Assessment of displacement ventilation systems in airborne infection risk in hospital rooms

RESEARCH ARTICLE Assessmen o displacemen en ila ion sys ems in ai bo ne in ec ion isk in hospi al ooms Jose ´Manuel Villa uelaID 1 *, Ine ´s Olmedo 2 , Fe ´lix A. Be langaID 2 , Manuel Ruiz de AdanaID 2 1ITAP, Depa men o Ene gy and Fluid Mechanics, Uni e si y o Valladolid, Valladolid, Spain, 2Depa men o Physical Chemis y and Applied The modynamics, Uni e si y o Co doba, Co ´ doba, Spain *[email p o ec ed] Abs ac E icien en ila ion in hospi al ai bo ne isola ion ooms is impo an is-à- is dec easing he isk o c oss in ec ion and educing ene gy consump ion. This pape analyses he sui abili y o using a displacemen en ila ion s a egy in ai bo ne in ec ion isola ion ooms, ocusing on heal h ca e wo ke exposu e o pa hogens exhaled by in ec ed pa ien s. The analysis is mainly based on nume ical simula ion esul s ob ained wi h he suppo o a 3-D ansien nume ical model alida ed using expe imen al da a. A he mal b ea hing manikin lying on a bed ep esen s he sou ce pa ien and ano he he mal b ea hing manikin ep esen s he exposed indi idual s anding beside he bed and acing he pa ien . A adian wall ep esen s an ex e nal wall exposed o sola adia ion. The ai change e iciency index and con aminan emo al e ec i eness indices and inhala ion by he heal h ca e wo ke o con aminan s exhaled by he pa ien a e conside ed in a ypical ai bo ne in ec ion isola ion oom se up wi h h ee ai enewal a es (6 h -1 , 9 h -1 and 12 h -1 ), wo exhaus opening posi ions and wo heal h ca e wo ke posi ions. Resul s show ha he adian wall signi ican ly a ec s he ai low pa e n and con aminan dispe sion. The lockup phenomenon occu s a he inhala ion heigh o he s anding manikin. Displacemen en ila ion enews he ai o he ai bo ne isola- ion oom and elimina es he exhaled pollu an s e icien ly, bu is a a disad an age com- pa ed o o he en ila ion s a egies when he isk o exposu e is aken in o accoun . In oduc ion Hospi al acili ies a e places wi h a high isk o c oss in ec ion be ween hei occupan s. S udies in Eu opean hospi als [1] indica e ha nosocomial in ec ions con ibu e signi ican ly o mo - bidi y and mo ali y a es and ha many o hese in ec ions a e ansmi ed by ai bo ne pa ho- gens [2]. E e yday pulmona y ac i i ies, like b ea hing [3], coughing [4,5], sneezing [6], alking [3], a e sou ces o bio-ae osols [7] ha may be laden wi h he pa hogens esponsible o in ec ious disease ansmission. Once he bio-ae osols lea e he in ec ed pe son hei a e depends on mul iple and complex ac o s [7–10]. One o he mos impo an ac o s is PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0211390 Janua y 30, 2019 1 / 18 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Ci a ion: Villa uela JM, Olmedo I, Be langa FA, Ruiz de Adana M (2019) Assessmen o displacemen en ila ion sys ems in ai bo ne in ec ion isk in hospi al ooms. PLoS ONE 14(1): e0211390. h ps://doi.o g/10.1371/jou nal. pone.0211390 Edi o : Es eban Tlelo-Cuau le, Ins i u o Nacional de As o isica Op ica y Elec onica, MEXICO Recei ed: Oc obe 31, 2018 Accep ed: Janua y 11, 2019 Published: Janua y 30, 2019 Copy igh : ©2019 Villa uela e al. This is an open access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal au ho and sou ce a e c edi ed. Da a A ailabili y S a emen : All ele an da a a e wi hin he manusc ip . Funding: The au ho s acknowledge he inancial suppo ecei ed om he Minis y o Finance and Compe i i eness, S a e Sec e a ia o Resea ch, De elopmen and Inno a ion, Spain, he Na ional R&D p ojec wi h e e ence DPI2014-55357-C2-1- R, en i led “Ven ila ion sys em in luence on ai bo ne ansmission o human exhaled bioae osols. C oss in ec ion isk e alua ion”. This p ojec is co- inanced by he Eu opean Regional undoub edly he ai low pa e n, bo h in he oom as a whole as well as in he mic oen i on- men a ound he sou ce pa ien and he ulne able indi idual. Ai bo ne in ec ion isola ion oom I he app op ia e measu es a e no aken, he bio-ae osols emi ed by pa ien s hospi alized wi h an ai bo ne disease may be dispe sed uncon ollably a ound he ai bo ne in ec ion isola- ion oom (AIIR) o he es o he hospi al [11]. Di e en me hods and echnologies a e a ail- able o p o ide adequa e p o ec ion o people who pass h ough a hospi al [12]. One o he ecommended measu es is o main ain a nega i e p essu e wi h espec o he su ounding a ea so ha ai lows in o he oom and no in he opposi e di ec ion when doo s a e open. Un o una ely, nega i e p essu e b ie ly disappea s du ing doo ope a ion and ai leakage is i ually ine i able [13]. Many guidelines and egula ions [14–18] ela ed o ai bo ne isola ion ooms (AIIR) ad ise o equi e ha access o he oom should be h ough an an e oom in o de o minimize escape o con amina ed ai [19,20]. Ye nei he he nega i e p essu e no he an e oom p e en s he isk o he pe son en e ing he AIIR. Only pe sonal sel -p o ec ion measu es and a sui able en ila ion s a egy educe he possibili y o con agion [21]. Wi h ega d o