Table 2 Material properties for the ideal gas law for atmospheric air

Density (kgm3) Specific heat at constant volume (JkgmiddotK) Gamma ratio Reference pressure (Pa)13 71675 14 1e5

Advances in Civil Engineering 7

symmetry plane was utilized by modelling only one half ofthe uid mesh

43 Numerical Results e objective of the numerical in-vestigation is to determine the eshyectiveness of the in-troduced meandering wall Figure 7 displays the behaviourof the propagated blast wave for the ldquoopen doorrdquo case anda ldquomeanderingrdquo case (D 1m E 1m) for a plane view atthe height of 15 meters A comparison of the two cases isperformed at several time frames in order to provide a clearrepresentation on the diversion of the blast wave from theadditional wall As already mentioned half of the geometrywas modelled in the numerical calculations so in order to

present the full model in Figure 7 special mirroring tech-niques are used in the postprocessing phase e rstsubgure presents the pressure eld distribution at 3msafter the detonation At that time frame the wave front startsbeing reected on the meandering wall as it is clear from thedishyerences between the two models For the ldquomeanderingrdquocase the blast wave starts to enter the protected zone at 4msand is split into two smaller waves propagating from bothsides of themeandering wall At 8ms it is evident that in theldquoopen doorrdquo case there is one strong wave front propagatinginto the protected zone while for the ldquomeanderingrdquo casethere are two weaker wave fronts e wave propagationcontinues in a similar way for both cases until 30ms after thedetonation where the two wave fronts of the ldquomeanderingrdquo

3 ms

10 ms 15 ms

Pressure

30 ms

4 ms 8 ms

10e+05 15e+05 20e+05 25e+05 30e+05

Figure 7 Pressure distribution of the blast wave for the ldquoopen doorrdquo model and a ldquomeanderingrdquo model (D 1m E 1m) for several timeframes (Pa)

500 10 20 30 40 50 60 70

100

150

200

250

300

Time (ms)

Pres

sure

(kPa

)

Open doorMeandering

(a)

0 10 20 30 40 50 60 70Time (ms)

Impu

lse (P

amiddots)

Open doorMeandering

0

500

1000

1500

(b)

Figure 8 Comparison of (a) pressure and (b) impulse values for a point located in the protected zone for an ldquoopen doorrdquo model anda ldquomeandering wallrdquo model (D 1m E 1m)

8 Advances in Civil Engineering

case are united into one But as can be noticed even in thatphase the wave front of the ldquomeanderingrdquo case is weaker thanthe one of the ldquoopen doorrdquo case

e meandering wall reects part of the initial blast waveand guides it back to the entrance area resulting in a slightlystronger blast wave into the entrance zone Since this zone hasalready been signicantly ashyected from the initial blast wave theadditional energy due to the presence of the ldquomeanderingrdquo walldoes not ashyect the injury risk probability Figure 8 presents thepressure and the impulse history for the abovementioned casesat a point that is located at the middle of the protected zone7 meters from the opening and 15 meters from the ground

level From the pressure-time history it is clear that in theldquomeanderingrdquo case the rst peak is eliminated while the secondpeak that results from the reection of the wave onto the lateralwalls is lower Similar results can be derived from comparingthe impulse-time histories as there is a signicant reduction ofthe total impulse when the meandering wall is present

e next gures present the inuence of the dimensionand location of the meandering wall on the blast wavepropagation As has already been described in previoussections areas with an impulse value larger than 1000 Pamiddotsare considered as high injury risk zones Figure 9 presentsthe injury risk areas for the protected zone of the examined

E 05

D1

D2Impulse

30e+03

20e+03

10e+03

00e+00

D3

E 1 E 2

Figure 9 Zone of impulse higher than 1000 Pamiddots for the ldquomeanderingrdquo studied cases and the ldquoopen doorrdquo model (parameters D and E in mand impulse in pamiddots)

105 151

E (m)

D (m

)

()

2

15

2

25

3

60

40

20

60

40

20

0

Figure 10 Percentage of ashyected area spectrum versus the investigated parameters for the ldquomeanderingrdquo model

Advances in Civil Engineering 9

building for several studied geometries (e results refer tothe same horizontal cut that is presented in Figure 7 but theoutput quantity is the calculated impulse Areas charac-terized by impulse values higher than 1000 Pamiddots are high-lighted D and E determine the location and the size of themeandering wall as described earlier

In the same figure the ldquoopen doorrdquo case is also includedfor comparison purposes In the ldquoopen doorrdquo case themajority of the protected zone results in a high injury riskarea In the ldquomeanderingrdquo case the size of the high injuryrisk area varies but is always smaller than the ldquoopen doorrdquocase It is clear that the area before the additional wall isalways unprotected irrespective of the presence of themeandering wall Moreover in some cases high impulsevalues are also recorded at the sides of the meandering walldue to the relatively strong wave fronts that are present atboth those areas Figure 10 summarizes the output of theparametric study by presenting the percentage of the injuryrisk area in the protected zone (above 1000 Pamiddots) versus thetwo examined parameters From comparing the severalldquomeanderingrdquo cases it can be concluded that if D is small(distance between the door and the added wall) the influ-ence of E (dimension of overlapping length) is of minorimportance However if D is large the influence of E issignificant (if E decreases the blast wave that enters into theprotected zone is stronger and the high injury risk areas arebecoming larger)

5 Conclusions

Access control points that are located inside or attached tobuildings are often adopted due to local constraints (eirdesign must consider possible terrorist attack scenarios(eg explosions) at the interior and exterior of these accesscontrol points (e mitigation of the consequences shouldsuch a terrorist attack materialize inside an access point andin the attached building can be achieved by using venting(release surfaces) and in particular meandering walls (einvestigation shows that the reduction of the pressure issignificant if meandering walls are used and that these wallscan therefore play an important role in separating accesscontrol points from the attached buildings Nevertheless it isobvious that meandering walls can be considered an effectiverisk reduction measure and not as a full protection solution(e risk reduction concerns both blast wave effects and forobvious reasons fragments resulting from explosions andshooting attacks Herein the idealized situation that is in-vestigated can be extended to include additional parametersAs has been presented in the current article the presence ofa rigid wall standing between an explosionrsquos detonationcentre and a target can effectively reduce the severity of theimpacting blast wave It is therefore recommended theaddition of such elements (varying in size and geometry) inorder to mitigate the risk of terrorist attacks

Conflicts of Interest

(e authors declare that there are no conflicts of interestregarding the publication of this paper

References

[1] A M Lewis ldquoState of the art report on European legislationrelating to explosives and explosive detection system for non-aviation configurationsrdquo Tech Rep JRC84076 Joint ResearchCentre Ispra Italy 2013

[2] B M Luccioni R D Ambrosini and R F Danesi ldquoAnalysisof building collapse under blast loadsrdquo Engineering Structuresvol 26 no 1 pp 63ndash71 2004

[3] M Larcher M Arrigoni C Bedon et al ldquoDesign of blast-loaded glazing windows and facades a review of essentialrequirements towards standardizationrdquo Advances in CivilEngineering vol 2016 Article ID 2604232 14 pages 2016

[4] R Hajek and M Foglar ldquoNumerical and experimentalanalysis of the effect of rigid barriers on blast wave propa-gationrdquo Journal of Structural Engineering vol 141 no 12p 04015061 2015

[5] C N Kingery and G Bulmash ldquoAirblast parameters fromTNT spherical air burst and hemispherical surface burstrdquoTech Rep ARBRL-TR-02555 Defence Technical InformationCenter Ballistic Research Laboratory Aberdeen ProvingGround MD USA Tech Rep ARBRL-TR-02555 1984

[6] M Larcher ldquoPressure-time functions for the description of airblast wavesrdquo Tech Rep JRC46829 Joint Research CentreIspra Italy 2008

[7] V Karlos and G Solomos ldquoCalculation of blast loads forapplication to structural componentsrdquo Tech Rep EUR26456EN Joint Research Centre Publications Office of theEuropean Union Luxembourg 2013

