ff~13o

E~UH\GY
;.\NO
BUJLOJ~IGS
ELSEVIER Energy and Buildings 24 ( 1996 ) 105- 115

Measurement of thermal stratification in large single-cell buildings

M.N.A. Said, R.A. MacDonald, G.C. Durrant
Institute for Research in Construction, Nati onal Research Council, Ottawa, Ont.. KIA OR6, Canada

Received 31March1995 ; revised 21October1995

Abstract

This paper describes measured temperature profiles and thermal stratification in eight aircraft hangar buildings during the heating season.
Presented also is the predicted impact of thermal stratification on heating energy requirements. The hangar buildings included two main ceiling
heights (9.35 and 17.1 m (31 and 56 ft)), two ceiling types (fiat and Quonset), two types of heating systems (vertical discharge forced warm
air and downdraft convective unit heaters), and various types of large external doors. Measured stratification, expressed by floor-to-ceiling
temperature di fferential, is in the range 4--11 °C ( 7 .2-20 °F). Two air layers existed in the hangar bay area, a warm upper air layer and a
cot>ler lower air laye r. The lower air layer (up to 2 m (7 ft) high from the floor) is characterised by a steep vertical temperature gradient in
the range 0.8-2.6 °C/m (0.4--1.4 °F/ft). In the upper air layer, above 2 m, the vertical temperature gradient shallows considerably to 0.5 °CI
m (0.3 °F/ft). Results indicate that thermal stratification can have a significant impact on the building's heating energy requirements. In the
hangar buildings studied, predicted excess heating energy requirements due to 8 °C floor-to-ceiling temperature differential can be as much
as 38% as compared to the case with no stratification.

Keywords: Field measurements; Stratification; Temperature profiles; Heating energy; Large space buildings; Aircraft hangars

1. Introduction Stratification is common in many buildings particularly

Warm air tends to rise under the influence of buoyancy
forces which gives rise to a positive vertical temperature
gradient between the floor and the ceiling. The air temperature
in the working zone would be lower than the air temperature
near the ceiling. These conditions produce what is known as
thermal stratification. During the heating season, the presence
of a warm air layer below the ceiling will increase the roof
heat losses and may increase the stack effect. The latter will
increase air infiltration and exfiltration. More heat would be
required to maintain comfort temperature in the working
zone. The higher temperature near the ceiling may also affect
the performance of some lighting fixtures and optical fire
detectors.
Stratification heating effects in aircraft hangars are the
prime subject of this study. In other buildings such as man-
ufacture plants, indoor air quality and cooling-season strati-
fication effects can be important. Stratification may be
beneficial to keep the working zone clean of pollutants. Cool-
ing-season stratification effect can reduce the cooling load
because the stratified warm air layer below the ceiling acts
like an insulating buffer which can reduce the roof and light-
ing heat gain components (e.g. see Refs. [ 1-4]).
those with a large non-compartmented volume (such as an
aircraft hangar, gymnasium, atria etc.) . Bottcher et al. [ 5]
reported measured stratification with floor-to-ceiling temper-
ature differences of up to 18 °C ( 32 °F) in a poultry building
(5.2 m ( 17 ft) ceiling ridge height) with a single space heater
mounted 3 m ( 10 ft) above the floor and ceiling fans running
at low speed. Hejazi-Hashemi [6] measured stratification of
up to 2 °C (3.6 °F) in atrium buildings (366 m2 (3940 ft2)
floor area and 9 to 13 m ( 30 to 43 ft) high) during the heating
season with the air circulation fans operating. During the
summer, when the atrium is naturally ventilated, stratification
reached up to 10 °C ( 18 °F).
Dean et al. [ 1] reported vertical temperature profiles for
both winter and summer operating conditions in a factory
building ( 3159 m2 ( 34 000 ft2) floor area and 6 m ( 20 ft)
ceiling height) . They indicated that measured stratification
was less than 5.6 °C (10 °F) during the heating season.
Bagheri and Gorton [7] measured temperature profiles in a
nuclearreactorfacility (334m 2 (3600ft2 ) floorareaand 12.5
m ( 41 ft) ceiling height) during the heating season. The
objective of their experiments was to determine the heating-
season performance of a system designed to produce stratified
air-conditioning during cooling-season operation. They also
discussed the effects on stratification due to outdoor temper-

0378-7788/96/$15.00 © 1996 Elsevier Science S.A. All rights reserved

SSDJ0378 -7788(95)00966-4
106 M.N.A. Saia et al. I Energy and Buildings 24 ( 1996) 105-115

ature, supply diffuser and return air elevation, magnitude of
internal load, and the air-handling unit fan and coil operation.
They indicated that thermal stratification was not significant
during the heating season. This suggests that, for the building
studied, an air-conditioning distribution system that is
designed to produce intentional stratification during summer
cooling operation will not produce significant stratification
during winter heating operation. Jones and O'Sullivan l 8]
reported temperature profiles and air movement in heated
factory buildings ( 180 to 1800 m2 ( 1940 to 19 400 ft 2 ) floor
area and 6 m (20 ft) ceiling height) with various heating and
control systems. The aim of their study was to assess the
performance of various heating and control systems in terms
of heat distribution and thermal comfort. They reported meas-
ured stratification as high as 20 °C ( 36 °F) for wall mounted,
near the ceiling, unit warm air heaters. Also related is the
work of Annex 26 of the International Energy Agency which
addresses air motion, temperature distribution, and energy
use of large buildings ranging from atria to industrial halls
[9].
In this study, vertical temperature profiles were measured
in eight aircraft hangar buildings during the heating season.
The main objectives of the measurements were to determine
the extent of thermal stratification and to provide an under-
standing of the factors contributing to it. Discussed also is
the predicted impact of thermal stratification on the building's
heating energy requirements.
2. Hangar buildings
The aircraft hangar buildings (referred to here as Hangar
Nos. 1-8) are located in Ottawa, Ont. ( 4634 °C-days below
18 °C). The hangars included two main ceiling heights (9.35
m ( == 31 ft) and 17 .1 m (56 ft)), two ceiling types (flat and
semicircular (Quonset)), two types of heating systems (ver-

