lable at ScienceDirect

Building and Environment 46 (2011) 235e244

Contents lists avai

Building and Environment

journal homepage: www.elsevier .com/locate/bui ldenv

Experimental and numerical study of room airflow under stratum ventilation

Zhang Lin a,*, Lin Tian a, Ting Yao a, Qiuwang Wang b, T.T. Chowa

aBuilding Energy & Environmental Technology Research Unit, School of Energy and Environment & Division of Building Science and Technology, City University of Hong Kong,Hong Kong SAR, Chinab State Key Lab of Multiphase Flow in Power Engineering, Xi’an Jiaotong University, Xi’an 710049, China

a r t i c l e i n f o

Article history:Received 1 May 2010Received in revised form16 July 2010Accepted 20 July 2010

Keywords:Stratum ventilationExperimentsElevated indoor temperatureCFD model validation

* Corresponding author. Tel.: þ852 2788 9805; fax:E-mail address: bsjzl@cityu.edu.hk (Z. Lin).

0360-1323/$ e see front matter � 2010 Elsevier Ltd.doi:10.1016/j.buildenv.2010.07.018

a b s t r a c t

This paper investigates the air movement, air temperature profile and gaseous contaminant trans-portation in an individual office with stratum ventilation. The room temperature is elevated comparedwith conventional standards. The experimental investigation is carried out in an environmental chamberwith the presence of heat generating rectangles used to simulate an occupant and a computer.Measurements of temperature, velocity, and CO2 concentration are carried out for nine plumb lines in thechamber. Up to sixteen points are measured along each plumb line. The experimental data of theaforesaid three parameters of the individual office in warm condition under stratum ventilation arepresented. The experimental data collected are used to validate a re-normalization group (RNG) ke3turbulence model used for the warm condition. The agreements between the predicted values andexperimental results are acceptable, which demonstrates the feasibility of simulating indoor airflows atelevated room temperature under stratum ventilation by the RNG ke3 turbulence model.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

The earth is already showing many signs of worldwide climatechange. The Intergovernmental Panel on Climate Change (IPCC) hasused its strongest language to link human activities to the currentplanetary warming [1]. There is international consensus to reduceCO2 emission (Kyoto Protocol, etc.). Minimizing the energyconsumption by air conditioning systems would help to reduce CO2emission. Proactive actions in this regard have been taken byseveral governments in East Asia. The Electrical and MechanicalServices Department (EMSD) of the Hong Kong S.A.R. governmentissues guidelines to ensure that normally room temperature isadjusted to 25.5 �C in summer [2]. The National Development andReform Commission (NDRC) of the Chinese State Council issueda similar guideline to set the indoor temperature to 26 �C in thecooling season [3]. The room temperature in the “Office of Presi-dent” in Taipei has been set to 27 �C after Mr. Ma Ying-jeou’sinauguration in May 2008 [4]. Similarly, in the Standard of EnergyManagement issued in last December, the Ministry of Knowledgeand Economy of the Republic of Korea recommended that roomtemperature of a building should be ranging from 26 to 28 �C insummer [5]. In a more radical move, the Ministry of the Environ-ment (MoE) of Japan has been encouraging people to set thetemperature of air conditioners at offices to 28 �C during the

þ852 2788 9716.

All rights reserved.

summer months [6]. A survey from The International FacilityManagement Association shows that many facility professionalsare adjusting the thermostat to higher settings in the summer tocut energy consumption in the United States [7].

The aforementioned policies seem to be in line with the prin-ciple of sustainable development. But at least a question stillremains ‑ Would such a practice deteriorate indoor environmentalquality, especially thermal comfort? To provide an answer, ANSI/ASHRAE Standard 55-2004 has been updated with new provisionsthat allow elevated air movement to broadly offset the need to coolthe air in warm conditions [8]. Arens et al. reported that when theroom air temperature is above 22.5 �C, there is a small risk of draftand a strong preference for more air movement [9]. In revising ENISO 7730, Olesen adopted Fountain and Arens’ (1993) theory thathigher air speed was required to offset increased indoor tempera-ture [10,11]. Stratum ventilationwas proposed by Lin et al. for smallto medium rooms to cope with the elevated room temperatures[12,13]. This air distribution method works by creating a layer offresher air in occupants’ breathing zone. It delivers supply airdirectly to the occupants with little attention to the indoor airquality (IAQ) and thermal comfort of the upper zone(approximately > 1.5 m from the floor if the occupants aremostly sedentary). The lower zone (approximately < 0.8 m fromthe floor) is not the target zone of IAQ performance either. This isrealized by positioning supply terminal(s) at the side-walls orcolumns slightly above the height of occupants, standing orsitting depending on the application, whereas for displacement

Fig. 1. Configurations of individual office.

