Numerical analysis of hybrid ventilation performance depending on climate characteristics

International Journal of Ventilation Volume 1 HybVent – Hybrid Ventilation Special Edition________________________________________________________________________________________________________________________

________________________________________________________________________________________________________________________

41

Numerical Analysis of Hybrid VentilationPerformance Depending on Climate Characteristics

Florence Cron, Christian Inard and Rafik Belarbi

LEPTAB, University of La Rochelle, Avenue Michel Crépeau,F-17042 La Rochelle Cedex 01, France

Abstract

This study, which formed part of the Annex 35 “Hybrid Ventilation in New and Retrofitted OfficeBuildings” project, was completed at LEPTAB and supported by the French Research Ministry and theADEME (Agence De l’Environnement et de la Maîtrise de l’Energie). It consisted of modelling a typicalclassroom and comparing different control strategies to estimate the performance of a hybrid ventilationsystem for different climates.

The intention was that investigated classrooms were assumed to be on the middle level of a three-storeybuilding, oriented South and surrounded by other classrooms subjected to the same conditions. Twomechanical ventilation systems were taken as references. These were: a mechanical exhaust system with alow consumption fan and without heat recovery, and a balanced mechanical ventilation system with two fansand some heat recovery. The hybrid ventilation approach investigated was a fan assisted natural ventilationsystem incorporating a demand control strategy based on indoor air temperature and CO2 concentration.

The performance of this hybrid ventilation system was analysed in terms of energy consumption, indoor airand dry resultant temperatures, and CO2 concentration level. Simulations of specific weeks in the year wereperformed for ten French cities and gave quite detailed patterns of behaviour. The study was extended toinclude yearly mean values of energy consumption. The results for both short and long time periods showedthe potential of this specific hybrid ventilation system according to climate and control strategy. Hybridventilation is shown to provide improved air quality. Also, in relation to delivered energy, energy savings arepossible but, except for Mediterranean cities, are not as much as with a mechanical system with heatrecovery.

Key words: hybrid ventilation, control strategy, thermal comfort, CO2 concentration, energy consumption.

1. Introduction

Recently, there has been increased interest in hybridventilation for the design of office buildings. Thework of the International Energy Agency (IEA)Annex 35 “Hybrid Ventilation in New andRetrofitted Office Buildings” has resulted in thepublication of a state of the art report (Annex 35,2000) and a booklet with a CD-ROM (Heiselberg,2002). These documents describe hybrid ventilationprinciples and cover design, control strategies,analysis methods and examples of existingbuildings.

It is at the early planning stage that the designershould evaluate which ventilation system is mostappropriate in terms of indoor air quality and energyconsumption. Several tools are available for such aventilation performance analysis (Delsante and Vik,

2000, Frascatoro and Perino, 2002). Various otherstudies on hybrid ventilation have already beenundertaken. Comparisons between results obtainedwith different evaluation tools were presented byDelsante et al (2002a and 2000b). The potential of ahybrid ventilation system compared to moretraditional ones was also presented by Cron et al(2002b) and by Elmankibi et al (2002) but waslimited to a couple of locations. The influence ofclimate on hybrid cooling systems was reported bySpindler et al (2002). Barriers to the implementationof hybrid ventilation in Canadian offices andeducational buildings due to regulations and climatehave been outlined by Bourgeois et al (2002).Simulations of existing buildings have also beenundertaken by Jeong and Haghighat (2002).However none of these studies have so far presentedan analysis of the potential of a hybrid ventilationsystem in relation to climatic characteristics.

F Cron, C Inard and R Belarbi________________________________________________________________________________________________________________________

________________________________________________________________________________________________________________________

42

Thus the purpose of this paper is to broadennumerical studies on hybrid ventilation to includeseveral cities and to draw some conclusions on thepotential of this type of ventilation system. We firstgive a description of the typical classroom that wassimulated for ten French cities. Then, after apresentation of the tools and the models used,results are given in terms of indoor air quality anddry resultant temperatures, CO2 concentration andenergy consumption. Simulations of specific weeksshow the room air flow and thermal behaviouraccording to occupancy and outdoor conditions.Results were also extended to a whole year. Theresults for all the locations allowed us to see thepotential of this hybrid ventilation system, given theclimate and the control strategy used.

2. Classroom Description

2.1 Description

The simulated room was a single classroom in themiddle level of a three-storey building. It was 9 mwide (along the external facade), 6 m deep and 3 mhigh. The room was oriented south and had a 1 mhigh double pane window along the whole facade.The external wall was assumed to consist of 80 mmof mineral wool and 150 mm of massive concrete,and the window to have a U-value of 2.7 W/(m2.°C).All external values were taken from meteorologicaldata, except the CO2 concentration that wasassumed to be constant and equal to 400 ppm.

Normal school hours were set from 8h00 to 15h00from Monday to Friday. Ventilation was providedduring this schedule by the opening of two inletsand one outlet. Heating hours were from 7h00 to15h00, from Monday to Friday. The set point airtemperature was 21 °C during the heating hours andwas reduced to 18 °C during the non-heating hours.Preheating brought the ventilation supply air to 18°C for all the ventilation systems. The heatingpower depended on the climate and aradiative/convective partitioning of 0.5/0.5 wasassumed. For the whole year study, we used theFrench regulation heating period i.e. from the 1st ofOctober until the 20th of May.