en ila ion, one common ecommenda ion is o use high enewal a es o dilu e and emo e pa hogens [22]. Howe e , his does no p e en he appea ance o s agnan zones and sho -ci cui ing, esul ing in “clean” and “pollu ed” a eas o exhaled pa hogens wi h he subsequen isk o high c oss-in ec ion a es. Se e al s udies indica e ha he design o a en ila ion sys em and he esul ing ai low pa e ns play a mo e impo an ole han jus ai enewal a es alone [23,24]. Ai low pa e ns gene a ed by en ila ion sys ems can be con- olled, and ecen esea ch has ocused on p o iding good ai dis ibu ion a he han on main aining high a es o ai enewal as a s a egy o educe he isk o ai bo ne con agion [25–30]. Displacemen en ila ion Va ious en ila ion s a egies such as mixing en ila ion (MV) and displacemen en ila ion (DV) o e di e en possibili ies o p o ec people om ai bo ne c oss in ec ion [10,31]. MV is he mos widely applied s a egy in hospi al pa ien ooms. Howe e , in ecen yea s DV has eme ged as an al e na i e. Some s udies ha e shown ha DV is mo e ene gy e icien [32]. S anda disa ion associa ions ha e de eloped DV guidelines and ecommenda ions o designe s. DV sys ems we e ini ially de eloped o emo e he mal loads in indus ial wa ehouses due o hei abili y o concen a e hea and pollu an s abo e he occupied zone. DV sys ems a e cha ac e ized by he mal and mass s a i ica ion such ha hey canno be modelled wi h he ully mixed oom ai app oach [33]. In DV, cool ai is supplied in o he lowe pa o he oom using low impulse di use s. This slow mo ing esh ai ills he oom om below, is hea ed and ises o he ceiling, whe e he exhaus is loca ed. The e mus be hea sou ces o DV o wo k. As b ea hing is also a hea and pollu an sou ce, con aminan s migh be anspo ed di ec ly o he uppe pa o he oom. DV o e s he possibili y o wo king wi h wo zones, a low zone wi h clean ai , and an uppe zone wi h pollu an s. Some au ho s epo ha i is pos- sible o design DV hospi al pa ien ooms ha ha e low human exposu e o bio-ae osols ha con aining pa hogens [32,34], al hough in ce ain si ua ions high exposu e may also exis in ooms wi h DV [35–37]. Assessmen o displacemen en ila ion sys ems in ai bo ne in ec ion isk in hospi al ooms PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0211390 Janua y 30, 2019 2 / 18 De elopmen Fund (ERDF). The unde s had no ole in s udy design, da a collec ion and analysis, decision o publish, o p epa a ion o he manusc ip . Compe ing in e es s: The au ho s ha e decla ed ha no compe ing in e es s exis . Abb e ia ions: ACH, Ai change pe hou , h -1 ; AIIR, Ai bo ne in ec ion isola ion oom; B, Bed; CFD, Compu a ional luid dynamics; DD, Displacemen di use ; DV, Displacemen en ila ion; E, Exhaus openings loca ed on he uppe pa o he eas wall; EN, Ven ila ion sys em con igu a ion combining E and N; ES, Ven ila ion sys em con igu a ion combining E and S; HCW, Heal h ca e wo ke ; IF, In ake ac ion index; IF PMV , In ake ac ion index wi h PMV; MV, Mixing en ila ion; N, Ex e nal wall hea gain on no h wall; P, Pa ien ; PMV, Pe ec mixing en ila ion; PDV, Pe ec displacemen en ila ion; Q, Ai low a e in he oom, m 3 /s; RANS, Reynolds-A e aged Na ie -S okes equa ions; S, Ex e nal wall hea gain on sou h wall; SM, Sou ce manikin (P); TM, Ta ge manikin (HCW); URANS, Uns eady-RANS; W, Exhaus openings loca ed on he uppe pa o he wes wall; WN, Ven ila ion sys em con igu a ion combining W and N; WS, Ven ila ion sys em con igu a ion combining W and S; c, A e age concen a ion o con aminan in he oom; c e , Con aminan concen a ion in he exhaus ai ; D a , Mass di usion coe icien , m 2 /s; ε a , Ai change e iciency index; ε c , Con aminan emo al e ec i eness; l 0 , La ges scale o he oom, m; τ�, Local mean age o he ai , s; τ n , Nominal ime cons an , s; V, Room olume, m 3 ; 0 , Cha ac e is ic ai eloci y, m/s; Y P , Mass ac ion o N 2 O in he ai exhaled by P; Y PMV , Mass ac ion o N 2 O in a oom wi h PMV. Objec i e and me hodology The aim o his wo k is o e alua e he sui abili y o applying he DV s a egy in AIIRs. The analysis is mainly based on nume ical simula ion esul s ob ained wi h he suppo o compu- a ional luid dynamics (CFD). The in e ac ion be ween he di e en ai lows -b ea hing lows [38], con ec i e lows a ound human bodies [39], he mal plumes abo e hea sou ces, ising bounda y laye low a he wa m wall oge he wi h la ge-scale ai mo emen s due o oom ai low ins abili ies- is so complex ha i is di icul o app oach he p oblem di ec ly. Ini ially, dispe sion o con aminan s exhaled by a single pe son s anding in an indoo en i onmen was s udied [40]. La e , a second pe son acing he i s was added o analyse he in e ac ion be ween he espi a ion lows o bo h people [41]. These wo p e ious s udies ha e enabled an adequa e p ocedu e o be es ablished o analysing he ole o en ila ion in he isk o c oss- in ec ion be ween pa ien and suscep ible heal h ca e wo ke caused by he ai bo ne pa hogens exhaled du ing b ea hing in an AIIR wi h DV. Using he alida ed model, wel e di e en nume ical es