Building and Environment 59 (2013) 336e348

Contents lists available at SciVerse ScienceDirect

Building and Environment

journal homepage: www.elsevier.com/locate/buildenv

Desk fans for the control of the convection ow around occupants using ceiling
mounted personalized ventilation
A. Makhoul*, K. Ghali, N. Ghaddar
Department of Mechanical Engineering, American University of Beirut, P.O. Box 11-0236, Beirut 1107-2020, Lebanon

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 13 July 2012
Received in revised form
29 August 2012
Accepted 29 August 2012

This study aims at studying by modeling and experimentation the performance enhancement of ceilingmounted personalized ventilation (PV) nozzle when assisted by small desk-mounted fans to reduce the
effect of thermal plume generated by the occupant. Detailed computational uid dynamics CFD simulations were performed to study the ow, thermal, and CO2 concentration elds around an occupant
using a single jet ceiling-mounted PV nozzle in the conditioned space while the fans are on. The CFD
model was integrated with a bioheat model to determine the corresponding human body segmental heat
uxes and convective plumes for predicting the occupant segmental and overall comfort. The CFD model
was validated with experiments using a cylindrical heat source and comparing measured velocity,
temperature and CO2 concentrations in the vicinity of the nozzle and the cylinder. Good agreement is
found between the measured and the numerically predicted values. The performance of the PV nozzle
assisted with desk fans is assessed by evaluation of ventilation effectiveness in the occupant microclimate and comfort conditions and results are compared with single jet PV and co-axial jet PV, both
without the use of desk fans. The desk-mounted fans were able to reduce the convection plumes around
the occupant and improved the performance of the single jet PV nozzle by doubling the ventilation
effectiveness and improving comfort. They permitted also to achieve a reduced energy saving by up to
13% when compared with conventional mixing ventilation systems.
2012 Elsevier Ltd. All rights reserved.

Keywords:
Personalized ventilation
Air quality
Thermal comfort
Energy efciency

1. Introduction
Means of reducing cooling requirements of spaces includes
localizing the ow around the occupant using variable angle
diffuser with central return vent as suggested by Lo et al. [1]. The
localizing ventilation system will permit to conne the conditioned
air in a canopy around each occupant separately thus allowing for
considerable energy savings when ofce spaces are partially
occupied. Although the mentioned system allows for reducing the
cooling requirements of ofce spaces, it does not provide the
occupants with effective fresh air delivery for enhanced indoor air
quality. For improved ventilation effectiveness, researchers suggested the usage of personalized ventilation (PV) modules with
adjustable ow rates and temperatures to meet the occupants
comfort needs [2,3]. Rather than installing the PV modules on ofce
desks as done conventionally, Yang [4] suggested combining the PV
devices with mixing ventilation systems using ceiling-mounted PV
for ofces in tropical climates. The nozzle was mounted in the

* Corresponding author. Tel.: 961 1 350000; fax: 961 1 744462.

E-mail address: ahm20@aub.edu.lb (A. Makhoul).
0360-1323/$ e see front matter 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.buildenv.2012.08.031

middle of a diffuser with conventional functionality. This design

permits elimination of the need to extend ducting system of the
task ventilators into the occupied zone, i.e. the raised oor system is
not required. The performance of ceiling PV (thermal comfort, air
quality and energy consumption) was studied by Yang et al. [5] who
reported 5% energy savings (for an occupant located directly
beneath a PV nozzle delivering 4.5 l/s of fresh air) when compared
to a mixed ventilation system that maintains an average room
temperature of 23.5
C and uses the same fresh air quantities of
4.5 l/s [5]. In spite of this limitation, the study gives insight into the
feasibility of using ceiling PV nozzles and into the effect of the
amount of fresh air on the operational energy savings.
For the ceiling PV nozzle to be effective, the fresh air of the PV jet
should reach the occupants breathing zone with minimal
entrainment of contaminated room air and it should have sufcient
terminal velocity to penetrate the occupants free convective
plumes. In ceiling mounted jets, the shear stress emanating
between the PV fresh air nozzle and the surrounding room stagnant
air reduces the effective delivery of fresh air to the occupants
breathing zone. The jet entrainment and mixing problems can be
circumvented by using a coaxial nozzle (primary nozzle supplying
fresh air surrounded by a concentric secondary nozzle supplying

A. Makhoul et al. / Building and Environment 59 (2013) 336e348

Nomenclature

Symbols
AQI
AC
C
D
D
IAQ
PV
t
y

air quality index

alternative current
CO2 concentration (ppmv)
diameter of the honeycomb holes (m)
diffusion of CO2 (m2/s)
indoor air quality
personalized ventilation
thickness of the honeycomb ow straighteners (m)
dimensionless wall distance

Greek symbols
3
effectiveness
r
density of CO2 (kg/m3)
m
viscosity of air (kg/m$s)
Subscripts
b
breathing zone
f
fresh air
r
return vent
t
turbulent
v
ventilation

recirculated air at the same velocity). Khalifa et al. [6,7] used coaxial horizontal PV jet and showed improvement in ventilation
effectiveness because of superior jet entrainment into the
breathing zone. This coaxial nozzle was initially intended to be
mounted on ofce desks for supplying air horizontally to the
breathing zone of the occupant. However, Makhoul et al. [8]
investigated the performance of the low-mixing coaxial nozzle
mounted in the ceiling and supplying vertical fresh air jet to the
occupant breathing zone. The nozzle was integrated with a diffuser
having a variable supply angle for ow localization. The effect of the
combined nozzle-diffuser on the energy consumption and the
occupants induced thermal comfort was assessed. Makhoul et al.
[8] reported that the coaxial nozzle achieved 3e4 times better
ventilation effectiveness when compared to a conventional
nozzle delivering the same amount of clean air in task ventilation
system (TVS) applications while attaining notable comfort levels
and achieving substantial energy savings up to 30%. Makhoul
et al. [8] system, although provides excellent ventilation effectiveness, it is more complex to design, operate, and control since slight
change in axial velocities may result in major change in performance. It is of interest to nd other means by which the simple PV
jet system of Yang et al. [4,5] can be modied without major
changes in the system design. The constraining parameter in the
design of Yang et al. [5] single jet PV stems from the ability of the jet
to penetrate down the rising occupant plume. If there is a way to
reduce the upward plume ow, then better ventilation effectiveness might be achieved near the occupant.
Bolashikov et al. [9,10] have used desk mounted fans to improve
the effectiveness of personalized ventilation devices by control of
convective ow around the occupant using desk mounted suction
fan to block free convection and hence enhance the penetration of
the horizontal PV supplied air to the face. The active control of
thermal plume used by Bolashikov et al. [9,10] made it possible to
provide 90% of clean air to the breathing zone of the occupant and
reduced PV supply ow rate. No work has been reported in the
literature to address the active control of free convection ow
above the occupant and at the occupant breathing zone when