Table 1Internal heat sources in office.

Items Heat (W)

Human simulators 75Computer box 180Lamps 72 � 2

Fig. 2. Positions of measuring points (V ¼ velocity

Z. Lin et al. / Building and Environment 46 (2011) 235e244236

ventilation, air for breathing is transported by the boundary layeraround the body of an occupant and the air quality is a weightedaverage of the air quality in the room from the breathing level tothe floor level. Ventilation efficiency can be expected to increase ifair is supplied directly into the breathing zone. The reason is thatthe ventilation efficiency could be defined as

Et ¼ ðtout � tinÞ=ðtbz � tinÞwhere Et is ventilation efficiency based on temperature; tin is supplyair temperature, �C; tbz is breathing zone averaged temperature, �C;tout is exhausted air temperature, �C.

, T ¼ temperature and C ¼ CO2 respectively).

Table 2Source and diffusion coefficients for each f.

f Gf Sf

1 0 0uj mþ mt �vp

vxj� rbgjðT � T0Þ

k ðmþ mtÞ=dk G� r3þ GB

3 ðmþ mtÞ=33 ðC31G� C32r3þ C33GBÞ3=k� RT m=d1 þ mt=dt STC mt=Sct SC

Where m is laminar viscosity mt ¼ rCmk2

3 is turbulent viscosity G ¼ mtvuivxjðvuivxj

þ vuj

vxiÞ is

the turbulent production GB ¼ �gibmtPrt

vTvxi

is the turbulent production due to

buoyancy R ¼ Cmrh3ð1�h=h0Þ1þbh3

32

k is the source term from re-normalization h ¼ Sk3,S ¼ð2Si;jSi;jÞ1=2,Si;j ¼ 1

2ðvuivxj

þ vuj

vxiÞCm ¼ 0:0845;C31 ¼ 1:42;C32 ¼ 1:68,C33 ¼ 1:68 are the

model constants dk ¼ 0:7194;d3 ¼ 0:7194;d1 ¼ 0:7,dt ¼ 0:9,Sct ¼ 0:7are Prandtlor Schmidt number

Z. Lin et al. / Building and Environment 46 (2011) 235e244 237

Because cooler supply air directly flows into the breathing zone/head level, the value of tbz would be the lowest in the room. If thesupply air temperature tin and the airflow rate are identical formixing ventilation, displacement ventilation and stratum ventila-tion, the value of tbz for stratum ventilation would be the lowest ofthe three. Therefore, the ventilation efficiency would be increased.This has been experimentally proven [14]. Therefore, comparedwith mixing ventilation and displacement ventilation, stratumventilation may offer higher ventilation efficiency. The stability ofthe air layer is determined by the thermal length of the supply airjets [15]. For the room air distribution, initial velocity is the controlvariable. Therefore, the range of face velocity and the locations ofair supply terminal should be carefully optimized to break theboundary layer of an occupant and to minimize the risks of draft.The required height of the breathing layer/zone depends on thenature of occupancy. The supply air terminals should be positionedat the occupants’ head level. At the same time, a quasi-stagnantzone is also formed between the breathing zone and the floor(0 < height < 0.8 m). The temperature within the quasi-stagnantzone should be reasonably controlled which would not be lowerthan that of displacement ventilation. The problem of “cold ankles”could be solved. Energy is also saved by avoiding over-cooling of thelower zone of a room. Indoor air should be mixed well and airtemperature gradient in the occupied zone should be lower thanthat of displacement ventilation. Because air is supplied directly tobreathing zone, the air supply temperature should be higher thanthat of mixing ventilation. This implies that the evaporatingtemperature of the refrigerating plant may also be elevatedaccordingly, which would result in a desirable higher coefficient ofperformance (COP) of the refrigerating plant.