Table I
Description of the hangar buildings
-
Hangar No . Floor dimensions ' Ceiling height Hangar door Roof Lights Heating system
(m (ft)) (m (ft))

49 X 67 (160X223) 17 .1 (56) horizontal-sliding Quonset halogen vertical discharge

folding doors, all single- forced warm air
pane glass, tail-gate system, 3500 cfm,
120500 Btu/h
2 49X67 (160 X 220) 17.1 (56) horizontal-sliding Quonset haloge n same as Hangar No . 1
folding doors , insulated,
1 row view ports, tail-
gate h
3 49X6'/ (160 X220) 17 .1 (56 ) horizontal-slitling Quonset halogen same as Hangar No. I
folding doors, insulated,
1 row view ports, tail-
gate
4 49 X 67 (1 60 X220 ) c 17.1 (56) same as Hangar No. 3, Quonset halogen same as Hangar No . I
•. : ... .. . _ _ ....._,.. L1- - 1-- ..l
vu;; w yv1Ll) u1ui...l\.i;:;:u

5 91.7X45.I 16.2 (53) can ti levered doors, top flat (steel truss) fluorescent downdraft convective
(301x148) half all glass, bottom unit heaters. ceiling
half insulated, 1 row mounted 6 units by
view ports door, each 530100
Btu/h, 2 units in
centre, each 317800
Btu/h
6 91.7X45.l 16.2 (53) same as Hangar No. 5 flat (steel truss) fluorescent same as Hangar No. 5
(301x148)
7 34.8 X28 .8 9.35 ( 30.67) over head door Hat (timber truss) incandescent downdraft convective
(114X94 .5) unit heaters, ceiling
mounted, 3 units each
347480 Btu/h, 4900
cfmct
R 12 X42 .7 (105 X 140) 9.35 (30.67) overhead door flat (timber truss) fluorescent downdraft convective
unit heaters, ceiling
mounted, each 209440
Btu/h, 3400 cfm •

'Floor area of measurement zone.

b North-side door includes two overhead doors with automatic door opener.
c Hangar bay partitioned in working zone height.

ct Equipped with four ceiling fans in measurement zone.

c Unit heaters fitted with canvas boots and equipped with seven ceiling fans in measurement zone.
 M.N.A. Said et al. I Energy and Buildings 24 ( 1996) 105-115
~1~~r;r~1 r steam heating coil, ducted underground, and discharged
(c)
Fig. 1. View of Hangar Nos. 1-4: (a) schematic view of Hangars Nos . 1-
4; (b) view of typical hangar door, Hangar Nos . 1-4 (2 south side); (c)
view of Hangar No. 2 north side overhead doors and sectional door.
ti cal discharge forced warm air and downdraft unit heaters),
and various types and sizes of large external doors. Table 1
lists the main features of the hangar buildings.
Hangar Nos. 1-4 have large horizontal-sliding folding
doors that span the entire 49 m ( 160 ft) width in both the
north and south ends of the hangar, Fig. 1. Each door consists
of six panels. There is also an overhead tail-gate door,
2.44 X 2.44 m ( 8 X 8 ft), at the centre in each end of the
hangar. The north-end door in Hangar No. 2 includes two
overhead doors with automatic door openers for transport
vehicles, Fig. 1( c). In Hangar Nos. 1-4, office space borders
the entire length of the hangar bay area on both the east and
west ends of the hangar.
Hangar Nos. 5 and 6 sandwich a four-storey office space
in the centre, Fig. 2(b). The north end of Hangar No. 5 and
the south end of Hangar No. 6 have large cantilevered doors
(five panels each) that span the entire 91.7 m (301 ft) width.
The door panels lift up to open, then swing to the inside. The
small hangars (Hangar Nos. 7 and 8) have overhead doors
and flat roofs. The hangar bay area in these two hangars is
partitioned, Fig. 3. All the hangars have concrete floors.
2.1. Heating systems
Two main heating system types are used in the hangar
buildings tested. A vertical discharge forced warm air heating
system is used in Hangar Nos. 1-4. This system is also known
107
as the 'door pocket system' because its main function is to
prevent the door mechanisms from freezing. The system con-
sists of two floor mounted steam heating coils (mounted at
0.91 m ( 3 ft) height from the floor), one in each of the north
and south ends of the hangar. Indoor air is drawn through the
through a floor grille in front of each hangar door. The average
supply air temperature was about 41 °C ( l 06 °F). The length
of throw is about 12.5 m ( 41 ft). The heaters are each con-
trolled by a thermostat mounted on the side of each heater at
a height of l.72 m (5.67 ft) and about 6.7 m (22 ft) from
the hangar door.
Downdraft convective steam unit heaters suspended from
the ceiling are used in Hangar Nos. 5-8. The indoor air enters
the unit heater from the top of the unit where it is heated and
blown downwards in a swirl motion. The unit heaters in
Hangar No. 8 are fitted with canvas boots (a vertical canvas
drop tube 3-3.6 m (10-12 ft) in length). The boots deliver
the warm air close to the working z0ne. Some unit heaters in
Hangar No. 7 are fitted with canvas !Joots. The unit heaters
are controlled by a thermostat mounteC.: on an interior perim-
eter wall. Hangar Nos. 7 and 8 were ,'lso equipped with
propeller type de-stratification ceiling fans 'S well as exhaust
fans . Hangar No. 7 has four l.42 m diamett. · ceiling fans in
the measurement zone while Hangar No. 8 ha~ ~even ceiling
fans in the measurement zone (Fig. 3). The ce1 '11g fans are
hung 0.9 m from the ceiling truss.
The office spaces bordering the hangar bay areal. "e their
own heating (steam radiators) and ventilating syste1. . The
office spaces were usually slightly pressurised with n •pect
to the hangar bay area. The steam is supplied to all bui mgs
from a central heating plant.
3. Measurements set-up
Air temperatures were measured using copper-constantan
thermocouples type T. The thermocouples were calibrated
using a calibrated digital thermometer and a water-glycol
bath calibration facility. The uncertainty in measured tem-
peratures was estimated to be ± 0.1 °C. Vertical temperature
profiles were measured using thermocouple trees, each con-
sisting of a vertical array of 16 thermocouples. The thermo-
couples were fastened to a nylon rope and protruded
approximately 5 cm (2 inches) from the tree. The lowest
thermocouple being 0.15 m'(6 inches) from floor level and
the highest 0.15 m ( 6 inches) below the ceiling. Table 2 lists
the positioning height of the remaining thermocouples. The
0.3 m ( 1 ft) high thermocouple on the thermocouple tree
mounted close to the hangar door was used to measure the
outdoor air temper&ture.
The thermocouple trees were hung in place from the ceil-
ing. Four thermocouple trees were used in each of Hangar
Nos. 1 and 2, while three thermocouple trees were used in
Hangar Nos. 3-8. Figs. 2 and 3 show the location of the
thermocouple trees. Care was taken at each mounting location
108 M.N.A. Said et al. I Energy and Buildings 24 (1996) 105-115