A teacher and 24 pupils were assumed to occupy theroom from 8h00 to 12h00 and from 13h00 to 15h00on Mondays to Fridays. On Tuesdays, from 10h00to 11h00, only 13 people were in the room, and at14h00 everybody left the school. There was noschool on Wednesday afternoons either. During

occupancy, lighting loads were 10 W/m2 andinternal heat loads were 80 W/person and CO2production was 18 l/(h.person). These sensibleinternal gains were assumed to be 50% radiative and50% convective. A solar shading factor was takeninto account, depending on the location and theseason, which was between 0.2 and 0.4 of the totalincident solar radiation when this was more than200 W/m2.

2.2 Ventilation Systems

Infiltration was assumed to be constant and equal to0.2 ach all the time. Window airing was defined bya constant air flow equal to 4 ach and could occurduring normal school hours only. This value wasadditional to the other air flow rates andindependent of the weather and the ventilationsystem. Window opening was scheduled when boththe indoor and outdoor air temperatures wererespectively higher than 23 °C and 12 °C, andclosed again as soon as the indoor air temperaturewent below 21 °C. Some night cooling wasprovided in hot periods of the year, includingSaturdays and Sundays. The night cooling was set tooperate between 22h00 and 7h00 when the indoortemperature was higher than 24 °C and when theindoor-outdoor temperature difference was higherthan 2 °C. It was turned off whenever the indoortemperature was below 18 °C.

Two mechanical ventilation systems were taken asreferences against which the performance of thehybrid ventilation system was compared. The firstmechanical ventilation alternative (MV1) was amechanical exhaust system in which a constantexhaust air flow of 0.15 m3/s was assumed andsupply air was provided through inlet grilles in thefacade. The system was operated during normalschool hours (from 8h00 to 15h00) and also, whennecessary, to provide night cooling. The low fanpower consumption was equal to 1000 W/(m3/s).

The second mechanical ventilation arrangement(MV2) was a balanced system that incorporatedboth an exhaust fan and a supply fan for the exhaustand supply of air at 0.15 m3/s. A heat recovery unitwith a temperature efficiency of 0.6 was includedbefore the preheating system. Both fans were alsooperated for night cooling. The fan combinedconsumption was 2500 W/(m3/s).

For both mechanical ventilation systems, the fan airflow rate was set to maintain an indoor CO2concentration below 1200 ppm during occupancy.


________________________________________________________________________________________________________________________

43

The hybrid ventilation approach was based on fanassisted natural ventilation. The opening of theinlets and outlets was not only dependent on theschool hours, but also on the CO2 concentration inthe classroom. The two inlet grilles were 0.5 mabove the classroom floor. The first one openedwhen the CO2 concentration was higher than 800ppm, the second when it was higher than 1000 ppm.Finally the fan was turned on when the CO2 reached1200 ppm. The dead band values were 100 ppm. Forthis hybrid ventilation system, the classroomincorporated a 4 m high exhaust chimney to enhancethe natural stack effect. No conductive heat transferwas taken into account through the chimney walls.Wind effects were taken into account: Cp values(wind pressure coefficients) from the AIVCliterature were used to make an interpolation (seeTable 1). For the chimney, the Cp was equal to -0.6and was independent of the wind direction. Theassisting fan was based on a low energyconsumption device equal to 200 W/(m3/s) andprovided 0.15 m3/s when it was on. Two nightcooling modes were considered: the first one withthe natural stack effect only (HVa) and the secondusing the fan (HVb).

Table 1. Cp values depending on the wind direction

Wind angle 0° 45° 90° 135° 180°

Cp value 0.25 0.06 -0.35 -0.60 -0.50

2.3 Meteorological Data and Locations

The ten French cities chosen for the simulation aregiven in Table 2.

Table 2. Cities chosen for simulations

French Cities Latitude ° Longitude °Agen 44.2 N 0.6 EAjaccio 41.9 N 8.7 ECarpentras 44.0 N 5.0 ELa Rochelle 46.2 N 1.2 WLimoges 45.8 N 1.3 EMâcon 46.3 N 4.8 ENancy 48.7 N 6.2 ENice 43.7 N 7.3 ERennes 48.1 N 1.7 WTrappes 48.8 N 2.0 E

France has a temperate climate covering an oceanicclimate, a Mediterranean climate and a semi-

continental climate. Thus government regulationdivides France into winter and summer climaticzones. Three zones are considered in winter fromthe coldest one (H1) to the warmest one (H3) andfour zones in summer from the coolest one (Ea) tothe warmest one (Ed), see Figure 1 and Figure 2.Degree-days based on 18 °C and also on 21 °C(temperature set point during heating hours) for theheating period are given in Figure 3.

Figure 1. Winter climatic zones in France

Figure 2. Summer climatic zones in France

Heating Degree-Days

1558

260624572874

27022663

208721302290

1631

3387

22302297

3343 355932903148

276527862964

0

500

1000

1500

2000

2500

3000

3500

4000

Ajaccio Nice

Agen

Carpen

tras

La Roc

helle

Rennes

Limog

es

Mâco

nNan

cy

Trappes

HD

D [°

C d

ays]

HDD 18°CHDD 21°C

Figure 3. Heating degree-days duringthe heating period


________________________________________________________________________________________________________________________

44

Cooling degree-days were defined by computing thedegree-days of outside temperatures above 26 °C,when air conditioning may be needed. The degree-days between 21 °C and 26 °C were also calculated.They were both taken during the non-heating period,see Figure 4.