From the viewpoint of heat removal ability, stratum ventilation,which can be used in elevated indoor air temperature condition,

Fig. 3. Visulized flow pattern form

does not have to remove so much heat as that of mixing ventilationand displacement ventilation. The need to remove heat is mainly toprovide thermal comfort to occupants in a non-industrial room.Although the supply air temperature for stratum ventilation ishigher than that of conventional ventilations, the distance betweenthe occupants and the air terminal is shorter, which results in:

1) reversed temperature gradient in the occupied zone ‑ thetemperature is lower at the head-chest level (the breathingzone) and higher at the ankle level, which effectively cools thebody parts that need cooling the most;

2) higher air speed at the head-chest level for the same airflowrate, which offsets the effect of higher temperature in thermalcomfort; and

3) reduced system cooling load ‑ Because of Points 1) and 2), themean room temperature for equal thermal comfort is signifi-cantly higher than that of conventional ventilations. The heattransmission and fresh air heat gain therefore decrease.

For both stratum ventilation and task station ventilation, occu-pants stay in the flow of supply air jet(s). This significantly improvesthe inhaled air quality. However, current task station ventilationdoes not provide reversed temperature gradient in the occupiedzone for energy saving. Task station ventilation is not designed forserving mobile occupants. Also the ducting and air terminals aresimpler for stratum ventilation. When used in a medium-sizedroom (e.g. a 9� 9m2 room), multiple supply terminals or long slotsmay be used to realize stratum ventilation. Depending on thethermal length of the supply air jets, nature of occupancy andfurniture layout, air can be supplied from one or two directions. Forthe latter, air can come from the opposite directions or theperpendicular directions.

2. Research methodology

For the study of indoor environment under a specified ventila-tion mode, experimental investigation and computational fluiddynamics (CFD) are the two main research approaches. Measure-ment of the data related to indoor airflow and contaminant diffu-sion directly provides first-hand and reliable information, includingair temperature, air velocity and contaminant concentration.Moreover, the measured data can also be used to validate themathematical models to be applied to similar situations. Generally,an environment chamber is capable of minimizing the influence ofoutdoor environment on the measured indoor parameters, but it isquite expensive and its physical size is limited. During themeasuring process, the supply airflow rate, temperature and theenvelope temperature should remain steady, which is not an easytask due to the variation of outdoor environment. In order to get

ed under stratum ventilation.

Fig. 4. Velocity and tempearture fields at y ¼ 1.45 m.

Z. Lin et al. / Building and Environment 46 (2011) 235e244238

detailed information about the flow field and contaminant distri-bution, themeasuring points need to be up to certain numbers. Thisis time consuming.

On the other hand, the airflow can be determined by compu-tationally solving a set of conservation equations describing the

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.00.00.20.40.60.81.01.21.41.61.82.02.22.42.6

Measured Simulated

z(m

)

v(m/s)

Line 1 (x = 0.6 m, y = 1.45 m)

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.00.00.20.40.60.81.01.21.41.61.82.02.22.42.6

Measured Simulated

z(m

)

v(m/s)

Line 3 (x = 1.85 m, y = 1.45 m)

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.00.00.20.40.60.81.01.21.41.61.82.02.22.42.6

Measured Simulated

z(m

)

v(m/s)

Line 5 (x = 2.95 m, y = 1.45 m)

a b

c d

e f

Fig. 5. Measuredand simulated velocity

flow and energy in a system. A pioneering work on indoor flowsimulation by CFD was conducted by Nielsen more than threedecades ago [15]. Due to the limitations of the experimentalapproach and the increase in performance and affordability ofhigh-speed computers, numerical solution of these conservation

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.00.00.20.40.60.81.01.21.41.61.82.02.22.42.6

Measured Simulated

z(m

)

v(m/s)

Line 2 (x = 1.65 m, y = 1.45 m)

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.00.00.20.40.60.81.01.21.41.61.82.02.22.42.6

Measured Simulated

z(m

)

v(m/s)

Line 4 (x = 2.15 m, y = 1.45 m)

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.00.00.20.40.60.81.01.21.41.61.82.02.22.42.6

Measured Simulated

z(m

)

v(m/s)

Line 6 (x = 3.2 m, y = 1.45 m)

profiles for various positions (m/s).

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.00.00.20.40.60.81.01.21.41.61.82.02.22.42.6

Measured Simulated

z(m

)

v(m/s)

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.00.00.20.40.60.81.01.21.41.61.82.02.22.42.6

Measured Simulated

z(m

)

v(m/s)

Line 7 (x = 2.15 m, y = 1.15 m) Line 8 (x = 1.85 m, y = 1.15 m)

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.00.00.20.40.60.81.01.21.41.61.82.02.22.42.6

Measured Simulated

z(m

)

v(m/s)

Line 9 (x = 1.325 m, y = 0.78 m)

g h

i

Fig. 5. (continued).