y
el= E1=ig
~ faj=
00 N
...
C\I""
.,..
l_!l!a
Han door

Steam heating
coil I fen room
Foroed warm alr floor grille
El~
"' !:1
..; -
Ne...::
! El..: : e1=
.,, !
I'
~ C<O-~E.t:~ Hengarll 5 ""'. ~
!i1 ...
se ~ ~
.B
i5 Nl ~ ii~
4th
thennocouple
tree
$ 4=....,
:l) · -·
14.3m
47ft Offices
91 .7m
301 II
El:::
~~
!:; ...

Steam healing
n.5m
coll I fen room Forced wami air floor grille
48ii'"
Ntree
Hangar#6
Hangar door
.! . ,.,. ._ _ ,.
c.. -. _...,"_ ·1~
El.:::
.....
... co

~
~l.;t~O.I"'° llVlll ltel.lf\:ICll ;:,lt,J\.1~11 UUvt 1 Ill \llJ
Hangar#
s c N
1 6.9 (22.5) 19.5 (64.0) 33.5 (111 .0)
2 5.2 (17.0) 18.9 (62.0) 32.6 (107.0)
3 72 (23.5) 20.3 (66.5) 33.7 (110.5) Hangar Door
(a) 4 7.0 (23.0) 21.1 (69.3) 34.0 (111 .5) (b)
Fig. 2. Thermocouple trees locations, Hangar Nos. 1-6 (notto scale): (a) Hangar Nos. 1-4 (Hangar No. I, fan rooms in east side); (b) Hangar Nos. 5 and 6.

to ensure that the thermocouples were placed away from draft,
electrical wires and ceiling light fixtures. Considerations were
also taken in positioning the thermocouple trees in order to
accommodate normal working activities in the hangar during
the measurements. All the thermocouple leads were con-
nected to a data acquisition systen1 which auto1T1aticuHy sarn-
pled the output from all thermocouples once every minute
and stored the values on a personal cumpult::r.
In the hangar buildings that are already equipped with de-
stratification ceiling fans, Hangar Nos. 7 and 8, the measure-
ments were conducted with and without the operation of the
ceiling fans. The exhaust fans were not operating during the
measurements presented here. The pressure differential
hetween the hangar bay area and its surroundings, the out-
doors and the bordering offices, was monitored using digital
micro manometers.
4. Results and discussion
Vertical temperature profiles were measured in the eight
h ~ no~r huilrlincrc
-·--··o-· . . . . - .. ·-·-..
o~· ..A- tot~I
.. .. - .. nf
~
'? 1 tl3C'teo
"' ... -· ..
_.LO ............................... ~ .......... L ..... ~
4.1. Temperature profiles
Representative measured temperature profiles are shown
in Figs. 4-6 for the 17.1 m (56 ft) high Quonset roof hangars,
in Fig. 7 for the 16.2 m (53 ft) high fiat roof hangars, and in
Fig. 8 and Fig. 9 for the 9.35 m (31 ft) high fiat roof hangars.
Graph (a) in these figures shows the hourly variation of
various indoor air temperatures as well as outdoor air tem-
perature. Graph ( b) shows the vertical temperature profiles
at selected times. Each vertical profile represents the average
temperature over the noted hours in the graph legend.
In all hangars, except Hangar No. 2, the temperature pro-
files measured by all thermocouple trees (see Figs. 2 and 3
for locations) were quite similar in each hangar. This indi-
cates that the air temperatures at a given height were uniform
in most hangars. In Hangar No. 2, there was some lateral
variation in temperatures measured by the north thermocou-
ple tree and the south thermocouple tree. This is because,
unlike the large sectional hangar door in the south end of the
hangar, the relatively smaller overhead doors in the north end
of the hangar are opened quite frequently and remain open
for short periods. The effect of the incoming cold air does
1u"lc r-Anr111r-t,.::1.rl rl11r~"""
~ ........... o f'IAt r1::u3r-h tht"l> ciri.11th
.a.av'" .._'"' ..... _..I.I ....... ..., ._,.._,~-..1.a
O.Tu-l n.f' tho hnTll'Yn'I'" Un.Tiro,...
'°'.I.a-.._,.._ th..,. .... "_.th ,....,..,:a
1,..1.&V .&.11.4.1164..4.Lo ..1...1.'-'Ll'-'\.I' '-''"-' llV.l\.IL-'-'ll\.I

January to March. The monitoring period in 1::a1,;h lt::sl varit::d warm air cin:ulatiun fan operated more frequently than the
between 24 and 70 h. south-end fan.
 M.N.A. Said et al. I Energy and Buildings 24(1996)105-115 109