337

ventilation is done via ceiling mounted single jet PV placed at the

center of the supply diffuser that localizes the ow around the
occupant forming a canopy. The combination of active method to
control thermal plume around the occupant and simple vertical PV
jet may result in improved performance.
The objective of this work is to evaluate the enhancement in
performance of the ceiling mounted single jet PV nozzle placed at
the center of supply diffuser when assisted by desk fans that control
the natural convection plumes generated by the human body. The
peripheral supply diffuser will have an added functionality over
and above the conventional diffuser that permits changing the
supply angle for better localization of the ow around the occupant.
The conditioned space and the occupant microclimate velocity,
temperature and CO2 concentration elds will be obtained from 3D CFD simulation model integrated with a bioheat model to
determine human segmental heat generation and the local and
overall comfort. Experiments will be performed on a heated
cylinder to validate thermal and ow eld and air quality around
the cylinder as measured by CO2 concentrations. This will be followed by a parametric study to determine the best combination of
jet ow rate and supply temperature to evaluate the extent of
mitigation of the occupant thermal plume ow rate and the
effectiveness of fresh air delivery when desk mounted fans are
used. The simple conguration of ceiling mounted single jet PV
nozzle assisted with desk-mounted fans will be compared with the
performance of the PV system when the fans are off similar to Yang
et al. [4,5] PV system and with the relatively complex system of
ceiling-mounted coaxial nozzles of Makhoul et al. [8]. The
comparison will be made on the bases of thermal comfort and
indoor air quality. In addition, at nearly equal comfort and air
quality, the performance of the systems will be assessed in terms of
energy savings and effective delivery of fresh air.
2. Methods
The objective of the proposed desk-mounted fans combined
with a ceiling-mounted personalized ventilator is to control the
convection ow resulting from the rising thermal plumes generated by the human body. This will enhance the performance of the
PV nozzle that supplies cool fresh air downwards at acceptable
velocity to reach the occupant providing individual thermal
comfort and good air quality zone. Fig. 1 shows schematics of (a) the
ceiling nozzle and diffuser and (b) the conditioned workspace with
the occupant in seated position below the diffuser. The air distribution system is composed of a primary central nozzle for fresh air
delivery and a peripheral angled diffuser to form a canopy for
localizing the ow around the occupant and maintain the room
macroclimate temperature. The ow localization technique
through the usage of the angled peripheral diffuser was suggested
by Makhoul et al. [8]. This will allow maintaining a desired air
temperature in the occupants microclimate that is lower than the
temperature attained in the macroclimate region that is away from
the human body. In addition to the ow localization, a fan delivering a nominal ow rate of 10 L/s is mounted under the desk to
control the convection plumes around the occupant.
To assess the performance of the proposed air distribution
system in providing localized thermal comfort and air quality to
occupants, 3-D detailed numerical simulations will be performed
using commercial software such as ANSYS Fluent [11] to obtain ow
and thermal elds in addition to concentration eld of carbon
dioxide for a typical ofce space. To determine the appropriate ow
and thermal parameters of the single jet nozzle that are capable in
providing the desired comfort and air quality, the bioheat model of
Salloum et al. [12] and the comfort models of Zhang et al. [13e16]
will be integrated with the CFD code to predict the human

338

A. Makhoul et al. / Building and Environment 59 (2013) 336e348

Fig. 1. (a) Frontal and top views of the proposed ceiling nozzle and (b) computational domain.

segmental heat ux conditions for the CFD model and segmental

and overall thermal sensation and comfort. The ndings of the
numerical model will be validated by experimentation on a thermal
cylindrical manikin constructed of a surface heated cylinder placed
inside a controlled climatic laboratory equipped with the retrotted fresh air distribution system to measure the velocity,
temperature and CO2 concentration at specied positions near the
cylinder and the ceiling nozzle in room. The experimental results
will also be used to determine the effectiveness of the proposed air
distribution system. The validated numerical models will then be
used to perform a parametric study using actual human body
model to determine best performance operating parameters of the
diffuser and the single nozzle jet ow rates, temperatures, and
diffuser angle.
Since this study is targeting the IAQ in non-residential spaces,
a full-scale computational domain representing a typical ofce
chamber measuring 3.4 m  3.4 m  2.6 m will be considered (see
Fig. 1b). The seated thermally heated manikin is included in the
middle of the room in the workspace. The manikin has a height of
1.2 m and a total surface area of 1.78 m2 and the dimensions of each

segment will be determined according to the data provided by

Gagge et al. [17]. The ceiling mounted nozzle integrated with a slot
diffuser is centered over the manikins head in addition to the two
symmetrically installed return vents and exhausts. A desktop
computer generating 93W [18] was placed on a desk in front of the
manikin. In addition, a lighting load of 12 W/m2 will be used
according to the recommendations of ASHRAE standard 90.1-2007
[19]. The remaining load is assumed distributed equally as heat ux
through the walls so that the total room load is around 60 W/m2 of
room area.
2.1. Airow modeling
The usage of a detailed CFD model that can accurately predict
the entrainment and mixing between the ceiling-mounted PV
nozzle and the surrounding air is important in order to capture the
interaction between the nozzle jet and the rising thermal plume
from the manikin representing the human body. Consequently,
proper modeling of ow physics is critical: turbulence models,
buoyancy effects, and boundary layers. Besides, controlling the grid

A. Makhoul et al. / Building and Environment 59 (2013) 336e348

resolution in some critical areas is essential to capture accurately

the shear-layer entrainment process and the different thermal
plumes as well as the uid/thermal boundary layers around the
heated manikin. The commercial CFD solver, ANSYS Fluent [11], is
used for numerical modeling to solve for the airow, thermal, and
species concentration elds in the room. By making use of the
symmetry plane that cuts through the center of the manikin,
nozzles and return vents as shown in Fig. 1, only one-half of the
room is considered for numerical modeling.
To create a CFD model that can capture the air jet properties
accurately, several methods are considered in the literature
among which we cite the box and momentum methods that rely
on describing the diffusers resulting airow [20]. Since the
primary concern of this study was the resulting airow delivered by the diffuser away from its outlet, the momentum
method was selected. Besides, the velocity prole and turbulent
intensity of the nozzle are of great importance to model the
mixing rate between the jet and the stagnant room air. These
velocity and turbulence intensity data cannot not be obtained
unless some experimental measurements of the ow characteristics at the outlet of the nozzle shown in Fig. 1a are performed. These values will be measured experimentally as will be
explained later in section 2.3. Thus, the experimentally obtained
nozzle jet characteristics will provide the necessary boundary
conditions or input to the CFD model and they serve to evaluate
its accuracy.
Once the nozzles jet characteristics are determined experimentally, the nozzle is included in a computational domain that
represents a typical ofce. A tetrahedral unstructured grid is
generated using different element sizing for the boundary faces and
element size of 5 mm is chosen for the nozzles resulting in 9000
elements. In order to resolve the boundary layer, the surface grid for
the manikin is created with elements of 1 cm. The total number of
elements on the half surface of the manikin is around 20,000 as
shown in Fig. 2. An ination boundary layer was created around the
manikin with a rst layer thickness of 1.5 mm, a growth rate of 1.2
and total of 4 layers. The corresponding y values ranged between
0.8 and 4 on the manikins surface.