[8] V Karlos G Solomos and M Larcher ldquoAnalysis of the blastwave decay coefficient using the Kingery-Bulmash datardquoInternational Journal of Protective Structures vol 7 no 3pp 409ndash429 2016

[9] M Larcher ldquoDevelopment of discrete cracks in concreteloaded by shock wavesrdquo International Journal for ImpactEngineering vol 36 pp 700ndash710 2009

[10] J Li and H Hao ldquoInternal and external pressure prediction ofvented gas explosion in large rooms by using analytical andCFD methodsrdquo Journal of Loss Prevention in the ProcessIndustries vol 49 pp 367ndash381 2017

[11] EN14494 ldquoGas explosion venting protective systemsrdquo 2007[12] M Larcher F Casadei and G Solomos ldquoInfluence of venting

areas on the air blast pressure inside tubular structures likerailway carriagesrdquo Journal of Hazardous Materials vol 183no 1-3 pp 839ndash846 2010

[13] P J Miller ldquoA reactive flow model with coupled reactionkinetics for detonation and combustion in non-ideal explo-sivesrdquo MRS Proceedings vol 418 1995

[14] M Larcher R Forsberg U Bjoernstig A Holgersson andG Solomos ldquoEffectiveness of finite element modelling ofdamage and injuries for explosions inside trainsrdquo Journal ofTransportation Safety and Security vol 8 pp 83ndash100 2015

[15] G Valsamos F Casadei M Larcher and G SolomosldquoImplementation of flying debris fatal risk calculation inEUROPLEXUSrdquo Tech Rep JRC94805 Joint Research CentrePublications Office of the European Union Luxembourg 2015

[16] E Gonzalez Ferradas F Dıaz Alonso M Doval MintildearroA Mintildeana Aznar J Ruiz Gimeno and J Sanchez PerezldquoConsequence analysis by means of characteristic curves todetermine the damage to humans from bursting sphericalvesselsrdquo Process Safety and Environmental Protection vol 86no 2 pp 121ndash129 2008

10 Advances in Civil Engineering

[17] S Mannan Leesrsquo Loss Prevention in the Process IndustriesHazard Identification Assessment and Control Vol 2 ElsevierAmsterdam 2005

[18] Department of Defence USA Unified Facilities Criteria UFC3-340ndash2 2008

[19] ldquoNATO STANAG 2280 design threat levels and handoverprocedure for temporary protective structuresrdquo 2008

[20] G F Kinney and K J Graham Explosive Shocks in AirSpringer Berlin Germany 1985

[21] D Bogosian J Ferritto and Y Shi ldquoMeasuring uncertaintyand conservatism in simplified blast modelsrdquo in Proceedingsof 30th Explosive Safety Seminar Atlanta Georgia August 2002

[22] M Larcher and F Casadei ldquoExplosions in complexgeometriesndasha comparison of several approachesrdquo InternationalJournal of Protective Structures vol 1 no 2 pp 169ndash195 2010

[23] M Larcher A Stolz L Rudiger and M Steyerer ldquoStatischeErsatzlasten und ldquoIngenieurverfahrenrdquordquo in Berichte ausdem Konstruktiven Ingenieurbau Workshop Bau-ProtectN Gebbeken K (oma M Rudiger and M Klaus Edspp 117ndash130 Universitat der Bundeswehr MunchenNeubiberg Germany 2012

[24] FEMA ldquoSite and urban design for securityndashguidance againstpotential terrorist attacksrdquo Tech Rep FEMA 430 FEMAWashington DC USA 2007

[25] Joint Research Centre and Commissariat a lrsquoenergie atomiqueet aux energies alternatives EUROPLEXUS httpwww-epxceafr

[26] Commissariat a lrsquoenergie atomique et aux energies alterna-tives CAST3M httpwww-cast3mceafr

[27] ParaView httpwwwparavieworg[28] F Daude and P Galon ldquoOn the computation of the baerndash

nunziato model using ALE formulation with HLL-and HLLC-type solvers towards fluidndashstructure interactionsrdquo Journal ofComputational Physics vol 324 pp 189ndash230 2016

[29] M Larcher F Casadei and G Solomos ldquoSimulation of blastwaves by usingmapping technology in EUROPLEXUSrdquo TechRep JRC91102 Joint Research Centre Publications Office ofthe European Union Luxembourg 2014

[30] E Lee J Horning and J Kury ldquoAdiabatic expansion of highexplosives detonation productsrdquo University of CaliforniaLawrence Livermore National Laboratory Livermore CAUSA 1968

[31] B Dobratz and P Crawford LLNL Explosives HandbookProperties of Chemical Explosives and Explosive SimulantsUniversity of California Lawrence Livermore NationalLaboratory Livermore CA USA 1985

Advances in Civil Engineering 11

As shown in Figure 6 the distance D between theopening and the introduced meandering wall and the di-mension E that refers to the overlapping between the ad-ditional blocking wall and the preexisting rigid wall are notexplicitly determined (ese two parameters are selected asvariables in the performed parametric study in order toquantify their influence on the blast wave propagationD ranges from 1 to 3 meters corresponding to constructionspecifications (eg wheelchair access and personnel flow)Similarly E ranges from 05 to 2 meters to allow a minimumdistance of 15 meters between each extremity of themeandering wall and the lateral walls of the building and toguarantee comfortable access to the attached protected zone

42NumericalModel (e structural parts of the building inthe case study are assumed rigid meaning that the walls theceiling and the floor of the structure are nondeformableA postanalysis investigation of the recorded maximum pres-sures indicated that thick stiff structural elements like concreteor steel walls respect this basic assumption (is hypothesisresults in decreased computational power demand sincethere is no need to include deformable structural parts orperform cumbersome fluid-structure interaction (FSI)calculations Moreover the absence of structural de-formations allows an accurate study of the influence of thetwo parameters (D and E) on the blast wave propagation

As already mentioned for the numerical calculationsa purely Eulerian approach is utilized where only the fluidpart of the model is considered (e Eulerian formulationconsiders the computational mesh fixed while the fluid(particles) moves relative to these fixed grid points (econservation laws of mass momentum and energy areexpressed in a spatial framework (e discretization is basedon cell-centred finite volume (FV) formulations in which thegoverning equations are solved in integral form (ismethod is conservative since the formulation ensures thatthe flux entering a given volume is identical to that leaving

the adjacent volume (e numerical fluxes between adjacentFVs were calculated using the Harten-Lax-van Leer-contact(HLLC) Reimann solver [28]

(e fluid domain was modelled with 3D cubic finitevolumes (CUVF) of 01 meter edge size (e size of the cubicvolumes is selected after a convergence parametric study Onthe extremity surfaces of the two different zones (yellow lineson Figure 6) absorbing boundaries have been introduced inorder to simulate nonreflecting atmospheric boundariesSuitable shell elements (CL3D) that allow the introduction ofan absorbing medium have been employed and froma mathematical point of view represent a supersonic outletAs already mentioned the rigid parts of the model aresimulated as reflecting boundaries For a 3D cubic FV a freesurface is considered by definition as a reflecting boundaryA simple duplication of the fluid nodes at the area where therigid structure is located results into free FV surfaces thatbehave as reflecting boundaries

(e representation of the blast load is performed viathe 1D to 3D mapping technique [29] during whicha 1D numerical calculation that simulates the blast load isperformed with a very fine mesh(e data obtained from the1D calculation are mapped into a spherical shape andinserted in the coarser 3D mesh An important detail thatshould be taken into account is that when transferring theresults of the 1D analysis into the 3D mesh the blast load isconsidered spherical so the transformation should beperformed before the wave comes to contact with any ob-stacles (is mapping technique can significantly acceleratethe calculation time without sacrificing the accuracy of theresults For the representation of the detonation load inthe 1D calculation the Jones-Wilkins-Lee EOS law is used[30](e material parameters for the TNTare proposed by[31] and are included in Table 1 (e perfect gas law isemployed for the representation of the atmospheric air inthe fluid simulation (e material parameters for theatmospheric air are shown in Table 2 Finally one