Z. Lin et al. / Building and Environment 46 (2011) 235e244 239

equations provides a practical option for computing the airflowand contaminants distribution in buildings. Generally, indoorairflow is a three-dimensional turbulent flow which is driven bymomentum and buoyancy. Eight different types of turbulencemodels were investigated by Chen to determine the mostappropriate model for indoor airflow computations [16]. Heconcluded that the re-normalization group (RNG) ke3 model wasthe most accurate model among the eddy-viscosity models tested.Rouaud and Havet compared the performance of RNG ke3 modeland standard ke3 model in simulation of clean room, they foundthat RNG ke3 model was more suitable [17]. Gebremedhin andWu investigated the performance of five turbulent modelsincluding standard ke3 model, RNG ke3 model, the Low-Rey-nolds-Number ke3 model, keu model and Reynolds Stress Model(RSM) by commercial software Phoeincs. With convergenceand computation stability as the evaluating criteria, they discov-ered that RNG ke3 model outperformed all the other turbulentmodels for the simulation of airflow in a ventilated room [18].Posner’s research also demonstrated that the simulated resultsfrom RNG ke3 model were much closer to experimental datacompared with those of standard ke3 model when the modelswere used to simulate indoor airflow [19]. An RNG ke3 model wasdeveloped for stratum ventilation, but the data used to validatethis model were not specifically collected for stratum ventilation[20]. The RNG ke3 model used for current study has beenvalidated under conventional conditions (mean occupied zonetemperature ¼ 24.2 �C) [21,22].

The experimental data of stratumventilationwere collected andreported under conventional room conditions previously. Whileconducting research on stratum ventilation, it has become clear tothe authors recently that it would be more beneficial to applystratum ventilation for elevated room temperatures. This is the

reason why this report presents the experimental results ofairflows in an individual office under stratum ventilation in warmcondition (mean occupied zone temperature ¼ 26.4 �C) differentfrom the previous experiments. CFD is a powerful tool for us toexpand our research and offer detailed information with less timeand financial costs. However, it is essential to validate the CFDmodel before it can be applied with confidence. At higher indoor airtemperature (26.4 �C), the new conditions of

(1) the load and supply air temperature,(2) the interactions between the air jet and the room air (thermal

length, Archimedes number, etc.) and(3) the temperature difference across the external wall lead to

significantly different velocity and temperature fields. Thesame RNG ke3 turbulence model is therefore needed to be re-validated for varied internal and boundary conditions usingexperimental results acquired under new conditions.

3. Experimental arrangement

3.1. Environmental chamber and measuring instruments

The experiment is carried out in an environmental chamber atXi’an Jiaotong University. It consists of three rooms: a plant roomaccommodates the air handling unit that is capable of controllingthe air temperature, relative humidity and airflow rate; a testchamber in which the measurements are made, and an envelopechamber which is used to prevent the influence of outdoor envi-ronment on the test chamber. The test chamber is enclosed by theenvelope chamber. An air conditioner is installed for the envelopechamber to maintain the temperature fluctuation within

20 21 22 23 24 25 26 27 28 29 300.00.20.40.60.81.01.21.41.61.82.02.22.42.6

Measured Simulated

z(m

)

t(oC)

20 21 22 23 24 25 26 27 28 29 300.00.20.40.60.81.01.21.41.61.82.02.22.42.6

Measured Simulated

z(m

)

t(oC)

Line 1 (x = 0.6 m, y = 1.45 m) Line 2 ( x = 1.65 m, y = 1.45 m)

20 21 22 23 24 25 26 27 28 29 300.00.20.40.60.81.01.21.41.61.82.02.22.42.6

Measured Simulated

z(m

)

t(oC)

20 21 22 23 24 25 26 27 28 29 300.00.20.40.60.81.01.21.41.61.82.02.22.42.6

Measured Simulated

z(m

)

t(oC)

Line 3 (x = 1.85 m, y = 1.45 m) Line 4 (x = 2.15 m, y = 1.45 m)

20 21 22 23 24 25 26 27 28 29 30 31 320.00.20.40.60.81.01.21.41.61.82.02.22.42.6

Measured Simulated

z(m

)

t(oC)

20 21 22 23 24 25 26 27 28 29 300.00.20.40.60.81.01.21.41.61.82.02.22.42.6

Measured Simulated

z(m

)

t(oC)

a b

c d

e f

Line 5 (x = 2.95 m, y = 1.45 m) Line 6 (x = 3.2 m, y = 1.45 m)

Fig. 6. Measured and simulated temperature profiles for various positions (�C).

