International Journal of Occupational Safety and

Ergonomics

ISSN: 1080-3548 (Print) 2376-9130 (Online) Journal homepage: http://www.tandfonline.com/loi/tose20

Numerical Simulations of Heat and Moisture

Transport in Thermal Protective Clothing Under
Flash Fire Conditions

Guowen Song, Patirop Chitrphiromsri & Dan Ding

To cite this article: Guowen Song, Patirop Chitrphiromsri & Dan Ding (2008) Numerical
Simulations of Heat and Moisture Transport in Thermal Protective Clothing Under Flash Fire
Conditions, International Journal of Occupational Safety and Ergonomics, 14:1, 89-106, DOI:
10.1080/10803548.2008.11076752

To link to this article: http://dx.doi.org/10.1080/10803548.2008.11076752

Published online: 08 Jan 2015.

Submit your article to this journal

Article views: 36

View related articles

Citing articles: 9 View citing articles

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=tose20

Download by: [71.220.158.2] Date: 23 October 2015, At: 01:22

 International Journal of Occupational Safety and Ergonomics (JOSE) 2008, Vol. 14, No. 1, 89–106

Numerical Simulations of Heat and Moisture

Transport in Thermal Protective Clothing
Under Flash Fire Conditions
Guowen Song

Department of Human Ecology, University of Alberta, Edmonton, Canada

Patirop Chitrphiromsri
International Journal of Occupational Safety and Ergonomics 2008.14:89-106.

College of Textiles, North Carolina State University, Raleigh, USA

Dan Ding

Department of Human Ecology, University of Alberta, Edmonton, Canada

A numerical model of heat and moisture transport in thermal protective clothing during exposure to a flash
fire was introduced. The model was developed with the assumption that textiles are treated as porous media.
The numerical model predictions were compared with experimental data from different fabric systems and
configurations. Additionally, with the introduction of a skin model, the parameters that affect the performance
of thermal protective clothing were investigated.

numerical model protective clothing moisture transport flash fire condition skin model

1. INTRODUCTION conditions. However, moisture transport in fabrics

and its effect on the protective performance of the
Heat and moisture transfer in porous media are garment have not been studied in sufficient detail.
common phenomena with wide applications in Textile fabric can be treated as a porous medium.
engineering fields, such as civil engineering, energy Heat and mass transport in wet porous media are
conservation, textiles, and functional garment coupled in a complicated way. Energy transport in
design [1]. Field study and research performed such a medium occurs by radiation and conduction
during the past 10 years have demonstrated that the in all phases as well as by convection within the
presence of moisture in textiles significantly affects liquid and gas phases. There are many models for
the performance of textiles and garments [2, 3, 4, analyzing multiphase transport in porous media.
5, 6]. In developing thermal protective clothing Vafai and Sözen summarized and compared
materials, such as firefighter garments, researchers these models [7]. Gibson’s model described
have traditionally focused on heat transfer in dry fabrics subjected to intensive heat [8]. Gibson

Data in this article were generated to characterize the properties of materials or assemblies in response to thermal exposure under
controlled laboratory conditions or in model predictions. They should not be used to appraise materials, products or assemblies under
actual fire conditions. They are not presented to predict all types of field conditions where the nature of thermal exposure can be
physically complicated and unqualified. We wish to emphasize that it is not our intention to recommend, exclude or predict the suitability
of any commercial product for a particular end-user.
Correspondence and requests for offprints should be sent to Guowen Song, Protective Clothing & Equipment Research Facility,
Department of Human Ecology, 331 Human Ecology Building, University of Alberta, Edmonton, AB T6G 2N1, Canada. E-mail:
<[email protected]>.
 90 G. SONG ET AL.

analyzed multiphase transport in hygroscopic garment system consists of three different fabric
porous textiles. However, Gibson’s model did layers, i.e., an outer shell, a moisture barrier, and
not consider radiation heat transfer within the a thermal liner, respectively, from the exterior to
fabric layer. Torvi developed a one-dimensional the interior of the clothing ensemble. An air gap
transient heat transfer model, which accounts for between the inner layer of the fabric and the heat
the penetrating radiative heat transfer through a flux sensor is introduced.
fabric [9]. In this research, a model that couples
heat and moisture transport for multilayer 2.1. Heat and Moisture Transport in
protective fabrics under flash fire conditions is Textile Material
introduced and the effect of heat and moisture
Textile fabrics can be modeled as hygroscopic
transport on the performance of thermal
porous media. The porous textile material is
protective clothing is analyzed.
a mixture of a solid phase consisting of solid
International Journal of Occupational Safety and Ergonomics 2008.14:89-106.

fibers and bound water absorbed by a solid

2. MATHEMATICAL MODEL polymer matrix, and a gaseous phase consisting
FORMULATION of water vapor and dry air. A schematic
diagram of the structure of a porous textile is
Figure 1 illustrates a multilayer protective fabric demonstrated in Figure 2. Gibson [8] developed
system exposed to a high-intensity flash fire. This a set of equations for modeling heat and mass

Figure 1. Schematic diagram of protective clothing fabric and the thermal protective performance
(TPP) sensor. Notes. conv.—convection, rad.—radiation.

JOSE 2008, Vol. 14, No. 1

 PROTECTIVE CLOTHING MODELING 91
International Journal of Occupational Safety and Ergonomics 2008.14:89-106.

Figure 2. Schematic diagram of a two-phase structure of porous textile media in the averaging
control volume.

transfer through textile materials by applying convection contributions caused by pressure

Whitaker’s [10] theory of coupled heat and differences, which can arise either due to body
mass transfer through porous media. Torvi’s movement or due to external air movement, are
model assumed that convective heat flux only neglected. Moreover, it is assumed that if there
applied to the surface of the fabric but radiative is any extra liquid sweat which builds up on the
heat flux could penetrate through the fabric up surface of the sensor (skin), it will either drip off
to a certain depth [9]. Morton and Hearle gave or wick into the fabric and then it will be absorbed
definitions of the physical properties of the by the fabric fibers and become bound water. In
fabric [11]. The convective heat transfer in the other words, free liquid water exists neither on the
air gap between the fabric and the sensor was surface of the sensor nor in the fabric layer.
simulated as natural convection in a horizontal The energy equation can be formulated as
enclosure, which was heated from below [12]. wT w § wT ·
Based on those assumptions, the energy balance Uc p
wt

 'hl  'hvap m sv  wx ©
¨ keff cc e Jx ,
¸  J ˜ qrad
wx ¹
in the infinitesimal element of the fabric can be (1)
wT w w§ wT ·
Uc pequation.
formulated in the form of a differential
wt
 
 'hl  'hvap m sv w ¨ keffbw  ¸  svJ ˜ qrad e ,
U
wxw©t
H 
wx ¹
m  cc
0. Jx

A partial differential equation is developed for

w w § wU ·
temperature distribution in a composite w
Uw fabricH bw   m sv where0.
wt
 
H J Uv  mdensity
ρ—effective sv
wx ¨ Deff thev ¸fabric,
of
wx
,
layer by combining Gibson’s and Torvi’swtmodels. cp—effective specific heat ©of the ¹fabric,
w w § wU · Dt—time, a HJ
Therefore, the energy equation is modeled
wt
H J Uv  
onm
sv 
T—temperature,
wx © ¨ Deff Dv eff¸ ,
wx
, ∆hl—enthalpy of
W from bound water to free
Gibson’s model and penetrating radiation, a transition per ¹unit mass
term described by Torvi’s model. Thermal Da HJ wTvap—enthalpy of evaporation
Deff ,
liquid water, keff ∆h  qconv
cc  qrad cc  x 0;
properties of all phases are accounted for in Wthe per unit mass, m w.xsv —mass
x 0 flux of vapor out of
model based on the relations given by Gibson. .
wT the fiber (or into the fiber if msv is negative),
kaccounted  qconv
cc  qrad cc  x 0;keff wT  qair,rad
cc cc
k qair,cond/conv  , with-air-gap
Radiative heat transfer in the fabric is eff
wx x 0 x—linear vertical wx x co-ordinate, eff—effectivex Lfab
Lfab
for by introducing in the energy equation a source thermal conductivity of the fabric, γ—extinction
wT
term similar to that in Torvi’s model. kGas
eff phase
wx x Lfab
 q cc
coefficient
air,rad cc  kof wthe
 qair,cond/conv
eff
T , with-air-gap
x Lfabric,
configuration;
fab x Lrad  Tsens ), without-air-gap co
hi (Tand qʺ —incident
wx x Lfab fab
fab

wT wUv
 keff hi (Tfab  Tsens (),Uwithout-air-gap
x Lfab hm,amb  Uv ) x2008, configuration;
 Deff14, No. 1;
wx v,amb JOSE 0 Vol.
x Lfab wx x 0
wUv wUv
 x 0wT
wx x 0 keff hkweffTwt (U  qcc U qair,rad cc)qcc wx ©qD cc; wU
air,cond/conv
wxv ¹
x ;L , with-air-gap c
wx wxx 0 x LDfaba HJ
m,amb v,amb conv v x rad 0 x 0 eff
wT wx x 0 fab
 keff
wx x Lfab
 qair,rad
cc cc Deff
 qair,cond/conv
wT wT
Wx Lfab, , with-air-gapwUconfiguration;
92 G. SONG ET AL.  keff hkm,gap eff (Uv  Uv,air qcch)ix(TLfab  qccL Deff  Tsensv ), without-air-gap , with-air-gapconf
, with-air-gap con
co
w x wxx LxwTLfab air,rad fab xair,cond/conv
fab
w x x Lfab