Table 1 Parameters for the JWL EOS for TNT (Dobratz and Crawford [31])

Parameters Description Parameters for explosive Parameters for airA (Pa) Material constant (experimental) 37121e11 37121e11B (Pa) Material constant (experimental) 323e9 323e9R1 Material constant (experimental) 415 415R2 Material constant (experimental) 095 095Beta Material constant (experimental) 025 025RHO (kgm3) Density 1630 13EINT (Jkg) Current internal energyPer unit mass 429e6 021978e6OMEGA Specific heat ratio 03 03D (ms) Detonation velocity 6930 mdash

Table 2 Material properties for the ideal gas law for atmospheric air

Density (kgm3) Specific heat at constant volume (JkgmiddotK) Gamma ratio Reference pressure (Pa)13 71675 14 1e5

Advances in Civil Engineering 7

symmetry plane was utilized by modelling only one half ofthe uid mesh

43 Numerical Results e objective of the numerical in-vestigation is to determine the eshyectiveness of the in-troduced meandering wall Figure 7 displays the behaviourof the propagated blast wave for the ldquoopen doorrdquo case anda ldquomeanderingrdquo case (D 1m E 1m) for a plane view atthe height of 15 meters A comparison of the two cases isperformed at several time frames in order to provide a clearrepresentation on the diversion of the blast wave from theadditional wall As already mentioned half of the geometrywas modelled in the numerical calculations so in order to

present the full model in Figure 7 special mirroring tech-niques are used in the postprocessing phase e rstsubgure presents the pressure eld distribution at 3msafter the detonation At that time frame the wave front startsbeing reected on the meandering wall as it is clear from thedishyerences between the two models For the ldquomeanderingrdquocase the blast wave starts to enter the protected zone at 4msand is split into two smaller waves propagating from bothsides of themeandering wall At 8ms it is evident that in theldquoopen doorrdquo case there is one strong wave front propagatinginto the protected zone while for the ldquomeanderingrdquo casethere are two weaker wave fronts e wave propagationcontinues in a similar way for both cases until 30ms after thedetonation where the two wave fronts of the ldquomeanderingrdquo

3 ms

10 ms 15 ms

Pressure

30 ms

4 ms 8 ms

10e+05 15e+05 20e+05 25e+05 30e+05

Figure 7 Pressure distribution of the blast wave for the ldquoopen doorrdquo model and a ldquomeanderingrdquo model (D 1m E 1m) for several timeframes (Pa)

500 10 20 30 40 50 60 70

100

150

200

250

300

Time (ms)

Pres

sure

(kPa

)

Open doorMeandering

(a)

0 10 20 30 40 50 60 70Time (ms)

Impu

lse (P

amiddots)

Open doorMeandering

0

500

1000

1500

(b)

Figure 8 Comparison of (a) pressure and (b) impulse values for a point located in the protected zone for an ldquoopen doorrdquo model anda ldquomeandering wallrdquo model (D 1m E 1m)

8 Advances in Civil Engineering

case are united into one But as can be noticed even in thatphase the wave front of the ldquomeanderingrdquo case is weaker thanthe one of the ldquoopen doorrdquo case

e meandering wall reects part of the initial blast waveand guides it back to the entrance area resulting in a slightlystronger blast wave into the entrance zone Since this zone hasalready been signicantly ashyected from the initial blast wave theadditional energy due to the presence of the ldquomeanderingrdquo walldoes not ashyect the injury risk probability Figure 8 presents thepressure and the impulse history for the abovementioned casesat a point that is located at the middle of the protected zone7 meters from the opening and 15 meters from the ground

level From the pressure-time history it is clear that in theldquomeanderingrdquo case the rst peak is eliminated while the secondpeak that results from the reection of the wave onto the lateralwalls is lower Similar results can be derived from comparingthe impulse-time histories as there is a signicant reduction ofthe total impulse when the meandering wall is present

e next gures present the inuence of the dimensionand location of the meandering wall on the blast wavepropagation As has already been described in previoussections areas with an impulse value larger than 1000 Pamiddotsare considered as high injury risk zones Figure 9 presentsthe injury risk areas for the protected zone of the examined

E 05

D1

D2Impulse

30e+03

20e+03

10e+03

00e+00

D3

E 1 E 2

Figure 9 Zone of impulse higher than 1000 Pamiddots for the ldquomeanderingrdquo studied cases and the ldquoopen doorrdquo model (parameters D and E in mand impulse in pamiddots)

105 151

E (m)

D (m

)

()

2

15

2

25

3

60

40

20

60

40

20

0

Figure 10 Percentage of ashyected area spectrum versus the investigated parameters for the ldquomeanderingrdquo model

Advances in Civil Engineering 9

building for several studied geometries (e results refer tothe same horizontal cut that is presented in Figure 7 but theoutput quantity is the calculated impulse Areas charac-terized by impulse values higher than 1000 Pamiddots are high-lighted D and E determine the location and the size of themeandering wall as described earlier

In the same figure the ldquoopen doorrdquo case is also includedfor comparison purposes In the ldquoopen doorrdquo case themajority of the protected zone results in a high injury riskarea In the ldquomeanderingrdquo case the size of the high injuryrisk area varies but is always smaller than the ldquoopen doorrdquocase It is clear that the area before the additional wall isalways unprotected irrespective of the presence of themeandering wall Moreover in some cases high impulsevalues are also recorded at the sides of the meandering walldue to the relatively strong wave fronts that are present atboth those areas Figure 10 summarizes the output of theparametric study by presenting the percentage of the injuryrisk area in the protected zone (above 1000 Pamiddots) versus thetwo examined parameters From comparing the severalldquomeanderingrdquo cases it can be concluded that if D is small(distance between the door and the added wall) the influ-ence of E (dimension of overlapping length) is of minorimportance However if D is large the influence of E issignificant (if E decreases the blast wave that enters into theprotected zone is stronger and the high injury risk areas arebecoming larger)

5 Conclusions

Access control points that are located inside or attached tobuildings are often adopted due to local constraints (eirdesign must consider possible terrorist attack scenarios(eg explosions) at the interior and exterior of these accesscontrol points (e mitigation of the consequences shouldsuch a terrorist attack materialize inside an access point andin the attached building can be achieved by using venting(release surfaces) and in particular meandering walls (einvestigation shows that the reduction of the pressure issignificant if meandering walls are used and that these wallscan therefore play an important role in separating accesscontrol points from the attached buildings Nevertheless it isobvious that meandering walls can be considered an effectiverisk reduction measure and not as a full protection solution(e risk reduction concerns both blast wave effects and forobvious reasons fragments resulting from explosions andshooting attacks Herein the idealized situation that is in-vestigated can be extended to include additional parametersAs has been presented in the current article the presence ofa rigid wall standing between an explosionrsquos detonationcentre and a target can effectively reduce the severity of theimpacting blast wave It is therefore recommended theaddition of such elements (varying in size and geometry) inorder to mitigate the risk of terrorist attacks

Conflicts of Interest

(e authors declare that there are no conflicts of interestregarding the publication of this paper

References

[1] A M Lewis ldquoState of the art report on European legislationrelating to explosives and explosive detection system for non-aviation configurationsrdquo Tech Rep JRC84076 Joint ResearchCentre Ispra Italy 2013

[2] B M Luccioni R D Ambrosini and R F Danesi ldquoAnalysisof building collapse under blast loadsrdquo Engineering Structuresvol 26 no 1 pp 63ndash71 2004

[3] M Larcher M Arrigoni C Bedon et al ldquoDesign of blast-loaded glazing windows and facades a review of essentialrequirements towards standardizationrdquo Advances in CivilEngineering vol 2016 Article ID 2604232 14 pages 2016

[4] R Hajek and M Foglar ldquoNumerical and experimentalanalysis of the effect of rigid barriers on blast wave propa-gationrdquo Journal of Structural Engineering vol 141 no 12p 04015061 2015