Z. Lin et al. / Building and Environment 46 (2011) 235e244240

reasonable range (�0.5 �C) during the experimental runs. The testchamber, 3.9 m � 2.9 m � 2.6 m, is ventilated via a 0.20 m � 0.17 mrectangular air grille located in the middle of right wall of the testchamber and a 0.60 m � 0.60 m perforated ceiling exhaust with15.3% as the effective area ratio, as shown in Fig. 1. The occupant inthe test chamber is simulated by a box, 0.25 m� 0.40 m� 1.20 mmand is heated by three 25 W light bulbs. One point source of CO2was introduced at the height of 1.10 m of the box to simulateexhalation from the occupant, with an initial velocity of 0.05 m/s inthe horizontal direction. A box, 0.40 m � 0.40 m � 0.40 m, locatedon the desk with three 60 W light bulbs, is used to simulatea personal computer [23,24]. Five rectangles, one as surface(0.68 m � 1.38 m � 0.06 m) with the top at the height of 0.76 m,two as drawer clusters (0.560 m � 0.380 m � 0.585 m), two as feet(0.430 m � 0.070 m � 0.115 m), and a thin vertical surface(0.52 m � 0.45 m) are used to construct the desk. Thetest chamber is illuminated by 2 sets of fluorescent lights,0.17 m � 1.24 m � 0.07 m, consuming a total of 2 � 72 W elec-trical power. A bookshelf, 0.40 m � 0.80 m � 1.85 m, is in the test

room. The ventilation rate is calibrated to be 5.5 air changesper hour (ACH), corresponding to the face velocity of 1.19 m/s at thesupply grille. This is realized by adjusting the damper of theHVAC system serving the test room. The volume flow rateis measured by an orifice flow meter. The supply temperature is21 �C, which is monitored by two thermocouples. Table 1summarizes the information on internal heat sources in the indi-vidual office.

A visualization system using smoke generator is used to showthe flow pattern formed by stratum ventilation. Once the smokegenerated by a theatre fog generator is injected into the testchamber, the flow pattern can be recorded by a camera. Thetemperature of smoke generated by theatre fog generator is farabove the temperature of the test room, so the smoke has to becooled by long tube before being injected into the test room toapproach neutral buoyancy.

The SWEMA hot-wire anemometer system is used to measureair velocity and temperature. For velocity, the measuring range is0.05 m/s to 3 m/s, the measuring error is �0.03 m/s for 0.05 m/s to

20 21 22 23 24 25 26 27 28 29 300.00.20.40.60.81.01.21.41.61.82.02.22.42.6

Measured Simulated

z(m

)

t( oC)

20 21 22 23 24 25 26 27 28 29 300.00.20.40.60.81.01.21.41.61.82.02.22.42.6

Measured Simulated

z(m

)

t( oC)

Line 7 (x = 2.15 m, y = 1.15 m) Line 8 (x = 1.85 m, y = 1.15 m)

20 21 22 23 24 25 26 27 28 29 300.00.20.40.60.81.01.21.41.61.82.02.22.42.6

Measured Simulated

z(m

)

t( oC)

Line 9 (x = 1.325 m, y = 0.78 m)

g h

i

Fig. 6. (continued).

Z. Lin et al. / Building and Environment 46 (2011) 235e244 241

1.00 m/s and �3% of readings for 1.00 m/s to 3.00 m/s, the dynamicresponding time is 0.2 s. The measuring error for air temperature is�0.3 �C. Copper-constantan thermocouples were used to measureair temperature and wall surface temperature of the test chamber.The measuring error for the system is 0.3 �C.

The multi-point non-dispersive infrared CO2 systemwas used tomeasure the concentration of CO2. The system contains 10 sensors.Two sensors are used to measure CO2 concentrations of the supplyair and exhaust air respectively. The measuring range is0e2000 ppm. The measuring error is �20 ppm þ 3% of readings.The dynamic responding time is less than 60 s.