s a e ca ied ou o analyse how ai enewal a es, he posi ion o he heal h ca e wo ke , and he posi ion o ai exhaus openings a ec he isk o c oss in ec ion. Tes oom and expe imen al se up The expe imen al s udy o a pa ien (P) lying on a hospi al bed and a heal h ca e wo ke (HCW) s anding close o he bed in a ypical AIIR oom (Fig 1) [14,15,42,43] was ca ied ou in a es oom a Co doba Uni e si y, 4.5 m (long), 3.3 m (wide), and 2.8 m (high). The wo he mal b ea hing manikins ha e he same geome y. The o al sensible hea emi ed o each manikin co esponds o a me abolic a e o 1 me o he HCW and 0.7 me o he pa ien , 80 W and 70 W, espec i ely. The e is an ex e nal hea gain o 500 W in he 4.5 m wall opposi e he HCW, which ep esen s an ex e nal wall exposed o sola adia ion. The emaining walls as well as he loo and ceiling a e adiaba ic as he chambe is inside a lab a he same empe a u e. A displacemen low di use (QLV-180-200-800, T ox, Ge many) a e used as supply ai uni o he hospi al oom, and wo exhaus openings we e loca ed on he opposi e wall, jus below he ceiling. The en ila ion sys em was se a h ee di e en ai change a es o 12, 9 and 6 ACH, supplying ai a 21.8˚C, 20.6˚C and 18.2˚C, espec i ely o main ain he same mean oom empe a u e. Pa o he e ec i e a ea o he displacemen di use is pa ially co e ed du ing he 9 ACH and 6 ACH es s in o de o main ain he same supply eloci y. In o ma ion abou b ea hing manikins, measu ing ins umen s and o he s de ails o hese expe imen s can be ound in [44]. Fig 1. Expe imen al se up. Nume ical se up o EN cases (Table 1). h ps://doi.o g/10.1371/jou nal.pone.0211390.g001 Assessmen o displacemen en ila ion sys ems in ai bo ne in ec ion isk in hospi al ooms PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0211390 Janua y 30, 2019 3 / 18 Nume ical simula ion Indices o quan i ying he en ila ion and in ec ion isk The en ila ion sys em o a oom can pu sue di e en aims: he mal com o , ai enewal, elimina ion o gaseous o suspended con aminan s, a oiding isk o in ec ion, e c. Depending on he ac i i ies ca ied ou in he oom, one aim o ano he will p e ail. Speci ic indices exis o quan i y he ex en o which each aim is achie ed. The mos basic index is ai changes pe hou (ACH). To calcula e ACH, only he ai olume o he oom and he ai low a e need o be known. This index is commonly used in guidelines and ecommenda ions. The ai change e iciency index (ε a ) is de ined as he a io be ween he minimum and he ac ual mean eplacemen imes and can be calcula ed om he exp ession: εa¼ n 2 ð1Þ whe e τ n =V/Q,Vis he oom olume and Qis he low a e o esh ai , i.e. τ n is he in e se o he numbe o ai changes pe second, and is he a e age age o he ai in he oom. Ai change e iciency only depends on he o e all ai low pa e n in he oom, and akes alues be ween 0 and 1. I , in addi ion o knowing he ACH and he ai low pa e n, he cha ac e is ics o he con aminan and he poin o emission a e also known, hen he con aminan emo al e ec- i eness index (ε c ) can be used. This index can ake any posi i e alue. Assuming ha he ai supplied o he oom is con aminan ee and ha he low is s eady, ε c is calcula ed by di iding he concen a ion o pollu an in he exhaus ai c e by he a e age concen a ion in oom c: εc¼ce cð2Þ Finally, i he a ea o be p o ec ed is known - he su ace o a p in ed ci cui du ing manu- ac u e, he ins umen able du ing a su gical ope a ion o he lungs o a pe son sha ing a oom wi h ano he in ec ed pe son- he in ake ac ion (IF) index may be used. This index is he low a e o con aminan ha c osses he su ace o be p o ec ed di ided by he low a e o con aminan ha en e s o is gene a ed inside he oom (Benne e al., 2002). In o de o assess he isk o c oss-in ec ion, he in ake ac ion is de ined as he p opo ion o he cumula- i e mass o con aminan inhaled by he HCW o he mass o con aminan emi ed in he pa ien ’s exhala ion du ing he same pe iod o ime. IF ¼RQHCWYHCW d RQPYPd ð3Þ Whe e Q HCW , and Q P a e he ins an aneous b ea hing low a e o HCW and P, espec- i ely, Y HCW is he ins an aneous mass ac ion o N 2 O in he HCW inhala ion ai , and Y P is he ins an aneous mass ac ion o N 2 O in he pa ien ’s exhala ion ai . Go e ning equa ions Ai bo ne c oss in ec ion be ween occupan s is uns eady, non-iso he mal and is a h ee-dimen- sional p oblem in ol ing wo species: ai and con aminan . As modelling ool, CFD has been applied o simula e he uns eady ai low using he URANS me hod oge he wi h he RNG k–e u bulence model equa ions, mean age o ai equa ion, he N 2 O mass ac ion equa ion and includes he e ec o he mal adia ion using comme cial so wa e Ansys Fluen . Assessmen o displacemen en ila ion sys ems in ai bo ne in ec ion isk in hospi al ooms PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0211390 Janua y 30, 2019 4 / 18 The local mean age o ai τin he whole luid ield is calcula ed sol ing he ollowing conse - a ion equa ion: @ @ þ !Da � �¼1ð4Þ whe e !is he ai eloci y and D a is he mass di usion coe icien [45]. A sub ou ine sol ing Eq (4) nume ically is w i