 keff
wT
hi (Tfab x L  Teff
k fab  qconv cc  qrad cc  x 0; x L fab

wT wU sensw), x without-air-gap configuration;

wU
wx x Lfab fab x 0
 keff hm,amb Deff (Uv,amb v
wTofwTwater

hi U(Tvfab )0, DTeffsens ),vwithout-air-gap
without-air-gap ;w §
air, wT ·configu
configuration;
radiation heat flux from the flame onto the fabric. ρv,amb —density
wx kxwU wU Lx c
eff vfab x w
p 
x vapor
'

h q
l 
x0 Lfab'h
cc
in thewmxambient

vap q cc sv x 0 
¨ keff , with-air-g ¸  J ˜ qrc
the airwgap. x ©x Lfabwx ¹
Lfab
ρv,air wx x;tofLfabwater vapor inwU
air,rad air,cond/conv
Chitrphiromsri and Kuznetsov provide hm,amb (Uv,ambon Uv ) xand
details 0 Deff —density
hqm,gap cc (Uwx xUw0 ) wUv v
the development of the model and its relevant hAccording (Uv,amb
air,rad
UwwT Uv,air
v to Torvi
v )Hx x0 Lfab Deff
[9] Deff
the radiation ; , and
with-air-gap con
bw  m 0.
wx wflame x x Lfab
m,amb
parameters [13]. convection cc  kwU
qair,cond/conv heat w tfluxes h
from ( Tsv
the x  0T sens without-air-gap
to ), the
hm,gap (Uv  Uv,air ) x Lfab  Deff eff v wx , with-air-gap i fab x Lconfiguration;
fab
wT continuity equation wcan § bewT · fabric wU wxvw x Lfab fab0, without-air-gap
x L wwU§ wUvwT · ·
The solid phase
described as p wt
Uc 
 'hl  'hvap m sv  wwU x©
¨ keff
wx ¹
¸  J h˜ m,gap ccTD
qrad U c
are
e(J
fab eff

UwxxvT,LfabU
w 
xw t
 '
v,air HhJ)Uxv L'fabh m 
 sv m
Deff ¨ vwDeff
 w x w x
§ configuration;
¨ k wUwx
, with-air-gap
¸ , ¸  J ˜ qrad cc econfig
Jx
,
( q
p
cc h w t  q cc x
( U) L l
fab  h
vap
U ()T
s

v
T © wD x x
© L
),
eff
fab wx ¹(8)
v ¹ ;
wT w § w · w T c,fl g w §fab x 0wT ·
 T
 
v
  0, without-air-gap configuration;
D m,amb v,amb v x 0 eff
wwTUc p
conv Jrad
qcc 'eh , 'hvap
x 0
wx xJ 0˜ qrad cc e , J x
¨fabk·eff Uc¸pccqwJair,rad  Da Hm
x
 'hl  'hvap eff m w sv§wx x(2) wLT ccJ˜ wU J sv ¨ keff ¸
UcUpw wt H bw'wht lm'svhvap0.m sv   ¨ k w x © ¸  w
J x˜ q¹ D U e t xwradv l
,  H D   m 0, § ,
without-air-gap
0.
Vsv¨ T heat transfer 4 w x © w x ·
¹
configuration;
4
 Tsenscoefficient
wU¸v
wx ¹ where ccwhwc,flwhtx—convective
eff rad eff eff
wt cc wx © Uv  U©Wv,airfab) x xLfabLfab  Deff
bw
w qair,rad w qair,cond/conv x L(fab ¹ , with-air-gap
The gasw phase  Hbw   wm sv §equation 0.wUv ·can Uw wair,rad
between q cc H bw x Lm
m,gap
the flame and the outer surface wx ,ofx Lthe
written asUwwtwt U Hcbw
w H UUw wdiffusivity
J v wtT m    m
svsv'hlw0. 
D
¨ h eff qcc wx m sv ¸¹
, w §
be
wT ·qair,rad

cc T
wt fab Hccx J U  Jxk
 sv wT0. w §
fab
effm  by radiation § 1 Dqconv wU
cc 1 Hvsens  ·
¸ , ¸ the
·
cc x 0; fabric
qrad fab
x ' ¨ keff fabric. ¸  Jwt˜ qThe rad e v ,wU sv
heat fluxwx x¨w0Hx ¨© eff Hwxfrom
Lfab
p © vapair,cond/conv ¹
w w t wT§ wUv ·© w x w x w¹qthe cc (qsensor cc ) x w § © airfab wUwithout-air-gap
· fabsens ),¹
    cc D v
DHa HJ U to H J Uconv m qacross 0 thehc,fl (T0,
gapg  v isT configuration;
Jv m svH wx wT ¨ Deff ¸, x 0
air,cond/conv
w w sv§ fabDxeff
wU
¨ v · Lfabwx (3) ¸, eff rad
Deff w
H J Uv  msv
wt Uw wWt  H bw  wxm¨©sv 0.
t w  , w
Deff © ¸ , x ¹
cc ) x 0 hc,fl T
w t D 
v
Hsens  keff, © fab D a J w xx L
wVx§¨ xTfab
wx  q¹air,rad
cc cc
 qair,cond/conv x Lfab , with-
D H
(qwx cc ¹ qrad
conv (Tfab g eff
x TLfab
D fab x 0W),
q Hcc
4
 Tsens 4 ·¸
wT w § wT · cc a air,rad L
¹ Tsens ),
  x ( qcc qair,cond/conv h © (Thc,gap x T(Lfab Tfab
fab
ccJ , ) x 0x· L fab
International Journal of Occupational Safety and Ergonomics 2008.14:89-106.

a J
Uc p D   ' hl wD wTeff

D a'Hh mqsccv,  qwof k§the  ·J ˜ qrad cc §e JD ,4effconv cc  qwrad
qair,rad fab x ),
, (9)
 
L
¨
cc wU¸ T x 0 fab
 
where —effective
k J vapdiffusivity gas
; phase 4 wTc,fl g
V  cc  
eff
wwTt eff Deff wxH J U,v W m sv wx ©§ ¨ D w w T · ¨ T W
q T ¸ cc § cc  H ·

x L
eff conv rad x 0x ¹ v
, J
 ¸J ˜ qcc e , x Lfabeff wx eff¹kair
fab x  k 
sens fab
k q 1 q h 1 ( T  ;  T ), without-air-
wT · inUcthe  'hlwt W'hxisvap air,cond/conv
cc wwTx ¹¸· ¹ rad© Jx ¨)4, rad 
w § msv [8]wfrom k eff wx § x(T conv sens
0 x4 ·
p wfabric,
Tte Jx , which 0defined
wx ©¨§x q©eff x 0V ,¨wT
i x fab ¸Lfab sens
v k
¨ eff
wx © ww§xk¹ wU
 J ˜
¸ UTc p·rad H 'hlJq cc
w

w
 ˜ qkrad cck ewm T
 wxsvkT' ,eff vap
w T
 w
h 0. m sv  qcc ¨ keffqcc x¸ Lfab;J ˜ qrad
DxaqHccJ  qccwwqconv
air,rad
Twx· x¹ 0  J§¨,xwith-air-gap
cc e 1,k 1c,gap w
h
cc wx Hsens
T Nu
Tfab x ·Lqcc © fab©qccfab H
Lgap radx Lxfab0; sens ¹¹ ,
x L fab H T sens ¸
vap m ¨ effUc w w¸wwTwttJbw   w;rad
xcc§ © qair,cond/conv qeff w conv
fab
w cc T
cc §(U)x1q0air,rad 0x),wU
cc configuration;
¸
w § wx ©wT ·Uwpxw ¹  H bw'Jeff hxl, wDmxw'effsvxxhvap ¨ keffx 0 ¸  J ˜ qxradLefab© H,where (4) Hsens cc ofh1the cc  ·T  D
sv
0.mxsconv
air,rad
eff  v0, air,rad rad kHeff x x L
(qconv
 sens ccfab
0
—emissivity
hmc  qrad
¹m,amb  (Hq)Tsensor.
U Torvi v , with-air-gap c
also ;
m sv k
¨ eff 
¸Uw wt rad J w˜
w tq cc e x0 Lfab
w § W wU w x ·© w x ¹ fab w x§ · ¨ v,ambdT  c,fl air,cond/conv
v sensxg 0 fab x eff
¸
Lfab
wx ¹ w H J UHv bw m   w T w T w x
p
 m cc
HxALqconfiguration; x TLfabsens
cc
¸qair,rad x Lfabkeff, with-air-gap
0. x L
wx © sv D v
,cc H sens qair,cond/convcc gives q q the ¨
heat ¸
flux
fab
H
 by
cc© fabc,gap h , H
(T
conduction/convection

cc sens  T ),
4 x L¹ fab
x
, 4with-air-gap 0 confi
wwt wwt D wsvTwkTeff wxT¨§ ofeffwater dt§¨qTair,cond/conv ·
sens
air,cond/conv fab
0. w©xDx qLhair,rad
wUv · Uwhere w HJH U keff
a—diffusivity
km msvkeffww0.
 wU
ccfab wx
 ccfab·¹vapor
(qTvconv , ccqrad cc inTxsens
qair,cond/conv the ; 
), air , with-air-gap
without-air-gap from H (Tfab
wx configuration;
the fabric LT©
x wconfiguration; to the
air,rad
¹fab senssensor V
© across
fab the
 T  D),eff air sensgap:wU ¸v
¹
xL§¨©fabx eff wU ¸¹ qair,cond/conv  keffqcc hm,gap ((UTvh)),i(TUfab without-air-gap con
bw
wwtwτ—fabric
eff
wxwxx wxwLwxfab i
ccfab fab
kair v,airx ) xLfab
x L
, with-air
0 wxv ·
t D x L 0 x L hc,gapsens Lxfab
Tsens

v sv
 
fab
, and H J Uv a HJ mtortuosity.
fab
cc w cc  cc Lfab sens ,
with-air-gap conf
§wx ¸¹ wUv · , sv k wwwx§wT¨©TforeffwU
D , x L fab h
cc wc,gap
q x air,rad
Nuq q , , w x
·h¸¹i (Tfab wU  Tsens ), without-air-gap T hc,gap (§Tfab 1 T x1),Lfab H sens ·), (10)
sens x x
air,radL L air,cond/conv
wD L fab  Tsens
fab Lgap
 kqeff
x Lfab
wBoundary wx fab
Deff
x ¨© wUv ·wx ¸¹
,
w D
t
teff D H conditions
H J Uv aWkm
h J svw
,   T eff ¨ Deff the
k ( Uw x w x  h U (
wx T )
v fabric:

¸ q, cc 
h D  x Lccfab v kair (T )
T q Nu), 
without-air-gap
; , ,
air,cond/conv
qw
with-air-gap
cc xconfiguration; h (
L fab
T
hLi (gap
x configuration; Tfab
configuration;
wU
 T
x L¨fab x sens
),© H

with-air-gap wU
H

without-air-gap ¸
¹
configuration;
configur
, Da HJ wx w©x x L fab ¹ x Lfabc,gap wx
eff
m,amb eff v,amb x iL v fab 0 air,rad eff air,cond/conv
sens x Lfab h sens (U x L i Dfab U vsens ) x 0 0, fab Deff v
sens ;
wx ¸¹
eff x fab
hc,gap§—convective (T· ) v dT without-air-gap configuratio
Dkeff wT W ,  qccx Lfab mc
kv,amb
eff
L wherem,amb heat transfer wx coefficient
 eff Da HWJ
w conv  qrad cc fab  ; x(5)
wU
0 gap
h q
c,gap cc
sens sens H Nu ¨
airpeff
¸ w ,x x L, wU vª x 0 1 º
$
T
x h w ( U x
 0
U ) 
wU wU
D v
; h
of the ( U air Hdue
 L AU ) to dt  Dfab
conduction and ; natural
h (U cc ) x 0;hv(TxD0D §v 
Deffk x ,0 cc Uqv,amb T © gap¹sens $
mc ; x· ),,xwithout-air-gap ª 1 708 º (xT«wU § Ra ·3 T »),
qk(m,amb eff m,amb v,amb v x 0 eff
Uconv v dT
»Dwis
Urad pw with-air-gap ccconfiguration;cc1.44
 x 0; keff wwT
eff wTWhm,amb
x x m,gap 0  qcc w
v v,air v x L cceffxeffLwfab
qsens ¨
Tsens
¸ , hNu (1Uinvqconfiguration;
qair,cond/conv  Uv,air )LxfabLfabwhich
x1gap,
hc,gap  x 0xv L
¨ sens
1» .
fab ¸ , with-air-gap con
 ccx 0 0 convection
 the air
Ra ¼wU x L©w5x830
eff
v,amb 0 fabi fab
x q cc ; w
x x cc
qsens§ mc
air,rad
cc «
· ¬dTqair,cond/conv cc
fab
«
 
m,gap eff
xH (6a)
A
with-air-gap dt configuration; qpair,rad ,¹with-air-gap confi
cc  x 0; cc xrad x0 Lfab
 qrad  keffwTwx conv qair,rad Lqfab © , wU ¹ cc cc¸ ) k ,(T
sens(U ¨  Uqair,cond/conv
qwith-air-gap ) eff v «¬ »
Lx L
keff wTx xx 0Lfab  qhccwU   qccv  xUair,cond/conv
ccU((UUrad
; ) x Lfab xwU
 LD wU
v sens
h,m,gap vh H  Av,air¹configuration; D
fab gap fab
, with-air-gap(11)¼ configu
 
fab
 keffwx xh 0 D (hUconv cc
) qair,cond/conv x Ldt fabair
cadcc  x 0; qvair,rad , with-air-gap configuration; © Nu ,
m,gap v,air eff
0
 q cc vq cc w x q cc
v ,with-air-gap cc wU
configuration; (  , with-air-gap ), w xconfiguration;
qair,cond/conv  x L , with-air-gap wwTx m,gapconfiguration; 0, Uwithout-air-gap
) D  D configuration;; q h T T with-air-gap configuration;
sens
: c,gap T v Lfab 'E
 k wT x Lfab w qxair,rad x wLxfab ,xwith-air-gap
eff m,amb v v,air
cc x L  qair,cond/conv
v,amb x Lfab v x 0 eff x Lfabeff
cc
sens air,rad x Lfab air,cond/conv d
 Deff Pfabexp(
sens x L i L 0,),without-air-gap configuration;
sens x Lfab
wfabx x 0 qccconfiguration; vqair,rad x Lfabª  Lgap , with-air-gap
fab gap
ccwx xxLqLfab cc gap
cc
r,rad  qair,cond/conv
fab
xkeffkLfabeffeffw,Twwith-air-gap
wx x Lfab hcci (configuration;
Tx xqccL wU   Dqair,rad
TwU fab 
fab
vx q L
 Tsensq),cc without-air-gap
ccair,cond/conv 0, 
without-air-gap h, ( T
L
iwith-air-gap  T configuration;
 ),
whereD
configuration;
sens dt
with-air-gap
configuration; wU Nu—Nusselt ( q ccconfiguration;
§ mc
0, 
fab RT
air,cond/conv
q cc
number;
· )
without-air-gap k
$hc,fl ( T 
(T)—thermal
T
configuration;
1
), º
$ configu
sens wUv
fab sens dT
x Lfab, with-air-gap
fab
, is a«§function Ra ·x3 0 »
(6b) air
THª«sens
conv p 1 708 º
rad x 0 g fab
cc
 qair,cond/conv keffwx x D fab configuration; hv ieff(T (Ufab wvxx xUL0,fab )Tsens ),xwithout-air-gap Lfab eff
configuration; ccwxhqsens cc Lof
 Tsens ), without-air-gap wwTx configuration; eff hm,gap
Lair,rad v,air x Lfab  Deff configuration;
without-air-gap cc Nu
, with-air-gap
qsens
conductivity qair,rad i1(xT 1.44 ¨the
configuration;
fab  ),
11air, ¸ with-air-gap
which$
» ¨ configuration; ¸ of  1T .
RaV ¼¨ Tfab © 5 830¹Tsens$4 »·¸
«
fab Lfab
 keff xqccLfab wxh (xTL  TwU ), without-air-gap wx x Lfabconfiguration; $ ª fab
© ¬ 3 ¹sensA º dt § 4
fab
Tfab x L  Tsens ), without-air-gap h wTw(xUv,amb x L qconfiguration;
air,cond/conv i fab
Ucc v ) x 0  Deff fab x Lfab sens v
; ª 1only; q708cc qºand cc «LgapRa § ·
—thickness» of the air« gap x L [9]. »¼¹
1  1.44 «1 configuration; » q«sens cc  qair,radcc¸  1» q.air,cond/conv cc$ © ª ¬  fab 3 º , with-air-gap
1
fab  km,amb q ccTfabfab hUair,rad i (Tfab wUxv LfabD Tsens wU wxv),Nu without-air-gap air,rad air,cond/conv
cc¨ qair,rad , c
wU
 Tvsens ), without-air-gap h eff
w x(
t U air,rad configuration;
xqD Lccfab ) x 0;
0, without-air-gap ¬ configuration;RaccNuT d¼fab :1 x«L©1.44 1 ¹'Efab » §«¨1
5 830 ª x 1 L 708 º § Ra ·
1 ¸fabH sens
x L  L »· .
m,amb xv,amb Lfab 'veff Ex 0 eff wU
w qair,cond/conv  gap
1
;
³
xDLfab
wU h: UPqv,amb
cc(exp( Uvq)ccx)dw)0tx. x D
air,cond/conv
wU x v (7a) ¬ P exp( «  Ra ), ¼ » ¨« 5 830  »¸
m,amb((U eff (T v xT0; , with-air-gap
fab
eff
wxDeffx 0 v h; m,gap v  conv
qccUv,air
air,cond/conv RT ) x radLfabx 0  D Lh fab
wUwvgx dx:fab ), 2.2. Heat
configuration; dt cc
ccTransfer
qsens cc h¬)i (TRT fabinto  Tsens ¼the
« H©fab ),H¹sens configuration;
Sensor
),©with-air-gap »¹
0 T
(U0 T tqUccv ) xx 0Lfab t Deff wU
c,fl
eff
wxv x ;Lfab P exp(
x 0 ' E T ( q  q h ( T ¬T ¼
wU wUv wx x h0m,amb hm,gap (Uv,amb  Ux v,air
fab
)  D
0
, with-air-gap
fab
), configuration;x Lfab conv rad x 0 c,fl g fab x 0
 D ; fab Lair,rad fab'EV T § wwU x ' E · The heat flux, H
 which applies to a ª4 ·
sensor, is1 º
$
v 2 x L eff
cc :  qPrad sens'h
³ RT (qdconv Ec,fl§(Tg 4Tfab
4 4
w vxdt 0T TT ccexp(
v
Dsolid / d cc cc
Deff eff wx wU , with-air-gap configuration;
³  ¨ x ¸
fab
U :  U ( q :  q  P) exp( h ( ) d t . (7b) ), ) ),$
h t
(wU(vqv cc v,air P exp( )
q'ccEqconfiguration; ) d
D t . , with-air-gap configuration;  V),¨ Tªfab 1 708 x 0 º Tsens«§ ¸ Ra ·
1»»),.
f conv rad fab
0 c,fl x L sens
fab x 0 3
wx x Lxfab0v :m,gap cc0 air,cond/conv cc)x0, © eff TgwxxRT
x LTfab fab ), x¹ 0
g