[5] C N Kingery and G Bulmash ldquoAirblast parameters fromTNT spherical air burst and hemispherical surface burstrdquoTech Rep ARBRL-TR-02555 Defence Technical InformationCenter Ballistic Research Laboratory Aberdeen ProvingGround MD USA Tech Rep ARBRL-TR-02555 1984

[6] M Larcher ldquoPressure-time functions for the description of airblast wavesrdquo Tech Rep JRC46829 Joint Research CentreIspra Italy 2008

[7] V Karlos and G Solomos ldquoCalculation of blast loads forapplication to structural componentsrdquo Tech Rep EUR26456EN Joint Research Centre Publications Office of theEuropean Union Luxembourg 2013

[8] V Karlos G Solomos and M Larcher ldquoAnalysis of the blastwave decay coefficient using the Kingery-Bulmash datardquoInternational Journal of Protective Structures vol 7 no 3pp 409ndash429 2016

[9] M Larcher ldquoDevelopment of discrete cracks in concreteloaded by shock wavesrdquo International Journal for ImpactEngineering vol 36 pp 700ndash710 2009

[10] J Li and H Hao ldquoInternal and external pressure prediction ofvented gas explosion in large rooms by using analytical andCFD methodsrdquo Journal of Loss Prevention in the ProcessIndustries vol 49 pp 367ndash381 2017

[11] EN14494 ldquoGas explosion venting protective systemsrdquo 2007[12] M Larcher F Casadei and G Solomos ldquoInfluence of venting

areas on the air blast pressure inside tubular structures likerailway carriagesrdquo Journal of Hazardous Materials vol 183no 1-3 pp 839ndash846 2010

[13] P J Miller ldquoA reactive flow model with coupled reactionkinetics for detonation and combustion in non-ideal explo-sivesrdquo MRS Proceedings vol 418 1995

[14] M Larcher R Forsberg U Bjoernstig A Holgersson andG Solomos ldquoEffectiveness of finite element modelling ofdamage and injuries for explosions inside trainsrdquo Journal ofTransportation Safety and Security vol 8 pp 83ndash100 2015

[15] G Valsamos F Casadei M Larcher and G SolomosldquoImplementation of flying debris fatal risk calculation inEUROPLEXUSrdquo Tech Rep JRC94805 Joint Research CentrePublications Office of the European Union Luxembourg 2015

[16] E Gonzalez Ferradas F Dıaz Alonso M Doval MintildearroA Mintildeana Aznar J Ruiz Gimeno and J Sanchez PerezldquoConsequence analysis by means of characteristic curves todetermine the damage to humans from bursting sphericalvesselsrdquo Process Safety and Environmental Protection vol 86no 2 pp 121ndash129 2008

10 Advances in Civil Engineering

[17] S Mannan Leesrsquo Loss Prevention in the Process IndustriesHazard Identification Assessment and Control Vol 2 ElsevierAmsterdam 2005

[18] Department of Defence USA Unified Facilities Criteria UFC3-340ndash2 2008

[19] ldquoNATO STANAG 2280 design threat levels and handoverprocedure for temporary protective structuresrdquo 2008

[20] G F Kinney and K J Graham Explosive Shocks in AirSpringer Berlin Germany 1985

[21] D Bogosian J Ferritto and Y Shi ldquoMeasuring uncertaintyand conservatism in simplified blast modelsrdquo in Proceedingsof 30th Explosive Safety Seminar Atlanta Georgia August 2002

[22] M Larcher and F Casadei ldquoExplosions in complexgeometriesndasha comparison of several approachesrdquo InternationalJournal of Protective Structures vol 1 no 2 pp 169ndash195 2010

[23] M Larcher A Stolz L Rudiger and M Steyerer ldquoStatischeErsatzlasten und ldquoIngenieurverfahrenrdquordquo in Berichte ausdem Konstruktiven Ingenieurbau Workshop Bau-ProtectN Gebbeken K (oma M Rudiger and M Klaus Edspp 117ndash130 Universitat der Bundeswehr MunchenNeubiberg Germany 2012

[24] FEMA ldquoSite and urban design for securityndashguidance againstpotential terrorist attacksrdquo Tech Rep FEMA 430 FEMAWashington DC USA 2007

[25] Joint Research Centre and Commissariat a lrsquoenergie atomiqueet aux energies alternatives EUROPLEXUS httpwww-epxceafr

[26] Commissariat a lrsquoenergie atomique et aux energies alterna-tives CAST3M httpwww-cast3mceafr

[27] ParaView httpwwwparavieworg[28] F Daude and P Galon ldquoOn the computation of the baerndash

nunziato model using ALE formulation with HLL-and HLLC-type solvers towards fluidndashstructure interactionsrdquo Journal ofComputational Physics vol 324 pp 189ndash230 2016

[29] M Larcher F Casadei and G Solomos ldquoSimulation of blastwaves by usingmapping technology in EUROPLEXUSrdquo TechRep JRC91102 Joint Research Centre Publications Office ofthe European Union Luxembourg 2014

[30] E Lee J Horning and J Kury ldquoAdiabatic expansion of highexplosives detonation productsrdquo University of CaliforniaLawrence Livermore National Laboratory Livermore CAUSA 1968

[31] B Dobratz and P Crawford LLNL Explosives HandbookProperties of Chemical Explosives and Explosive SimulantsUniversity of California Lawrence Livermore NationalLaboratory Livermore CA USA 1985

Advances in Civil Engineering 11

symmetry plane was utilized by modelling only one half ofthe uid mesh

43 Numerical Results e objective of the numerical in-vestigation is to determine the eshyectiveness of the in-troduced meandering wall Figure 7 displays the behaviourof the propagated blast wave for the ldquoopen doorrdquo case anda ldquomeanderingrdquo case (D 1m E 1m) for a plane view atthe height of 15 meters A comparison of the two cases isperformed at several time frames in order to provide a clearrepresentation on the diversion of the blast wave from theadditional wall As already mentioned half of the geometrywas modelled in the numerical calculations so in order to

present the full model in Figure 7 special mirroring tech-niques are used in the postprocessing phase e rstsubgure presents the pressure eld distribution at 3msafter the detonation At that time frame the wave front startsbeing reected on the meandering wall as it is clear from thedishyerences between the two models For the ldquomeanderingrdquocase the blast wave starts to enter the protected zone at 4msand is split into two smaller waves propagating from bothsides of themeandering wall At 8ms it is evident that in theldquoopen doorrdquo case there is one strong wave front propagatinginto the protected zone while for the ldquomeanderingrdquo casethere are two weaker wave fronts e wave propagationcontinues in a similar way for both cases until 30ms after thedetonation where the two wave fronts of the ldquomeanderingrdquo

3 ms

10 ms 15 ms

Pressure

30 ms

4 ms 8 ms

10e+05 15e+05 20e+05 25e+05 30e+05

Figure 7 Pressure distribution of the blast wave for the ldquoopen doorrdquo model and a ldquomeanderingrdquo model (D 1m E 1m) for several timeframes (Pa)

500 10 20 30 40 50 60 70

100

150

200

250

300

Time (ms)

Pres

sure

(kPa

)

Open doorMeandering

(a)

0 10 20 30 40 50 60 70Time (ms)

Impu

lse (P

amiddots)

Open doorMeandering

0

500

1000

1500

(b)

Figure 8 Comparison of (a) pressure and (b) impulse values for a point located in the protected zone for an ldquoopen doorrdquo model anda ldquomeandering wallrdquo model (D 1m E 1m)