339

In order to capture the entrainment process between the fresh

air jet and the polluted surrounding air, a clustered grid at the
interface of the jet and the stagnant air would be necessary.
However, since buoyancy effects are involved and jet deections
can occur, the trajectory of the jet is unpredictable. For this reason,
four spheres of inuence of 1 m diameter were integrated between
the nozzles in the ceiling and the thermal manikin: the grid is
clustered in this region and the elements size is chosen to be 2 cm.
For the rest of the domain, an unstructured grid was used which
resulted in a total number of cells of 1,200,000, Fig. 2. It is important
to note that the total number of cells is controlled by the grid size in
the mixing region and not by the manikins surface grid or the size
of the domain.
Boundary conditions employed in the CFD model were similar
to the conditions obtained in the experimental setup. A velocity
prole obtained experimentally will be used for the nozzle jet with
the corresponding airow rate. The velocities are assumed normal
to the boundary for the PV nozzle, the peripheral diffuser had an
angled velocity with respect to the normal vector, and a pressure
outlet was used at the return vent with a value of zero gage. The
fan boundary condition was used at the inlet of desk-mounted fan
with a predened pressure jump in order to obtain the desired ow
rate. Heat uxes are applied at the boundary of the walls, lighting,
and computer while the manikin surface temperatures are updated
from the Bioheat model [12] at each iteration using the method
initially introduced by Makhoul et al. [21].
The CFD model is coupled with the Bioheat model [12] to assess
the thermal comfort of the occupant and to get the segmental skin
temperatures of the different body segments that in turn inuence
the rising thermal plume, Fig. 3. The model divides the body into 11
segments (head, chest, back, abdomen, buttocks, upper arm, lower
arm, hands, thighs, calves and feet) to simulate the physiological
responses of each body segment for predicting its thermal state of
comfort based on Zhangs model [13e16] in a high transient
thermal environment. This interdependence between Fluent and
the Bioheat model is important to capture accurately the ow
characteristics and the human thermal response. Initial segmental
skin temperatures are estimated for the manikin and the

Fig. 2. Generated grid at the symmetry plane.

340

A. Makhoul et al. / Building and Environment 59 (2013) 336e348

Fig. 3. Simulation models diagram.

corresponding convection coefcient and air temperature taken

from Fluent are used to update the skin temperatures using the
body model, as described by Cropper et al. [22]. This procedure is
repeated until convergence was achieved and the corresponding
results are considered converged.
The commercial solver, ANSYS Fluent [11] used to solve the
NaviereStokes equations has several options for the turbulence
model. Among them, the Realizable ke3 model is selected along
with the enhanced wall treatment and the full buoyancy effect
options. One of the main advantages of the Realizable ke3 model is
that it predicts more accurately the spreading rate of planar and
round jets and that it provides superior prediction for ows
involving separation and recirculation. The enhanced wall treatment option will allow switching between the two-layer model and
the enhanced wall functions according to the grid size. In regions
where y is of the order of unity, the two-layer model is used and
the laminar sublayer is resolved. However, in regions where the
boundary layer is not fully resolved (3 < y < 10), an enhanced wall
function is used. Besides, the full buoyancy effect option enables to
include the turbulence generation due to buoyancy on the 3 term.
The incompressible ideal gas law was used for the density
modeling in order to account for the buoyancy effects. The species
transport equation is also enabled to compute the concentration of
the CO2 and assess the indoor air quality (IAQ) in the domain. In
addition, a second-order upwind discretization scheme is
employed to solve for the momentum, k, 3 , and energy equation as
well as the species transport equation, and the PRESTO! scheme
is used for the pressure. Numerical convergence was judged using
several criteria: Scaled residuals reported by Fluent should reach
a minimum value of 105, but convergence was not reached until
the net heat ux of the computational domain was below 1% of the
total heat gain and the CO2 concentration in the breathing zone is
stabilized.
2.2. Air quality modeling
To investigate the level of mixing between the supplied fresh air
and the stagnant contaminated room air the ventilation

effectiveness is assessed by using CO2 as a tracer gas. The fresh air

nozzle in the ceiling has a CO2 concentration of 400 ppm which is
the same concentration found in outdoor clean air. The amount of
fresh air supplied to the room is equal to the amount of room air
exhausted thus preventing the build-up of CO2 in the space. The
ventilation effectiveness, also known as the air quality index (AQI),
at the breathing zone is dened as
3v

Cr  Cb
Cr  Cf

(1)

where Cr is the CO2 concentration at the return vent, Cb is the CO2

concentration at the breathing level, and Cf is the fresh air nozzle
CO2 concentration. The higher the ventilation effectiveness is, the
better the air quality would be at the breathing zone. The breathing
zone CO2 concentration is dened as the volume average concentration in the spherical volume of 2 cm diameter placed at 2.5 cm
from the manikins nose.
The species transport equation is activated to compute the
concentration of the pollutants and their distribution in the room.
The turbulent Schmidt number, which is dened as the ratio of the
turbulent viscosity of air mt to the product of the density and
turbulent diffusion of CO2 rDt in the air, will allow to calculate the
turbulent diffusivity for species transport. In a study on gas phase
reactions, Sorensen et al. [23] reported that values of turbulent
Schmidt number between 0.9 and 1.0 give better agreement with
experimental data. Thus, an average value of 0.95 will be adopted
for this work.
2.3. Experimental setup
The performance evaluation of the proposed air conditioning
system for the effective localization of thermal comfort and indoor
air quality at minimal energy cost in an open ofce setting was
investigated by experimental studies on a heated cylinder of 0.5 m
diameter and 1 m height representing the human body with an
approximate surface area of 1.8 m2. Therefore, in order to validate
the performance of the CFD model in predicting the inhaled IAQ,