The hourly variation of indonr temperatures, on average, Table 2

Thermocouples positioning height in m (ft) from the floor
remains uniform over time, see Fig. 4(a) for example. Var-
iation in outdoor temperature appears to have little effect on Hangar Nos. 1-4 Hangar Nos . 5 and 6 Hangar Nos . 7 and 8
the hourly variation of indoor temperatures. The main inter-
ruption to the indoor temperature is during openings of the 16.92 (55.5) 16.00 (52.5) 9. 14 (30.0)
large doors of the hangar. During such events, due to stack 15.85 (52.0) 14.63 (48.0) 8.53 (28.0)
14.63 (48.0) 13.41 (44.0) 7.32 (24.0)
effect, masses of cold air enter the hangar primarily at a low
13.41 (44.0) 12.19 (40.0) 6. 10 (20.0)
level and warm air exits the hangar at a high level. The indoor 12.19 (40.0) 10.97 (36.0) 4.88 ( 16.0)
air temperature in the working zone drops sharply depending 10.97 (36.0) 9.75 (32.0) 3.66 ( 12.Q)
on outdoor temperature. The impact of the incoming cold air 9.75 (32.0) 8.53 (28.0) 3.05 (10.0)
on the air temperature near the ceiling depends on the size 8.53 (28.0) 7.32 (24.0) 2.44 (8.0)
7.32 (24.0) 6.10 (20.0) 1.83 ( 6.0)
and duration of the door opening. Once the door is closed,
6.10 (20.0) 4.88 ( 16.0) 1.22 ( 4.0)
the indoor temperature returns to its normal pattern. It is noted 4.88 (16.0) 3.66 ( 12.0) 0.30 ( 1.0)
that in Hangar No. 2, Fig. 5(a), the 0.15 m (0.5 ft) hourly 3.66 ( 12.0) 2.44 (8.0) 0. 15 (0.5)
variation of the indoor temperature shows the effect of the 2.44 (8.0) 1.83 (6.0)
frequent opening of the overhead doors (in the north end of 1.22 ( 4.0) 1.22 ( 4.0)
the hangar) mostly between 07:00 and 20:00 h. 0.30 ( 1.0) 0.30 ( 1.0)
0.15 (0.5) 0.15 (0.5)
The shape of the vertical temperature profile, in most part,
is similar in all hangars. Basically two air layers exist in the
hangar, a warm upper air layer and a cooler lower air layer. size, see graph (b) in Figs. 4-9. In Hangar No. 5, Fig. 7(b),
The lower air layer, up to 2 m (7 ft) high from the floor is the lower air layer extends to 8 m high from the floor which
characterised by a steep vertical temperature gradient, is thought to be due to a shallow air discharge angle of the
whereas in the upper air layer, above 2 m, the gradient shal- ceiling mounted unit heaters as well as excessive cold air
lows dramatically. This trend is common to most hangars infiltration from a broken view port. This suggests that in the
tested regardless of hangar ceiling height, roof type or door hangars tested, the noted parameters (ceiling height, roof type

y
fjl.1- ""-'~ ("" I 47.Bm(15711) -- • IJ
28.8 m(94.5 fl) ~ Offices I· azm ,.1
- 105ft
Olllces
Ne
@ @
0 0
E;:: Screened-in area 0 0
co..,.
c:i;;: @
Ce @ @
~
@
Overhoa< 0
door Se@
N9
0 ® ®
0 Measurement zone
Overhead
door
I!8 E= I Ce Jj .. ~
":~ @
~-
Offices Overhead
I ~
t
Overhead
door(Oui
operation 24.4m
BO I
door I I
@ Se ®
-
l
i!
:c

@
0 0
©
Offices
(a) (b)
OVerflead door 5.5 x 4.9 m (18 x 16 ft)
e Thennocouple tree
O Unltheater e Thermocouple tree O Unftheater
@ Cellingran (j) Thermostet location @CeHJngtan
Fig. 3. Thermocouple trees locations, Hangar Nos. 7 (a) and 8 (b).
110 M.N.A . Saia et al. I Energy and Buildings 24 (1996) 105-115

- 14.6m·.. --B.5m ···0.15m -·outdoor

30.-----.---.---.--.----..----.--~-.,..----..--~--.----.

- 14,6 m ... · 8.5 m --- 0.15 m -- Outdoor

30 . . 20~
14.6 m --l--.. : u
20~-.,~~cri: ~ 10
E
u
E

I T l
g> 10

. ---
! :: }--.L-- -iJ;J]£; --'----':.±1
~
.~ "f:~---"-----r----
~
-10
u

l !
-20 ,
:
:-
] ; ; i ! l ! !
·---i-··-r·-- .
~ !
:
i
J
·--·"·----:----..--·-·r-----·----:-----~----"·--;.
1
!
j
:
f
:
I
l

----~----- r·· ·- 1- ----t·· ·- 1 -···i····-t···-1--· · -

i.
l
:

i
:
;

;.
:
:

:
:

:.

-", (•) i ----~----, ' ·t

. (a) ! i : i i
-30 . . •

:----:--1--+-:----r--- 0 4 8 12
'Time(hr)
16 20 24

4 8 12 16 20 24
Time(hr) --- 00:00-04:00 hr ··•" 10:30-12:15 hr-• 21:00-24:00 hr
20 .---------------------~

-- 02:00-07:00 hr -'r-11:30-14:00 hr-• 20:00-24:00 hr

20 .------------------ -----. 16

16 §: 12
I
§: 12 ~ 8
I
l 8 4
(b)
4 0I • • __.,,r:r:=-. I I •

(b) 12 14 16 18 20 22 24 26 28
0 I I ! • . . . ,. . - -....-- I I I I Temperature (deg C)
iu l.l iO 10 Lli LL .l4 LO
'" Temperature (deg C)
Fig. 6. Temperature profiles, Hangar No. 3, north 'N' thermocouple tree.
(a) Hourly variation of air temperatures at height 14.6, 8.5 and 0.15 m ( 48,
Fig. 4. Temperature profiles, Hangar No. 1, north 'N' thermocouple tree. 28 and 0.5 ft) from floor level. (b) Vertical temperature profiles at various
(a) Hourly variation of air temperatures at height 14.6, 8.5 and 0.15 m (48, hours of the day.
28 and 0.5 ft) from floor level. (b) Vertical temperature profiles at various
hours of the day. - H.6 m ·-· • 6Ji m • • • 0, 16 m - • autd0ot
30 .----.---.---~-.----.--~-.----..--~-~-.,..---.