Cooling Degree-Days

54

5 8 13 7 6

263291

163183

143

88 101 97

14 132

7

135104

0

50

100

150

200

250

300

350

Ajaccio Nice

Agen

Carpen

tras

La Roc

helle

Rennes

Limog

es

Mâco

nNan

cy

Trappes

CD

D [°

C-D

ays]

CDD 26°CDD 21°C-26°C

Figure4. Cooling degree-days and degree-daysbetween 21 °C and 26 °C during

the non-heating period

The simulation goals were to determine indoortemperatures, indoor air quality and performanceresults for specific weeks and for a whole year

3. Models used for the Simulations

The object oriented solver SPARK was used toperform the whole problem with differential andnon linear equations. They were all implemented inSPARK for a previous study undertaken withintIEA Annex 35 (Cron and Inard, 2002a) with a timestep equal to 60 s. Comparisons between simulationresults for another case study and experimental datawere also reported earlier (Cron et al, 2002b).

3.1 The Room Air Flow and Thermal Models

A single well-mixed zone with pressure varyinghydrostatically was assumed for the models.

For a zone i with n openings, the “pure” air massbalance and the pollutant mass balance equationswere:

∑∑==

=−n

jijas

n

jjias mm

110 (1)

∑ ∑= =

=+−n

j

n

j

esesijesjies dt

dVSmm

1 1

ρ (2)

As described previously, the infiltration and thewindow airing air flows were assumed to beconstant. For both mechanical ventilation systems,the air flow rate was equal to 0.15 m3/s.

For the hybrid ventilation system, when the fan wasturned off, a simple power law relation was used forthe inlet grilles and the chimney:

( ) ( ) 5.05.0ijji PPCPCQ −=∆= (3)

whereC = 0.053 m3/(Pa 0.5.s) for each grilleC = 0.088 m3/(Pa 0.5.s) for the chimney.

From Equation (3), the thermal buoyancy differencewas incorporated to obtain the mass air flow rate,using a correction factor KQ (Feustel et al, 1990):

( ) 5.0ijQijjiQijjias PPCKQKm −== ρρ (4)

with 2

jiij

ρρρ

+= the mean air density of the zone

i and the zone j.

The thermal balance equation was:

( ) ( )( )

( )dt

dVTcp

dtdTVcpcp

PPTTcpS

TcpmcpmTcpmcpm

esesesesasas

convconvloadconvheateseses

n

j

n

jesijesasijasjesjiesasjias

ρρρ ++=

Φ+++−+

+−+∑ ∑= =1 1

(5)

3.2 The Room Envelope Models

Conduction heat transfer was described by anelectrical 2R-3C model (Rumianowsky et al, 1989)that gives a good response to a high frequencyexcitation indoors, see Figure 5.

TM(t)

CA CM CB

TA(t) TB(t)

RARB

Indoor Layer i Layer j Outdoor

eA , nA eB , nB

TA(t)Ti-1(t) Ti(t)

TM(t) Tj-1(t) Tj(t)TB(t)

Figure 5. Description of the 2R-3C model


________________________________________________________________________________________________________________________

45

Considering the steady state, we have:

∑=

=An

k k

kA S

eR

1 λ (6)

∑+=

=Bn

nAk k

kB S

eR

1 λ (7)

with nA + nB = n, the number of layers.

Accumulation of heat in the wall was expressed by:

BBMMAAaccum dTCdTCdTCdQ ++= (8)

where:

( )∑=

−=An

jjjjjA SecpC

11 βρ (9)

∑+=

=B

A

n

niiiiiB SecpC

1δρ (10)

( )∑∑+==

−+=B

A

A n

niiiii

n

jjjjjM SecpSecpC

111 δρβρ (11)

A

j

k j

j

k

k

j RS

eS

e∑

−

=+

=

1

1 2λλβ (12)

B

i

nk i

i

k

k

i R

Se

Se

A

∑−

+=+

=

1

1 2λλδ (13)

A comparison between this model response and afinite difference model response was undertaken.The position of the intermediate node was set to theoptimum value of eA that provided the minimumerror EeA where EeA was defined by:

∫∞

−=0

dtTTE AdifferencefiniteAeA (14)

The simulated room was assumed to be surroundedby rooms at the same conditions.

The outdoor long-wave radiation was given by:

( ) ( )( ) ( )

−++

−−=Φ

44

44

2cos1

2cos1

osvoo

osoooonetGLO

TTS

TTS

βσε

βσε (15)

The short-wave radiation absorbed by the externalwall surface was computed from the incident solarradiation and the solar properties of the surface.

The convective heat transfer at the outdoor surfacewas calculated by:

( )osoooconvoconv TTSh −=Φ (16)

nmetoconv dUch += (17)

where c = 2.5, d = 3.5 and n = 1 according to Ferries(1980).

