PLEA2009 - 26th Conference on Passive and Low Energy Architecture, Quebec City, Canada, 22-24 June 2009
Displacement Ventilation and Passive Cooling Strategies
PAUL CAREW1, BERNARD BEKKER1
1
PJCarew Consulting, Cape Town, South Africa
ABSTRACT: The publication of the ASHRAE “System Performance Evaluation and Design Guidelines for Displacement
Ventilation” [0] has contributed to the wider acceptance of displacement ventilation (DV) as a ventilation strategy, by
offering clear guidelines from an established organisation. A significant advantage of DVis that it lowers the supply air
quantities required for cooling compared to conventional mixing ventilation (MV) at the same supply air temperature.This
opens up opportunities for the use of passive (non refrigeration-cycle based) cooling sources, which typically are limited
in supply air temperature and based on 100% fresh air supply when compared to conventional refrigeration-cycle based
sources. This paper quantifies the impacts of using DVin comparison to MV on the peak capacity, size and humidity levels
associated with the following passive cooling sources: evaporative cooling, two stage evaporative cooling, thermal stores
and air to ground heat exchangers. Ageneric office building in Johannesburg, South Africa, is used as a model. The paper
illustratesthe extent to which the use of DVexpands the ability of passive cooling strategies to serve spaces previously
considered as having too high a heat load (when calculated using MV system guidelines). The paper however also
recognises that passivecooling strategies are unlikely to be widely implemented until design guidelines exist from
organizations similar to ASHRAE.
Keywords: passive design strategies, displacement ventilation
INTRODUCTION
Based on the authors’ experiences as consultants in the
built environment, conventional Heating Ventilation and
Air Conditioning (HVAC) engineers keep to using clear
guidelines and calculations methods provided by
established organisations. It is unlikely that
unconventional strategies and systems will be
incorporated in a project in the absence of such
guidelines or methods, unless it is a showcase project
with sufficient resources to adequately investigate
alternatives.
Displacement ventilation (DV) can be seen as an
example of this. While popular in Scandinavia [0], the
American Society for Heating, Refrigeration and Airconditioning Engineers (ASHRAE) only recently
released “System Performance Evaluation and Design
Guidelines for Displacement Ventilation” [0], which
provides generic DV sizing and design guidelines. This
publication differs from existing supplier-specific DV
design guidelines in that itprovides a simple ten-step
guide to design the DV system [0]. While the authors
have not found literature that documents the worldwide
growth in the number of installed DV systems, since the
ASHRAE publication was released they have found
HVAC engineers in general to be more receptive to the
use of DV for conventional projects.
The aim of this paper is to illustrate the impacts of
using DV in comparison tomixing ventilation (MV) on
the peak capacity, size and humidity levels associated
with various passive cooling strategies, informed by the
ASHRAE DV guidelines. The intention is NOT to
provide clear design guidelines for these passive
strategies, but merely to demonstrate the implications of
these strategies through quantitative analysis.
Through this analysis it is demonstrated that a
number of passive cooling strategies, considered
impractical for MV applications, can potentially be of
benefit in building designs utilising DV.The comparison
also highlights the risk of over-design when conventional
MV guidelines are used to design cooling sources for DV
applications.
DISPLACEMENT VS. MIXED VENTILATION
DV introduces air at low level and low velocity, and at
high supply air temperature (Ts),typically around
18°C,when compared to conventional air conditioning
systems which use MV. The air that is slightly cooler
than the intended room temperature runs along the floor
until it reaches a heat load (Fig. 1). The heat load induces
a plume of warmer air that rises due to lower density.
This induces stratification in room temperature with the
occupied area of the room within comfort conditions and
the space near the ceiling at higher temperature
conditions. The air near the ceiling is continually
exhausted to prevent a build up of warm air into the
occupied zone.