³ d)t .xwithout-air-gap
0 0 hc,fl (wU determined
dt cc Nufrom a1.44 test © «temperature
1  x Lfabh » sensor (T«fab¨ ¹x on the Tsens
Lfab
Lfab  Deff İD, with-air-gap
eff(P exp( qconv rad RT v xfab configuration; , cc1 RT
qair,cond/conv L fab ¸
wU w x Lm,gap x h U wUw x Uair,rad
RT ) x L  D §
§2 V1w¨xT 1 configuration;
Lfab x 0
, with-air-gap
H sens ·t Tsens ¸ basis air,rad · configuration; q § 4 ¬ x L fab Ra c,gap ·
¼ ¸ © of the
4 5 ,830 ¹
 Deff v 
fabD 0
v v x v,air LT x Lfab fab
2 x0,Lwithout-air-gap
eff 4 4
of a xlumped Lfab V ¨ Tfab heat §x 1Lcapacity 1Tsens Hanalysis ·«¬ »¼
, with-air-gap
IJ eff wU configuration; fab V §/T ¨f 4 © fab x TLfab 4¸ ·
fabfab
v x /L dqfcc Dsolid
Dxsolid dfab  ¹ ' E © ¹
x Lfab fab( sens
ut-air-gap wx configuration; w ¨ ¸ k T )
x Lfab q
without-air-gap configuration;

Dsolid
q
cc D
air,rad eff/ d 2
cc x L(qconv
wUwİfvqxair,rad
fab
cc İxqrad
air,rad 0, without-air-gap
ccLfab © H
©) x 0 h§c,flfab(Tg 1TfabH ¹ , ),·
fab x L H

³
sens
1configuration;
:
configuration;
sens
¹ P
x 0
exp( , q cc
insulated ) d
RT x Ldfab: c,gap
air,rad t . copper h § Nu sensor.
P exp(
¨
1 © H'
The air heat flux, is modeled
, · ¹
1EL), H sensH sens
 fab ¸
cceff
Dqair,rad
q cc cc H  x L0,
fab fabwithout-air-gap § 1 ¨ 1  H  · 0sens ¸ as ¨  gap ¸
İ air,cond/conv H sens dt
cccc wxIJ x Lfab IJ © H§fabmcRT·H sens ¹
sens
¨  V §¨ T© H fab
hout-air-gap configuration; cond sens H  sens(7c)2¹4 ·
q 4
D ¸  T/ d ¸
IJTqfab air,radcc
qair,rad cc
qair,cond/conv © H fab © (THfabsens
hc,gap fab 
x xLfab L¹
solid sens Tfsens¹), cc
H sens qair,cond/conv qsenscc p
h ¸ (Tfab x , L
dT
 Tsens(12) ),
ccqair,cond/conv
qair,rad cc x Lfab qHair,rad
ccsens x L fab İ fab , x L fab¨  c,gap
Tfab
where q cc
air,cond/conv q Hcc
 —convective q xcc Lfab heat§ flux from
 Hsens the · © H A ¹ dt fab
(ccqconv ccair,cond/conv
qqair,cond/conv x  Lfab cc ) x 0qcc hc,flcond
sens
qcond k(airTg( T T) fab x¨ 0h1),c,gap(1TIJfab 
¸  Tsens cc),
qair,cond/conv kairh(c,gap
sens
Tq)cc (Tfab x qLcc fab  Tsens ),
 xp)Lsens
rad
T
flame to cc
thecc h fabric, Nu
air,cond/conv
q cc x L ,fab flux
—heat H by radiation
H cc
cond
(qTfabfab
cc xfab q q cc ) x 0 hxc,flair,rad
c,gap
Lggap Tfab h c,gap© fab ( 
T  x L fab
 T ¹ ) , where m
hc,gapsens Nu x Lqfab
—density , the air,cond/conv
of sensor, (c — , with-air-g
Lq § TtheL4(fab T ),
x L air,rad
air,cond/conv fab sens sens sens air,rad
Tqfrom cc conv rad x 0 4 ·x L fab L fab  Lgap
Tfab x 0 ), (qconv
fabair,rad ccx Lfabthe
fab
fabric
cc ) x H0 ¨ hfab
qqair,cond/conv
ccrad V to sensor 
cc (Txg LkTairfab(T )sens), ¸ qcc
qair,cond/conv across
T the air gap, specific heat k of (T )the sensor, and Asens—surface
gap
hsens §©¨flux
c,fl
kairp4by Nu
(·T )conduction/convection x 0, 4 · ¹ , cond hc,gap Nu air qsenscc , hi (Tfab  Tsens ), with-air-gap configuration
,fl (Tg  Tfab x 0 ), (qqq cc mc dT
fab
cc —heat V c,gap T  T ¸ §
 4· cc
conv  q
air,rad
air,cond/conv xTradcc
hLc,gapqfab ccx 0qccNu
)sens ¨ hc,flT
§ § fab(T4g¸xLT,fab

Lgap
H
), 4 · (T
x ,0 h·c,gap
sens
cc 
area of the sensor.
L mc gapp · dT
T
 T sens ), ¸ cc the fabric
qair,rad fab Vto©¨¨Tthe ©
air,cond/conv 1
fab
HfabALgap x L1 fab 
dtacross Tsens¸ ¸the ¹ q
, air,rad
fab T sens ) , q cc ¨conditions ¸ , $
ab4 fab x 0 ¹
 4 ·T from x L fab
xL Lfabfab
sensor
¹mc sens airx gap, L fab
Boundary sens
H
 A dtfor the sensor: $ ª 1 º
T ¸ § © § § xsens · · · ¹ § mc © · ¹
, © ccH 1fab·  k1pHccsens HTsensdT
x Lfab
fab
cc x L 4 fab  4
) ·¹¸ ¸¹, , the cc inner dT ª 1 708 º «§ Ra · ».
cc Vqconductance
qair,cond/conv 3
¹hq—interfacial
ab sens sens
ccT
p
§¨©sens  cc
fab
 Hxsens Lfab·
 T¸sens 4 ·¸ q,ccHi
air,rad mc p1¨x Lfab
§¨qfabair,rad dT ¸(Hbetween
T , with-air-gap qsens ¨ configuration; Nu¸ 1 ,1.44 «1   «¨ ¸  1»conf
x Lq qair,cond/conv
 
sens
q cc fabsens h H Nu H
 A1 air
H , , dt q cc H
 A q cc dt  q cc » , with-air-gap
(13a)
the ¨c,gap © fab¸©and ¹ , Tfabfab gap © ¹air,rad ¬ Rax Lfab¼  Lgap © 5 830 ¹

1HLfab 1  H sens¹ · surface sens x Lof fabric  Lthe ¹sens sens
senssensor,
x L L
§ H¨©A1H fab¹sens1 dt · ¸¹
sens sens air,cond/conv
 ¹ ,¸ 
air,rad
fab © HHgap x L fab
sens
«¬ »¼
 
sens
1  H
 H
 · ¹
Hqsens
cc
—temperature q cc
sens q oncc¨h i ( hTthefab  q cc
( T
inside
T sens ),
sens

sens
q with-air-gap
¸cc
surface  T of
) 
, configuration;
the , q cc
with-air-gap
sens q cc q cc hid(: 
configuration;
q cc  Tsens ), with-air-gap
Tfab air,cond/conv 'xE Lfab  Lgap configuration;
, with-air-gap configura
ccq©air,rad xTsens
air,rad
air,cond/conv x L sens
ccH fabc,gap  qpHair,cond/conv
air,rad
fab x L air,cond/conv
fab senssens ¸ Hq sens cc
qsens fab § mc · dT
ccsens ¹ surface
fab x Lfab  Lgap
, with-air-gap configuration;
sens
H sens ¹ fabric,cc Tsens —temperature
q h ¨ ( T on ¸ the 
, Lfab),of
ª Lgapthe 1
ºqcc
$ P exp(  ),
 Tsens ), Hsensor,
air,cond/conv
cc
xkL fab q sens
( cc
T ) c,gap
© h H
( T
A
fab
¹ x T L fab ),
dt
sens
$ with-air-gap
§ · configuration; h ( T dt T ), without-air-gap
with-air-gap RT configuration.
ª configuration; 1 º
$
ab x L qhair,cond/conv
sens h Nu —convective
ccNu air sens h
(Tfab  T1sens
, i ª(
mass Tfab 1 708
transfer sens º  T «
coefficient, ) , Ra 3
» sens i fab sens
ª 1 708 º «§ Ra · $ »
$
fab
hc,gap (Tfab x L  Tsens c,gap ), q msens x L fab h1i 
(ccT )h1.44
c,gap fab  ),x with-air-gap
sens L fab  ¨ configuration;
sens
¸  1»1 . Nu
3
cch kair
Lqgap  T«¬fab xLRa »¼ Tsens«)©$, 5 830 1 configuration;
1.44 «1  »ª  «¨ 5 830 º  1» .
¸(13b)
$
fab qair,cond/conv Nu , q cc
(air,rad q cc ª ¹ , with-air-gap º Ra
1
(Tfab x L  Tsens ), c,gap x k
L fab sens
( T )
c,gap
ª fab air,cond/conv
1 708 ª «
¼ x
«
L§ 1
RaL
º ·
$
»
¼ » ª 1¬ 708 º
$
¼ « § Ra © · 3 ¹» »
L
air fab gap 3
 1»1. 1.44 «1  «
¬ ¸ 1 . ¼
hc,gap 2008, §Nu mcVol. p ·(T ,No. 1  11.44 1 º «§» Ra  «¨ · 3 »  «¨ 5 830
$
Nu »¸ Nu
JOSE gap 14,dT
cc) 1.44h,iª(1T'fab 1708 «T
fab
cc
qsens ¨§Nu dk: airL¸ 1qgap  E ¬ ),Ra with-air-gap
¼ 5 configuration;