A. Makhoul et al. / Building and Environment 59 (2013) 336e348

the numerical model was modied to include the heated cylinder

placed on a supportive base in front of a desk where the fans are
mounted as seen in Fig. 4a. Similar turbulence models, grid
generation process, and boundary conditions were applied to this
domain as the ones described previously in section 2.1. The aim of
the experiments was to validate the predicted CFD model results on
ow, thermal and CO2 concentrations and to provide measurements for more accurate modeling of the boundary conditions.
Extensive measurements of the ow characteristics were conducted in the vicinity of the nozzles outlet and the installed heated
cylinder representing the human body. The experiments included
measurement of the velocity, temperature and carbon dioxide
concentration elds in specied positions of the proposed air
conditioning system.
The experimental room (3.4 m  3.4 m  2.8 m) is constructed
of highly insulated walls with an internal layout for one working
ofce station as shown in Fig. 4. The cooling load of the experimental space was mainly due to internal loads such as lighting
(12 W/m2) and the heated cylinder in addition to the external load

Fig. 4. (a) Frontal and (b) top view of the experimental setup for the validation process.

341

through the walls exposed to an outer air temperature of 32
C

representing typical outdoor summer conditions. Two Honeywell
(9000 Btu/h) air-handling units are used for the control of the space
air conditioning: one unit supplies conditioned fresh air though the
nozzle and the other conditioning unit delivers recirculated air to
the peripheral ceiling diffuser (Fig. 5). The recirculation airow is
controlled by motorized dampers and by varying the air-handling
units fan speed.
The heated cylinder representing the human body was placed
on a supportive base at a distance of 10 cm from the desk where the
fans responsible for controlling the convection ow around the
cylinder are installed. The desk-mounted fans suppress the thermal
plumes by creating a negative pressure gradient that diminishes
when moving away from the inlet of the fans. Therefore, the fans
operate more efciently when they are installed closer to the
human body. However, in order to simulate realistic ofce conditions, a typical value of 10 cm representing the common distance
separating the occupant from the desk was considered. The AC
motorized (220 V) fans have a nominal power of 10 W and were
operated at their maximum capacity of 10 L/s ow rate. The fans
were installed in a wooden short tunnel and xed to the internal
upper part of the desk as shown in Fig. 4 to direct the ow away
from the occupant and suppress any interference or recirculation
that might occur in the vicinity of the fans. The outlet of the fans
was kept free and away from any obstruction to ensure effective
operation. Installing the fans in this location, which was the same
position adopted by Bolashikov [9], was chosen since it was proven
to be the most efcient and least cumbersome. When other positions where investigated, such as installing the fans on the side and
frontal walls of the desk, it turned out to be much less effective.

Fig. 5. Picture showing the experimental setup with the heated cylinder and the deskmounted fan.

342

A. Makhoul et al. / Building and Environment 59 (2013) 336e348

The temperature and ow rate of the PV and diffuser ow are

controlled by installed dampers and in-ow electric heaters. To
reduce the ow turbulence intensity and mixing in the airow, the
nozzle was tted with honeycomb ow straighteners and two
screens. The honeycomb holes diameter is 4 mm while their
thickness is 30 mm such that the thickness over diameter ratio (t/D)
is equal to 7.5, which is in the recommended range for optimal
results [24]. Two screens with 60% porosity were installed
upstream and downstream of the honeycombs. The human body is
replaced by a surface heated cylinder representing a sitting person.
The cylinder surface was heated using electric heaters generating
70 W (z39 W/m2) of heat, which is equivalent to the average
sensible heat generated by a sitting person as recommended by the
2009 ASHRAE Handbook-Fundamentals [18]. The base of the
cylinder was insulated to prevent any heat dissipation through the
oor. The heated cylinder will permit to test the blockage effect of
the thermal rising plumes and the ability of the fresh airow of the
nozzle to penetrate the buoyant cylinder plumes and assess the
performance of the desk-mounted fans. The success of the
proposed air distribution system relies on the concept that the
terminal velocity of the fresh air jet should be above 0.3 m/s [25] to
reach and offset the free upward convective ows around the
occupant. Hence, the experiments will focus on the concept that
the thermal draft that could result from the high jet velocity can be
only minimized if the ambient temperature surrounding the
human body is raised. As a matter of fact, the work of a number of
researchers suggested that the human acceptability to higher air
velocities increases with ambient temperature [26e28]. In addition, there is no need for ceiling jet to reach the lower body
segments as it was also found from the work of other researchers
that the upper body segments are the most sensitive segments for
the cooling needs [29].
The experimental setup allows for variations in the ambient
room temperatures, fresh air temperature and ow rate of the PV
nozzle, and the issue angles of the peripheral diffuser ow. The
velocity and turbulence measurements at the outlet of the PV
nozzle were taken using omni-directional hot-wire anemometers
manufactured by TSI having a 3% accuracy and that were calibrated for low velocity range measurements (<2 m/s). The
anemometers were placed on a horizontal rod directly at the outlet
of the nozzle and were triggered to measure the air velocity and
turbulence intensity every 5 mm from the center of the primary
nozzle to the bounding wall. Similarly, to validate the simulation
results of the peripheral diffuser jet, the velocity was measured at
nine different points labeled on the jet prole of Fig. 6.
For air quality assessment, CO2 sampling tests were conducted
by measuring the CO2 concentration using CO2 sensors manufactured by Alphasense installed at a rake passing over the heated
cylinder. The heated cylinder representing the human body was

placed underneath the nozzles while the return duct was seeded
with 0.0035 L/s ow rate of CO2 to mimic indoor pollutants (Fig. 4).
At steady state, the data was compared with CO2 measurements
taken at the inlet of the return vent and the air quality was evaluated using the ventilation effectiveness equation. It is important to
note that the CO2 sensors were calibrated at the outdoor conditions
before starting the data sampling. The difference in the CO2
concentration between the breathing zone and the macroclimate is
required to assess the IAQ. In addition, and in order to be able to
validate the thermal ow eld obtained numerically, the
measurement of the temperature eld around the human body is
necessary. A set of 9 T-type thermocouples with 0.3
C accuracy
manufactured by Omega were distributed around the cylinder as
shown in Fig. 4a to monitor the variations in the temperature eld.
These thermocouples were connected to an OMEGA DaqPro data
logger to store the data. The measurements were recorded at steady
state, which is assumed to be reached when the monitored
temperatures reach nearly constant values with variation within
0.3
C.
For the model validation, eighteen experiments were performed
at three different fresh air ow rates (5 L/s, 7.5 L/s, 10 L/s) in addition to three associated temperatures (24
C, 20
C, 16
C) for each
of the fans status (ON or OFF). The peripheral diffusers delivered air
at 45
from the ceiling plane at a temperature of 16
C. The CFD
simulations have shown that a peripheral supply angle of 45
is
optimal to obtain a localized ow around the occupant. The CO2
concentration in the air was measured at the horizontal rake that is
situated at 1.2 m from the oor. The height of the ceiling was 2.6 m,
thus the distance separating the inhaled air quality measurements
from the nozzles supply is 1.4 m. The desk fans were mounted in
the middle under the desktop at a height of 0.7 m from the oor
and at a distance of 0.1 m from the heated cylinder. The macroclimate air temperature (which is considered the temperature of
the air away from the human microclimate) was controlled and
maintained at 26
C for all the simulated cases. It is important to
note that the experiments were focused on validating the CFD
model of the single jet ceiling nozzle with and without the deskmounted fans. The experimental validation of the coaxial ceiling
nozzle was performed and reported in the work of Makhoul et al.
[8].
3. Results and discussions
3.1. Validation of the CFD model
Fig. 7 shows the velocity and turbulence intensity measured at
the outlet of the PV nozzle and that were later used as inputs to the
CFD model to dene the nozzle boundary conditions. The velocity
measurements showed steady amplitude at the nozzle outlet that