PLEA2009 - 26th
2
Conference onn Passive and Low
w Energy Architecture, Quebec City, Canada, 22-24 Juune 2009
HOTTER
ED
UNOCCUPIE
ZONE
OCCUPIED
D
ZONE
t occupants, low level ligghts and
as the loads from the
ment (kW), Ql as the load froom the overheaad lights
equipm
(kW) and
a Qex as the envelope loadss with aoe, al, aex as the
fractioons of the reespective loadds that occur in the
occupiied zone(0.2955, 0.132 and 0.185 in [1]). The
resultinng supply air flow
f
rate is calcculated as
Vh =
Qdis
ΔThf ρC p
…(3)
withΔT
Thf as the tempperature differeence between head
h
and
feet annd set at 2°C.
Figure 1: Dispplacement ventillation with indiccative temperatuure
gradient relatiive to height
When usiing MV the co
ool air - typicaally around 12 to
14°C - is inttroduced at hiigher velocitiees and at a higgh
level, inducinng room air to mix with it as shown in Fig. 2.
There shouldd be a relatively
y consistent temperature in thhe
room.
Exhaust air teemperatureW
When using DV
D
the
exhausst air temperaature (Te) is inn general highher than
when using MV. Inn the case of MV
M the Te shhould be
close to the desiggn room tem
mperature, whhile the
ASHR
RAE DV guidelines supply thhe following equations
to calcculate Te for DV
V systems:
Te = Ts +
Qt
ρC pV
where
WELL-MIXE
ED
AIR
Ts = Th − ΔThf −
… (4)
θ f Qt
ρC p V
… (5)
from
θf =
1
Vρ C p
A
Figure 2: Mixing
M
ventilatio
on with indicaative temperatuure
difference relaative to height sh
hown as constannt
Supply air flow ratee calculationssA fundamenttal
difference inn the calculatio
on of the suppply air flow raate
between DV
V and MV beco
omes apparent in the ASHRA
AE
DV guidelinees. Assuming no
n latent coolinng, with MV thhe
rate is calculated according
g to the entire room
r
load withh
Qt
V=
ρC p (Th − Ts )
…(1)
3
where V is the supply airr rate (m /s), Qt is total rooom
d
(kgdryaiir/m3), Cp is thhe
load (kW), ρ is the air density
specific heaat capacity off air (kJ/kgdryaair), Th is rooom
design tempeerature (°C) an
nd Ts is supplyy air temperatuure
(°C.)
Howeverr with DV the flow rate is
i based on thhe
portion of the
t total room
m load that iss present in thhe
occupied zonne using the folllowing breakddown [0]:
Qdis = aoe Qoe + al Ql + aex Qex
(
1
αr
+
1
α cf
… (6)
) +1
whereT
Tf is the flooor temperatuure (°C), Th is the
temperrature at the heead or room deesign temperatuure (°C),
θf is thhe dimensionlless temperatuure, αr is the radiative
r
heat transfer coefficient of the floorr (W/°C m2) annd αcfthe
convecctive heat trannsfer coefficiennt of the flooor (W/°C
m2), booth put forwardd as 5 by ASHRAE [0].
BACK
KGROUND TO
O MODEL
For thee purpose of thhis paper a genneric 100m2 oppen plan
office space in Johannnesburg, Souuth Africa, is used
u
as a
model onto which different venntilation and passive
coolingg strategies cann be applied. The
T peak coolinng loads
within this office spaace areshown inn Table 1.
Table 1:Peak
1
cooling loads of the 100m2 space (7W/m2 for
2
people, 13W/m for equipment and 10 W/m
W 2 for lightingg)
Occuupants and equippment
Qoe
2 kW
k
Overrhead lighting
Ql
1 kW
k
Exteernal loads
Qex
4 kW
k
Totaal loads
Qt (ffor MV)
7 kW
k
Loadds with occupiedd zone
Qdiss (for DV)
1.4462kW
…(2)
withQdis as the
t load betweeen the head and
a the feet off a
sedentary occcupant being served
s
by the DV system, Qoe
Thrree ventilationn systems will now be consiidered: a
DV syystem, as weell as two MV
M systems, one
o
that
PLEA2009 - 26th
2
Conference onn Passive and Low
w Energy Architecture, Quebec City, Canada, 22-24 Juune 2009
supplies air at the same high temperatture as the DV
D
system (MV
Vht) and the otther at a mucch lower suppply
temperature (MVlt). Using Eq. 1 to 3, thee required suppply
air flow rate into the room
m for each of thhese systems caan
now be calcuulated, as show
wn in Table 2.