en, and he sub ou ine is buil in o he CFD-p og am. Once he a e age age o he ai in he whole oom is calcula ed, i is possible o di ec ly e alua e ε a acco ding o Eq (1). is he a e age o τin he whole oom and τ n can be calcula ed as he in e se o he numbe o ai changes pe second o as he a e age o τin he ai ex ac ions. The wo alues coincide. The CFD-p og am models he mixing and anspo o wo chemical species, ai and N 2 O, by sol ing equa ions desc ibing con ec ion and di usion o each componen species wi hou eac ions. ε c and IF can be calcula ed di ec ly om hei de ini ions. Since he model is an- sien , he ime e olu ion o ε c and IF can be calcula ed. Radia ion is in oduced in o he CFD model using he su ace- o-su ace adia ion model [46]. The impo ance o he mal adia ion in ai low wi h DV was examined expe imen ally in [47]. The RNG k–εmodel ha akes in o accoun he low Reynolds-numbe e ec s in con- junc ion wi h enhanced wall ea men ha combines a wo-laye model wi h enhanced wall unc ions a e used in hese simula ions. P essu e- eloci y coupling was esol ed using he PISO scheme. A second-o de implici ansien o mula ion is chosen which is uncondi ion- ally s able wi h espec o ime-s ep size. A second-o de upwind disc e iza ion scheme is used o all equa ions [46]. In ansien simula ions (also in expe imen s) an e o migh occu du ing s a -up and when le ing simula ions un su icien ime o achie e cha ac e is ic la ge eddy u no e ime [48]. The ini ial condi ions o non-s eady compu a ions a e ob ained o a s eady simula ion. The i s 30 minu es o he ansien simula ion a e disca ded. La ge eddy u no e ime is a cha ac e is ic imescale o he domain l 0 / 0 whe e l 0 is he la ges scale o he oom and 0 is he cha ac e is ic eloci y. An es ima ion o 0 can be made by di iding he en ila ion low a e by a hal sec ion o he es chambe , which gi es a la ge eddy u no e ime o i e min- u es o 6 ACH. A la ge eddy u no e ime is in e sely p opo ional o ACH i he emaining Table 1. Condi ions o he CFD es pe o med. Simula ion nomencla u e Exhaus openings wall Radian wall ACH Supply ai low (m 3 /h) Supply ai empe a u e (˚C) Simula ion ime (min) EN_12�Eas No h 12 500 21.8 10 ES_12 Eas Sou h WN_12 Wes No h WS_12 Wes Sou h EN_09�Eas No h 9 375 20.6 15 ES_09 Eas Sou h WN_09 Wes No h WS_09 Wes Sou h EN_06�Eas No h 6 250 18.2 20 ES_06 Eas Sou h WN_06 Wes No h WS_06 Wes Sou h (�) Expe imen al es h ps://doi.o g/10.1371/jou nal.pone.0211390. 001 Assessmen o displacemen en ila ion sys ems in ai bo ne in ec ion isk in hospi al ooms PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0211390 Janua y 30, 2019 5 / 18 pa ame e s emain unchanged. In o de o ob ain a sui able empo al a e age, o al imes o 20, 15 and 10 minu es a e simula ed o 6, 9 and 12 ACH, espec i ely (Table 1). To cap u e he e ec s o he smalle ime scales ela ed o he b ea hing p ocess, a ime s ep o 0.02 s is selec ed. Compu a ional domain The domain o he compu a ional model mimics in de ail he expe imen al geome y o he li e-size hospi al isola ion oom. Mos o he domain is buil wi h a hexahed on mesh. A e a- hed al mesh has been employed nea he di use and nea o he manikin’s su ace due o hei geome y complexi y. Mesh e inemen was pe o med a ound he manikins, he exhaus s, he walls, and he displacemen di use since high eloci y and empe a u e g adi- en s a e expec ed. The high concen a ion, eloci y and empe a u e g adien s equi e a e y ine local g id sys em a he manikins’ aces and in he exhala ion zones [49]. The shape o he manikins o he nume ical simula ions ep oduces a he mal b ea hing manikin used by o h- e s au ho s [50,51]. De ailed in o ma ion abou he manikins’ shape and mesh can be ound in [41]. A sensi i i y s udy was ca ied ou wi h successi e e inemen s o he exhala ion zones and a ound he displacemen di use . A inal mesh o nea ly one and a hal million cells is used. Bounda y condi ions The CFD model o a pa ien and a HCW in AIIR ai h ully ep oduces he expe imen al condi- ions [44]. The lying manikin (also known as he pa ien o sou ce manikin, SM) exhales h ough he mou h and inhales h ough he nose. The s anding manikin (also known as he HCW o a ge manikin, TM) exhales and inhales h ough he nose. B ea hing unc ions a e a e y impo an poin in hese simula ions [52]. The wo manikins b ea he ollowing a sinusoi- dal unc ion. Fo he pa ien manikin, he idal olume is 0.57 li es and he b ea hing e- quency is 20 b ea hs/minu e. Fo he HCW manikin, he idal olume is 0.66 li es and he b ea hing equency is 15 b ea hs/minu e. The pa ien manikin hus pe o ms ou ull b ea hs in a 12-second pe iod and he HCW manikin h ee ull b ea hs du ing he same pe iod. Veloc- i y is no mal and uni o m in he HCW’s nos ils and in he pa ien ’s mou h. The empe a u e o expi ed ai in he wo manikins is 34˚C. The mass ac ion o N 2 O in he exhaled ai o he pa ien manikin is Y P = 0.027. The bounda y condi ion o he displacemen di use is a uni o m eloci y o 0.926 m/s, which is no mal o e ical di use su aces. Fo he 12 ACH simula