20
u
-14.6 m -- - 8.5 m ·-- 0.15 m -- Outdoor
!
jot±+ -:-·rt
30.----.---.---.--.,..----.---.---.--.----.---.---.---. 10
14.6

20 b=:;:t...mv~-
: ~ ---·i>=»-~~ ti -~
....,,,,,.j..~~
--- •--· ··'c _ ,. ., __,,......
u & ' . . ..
gt 10 .~ -10 !· - ---~·-··
E
~ o r--;---;----,;--_____,,_::.:~:.::...._.+--+---i----i---+--+-J -20 : : ;__ - - ---t-- --r----;-----r---- 1----r--· t·---1----
~ -30
\U) :
.
: :
•
:
•
:
.
:
•
:
.
:
•
: ;
'
~ -10 0 4 8 12 16 20 24
~l~ · ·,, f 1 nme(hr)
-
20
-30
-(~;r··r-·T ·--r-·r· ·
. . .
-----1----
:
-·-- :---- -~-- --
: .
----- 20 .--------------------~

--- 00:00-08:00 hr ... · 13:00-16:00 hr·• 21:0~24:00 hr

0 4 8 12 16 20 24
nme(hr) 16

--- 04:00-07:00 hr .... 08:00-12:00 hr-• 12:00-16:00 hr -11:- 16:00-20:00 hr §: 12

20 I
~ 8
16 ~
.r? -~

di
4

i
~
12
8
t o'
10
(b)
I

12 14
•......c:iii2•)!t'C
16 18 20
I I

22
r !

24
I

26

4 f
(b) -~-- YJ Fig. 7. Temperature profiles, Hangar No. 5, centre 'C' thermocouple tree.
(a) Hourly variation of air temperatures at height 14.6, 8.5 and 0.15 m ( 48,
Temperature (deg C)

o L......
28 and 0.5 ft) from floor level. (b) Vertical temperature profiles at various
10 12 14 16 18 20 22 24 26
Temperature (deg C) hours of the day.
Fig. 5. Temperature profiles, Hangar No. 2, north 'N' thermocouple tree. ......- ...l ..... ..-..- ,..,!~,.. \ ...l!...l .......... 11- t... ............ ,.... .... !.-.... !.t::: ............. ,. !-£:1 ....... - .......... ~- .a.L ..... - L - - -
VJ. UVVl ,;)1£..V} UJU J.IVI. Ila vv a .:).l,511111 ..... CUll. JlJUU'-'llV\.I v u Ul~ .:)llal'v
(a) Hourly variation ofairtemperatures at height 14.6, 8.5 and 0.15 m (48,
28 and 0.5 ft) from floor level. (b) Vertical temperature profiles at various of the floor-to-ceiling temperature profile. However these
hours of the day. parameters, as discussed later, might have an influence on the
 M.N.A. Said et al. I Energy and Buildings 24 ( 1996) 105-115 111

- B,5m - · • .Sm -·· D.15m - - outdoor

30.----,-~~~~~--:,----.-~.,.--.-~-,----:~--,---.
ceiling fans on

w I . ' , j, ..c:r ; ...r ......i ·r'.•~.:.~f "·.(. ,. ;--r-· J 20

: :

G G
go 10 ~ 10
:e :e
i 0
i~ 0

~ -10 -10
~ ~
-20 -20

-30
4 8 12 16 20 24 0 4 8 12 16 20 24
Time(hr] Time(hr]

20 20.-~~~~~~~~~~~~~~~~~~-,

- . Fans off (00:00-05:00 hr) · "•'· Fans off (11:30-12:30 hr]

16 16 ~ ·• Fens on (16:00-18:00 hr) -E· Fens on (19:00-24:00 hr)

I
I:
"'
·o;
:I:
12

4
I 12
I:
"' 8
~

4
I .
..'
,,,. .../"
.

.J' ,.l
o~
o' (b)
_...-:-.£'~
I

12 14 16 18 20 22 24 26
Tempereture (deg CJ
Fig. 8. Temperature profiles, Hangar No. 7, south 'S' thermocouple tree, de-
stratification fans off, week-end conditions. (a) Hourly variation of air
temperatures at height 8.5, 4.9 and 0.15 m (28, 16 and 0.5 ft) from floor
level. (b) Vertical temperature profile.
temperature stratification level. The operation of the de-strat-
ification ceiling fans in Hangars Nos. 7 and 8 appears to have
little effect on the shape of the vertical temperature profile,
Fig. 9 ( b). It is noted that the shape of the vertical temperature
profile is similar to that measured by Bagheri and Gorton [7]
in a nuclear reactor facility.
The air temperature in the upper air layer is almost uniform
in Hangar No. 1 (Fig. 4(b)) where the hangar doors are
mostly glass, and also in Hangar No. 5 (Fig. 7 (b)) where
the top half of the door is all glass. This is because of the low
R value of the glass which dissipates heat in the upper air
layer. In contrast, a higher thermal insulation R value traps
heat in and tends to maintain stratification.
The frequent opening of the overhead doors in the north
end of Hangar No. 2 caused the north-end warm air circula-
tion fan to operate frequently during working hours (Fig.
S(a), 08:00 to 20:00 h). As a result, the air temperatures in
the upper air layer, and hence stratification, steadily increased
during working hours .
4.2. Thermal stratification
Table 3 lists measured stratification levels (expressed by
floor-to-ceiling temperature differential, °C) as well as ver-
tical temperature gradients, °C/m. The values are averaged
over each test period excluding the periods in which the
hangar door was opened. The stratification varied from 4 °C
( 7 .2 °F) to 11 °C ( 20 °F) . The average vertical temperature
gradient varied between 0.8 and 2.6 °C/m (0.4 to 1.4 °F/ft)
in the lower air layer (working zone), up to 2 m (7 ft) high
10 I ' r I -'l'i" I • I
28 12 14 16 18 20 22 24 26
Temperature (deg C)
Fig. 9. Temperature profiles, Hangar No. 8, south 'S' thermocouple tree,
effect of de-stratification ceiling fans operation. (a) Hourly variation of air
temperatures at height 8.5, 4.9 and 0.15 m (28, 16 and 0.5 ft) from floor
level. (b) Vertical temperature profiles, effect of de-stratification ceiling
fans.
from the floor. In the upper layer, above 2 m, the vertical
temperature gradient is about 0.5 °C/m (0.3 °F/ft) in all
hangars. It is noted that several unit heaters in Hangar No. 6
were out of service during the measurements. This would
result in a higher than usual stratification level because the
downward throw of these unit heaters helps in distributing
the warm air down to lower levels.
4.2.1. Outdoor temperature effect
Outdoor temperature variation, with no openings of the
hangar door, seems to have little effect on stratification,
Fig. 10( a). The main interruption to the stratification is dur-
ing opening of the large doors at either end of the hangar,
Fig. lO(b). During such events, the air temperature near the
floor drops sharply with much less disturbance near the ceil-
ing, sending the floor-to-ceiling temperature differential up
to 30 °C ( 54 °F) depending on outdoor temperature. Once
the door is closed the stratification quickly returns to its nor-
mal uniform pattern.
4.2.2. Ceiling fans effect
The de-stratification ceiling fans used in Hangar Nos. 7
and 8 appear to have little effect on stratification in these two
hangars, Fig. 9. This could be due to the fact that these two
hangars had low stratification to start with ( 4 °C). The unit
heaters which are fitted with canvas boots appear to have the
dominant effect on reducing stratification in these two han-
gars because the canvas boots distribute the wann air directly
to the working zone.
112 M.N.A. Said et al. I Energy and Buildings 24 ( 1996) 105-115