A mean radiative temperature model was chosen todescribe the long-wave heat transfer indoors. Part ofthe internal loads and the heating system was takeninto account in the model. Given a surface i wemodelled this exchange as:

( )isrmiirminetGLO TTSh −=Φ (18)3

24

+= rmis

iirmTT

h σε (19)

( )

∑

∑

=

=++

= n

iiirm

n

iradloadradheatisiirm

rm

Sh

PPTShT

1

1 (20)

Indoors, solar radiation transmitted through thewindow was assumed to be entirely incident on thefloor. One part was absorbed, depending on thesolar properties of the floor, the other part wasassumed to be reflected in a diffuse way. Both thetransmitted and the reflected solar radiation werethen absorbed over all the surfaces, depending onthe surface ratio.

Convection indoors was obtained by:

( )isiiconviconv TTSh −=Φ (21)

where b

isiconv TTah −= (22)

a and b were taken from Allard (1987):

a = 3 and b = 0 for a horizontal surfacea = 1.5 and b = 0.33 for a vertical surface.


________________________________________________________________________________________________________________________

46

The dry resultant temperature is a simplifiedindication of thermal comfort and was obtained by:

rmconv

rmrmconvrs hh

ThThT

++

= (23)

3

24

+= rmrs

rmTT

h σε (24)

brs

conv DTT

ah−

= (25)

hconv is here the globe convection heat transfer, a andb are coefficients equal to 1.4 and 0.25 respectivelyand D is the globe diameter equal to 0.15 m.

4. Simulation Results and Discussion

Simulations of specific weeks for different seasonswere performed for each city and results are given interms of indoor and dry resultant temperatures, CO2concentration and energy consumption. Simulationswere extended over a whole year, to better estimatethe potential of this hybrid ventilation systemaccording to the climate specifications. For everysimulation, the two mechanical ventilation systemsgive the same temperatures and CO2 concentrationssince they differ only by the heat recovery and thefan consumption. Hybrid ventilation HVa and HVbare identical except when there is some nightcooling, but in that case, results remain quite close.So we consider, except for energy consumption,only the results for mechanical ventilation (MV) andhybrid ventilation with the mechanical night cooling(HV).

4.1 Short Period Analysis

Results for specific days for both Trappes and Niceare given here: the first day is a Tuesday in winter(the 16th of January) and the second is a Thursday insummer (the 28th of June).

In winter, Figure 6 and Figure 7 show the set pointtemperature change in the morning (at 7h00) and inthe afternoon (at 15h00). For both Trappes andNice, the maximum temperature occurs with hybridventilation since, when the pupils enter the room at8h00 or 13h00, there is at first no ventilation due tothe low CO2 level. The first grille opens when theCO2 concentration reaches 800 ppm. Half of thepupils are in the classroom from 10h to 11h and wesee a little temperature decrease due to lowerinternal loads.

Indoor air temperature for Trappes in winter

18

18.5

19

19.5

20

20.5

21

21.5

22

22.5

6 7 8 9 10 11 12 13 14 15 16Time [h]

Tem

pera

ture

[°C

]

MV temperatureHV temperature

Indoor CO2 concentration for Trappes in winter

300400500600700800900

1000110012001300

6 7 8 9 10 11 12 13 14 15 16Time [h]

Con

cent

ratio

n [p

pm]

MV CO2 HV CO2

Figure 6 . Temperatures and CO2 concentration in winterfor Trappes (MV and HV systems)

Indoor air temperature for Nice in winter

18

18.519

19.5

20

20.5

21

21.5

22

22.5

6 7 8 9 10 11 12 13 14 15 16Time [h]

Tem

pera

ture

[°C

]


Indoor CO2 concentration for Nice in winter

300400500600700800900

1000110012001300

6 7 8 9 10 11 12 13 14 15 16Time [h]

Con

cent

ratio

n [p

pm]

MV CO2HV CO2

Figure 7. Temperatures and CO2 concentration in winterfor Nice (MV and HV systems)


________________________________________________________________________________________________________________________

47

During lunch time and between 14h and 15h, thereis no occupancy and, whereas the fans are stillworking for the mechanical mode, the inlet grillesclose for the hybrid system because of the low CO2concentration in the room. This gives, of course,some energy savings.

CO2 concentration evolution is quite differentbetween mechanical and hybrid ventilation modes.With all the pupils, the CO2 tends to the constantvalue of 1200 ppm with mechanical ventilation,whereas with hybrid ventilation, the ventilation ratevaries both according to CO2 control and the stackeffect. The latter is sufficient to ensure good indoorair quality and the fan doesn’t need to be turned on.As a consequence CO2 concentration is higher withmechanical than with hybrid ventilation. Also, in themorning, a lower outside temperature in Trappesthan in Nice induces a more important stack effectand thus a lower CO2 level for Trappes in themorning. When half of the pupils are in the class,the CO2 level decreases and goes to around 800 ppmwith mechanical ventilation. The hybrid ventilationcontrol closes one inlet grille at 900 ppm, so the airflow rate is reduced, whereas mechanical ventilationcontinues to provide the constant 0.15 m3/s. Overall,the hybrid system leads to a concentration ofbetween 900 and 1000 ppm which is, on average,higher than with the mechanical mode. However, wecan say that for this specific winter day, themaximum value of CO2 concentration is lower withhybrid ventilation than with mechanical ventilationfor both cities. There is also less exposure to aconcentration higher than 1000 ppm for Nice (where1000 ppm is the value given in the Frenchregulation).