830 1 . ¬ Ra ¼ »
hc,gap Nu mc  Ap ¹·sens sens , « ),Ra »¼ «© 5 830
sens ¨ ©
«¬ ¹ ¸ »¹ »¼d$ : P exp( 'E ), «¬ © ¹ »¼
cc © H P exp( dT
dt ¬
qsens ¨§ mcdtLp gap ¸· dT , RT «¬ ª »¼1 º dt RT
 fab x L
fab

cc  qrad
(qconv cc ) x 0 hc,fl (Tg  Tfab x 0
),

V §¨ Tfab 4  Tsens 4 ·¸ PROTECTIVE CLOTHING MODELING 93

cc
qair,rad © x Lfab ¹,
x Lfab § 1 1  Hsens
 ·
¨   ¸
H
© fab
2.3. Natural Convection Hsens ¹ Air Gap
in the Nominal fabric thickness is chosen at the pressure
H sens Between the Fabric and the Sensor of 10 gf/cm2.
ccFor modeling the
qair,cond/conv hc,gap (Tfab x response
thermal L fab
 Tsensof), thermal
x L fab
protectivek clothing exposed to a flash fire in 3. MODEL VALIDATION
air (T )
hc,gap Nu ,
thermal protective performance (TPP) tests,
Lgap
convective heat transfer in the air gap between To compare the prediction based on the model
§ mc · dT with experimental data, different fabric systems,
ccthe fabric
qsens ¨
p and the sensor
¸ , is simulated as natural
© H A ¹sens
convection in dt a horizontal enclosure heated from one- and two- to multilayer systems, were
ccfrom qbelow.
qsens cc
air,rad  qHollands,
cc
air,cond/conv xUnny, Raithby,
, with-air-gap selected and tested in different configurations.
and configuration;
Lfab  Lgap
Konicek [12] presented the correlation for air in a The multilayer system consisted of three different
cchorizontal
qsens  Tsens ), with-air-gap
hi (Tfabenclosure heated from below: configuration; fabric layers: an outer shell, a moisture barrier,
ª 1 º
$ and a thermal liner, from the exterior to the
$
ª 1 708 º «§ Ra · 3 » (14) interior of clothing ensembles, respectively.
International Journal of Occupational Safety and Ergonomics 2008.14:89-106.

Nu 1  1.44 «1  »  «¨ ¸  1» .
¬ Ra ¼ 5 830 ¹ The two-layer system included an outer shell
«¬ © »¼
and a moisture barrier, and the one-layer system
d : The notation 'E[ ]° indicates that if the argument consisted of a shell only. These fabric systems
P exp( ),
dtin the squareRTbrackets is negative, the quantity were exposed to a high-intensity flash fire
should be taken as zero. simulation in a TPP [16] test configuration as
shown in Figure 1. Kombat™ (60/40 Kevlar®/
2.4. Numerical Procedure polybenzimidazole [PBI] blend) was the shell
fabric of the garment system, ComfortZone™
The finite volume method is adopted to solve
(flame retardant polyurethane film on Basofil®/
the differential equations, which are the energy
Aramid blend spunlace) was the moisture
equation for the fabric, the solid phase continuity
barrier, and Aralite (Aramid batt quilted to
equation, the gas phase diffusivity equation, and
Nomex®) was the thermal liner. The tests were
the heat transfer equation for the sensor [14,
performed under two test configurations: with
15]. The Crank-Nicholson scheme is applied
and without an air gap. Nominal thickness of
to discretize the transient partial differential
the air gap (0.00635 m; 1/4 in.) was adopted for
equations. Due to nonlinearities in this system,
the TPP test. Table 1 shows the thermophysical/
the Gauss-Seidel point-by-point iterative scheme
geometrical properties of the fabric utilized in
is employed to solve these equations. To avoid
the computations. Table 2 lists the radiation
divergence of the iteration method, the under-
parameters used in the computations [9]. Table 3
relaxation procedure is utilized. The value of the
gives the thermophysical/geometrical properties
under-relaxation parameter is 0.8. The solution
of the TPP sensor. Table 4 details the thermal
procedure is as follows: all variables are known
properties of the flame and the ambient air, and
at the initial state, then the program progresses
the initial data of the fabric and the air gap.
in given time increments. The variables at the
Figures 3 and 4 show the comparisons of
previous time step are used as guessed values for
computational and experimental results of
the variables at the current time step. The new
temperature histories on the surface of the sensor
values of variables are computed by visiting each
for the one-layer shell fabric with and without
grid point in a certain order. Then the iterations
an air gap, respectively. Six-second exposure
are repeated until the changes in the solutions
was applied to the test configuration with an air
become smaller than 10–6.
gap and 3-s exposure to the one without an air
gap. Overall agreement can be observed from
2.5. Determination of Fabric Thickness these comparisons of model prediction and
Fabric thickness is evaluated by performing experimental results. However, in Figure 4 (the
measurements with the Kawabata instrument. configuration without an air gap) the temperature
Thickness change depends on the applied load. rise predicted by the model is relatively lower

JOSE 2008, Vol. 14, No. 1

 94 G. SONG ET AL.

TABLE 1. Thermophysical/Geometrical Properties of the Fabric [2]

Outer Shell: Moisture Barrier: Thermal Liner:
Property Symbol Unit Kombat 7.5 oz/yd2 ComfortZone Aralite®
Fiber density ρds kg⋅m–3 1 384 1 295 1 380
Fiber specific heat (cp)ds J⋅kg–1⋅K–1 1 420 1 325 1 200
Fiber thermal conductivity kds W⋅m–1⋅K–1 0.179 0.144 0.130
–3
Thickness L m 0.56 × 10 0.73 × 10–3 1.66 × 10–3
Fiber volume fraction εds 0.334 0.186 0.115
t
Fiber regain at 65% relative R 'E 0.084 0.038 0.045
humidity : f P exp( 
RT ³
) dt .
Tortuosity τ 0 1.50 1.25 1.00
2 –1
Ratio of effective diffusivity Dsolid / d f s 2.34 × 10–4 2.34 × 10–4 2.34 × 10–4
of bound water in the fiber
to average fiber diameter İ
IJ hi
International Journal of Occupational Safety and Ergonomics 2008.14:89-106.

Interfacial conductance W⋅m–2⋅K–1 55.0 70.0 40.0

between inner surface of
fabric and sensor
cc
qcond
Notes. 7.5 oz/yd2 = 254.3 g/m2.
cc
qair,rad
t cc 'E
qair,cond/conv
: P ³
exp(
TABLE 2. Radiation Parameters  [3] )dt .
Tfab xRT L fab
TABLE 3. Thermophysical/Geometrical Proper­
0 ties of the Thermal Protective Performance
2 TPP (TPP) sensor
Property Dsolid /Fabric
Symbol df Flame Sensor
Emissivity İ 0.9 0.02 0.95 TPP
Property Symbol Unit Sensor
Transmissivity IJ 0.01 — —
Mass of copper m g 17.8
Notes. TPP—thermal protective performance.
cc
qcond Specific heat of copper cp cal/g⋅°C 0.0927
2
cc Surface area of copper A cm 12.56
qair,rad
cc
qair,cond/conv
TABLE 4. Initial Conditions for the Fabric and the Air Gap of the Thermal Protective Performance
Tfab Test,
(TPP) Configuration x L faband the Thermal Properties of the Flame and the Ambient Air

Condition/Property Symbol (Unit) Unit Value

Initial fabric temperature T0,fap °C 26.0
Initial relative humidity in fabric φ0,fab — 0.68
Initial relative humidity in air gap φ0,gap — 0.68
Initial sensor temperature T0,sens °C 28.0
Ambient temperature Tamb °C 26.0
Ambient relative humidity φamb 0.68
Convective heat transfer coefficient between hc,fl W⋅m–2⋅K–1 120
flame and outer surface of fabric
Flame temperature Tfl °C 1 450
–1
Convective mass transfer coefficient in air gap hm,gap m⋅s 0.021
Ambient convective mass transfer coefficient hm,amb m⋅s–1 0.021
Total gas phase pressure pγ N⋅m–2 1.01325 × 105
Thickness of air gap Lgap m 6.35 × 10–3

than the experimental data before 10 s and was condensation water on the surface of the
higher after 10 s. This may be because vaporized sensor. Figures 5 and 6 compare computational
moisture as a result of temperature gradient and experimental results for the two-layer
moves towards the sensor and condensates on fabric system. The exposure times applied in
the surface of the sensor. In the experiment, there Figure 5 (with an air gap) and Figure 6 (without

JOSE 2008, Vol. 14, No. 1

 PROTECTIVE CLOTHING MODELING 95

an air gap) are 7.2 and 14.2 s, respectively. The 23 s for the configuration with an air gap and 15 s
computation curve predicts a higher temperature for the one without an air gap. The comparisons
rise comparing with experimental data in both in Figures 7 and 8 demonstrate good agreement
configurations, particularly after 10–20 s from between model predictions and experimental
the start of exposure. The moisture barrier in this tests under both test configurations. The effect of
fabric system may contribute to the difference the moisture barrier layer, like in the two-layer
between computational and experimental results. system, was minimized in the multilayer system
In the model, the nature of moisture transfer as a result of the addition of a thicker thermal
in the barrier layer was not considered and the insulation layer. The exposure times for these
assumption was the same as that used in the case fabric systems were selected on the basis of the
of textile layers. In Figures 7 and 8, computational estimation of energy transferred through a fabric
and experimental results for the multilayer fabric system and air layers that could cause second-
are compared for different test configurations. degree burn.
International Journal of Occupational Safety and Ergonomics 2008.14:89-106.