Table 2: Resulting air flow fro
om MV (low andd high Ts) and DV
D
systems with % relative to lo
ow (and high teemperature mixinng
ventilation (M
MVlt and MVht)
Mode
Ts (°C)
Th/Te (°C)
V (m3/s)
% ofMV
Vlt
% ofMV
Vht
MV
Vlt
14
26
0.59
9
100%
37%
%
MVht
22
26
1.61
273%
100%
DV
22
26/31
0.74
125%
46%
From Tabble 2 it is clearr that the air flow rate requireed
for DVislesss than half th
hat required for MVht, eveen
though both ventilation sy
ystems have the same suppply
temperature. The airflow rates for DV
V and MVlt are
a
similar, but MV
M lttypically require a mucch larger coolinng
source than DV
D in order to obtain the low
w Ts.
u
the
Evvaporative coooling Evaporaative cooling uses
changee of phase of
o the water contained in an air
stream
m/volume to cool down the aiir. Energy from
m the air
and noon-evaporatingg water is usedd to supply thhe latent
energyy required. This results in a teemperature droop in the
air andd residual watter, and an inccrease in absolute and
relative humidity of the
t air.
The minimum temperature achievable through
evaporrative coolingg is expresseed by the wet
w bulb
temperrature (Twb) off the air. In reallity the air cann only be
cooledd a part of the way between the dry bulb (Tdb) and
wet bulb
b
temperatture and thiss is known as the
evaporrative cooling efficiency
e
(hevaap), assumed ass 80%.
Ts = Tdb − η evap (Tdb
d − Twb )
Figuree 4 provides thhe probability of Ts throughhout the
year baased on eq. 7 and
a on annual climatic
c
data [44].
%
The compparison in Tab
ble 2 highlightss the efficienciies
introduced by
b using DV in
nstead of MV.The rest of thhis
paper will explore
e
how th
hese efficienciies impact on a
selection off passive coo
oling strategies.A number of
passive desiggn strategies arre potentially applicable
a
in thhe
Johannesburgg climate baseed on Fig. 3, which
w
indicatess a
wide diurnaal temperaturee range and low wet buulb
temperature for much of th
he year.This ciity was therefoore
chosen as thee context for th
he simulations presented in thhis
paper. Typical office hourss were assumedd.
dry bulb temp
…(7)
2-stage
e
evaporative
cooling
single stag
ge
evaporativ
ve
cooling
Ts
Figure 4: Cumulative distributed freqquency of Ts forr 2-stage
evaporaative cooling (1
( st line) and evaporative
e
coooling (2nd
line)[0,0]
Altthough evaporrative cooling might be clim
matically
suitablle, the comfoort expectationns and room loading
represeent the actual limitations
l
to thhis strategy. As
A shown
in Fig. 5, evaporativee cooling resullts in an increase in the
absoluute and relativve humidity of the space, which
impactts on comfort.
w bulb temp
wet
Figure 3:Dry and wet bulb temperatures for
f Johannesburrg.
[0,0]
IMPACT OF
O DISPLACEMENT VENTILATIO
V
ON
ON PASSIV
VE COOLING
G STRATEGIES
The applicaability of passsive cooling strategies is a
function of climate,
c
comfo
ort expectationns related to thhe
activity in thhe space, interrnal loading of
o the space annd
equipment/deesign implicattions – often relating
r
back to
cost. What thhis paper puts forward is thaat the method of
introducing the
t cooling into
o the space also impacts on thhe
suitability. The
T impact is quite specific to each passivve
cooling strattegies, and will
w now be explored
e
for thhe
following sttrategies: evap
porative coolinng – single annd
two stage, thermal storag
ge based on night time air
a
temperaturess and air to ground heat exchaangers.