In summer, Figure 8 and Figure 9 show the resultsof simulations. For Trappes, as there is at first noventilation for hybrid ventilation, the temperature isseen to rise quite quickly and window airing isprovided earlier for the hybrid mode than for themechanical one. Nice has higher temperatures, sowindow airing is provided as soon as the pupils arein the room. When the window is opened for bothsystems, the hybrid ventilation temperature is higherthan the mechanical ventilation one. This is initiallydue to a lower stack effect in summer compared towinter. Then, with the window airing, only one inletgrille is opened for hybrid ventilation. Hence thereis a lower air flow rate with the hybrid system thanwith the mechanical one. For these reasons, themaximum value of temperature is obtained withhybrid ventilation, but doesn’t exceed 26 °C in both

Indoor air temperature for Trappes in summer

2020.5

2121.5

2222.5

2323.5

2424.5

2525.5

6 8 10 12 14 16 18 20 22 24Time [h]

Tem

pera

ture

[°C

]


Indoor CO2 concentration for Trappes in summer

300400500600700800900

1000110012001300

6 8 10 12 14 16 18 20 22 24Time [h]

Con

cent

ratio

n [p

pm]

MV CO2 HV CO2

Figure 8. Temperatures and CO2 concentration insummer for Trappes (MV and HV systems)

Indoor air temperature for Nice in summer

2020.5

2121.5

2222.5

2323.5

2424.5

2525.5

6 8 10 12 14 16 18 20 22 24Time [h]

Tem

pera

ture

[°C

]


Indoor CO2 concentration in Nice in summer

300400500600700800900

1000110012001300

6 8 10 12 14 16 18 20 22 24Time [h]

Con

cent

ratio

n [p

pm]

MV CO2 HV CO2

Figure 9. Temperatures and CO2 concentration insummer for Nice (MV and HV systems)


________________________________________________________________________________________________________________________

48

cities. At lunch time, window airing provides somefree cooling. After 17h, solar shading is removed forboth cities which explains the slight temperature riseat that moment. Nice is warm enough to have nightcooling on the night before and the night after thisday, whereas Trappes has some night cooling onlyduring the following night.

Since the window is opened almost immediatelywith hybrid ventilation, CO2 concentration staysbelow 1000 ppm, and is lower than that given bymechanical ventilation where there is no windowairing. But, when the mechanical mode has somewindow airing, the CO2 concentration is lower thanwith the hybrid mode. The example of Trappesshows that, even if the mean CO2 concentration ishigher with the hybrid ventilation, the time exposureto a concentration higher than 1000 ppm remainslower with hybrid ventilation.

4.2 Yearly Simulations

As presented by Ghiaus and Allard (2002), twobalance temperatures for the room equipped withmechanical ventilation have been calculated. Thefirst is with the window closed and the second iswith the window open. Given the requirements ofindoor comfort, Figure 10 presents the resultssubject to the required heating, cooling and air–conditioning needs.

0

10

20

30

40

50

60

0 5 10 15 20 25 30 35 40Outdoor Air Temperature To [°C]

Tem

pera

ture

[°C

]

Balance temperature with no window airing

To

Heating

Window 100% opened

Air conditioningZone of comfort

Window closed

Balance temperature with window airing

12 26

Heating, cooling and air-conditioning needs

Windowairing

Cooling

Figure 10. Heating, cooling and air-conditioning needsduring occupancy

For mechanical ventilation, Figure 10 shows someheating is needed whenever the outdoor airtemperature is below 10.5 °C. When the outdoortemperature is equal to 12 °C, the window opensand can be successively opened and closed between12 °C and 16 °C, depending on the solar gains thatweren’t taken into account for this diagram. Whenthe outside temperature is higher than 16 °C, the

window is opened all the time during school hours.Even if window airing can provide some freecooling, this part of the figure begins to be critical interms of thermal comfort when considering solarradiation gains and the lower stack effect for hybridventilation in this range of temperatures. For anoutdoor air temperature higher than 26 °C, the roomshould be air-conditioned to provide satisfactorythermal comfort. Figure 10 shows the importance ofthe needs of thermal comfort conditions in relationto the outdoor air temperature, thus the annualoutdoor air temperature distribution duringoccupancy was calculated from the weather datafiles, see Table 3.

Table 3. Annual distribution (percentage) of the outdoorair temperature during occupancy

Cities To<12°C 12°C≤≤≤≤To<16°C

16°C≤≤≤≤To<26°C

26°C≤≤≤≤To

Ajaccio 17 34 46 3Nice 25 30 45 0Agen 50 19 27 4

Carpentras 44 17 33 6La Rochelle 51 23 26 0

Rennes 59 22 19 0Limoges 60 16 24 0Mâcon 59 14 26 1Nancy 62 14 23 1

Trappes 65 14 21 0

Several categories of cities can be distinguishedhere. First of all, Ajaccio and Nice have an annualoutdoor air temperature higher than 16 °C duringmore than 45 % of the occupancy time and have anoutside temperature lower than 12 °C for less than25% of the occupancy. These two cities are also theones that have the lowest heating degree-days andthe highest degree-days between 21 °C and 26 °C.