Fig. 6. Validation of CFD diffuser jet prole with experimental data.

A. Makhoul et al. / Building and Environment 59 (2013) 336e348

343

Fig. 7. Velocity and turbulence intensity measurements at the PV nozzle outlet.

decreased when approaching the walls representing a top hat

prole as seen in Fig. 7. The zero velocity at the nozzle wall indicates
a no-slip condition regarding the airow movement in the viscous
sublayer. Fig. 7 shows that the turbulent intensity increases when
approaching the nozzle wall, which is due to the decrease in the
velocity in the wall boundary layer, thus the ratio of the velocity
uctuations to the mean ow velocity is higher. Values of the
turbulent intensity varied between 2.4% and 3%, which resulted in
adopting the value of 2.7% for the CFD simulations.
The air quality was assessed using the ventilation effectiveness
(3 v) and the results for the three ow rates are shown in the graph
of Fig. 8. The most efcient cases obtained numerically were chosen
for presentation in Fig. 8: these cases correspond to a jet temperature of 16
C for all the three ow rates of 5 L/s, 7.5 L/s and 10 L/s.
The ventilation effectiveness is plotted versus the distance from the
breathing level center point along a line in the symmetry plane.
X 0 represents the center of the line located directly above the
center of the heated cylinder. Fig. 8 shows an increasing air quality
with the increasing ow rate. The ventilation effectiveness reached
a value of 0.26 for a PV ow rate of 10 L/s. These predicted ventilation effectiveness values were compared with experimental data
using the CO2 sensors as described in the experimental protocol.
The results showed good agreement with the CFD data and the
matching was better for higher ow rates especially for the case of
10 L/s. At the lower ow rate of 5 L/s, the ventilation effectiveness
was less evident which is due to the instability of the jet having
a low momentum and due to the interaction with the rising
thermal plume hence the deection of the jet as seen later.
In addition to the ventilation effectiveness measurements and
validation, temperature measurements at 9 points located around
the heated cylinders were performed according to the schematic of
Fig. 4a. For the temperature eld validation, the case where the
macroclimate temperature was maintained at 26
C with a fresh air
ow rate of 10 L/s and temperature of 16
C was considered. The
results of these measurements are compared to the predicted
temperature values obtained through numerical simulations and
are shown in Fig. 9. The experimental and predicted results agreed
well and the noticed error did not exceed the instrumentation
uncertainty.
Furthermore, since the momentum method was used to model
the diffusers jet numerically, experimental measurements of the
diffusers velocity prole were necessary to validate the predicted
numerical values. The result of the validation of the angled jet
diffuser velocity prole is shown in Fig. 6. The gure shows the
velocity measured at several points within the jet and compares the
data with the predicted CFD values. The results show good

Fig. 8. Measured and predicted ventilation effectiveness for a PV fresh air ow rate of
(a) 5 L/s, (b) 7.5 L/s and (c) 10 L/s.

agreement between the measured and predicted values since most

of the noticed differences are within the instrument uncertainty.
This shows that the considered diffuser with the corresponding
geometry and grid density is modeled accurately and can be used to
simulate the jet behavior further in the study.
3.2. Nozzles performance comparison
The aim of this part is to compare the performance of three
types of personalized ventilation systems: a) the ceiling-mounted
nozzle with desk fans turned OFF, b) the ceiling-mounted nozzles
with desk fans turned ON, and c) the low-mixing coaxial ceilingmounted nozzle. The performance of each nozzle type was
assessed by observing the achieved air quality, the perceived
thermal comfort, the ow localization and consequently the energy

344

A. Makhoul et al. / Building and Environment 59 (2013) 336e348

Fig. 9. Validation of temperature prole in the vicinity of the heated cylinder with
experimental data using a 10 L/s jet ow rate and 16
C temperature.

savings. Different parameters will affect the performance of the PV

nozzles and the most prominent are the airow rate and temperature. For that reason, several simulations were performed for
a typical ofce load of 60 W/m2 with three different fresh airow
rates in addition to three associated temperatures resulting in nine
test cases as seen in Table 1. The room air temperature was maintained at 26
C by varying the peripheral diffuser ow rate while
maintaining the temperature of the air it delivers at 16
C. For each
simulation case, the room is divided into two zones: the microclimate and the macroclimate. The boundary between the two zones
is represented by the vertical lines shown in Fig. 2.
A relatively low peripheral diffuser temperature of 16
C was
used so that the system will be able to remove the cooling load of
the room. The Bioheat model was coupled with the CFD model to
generate the human response to the environmental conditions. The
skin and core temperatures and their rate of change are then
integrated into Zhangs thermal comfort model [13e16] to generate
the segmental and overall thermal comfort and thermal sensation.
3.2.1. Air quality and comfort comparison
Table 1 shows the terminal velocity and terminal temperature,
which are the velocities and temperatures attained by the jet when
it reaches the human head, in addition to the ventilation effectiveness factor. It is clear that higher velocities (up to 0.58 m/s) and
lower temperatures (a minimum of 24.63
C) are attained when the
desk-mounted fans are turned ON due to the reduced ow rate of
the thermal rising plume. Hence, less resistance is opposed to the
PV jet and it can easier penetrate the mitigated plume generated by
the human body. In addition, it is important to note the effect of