3.2. Test procedure

Measurements are taken for steady state condition. When thetemperature, velocity and CO2 concentration of the test room arealmost unchanged, a steady state condition is believed to beachieved. This process would almost take more than 12 h afterthe HVAC system starts to work [23,24]. It takes one to 2 h forthe test chamber to reach steady state condition again after re-location of a probe-holding pole [24]. Fig. 1 shows the configu-ration of the test room with stratum ventilation. Due to thelimited number of anemometer and CO2 concentration sensor,the temperature, velocity and concentration along nine plumblines in the test chamber are measured line by line by shiftingthe probe-holding pole with hot-wire anemometers, thermo-couples and CO2 sensor. The locations of the plumb lines areshown in Fig. 2. For Plumb Lines 1, 2, 5, 6 and 9, the probe-holding pole used is almost as long as the height of the room.For Plumb Lines 3, 4, 7 and 8, the pole is erected on the desk,and is therefore shorter.

4. CFD modelling

In this study, an RNG ke3 model based on the commercialprogram of FLUENT is used to perform the analysis [25]. Generalgoverning equations for the RNG ke3 model are as follows [26]:

divðrVfÞ ¼ div�Gf grad f

�þ Sf (1)

For steady flow where f represents each of the three velocitycomponents u, v, w, k is the kinetic energy of turbulence, 3 is thedissipation rate of the kinetic energy of turbulence, T is the airtemperature, c is the CO2 concentration, Gf is the effective diffusioncoefficient and Sf is the source term of the general equation. Thesource term and the diffusion coefficient corresponding to eachvariable f in this study are given in Table 2. The velocity and itscomponents and the temperature affect thermal comfort andenergy consumption. The CO2 concentration is closely related toIAQ.

The equations are discretized into algebraic equations by thesecond-order upwind scheme, which couples pressure and velocityflow fields. The Boussinesq model is employed to consider thebuoyancy effect. The air inlet is defined as an opening with uniformairflow velocity. The outlet boundary conditions are set as outflowof room air. The walls provide heat sources with constant heat flux.The discrete ordinates (DO) radiation model is used to simulateradiation of the walls [25,27]. The standard wall function is used todescribe the turbulent flow properties in the near wall region. Moredetails could be found from Launder and Spalding [28].

Because the existence of the lamps, bookshelf, dummy,computer and desk plate, the geometrical configuration of flowregion is quite complicated. Therefore, unstructured grid is adoptedfor the simulation of the flow field. The grid numbers is 587344,which is checked to be grid independent. The converged residuals

400 500 600 700 800 900 10000.00.20.40.60.81.01.21.41.61.82.02.22.42.6

z(m

)

c(ppm)

Measured Simulated

400 500 600 700 800 900 10000.00.20.40.60.81.01.21.41.61.82.02.22.42.6

z(m

)

c(ppm)

Measured Simulated

Line 1 (x = 0.6 m, y = 1.45 m) Line 2 (x = 1.65 m, y = 1.45 m)

400 500 600 700 800 900 10000.00.20.40.60.81.01.21.41.61.82.02.22.42.6

z(m

)

c(ppm)

Measured Simulated

400 500 600 700 800 900 10000.00.20.40.60.81.01.21.41.61.82.02.22.42.6

z(m

)

c(ppm)

Measured Simulated

Line 3 (x = 1.85 m, y = 1.45 m) Line 4 (x = 2.15 m, y = 1.45 m)

400 500 600 700 800 900 10000.00.20.40.60.81.01.21.41.61.82.02.22.42.6

z(m

)

c(ppm)

Measured Simulated

400 500 600 700 800 900 10000.00.20.40.60.81.01.21.41.61.82.02.22.42.6

z(m

)

c(ppm)

Measured Simulated

Line 5 (x = 2.95 m, y = 1.45 m) Line 6 (x = 3.2 m, y = 1.45 m)

a b

c d

e f

Fig. 7. Measured and simulated CO2 concentration profiles for various positions (ppm).

Z. Lin et al. / Building and Environment 46 (2011) 235e244242

for continuity equation, uj, k and 3 are less than 10�3. The convergedresidual for T and c is 10�6. During the computation, the inlet flowrate and outlet flow rate are also monitored. The difference is lessthan 10�6, which also demonstrates the convergence of thecomputation.