Although Agen and Carpentras experience outsidetemperatures higher than 16°C during occupancyless frequently than Ajaccio and Nice (30% and40%), they have more cooling degree-days andfewer degree-days between 21 °C and 26 °C thanNice and Ajaccio. So, Agen and Carpentras belongto a second category of cities.

Although Agen and La Rochelle are in the sameclimatic zones (H2 and Ec) they show climaticdifferences, La Rochelle is cooler since the outsidetemperature is never above 26 °C during theoccupancy time. Also, although La Rochelle andCarpentras have about the same heating degree-


________________________________________________________________________________________________________________________

49

days, the cooling degree-days and the annualtemperature distribution are different. In fact bothLa Rochelle and Rennes have a quite mild climate,with the Atlantic Ocean influence. They have aboutthe same frequency of outside temperature between12 °C and 16 °C as between 16 °C and 26 °C. ButRennes has a cooler summer since the coolingdegree-days and the degree-days between 21 °C and26 °C are both lower.

The last category includes Limoges, Mâcon, Nancyand Trappes that have an outside temperature lowerthan 12 °C for about 60% of the occupancy time andhave the highest heating degree-day values.

The global performances of hybrid ventilation onindoor thermal comfort and air quality wereanalysed over the whole year. For all the cases, theindoor air and the mean dry resultant temperatureswith the two hybrid systems are both less than 0.4°C higher than with the mechanical mode (seeFigure 11). Air flow rates are in fact lower withhybrid ventilation for reasons explained previouslywhen we analysed specific days. These differencesappear to be a little more important for cities with awarmer climate in the summer zone Ed. Figure 11shows that Carpentras (zone Ed) has the highestmean dry resultant temperature. The reasons are thatCarpentras is the city with the highest coolingdegree-days and it has an outdoor temperaturehigher than 26 °C during 6 % of the occupancytime. On the other hand Rennes (zone Ea – thecoolest zone) has the lowest mean dry resultanttemperature value and has in fact, as shownpreviously, the lowest number of cooling degree-days and degree-days between 21 °C and 26 °C.

Annual mean dry bulb temperature during occupancy

23.923.824.124.3

23.8

24.8

25.9

24.424.224.3 242424.3

24.6

23.8

24.9

26.2

24.624.6 24.4

21

22

23

24

25

26

27

Ajaccio Nice AgenCarpentrasLa Rochelle

RennesLimoges Mâcon NancyTrappes

Temperature[°

MV1 HVb

Figure 11. Annual mean dry resultanttemperature during occupancy

The annual mean CO2 concentration results show areal difference between the warm climates and the

mild or colder ones. For Ajaccio, Nice, Agen andCarpentras, the value of the mean CO2 concentrationis clearly higher with the hybrid system and it is theopposite for all the other cities (the difference forMâcon is too small to give a conclusion). Mean CO2concentration is, in fact, strongly associated withwindow airing (i.e. periods with an outdoortemperature higher than 12 °C). There are also somecontrary effects between window airing and thestack effect. As shown previously, there is a lowerair flow rate when the window is opened for hybridventilation since one inlet grille is closed. Hence forAjaccio, Nice, Agen and Carpentras, the windowairing effect prevails over the stack effect. ForMâcon, these effects are approximately equal. Forall the other cities, the stack effect prevails overwindow airing.

Annual mean CO2 concentration during occupancy

9971009 1013

955938 946

979

1031

969

926

973

907904

984983984979

999

948

987

800

850

900

950

1000

1050

Ajaccio Nice

Agen

Carpen

tras

La Roc

helle

Rennes

Limog

es

Mâco

nNan

cy

Trappes

CO

2 con

cent

ratio

n [p

pm] MV1 HVb

Exposure to a CO2 concentration higher than 1000ppm

362

509

656

577537

607 616

405

517

353

491515473457

542

368316

486

299324

0

120

240

360

480

600

720

Ajaccio Nice

Agen

Carpen

tras

La Roc

helle

Rennes

Limog

es

Mâco

nNan

cy

Trappes

Exp

osur

e tim

e [h

]

MV1 HVb

Figure 12. Annual mean CO2 concentration and exposureto a CO2 level higher than 1000 ppm

Even if Mâcon and Rennes both have an outdoortemperature lower than 12 °C during 59 % of theoccupancy time, they are respectively in the summerzones Ec and Ea which explains the differences inFigure 12. Likewise La Rochelle and Agen have asimilar frequency of outdoor temperature lower than12 °C and are both in the zones H2 and Ec.However they have a difference in terms of meanCO2 concentration. In fact, according to the degree-days, La Rochelle is less warm in summer and lesscold in winter than Agen.


________________________________________________________________________________________________________________________

50

Although in some cities, hybrid ventilation seems tobe a priori less efficient with regard to indoor airquality, Figure 12 shows that there is less exposureto high CO2 concentrations in all cases. As shown inthe specific day analysis (Figures 6, 7, 8 and 9),there can be an exposure to a concentration higherthan 1000 ppm for mechanical ventilation (when thewindow is still closed), but the mean concentrationvalue is lower for mechanical ventilation than forhybrid ventilation (when the window is opened).Here again this is due to the antagonist effects ofwindow airing and stack effect. The importantconclusion is that, irrespective of climate, whenapplying this specific control strategy, indoor airquality, in terms of CO2 concentration, is better withhybrid ventilation than with mechanical ventilation.