reducing the PV air supply temperature on acting positively on the

attained air quality since the buoyancy forces generated by the cold
jet will be acting downward. Obviously, this will also allow an
increased amount of fresh air to reach the occupants breathing
zone: this can be observed when monitoring the variation of the
ventilation effectiveness index. In fact, a remarkable improvement
is noticed in terms of indoor air quality when the desk-mounted
fans are turned ON. A ventilation effectiveness value of 22% could
be reached with a fresh air ow rate of 10 L/s supplied at
a temperature of 16
C and this is compared to an effectiveness of
10.4% obtained for the same conguration when the fans are turned
OFF. Similarly, Yang [4] reported an effectiveness of around 10% for
a ceiling PV nozzle integrated with a conventional diffuser. On the
other hand, a remarkable 11 times improvement in the ventilation
effectiveness (going from 1% to 11%) is observed when the fans are
turned ON for the 5 L/s case with a supply temperature of 16
C.
This nding highlights the performance of the desk-mounted fans
in suppressing the thermal plumes and permitting the PV jet to
provide the occupant with fresh air more effectively. However,
when compared with the coaxial PV nozzle, a better IAQ and higher
ventilation effectiveness (up to 33%) is observed for the latter that
could reach a two-fold value for some cases. This brings to the
picture the high efciency of the low-mixing coaxial nozzle in
providing fresh air to the breathing zone in an efcient way.
On the other hand, when examining the performance of an air
conditioning system, it is important also to check its ability in
providing the occupants with the desired thermal comfort. Fig. 10a
and b shows the overall thermal comfort perceived by the occupant
for different ow rate and temperature combinations when the
desk-mounted fans are turned OFF or turned ON, respectively. It is
clear that the comfort increases with an increasing airow rate and
a decreasing air temperature. When comparing the effect of the
desk-mounted fans on the induced thermal comfort, a net
improvement is observed when the fans are turned ON. This effect
increases with the decreasing PV supply air temperature letting the
jet take advantage of the buoyancy forces acting downwards and
supporting the jet in penetrating the thermal plumes as seen in
Fig. 11. Fig. 10 shows also a remarkable improvement in overall
comfort especially for the 5 L/s ow rate cases where the PV jet was
so weak to effectively reach the human body without the usage of
the desk-mounted fans. However, Fig. 10c, which corresponds to
the ceiling-mounted coaxial jet, shows a remarkably improved
thermal comfort especially for the 5 L/s and 7.5 L/s cases. A nearly
similar or slightly lower comfort is observed for the 10 L/s case
associated with a supply temperature of 16
C, which is due to the
thermal draft and cold sensation felt under these conditions.
In order to examine and analyze in details the perceived thermal
comfort under certain operating conditions of the PV jet, observing

Table 1
Comparison of the terminal temperature, velocity and ventilation effectiveness for the different cases when the desk mounted fans are turned ON or OFF and when using
a coaxial nozzle.
Case

F
F
F
F
F
F
F
F
F

5 L/s, T 24
C
5 L/s, T 20
C
5 L/s, T 16
C
7.5 L/s, T 24
C
7.5 L/s, T 20
C
7.5 L/s, T 16
C
10 L/s, T 24
C
10 L/s, T 20
C
10 L/s, T 16
C

Fans OFF

Fans ON

Coaxial nozzle

Terminal
temp. [
C]

Terminal
velocity [m/s]

3v

Terminal
temp. [
C]

Terminal
velocity [m/s]

3v

Terminal temp. [
C]

Terminal
velocity [m/s]

3v

26.53
26.47
26.42
26.06
26.00
25.85
25.68
25.35
25.17

0.26
0.26
0.27
0.33
0.36
0.36
0.40
0.46
0.48

0.13%
0.10%
0.94%
0.03%
0.38%
8.60%
0.84%
2.01%
10.38%

26.39
26.33
26.21
26.01
25.96
24.72
25.64
25.23
24.63

0.28
0.31
0.34
0.38
0.43
0.48
0.47
0.52
0.58

1.55%
1.61%
11.14%
2.04%
8.44%
20.67%
2.95%
14.56%
22.05%

26.65
25.85
25.32
26.20
25.44
24.68
25.60
25.16
24.22

0.23
0.31
0.40
0.55
0.62
0.67
0.86
0.90
0.93

10.64%
21.09%
21.48%
27.32%
27.84%
28.12%
32.29%
32.66%
32.98%

A. Makhoul et al. / Building and Environment 59 (2013) 336e348

345

improvement in comfort is accompanied by a decrease in the

thermal sensation since the affected segments are exposed to
cooler air with higher velocity. The rest of the body parts are less
affected since they are exposed to nearly the same climatic conditions. The enhanced comfort conditions are associated with up to
20% increase in the terminal jet velocity and with reduction of up to
0.6
C in the terminal temperature for the case of 10 L/s supplied at
16
C (see Table 1).
Besides, since higher macroclimate temperatures are maintained inside the space, the probability of any possible thermal
draft sensation due to the PV jet velocity is minimized [26e28].
Meanwhile, when considering the effect of the fans on the lower
parts of the human body, the segmental comfort shown in Table 2
shows a negligible change in comfort by less than 0.1 which is
actually accompanied by an increase in the overall body comfort by
up to 0.54. In fact, and as suggested by a number of researchers, the
lower body parts are less sensitive for cooling needs [29]. Therefore,
the impact of the air movement created by the desk-mounted fans
on the thermal sensation and comfort perceived by the lower body
parts can be neglected.

Fig. 10. Overall thermal comfort with the desk mounted fans turned (a) OFF, (b) ON,
and (c) for the coaxial PV nozzle.

the segmental comfort of the different body parts is important.

Table 2 shows the segmental comfort and segmental thermal
sensation predicted by the Bioheat model for the different human
body segments. Notice that the most affected segments by the
operation of the desk-mounted fans are located in the upper body
part (head, chest, back, and abdomen). The head is the most highly
affected since it is the rst body part hit by the jet before it spreads
over the rest of the upper body, as seen in Fig. 11. An improvement
in the segmental comfort, accompanied by a decrease in the
thermal sensation, is noticed when operating the desk-mounted
fans comparing to a situation where the fans are turned OFF. This