5. Results and discussions

Fig. 3 shows the flowpattern of stratumventilation visualised bysmoke. Colour smoke is not used due to the density difference. Theeffort has been devoted to make the temperature of the smoke asclose to that of the supply air as practically possible. Fig. 4 showsthe computed velocity and temperature fields at the section ofy ¼ 1.45 m. Both the experimental and numerical results show thatthe cooler air stream from the supply grille, while its jet axis dipsa bit along the way, spreads into the breathing zone at the head-echest level (the layer between 0.8 m and 1.2 m above floor level)in front of the occupant and then gradually flows downward due tobuoyancy, which enables the air quality of the breathing zone to bebetter than the other regions. Because of the existence of occupant

and desk, several eddies are formed. In the vicinities of the occu-pant, plumes form due to natural convection. The temperature inthe breathing zonewhere the air streampasses is lower than that inthe other regions. Along the axis of the supply air jet, the distancesbetween the grille and the measuring positions are in ascendingorder from Plumb Line 6 to Plumb Line 1 (Fig. 6aef). The charac-teristic curvity (at the breathing level) of the temperature profilesof stratum ventilation is descending due to entrainment of room airby the supply airflow. The temperature in the occupied zone islower than that in the upper zone. The temperaure gradient in thelower zone (within the height of 0.8m from the floor) is quite small,which implies that the air in this zone is either well mixed orstagnant. All the phenomena observed match the characteristics ofstratum ventilation discussed in the Introduction section.

Both measured and computed values of velocity, temperatureand CO2 concentration for measuring points along Plumb Lines 1e9are shown in Figs. 5e7. In terms of velocity, the accuracy is such thatthe largest errors are 0.18 m/s for both the cases of simulationunderestimating and overestimating the experiment. However theexperimental results are themselves limited by the accuracy of the

400 500 600 700 800 900 10000.00.20.40.60.81.01.21.41.61.82.02.22.42.6

z(m

)

c(ppm)

Measured Simulated

400 500 600 700 800 900 10000.00.20.40.60.81.01.21.41.61.82.02.22.42.6

z(m

)

c(ppm)

Measured Simulated

Line 7 (x = 2.15 m, y = 1.15 m) Line 8 (x = 1.85 m, y = 1.15 m)

400 500 600 700 800 900 10000.00.20.40.60.81.01.21.41.61.82.02.22.42.6

z(m

)

c(ppm)

Measured Simulated

Line 9 (x = 1.325 m, y = 0.78 m)

g h

i

Fig. 7. (continued).

Z. Lin et al. / Building and Environment 46 (2011) 235e244 243

hot wire anemometer at low velocities. In terms of temperature,the computed results match the experimental results well except inthe big void under the desk that complicates the airflows. At theopen edge of the void (x¼ 1.65m, y¼ 1.45m, z¼ 0.8m), the error is2.5 �C with the simulation overestimating the experimental result.The discrepancies between measured and simulated CO2 concen-trations are significant at some locations, which are probably due tothe difference between the actual background (outdoor) CO2concentration during the experiments and the value used in thesimulations. The background CO2 concentration in Xi’an is fluctu-ating during the measurement of the nine plumb lines. An exper-imental run lasts for nearly 30 h. For simulation of CO2concentration distribution in the room, the mean background CO2concentration for the 30 h is adopted as the value in the supply air.

6. Conclusions

The experimental data of temperature, velocity and CO2concentration of an individual office in warm condition understratum ventilation are presented in this paper, which demonstratethe characteristics of stratum ventilation. The simulated resultsfrom the RNG ke3 turbulence model are in acceptable agreementwith the experimental data, which shows that the current modelcan be used to execute further numerical studies of stratumventilation.

Acknowledgements

The work described in this paper is supported by a grant fromthe Natural Science Foundation of China (Grant No. 50821604) andby a Strategic Research Grant of the City University of Hong Kong,Hong Kong Special Administrative Region, China (Project No.7002481).

References

[1] Pachauri RK. IPCC Director-general, the IPCC Fourth Assessment workinggroup reports: key findings. New York City: United Nations Headquarters;24th September 2007.

[2] Electrical and Mechanical Services Department (EMSD). Government of HongKong S.A.R., Energy Conservation Charter 2006 e suitable room temperature,http://www.energyland.emsd.gov.hk/blue_sky/eng/index.htm.

[3] National Development and Reform Commission (NDRC). The State Council ofChina, Citizens’ pact for deed of thrift, <http://www.ndrc.gov.cn/gzdt/t20051223_54462.htm>.

[4] NOW news, Taiwan. Campaign in office of president for saving energy andreducing carbon emissions, <http://www.nownews.com/2008/05/24/10844-227 9584.htm>.