Table 4 illustrates that hybrid ventilation providessome energy savings on heating and preheating forall cases, mostly because the inlets are closed duringnon-occupancy. The energy saving difference is dueto a higher stack effect for colder cities since theyinduce a higher air flow rate and thus more heatingand preheating consumption. In relation to fanenergy savings, warm cities and especially Ajaccioand Nice, have more fan consumption than coldcities, due to more night cooling requirements.Differences between the two hybrid modes are notsignificant, so results are only presented for HVb.

Table 4. Annual energy savings (percentage) between HVcompared to MV1

Energysavings

HeatingHVb/MV1

PreheatingHVb/MV1

FanHVb/MV1

Ajaccio 27 27 92Nice 21 16 92Agen 11 14 94

Carpentras 24 12 93La Rochelle 16 7 96

Rennes 11 7 98Limoges 7 7 97Mâcon 6 11 95Nancy 5 8 97

Trappes 6 6 98

Table 5 correlates the previous conclusions: Ajaccioand Nice have less preheating and more energyconsumption for the fan than the other cities, thushybrid ventilation with its low fan power leads tohigh energy savings compared to both mechanicalmodes. For the other cities, hybrid ventilation

provides energy savings for heating, preheating andthe fan compared to the first mechanical mode (i.e.extract only) but energy consumption is higher thanfor the mechanical system with heat recovery. Thereason is that preheating prevails over the fan interms of consumption for these cities. Finally allthese characteristics appear also in the annual totalenergy consumption for all the ventilation modes(Figure 13).

Table 5. Fan consumption over preheating consumptionfor MV1. Comparison of consumption between hybrid

ventilation and the mechanical ventilation modes(percentage)

Cities Fan /Preheating

(MV1) x 100

HVbcompared

to MV1

HVbcompared

to MV2Ajaccio 41 -42 -28

Nice 36 -33 -14Agen 17 -23 +31

Carpentras 24 -27 +27La Rochelle 16 -19 +65

Rennes 12 -16 +58Limoges 12 -15 +61Mâcon 13 -16 +33Nancy 11 -13 +44

Trappes 11 -13 +48

Annual total energy consumption

3060

2877

2860

2353

2269

1856

1521

2106

1370

1182

169918

52

1796

1239

1210

91612

32

1063

872

956

2512

1907

2660

2388

1989

1510

1108

1619

913

683

0

500

1000

1500

2000

2500

3000

3500

Ajaccio Nice

Agen

Carpen

tras

La Roc

helle

Rennes

Limog

es

Mâco

nNan

cy

Trappes

Con

sum

ptio

n [k

Wh]

MV1 MV2 HVb

Figure 13. Annual total energy consumptionfor the different ventilation modes

Finally Figure 13 shows that the performance of fanassisted natural ventilation is dependent on theheating needs. The cities that require more nightcooling have a better performance with the hybridsystem, whereas the ones that have high heatingdegree-day values have the highest energy savingswith the heat recovery system. These results may begeneralised to other buildings, as soon as thetemperature range distribution over the year is


________________________________________________________________________________________________________________________

51

known. With these results, the relation betweenclimate specifications and energy savings over awhole year can be finally summarise, see Table 6.

Table 6. Hybrid ventilation energy savings estimatedfrom the climatic conditions

Heatingneeds

Ventilationneeds

Windowairing needs

Energysavings

�� ☺☺☺☺�� ☺☺☺��

�

�� ☺☺

��

��

� �� ☺

� : ca. 10 % of the occupancy time (To < 12 °C)☺: ca. 10 % of energy savings compared to amechanical ventilation mode like MV1

5. Conclusions

This study has shown the general behaviour of fanassisted natural ventilation with control based ontemperature and CO2 concentration. This approachprovides air flow on demand and, even if the meanCO2 concentration is a little higher with the hybridsystem than with the mechanical system, overallindoor air quality is improved since periods ofexposure to CO2 above 1000 ppm are reduced. Adrawback may be that the room air temperature is alittle higher and may be uncomfortable, especially inreally warm cities. This specific hybrid ventilationprovides some energy savings, but not as much as amechanical system with heat recovery, except forMediterranean cities. Heating and cooling degree-days give an initial idea about climate characteristicsand the resultant impact on indoor conditions.However degree-day data alone are not sufficient toestimate hybrid system performance. Hence moreinformation such as the annual outdoor temperaturedistribution is needed to complete the analysis sincethis gives better indications on overall room needs.

This type of analysis may be useful for designers atthe beginning of a project since it providesknowledge about the technical solution that may beappropriate for a specific building. Designers alsohave to decide between improving indoor air qualityand reducing energy consumption to as low aspossible. An interesting future work would be toextend these simulations to other specific climatessuch as humid, very cold or very warm climates.This would enable an atlas of suitability of hybridventilation performance to be developed.