ions, he en i e on a ea o he displacemen di use is used as an inle . The uppe qua e and uppe hal o he displacemen di use on a ea a e conside ed as walls o he 9 and 6 ACH simula ions, espec i ely, in o de o main ain he same inle eloci y [53]. The e ec i e a ea o he displace- men di use is aken in o accoun adding he co esponding momen um/ olume sou ce in a sub-domain in on o he di use [54]. The ai lea es he oom h ough wo exhaus openings loca ed in he same wall as he dis- placemen di use (wes ) o in he opposi e wall (eas ). A p essu e-ou le bounda y condi ion is imposed in he exhaus openings. In o de o main ain he same mean indoo empe a u e o all es s, he ai is supplied a a empe a u e o 21.8˚C, 20.6˚C, and 18.2˚C o 12, 9 and 6 ACH, espec i ely. The ceiling, loo and all he walls excep one 4.5 m ×2.8 m wall (no h o sou h wall) a e conside ed adiaba ic. In his non-adiaba ic wall a hea low o 39.7 W/m 2 is imposed as he bounda y condi ion. This ep esen s a glazed wall wi h a ansmission coe icien o 4 Wm -2 K -1 wi h 10˚C Assessmen o displacemen en ila ion sys ems in ai bo ne in ec ion isk in hospi al ooms PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0211390 Janua y 30, 2019 6 / 18 empe a u e di e ence be ween indoo and ou doo ai . The lying and s anding manikins ha e a he mal load o 70 W and 80 W, espec i ely. Valida ion o nume ical esul s wi h expe imen al da a In ooms wi h displacemen en ila ion, he mal s a i ica ion is gene a ed. The empe a u e o he exhaled ai is highe han he empe a u e o he ambien ai . The e ec s o buoyancy will play an impo an ole in bo h he e olu ion o exhaled ai and he dispe sion o he associ- a ed pa hogens. I is he e o e c ucial o co ec ly p edic he oom’s empe a u e ield. This nume ical model was p e iously used o s udy he dispe sion o exhaled con aminan s by he mou h o a pe son s anding in a oom wi h DV. The nume ical esul s we e compa ed o da a om expe imen s pe o med in a ull-scale labo a o y a he Uni e si y o Aalbo g [55]. I was ound ha he nume ical model was able o accu a ely ep oduce bo h he he mal s a i ica- ion in he oom and he de lec ion o he exhaled ai je . De ails on he alida ion o he em- pe a u e ield a e shown in [40]. Exhaled pa hogens do no dispe se in he same way when expi ed h ough he nose o mou h. How he ulne able indi idual b ea hes also a ec s he mic oen i onmen nea he ace and, he e o e, he isk o inhaling pa hogens. This CFD model was used in a p e ious s udy o analyse how he way o b ea hing in luences he isk o c oss-in ec ion be ween wo people s anding acing each o he a di e en dis ances in a oom wi h DV. The esul s o he CFD simula ions we e compa ed o expe imen al da a ob ained in he Albo g labo a o y wi h wo manikins a di e en dis ances and exhaling h ough he mou h and inhaling h ough he nose. De ails o his new alida ion a e shown in [41]. The nume ical model is now alida ed again wi h he expe imen al esul s ob ained in a eal-scale labo a o y a he Uni e si y o Co ´ doba [44]. Compa ing he esul s o wo global en ila ion e iciency indices, he ai change e iciency ε a and he con aminan emo al e ec- i eness ε c clea ly e idences how he nume ical model is able o cap u e he expe imen al en- dencies and o ep oduce he alues o a easonable deg ee o concu ence (Fig 2). The expe imen al alues o con aminan emo al e ec i eness a e sligh ly highe han he nume i- cal alues. The possibili y ha hese di e ences a e due o di icul ies ha eme ge when mea- su ing hese indexes using pho oacous ic spec oscopy canno be uled ou . Resul s and discussion Tempo al e olu ion o con aminan inhaled The exhala ion, dispe sion and inhala ion o con aminan s a e ansien phenomenon. Nume ical da a p o ide a de ailed empo al e olu ion o he amoun o con aminan inhaled Fig 2. Expe imen al and nume ical ai change e iciency index. a) Ai change e iciency. b) Con aminan emo al e ec i eness. EN cases. h ps://doi.o g/10.1371/jou nal.pone.0211390.g002 Assessmen o displacemen en ila ion sys ems in ai bo ne in ec ion isk in hospi al ooms PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0211390 Janua y 30, 2019 7 / 18 by he HCW. A b ea hing cycle (inhala ion and exhala ion) las s h ee seconds o he P and ou seconds o he HCW. Each 12 seconds bo h s a a he same ime, al hough he wo b ea hs a e ou o phase h oughou he cycle. This p og essi e phase displacemen migh cause he amoun o con aminan inhaled by he HCW in each o he h ee cycle inhala ions o di e . The phase-a e aged me hod was used in Fig 3 o show he concen a ion o N 2 O inhaled by he HCW in 12-second cycles. I is wo h no icing ha he a e age amoun o con aminan inhaled in he h ee inhala ions is he same. This is a clea indica ion ha he in e ac ion be ween he b ea hs o he wo manikins is e y weak due o he g ea dis ance be ween he b ea hing zones o he wo manikins [41]. Be ween he mou h o he pa ien and he nose o he HCW he e is a dis ance o 0.94 m in a s aigh line. The nose o he HCW is 0.53 m away om he pa ien ’s exhala ion axis. I should also