3.2.2. Energy analysis

One of the main advantages of the proposed system is its ability
to provide a good inhaled air quality with less amount of fresh air.
The usage of the desk-mounted fans will permit to reduce the
blockage effect created by the rising thermal plumes and thus allow
for a more effective delivery of fresh air. In addition, it permits
having a higher room macroclimate temperature while maintaining an acceptable thermal comfort due to the personalized ventilator jet that creates a comfortable localized air region around the
occupant. These preceding characteristics of the system could lead
up to a high energy saving potential comparing to the conventional
mixing ventilation systems where the whole room is maintained at
a homogeneous air temperature and the fresh air is mixed with the
recirculated air before being supplied through the ceiling diffuser.
In order to evaluate the energy savings associated with the alternate operation of the desk-mounted fans, an energy analysis was
performed in order to estimate the cooling loads for selected cases.
The calculations were based on equal thermal sensation bases for
each simulated case separately so that the occupant will have
nearly the same sensation in the mixing ventilation system and the
ceiling-mounted PV system (see Table 3). In addition, the equivalent amount of fresh air that gives the same inhaled air quality for
each case was calculated and used in the analysis. A chiller COP of 5
was assumed to obtain the electrical loads from the calculated
cooling loads. For the PV system, a fan of 15 W of nominal power
was assumed to supply the air through the ceiling-mounted nozzle;
thus, the electrical energy consumed by the fan is added to the
chiller electrical load. When considering the coaxial nozzle an
additional fan of the same power is considered for supplying the
recirculated air through the secondary nozzle. Besides, the deskmounted fan is assumed to have a nominal electrical power of
10 W. Having calculated the electrical energy consumption for each
case separately, it was compared with the energy consumption of
the equivalent mixing ventilation case and the potential energy
savings are assessed.
For each fans status (ON or OFF), four cases achieving high
ventilation effectiveness were chosen for the analyses. The corresponding thermal comfort and thermal sensation are indicated
next to each case as seen in Table 3. The savings ranged between 5%
and 13% for the single core jets depending on the fans status, while
it varied between 25% and 30% for the coaxial nozzle. When
examining the cases where the fans are turned OFF, the highest
energy saving that could be achieved was 8.25% for a fresh airow
rate of 10 L/s and temperature of 16
C. These savings are mainly

346

A. Makhoul et al. / Building and Environment 59 (2013) 336e348

Fig. 11. (a) Temperature, (b) velocity, and (c) CO2 contours for the vertical jet with 7.5 L/s fresh air ow rate supplied at 20
C with a macroclimate temperature of 27
C.

due to the enhanced indoor air quality achieved with a reduced

amount of fresh air and to the improved thermal comfort
comparing to the mixing ventilation systems. However, for the
same case and when the fans are turned ON, an increase in the
energy savings is obtained up to 13.25%. This improvement is due to
the effective delivery of fresh air and to improved induced thermal

comfort when the desk-mounted fans are turned ON. However, in

spite of the benets these fans bring, they consume additional
electrical energy when operating. For that reason, the energy
savings associated with the same case did not double when the fans
were turned ON even though the ventilation effectiveness itself
doubled as seen in Table 3. However, when observing the energy

A. Makhoul et al. / Building and Environment 59 (2013) 336e348

347

Table 2
Segmental thermal comfort for selected cases.
Case
Fans OFF

F 10 L/s T 20
C
F 10 L/s T 16
C

Fans ON

F 10 L/s T 20
C

F 10 L/s T 16 C

Head

Chest

Back

Abdomen

Buttocks

U. arm

L. arm

Hands

Thighs

Calves

Feet

Overall

Comfort
Sensation
Comfort
Sensation

1.34
2.33
1.41
0.60

1.91
1.19
1.43
1.20

0.46
1.28
0.25
1.34

0.43
0.41
0.66
0.46

0.23
0.21
0.66
0.26

1.36
0.83
1.30
0.87

0.31
1.42
0.32
1.49

0.31
0.71
0.73
0.59

1.22
0.51
1.18
0.53

0.64
0.66
0.65
0.68

1.82
0.69
1.68
0.71

0.04
1.30
0.59
0.69

Comfort
Sensation
Comfort
Sensation

0.64
0.66
1.17
1.28

1.74
1.09
1.66
0.61

0.35
1.34
0.79
0.85

0.61
0.42
0.90
0.48

0.50
0.25
1.02
0.34

1.36
0.83
1.24
0.90

0.42
1.41
0.34
1.54

0.44
0.73
0.85
0.70

1.21
0.52
1.18
0.53

0.58
0.73
0.57
0.77

1.73
0.70
1.58
0.69

0.60
0.91
0.86
0.15

Table 3
Energy savings at equivalent thermal sensation and inhaled air quality.
Case

Thermal
comfort

Thermal
sensation

3v

Mix vent. equ.

ow [L/s]

Mix vent. equ.

temp [
C]

Energy
savings

Fans OFF

F
F
F
F

7.5 L/s, T 20
C
7.5 L/s, T 16
C
10 L/s, T 20
C
10 L/s, T 16
C

0.06
0.52
0.04
0.59

1.38
0.83
1.3
0.69

0.38%
8.60%
2.01%
10.38%

7.53
8.21
10.23
11.18

25.88
25.33
25.8
25.19

4.67%
6.99%
6.66%
8.25%

Fans ON

F
F
F
F

7.5 L/s, T 20
C
7.5 L/s, T 16
C
10 L/s, T 20
C
10 L/s, T 16
C

0.38
0.67
0.60
0.86

0.99
0.27
0.91
0.15

8.44%
20.67%
14.56%
22.05%

8.20
9.46
11.73
12.86

25.49
24.77
25.41
24.65

6.50%
11.68%
10.51%
13.25%

Coaxial nozzle

F
F
F
F

7.5 L/s, T 20
C
7.5 L/s, T 16
C
10 L/s, T 20
C
10 L/s, T 16
C

1.42
1.19
1.12
0.79

0.78
0.96
0.99
1.16

26.59%
27.06%
31.85%
32.12%

10.22
10.29
14.71
14.76

23.44
23.08
23.02
22.68

26.78%
24.93%
30.08%
27.08%

savings associated with the operation of the ceiling-mounted

coaxial nozzle, a remarkable improvement of up to two or three
times is achieved for some cases with savings that could attain
values of 30%. This is due to the enhanced air quality and better
comfort that can be provided by the coaxial nozzle which imply
effective ventilation with reduced amount of fresh air and correspond to a lower mixing-ventilation equivalent air temperature.

jet coverage provided by the primary and secondary nozzles

permitted to obtain improved overall thermal comfort when
compared to the single jet cases. This enhanced performance of the
coaxial PV nozzles lead up to remarkable cut down in energy
consumption (up to 30%) when compared to the mixing ventilation
system relying on a homogeneous temperature distribution in the
space.