Table 3
Stratification (floor-to-ceiling temperature difference ( °C)) and vertical temperature gradient (°C/ m)

Hangar Hangar Hangar Hangar Hangar Hangar Hangar Hangar

No . I No . 2 No. 3 No. 4 No . 5 No . 6 No. 7 No. 8

Average stratification
Normal (no door 4 8 9 6 9 11 4 4
opening)
Door open 15 25 25 20 25 28 15 13
Average vertical
temperature gradient
height < 1.8 m 0.8 1.5 2.3 1.62 1.1 2.6 1.1 I
height > 1.8 m 0.3 0.5 0.5 0.3 0.5 0.5 0.4 0.4
Notes full glass insulated doors, many door insulated insulated doors, some unit ceiling ceiling
doors openings doors a broken view heaters out of fans off fans on
port service

4.2.3. Ceiling height effect assumed to happen mostly from the top zone, zone 3 ( 15 000
The ceiling heights (9 .35, 16.2 and 17 .1 m ( 31, 53 and 56 II s or 7500 cfm). The middle zone, zone 2, was assumed to
ft)) studied do not appear to be a contributing factor to have moderate exfiltration (5000 l/s or 2500 cfm) and little
thermal stratification. As can be seen from Figs. 4-9, most of infiltration (800 1/s or 400 cfm). The air infiltration and
the vertical temperature change occurs in the working zone. exfiltration for each zone were based on the apparent gaps
The ceiling type (flat or Quonset) also does not appear to be underneath the doors, between door sections, and around the
a factor that fosters stratification. Thergfore, to estimate the tail-gate door at the top. A multizone computer program [ 11]
tALtllL u[ d1t1111ai sua1iilcaLiu11 in a iarge space high ceiiing was also used to provide guidance in the distribution of the
building, it is sufficient to measure the temperature at about air infiltration and exfiltration. The airflow between the zones
0.15 m (0.5 ft) and 3 m (or 10 ft) height from the floor in
was set to maintain conservation of mass in each zone. The
several locations in the building. Above 3 m height, a 0.5 °C
air flowed from zone 1 to zone 2 to zone 3. R values of the
per meter height (0.3 °F per foot height) can be assumed.
walls and the roof of the hangars were estimated according
to construction details. The following are the R values used
4.2.4. Heating energy implications in the calculations for Hangar No. 1.
Hangar Nos. 1 and 2 were used to assess the effect of
thermal stratification on the building's heating energy requi-
rements because these two buildings were used for additional 30 . . . . . ' . .
t i ! ! : ! ! l :
measurements that included the air change rate per hour and _ 20 r-- :~--.- +----f----+---+--+--~+- ·+···+···· ______ ___
air flow patterns. The measured average stratification, T::ihle ~
&
10 ____ J.. .. .L~~~~~~~~-:-±. :-L ...JL.J....._.ull...J .............
L..l........~l... _J__.._ __
3, is 4 °C in Hangar No. 1 and 8 °C in Hangar No. 2. The O I

t=: 0 : : : ! : : : • ; I • I

Building Loads Analysis and System Thermodynamics ~ • I :11 I : : j, ,I • • • •

(BLAST) computer program r101 was used to calculate the ~
E -10 !~:
·-··~····-;-·· ! i : '.~ .~)~,;
· . • , -~-·--·:··~•J,...,t[i:'!.:. .<,-1, .•. ..;i..... iL•••• ;i••• _.
building's heating energy requirements. The calculations
.,
1--
·· I
: i: :: . : i. /j! .l>., ......I ,,~. !: ;:
. ~/R"°"
-~ ··- · . t . ... ~· . '.l . . • ! . . '.,,ji(., • " .
were conducted on a 486-50 MHz desktop computer.
-20
~ • : :- 1 outdoor temperature :· • •• • ~1+t>:-t~"!·

For ease of simulation the Quonset roof was simulated to -30

(a) i' • :. . :; ! : : . . .
': i .
!: . !': ·;'
0 4 8 12 16 20 24
be flat at an equivalent height of 14.6 m or 48 ft (same nme(hr)
building air volume). To simulate the measured vertical tem-
perature distribution, the hangar was subdivided into 3 ver- 30 '

tical zones: 0-1.8, 1.8-8.5 and 8.5-14.6 m (or 0--6, 6-28 and
"'s~c~a~···t.
20
28-48 ft). These zones were chosen based on the distinct u .. ... ... ...
'
g> 10 ·-~ - ---- :-.- - i-· · :
zones noted in the measured vertical temperature profile in E. -- ! I
-
'
-
, :

[ j l;.::;:.~.1··--·~--·
the hangar. 0
For Hangar No. 1, the air temperatures in zones 1, 2 and 3 i~-10 i ; i i··-··:····~-----, ... L... L. .. L. __
--·ovtdoonempereture _..,.f-..