8 Advances in Civil Engineering

case are united into one But as can be noticed even in thatphase the wave front of the ldquomeanderingrdquo case is weaker thanthe one of the ldquoopen doorrdquo case

e meandering wall reects part of the initial blast waveand guides it back to the entrance area resulting in a slightlystronger blast wave into the entrance zone Since this zone hasalready been signicantly ashyected from the initial blast wave theadditional energy due to the presence of the ldquomeanderingrdquo walldoes not ashyect the injury risk probability Figure 8 presents thepressure and the impulse history for the abovementioned casesat a point that is located at the middle of the protected zone7 meters from the opening and 15 meters from the ground

level From the pressure-time history it is clear that in theldquomeanderingrdquo case the rst peak is eliminated while the secondpeak that results from the reection of the wave onto the lateralwalls is lower Similar results can be derived from comparingthe impulse-time histories as there is a signicant reduction ofthe total impulse when the meandering wall is present

e next gures present the inuence of the dimensionand location of the meandering wall on the blast wavepropagation As has already been described in previoussections areas with an impulse value larger than 1000 Pamiddotsare considered as high injury risk zones Figure 9 presentsthe injury risk areas for the protected zone of the examined

E 05

D1

D2Impulse

30e+03

20e+03

10e+03

00e+00

D3

E 1 E 2

Figure 9 Zone of impulse higher than 1000 Pamiddots for the ldquomeanderingrdquo studied cases and the ldquoopen doorrdquo model (parameters D and E in mand impulse in pamiddots)

105 151

E (m)

D (m

)

()

2

15

2

25

3

60

40

20

60

40

20

0

Figure 10 Percentage of ashyected area spectrum versus the investigated parameters for the ldquomeanderingrdquo model

Advances in Civil Engineering 9

building for several studied geometries (e results refer tothe same horizontal cut that is presented in Figure 7 but theoutput quantity is the calculated impulse Areas charac-terized by impulse values higher than 1000 Pamiddots are high-lighted D and E determine the location and the size of themeandering wall as described earlier

In the same figure the ldquoopen doorrdquo case is also includedfor comparison purposes In the ldquoopen doorrdquo case themajority of the protected zone results in a high injury riskarea In the ldquomeanderingrdquo case the size of the high injuryrisk area varies but is always smaller than the ldquoopen doorrdquocase It is clear that the area before the additional wall isalways unprotected irrespective of the presence of themeandering wall Moreover in some cases high impulsevalues are also recorded at the sides of the meandering walldue to the relatively strong wave fronts that are present atboth those areas Figure 10 summarizes the output of theparametric study by presenting the percentage of the injuryrisk area in the protected zone (above 1000 Pamiddots) versus thetwo examined parameters From comparing the severalldquomeanderingrdquo cases it can be concluded that if D is small(distance between the door and the added wall) the influ-ence of E (dimension of overlapping length) is of minorimportance However if D is large the influence of E issignificant (if E decreases the blast wave that enters into theprotected zone is stronger and the high injury risk areas arebecoming larger)

5 Conclusions

Access control points that are located inside or attached tobuildings are often adopted due to local constraints (eirdesign must consider possible terrorist attack scenarios(eg explosions) at the interior and exterior of these accesscontrol points (e mitigation of the consequences shouldsuch a terrorist attack materialize inside an access point andin the attached building can be achieved by using venting(release surfaces) and in particular meandering walls (einvestigation shows that the reduction of the pressure issignificant if meandering walls are used and that these wallscan therefore play an important role in separating accesscontrol points from the attached buildings Nevertheless it isobvious that meandering walls can be considered an effectiverisk reduction measure and not as a full protection solution(e risk reduction concerns both blast wave effects and forobvious reasons fragments resulting from explosions andshooting attacks Herein the idealized situation that is in-vestigated can be extended to include additional parametersAs has been presented in the current article the presence ofa rigid wall standing between an explosionrsquos detonationcentre and a target can effectively reduce the severity of theimpacting blast wave It is therefore recommended theaddition of such elements (varying in size and geometry) inorder to mitigate the risk of terrorist attacks

Conflicts of Interest

(e authors declare that there are no conflicts of interestregarding the publication of this paper

References

[1] A M Lewis ldquoState of the art report on European legislationrelating to explosives and explosive detection system for non-aviation configurationsrdquo Tech Rep JRC84076 Joint ResearchCentre Ispra Italy 2013

[2] B M Luccioni R D Ambrosini and R F Danesi ldquoAnalysisof building collapse under blast loadsrdquo Engineering Structuresvol 26 no 1 pp 63ndash71 2004

[3] M Larcher M Arrigoni C Bedon et al ldquoDesign of blast-loaded glazing windows and facades a review of essentialrequirements towards standardizationrdquo Advances in CivilEngineering vol 2016 Article ID 2604232 14 pages 2016

[4] R Hajek and M Foglar ldquoNumerical and experimentalanalysis of the effect of rigid barriers on blast wave propa-gationrdquo Journal of Structural Engineering vol 141 no 12p 04015061 2015

[5] C N Kingery and G Bulmash ldquoAirblast parameters fromTNT spherical air burst and hemispherical surface burstrdquoTech Rep ARBRL-TR-02555 Defence Technical InformationCenter Ballistic Research Laboratory Aberdeen ProvingGround MD USA Tech Rep ARBRL-TR-02555 1984

[6] M Larcher ldquoPressure-time functions for the description of airblast wavesrdquo Tech Rep JRC46829 Joint Research CentreIspra Italy 2008

[7] V Karlos and G Solomos ldquoCalculation of blast loads forapplication to structural componentsrdquo Tech Rep EUR26456EN Joint Research Centre Publications Office of theEuropean Union Luxembourg 2013

[8] V Karlos G Solomos and M Larcher ldquoAnalysis of the blastwave decay coefficient using the Kingery-Bulmash datardquoInternational Journal of Protective Structures vol 7 no 3pp 409ndash429 2016

[9] M Larcher ldquoDevelopment of discrete cracks in concreteloaded by shock wavesrdquo International Journal for ImpactEngineering vol 36 pp 700ndash710 2009

[10] J Li and H Hao ldquoInternal and external pressure prediction ofvented gas explosion in large rooms by using analytical andCFD methodsrdquo Journal of Loss Prevention in the ProcessIndustries vol 49 pp 367ndash381 2017

[11] EN14494 ldquoGas explosion venting protective systemsrdquo 2007[12] M Larcher F Casadei and G Solomos ldquoInfluence of venting

areas on the air blast pressure inside tubular structures likerailway carriagesrdquo Journal of Hazardous Materials vol 183no 1-3 pp 839ndash846 2010

[13] P J Miller ldquoA reactive flow model with coupled reactionkinetics for detonation and combustion in non-ideal explo-sivesrdquo MRS Proceedings vol 418 1995

[14] M Larcher R Forsberg U Bjoernstig A Holgersson andG Solomos ldquoEffectiveness of finite element modelling ofdamage and injuries for explosions inside trainsrdquo Journal ofTransportation Safety and Security vol 8 pp 83ndash100 2015

[15] G Valsamos F Casadei M Larcher and G SolomosldquoImplementation of flying debris fatal risk calculation inEUROPLEXUSrdquo Tech Rep JRC94805 Joint Research CentrePublications Office of the European Union Luxembourg 2015

[16] E Gonzalez Ferradas F Dıaz Alonso M Doval MintildearroA Mintildeana Aznar J Ruiz Gimeno and J Sanchez PerezldquoConsequence analysis by means of characteristic curves todetermine the damage to humans from bursting sphericalvesselsrdquo Process Safety and Environmental Protection vol 86no 2 pp 121ndash129 2008

10 Advances in Civil Engineering

[17] S Mannan Leesrsquo Loss Prevention in the Process IndustriesHazard Identification Assessment and Control Vol 2 ElsevierAmsterdam 2005

[18] Department of Defence USA Unified Facilities Criteria UFC3-340ndash2 2008

[19] ldquoNATO STANAG 2280 design threat levels and handoverprocedure for temporary protective structuresrdquo 2008

[20] G F Kinney and K J Graham Explosive Shocks in AirSpringer Berlin Germany 1985

[21] D Bogosian J Ferritto and Y Shi ldquoMeasuring uncertaintyand conservatism in simplified blast modelsrdquo in Proceedingsof 30th Explosive Safety Seminar Atlanta Georgia August 2002

[22] M Larcher and F Casadei ldquoExplosions in complexgeometriesndasha comparison of several approachesrdquo InternationalJournal of Protective Structures vol 1 no 2 pp 169ndash195 2010