be men ioned ha he amoun o pollu an inhaled in each 12-second cycle changes signi ican ly. In o he wo ds, cyclical dispe sion is e y no iceable, especially o low ACH. This esul shows ha he dispe sion o exhaled con aminan s and hei subsequen inhala ion a e ansien phenomena due o he ansien na u e o he ai low pa e n in he a ea be ween P and HCW. The o e all ai low in he oom also exhibi s a ansien beha iou wi h oom ai low equencies lowe han he manikins’ b ea hing equencies. These esul s a e in line wi h p e ious wo ks [41,56]. Ai low pa e n. Ai change e iciency In a pe ec mixing en ila ion (PMV) low, ai composi ion is equal h oughou he whole oom and no con aminan concen a ion g adien s a e p esen . PMV implies in ini ely apid di usion and pe ec displacemen en ila ion PDV implies absolu e absence o di usion [45]. PMV and PDV a e idealized heo e ical low pa e ns ha ne e occu in p ac ice bu which a e, ne e heless, use ul concep s o compa a i e pu poses. In a pe ec mixing en ila ion low ε a = 0.5. In a pe ec displacemen en ila ion low ε a = 1. As expec ed, in all he cases analysed, 0.5<ε a <1.0, when using he DV s a egy. The e is a clea co ela ion be ween he en ila ion e iciency and he posi ion o he ai exhaus openings Fig 3. Concen a ion o N 2 O inhaled by he HCW o he case o EN_09. Solid lines: phase-a e age alues, do ed lines: maximum alue, dashed lines: minimum alue co esponding o 75 cycles o 12s. h ps://doi.o g/10.1371/jou nal.pone.0211390.g003 Assessmen o displacemen en ila ion sys ems in ai bo ne in ec ion isk in hospi al ooms PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0211390 Janua y 30, 2019 8 / 18 (Fig 4). Wi h he exhaus openings in on o he displacemen di use , in he wall o he bed, a e age en ila ion e iciency is ε a = 0.56. Wi h he exhaus openings in he wall opposi e he bed he a e age e iciency inc eases signi ican ly, ε a = 0.69. The ai low pa e n in he oom is close o ha o PDV. This beha iou is in line wi h o he au ho s’ obse a ions. The ai change e iciency index is also co ela ed wi h he posi ion o he adian wall. In all cases, when he adian wall is behind he HCW (emp y symbols) ai change e iciency is g ea e han when i is opposi e he HCW ( ull symbols), p obably because he mal loads a e mo e balanced compa ed o he oom’s plane o symme y. No clea co ela ion can be es ab- lished be ween ai change e iciency and ACH. Dispe sion o he con aminan . Con aminan emo al e ec i eness DV p o ides e ical he mal s a i ica ion ha de e mines con aminan dispe sion in he oom. In o de o main ain he same mean empe a u e in he oom wi h less ACH, he di e - ence be ween he ai inle empe a u e and he a e age empe a u e o he oom is inc eased. Up o a heigh o 1.1 m, he e ical empe a u e g adien is g ea e o low ACH (Fig 5A), whe eas om 1.1 m o he ceiling he g adien s a e almos equal o all ACH. This beha iou can also be seen o he o he simula ed cases. These nume ical esul s a e consis en wi h expe imen al obse a ions [57,58]. As he con aminan sou ce is also a sou ce o hea , he con aminan will be anspo ed di ec ly o he op o he oom [59]. The pollu an is expec ed o be dis ibu ed acco ding o a wo-zone model [35], a clean zone in he lowe pa (y<y s ) o he oom and an unclean zone in he uppe pa (y>y s ). The mean concen a ion o N 2 O in ho izon al planes is shown in Fig Fig 4. Ai change e iciency index. Blue symbols: exhaus openings on he wall o he DD. Red symbols: exhaus openings on he wall o he bed. Full symbols: adian wall in on o he HCW. Emp y symbols: adian wall behind he HCW. h ps://doi.o g/10.1371/jou nal.pone.0211390.g004 Assessmen o displacemen en ila ion sys ems in ai bo ne in ec ion isk in hospi al ooms PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0211390 Janua y 30, 2019 9 / 18 15. VACIC. Guidelines o he classi ica ion and design o isola ion ooms in heal h ca e acili ies [In e ne ]. Vic o ian Ad iso y Commi ee on In ec ion Con ol. Melbou ne, Vic o ia, Aus alia.: Depa men o Human Se ices—Vic o ian Ad iso y Commi ee on In ec ion Con ol; 2007. A ailable: h p://www. heal h. ic.go .au/ideas/ egula ions 16. AusHFG. Aus alasian Heal h Facili y Guidelines [In e ne ]. 2009. A ailable: h p://www. heal h acili yguidelines.com.au/guidelines.h m 17. NORDEN. Bes p ac ice in design and es ing o isola ion ooms in No dic hospi als. Rydock JP, Eian PK, Lindq is C, Welling I, Welling I, edi o s. No dic Inno a ion Cen e. OSLO: No dic Inno a ion Cen- e; 2004. p. 57. 18. NTC. Tube culosis In ec ion Con ol: A P ac ical Manual o P e en ing TB [In e ne ]. F ancis J. Cu y Na ional Tube culosis Cen e . 