[5] Ministry of Knowledge and Economy. The Republic of Korea. The standard ofenergy management, http://www.mke.go.kr/info/law/gosiView.jsp?; December29, 2008. Item No. 2008-219.

[6] Japanese Ministry of the Environment (MoE). National campaign to fightglobal warming, http://www.env.go.jp/en/press/2005/0428b.html.

[7] International Facility Management Association. Temperature wars e savingsvs. comfort, http://www.ifma.org; July 2009.

[8] ASHRAE. Thermal environmental conditions for human occupancy. , Atlanta:American Society of Heating, Refrigerating, and Airconditioning Engineers,Inc; 2008. Addenda d, e, f, and g to ANSI/ASHRAE Standard 55-2004.

[9] Arens E, Turner S, Zhang H, Paliaga G. Moving air for comfort. AmericanSociety of Heating, Refrigerating, and Airconditioning Engineers Journal2009;51(5):18e28.

[10] Olesen BW. Introduction to the new revised draft of EN ISO 7730: Wirsbo-VELTA GmbH, Norderstedt/Germany.

[11] Fountain ME, Arens EA. Air movement and thermal comfort. American Societyof Heating, Refrigerating, and Airconditioning Engineers Journal 1993;35(8):26e30.

[12] Lin Z, Chow TT, Tsang CF. Stratum ventilation? A conceptual introduction. In:Proceedings of the tenth international conference on indoor air quality andclimate, Beijing, China; 4e9 September 2005. p. 3260e4.

[13] Lin Z, Chow TT, Tsang CF, Fang KF, Chan LS. Stratum ventilation e a potentialsolution to elevated indoor temperature. Building and Environment 2009;44(11):2256e69.

[14] Lin Z, Tian L, Wang Q, Chow TT. Experimental investigation of thermal andventilation performances of stratum ventilation, manuscript submitted toBuilding and Environment.

[15] Nielsen PV. Flow in air-conditioned rooms, Ph.D. thesis. Technical Universityof Denmark: Copenhagen; 1974.

Z. Lin et al. / Building and Environment 46 (2011) 235e244244

[16] Chen Q. Comparison of different k � 3 models for indoor airflow computa-tions. Numerical Heat Transfer 1995;28(Part B):353e69.

[17] Rouaud O, Havet M. Computation of airflow in a pilot scale clean room usingturbulence models for indoor airflow computations, Part B. Fundamentals OfNumerical Heat Transfer 1999;28(3):353e69.

[18] Gebremedhin KG, Wu B. Characterization of flow field in a ventilated spaceand simulation of heat exchange between cows and their environment.Journal of Thermal Biology 2003;28(4):301e19.

[19] Posner JD, Buchanan CR, Dunn-Rankin D. Measurement and prediction ofindoor air flow in a model room. Energy and Buildings 2003;35(5):515e26.

[20] Lin Z, Chow TT, Tsang CF. Validation of CFD model for research into stratumventilation. International Journal of Ventilation 2006;5(3):345e63.

[21] Tian L, Lin Z, Wang QW, Liu J. Numerical investigation of indoor aerosolparticle dispersion under stratum ventilation and under displacement venti-lation. Indoor and Built Environment 2009;18(4):360e75.

[22] Tian L, Lin Z, Wang QW. Comparison of gaseous contaminant diffusion understratum ventilation and under displacement ventilation. Building and Envi-ronment 2010;45(9):2035e46.

[23] Yuan X, Chen Q, Glicksman LR. Measurements and computations of room airflow with displacement ventilation. American Society of Heating, Refriger-ating, and Airconditioning Engineers Trans 1999;105(1):340e52.

[24] He G. Modeling indoor pollutant exposures under different ventilationschemes, Ph.D. thesis. University of Miami, USA; 2003.p. 215.

[25] FLUENT. Fluent 6.2 user’s guide. Lebanon, NH: Fluent Inc; 2005.[26] Choudhury D. Introduction to the renormalization group method and turbu-

lence modeling. Fluent Inc; 1993. Technical Memorandum TM-107.[27] Abanto J, Barrero D, Reggio M, Ozell B. Airflow modeling in a computer room.

Building and Environment 2004;39(12):1393e402.[28] Launder BE, Spalding DB. The numerical computation of turbulent flows.

Computer Methods in Applied Mechanics and Energy 1974;3:268e389.