[23] M Larcher A Stolz L Rudiger and M Steyerer ldquoStatischeErsatzlasten und ldquoIngenieurverfahrenrdquordquo in Berichte ausdem Konstruktiven Ingenieurbau Workshop Bau-ProtectN Gebbeken K (oma M Rudiger and M Klaus Edspp 117ndash130 Universitat der Bundeswehr MunchenNeubiberg Germany 2012

[24] FEMA ldquoSite and urban design for securityndashguidance againstpotential terrorist attacksrdquo Tech Rep FEMA 430 FEMAWashington DC USA 2007

[25] Joint Research Centre and Commissariat a lrsquoenergie atomiqueet aux energies alternatives EUROPLEXUS httpwww-epxceafr

[26] Commissariat a lrsquoenergie atomique et aux energies alterna-tives CAST3M httpwww-cast3mceafr

[27] ParaView httpwwwparavieworg[28] F Daude and P Galon ldquoOn the computation of the baerndash

nunziato model using ALE formulation with HLL-and HLLC-type solvers towards fluidndashstructure interactionsrdquo Journal ofComputational Physics vol 324 pp 189ndash230 2016

[29] M Larcher F Casadei and G Solomos ldquoSimulation of blastwaves by usingmapping technology in EUROPLEXUSrdquo TechRep JRC91102 Joint Research Centre Publications Office ofthe European Union Luxembourg 2014

[30] E Lee J Horning and J Kury ldquoAdiabatic expansion of highexplosives detonation productsrdquo University of CaliforniaLawrence Livermore National Laboratory Livermore CAUSA 1968

[31] B Dobratz and P Crawford LLNL Explosives HandbookProperties of Chemical Explosives and Explosive SimulantsUniversity of California Lawrence Livermore NationalLaboratory Livermore CA USA 1985

Advances in Civil Engineering 11

case are united into one But as can be noticed even in thatphase the wave front of the ldquomeanderingrdquo case is weaker thanthe one of the ldquoopen doorrdquo case

e meandering wall reects part of the initial blast waveand guides it back to the entrance area resulting in a slightlystronger blast wave into the entrance zone Since this zone hasalready been signicantly ashyected from the initial blast wave theadditional energy due to the presence of the ldquomeanderingrdquo walldoes not ashyect the injury risk probability Figure 8 presents thepressure and the impulse history for the abovementioned casesat a point that is located at the middle of the protected zone7 meters from the opening and 15 meters from the ground

level From the pressure-time history it is clear that in theldquomeanderingrdquo case the rst peak is eliminated while the secondpeak that results from the reection of the wave onto the lateralwalls is lower Similar results can be derived from comparingthe impulse-time histories as there is a signicant reduction ofthe total impulse when the meandering wall is present

e next gures present the inuence of the dimensionand location of the meandering wall on the blast wavepropagation As has already been described in previoussections areas with an impulse value larger than 1000 Pamiddotsare considered as high injury risk zones Figure 9 presentsthe injury risk areas for the protected zone of the examined

E 05

D1

D2Impulse

30e+03

20e+03

10e+03

00e+00

D3

E 1 E 2

Figure 9 Zone of impulse higher than 1000 Pamiddots for the ldquomeanderingrdquo studied cases and the ldquoopen doorrdquo model (parameters D and E in mand impulse in pamiddots)

105 151

E (m)

D (m

)

()

2

15

2

25

3

60

40

20

60

40

20

0

Figure 10 Percentage of ashyected area spectrum versus the investigated parameters for the ldquomeanderingrdquo model

Advances in Civil Engineering 9

building for several studied geometries (e results refer tothe same horizontal cut that is presented in Figure 7 but theoutput quantity is the calculated impulse Areas charac-terized by impulse values higher than 1000 Pamiddots are high-lighted D and E determine the location and the size of themeandering wall as described earlier

In the same figure the ldquoopen doorrdquo case is also includedfor comparison purposes In the ldquoopen doorrdquo case themajority of the protected zone results in a high injury riskarea In the ldquomeanderingrdquo case the size of the high injuryrisk area varies but is always smaller than the ldquoopen doorrdquocase It is clear that the area before the additional wall isalways unprotected irrespective of the presence of themeandering wall Moreover in some cases high impulsevalues are also recorded at the sides of the meandering walldue to the relatively strong wave fronts that are present atboth those areas Figure 10 summarizes the output of theparametric study by presenting the percentage of the injuryrisk area in the protected zone (above 1000 Pamiddots) versus thetwo examined parameters From comparing the severalldquomeanderingrdquo cases it can be concluded that if D is small(distance between the door and the added wall) the influ-ence of E (dimension of overlapping length) is of minorimportance However if D is large the influence of E issignificant (if E decreases the blast wave that enters into theprotected zone is stronger and the high injury risk areas arebecoming larger)

5 Conclusions

Access control points that are located inside or attached tobuildings are often adopted due to local constraints (eirdesign must consider possible terrorist attack scenarios(eg explosions) at the interior and exterior of these accesscontrol points (e mitigation of the consequences shouldsuch a terrorist attack materialize inside an access point andin the attached building can be achieved by using venting(release surfaces) and in particular meandering walls (einvestigation shows that the reduction of the pressure issignificant if meandering walls are used and that these wallscan therefore play an important role in separating accesscontrol points from the attached buildings Nevertheless it isobvious that meandering walls can be considered an effectiverisk reduction measure and not as a full protection solution(e risk reduction concerns both blast wave effects and forobvious reasons fragments resulting from explosions andshooting attacks Herein the idealized situation that is in-vestigated can be extended to include additional parametersAs has been presented in the current article the presence ofa rigid wall standing between an explosionrsquos detonationcentre and a target can effectively reduce the severity of theimpacting blast wave It is therefore recommended theaddition of such elements (varying in size and geometry) inorder to mitigate the risk of terrorist attacks

Conflicts of Interest

(e authors declare that there are no conflicts of interestregarding the publication of this paper

References

[1] A M Lewis ldquoState of the art report on European legislationrelating to explosives and explosive detection system for non-aviation configurationsrdquo Tech Rep JRC84076 Joint ResearchCentre Ispra Italy 2013

[2] B M Luccioni R D Ambrosini and R F Danesi ldquoAnalysisof building collapse under blast loadsrdquo Engineering Structuresvol 26 no 1 pp 63ndash71 2004

[3] M Larcher M Arrigoni C Bedon et al ldquoDesign of blast-loaded glazing windows and facades a review of essentialrequirements towards standardizationrdquo Advances in CivilEngineering vol 2016 Article ID 2604232 14 pages 2016

[4] R Hajek and M Foglar ldquoNumerical and experimentalanalysis of the effect of rigid barriers on blast wave propa-gationrdquo Journal of Structural Engineering vol 141 no 12p 04015061 2015

[5] C N Kingery and G Bulmash ldquoAirblast parameters fromTNT spherical air burst and hemispherical surface burstrdquoTech Rep ARBRL-TR-02555 Defence Technical InformationCenter Ballistic Research Laboratory Aberdeen ProvingGround MD USA Tech Rep ARBRL-TR-02555 1984

[6] M Larcher ldquoPressure-time functions for the description of airblast wavesrdquo Tech Rep JRC46829 Joint Research CentreIspra Italy 2008

[7] V Karlos and G Solomos ldquoCalculation of blast loads forapplication to structural componentsrdquo Tech Rep EUR26456EN Joint Research Centre Publications Office of theEuropean Union Luxembourg 2013

[8] V Karlos G Solomos and M Larcher ldquoAnalysis of the blastwave decay coefficient using the Kingery-Bulmash datardquoInternational Journal of Protective Structures vol 7 no 3pp 409ndash429 2016

[9] M Larcher ldquoDevelopment of discrete cracks in concreteloaded by shock wavesrdquo International Journal for ImpactEngineering vol 36 pp 700ndash710 2009