The exposure times for the multilayer fabric were

60
exposure time = 6.0 s model prediction
at the sensor with spacer replicate No. 1
replicate No. 2
55

50
Temperature (°C)

45

40

35

30

25
0 10 20 30 40 50 60
Time (s)

Figure 3. Comparison of computational and experimental results of temperature histories on the

surface of the sensor for a one-layer system (Kombat™ 7.5 oz/yd2) with an air gap.
Notes. 7.5 oz/yd2 = 254.3 g/m2.

JOSE 2008, Vol. 14, No. 1

 96 G. SONG ET AL.

50
model prediction
exposure time = 3.0 s
replicate No. 1
at the sensor with spacer replicate No. 2
replicate No. 3
45
Temperature (°C)

40

35
International Journal of Occupational Safety and Ergonomics 2008.14:89-106.

30

25
0 10 20 30 40 50 60
Time (s)
Figure 4. Comparison of computational and experimental results of temperature histories on the
surface of the sensor for a one-layer system (Kombat™ 7.5 oz/yd2) without an air gap.
Notes. 7.5 oz/yd2 = 254.3 g/m2.

70
exposure time = 14.2 s model prediction
at the sensor with spacer replicate No. 1
65 replicate No. 2
replicate No. 3

60
Temperature (°C)

55

50

45

40

35

30

25
0 10 20 30 40 50 60
Time (s)
Figure 5. Comparison of computational and experimental results of temperature histories on the
surface of the sensor for a two-layer (Kombat™ 7.5 oz/yd2, ComfortZone™) system with an air gap.
Notes. 7.5 oz/yd2 = 254.3 g/m2.

JOSE 2008, Vol. 14, No. 1

 PROTECTIVE CLOTHING MODELING 97

70
exposure time = 7.2 s model prediction
at the sensor without spacer replicate No. 1
65 replicate No. 2
replicate No. 3

60
Temperature (°C)

55

50

45

40
International Journal of Occupational Safety and Ergonomics 2008.14:89-106.

35

30

25
0 10 20 30 40 50 60
Time (s)
Figure 6. Comparison of computational and experimental results of temperature histories on the
surface of the sensor for a two-layer system (Kombat™ 7.5 oz/yd2, ComfortZone™) without an air
2 2
gap. Notes. 7.5 oz/yd  = 254.3 g/m .

70
exposure time = 23.0 s model prediction
at the sensor replicate No. 1
65 replicate No. 2
replicate No. 3
replicate No. 4
60
Temperature (°C)

55

50

45

40

35

30

25
0 10 20 30 40 50 60
Time (s)

Figure 7. Comparison of computational and experimental results of temperature histories on the

surface of the sensor for a two-layer (Kombat™ 7.5 oz/yd2, ComfortZone™, Aralite®) system with an
air gap. Notes. 7.5 oz/yd2 = 254.3 g/m2.

JOSE 2008, Vol. 14, No. 1

 98 G. SONG ET AL.

70
exposure time = 15.0 s model prediction
no spacer replicate No. 1
65 at the sensor
replicate No. 2
replicate No. 3
replicate No. 4

Temperature (°C) 60

55

50

45

40
International Journal of Occupational Safety and Ergonomics 2008.14:89-106.

35

30

25
0 10 20 30 40 50 60
Time (s)
Figure 8. Comparison of computational and experimental results of temperature histories on the
surface of the sensor for a three-layer (Kombat™ 7.5 oz/yd2, ComfortZone™, Aralite®) system without
an air gap. Notes. 7.5 oz/yd2 = 254.3 g/m2.

4. MODEL APPLICATION convection is negligible. Therefore, radiation

and conduction heat transfer will dominate heat
transfer mechanisms across the air gap.
4.1. Temperature and Moisture Figure 9 describes temperature distribution
Distribution predicted by the numerical model in a
multilayer fabric, the air gap, and the sensor at
Predictions from the established model are
different moments of time. Distance in Figure 9
performed for the duration of 4-s exposure to a
represents the thickness of the fabric composite
flash fire, with an introduction of an air gap. The
(0–2.95 mm) and the size of the air gap
air gap is 0.00635 m (1/4 in.) long. To examine
(2.95–9.20 mm). The end of the distance, at
the heat and moisture transfer feature during the
9.20 mm, represents the surface of the sensor.
cooling period (postexposure), computations
During 4-s exposure, the surface temperature of
continue until 60 s after the end of exposure.
the fabric of the outer shell rises fast, from room
The temperature of the hot gas and the ambient
temperature to more than 400 °C, and maximum
temperature gradually decrease after 4 s of
temperature can be observed at the outer surface
burning [17]. From the expression for the Nusselt
of the fabric at ~4 s. After exposure, surface
number, natural convection will contribute to
temperature falls slowly and in ~60 s surface
heat transfer across the enclosure when the
temperature drops back to a point that is very
Rayleigh number, Ra, is greater than 1 708. The
close to room temperature. In comparison to the
maximum Rayleigh number, for all computed
temperature on the outer surface of the fabric,
cases of the TPP test was 1 123. Therefore, for
the temperature on the inner surface of the fabric
the case computed in this paper, Ra is always
increases slowly and maximum temperature is
smaller than 1 708, which means that natural
predicted at ~10 s in 4-s exposure to flash fire.

JOSE 2008, Vol. 14, No. 1

 PROTECTIVE CLOTHING MODELING 99

time = 0.0 s
50 0
time = 0.5 s
time = 1.0 s
time = 2.0 s
time = 4.0 s
40 0
time = 5.0 s
time = 10.0 s
time = 20.0 s
time = 40.0 s
Temperature (°C)

30 0 time = 60.0 s

20 0
International Journal of Occupational Safety and Ergonomics 2008.14:89-106.

10 0

0
0 0.002 0.004 0.006 0.008 0.010
Distance (m)
Figure 9. Temperature distribution in the fabric, the air gap, and the sensor at different times.

time = 0.0 s
time = 0.5 s
time = 1.0 s
1.5
time = 2.0 s
1.4 time = 4.0 s
time = 5.0 s
1.3
time = 10.0 s
1.2 time = 20.0 s
1.1 time = 40.0 s
time = 60.0 s
1.0
Relative Humidity

0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0 0.0005 0.0010 0.0015 0.0020 0.0025 0.0030
Distance (m)

Figure 10. Distribution of relative humidity in the fabric at different moments of time.

JOSE 2008, Vol. 14, No. 1

 100 G. SONG ET AL.

This phenomenon indicates that energy stored In Figure 11 the distribution of fiber moisture
during exposure moves towards the back of the regain in the fabric (bound water in the solid
fabric layers during the postexposure period. A phase as shown in Figure 2) at different moments
significant temperature difference is predicted of time are investigated. Regain is defined as the
between the inner layer of the fabric and the ratio of the mass of absorbed water in the fiber to
surface of the sensor. the mass of dry fiber. At the beginning the fiber
Figure 10 depicts distribution of relative in the fabric is in the state of equilibrium. Fiber
humidity in the fabric (vapor in gas phase in the regain in the fabric decreases during 4-s exposure
fabric as shown in Figure 2) at different moments to the flash fire and it continues to decrease until
of time. Relative humidity in the outer layer of minimum values are achieved, which corresponds
garment decreases significantly when exposure to equilibrium at those temperature and relative
to flash fire starts and it increases in the latter humidity conditions. If the cooling time is long
layers. This indicates that the moisture vaporized enough, both fiber regain and moisture content in
International Journal of Occupational Safety and Ergonomics 2008.14:89-106.

in the outer layer moves towards the inside fabric the fabric will go back to what they were at the
layer and eventually into the air gap because of initial state.
the temperature gradient as described in Figure 9. Figure 12 presents distribution of vapor density
At ~10 s, the model predicts minimum relative in the multilayer fabric at different moments
humidity in the fabric system; then the fabric of time. Vapor density increases because
system gradually regains its moisture. After temperature increases and causes the phase
the temperatures on the outer surface and the transition from bound water to water vapor. In
inner surface of the fabric become low enough, addition, as the temperature gradient develops
relative humidity starts growing back to its initial during exposure, moisture in the outer layer
distribution. moves towards the inner layer and contributes to
the change in vapor density.

time = 0.0 s
time = 0.5 s
time = 1.0 s
0.15
time = 2.0 s
0.14 time = 4.0 s
time = 5.0 s
0.13
time = 10.0 s
0.12 time = 20.0 s
0.11 time = 40.0 s
time = 60.0 s
0.10
Fiber Regain

0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0
0 0.0005 0.0010 0.0015 0.0020 0.0025 0.0030
Distance (m)

Figure 11. Distribution of fiber moisture regain in the fabric at different moments of time.

JOSE 2008, Vol. 14, No. 1

 PROTECTIVE CLOTHING MODELING 101

time = 0.0 s
time = 0.5 s
time = 1.0 s
0.20
time = 2.0 s
time = 4.0 s
0.18
time = 5.0 s
time = 10.0 s
0.16
time = 20.0 s
time = 40.0 s
0.14 time = 60.0 s
Vapor Density (kg/m )
2

0.12

0.10
International Journal of Occupational Safety and Ergonomics 2008.14:89-106.

0.08

0.06

0.04

0.02

0
0 0.0005 0.0010 0.0015 0.0020 0.0025 0.0030
Distance (m)

Figure 12. Distribution of vapor density in the fabric at different moments of time.

0.65

0.64

0.63
Fabric Weight per Unit Area (kg/m )
2

0.62

0.61

0 .60

0.59

0.58

0.57

0.56

0.55
0 10 20 30 40 50 60
Time (s)

Figure 13. Predicted change in fabric weight during and after exposure.

JOSE 2008, Vol. 14, No. 1

 Da HJ
Deff ,
W
wT
102 G. SONG ET AL. keff  qconv cc  x
cc  qrad ;
wx x 0
0

wT
1.0  keff
wx
 qair,rad
cc cc
 qair,cond/conv x Lfab
, with-ai
x Lfab

0.9 wT
 keff hi (Tfab  Tsens ), without-air-ga
wx x Lfab
x Lfab
0.8
wUv
hm,amb (Uv,amb  Uv ) x  Deff ;
Relative Himidity of Air Gap
0
0.7 wx x 0
wUv
0.6 hm,gap (Uv  Uv,air ) x Lfab  Deff , with-air-g
wx x Lfab

0.5 wUv
 Deff 0, without-air-gap configuration;
wx x Lfab
0.4
cc
qair,rad
0.3 cc
qair,cond/conv
International Journal of Occupational Safety and Ergonomics 2008.14:89-106.

Tfab
0.2 x Lfab

cc  qrad
(qconv cc ) x 0 hc,fl (Tg  Tfab x 0
),
0.1
V §¨ Tfab 4  Tsens 4 ·¸
0 qcc © x L fab ¹,
0 10 20 30 4 0air,rad x 50
Lfab § 160 1  H sens ·
Time (s) ¨   ¸
© H fab Hsens ¹
H
Figure 14. Predicted change in relative humidity in the air gap during and after exposure.
sens
cc
qair,cond/conv hc,gap (Tfab x L fab
 Tsens ),
x L fab

The calculated fabric weight per unit area 4.2. Effects of Variations
kair (T )on Clothing
during and after exposure is examined in h Nu
Protective Performance
c,gap ,
L gap
Figure 13. It shows that fabric weight decreases
The three-layer skin § mcmodel was
dTused to predict the
significantly during the first 20 s. This is mainly p ·
performance ccof a thermal
qsens ¨ , fabric system
¸ protective
because when temperature increases, moisture © H A ¹sens dt
in terms of skin burn. It was based on Henriques
(bound water) in the fiber vaporizes and moves
and Moritz’s cc work,
qsens  qwhich
cc
air,rad x Lfab  Lgap , with-air-ga
cc that destruction
 shows
qair,cond/conv
into the air gap. This phenomenon relates well
of the tissueqcclayer hlocated at the epidermis/dermis
to the change in fiber moisture regain as shown sens i (Tfab  Tsens ), with-air-gap configuration;
interface in human skin starts when the tissue $
in Figure 11, which shows that minimum fiber ª 44 °C 1 º
temperature of the basal layer1 rises above
$
708 º «§ Ra ·
moisture regain can be obtained after 20 s. ª 3
»
Nu 1  1.44
[18]. The destruction 1  be modeled
rate«can »  «¨using a¸  1» .
¬ Ra ¼ 5 830
Figure 14 presents the predicted change in
first-order chemical reaction, i.e., «¬© ¹ »¼
relative humidity in the air gap during and after
d: 'E
exposure. After 4-s exposure relative humidity P exp( ), (15)
drops significantly and reaches the minimum dt RT
at ~10 s from the start of exposure. From where Ω—quantitative measure of burn damage
Figure 9 the model predicts temperature change at the basal layer or at any depth in the dermis;
in the air gap and higher temperature can be P—frequency factor, s–1; E—activation energy
observed at ~10 s. When temperature increases, for skin, J/mol; R—universal gas constant,
saturation pressure also increases. Therefore, 8.315 J/kmol⋅K; T—absolute temperature at
relative humidity in the air gap decreases even the basal layer or at any depth in the dermis,
though the moisture content in the air gap K; t—total time for which T is above 44 °C
increases. (317.15 K).
Integration of this equation yields
t
'E
: ³ P exp( RT )dt.
0
(16)

Dsolid / d f2
JOSE 2008, Vol. 14, No. 1 İ
IJ
 PROTECTIVE CLOTHING MODELING 103

Integration is performed from the time when The effect of fiber density, conductivity,
the temperature of the basal layer of the skin, T, and capacity on the protective performance of
exceeds or equals 44 °C. Henriques found that clothing was investigated. As different fibers
if Ω ≤ 0.5, no damage will occur at the basal were used in shell fabrics and thermal liners,
layer [18]. If Ω is 0.5–1.0, there will be first- different ranges of values of properties were
degree burns, whereas if Ω > 1.0, second-degree selected. Table 6 demonstrates the effect of
burns will result. The damage criteria can be a change in fiber density on the protective
applied to any depth of skin provided appropriate performance of clothing. The predictions were
values of P and ∆E are used. Mathematically, a made by changing the fiber density in the
second-degree burn has been defined as Ω > 1.0 numerical model while other parameters were
at the epidermis/dermis interface. Details about the same as those employed in the model. The
the structure of the skin model as well as the results showed that increasing fiber density of
properties and parameters used for Henriques’ shell fabrics and thermal liners, improved thermal
International Journal of Occupational Safety and Ergonomics 2008.14:89-106.

burn integral are given in the Appendix on p. 106. protective performance. In the model the effective
Table 5 illustrates the prediction of protective density of the fabric was employed, which is a
performance for the three-layer clothing system combination of the density of the fiber, moisture,
with different sizes of the air gap. Table 1 lists the and trapped air in the fabric. Therefore, a change
properties used in the numerical model. Minimum in fiber density affects the effective density of
exposure time required to develop a second- the fabric. An increase in fabric effective density
degree burn was predicted. The results show causes lower fabric thermal diffusivity, hence
that with the typical air gap size of 0.00635 m heat transfer in the fabric system slows down and
(1/4 in.), minimum exposure time was predicted consequently improved protective performance
at 10.3 s under exposure to flash fire. For the is predicted. A change in fiber conductivity
selected air gap range of 1.59–6.35 mm, the and thermal capacity affects thermal protective
model predicts a significant decrease in minimum performance as illustrated in Tables 7 and 8. This
exposure time when the size of the air gap was achieved by changing one parameter and
decreases. As air is a good insulator, decreasing leaving the others the same as in the numerical
the size of the air gap will decrease the value of model. As described in the development of the
insulation provided by the air layer between the model, the fabric was modeled as porous media
fabric and the skin. However, when the air gap and the system consisted of fiber, trapped air,
was increased beyond 6.35 mm, no significant and moisture. A large part of the insulation
change in minimum exposure time was observed. of the thermal liner was provided by the air
This agrees with Song’s [19] study that during trapped in the fabric. Therefore, change in fiber
exposure to a flash fire (84 kW/m2) natural conductivity and specific heat capacity within
convection may occur in the air gap between the a certain range did not significantly affect heat
clothing and human skin (sensor). When the air transfer as expected. The results obtained from
gap is large, convection increases the amount of the numerical model, as shown in Tables 7 and
energy transferred to the skin. 8, confirm this. However, the overall trend shows
that increasing fiber conductivity affects heat
TABLE 5. Air Gap Size and Minimum Exposure
Time Required to Generate a Second-Degree flow in the system and causes skin temperature to
Burn rise faster. As a result, minimum exposure time
Air Gap Size (m) Minimum Exposure Time (s) decreases as fiber thermal conductivity increases.
0.00159 7.7 For fiber specific heat capacity, the model results
0.00318 9.1 indicate that minimum exposure time increases
0.00635 10.3 with an increase in fiber specific heat capacity.
0.01270 10.7 It should be noted that the model analysis does
not include fiber decomposition reactions and its
effect on heat and mass transfer.

JOSE 2008, Vol. 14, No. 1

 104 G. SONG ET AL.

TABLE 6. Fiber Density and Predicted Minimum minimum exposure time. An increase in fabric
Exposure Time Required to Generate a Second- thickness provides better insulation and when
Degree Burn
there is exposure, the temperature difference
Fiber Density (kg⋅m–3) Minimum between the exposed and skin sides of the fabric
Shell Fabric Thermal Liner Exposure Time (s) increases. Additionally, during exposure, a thicker
1 107 1 104 9.7
fabric normally stores more thermal energy than
1 384 1 380 10.3
a thinner one; therefore, energy transfer to human
1 661 1 656 11.0
skin (or a sensor) is lower.
2 076 2 070 12.1
The effects of the initial temperature of a
fabric and of environmental conditions on the
TABLE 7. Fiber Thermal Conductivity and
Predicted Minimum Exposure Time Required to performance of protective clothing were examined
Generate a Second-Degree Burn using the numerical model. Table 10 lists the
Fiber Thermal Conductivity results of model predictions with different initial
International Journal of Occupational Safety and Ergonomics 2008.14:89-106.

(W⋅m–1⋅K–1) Minimum temperatures of a fabric; a clothing system can

Shell Fabric Thermal Liner Exposure Time (s) provide more protection if the initial temperature
0.143 0.104 10.8 of the fabric is relatively low. Comparatively,
0.179 0.130 10.3 when exposed to a thermal hazard, the fabric at
0.215 0.156 10.0 a lower temperature will store more energy than
0.269 0.195 9.6 a fabric system at a higher temperature and it will
generate a slow temperature rise on the side of the
TABLE 8. Fiber Specific Heat Capacity and skin. As a result, the fabric system will perform
Predicted Minimum Exposure Time Required to
Generate a Second-Degree Burn
better. Moisture regain in the fiber changes both
fiber conductivity and specific heat, and during
Fiber Specific Heat Capacity
(J⋅kg–1⋅K–1)
intense thermal conditions a hot stream, which
Minimum
Shell Fabric Thermal Liner Exposure Time (s)
contributes to human skin burn, is generated.
1 136 960 9.8
Table 11 shows that the effect of moisture on
minimum exposure time is minor when fiber
1 420 1 200 10.3
moisture regain is relatively low. When it is high,
1 704 1 440 10.9
however, minimum exposure time is higher.
2 130 1 800 11.8
TABLE 10. Initial Temperature of the Fabric and
The results obtained from the numerical Predicted Minimum Exposure Time Required to
Generate a Second-Degree Burn
model suggest that the thickness of the fabric
significantly affects its thermal protective Initial Temperature (°C) Minimum Exposure Time (s)
performance. Table 9 reports on the thermal 5 11.4
protective performance of fabric systems with 26 10.3
different thickness of shell and thermal liners: an 35 9.7
increase in fabric thickness significantly increases 50 8.5

TABLE 9. Fabric Thickness and Predicted TABLE 11. Fabric Moisture Regain and
Minimum Exposure Time Required to Generate Predicted Minimum Exposure Time Required to
a Second-Degree Burn Generate a Second-Degree Burn

Thickness (mm) Fiber Moisture Regain (%) Minimum

Minimum Exposure
Shell Fabric Thermal Liner Time (s) Shell Fabric Thermal Liner Exposure Time (s)

0.45 1.33 7.8 16.8 9.0 10.9

0.56 1.66 10.3 8.4 4.5 10.3

0.67 1.99 13.3 4.8 2.3 10.2

0.84 2.46 21.0 2.4 1.1 10.1

JOSE 2008, Vol. 14, No. 1

 PROTECTIVE CLOTHING MODELING 105

5. CONCLUSIONS PA, USA: American Society for Testing

and Materials (ASTM); 1997. p. 334–52.
A numerical model of coupled heat and moisture 6. Crow RM, Osczevski RJ. The interaction of
transport in protective clothing during exposure to water with fabrics. Text Res J. 1998;68(4):
a flash fire was investigated and compared with 280–8.
experimental results. This model can be used to 7. Vafai K, Sözen M. A comparative analysis
predict the thermal response of a protective fabric of multiphase transport models in porous
when it is exposed to a flash fire. It can also be media. In: Tien CL, editor. Annual Review
used to obtain the distribution of temperature of Heat Transfer. Vol. 3. New York, NY,
USA: Hemisphere; 1990. p. 145–62.
and moisture as well as information on the effect
of some properties of fabrics and an air gap on 8. Gibson PW. Multiphase heat and mass
transfer through hygroscopic porous media
protective performance under exposure to a flash
with applications to clothing materials
fire. A typical three-layer protective clothing
(Technical report Natick/TR-97/005).
International Journal of Occupational Safety and Ergonomics 2008.14:89-106.

system provides protection during ~9–12-s Natick, MA, USA: U.S. Army Natick
exposure to a flash fire, without causing a second- Research, Development, and Engineering
degree burn. An air gap layer between the fabric Center; 1996.
and the skin provides extra insulation and slows 9. Torvi DA. Heat transfer in thin fibrous
down heat transfer to the skin. The thickness of materials under high heat flux conditions
the fabric system is an important factor in the [doctoral dissertation]. Edmonton, Canada:
performance of thermal protective clothing. University of Alberta; 1997.
10. Whitaker S. Simultaneous heat, mass, and
momentum transfer in porous media: a
REFERENCES theory of drying. In: Hartnett JP, Irvine Jr TF,
editors. Advances in Heat Transfer. Vol. 13.
1. Bouddour A, Auriault JL, Mhamdi- New York, NY, USA: Academic Press;
Alaoui M. Heat and mass transfer in wet 1977. p. 119–203.
porous media in presence of evaporation–
11. Morton W, Hearle J. Physical properties of
condensation. Int J Heat Mass Transfer.
textile fibres. Manchester, UK: The Textile
1998;41(15):2263–77.
Institute; 1993.
2. Stull JO. The effect of moisture on
12. Hollands KGT, Unny TE, Raithby GD,
firefighter protective clothing thermal
Konicek L. Free convective heat transfer
insulation: review of industry research.
across inclined air layers. J Heat Transfer.
In: Nelson CN, Henry NW, editors.
1976;98:189–93.
Performance of protective clothing (ASTM
STP 1386). West Conshohocken, PA, USA: 13. Chitrphiromsri P, Kuznetsov AV.
American Society for Testing and Materials Modeling heat and moisture transport
(ASTM); 2000. p. 557–76. in firefighter protective clothing during
flash fire exposure. Heat Mass Transfer.
3. Lee YM, Barker RL. Effect of moisture
2005;41:206–15.
on the thermal protective performance of
heat-resistant fabrics. J Fire Sci. 1986;4(5): 14. Patankar SV. Numerical heat transfer and
315–31. fluid flow. Washington, DC, USA: Taylor
& Francis; 1980.
4. Barker RL, Guerth-Schacher C, Grimes RV,
Hamouda H. Effects of mois­ture on the 15. Tannehill JC, Anderson DA, Pletcher RH.
thermal protective performance of firefighter Computational fluid mechanics and heat
protective clothing in low-level radiant heat transfer. Washington, DC, USA: Taylor &
exposures. Text Res J. 1996;76(1):27–31. Francis; 1997.
5. Lawson JR. Fire fighters’ protective cloth­ 16. American Society for Testing and Materials
ing and thermal environments of structural (ASTM) Standard test method for thermal
fire fighting. In: Stull JO, Schwope AD, protective performance of materials for
editors. Performance of protective clothing clothing by open-flame method (ASTM D
(ASTM STP 1273). West Conshohocken,

JOSE 2008, Vol. 14, No. 1

 106 G. SONG ET AL.

4108-87). West Conshohocken, PA, USA: experimental investigation. Am J Pathol.

ASTM; 1987. 1947;23:531–49.
17. Song G, Barker R, Hamouda H, 19. Song G. Clothing air gap layers and thermal
Kuznetsov A, Chitrphiromsri P, Grimes R. protective performance in single layer
Modeling the thermal protective performance garment. J Ind Text. 2007;36(3):139–205.
of heat resistant garments in flash fire 20. Stoll AM, Greene LC. Relationship between
exposures. Text Res J. 2004; 74(12):1033–40. pain and tissue damage due to thermal
18. Henriques FC Jr, Moritz AR. Studies of radiation. J Appl Physiol. 1959;14(3):373–82.
thermal injuries I: the conduction of heat 21. Takata A. Development of criterion for skin
to and through skin and the temperatures burns. Aerosp Med. 1974;45(6):634–7.
attained therein. A theoretical and
International Journal of Occupational Safety and Ergonomics 2008.14:89-106.

APPENDIX

A three-layer skin model was used in this study. Table 12 summarizes the thermophysical properties of
the skin layer.
TABLE 12. Skin Properties in a Three-Layer Skin Model
Human Skin Property Value
Epidermis thermal conductivity (W/m⋅°C) 0.255
3
density (kg/m ) 1 200
specific heat (J/kg⋅°C) 3 598
–5
thickness (m) 8.0 × 10
Dermis thermal conductivity (W/m⋅°C) 0.523
3
density (kg/m ) 1 200
specific heat (J/kg⋅°C) 3 222
–3
thickness (m) 2.0 × 10
Subcutis thermal conductivity (W/m⋅°C) 0.167
3
density (kg/m ) 1 000
specific heat (J/kg⋅°C) 2 760
–2
thickness (m) 1.0 × 10

The values of P and ∆E in Equation 16 are from Stoll and Greene [20] and Takata [21]:

epidermis
for T < 50 ºC P = 2.185 × 10124 s–1, ∆E/R = 93 534.9 K;
for T ≥ 50 ºC P = 1.823 × 1051 s–1, ∆E/R = 39 109.8 K;
dermis
for T < 50 ºC P = 4.32 × 1064 s–1, ∆E/R = 50 000 K;
for T ≥ 50 ºC P = 9.39 × 10104 s–1, ∆E/R = 80 000 K.

JOSE 2008, Vol. 14, No. 1