Acknowledgements

This study has been carried out within theframework of the Annex 35 “Hybrid Ventilation inNew and Retrofitted Buildings” and was supportedby the French Ministry and also by the Frenchenvironmental and energy agency ADEME (AgenceDe l’Environnement et de la Maîtrise de l’Energie).The author would like to thank also the LBNL(Lawrence Berkeley National Laboratory, USA) andespecially Ashok Gadgil and Dimitri Curtil for theirkindness, availability and help

Nomenclature

cp specific heat [J/(kg°C)]e thickness of a layer [m]h heat transfer coefficient [W/(m2°C)]m mass flow rate [kg/s]P internal gains (heating system or

internal loads) [W]Q air flow rate [m3/s]R thermal resistance [°C/W]S surface [m2]Ses internal pollutant mass production [kg/s]t time [s]T temperature [°C] or [K]U wind speed [m/s]V volume of the zone [m3]

Greek Symbols

β angle between the wall and the horizontalplane [°]

∆P pressure difference [Pa]ε radiative emissivity [-]λ material conductivity [W/(m.°C)]Φ heat transfer at the surface of a wall [W]ρ specie density [kg/m3]σ Stephan-Boltzmann constant [W/(m2.K4)]

Subscripts

A node Aas pure airB node Bconv convectivees pollutantGLO long-waveheat heating systemi indoors or zone iij from the zone i to the zone jj zone jji from the zone j to the zone i


________________________________________________________________________________________________________________________

52

load internal loadsM node Mmet meteorologicalnet net radiative heat transfero outdoorsrad radiativerm mean radiativers dry resultants surfacev sky vault

References

Allard F. (1987) Contribution à l’étude destransferts de chaleur dans les cavités thermiquemententraînées à grand nombre de Rayleigh, Thèsed’Etat, INSA de Lyon, 1987.

Annex 35 (2000) CD with the State of the art ofHybrid Ventilation, IEA Annex 35, 2000.

Bourgeois D., Haghighat F. and Potvin A. (2002)“On the applicability of hybrid ventilation inCanadian office and educational buildings: Part 2 –Implementing Annex 35 pilots study projects inCanada”, Proceedings of the 4th InternationalForum on Hybrid Ventilation, Montreal, Canada,14-15 May 2002, pp24-29.

Cron F. and Inard C. (2002a) “Hybrid VentilationSimulation of Control Strategies – WG A1 input tofirst parameter study simulation in WG B7”,Technical report F7-081, IEA Annex 35, 2002.

Cron F., El Mankibi M., Inard C. and Michel P.(2000b) “Experimental and numerical analysis of ahybrid ventilated room”, Proceedings of the 8th

International Conference on Air Distribution inRooms, Copenhagen, Denmark, 8-11 September2002, pp493-496.

Delsante AE and Vik TA (2000) “State-of-the-ArtReport on Hybrid Ventilation”, IEA Annex 35report, 2000.

Delsante AE, Aggerholm S., Citterio M., Cron F., ElMankibi M. (2002a) “The Use of simulation tools toevaluate ventilation systems and control strategies”,Proceedings of the 4th International Forum onHybrid Ventilation, Montreal, Canada, 14-15 May2002, pp103-110.

Delsante AE, Aggerholm S. (2002b) “The use ofsimulation tools to evaluate hybrid ventilation

control strategies”, IEA Annex 35 Technical Report,2002.

El Mankibi M., Cron F., Michel P. and Inard C.(2002) “Control strategies for hybrid ventilationsimulations”, Proceedings of the EPIC 2002 AIVCConference, Lyon, France, 23-26 October 2002,pp437-442.

Ferries B. (1980) Contribution à l’étude desenveloppes climatiques et aide à leur conceptionpar micro-ordinateur, Thèse Doct.-Ing., INSA deLyon, 1980.

Feustel HE and Rayner-Hooson A., Ed (1990)“COMIS fundamentals”, Technical report LBNL-28560, Lawrence Berkeley National Laboratory,1990.

Frascatoro GV and Perino M. (2002) “Natural vs.Mechanical Ventilation – A Tool to Help Making aChoice”, International Journal of Ventilation, 1,No.2, pp101-108.

Jeong Y. and Haghighat F. (2002) “Modelling of aHybrid-Ventilated Building – Using ESP-r”,International Journal of Ventilation, 1, No.2,pp127-139.

Heiselberg P., Ed. (2002) Principles of HybridVentilation, IEA Energy Conservation in Buildingsand Community Systems Program Annex 35:Hybrid Ventilation in New and Retrofitted OfficeBuildings, 2002.

Ghiaus C. and Allard F. (2002) “Assessing climaticsuitability to natural ventilation by using global andsatellite climatic data”, Proceedings of the 8th

International Conference on Air Distribution inRooms, Copenhagen, Denmark, 8-11 September2002, pp625-628.

Rumianowski P., Brau J. and Roux JJ (1989) “Anadapted model for simulation of the interactionbetween a wall and the building heating system”,Thermal performance of the exterior envelopes ofbuildings IV Conference, Orlando, USA, 1989.

Spindler H., Glicksman L., Norford L. (2002) “Thepotential for natural and hybrid cooling strategies toreduce cooling energy consumption in the UnitedStates”, Proceedings of the 8th InternationalConference on Air Distribution in Rooms,Copenhagen, Denmark, 8-11 September 2002,pp517-520.

Numerical analysis of hybrid ventilation performance depending on climate characteristics

Documents

Transcript of Numerical analysis of hybrid ventilation performance depending on climate characteristics