2007. A ailable: h p://www.na ional bcen e .edu/TB_IC/ 19. Subhash SS, Ba acco G, Fennelly KP, Hodgson M, Radono ich LJ J . Isola ion an e ooms: Impo an componen s o ai bo ne in ec ion con ol. Am J In ec Con ol. Else ie ; 2013; 41: 452–455. h ps://doi. o g/10.1016/j.ajic.2012.06.004 PMID: 23036479 20. Saa inen PE, Kallioma ¨ki P, Tang JW, Koskela H. La ge Eddy Simula ion o Ai Escape h ough a Hospi- al Isola ion Room Single Hinged Doo way—Valida ion by Using T ace Gases and Simula ed Smoke Videos. Szni man J, edi o . PLoS One. 2015; 10: e0130667. h ps://doi.o g/10.1371/jou nal.pone. 0130667 PMID: 26151865 21. Subhash SS, Ba acco G, Fennelly KP, Hodgson M, Radono ich LJ. Isola ion an e ooms: Impo an componen s o ai bo ne in ec ion con ol. Am J In ec Con ol. Mosby; 2013; 41: 452–455. h ps://doi. o g/10.1016/j.ajic.2012.06.004 PMID: 23036479 22. Hy inen M, Rau io A, Pasanen P, Reponen T, Ea nes GS, S ei el A, e al. Ai bo ne In ec ion Isola ion Rooms–A Re iew o Expe imen al S udies. Indoo Buil En i on. 2011; 20: 584–594. h ps://doi.o g/10. 1177/1420326X11409452 23. Pan elic J, Tham KW. Adequacy o ai change a e as he sole indica o o an ai dis ibu ion sys em’s e ec i eness o mi iga e ai bo ne in ec ious disease ansmission caused by a cough elease in he oom wi h o e head mixing en ila ion: A case s udy. HVAC&R Res. Taylo & F ancis; 2013; 19: 947– 961. h ps://doi.o g/10.1080/10789669.2013.842447 24. Eames I, Tang JW, Li Y, Wilson P. Ai bo ne ansmission o disease in hospi als. J R Soc In e ace. 2009; 6: S697. h ps://doi.o g/10.1098/ si .2009.0407. ocus PMID: 19828499 25. Mema zadeh F, Xu W. Role o ai changes pe hou (ACH) in possible ansmission o ai bo ne in ec- ions. Build Simul. Rock ille Pike, Be hesda, Ma yland, USA: Tsinghua Uni e si y P ess; 2012; 5: 15– 28. h ps://doi.o g/10.1007/s12273-011-0053-4 26. Me ´ndez C, San Jose ´JF, Villa uela JM, Cas o F. Op imiza ion o a hospi al oom by means o CFD o mo e e icien en ila ion. Ene gy Build. 2008; 40: 849–854. h ps://doi.o g/10.1016/j.enbuild.2007.06. 003 27. Villa uela JM, Cas o F, San Jose ´JF, Sain -Ma in J. Compa ison o ai change e iciency, con aminan emo al e ec i eness and in ec ion isk as IAQ indices in isola ion ooms. Ene gy Build. Else ie B.V.; 2013; 57: 210–219. h ps://doi.o g/10.1016/j.enbuild.2012.10.053 28. Khan JA, Feigley CE, Lee E, Ahmed MR, Tamanna S. E ec s o inle and exhaus loca ions and emi ed gas densi y on indoo ai con aminan concen a ions. Build En i on. 2006; 41: 851–863. h ps://doi.o g/ 10.1016/j.builden .2005.04.002 29. Hayashi T, Ishizu Y, Ka o S, Mu akami S. CFD analysis on cha ac e is ics o con amina ed indoo ai en ila ion and i s applica ion in he e alua ion o he e ec s o con aminan inhala ion by a human occu- pan . Build En i on. 2002; 37: 219–230. h ps://doi.o g/10.1016/S0360-1323(01)00029-4 30. Khanka i K. Ai low dynamics o a pa ien oom. REHVA J. 2017; 5: 52–58. 31. Ju elionis A, Gagy ėL, P asauskas T, Čiuz ˇas D, K ugly E, S ˇeduiky ėL, e al. The impac o he ai dis i- bu ion me hod in en ila ed ooms on he ae osol pa icle dispe sion and emo al: The expe imen al app oach. Ene gy Build. 2015; 86: 305–313. h ps://doi.o g/10.1016/j.enbuild.2014.10.014 32. Meliko A, Mund E, Mus akallio P, Nielsen P V. Displacemen Ven ila ion, REHVA Guidebook 23. Kosonen R, edi o . REHVA Fede a ion o Eu opean Hea ing, Ven ila ion and Ai Condi ioning Associa- ions; 2004. 33. Ma eus NM, Ca ilho da G ac¸a G. A alida ed h ee-node model o displacemen en ila ion. Build En i- on. 2015; 84: 50–59. h ps://doi.o g/10.1016/j.builden .2014.10.029 34. ASHRAE. S anda d 170–2017—Ven ila ion o Heal h Ca e Facili ie. 2017. 35. Bjø n E, Nielsen P V. Dispe sal o exhaled ai and pe sonal exposu e in displacemen en ila ed ooms. Indoo Ai . 2002; 12: 147–164. h ps://doi.o g/10.1034/j.1600-0668.2002.08126.x PMID: 12244745 Assessmen o displacemen en ila ion sys ems in ai bo ne in ec ion isk in hospi al ooms PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0211390 Janua y 30, 2019 16 / 18 36. Li Y, Huang X, Yu ITS, Wong TW, Qian H. Role o ai dis ibu ion in SARS ansmission du ing he la g- es nosocomial ou b eak in Hong Kong. Indoo Ai . Munksgaa d In e na ional Publishe s; 2004; 15: 83– 95. h ps://doi.o g/10.1111/j.1600-0668.2004.00317.x PMID: 15737151 37. Qian H, Nielsen P V, Li Y, Hyldgaa d CE. Ai low and con aminan dis ibu ion in hospi al wa ds wi h a displacemen en ila ion sys em. P oceedings o BEPH 2004—The 2nd In e na ional Con e ence on Buil En i onmen and Public Heal h. Beijing: China En i onmen al Science P ess; 2004. p. 355. 38. Bi ola o a M, Kie a W, Za l E, Popiolek Z, Meliko A. E ec o ai low in e ac ion in he b ea hing zone on exposu e o bio-e luen s. Build En i on. Pe gamon; 2017; 125: 216–226. h ps://doi.o g/10.1016/j. builden .2017.08.043 39. Licina D, Meliko A, Sekha C, Tham KW. Human con ec i e bounda y laye and i s in e ac ion wi h oom en ila ion low. Indoo Ai . 2015; 25: 21–35. h ps://doi.o g/10.1111/ina.12120 PMID: 24750235 40. Villa uela JM, Olmedo I, Ruiz de Adana M, Me ´ndez C, Nielsen PV. CFD analysis o he human exhala- ion low using di e en bounda y condi ions and en ila ion s a egies. Build En i on. 2013; 62. h ps:// doi.o g/10.1016/j.builden .2013.01.022 41. Villa uela JM, Olmedo I, San-Jose ´JF. In luence o human b ea hing modes on ai bo ne c oss in ec ion isk. Build En i on. 2016; h ps://doi.o g/10.1016/j.builden .2016.07.005 42. Siegel JD, Rhineha E, Jackson M, Chia ello L, Commi ee he HICPA. 