[10] J Li and H Hao ldquoInternal and external pressure prediction ofvented gas explosion in large rooms by using analytical andCFD methodsrdquo Journal of Loss Prevention in the ProcessIndustries vol 49 pp 367ndash381 2017

[11] EN14494 ldquoGas explosion venting protective systemsrdquo 2007[12] M Larcher F Casadei and G Solomos ldquoInfluence of venting

areas on the air blast pressure inside tubular structures likerailway carriagesrdquo Journal of Hazardous Materials vol 183no 1-3 pp 839ndash846 2010

[13] P J Miller ldquoA reactive flow model with coupled reactionkinetics for detonation and combustion in non-ideal explo-sivesrdquo MRS Proceedings vol 418 1995

[14] M Larcher R Forsberg U Bjoernstig A Holgersson andG Solomos ldquoEffectiveness of finite element modelling ofdamage and injuries for explosions inside trainsrdquo Journal ofTransportation Safety and Security vol 8 pp 83ndash100 2015

[15] G Valsamos F Casadei M Larcher and G SolomosldquoImplementation of flying debris fatal risk calculation inEUROPLEXUSrdquo Tech Rep JRC94805 Joint Research CentrePublications Office of the European Union Luxembourg 2015

[16] E Gonzalez Ferradas F Dıaz Alonso M Doval MintildearroA Mintildeana Aznar J Ruiz Gimeno and J Sanchez PerezldquoConsequence analysis by means of characteristic curves todetermine the damage to humans from bursting sphericalvesselsrdquo Process Safety and Environmental Protection vol 86no 2 pp 121ndash129 2008

10 Advances in Civil Engineering

[17] S Mannan Leesrsquo Loss Prevention in the Process IndustriesHazard Identification Assessment and Control Vol 2 ElsevierAmsterdam 2005

[18] Department of Defence USA Unified Facilities Criteria UFC3-340ndash2 2008

[19] ldquoNATO STANAG 2280 design threat levels and handoverprocedure for temporary protective structuresrdquo 2008

[20] G F Kinney and K J Graham Explosive Shocks in AirSpringer Berlin Germany 1985

[21] D Bogosian J Ferritto and Y Shi ldquoMeasuring uncertaintyand conservatism in simplified blast modelsrdquo in Proceedingsof 30th Explosive Safety Seminar Atlanta Georgia August 2002

[22] M Larcher and F Casadei ldquoExplosions in complexgeometriesndasha comparison of several approachesrdquo InternationalJournal of Protective Structures vol 1 no 2 pp 169ndash195 2010

[23] M Larcher A Stolz L Rudiger and M Steyerer ldquoStatischeErsatzlasten und ldquoIngenieurverfahrenrdquordquo in Berichte ausdem Konstruktiven Ingenieurbau Workshop Bau-ProtectN Gebbeken K (oma M Rudiger and M Klaus Edspp 117ndash130 Universitat der Bundeswehr MunchenNeubiberg Germany 2012

[24] FEMA ldquoSite and urban design for securityndashguidance againstpotential terrorist attacksrdquo Tech Rep FEMA 430 FEMAWashington DC USA 2007

[25] Joint Research Centre and Commissariat a lrsquoenergie atomiqueet aux energies alternatives EUROPLEXUS httpwww-epxceafr

[26] Commissariat a lrsquoenergie atomique et aux energies alterna-tives CAST3M httpwww-cast3mceafr

[27] ParaView httpwwwparavieworg[28] F Daude and P Galon ldquoOn the computation of the baerndash

nunziato model using ALE formulation with HLL-and HLLC-type solvers towards fluidndashstructure interactionsrdquo Journal ofComputational Physics vol 324 pp 189ndash230 2016

[29] M Larcher F Casadei and G Solomos ldquoSimulation of blastwaves by usingmapping technology in EUROPLEXUSrdquo TechRep JRC91102 Joint Research Centre Publications Office ofthe European Union Luxembourg 2014

[30] E Lee J Horning and J Kury ldquoAdiabatic expansion of highexplosives detonation productsrdquo University of CaliforniaLawrence Livermore National Laboratory Livermore CAUSA 1968

[31] B Dobratz and P Crawford LLNL Explosives HandbookProperties of Chemical Explosives and Explosive SimulantsUniversity of California Lawrence Livermore NationalLaboratory Livermore CA USA 1985

Advances in Civil Engineering 11

building for several studied geometries (e results refer tothe same horizontal cut that is presented in Figure 7 but theoutput quantity is the calculated impulse Areas charac-terized by impulse values higher than 1000 Pamiddots are high-lighted D and E determine the location and the size of themeandering wall as described earlier

In the same figure the ldquoopen doorrdquo case is also includedfor comparison purposes In the ldquoopen doorrdquo case themajority of the protected zone results in a high injury riskarea In the ldquomeanderingrdquo case the size of the high injuryrisk area varies but is always smaller than the ldquoopen doorrdquocase It is clear that the area before the additional wall isalways unprotected irrespective of the presence of themeandering wall Moreover in some cases high impulsevalues are also recorded at the sides of the meandering walldue to the relatively strong wave fronts that are present atboth those areas Figure 10 summarizes the output of theparametric study by presenting the percentage of the injuryrisk area in the protected zone (above 1000 Pamiddots) versus thetwo examined parameters From comparing the severalldquomeanderingrdquo cases it can be concluded that if D is small(distance between the door and the added wall) the influ-ence of E (dimension of overlapping length) is of minorimportance However if D is large the influence of E issignificant (if E decreases the blast wave that enters into theprotected zone is stronger and the high injury risk areas arebecoming larger)

5 Conclusions

Access control points that are located inside or attached tobuildings are often adopted due to local constraints (eirdesign must consider possible terrorist attack scenarios(eg explosions) at the interior and exterior of these accesscontrol points (e mitigation of the consequences shouldsuch a terrorist attack materialize inside an access point andin the attached building can be achieved by using venting(release surfaces) and in particular meandering walls (einvestigation shows that the reduction of the pressure issignificant if meandering walls are used and that these wallscan therefore play an important role in separating accesscontrol points from the attached buildings Nevertheless it isobvious that meandering walls can be considered an effectiverisk reduction measure and not as a full protection solution(e risk reduction concerns both blast wave effects and forobvious reasons fragments resulting from explosions andshooting attacks Herein the idealized situation that is in-vestigated can be extended to include additional parametersAs has been presented in the current article the presence ofa rigid wall standing between an explosionrsquos detonationcentre and a target can effectively reduce the severity of theimpacting blast wave It is therefore recommended theaddition of such elements (varying in size and geometry) inorder to mitigate the risk of terrorist attacks

Conflicts of Interest

(e authors declare that there are no conflicts of interestregarding the publication of this paper

References

[1] A M Lewis ldquoState of the art report on European legislationrelating to explosives and explosive detection system for non-aviation configurationsrdquo Tech Rep JRC84076 Joint ResearchCentre Ispra Italy 2013

[2] B M Luccioni R D Ambrosini and R F Danesi ldquoAnalysisof building collapse under blast loadsrdquo Engineering Structuresvol 26 no 1 pp 63ndash71 2004

[3] M Larcher M Arrigoni C Bedon et al ldquoDesign of blast-loaded glazing windows and facades a review of essentialrequirements towards standardizationrdquo Advances in CivilEngineering vol 2016 Article ID 2604232 14 pages 2016

[4] R Hajek and M Foglar ldquoNumerical and experimentalanalysis of the effect of rigid barriers on blast wave propa-gationrdquo Journal of Structural Engineering vol 141 no 12p 04015061 2015

[5] C N Kingery and G Bulmash ldquoAirblast parameters fromTNT spherical air burst and hemispherical surface burstrdquoTech Rep ARBRL-TR-02555 Defence Technical InformationCenter Ballistic Research Laboratory Aberdeen ProvingGround MD USA Tech Rep ARBRL-TR-02555 1984

[6] M Larcher ldquoPressure-time functions for the description of airblast wavesrdquo Tech Rep JRC46829 Joint Research CentreIspra Italy 2008