~(~;·t,·-·~+--1~"'·'f:::f=T-··-f···-1··-i····-~----1---·1
were set at 17, 20.4 and 21 °C ( 62.5, 69 and 70 °F), respec-
tively. Within each zone, the air temperature was assumed to 1-- -20
be uniform. The air infiltration rate for Hangar No. 1 (20 000 -30 . • .

l/s or 10 000 cfm) was calculated based on the measured air 0 4 8 12 16 20 24

nme(hr)
change rate (using decay of sulfur hexafluoride, SF"). It was l\~;.
iG. 1-iumly vaiiai.ivu ul siiuiiJit;uiiuu (iluur-Lo-ceiiing t~mperacure Liif-
simnl::itecl to mostly infiltrate into the building in the bottom f~rence),effect of outdoor temperature: (a) Hangar No. 7 (no hangar door
zone, zone 1 ( 19 200 I Is or 9600 cfm) . Exfiltrati on was openings); (b) HangarNo.4.
 M.N.A. Said et al. I Energy and Buildings 24 (1996) 105-115 113

Hangar door overall R value 0.3 m2 °C/W while Hangar No. 1 doors are mostly single-pane glass (R
(north and south walls) ( 1.6 h ft2 °F/Btu) value = 0.3 m2 °C/W), Hangar No. 2 doors are fully insu-
Brick walls to bordering 0.7 m 2 °C/W lated (R value = 1.56 m2 °C/W or 8.85 h ft 2 °F/Btu) except
offices (3.8 h ft 2 °F/Btu) for a row of view ports. The air temperatures in zones 1, 2
(east and west walls) and 3 were set at 17, 23.8 and 25 °C (62.5, 75, and 77 °F),
Floor (concrete slab on 0.24 m 2 °C/W respectively. The air infiltration rate for Hangar No. 2 ( 24 000
grade) ( 1.4 h ft 2 °F/Btu) I/ s or 12 000 cfm) was calculated based on the measured air
Roof 1.4 m2 °C/W change rate per hour. The heating energy requirement for
(8 h ft 2 °F/Btu) Hangar No. 2 is calculated to be an average of 405 kWh per
h over the three month period January to March.
To assess the impact of stratification on heating energy,
The hourly heating energy requirement was simulated over calculations were conducted using temperature differences
a three month period, from 1 Jan. to 31 Mar., using 1976 from 0 (no stratification) to 13 °C between zones 1 and 3.
hourly weather data for Ottawa as measured by the national The air temperature in zone 1 was set at 17 °C ( 62.5 °F)
weather service. This period of time coincides with the meas- assuming that this temperature is fixed by the thermostat
urement period in the hangars. The heating energy require- setting. The temperature in zone 3 was varied from 17 to 30
ment for Hangar No. 1 is calculated to be an average of 396 °C in increments of 1 °C. Based on the trend noted in measured
kWh per h over the three month period January to March. vertical temperature profiles, the temperature in the inter-
This value is the same order of magnitude as the energy mediate zone 2 was set to closely follow the temperature of
consumption monitoring results by the National Research zone 3. Table 4 lists the temperatures used in zones 2 and 3.
Council of Canada (NRC) for a similar hangar near Hangar Table 5 lists the impact of stratification on the building's
No. I. The monitored energy usage ofNRC's hangar is esti- heating energy requirements. Fig. 11 shows the results in a
mated to be an average of 590 kWh per h for the months Nov. graphical form. The estimated heating energy requirement
1993 to Mar. 1994. This indicates that the estimated heating for Hangar No. 1 is represented at !l. T of 4 °C (396 kWh per
energy requirements as well as the assumptions made in the h) which amounts to about 20% in excess heating energy
simulation are realistic. It is noted that the calculated heating requirements as compared to the case with no stratification
energy requirement for Hangar No. 1 (396 kWh per h) does ( !l. T = 0 °C). For every 1 °C temperature stratification, there
not account for the effect of the hangar door opening events. is an average (over the three month period January to March)
Thus the calculated heating energy requirements would be of 15.6 kWh per h (or 4.8%) associated increase in heating
underestimated as the occasional inrush of cold air when the energy requirements in Hangar No. I. For Hangar No. 2, the
hangar door is opened and the subsequent recovery of indoor excess heating energy requirement due to the 8 °C tempera-
air temperature are not accounted for in the calculations. ture stratification is 38.4% as compared to the case with no
Hangar No. 2 was simulated in a similar fashion. The main stratification. For every 1 °C temperature stratification in Han-
difference between Hangar No. 2 and Hangar No. 1 (that is gar No. 2, there is an average of 13.9 kWh per h associated
pertinent to estimating heating energy requirements) is that increase in heating energy requirements. This indicates that
a well-insulated building can. support a greater degree of
stratification than a poorly-insulated building for the same
Table4 energy budget.
Air temperature in zones 2 and 3 used in estimating heating energy (zone I The decision as to when corrective de-stratification meas-
temperature= 17 °C (62.5 °F))
ures are warranted depends on the pay-back criteria and the
Stratification (°C) Zone temperature (°C) (°F) suitability of the de-stratification device to the building. For
ti.T(zone 1 tozone3) discussion purposes, if a suitable de-stratification device
Zone2 Zone3 reduced the stratification in Hangar No. 2 by say 2 °C (i.e.
from 8 to 6 °C), the building's heating requirements would
0 17.0 (62.5) 17.0 (62.5)
1 17.8 (64.0) 18.0 ( 64.4)
be reduced by about 27 kWh per h or 7%. This amounts to
2 18.7 (65.7) 19.0 (66.2) 90 720 kWh ( 310 million Btu or 310 000 cubic feet of natural
3 19.5 (67.0) 20.0 (68.0) gas) over a five month heating season, less the operating
4 20.4 (69.0) 21.0 (70.0) power consumption of the de-stratification device which is
5 21.2 (70.2) 22.0 (71.6) relatively small. The savings in energy budget will be higher
6 22.I (71.8) 23.0 (73.4)
24.0 (75.2)
for a greater reduction in stratification and in locations with
7 22.9 (73.2)
8 23.8 (74.8) 25.0 (77.Q) higher associated energy cost.
9 24.6 (76.3) 26.0 (78.8)
10 25.5 (78.0) 27.0 (80.6)
28.0 (82.4)
5. Conclusions
11 26.3 (79.3)
12 27.2 (81.0) 29.0 (84.2)
The study provides guidance to the extent and how to
13 28.0 (82.4) 30.0 (86.0)
assess thermal stratification in aircraft hangar buildings dur-
114 M.N.A. Saia et al. I Energy and Buildings 24 ( 1996) 105-115