Figure 5:Psychrometrric chartsshowinng(left) the evaaporative
coolingg process with 1 as the outdoorr, 2 the supplyaand 3 the
room air
a conditions with
w increasing absolute
a
humidity during
the dayy, and (right) thhe 2-stage evapporative coolingg process
indicatiing the sensibble(1a to 1b) as well as adiabatic
a
componnent (1b to 2)
The increase in humidity in thhe space overr time is
relative to the amoount of vapourr being addedd to the
space. The amount of
o vapour enterring a space is directly
relatedd to the supply air flow. Therrefore the loweer the air
flow, the
t lower the abbsolute amounnt of vapour inttroduced
PLEA2009 - 26th
2
Conference onn Passive and Low
w Energy Architecture, Quebec City, Canada, 22-24 Juune 2009
into the spaace, slowing th
he build up of vapour of thhe
space duringg the day. Fig
g. 8 indicatess a significanttly
lower inflow
w of humidity fo
or DV compareed to MV.
Figure 6: Fann driven evaporative cooling
This meaans that the comfort
c
and cooling
c
capaciity
(and therefoore also cost)) requirementss of a speciffic
project mighht be met thrrough the use of evaporativve
cooling utilissing DV, wherre it might havve failed utilisinng
MV. In other words, using
g evaporative cooling
c
with DV
D
rather than with
w MV mean
ns that a higheer room load caan
be met (moree cooling for th
he same humiddity increase), or
alternatively a lower increaase in humidityy can be realiseed
for the same room load.
Two-stagge
evaporative
cooolingTwo-stagge
evaporative cooling add
ds a sensibble pre-coolinng
component to
t the cooling process,
p
whichh lowers the wet
w
bulb temperrature before applying dirrect evaporativve
cooling, as shhown in Figs. 5 and 7:
T1bdb = T1addb − η sens (T1addb − T1awb )
…(8)
from which
TS = T1bdb − η evap (T1bdb − T1bwb )
…(9)
wherehsensis taken
t
to be 90%
% and hevap 80%
% [0]
This sennsible pre-coo
oling componeent is typicallly
provided byy water that is
i cooled thouugh evaporativve
cooling, usinng residual waater from the evaporative
e
paart.
Other strateggies can also bee used for sensible pre-coolinng,
e.g. passing air through an
n air-to-groundd heat exchangger
before applying an evaporaative cooling coomponent.
The resullt impacts in two
t
ways; air is
i delivered att a
lower temperrature to the sp
pace and/or theere is an increaase
in duration in
i the year thaat evaporative cooling can be
b
used (Fig. 4) based on a maaximum Ts of 22
2 °C.
This influuences the sizee of the room load that can be
b
served by thiis strategy and
d the comfort conditions
c
of thhe
occupants ass less humidity is being addedd to the space as
indicated in Fig.
F 9.
Figure 7: Diagram indicating
i
2-staage evaporativee cooling
ment
equipm
Figure 8: Air flow ratees(V) and humiddity inflow (g) ovver a day
for singgle(Evap) and 2-stage (2S) evvaporative cooliing for a
space with
w a peak totaal cooling load of 70W/m2 indiccated the
differennce between DV
V and MV.
Diu
urnal range and thermaal store Night time
ventilaation and exposed thermal mass
m
is a relativvely well
knownn passive coolinng strategy. It uses
u low nightt time air
temperratures to flushh the structure of the buildingg of heat
built up
u during the previous
p
day. The
T same princciple can
be useed in the introdduction of a reemote thermal store in
the forrm of a packedd bed, as show
wn in Fig 9. This bed is
cooledd down at nightt by flushing with
w outside air through
mechaanical ventilatioon. During the day outside aiir is then
introduuced through this bed,whichh cools the aiir before
introduuction into thhe space. Thee packed bed can be
construucted by incorpporating a variiety of thermall storage
materials, including ceramic ballss and tiles, roocks and
possibly phase changge materials.
The climatic lim
mitation of thee packed bed thermal
store strategy is thhe number off days that thhe night
temperrature drops to a usable tempperature. Basedd on a Ts
of 22°C,
C and a miniimum temperature differencee of 3°C
betweeen the store material and the supply air, the
PLEA2009 - 26th
2
Conference onn Passive and Low
w Energy Architecture, Quebec City, Canada, 22-24 Juune 2009
duration thatt the strategy iss applicable is indicated in Fig
F
10. It is cleear that the ambient
a
tempeerature betweeen
24h00 and 05h00
0
is lowerr than 19°C foor 98.5% of thhe
year.