2007 Guideline o Isola ion P ecau ions: P e en ing T ansmission o In ec ious Agen s in Heal hca e Se ings [In e ne ]. 2007. A ailable: h p://www.cdc.go /ncidod/dhqp/pd /isola ion2007.pd 43. Sehuls e LM, Chinn RYW, A duino MJ, Ca pen e J, Donlan R, Ash o d D, e al. Guidelines o en i on- men al in ec ion con ol in heal h-ca e acili ies. A lan a, GA 30333: U.S. Depa men o Heal h and Human Se ices Cen e s o Disease Con ol and P e en ion (CDC); 2003. p. 249. 44. Be langa FA, de Adana MR, Olmedo I, Villa uela JM, San Jose ´JF, Cas o F. Expe imen al e alua ion o he mal com o , en ila ion pe o mance indices and exposu e o ai bo ne con aminan in an ai bo ne in ec ion isola ion oom equipped wi h a displacemen ai dis ibu ion sys em. Ene gy Build. 2018; 158. h ps://doi.o g/10.1016/j.enbuild.2017.09.100 45. Roos A. On he e ec i eness o en ila ion. Eindho en Uni e si y o Technology. 1999. 46. FLUENT. Fluen 12.0 use ’s guide. Lebanon NFI, edi o . Lebanon, NH: Fluen Inc.; 2009. 47. Li Y, Sandbe g M, Fuchs L. E ec s o he mal adia ion on ai low wi h displacemen en ila ion: an expe imen al in es iga ion. Ene gy Build. Else ie ; 1993; 19: 263–274. h ps://doi.o g/10.1016/0378- 7788(93)90011-I 48. Knigh KJ, Debus KK. P ac ical applica ion o he LES me hod o mixing in la ge indoo spaces. ASME 2005 In e na ional Mechanical Enginee ing Cong ess and Exposi ion. O lando, Flo ida USA; 2005. pp. 587–594. 49. Liu L, Li Y. Simula ion o In e pe sonal T anspo o Expi a o y D ople s and D ople Nuclei be ween Two S anding manikins. BuildSim-No dic. Lilles øm, No way; 2012. 50. Xu C, Nielsen P V, Gong G, Jensen RL, Liu L. In luence o ai s abili y and me abolic a e on exhaled low. Indoo Ai . 2014; 25: 198–209. h ps://doi.o g/10.1111/ina.12135 PMID: 24920328 51. Nielsen P V. Con ol o ai bo ne in ec ious diseases in en ila ed spaces. J R Soc In e ace. 2009; 6: S747–S755. h ps://doi.o g/10.1098/ si .2009.0228. ocus PMID: 19740921 52. Bulinska A, Bulinski Z. A CFD analysis o di e en human b ea hing models and i s in luence on spa ial dis ibu ion o indoo ai pa ame e s. Compu Assis Me hods Eng Sci. 2015; 22: 213–227. A ailable: h p://h p//cames.ipp .go .pl/index.php/cames/a icle/ iew/23 53. Be langa FA, Olmedo I, de Adana MR, Villa uela JM, Jose ´JFS, Cas o F. Expe imen al assessmen o di e en mixing ai en ila ion sys ems on en ila ion pe o mance and exposu e o exhaled con ami- nan s in hospi al ooms. Ene gy Build. Else ie ; 2018; 177: 207–219. h ps://doi.o g/10.1016/j.enbuild. 2018.07.053 54. Einbe g G, Hags o ¨m K, Mus akallio P, Koskela H, Holmbe g S. CFD modelling o an indus ial ai di - use —p edic ing eloci y and empe a u e in he nea zone. Build En i on. 2005; 40: 601–615. h ps:// doi.o g/10.1016/j.builden .2004.08.020 55. Olmedo I, Nielsen P V, Ruiz de Adana M, Jensen RL, G zelecki P. Dis ibu ion o exhaled con aminan s and pe sonal exposu e in a oom using h ee di e en ai dis ibu ion s a egies. Indoo Ai . 2012; 22: 64–76. h ps://doi.o g/10.1111/j.1600-0668.2011.00736.x PMID: 21815935 56. Liu L, Li Y, Nielsen P V, Wei J, Jensen RL. Sho - ange ai bo ne ansmission o expi a o y d ople s be ween wo people. Indoo Ai . 2016; h ps://doi.o g/10.1111/ina.12314 PMID: 27287598 57. Qian H, Li Y, Nielsen P V, Hyldgaa d CE. Dispe sion o exhala ion pollu an s in a wo-bed hospi al wa d wi h a downwa d en ila ion sys em. Indoo Ai 2005 Model Assessmen , Con ol Indoo Ai Qual. 2008; 43: 344–354. h ps://doi.o g/10.1016/j.builden .2006.03.025 Assessmen o displacemen en ila ion sys ems in ai bo ne in ec ion isk in hospi al ooms PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0211390 Janua y 30, 2019 17 / 18 58. Li Y, Sandbe g M, Fuchs L. Ve ical Tempe a u e P o iles in Rooms Ven ila ed by Displacemen : Full- Scale Measu emen and Nodal Modelling. Indoo Ai . 1992; 2: 225–243. h ps://doi.o g/10.1111/j.1600- 0668.1992.00005.x 59. Gao N, He Q, Niu J. Nume ical s udy o he lock-up phenomenon o human exhaled d ople s unde a displacemen en ila ed oom. Build Simul. 2012; 5: 51–60. h ps://doi.o g/10.1007/s12273-012-0068-5 60. Zhou Q, Qian H, Ren H, Li Y, Nielsen P V. The lock-up phenomenon o exhaled low in a s able he - mally-s a i ied indoo en i onmen . Build En i on. 2017; h ps://doi.o g/10.1016/j.builden .2017.02.010 61. Meliko AK, Bolashiko ZD, Kie a W, Popiolek Z, B and M. Does inc eased en ila ion help educe c oss-in ec ion in isola ion hospi al wa ds? P oceedings o CLIMA 2010: Pape R7-TS39-OP02. An alya, Tu key; 2010. 62. Bolashiko ZD, Meliko AK, Kie a W, Popiołek Z, B and M. Exposu e o heal h ca e wo ke s and occu- pan s o coughed ai bo ne pa hogens in a double-bed hospi al pa ien oom wi h o e head mixing en i- la ion. HVAC&R Res. Taylo & F ancis; 2012; 18: 602–615. h ps://doi.o g/10.1080/10789669.2012. 682692 Assessmen o displacemen en ila ion sys ems in ai bo ne in ec ion isk in hospi al ooms PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0211390 Janua y 30, 2019 18 / 18