Table 5
Effect of temperature stratification on heating energy requirements Q (kWh per h)

Stratification (°C) Hangar No. 1 Hangar No. 2

t:,.1· (zone J to zone~) Q IOO(Q- Qo) !Qo Q 100( Q- Q0 ) /Qo

0 327.54 (Q0 ) 0.0 292.67 (Q0 ) 0.0

346.99 5.9 309.84 5.9
2 371.14 13.3 324.34 10.8
3 383.02 16.9 344.06 17 .6
4 395.77 20.8 354.53 21.1
5 409.19 24.9 365.13 24.8
6 422.19 28.9 377.74 29.1
7 437.30 33.5 391.16 33.7
8 452.88 38.3 405.12 38.4
9 467.99 42.9 418.40 43.0
10 483.57 47,6 432.23 47.7
11 498.37 52.2 445.51 52.2
12 514.75 57.2 459.60 57.0
13 530.45 61.9 473.96 61.9

550
Thermal stratification can have a significant impact on the

500 :;-~~~rNo.~.J_~---
W1! -~- · ~-- building's heating energy requirements. In the hangar build-
ings studied, excess heating energy requirements due to 8 °C
~ .....
~
•r~v
!
I
I
..
j I :
I
i ~I __JI
.. floor-to-ceiling temperature differential can be as much as
! 38% as compared to the case with no stratification. In average
~ I !
~ 400
UJ
-------1---·---·--
! ! - this amounts to 4.8% excess heating energy requirements for
each 1 °C floor-to-ceiling temperature differential. This does
~~1fl-L+-1-~
"'
c:
·ai
-r i l·-·r-+----
350 not account for the effect of hangar door opening events.
(I)
I
300 •.
i I 1
--+-·L.... ,_ .
I I • Assessing temperature stratification and exploring the pos-
I I : : ! sibilities for reducing it, can be a profitable activity.
250
l ! i : j

0 2 4 6 8 10 12 14
Temperature Stratification (deg CJ
Acknowledgements
Fig. 11. Effect of temperature stratification on heating energy.

This work was partially supported by the Department of

ing the heating season. Measured thermal stratification,
expressed by floor-to-ceiling temperature differential, varied
from 4 °C (7.2 °F) to 11 °C (20 °F) in the buildings tested.
Two air layers existed in the hangar bay area, a warm upper
air layer and a cooler lower air layer. The lower air layer, up
to 2 m or 7 ft high from the floor, is characterised by a steep
vertical temperature gradient in therange0.8-2.6 °C/m (0.4-
1.4 °F/ft), whereas in the upper air layer a 0.5 °C per meter
height (0.3 °F per foot height) can be assumed.
For the ceiling heights (9 .35, 16.2 and 17.1m(31,53 and
56 ft)) studied, the ceiling height did not appear to be a
contributing factor to thermal stratification. For the three ceil-
ing heights, most of the vertical temperature change occurred
in the working zone from the floor to 2 m height. To estimate
the extent of thermal stratification in an aircraft hangar build-
ing, it is sufficient to measure the temperature at about 0. 15
m (0.5 ft) and 3 m ( 10 ft) height from the floor in several
locations in the building. Above that, add 0.5 °C per meter
height (or 0.3 °F per foot height). Stratification also appears
to be independent of the ceiling types studied (flat or
Quonset).
the National Defence of Cunudu. The authors gratefully
acknowledge the assistance and cooperation of the members
of the Department of National Defence during the site mon-
itoring. The authors acknowledge the contribution of Mr L.M.
McSheffrey in the field measurements. The authors are also
thankful to Dr S.A. Barakat for his suggestions and discussion
during the preparation of this paper.
References
[ l] R.H. Dean, A.K. Gupta and R.M. Van Becelaere, The effect of thermal
stratification on factory air-conditioning load, ASHRAE Trans., 82
(1976) 584--599.
[2] H.M. Bagheri and R.L. Gorton, Verification of stratified air-
conditioning design, ASHRAE Trans., 93 ( 1987) 211-227.
[3] R.A. Beier and R.L. Gorton, Thermal stratification in factories -
cooling loads and temperature profiles, ASHRAE Trans., 84 (1978)
325-339.
[ 4] J.B. Olivieri and T. Singh, The effect of supply and return outlets on
stratification: Part 2, ASHRAE Trans., 88 (1982) 45--61.
[5j R.W. Boucher, G.R. Baughman and L.B . Driggers, Temperature
stratification in broiler houses and the effects of ceiling fans, Am. Soc.
Agric. Eng., Appl. Eng. Agric., 4 ( 1) (1988) 66-71.
 M.N.A. Said et al. I Energy and Buildings 24 ( 1996) 105-115
[6] M.G. Hejazi-Hashemi, Thermal behaviour and energy performance of
atrium buildings, in P. Novak, B. Todorovic and E. Kulic (eds.), Proc.
Clima 2000 Second World Congr. Heating Ventilating, and Air
Conditioning, Sarajevo, Yugoslavia, 27 Aug.-27 Sept. 1989, Vol. 4,
pp. 17-23.
[ 7] H.M. Bagheri and R.L. Gorton, Performance characteristics of a system
designed for stratified cooling operating during the heating season,
ASHRAE Trans., 93 ( 1987) 367-381.
[8] P.J. Jones and P.E. O'Sullivan, Air movement and heat distribution in
heated factory spaces, A. Faist et al . (eds.), Proc. 3rd Int. Congr.
Building Energy Management, JCBEM '87, Switzerland, Oct. 1987,
Vol. 3, pp. 224-231.
115
[9] IBA Annex 26, Energy-Efficient Ventilation of Large Enclosures, The
Energy Conservation in Buildings and Community Systems,
International Energy Agency, Operating Agent: A. Moser, ETH-
:lentrum, LOW E2, Zurich, Switzerland .
[ 10] BLAST, The Building Loads Analysis and System Thermodynamics
Program, Version 3.0, Level 193, BLAST Support Office, Department
of Mechanical and Industrial Engineering, University of Illinois at
Urbana-Champaign, 1991.
[ 11] M.N.A. Said, NRC Smoke Control Model 2.0, User Manual, Rep. No.
652, National Research Council Canada, Institute for Research in
Construction, Oct. 1993.