Figure 9: Packed bed thermall store
Figure 10: DVIEW
D
indicatees duration thaat the night tim
me
cooling strateggy is applicable in Johannesburgg.
bed for DV can be rooughly estimatted at 27m3, coompared
to a beed of 58m3 for MV
M ht.
Usiing DV ratheer than MV therefore
t
signnificantly
reducees the required size of the paccked bed therm
mal store,
impactting on both material
m
costs and more impportantly
the spaace that the sttore takes up. This last impaact is, in
the auuthors’ experiennce, the comm
mon factor thaat makes
the strrategy unfeasibble, due to thee high cost off making
the spaace available. An additionall benefit derivved from
the low
wer air flow raate is the reducction in fan poower and
therefoore energy consumption.
Airr to ground heat
h
exchangeersAir to grouund heat
exchanngers consist of
o two distincttly different strategies.
The firrst one, seasonnal ground souurce cooling (F
Fig. 12),
utilisess thedifferencee between thee ambient tem
mperature
and thhe ground tem
mperature at deepths typicallyy greater
than 3m.Fig. 14indiccates how the amplitude of seasonal
variation in the soil temperature decreases,
d
and shifts in
phase, as the depth increases.The second strateggy, daily
thermaal storage (Figg. 13), uses thee lower night time air
temperrature in a way
w
similar too the thermal storage
strateggy described in the previoous section, with
w
the
storagee medium beiing the groundd immediatelyy around
the pipping. Compared to seasonal ground
g
source cooling,
the pipping for daily thermal
t
storagee can be locateed nearer
to the surface, and the piping runns are typicallly closer
togetheer.
The feasiibility of the sttrategy is furthher influenced by
b
the cost and size of packeed bed, which is a function of
the storage capacity off the thermall material, thhe
efficiency of
o the heat transfer
t
betweeen the storagge
medium and the air, and thee air flow rate through the beed.
Hollmuller et
e al. [0] notees that ideallyy a packed beed
should causee a 180° phasee shift in the daily temperatuure
profile, as shhown in Fig 11
1. For such a 180°
1
phase shift,
they suggestt a packed beed volume of roughly1m3 for
f
every 100m3/h air flow raate. They furthher note that thhe
amplitude change (transm
mission) of thhe phase-shifteed
output is rellated to the th
hermal storage material that is
used:
Figure 12: Seasonal grround source coooling
Figure 11: Thhe input and outp
tput temperaturees of a packed bed
b
thermal store over time, illustrating the effecct of a 180 degrree
shift in phase. Based on [0].
Figure 13: Daily grounnd source coolingg.
for a cost-efffective materiaal like gravel the
t transmissioon
of the input amplitude
a
is arround 50%. Ussing the volum
meairflow relatiionship, the sizze of a 180° phhase shift packeed
The climatic appplicability of thhe daily groundd source
coolingg strategy is similar to tthat describedd in the
PLEA2009 - 26th
2
Conference onn Passive and Low
w Energy Architecture, Quebec City, Canada, 22-24 Juune 2009
previous seection on thermal stores. The climattic
applicability of the season
nal ground soource cooling is
demonstratedd in Figs. 14 an
nd 15.The perccentage of office
hours that this
t
strategy is applicable in
i Johannesbuurg
depends on a variety of factors, incluuding the deptth,
a the air floow
diameter andd length of thee exchanger, and
rate. Fig. 15 illustrates how
w the percentaage applicabiliity
of this strategy can be up
u to 100% at
a a Ts of 22°C,
C
depending on
o the length of
o the exchangger, in this caase
shown for ann air velocity of 5m/s and a deepth of 3m.
Evaaporative coolling – significcantly less hum
midity is
being introduced
i
intoo the space duee to the lower air
a flow.
2-sstage evaporaative cooling–– intrinsically 2-stage
evaporrative coolingg introduces less humiditty than
convenntional evaporrative coolingg due to the sensible
coolingg component. This advantagge is further am
mplified
by DV
V in the lower supply
s
air flow
w required com
mpared to
MV making
m
the strattegy applicablee to more projects due
to com
mfort levels.
Theermal storagee – the storage volume is reduced,
r
with thhe main impacct being the reeduction in thee cost of
providding the space for
f the store.
Airr to ground heeat exchangerss – the pipe length
l
is
reduceed dramaticallyy.
Figure 14: Annual
A
air tem
mperature and expected grounnd
temperature at
a 3m, based on [0]
[
Figure 15: Peercentage of offfice hours (6h too 18h) that Tout of
the exchangerr is below maxim
mum Ts of the loaad, for air velocity
of 5 m/s, pipe diameter of 0.38
8m, and depth off 3m. Based on [0]
[
Using thee methodology
y described by Mihalakakou et
al. [0] to esttimate the leng
gth of exchanger required for
f
the simulateed Johannesb
burg office, with
w
the sam
me
assumptions as in Fig. 15,, results in thee following: DV
D
mpared to MV
Vht
requires rouughly 80m off pipes, com
requiring rouughly 170m.
The abovve estimation illustrates thatt, as DV loweers
the volume air
a flow rate reequired, it signiificantly reducces
the pipe lenngths required
d and thereforre increases thhe
technical annd financial feeasibility of thhe strategy. An
A
additional beenefit is a relatiive reduction in
i fan power annd
therefore eneergy consumptiion.
CONCLUSIIONS
DV significaantly lowers th
he required suppply air flow to
serve a room
m heat load, when
w
compareed to MV at thhe
same Ts. Thhis reduction im
mpacts the feaasibility of usinng
passive cooliing strategies in
n projects in a various ways:
The publication of the ASH
HRAE Guideliines has
assisteed in determinning with connfidence the fllow rate
requireed for DV:siignificantly loower than previously
estimaated using MV
V calculations.. However, thhe sizing
and caalculations of the
t various passsive coolingsttrategies
(with exception
e
of evaporative
e
annd 2 stage evaaporative
coolingg) still dependds on non-stanndardised methhods that
would typically noot be consideered by convventional
HVAC
C engineers. Establishing
E
g
guidelines
for passive
coolingg, which are enndorsed by orgganisations thatt HVAC
engineeers refer to inn confidencee.g. ASHRAE
E, would
assist greatly with gaining
g
furtherr penetration of these
strateggies beyond envvironmentally extreme projeccts.
ERENCES
REFE
1. Chen, Q. and Glicksman, L. (20003) System Perfformance
Evaluattion and Deesign Guidelinnes for Dispplacement
Ventilaation. ASHRAE
2. Svvensson, A.G.L
L. (1989). Nordic
N
experiennces of
displacement ventilatioon systems. ASH
HRAE Trans. 95((2).
3. Grapph produced usinng DView Versioon 1.09 developped by the
Nationaal Renewable Energy Labboratory and Mistaya
Engineering Inc. using IWEC weather file
4. Inteernational Weatther for Energyy Calculations (IWEC)
weatherr
data
file
from
sourced
http://w
www.eere.energyy.gov/buildings/eenergyplus/cfm//weather_
data3.cfm, last visited on 22nd May 20008
5. IWA
AC Two stage evvaporative cooliing units producct manual
from PROTEK
P
CC, Box
B 1943, Halfw
way House 16885, South
Africa
6. Holllmuller et al.(2006) - A new ventilation andd thermal
storagee technique forrpassive coolingg of buildings: thermal
phase-sshifting, PLEA, Geneva,
G
6-8 Sepptember 2006
7. Kusuuda, I. and Acheenbach, P. (19655), Earth temperature and
thermall diffusivity at seelected stations in
i USA, Nationaal Bureau
of Stanndards, Building Research Div., Washington
W
D.C
C.
8. Mihalakakou, G. et al, Parametric prediction of thhe buried
pipes cooling
c
potentiall for passive coooling applicatioons, Solar
Energyy