4. Conclusion

Acknowledgements

Providing clean air to the breathing zone of the occupant in an

ofce space is an important issue to consider for healthy buildings
design. In addition to the improved air quality, it permits to cut
down the air-conditioning electrical costs by supplying the fresh air
more efciently. Since personalized ventilation is considered as one
of the most effective solutions to the buildings ventilation problems, several add-ons and improvements were considered in the
study and the corresponding resulting impacts were analyzed.
Installing desk-mounted fans as an option to control the convective
plumes generated by the human body has shown a good performance enhancement potential. It permits to mitigate the rising
thermal plumes and reduce the resistance opposed to the vertical
PV jet thus allowing for a more effective delivery of fresh air. When
operating the desk-mounted fans the ventilation effectiveness
almost doubled (22.05%) when compared to a case where the fans
where witched OFF (10.38%). When comparing these PV systems
with the mixing ventilation system, energy savings up to 8% and
13% could be achieved when the desk-mounted fans are turned OFF
and ON, respectively.
The low-mixing ceiling-mounted coaxial PV nozzles where also
considered in this study to assess their performance and compare
with that obtained with the desk-mounted fans. They showed
a higher ability to provide good air quality to the occupants with
increased values of ventilation effectiveness up to 32%. The larger

The authors would like to thank the Lebanese National Council

for Scientic Research (CNRS) for their nancial support. The
support of the University Research Board (URB) at the American
University of Beirut is also acknowledged.
References
[1] Lo LJ, Novoselac A. Localized air-conditioning with occupancy control in an
open ofce. Energy Buildings 2010;42:1120e8.
[2] Halvonova B, Melikov A. Performance of ductless personalized ventilation in
conjunction with displacement ventilation: impact of intake height. Building
Environ 2010;45(4):996e1005.
[3] Li R, Sekhar S, Melikov A. Thermal comfort and IAQ assessment of under-oor
air distribution system integrated with personalized ventilation in hot and
humid climate. Building Environ 2010;45(9):1906e13.
[4] Yang B. Thermal comfort and indoor air quality evaluation of a ceiling
mounted personalized ventilation system integrated with an ambient mixing
ventilation system, PhD thesis. National University of Singapore. Singapore;
2009.
[5] Yang B, Chandra S, Melikov A. Ceiling mounted personalized ventilation
system in hot and humid climate e an energy analysis. Energy Buildings 2010;
42:2304e8.
[6] Khalifa HE, Janos MI, Dannenhoffer JF. Experimental investigation of reducedmixing personal ventilation jets. Building Environ 2009;44:1551e8.
[7] Jackie S, Russo TQ. Computational analysis of reduced-mixing personal
ventilation jets. Building Environ 2009;44:1559e67.
[8] Makhoul A. Ceiling-mounted fresh air personalized ventilator for occupantcontrolled microenvironment. Proceedings of the ASME 2012 International
Mechanical Engineering Congress and Exposition; IMECE 2012.

348

A. Makhoul et al. / Building and Environment 59 (2013) 336e348

[9] Bolashikov Z, Melikov A, Krenek M. Control of the free convective ow around

the human body for enhanced inhaled air quality: application to a seatincorporated personalized ventilation unit. HVAC R Res 2010;16(2):161e88.
[10] Bolashikov ZD, Melikov A, Krenek M. Improved performance of personalized
ventilation by control of the convection ow around occupant body. ASHRAE
Trans 2009;115(2):421e31.
[11] ANSYS Software. ANSYS Inc. http://www.ansys.com/.
[12] Salloum M, Ghaddar N, Ghali K. A new transient bioheat model of the
human body and its integration to clothing models. Int J Thermal Sci 2007;46:
371e84.
[13] Zhang H, Arens E, Kim D, BuchBerger E, Bauman F, Huizanga C. Comfort,
perceived air quality, and work performance in a low-power task e ambient
conditioning system. Building Environ 2010;45(1):29e39.
[14] Zhang H, Arens E, Huizenga C, Han T. Thermal sensation and comfort models
for non-uniform and transient environments: Part I: Local sensation of individual body parts. Building Environ 2010;45(2):380e8.
[15] Zhang H, Arens E, Huizenga C, Han T. Thermal sensation and comfort models
for non-uniform and transient environments: Part II: local comfort of individual body parts. Building Environ 2010;45(2):389e98.
[16] Zhang H, Arens E, Huizenga C, Han T. Thermal sensation and comfort models
for non-uniform and transient environments: Part III: whole-body sensation
and comfort. Building Environ 2010;45(2):399e410.
[17] Gagge AP, Fobelets AP, Berglund LG. A standard predictive index of human
response to the thermal environment. ASHRAE Trans 1986;92(2).
[18] ASHRAE. Handbook-fundamentals. Atlanta: American Society of Heating Airconditioning and Refrigeration Engineers, Inc; 2009.
[19] ANSI/ASHRAE. Standard 90.1-2007, energy standard for buildings except lowRise residential buildings. Atlanta: American Society of Heating, AirConditioning and Refrigeration Engineers, Inc; 2007.

[20] Chen Q, Jiang Zh. Simulation of a complex air diffuser with CFD techniques.
Proc Roomvent 1996;1:227e34.
[21] Makhoul A, Ghali K, Ghaddar N. The energy saving potential and the associated thermal comfort of displacement ventilation systems assisted by
personalized ventilation. Indoor Built Environ [published online 05.04.12],
http://dx.doi.org/10.1177/1420326X12443847.
[22] Cropper P, Yang T, Cook M, Fiala D, Yousaf R. Simulating the effect of complex
indoor environmental conditions on human thermal comfort. Building Simulat 2009. Scotland.
[23] Sorensen DN, Weschler CJ. Modeling-gas phase reactions in indoor environments using computational uid dynamics. Atmos Environ 2002;36:9e18.
[24] Kulkarni V, Sahoo N, Chavan S. Simulation of honeycombescreen combinations for turbulence management in a subsonic wind tunnel. J Wind Eng
Indust Aerodyn 2011;99:37e45.
[25] Bolashikov ZD, Nikolaev L, Melikov AK, Kaczmarczyk J, Fanger PO. Personalized ventilation: air terminal devices with high efciency. Proceedings of
Healthy Building, 2003; 850e55.
[26] Ghaddar N, Ghali K, Chehaitly S. Assessing thermal comfort of active people in
transitional spaces in presence of air movement. Energy Buildings 2011;
43(10):2832e42.
[27] Ghali K, Ghaddar N, Bizri M. The Inuence of wind on outdoor thermal
comfort in the city of Beirut: a theoretical and eld study. Int J HVAC R Res
2011;17(5):813e28.
[28] de Dear Cndido RJ, Lamberts R, Bittencourt L. Air movement acceptability
limits and thermal comfort in Brazils hot humid climate zone. Building
Environ 2010;45(1):222e9.
[29] Newsham G, Brand J, Donnelly C, Veitch J, Aries M, Charles K. Linking indoor
environment conditions to job satisfaction: a eld study. Building Res Inf
2009;37(2):129e47.