[7] V Karlos and G Solomos ldquoCalculation of blast loads forapplication to structural componentsrdquo Tech Rep EUR26456EN Joint Research Centre Publications Office of theEuropean Union Luxembourg 2013

[8] V Karlos G Solomos and M Larcher ldquoAnalysis of the blastwave decay coefficient using the Kingery-Bulmash datardquoInternational Journal of Protective Structures vol 7 no 3pp 409ndash429 2016

[9] M Larcher ldquoDevelopment of discrete cracks in concreteloaded by shock wavesrdquo International Journal for ImpactEngineering vol 36 pp 700ndash710 2009

[10] J Li and H Hao ldquoInternal and external pressure prediction ofvented gas explosion in large rooms by using analytical andCFD methodsrdquo Journal of Loss Prevention in the ProcessIndustries vol 49 pp 367ndash381 2017

[11] EN14494 ldquoGas explosion venting protective systemsrdquo 2007[12] M Larcher F Casadei and G Solomos ldquoInfluence of venting

areas on the air blast pressure inside tubular structures likerailway carriagesrdquo Journal of Hazardous Materials vol 183no 1-3 pp 839ndash846 2010

[13] P J Miller ldquoA reactive flow model with coupled reactionkinetics for detonation and combustion in non-ideal explo-sivesrdquo MRS Proceedings vol 418 1995

[14] M Larcher R Forsberg U Bjoernstig A Holgersson andG Solomos ldquoEffectiveness of finite element modelling ofdamage and injuries for explosions inside trainsrdquo Journal ofTransportation Safety and Security vol 8 pp 83ndash100 2015

[15] G Valsamos F Casadei M Larcher and G SolomosldquoImplementation of flying debris fatal risk calculation inEUROPLEXUSrdquo Tech Rep JRC94805 Joint Research CentrePublications Office of the European Union Luxembourg 2015

[16] E Gonzalez Ferradas F Dıaz Alonso M Doval MintildearroA Mintildeana Aznar J Ruiz Gimeno and J Sanchez PerezldquoConsequence analysis by means of characteristic curves todetermine the damage to humans from bursting sphericalvesselsrdquo Process Safety and Environmental Protection vol 86no 2 pp 121ndash129 2008

10 Advances in Civil Engineering

[17] S Mannan Leesrsquo Loss Prevention in the Process IndustriesHazard Identification Assessment and Control Vol 2 ElsevierAmsterdam 2005

[18] Department of Defence USA Unified Facilities Criteria UFC3-340ndash2 2008

[19] ldquoNATO STANAG 2280 design threat levels and handoverprocedure for temporary protective structuresrdquo 2008

[20] G F Kinney and K J Graham Explosive Shocks in AirSpringer Berlin Germany 1985

[21] D Bogosian J Ferritto and Y Shi ldquoMeasuring uncertaintyand conservatism in simplified blast modelsrdquo in Proceedingsof 30th Explosive Safety Seminar Atlanta Georgia August 2002

[22] M Larcher and F Casadei ldquoExplosions in complexgeometriesndasha comparison of several approachesrdquo InternationalJournal of Protective Structures vol 1 no 2 pp 169ndash195 2010

[23] M Larcher A Stolz L Rudiger and M Steyerer ldquoStatischeErsatzlasten und ldquoIngenieurverfahrenrdquordquo in Berichte ausdem Konstruktiven Ingenieurbau Workshop Bau-ProtectN Gebbeken K (oma M Rudiger and M Klaus Edspp 117ndash130 Universitat der Bundeswehr MunchenNeubiberg Germany 2012

[24] FEMA ldquoSite and urban design for securityndashguidance againstpotential terrorist attacksrdquo Tech Rep FEMA 430 FEMAWashington DC USA 2007

[25] Joint Research Centre and Commissariat a lrsquoenergie atomiqueet aux energies alternatives EUROPLEXUS httpwww-epxceafr

[26] Commissariat a lrsquoenergie atomique et aux energies alterna-tives CAST3M httpwww-cast3mceafr

[27] ParaView httpwwwparavieworg[28] F Daude and P Galon ldquoOn the computation of the baerndash

nunziato model using ALE formulation with HLL-and HLLC-type solvers towards fluidndashstructure interactionsrdquo Journal ofComputational Physics vol 324 pp 189ndash230 2016

[29] M Larcher F Casadei and G Solomos ldquoSimulation of blastwaves by usingmapping technology in EUROPLEXUSrdquo TechRep JRC91102 Joint Research Centre Publications Office ofthe European Union Luxembourg 2014

[30] E Lee J Horning and J Kury ldquoAdiabatic expansion of highexplosives detonation productsrdquo University of CaliforniaLawrence Livermore National Laboratory Livermore CAUSA 1968

[31] B Dobratz and P Crawford LLNL Explosives HandbookProperties of Chemical Explosives and Explosive SimulantsUniversity of California Lawrence Livermore NationalLaboratory Livermore CA USA 1985

Advances in Civil Engineering 11

[17] S Mannan Leesrsquo Loss Prevention in the Process IndustriesHazard Identification Assessment and Control Vol 2 ElsevierAmsterdam 2005

[18] Department of Defence USA Unified Facilities Criteria UFC3-340ndash2 2008

[19] ldquoNATO STANAG 2280 design threat levels and handoverprocedure for temporary protective structuresrdquo 2008

[20] G F Kinney and K J Graham Explosive Shocks in AirSpringer Berlin Germany 1985

[21] D Bogosian J Ferritto and Y Shi ldquoMeasuring uncertaintyand conservatism in simplified blast modelsrdquo in Proceedingsof 30th Explosive Safety Seminar Atlanta Georgia August 2002

[22] M Larcher and F Casadei ldquoExplosions in complexgeometriesndasha comparison of several approachesrdquo InternationalJournal of Protective Structures vol 1 no 2 pp 169ndash195 2010

[23] M Larcher A Stolz L Rudiger and M Steyerer ldquoStatischeErsatzlasten und ldquoIngenieurverfahrenrdquordquo in Berichte ausdem Konstruktiven Ingenieurbau Workshop Bau-ProtectN Gebbeken K (oma M Rudiger and M Klaus Edspp 117ndash130 Universitat der Bundeswehr MunchenNeubiberg Germany 2012

[24] FEMA ldquoSite and urban design for securityndashguidance againstpotential terrorist attacksrdquo Tech Rep FEMA 430 FEMAWashington DC USA 2007

[25] Joint Research Centre and Commissariat a lrsquoenergie atomiqueet aux energies alternatives EUROPLEXUS httpwww-epxceafr

[26] Commissariat a lrsquoenergie atomique et aux energies alterna-tives CAST3M httpwww-cast3mceafr

[27] ParaView httpwwwparavieworg[28] F Daude and P Galon ldquoOn the computation of the baerndash

nunziato model using ALE formulation with HLL-and HLLC-type solvers towards fluidndashstructure interactionsrdquo Journal ofComputational Physics vol 324 pp 189ndash230 2016

[29] M Larcher F Casadei and G Solomos ldquoSimulation of blastwaves by usingmapping technology in EUROPLEXUSrdquo TechRep JRC91102 Joint Research Centre Publications Office ofthe European Union Luxembourg 2014

[30] E Lee J Horning and J Kury ldquoAdiabatic expansion of highexplosives detonation productsrdquo University of CaliforniaLawrence Livermore National Laboratory Livermore CAUSA 1968

[31] B Dobratz and P Crawford LLNL Explosives HandbookProperties of Chemical Explosives and Explosive SimulantsUniversity of California Lawrence Livermore NationalLaboratory Livermore CA USA 1985

Advances in Civil Engineering 11

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Active and Passive Electronic Components

VLSI Design

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Shock and Vibration

Hindawiwwwhindawicom Volume 2018

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom

Journal ofEngineeringVolume 2018

SensorsJournal of

Hindawiwwwhindawicom Volume 2018

International Journal of

RotatingMachinery

Hindawiwwwhindawicom Volume 2018

Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Navigation and Observation

International Journal of

Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom