Building and Environment 87 (2015) 327e341

Contents lists available at ScienceDirect

Building and Environment

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

Air ow models for sub-slab depressurization systems design

Thierno M.O. Diallo a, *, Bernard Collignan b, Francis Allard a
a
b

University of La Rochelle-(LaSIE, CNRS UMR 7356), Avenue Michel Cr

epeau, 17042 La Rochelle, France
CSTB Health Division, 84 avenue Jean-Jaur
es, 77447 Marne-la-Vall

ee, France

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 3 September 2014
Received in revised form
14 January 2015
Accepted 16 January 2015
Available online 28 January 2015

The entry of soil gas pollutants (Radon, VOCs, ) into buildings can cause serious health risks to inhabitants. Various systems have been developed to limit this risk. Soil Depressurization System (SDS) is
one of the most efcient mitigation systems to prevent buildings against these pollutants. Two operating
modes of SDS are currently used: active and passive systems. Active systems use a fan which enables to
extract air from the sub-slab. Passive systems use the stack effect and the wind to extract air from the
sub-slab. Until now, no airow model has been developed that leads to the effective design of these
systems. In this paper, an analytical method has been used to develop airow models to design these
systems. The developed models take into account different type of substructures. These models are integrated in a multizone airow and heat transfer building code. This integration permits to take into
account climate conditions (stack effect, wind), building envelope characteristics and ventilation systems. Preliminary eld verication results for the extracted ow by the SDS in an experimental building
are presented and discussed. The results show that the airow models are accurate to design SDS. A rst
application of the models is illustrated by the impact of climate conditions on the operation of the
passive SDS. A second application is illustrated by the study of the impact of the ventilation strategies
(natural ventilation, exhaust ventilation, supply ventilation and balanced ventilation systems) on the
passive SDS operation.
2015 Elsevier Ltd. All rights reserved.

Keywords:
Soil gas
Passive sub-slab depressurisation
Active sub-slab depressurization
Airow model
Design
Ventilation systems

1. Introduction
The entry of soil gas pollutants (Radon, VOCs,) into buildings
can be an important exposure pathway for human health,
increasing the risk of lung-cancer. The most effective system to
reduce the soil gas pollutants entry into buildings is Soil
Depressurisation System. It prevents the convective transfer of
soil gas pollutants into buildings [1e6]. This system is generally
installed with an exhaust fan, which enables to maintain a constant depressurization beneath the building. Sometimes this
depressurization can be obtained naturally by using stack effect
and wind effect. The advantage of this system is a low cost of
operation and low maintenance. The performance measurements
of the passive SDS in several houses have given an average performance of 55% [7]. The ability and the efciency of the passive
SDS are not properly characterized and need to be tested [8]. In

* Corresponding author.
E-mail address: thierno.diallo@univ-lr.fr (T.M.O. Diallo).
http://dx.doi.org/10.1016/j.buildenv.2015.01.017
0360-1323/ 2015 Elsevier Ltd. All rights reserved.

the literature, some analytical [9,10] and numerical [11e15]

models have been developed to characterize such systems.
Reddy et al. [9] used an analytical airow SDS model based on the
non-Darcy ow. As explained by Gadjil et al. [11], this model has a
certain limit by the assumption that there is no slab peripheral
crack and the sub-slab/soil interface is impermeable. In fact, the
presence of peripheral slab crack can have a signicant impact on
the pressure eld and gas transfer in the sub-slab gravel layer.
According to experiments of Turk [16], 40e90% of the air drawn by
the SDS comes from inside of the building. The Reddy et al. model
is valid only when soil permeability is very lower than gravel
permeability under the slab. If the soil is permeable, this model is
not correct. Cripps [10] has developed also an analytical model
based on the Darcy-Forchheimer equation [11,12]. Unlike Reddy
et al. model, this model includes a peripheral crack of the slab.
However, as the Reddy et al. model, it considers that the sub-slab/
soil interface is impermeable. Gadjil et al. [11] and Bonnefous [12]
used a 3-D nite element model to study the performance of SDS
systems. This model takes into account the diffusive and
convective transports in the soil and the sub-slab gravel. However,

328

T.M.O. Diallo et al. / Building and Environment 87 (2015) 327e341

it does not consider the transport (convection and diffusion)

through the slab. It considers only the airow through the slab
peripheral crack. Otherwise, this model permits to identify the
mechanisms and factors contributing to the performance of SDS.
To study the effectiveness of passive SDS to reduce the radon entry
into buildings, Holford and Freeman [14] used the nite element
model (Rn3D), which simulates the convective and diffusive
transport in a porous medium. The transport in the porous medium is governed by Fick's law (diffusion) and Darcy's law (convection). This model considers the mass ux through the slab and
the peripheral. Compared to numerical models, analytical models
are more attractive in terms of use by professionals, even if they
represent fewer phenomenon involved.
Furthermore, a one-year follow up of a passive SDS has been
performed in an experimental dwelling on CSTB site [8,17]. The
efciency of the system has been monitored. These experimental
results show the potential advantage of passive SDS to protect
buildings against the soil gas pollutants. However, this efciency
depends on meteorological conditions and some building characteristics. In this context, it appeared necessary to develop a tool for
the design of such a system, which considers these relevant parameters meteorological conditions and building characteristics.
Such a tool could enable to test the efciency of passive SDS in a
specic context.
In this paper, an analytical model to characterize mechanical
operation of passive and active SDS is presented. Compared to
existing SDS analytical airow models, this model considers the
airow in the soil, which can affect the SDS efciency. It considers
both the slab permeability and the slab peripheral crack. The

airow in the sub-slab gravel layer is governed by the DarcyeForchheimer equation. Two slab substructures are considered:
the oating slab and the supported slab. This model is coupled to a
multizone airow and heat transfer building code and enables to
evaluate the efciency of passive SDS along a year in specic
meteorological conditions and building characteristics. The results
obtained are compared with experimental results and some
sensitivity studies are performed.
2. Presentation of previous experimental study
A one-year follow up of a passive Soil Depressurisation System
has been performed in an experimental dwelling on CSTB site
[8,17]. The objective of this study was to assess the mechanical
efciency of such a system along a year. This experiment and main
results obtained are summarized below (Fig. 1). During its construction, a passive SDS has been installed in the experimental
house MARIA. To analyze the performance of the system, the air
velocity and entrance duct temperature have been measured each
minute with a probe introduced in the duct. The indoor temperatures considered in the following analyses are average temperatures measured in different locals (rooms, living room and kitchen).
Likewise, basement depressurization has been measured between
the gravel layer and indoor environment at oor level with a differential manometer. Wind (velocity and direction) and external
temperature have been recorded from a meteorological station
located close to the experimental building.
Results obtained consist on an important database of physical
variables measured each minute along the year. Main results

Fig. 1. a) Parameters measured during the one year follow up b) Evolution of air ow extracted and soil depressurization c) Comparison of basement depressurization function of
basement extract ow for natural and mechanical extraction d) Percentage of running time of the system along year above three thresholds [8].

T.M.O. Diallo et al. / Building and Environment 87 (2015) 327e341

obtained during this experiment are presented in Fig. 1b and c.

Fig. 1b shows that the time evolution of the depression and the
extracted airow rate from the sub-slab are correlated. Furthermore, looking at basement permeability characterization (Fig. 1c), it
appears that characterization obtained with passive mode SDS is
quite similar to that obtained using a mechanical mode. Discrepancy could be due to a more leaky basement due to necessary adaptations undertaken to conduct passive SDS experiments. Fig. 1d
shows the percentage of running time of the system along the year
above three thresholds of extracted airow rates from the basement. These experimental results have shown the interest of passive S.D.S., on the one hand it protects the building against the soil
gas pollutants. On the other hand, this technique works at a marginal cost, unlike mechanical S.D.S. It appears also that the efciency of the S.D.S. is highly variable along the year. However, in
these experiments, the percentage of running time could be signicant and mainly during winter season. This is an interesting
result because preventive solutions are mainly needed during the
cold period to ght against soil gas pollutants entrance due to
convection. A secondary result showed the interest to install an
efcient static extractor at the exit of the duct to ameliorate the
running of the system [8]. As a conclusion, it could be said, that
efcient running of passive SDS can be achieved but it depends on
climatic conditions and some building and environmental parameters. Thus, it appears interesting to develop a model that can
predict the performances of passive SDS in specic conditions and
optimize the design of such system.

329

3. Development of an analytical air ow model for passive

SDS
An analytical model has been developed to determine the mechanical running characteristics of a passive SDS (airow and
depressurization) as a function of building characteristics and
meteorological conditions (wind and temperature difference).
Conceptual model to quantify airow through porous media is
based on the analogy between the conduction heat transfer and the
airow as presented in previous works [18,19]. The airow through
porous media follows Darcy law except in gravel layer where DarcyeForchheimer law is used. Airow through passive SDS duct is
due to pressure difference between gravel layer and the environment. It can come from indoors and/or from the soil:

Q SDS Q ind Q soil

(1)

Qsds, Qind and Qsoil (in m3/s) are respectively the airows into the
duct, from indoors and from the soil respectively. There is an
interaction between airows from indoor Qint and soil Qsoil. These
airows inuence both the sub-slab pressure level. The modication of the sub-slab pressure level modies Qind and Qsoil. For
example, if the airow from indoor changes, the sub-slab pressure
level changes and the airow from outdoor changes because it
depends in part to the pressure difference between the outdoor P0out
and under slab Pug . Fig. 2 shows a scheme of passive SDS integrated
in its environment.

Fig. 2. SDS integrated in a house, with different substructures: A) supported slab substructure B) oating slab substructure.

330

T.M.O. Diallo et al. / Building and Environment 87 (2015) 327e341

3.1. Determination of air ow from indoors (Qind) for supported slab

(Fig. 2a)

The resistance of the slab Rslab is given by eq. (4) and the
resistance of the crack is:

Driving pressure loss between indoors and duct entrance can be

written as followed:

Rcrack

Pind  Pe Pind  PUS PUS  Pe

(2)

with Pind the driving pressure above the oor, Pe the driving pressure at duct entrance and PUS the driving pressure under the slab.
Driving pressure drops between indoors and under slab can be
expressed:

Pind  PUS Rslab  Q ind

(3)

with (Diallo et al., 2012):

Rslab

eslab  m
kslab  Sslab

(4)

with eslab (m) the thickness of the slab, m (Pa$s) the dynamic viscosity of air, kslab (m2) the air permeability of slab and Sslab (m2) the
surface of the slab.
Based on expressions presented in Annex for the determination
of driving pressure drop in the gravel slab between two parallel
surfaces, the driving pressure drop between under slab and duct
entrance is as followed:
2
PUS  Pe Rg1 cA1
h Q ind Rg1 Q ind

(5)

Where c is the Forchheimer coefcient, Rg1 (Pa/m3/s) the gravel

resistance between under slab and duct entrance and Ah (m2) the
area of the hemisphere at the entrance of the duct. By replacing eq.
(3) and eq. (5) in eq. (2), the following expression is obtained:




Rg1 c
Q 2ind Rg1 Rslab Q ind  Pind  Pe 0
Ah

(6)

Solving eq. (6) with positive discriminant, airow from indoors

can be expressed:




Q ind  Rg1 Rslab




0:5 



2
Rg1 c
Rg1 c 1
Pind Pe
2
Rg1 Rslab 4
Ah
Ah
(7)

12  eslab  m

(10)

d3 P

With d (m) the width crack, P (m) the perimeter of the slab.

3.3. Determination of air ow from the soil (QSOIL)

Driving pressure drop between outdoors and duct entrance can
be written as follows:




P0out  Pe P0out  Pug Pug  Pe

(11)

With P0out and Pug outdoor driving pressure at ground level and
driving pressure under gravel layer respectively, in Pa. The pressure
drop between outdoors and under slab can be expressed by:



P0out  Pug Rsoil  Q soil

(12)

With [16]:

p
2 Lwf Lbw

B1
6
6ksoil P B
B
ln
Rsoil 6
B
6 pm
@
4
1

!

p
Lwf Lbw

Lslab em
2

!

em
2

131
C7
C7
C7
C7
A5

(13)

where ksoil (m2) is the air permeability of the soil, P (m) the
perimeter of the slab, Lwf (m) the length of foundation, Lslab the
length of the slab and em (m) the thickness of the foundation wall.
Lbw (m) is the basement wall height. If Lbw 0 in eq., we get slabon-grade substructure.
Based on expressions presented in the annex for the determination of pressure drop in gravel between two surfaces, driving
pressure drop between under gravel and duct entrance is as
follows:
2
Pug  Pe Rg2 c A1
duct Q soil Rg2 Q soil

(14)

Where Rg2 (Pa/m3/s) is the gravel resistance between under gravel

and duct entrance and Aduct (m2) the surface of the cylinder at the
entrance of the duct. Replacing eq. (12) and eq. (14) in eq. (11), we
obtain:

3.2. Determination of air ow from indoors (QIND) for oating slab

(Fig. 2b)






Rg2 c
Q 2soil Rg2 Rsoil Q soil  P0out  Pe 0
Aduct

To determine the air ow indoors for oating slab, the resistance

of the slab Rslab in eq. (7) is just replaced by the total resistance Rtot
of the slab and the peripheral crack which are in parallel.

Solving eq. (15) with positive discriminant, the air ow rate

from the soil can be expressed:




Q soil  Rg2 Rsoil



Q ind  Rg1 Rtot






0:5 


2
Rg1 c
Rg1 c 1
Pind  Pe
2
Rg1 Rtot 4
Ah
Ah
(8)

The total resistance Rtot is:

i1
h
RT Rcrack 1 Rslab 1

(15)

(9)




Rg2 Rsoil

2




0:5 
Rg2 c  0
Pout  Pe
4
Aduct

(16)




Rg2 c 1
 2
Aduct

Replacing eq. (16) and eq. (7) in eq. (1), the air ow rate into the
duct for supported slab can be deduced:

T.M.O. Diallo et al. / Building and Environment 87 (2015) 327e341

331



0:5 









2

Rg1 c
Rg1 c 1
Pind  Pe
2
 Rg1 Rslab Rg1 Rslab 4
 Rg2 Rsoil
Ah
Ah




0:5 
R c1
2

Rg2 c  0
g2
Pout  Pe
2
Rg2 Rsoil 4
Aduct
Aduct


Q SDS

Finally, by replacing Rslab by Rtot in eq. (17), the air ow rate into
the duct for oating slab is obtained.
3.3.1. Determination of driving pressure difference into the duct
This driving pressure difference is the result of the equilibrium
between pressure losses and stack effect. It is commonly expressed
as:

Pe  Ps DPfriction DPsingularity DPstack

(18)

with pressure losses due to friction into the duct and due to singularity, respectively:


H
1
Q SDS 2
 r
;
DH
2
Aduct



X
1
Q SDS 2

zi  r
2 Aduct

3.3.2. Determination of pressure loss at the exit of the duct (static

extractor)
In presence of wind, shape of the extractor creates a depression
at the exit of the duct and we have:

PS  PH
out

1
rC
U2
2 p;out

(23)

with PH
out the outdoor driving pressure at the exit duct level in Pa,
Cp,out a suction coefcient depending on the shape of the extractor
and U (m/s) wind velocity.

3.3.3. Determination of pressure loss at the exit of the duct

(mechanical extractor)
The depression created by a mechanical extractor can be
expressed by a quadratic law [21]:

DPfriction l
DPsingularity

(17)

with l and xi are the frictional and singular loss coefcients

respectively [20]. DH (m) is the hydraulic diameter of the duct, r
(kg/m3) the volumic mass of the air and h (m) the height of the
duct. The stack effect into the duct DPStack is:

DPstack re  rs g H

(19)

where re and rs (kg/m3) are respectively the air densities at the

entrance and at the exit of the duct. It is assumed that Te and re are
known. The unknown is:

rs re Te =Ts

(20)

To determine Ts and based on enthalpy balance into the duct,

temperature along the duct for a given height (h) can be written as
[21]:

Th Tint Te  Tint expu am h; with : u

pD
rQ SDS CP
(21)

In eq. (21), Tint is supposed to be known. am is a global exchange

coefcient between indoors and air ow into the duct. It could be
determined considering three resistances in parallel as followed:

1
1
lnr1 =r2
1

am Sint Sint hint

2plH
hext Sext

DPext

r 
Px Cx Q 2sds
r0

(24)

Where Px and Cx are coefcients from the overall characteristic of

the extractor used. r0 and r are respectively the reference density of
air at 20  C and the actual density. This depression can also be
obtained by a polynomial law from the fan characteristic [23]. This
kind of extractor is not used in this study; this extractor model is
presented just to show that the SDS model developed can be
adapted for this type of extractor as well. In this study, we focus on
the passive extractor.

3.3.4. Determination of QSDS

Using eq. (18) to eq. (23), pressure difference between the
entrance of the duct and outdoors can be written as:





H
Pe  PH
out Pe  Ps Ps  Pout

X
L
l

zi
DH


1
Q SDS 2
1
 rs  re g H rCp;out U2
 r
2
2
Aduct
(25)

(22)

where Sint (m2) and Sext (m2) are the internal and external area of
the duct respectively, r1 (m) and r2 (m) the internal and external
radius of the duct respectively. hint and hext are classical heat exchange coefcients for natural convection that can be found in
Elenbass [22]. Eq. (21) enables to determine T(H) (which corresponds to Ts). Once Ts determined, as a function of QSDS, rs can be
written as a function of QSDS with eq. (20) and as a function of
Pind ; Pe and P0out using eq. (17).

From eq. (20), air ow into the duct can be deduced:



1
2
rC

r

r
g
H

U
Pe  PH
e
s
out
2 p;out
!
!
1 0:5

X
L

zi
 r l
DH

Q SDS Aduct 2



(26)
Using the two expressions of QSDS (eqs. (17) and (26)), we
obtain:

332

"

T.M.O. Diallo et al. / Building and Environment 87 (2015) 327e341

 Rg1 Rslab




0:5 #

2
Rg1 c
Pind  Pe
Rg1 Rslab 4
Ah
"

#1 "


Rg1 c
 Rg2 Rsoil
 2
Ah
#




0:5

2
Rg2 c  0
Pout  Pe
Rg2 Rsoil 4
Aduct
"

#1


Rg2 c
 Aduct 2 Pe  PH
 2
out
Aduct

1
re  rs g H  rCp;out U2
2
"
!#1 !0:5
X
L
 r l

zi
0
DH


(27)
P0out

PH
out

In this equation, if Pind ;

and
are known. As explained
previously in x 2.3.1, rs can be expressed as a function
ofPind ; Pe and P0out . The only unknown is Pe. Once Pe is calculated,
QSDS can also be calculated for given conditions.
3.4. Integration in a ventilation model
For this study a numerical ventilation model developed under
Mathlab-Simulink environment has been used [23]. Equation (27)
can be integrated. This equation can be solved using Newton
method. For a given time step, Pind ; P0out and PH
out are given by
ventilation model. Pe, QSDS, Qsoil and Qind are calculated. Determination of Qind could modify indoor mass balance, so it is needed to
have a loop on mass balance of ventilation model; to obtain a
converged result.
This integration enables to obtain mechanical running characteristics of a passive SDS all along the year for given environmental
conditions (meteorology) and building characteristics (dimensions,
ventilation system, air permeability). It makes it possible to
conduct relevant sensitivity studies on given parameters to test the
ability of passive SDS to run in given conditions and to dimension it.
4. Numerical and experimental validation of simplied
airow resistances
The simplied solutions used in this paper to quantify airows
from soil into substructures have been numerically validated in the

Table 1
Comparison between analogical and numerical resistance.

DT

Rana (k/W)

Rnum (K/W)

Rnum/Rana

761
77
3.04

4.64$103
4.64$103
4.64$103

1.1$103
1.1$103
1.1$103

2.4
2.4
2.4

Diallo et al. study [19]. For example, the quantication of the

resistance of the airow from outdoor to the sub-slab gravel layer is
obtained by using the analytical model developed to estimate the
airow from outdoor to bare soil (Fig. 3).
Furthermore, to quantify the airow from the soil/gravel interface to the SDS entry duct, the resistance of airow from an innite
surface and entering into a cylinder has been used. This resistance
is found from the analogical resistance given by Sunderland and
Johnson [24]. However, by using this resistance, we found that it
underestimates the resistance of the airow entering the cylinder.
Subsequently, we performed a numerical study using Comsol 3D
code [25] to correct this resistance. The following Table 1 shows the
comparison between the analogical thermal resistance Rana (K/W)
and numerical resistance Rnum resistance (K/W). Calculations are
performed for a heat ow from an area of 60  60 m2 (this area is
used to represent an innite surface) and entering into cylinder
with 10 cm diameter. The thermal conductivity of the solid used is
400 W m1 K1 with a thickness of 10 cm Table 2.
We found a factor of 2.4 between the resistances. The analytical
resistance is then corrected by this factor in the following
calculations.
To validate experimentally the airow analytical model used to
quantify the airow from the outdoor to the SDS duct, the experimental study performed by Robinson and Sextro [26] has been
exploited. The experimental facility consists in a representative
rectangular room of the building (Fig. 4a). Interior dimensions are
2 m  3.2 m in surface and 2 m in height. The underground walls,
the oor and foundation walls are made with concrete of 15 cm of
thickness. Only 0.1 m from the wall of the room is not buried. The
permeabilities of soil, gravel and backll around the room were
characterized. All openings in the bottom oor were carefully
closed except one of 3.8 cm diameter that is at the center of the
oor. To measure the airow through the hole, the chamber was
depressurized by a pump. Some experiments have been conducted
to establish a relationship between the ow through this hole and
the depression measured in the chamber. These tests give a linear
relationship between the measured ow rate and the measured
depression (Fig. 4c).

Fig. 3. Numerical validation of airow from outdoor to bare soil [19].

T.M.O. Diallo et al. / Building and Environment 87 (2015) 327e341

333

Table 2
Correction of the analogical resistance.
Conguration

Analogical resistance
Rana l

4:44 D

1

D
5:66Z

Corrected analogical resistance

Rnum 2:4 l

4:44 D

1

D
5:66Z

1:85 D

1

D
5:66Z

Sunderland and Johnson [24]

Fig. 4. Experimental validation of the airow model from outdoor to SDS duct. a) experimental facility [26]. b) Assumption about streamlines to quantify the airow from outdoor
to the hole [18] c) Comparison between analytical and experimental airows [18].

Fig. 4c shows that the analytical and experimental ow rate

have the same behavior, they evolve linearly with the depression of
the experimental chamber. However, the analytical rate is underestimated by 28% compared to the experimental one. This
discrepancy could be explained at rst by the assumptions made in
analytical development and secondly by uncertainties about the
different measurements. However, despite this difference the result
is satisfactory compared to the uncertainties that may be encountered. Most important in this comparison is to have the order of
magnitude of the experimental ow.

5. Application of the model

To our knowledge, no study has been conducted to evaluate the
impact of these ventilation systems on the efciency of passive SDS
systems. In this paper, the impact of four ventilation systems on the
operation of the passive SDS is evaluated: natural ventilation, mechanical exhaust ventilation, mechanical supply ventilation and
balanced ventilation. For this purpose, the air ow model developed to design SDS systems has been used [23]. This model is
coupled with a multizone, building energy model [23]. This

Fig. 5. The model concept.

334

T.M.O. Diallo et al. / Building and Environment 87 (2015) 327e341

integration enables to take into account the impact of various parameters such as meteorological conditions (stack effect, wind),
building characteristics (e.g. building envelope, airtightness of the
envelope, building height) and ventilation systems. When the SDS
is operating, according to its efciency, the soil gas pollutant can
enter into the building. In this paper, the impact of ventilation
systems on the operation of SDS is studied regarding the evolution
of the indoor pollutant concentration when the system is operating.
To illustrate our study, radon gas (unit in Bq/m3) is chosen as soil
pollutant. The concept of the model is presented on Fig. 5.
For this study, it is supposed that the transport of the pollutant
in the soil is governed by convection and diffusion. In the gravel
layer under the slab and in the slab, the pollutant transport is
governed only by convection. Specially, air ow in a gravel layer is
approached by the nonlinear equation of DarcyeForchheimer
[11,12].
Determination of mass ux J1: The convective and diffusive
ux of the pollutant from the soil to the gravel layer is estimated by
a semi-empirical equation developed by Diallo [18]. For soil with air
permeability lower than 1012 m2, the transport is mainly diffusive
and the pollutant ux is given by:

J1 Dsoil Ab

Cs  Cgr
L

(28)

Where Cs (Bq/m3) is the pollutant source, Dsoil (m2/s) is the diffusion coefcient of the pollutant in the soil, Ab (m2) is the slab area, L
(m) is the depth of pollutant source, Cgr (Bq/m3) is the pollutant
concentration in the gravel layer under the slab. For soils with air
permeability greater than 1012 m2, the transfert mode is mixed
(convective and diffusive) and the pollutant ux J1 is giving by
Diallo [18].

J1 Cs Q ref

Dsoil
Dsoilref

In a previous study [18], this mass ux has been compared with

a numerical model Comsol [25] and an existing analytical widely
used vapor intrusion model Volasoil [27].
Fig. 6 shows a comparison between the semi-empirical model
developed and Volasoil model for a pollutant from soil entering into
crawlspace. For the low soil permeability, ksoil < 1012 m2, the results
of semi-empirical and numerical models are consistent with the
Volasoil prediction. This is explained by the fact that this model is
based on the analytical solution of one-dimensional convectiondiffusion equation and for these levels of soil permeabilities, pollutant
ux is exclusively diffusive. However, for high permeabilities
ksoil > 1012 m2, the Volasoil model deviates from the semi-empirical
and numerical models, and this gap increases with soil permeability.
The Volasoil model overestimates the mass ux of the pollutant into
building. This overestimation can be explained partly by neglecting
the effect of 2-D and the inuence of convective airow on the concentration prole from foundations. Moreover, this overestimation
can also be explained because Volasoil model doesn't consider the
possible airow of pollutants towards the atmosphere.
Determination of mass ux J1: The convective ux J2 of the
pollutant that can enter the building through the sub-slab can be
written as:

!


L
b
aQ s =Ab
Dsoil

(29)

Dsoil (m2/s) is the diffusion coefcient of the soil, Dsoilref is the

reference pollutant diffusion coefcient (106 m2/s), Qref (1 vol/h) is
the reference ventilation rate, Qs (m3/s) is the air ow from the soil
to the gravel layer, Ab (m2) is the slab area, L (m) is the depth of
pollutant source. a et b are parameters that depend on the
permeability of the soil. These parameters are given in the
following Table 3.
Qs the airow from the soil surface outside the building to the
gravel layer is [18]:

p
Lbw 2 Lwf

B1
k

B
Q s P0out  Pe soil P lnB

B
pm
@ 1

p
Lbw 2 Lbw

Lslab ew
2

C
C
 C
C
ew
A

J2 Cg

Pe  Pint
Rslab

(31)

Cgr is the concentration of the pollutant in the sub-slab gravel

layer. Pind is the indoor pressure. Rslab is the resistance of the slab.

Rslab

eslab  m
kslab  Sslab

(32)

Where eslab is the slab thickness, kslab is the slab permeability, Sslab
is the slab surface. If the pressure difference Pe  Pint in eq. (31) is
positive, no ux enters the building. If it becomes negative the air
ow can drive the pollutant into the building. In the gravel layer, it
is assumed that the pollutant is diluted by the airow of SDS (QSDS),
Its concentration in the gravel layer is assumed to be homogeneous
and is given by:

Cgr

J1
Q sds

(33)

Where J1 (Bq/s) is given by eq. (29) and Qsds, the air extracted by the
SDS is given by the airow model developed in the Diallo et al.
study [19].

(30)

Where Pe (Pa) is the pressure in the gravel layer, P0out (Pa) the
outside pressure, m (Pa$s) the dynamic air viscosity, Lslab (m) the
slab width, ew the wall thickness, Lwf (m) the wall foundation depth,
Lbw (m) the basement wall depth and P (m) the building perimeter.
Table 3
a and b parameters.
1012 < ksol <1010 m2


a 103 10

q
Lslab
L

ksol > 1010 m2



1:5

 q

b 103 2:4 Lslab
L  1:3




1:8
1:76
a 107 5:79 Lslab
L



0:36
 0:094
b 103 0:36 Lslab
L
Fig. 6. Comparison of semi-empirical and numerical models with Volasoil model [18].

T.M.O. Diallo et al. / Building and Environment 87 (2015) 327e341

6. Results and discussion

Table 4
Main model input parameters for the experimental validation.
Building substructure
Time step
Building tightness (measured)
Pressure coefcient of the passive
extractor (measured)
Soil permeability directly underneath
the building ksoil (measured)
Forcheimer coefcient
Gravel permeability kgr (m2)
Slab effective permeability kslab (measured)

Supported slab
1h
0.9 m3/h/m2
0.4
2.61$1013 m2
11.5 s/m [12]
107 m2
1.69$1010 m2

Mass transfer in the building: To consider the transport of

pollutant in the building, the ux of the pollutant J2 (Bq/s) through
the slab is considered as a source term in the mass conservation
equation. Then, the equation solved in the thermo-aeraulic model is:
Nz Nkj;i

X
X
dmi
_ jik
1  hjik  m
J2
dt
j0

k1

n Nki;j
C
X
Cj X
_ ijk ki  i


m
rj
ri
j0

335

(34)

k1

mjik is the airow rate from zone j to zone i in orice k (Bq/s).

The hjik and ki factors have been introduced by Allard and Rodriguez [28]. The hjik represents the ltering of pollutant's concentration from zone i to zone j through the hole k. 1  hjik represents
the concentration ratio up to zone i, and therefore affects the
concentration of transported pollutant. It can be used to represent
the material's pollutants absorption, or any other reaction (chemical reaction, phase change ) along the air path. At this step, with
the equation (34), the concentration of the pollutant in the building
when the SDS is operating can be evaluated. The simulations have
been performed with the following input parameters. Table 4
It is important to note that the effect of the moisture on the soil
permeability is not considered in the model. The variation of the soil
humidity can affect the soil permeability. In the unsaturated zone, the
pores can partially be occupied by the air and the water. In the saturated
zone, the pores are fully occupied by the water. The variation of the
humidity can also affect the diffusion of pollutants, indeed for pollutants with a weak constant Henry. In this study in a rst approximation,
the variation of the soil humidity can be taken into account via the
effective diffusion coefcient and the permeability of the soil. In the
eld studies, the effect of soil moisture is taken in account because the
measurements have been performed for whole the year.

6.1. Confrontation with experimental results

Passive SDS model presented in x 2 has been confronted with
experimental results of the experiment presented in x 1. For this
confrontation, it was needed to have some additional data as soil
and slab permeability. For slab permeability, the value obtained
during complementary experiments using tracer gas [29] is used.
Soil permeability was measured in situ thanks to a collaboration
with IRSN [30]. It appeared that soil permeability around the
experimental house was relatively heterogeneous. The average
permeability at different points has been used in our model. Fig. 7
presents a confrontation of experimental and numerical results
obtained for air ow through passive SDS and for gravel depressurization between July 2007 and February 2008 in the experiment.
Based on these results, it can be said that numerical results
obtained with our model are relevant. However, it is observed that
experimental results are more variable than numerical ones.
Although, there is generally an overestimation of airow through
the SDS duct with calculations and by consequence, an overestimation of numerical depressurization in gravel layer, these
ndings could be explained by different reasons. At rst, from a
numerical point of view, wind effect on the extractor is always
benecial. It is not necessary the case in a real environment due to
the possible angle between wind direction and exit of the duct
which can block the air ow coming from the duct. Furthermore,
turbulence and uctuations of wind can have a negative impact on
the extracted air ow and there phenomena are not taken into
account our in calculation. Also, results obtained with the model
can be sensible to variation of some relevant parameters used (slab
permeability, soil permeability and extractor suction coefcient).
Fig. 8 show the impact of the variation of these parameters on
averaged air ow from duct on considered period.
Despite assumptions of the model and some uncertainties due
to a lack of knowledge on some relevant parameters, it can be
concluded that numerical results obtained are quite satisfactory
compared to experimental one. As complementary results, and
with an analogy with experimental results presented in Fig. 1d,
Fig. 9 shows the percentage of running time of the system along
the considered period above three thresholds. Same analysis can
be conducted with numerical results than with experimental
results.
For example, it can also be numerically observed that installing a
static extractor with an optimized shape instead of a classical one
enables to take advantage of wind effect on duct exhaust, inducing
an enhancement of exhaust ow in averaged.

Fig. 7. Comparison between experimental and numerical airows for passive SDS (a) and gravel depressurization (b) between July 2007 and February 2008.

336

T.M.O. Diallo et al. / Building and Environment 87 (2015) 327e341

Fig. 8. Impact of variations of (a) slab permeability, (b) soil permeability and (c) extractor suction coefcient on averaged air ow from the SDS duct.

The predicted exhaust air ow using the static extractor with

optimized shape has also been compared with the experimental one.
Fig. 10a shows that the simulated air ow extracted follows the
experimental ow. Compared to the case of the passive SDS with
classical extractor shown in Fig. 4, comparison between the experimental results and the predicted airow ow rates is more satisfactory. The simulated airow follows better the behavior of the

experimental air ow. The simulated uctuations and measured air

ows have approximately the same evolution. The extracted airows
uctuate less than those obtained with the passive SDS. This
behavior can be explained by the fact that the static extractor with
optimized shape smoothes the uctuations and turbulence of wind
at the exhaust of SDS duct, stabilizing the depression generated by
wind and the exhaust airow. Furthermore, Fig. 10b shows that the

Fig. 9. Percentage of running time of the system along the considered period above three thresholds.

T.M.O. Diallo et al. / Building and Environment 87 (2015) 327e341

337

Fig. 10. Comparison between experimental and numerical for air ow through passive SDS (a) and gravel depressurization (b) between March 2008 and May 2008.

experimental sub-slab depression uctuates more than simulated

depression. It is found that the average experimental depression is
more important in March (time <750 h) because during this month
the effect of thermal buoyancy is more important and thus favors the
whole operation of the SDS. In April and May (time> 750 h), we
approach the summer, the stack effect is less important and the SDS
depression is weaker. Compared to the results of Fig. 4, simulated
SDS depression is closer than that measured, which explains that the
measured and simulated airows are closer. In general the various
confrontations are satisfactory. They show that despite the uncertainties and model assumptions, Its possible to approach the
reality.
6.2. Impact of meteorological conditions on the passive SDS
operation
A sensitivity study has been conducted to analyze the impact of
different meteorological conditions on mechanical running characteristics of passive SDS, for the studied building (Fig. 11). Fig. 11
shows meteorological conditions used for this study.
Meteorological conditions of these two cities were chosen
because they correspond to specic and very different climates that
could be observed in France.
It can be seen that external temperature is always lower in
Nancy than in Nice and wind generally higher in monthly averaged.
Fig. 12 presents numerical results obtained for airow through
passive SDS and for gravel depressurization along time for these
two conditions and during the considered period.
Fig. 13 shows the percentage of running time of the system
along the considered period above three thresholds for the two
towns. Passive SDS is more efcient in Nancy than in Nice.

This gure enables to say that for the building considered in

these calculations, passive SDS can be very efcient is in Nancy and
less efcient in Nice. As shown in Fig. 11, the reason is that Nancy
has a colder climate with wind generally slightly higher. This rst
sensitivity study show the potential interest of the model developed to test the ability of the passive SDS to be efcient for a
considered building in given meteorological conditions.
6.3. Impact of ventilation systems strategies on the passive SDS
operation
The studied building (Fig. 14) is a three-level full-scale house
with four zones [8,29]:
 Zone 1: It consists of four bedrooms, a shower and a bathroom
on rst-oor.
 Zone 2: It consists of the living-room, the kitchen and the toilets
situated on ground-oor
 Zone 3: It consists of a hall (stairs)
 Zone 4: It consists of a garage on the basement
The hall (Zone 3) is considered as transit space. It is connected to
all zones across the undercut of closed doors. No direct link is
considered between the different zones (Fig. 14).
The undercut doors is presented in the following Table 5.
The air inlets and permeabilities of the building envelope have
been represented similarly to those of the experimental building
Maria of CSTB [29] and used by Kof [23] for his work:
 4 air inlets at a height of 7.3 m, one to the west, another to the
east and the last two in the South;

Fig. 11. External temperature (a) and wind velocity (b) for two different cities, Nancy and Nice, in France (monthly averaged).

338

T.M.O. Diallo et al. / Building and Environment 87 (2015) 327e341

Fig. 12. Numerical air ow through passive SDS (a) and gravel depressurization (b) for Nice and Nancy meteorological conditions.

Fig. 13. Percentage of running time of the system along the considered period above two thresholds for Nice and Nancy.

 2 air inlets at a height of 4.8 m, one to the south and the other to
the east.
 20 air permeabilities of 1.2 m3/h/m2 under 4 Pa, including:
C 4 air permeabilities for each building faade
C 4 permeabilities at the basement, with two on the east and
two on the west.
The permeability of ceilings and oors are not taken into account. Air inlets characteristics and ventilation systems are
described in Kof's study [23]. The design of ventilation systems in

Fig. 14. Studied building.

the building MARIA is also based on Kof's study. Whatever the

ventilation system, the principle of ventilation is a sweeping
ventilation: air enters in the living rooms, passes through the
building and is extracted into the damp rooms. Four ventilation
systems are studied: mechanical exhaust ventilation, balanced
ventilation (mechanical exhaust and supply with heat recovery),
mechanical supply ventilation and natural ventilation. The ventilation systems are described as follows:
Mechanical exhaust ventilation: The ducts and extraction vents
have a diameter of 125 mm in the kitchen and 80 mm in other service
rooms. An input of self-adjustable air module 20 m3/h at 20 Pa is used
in each bedroom and two in the living room. The total air ow
extracted is 120 m3/h and distributed as follows (design airows):
kitchen 40 m3/h; Bathroom 30 m3/h; shower 30 m3/h; toilets 20 m3/h.
Supply ventilation system: fresh air taken from the outside at the
roof level is blown into the rooms. The system has two xed vents
blowing air in the living rooms. The air inlets are self-adjustable vents
on the building faades. The total air ow blown is 120 m3/h and 20 m3/
h for each of the four bedrooms and 2  20 m3/h in the living room.
Balanced ventilation system: fresh air taken from outside is
mechanically injected into the living rooms with two blowing
vents, while the air is extracted in the damp rooms. The airow
blown is 120 m3/h in the living rooms: 20 m3/h for each of the four
bedrooms, 2  20 m3/h in the living room. The extraction ow rate
is 120 m3/h distributed as follows: 40 m3/h for the kitchen, 30 m3/h
for the bathroom, 30 m3/h for the shower and 20 m3/h for the toilet.
The system has the same extraction rate in the damp rooms. The
ducts of air extraction system are 125 mm in diameter. The insufation ducts of fresh air have a diameter of 80 mm.

T.M.O. Diallo et al. / Building and Environment 87 (2015) 327e341

339

Table 5
Undercut doors between different building zones.
Undercut zones 1e3

Undercut zones 2e3

Undercut zones 4e3

6  80 cm2

2  80 145 cm2

80 cm2

Natural ventilation: For natural ventilation, the air inlets are in

the living rooms by self-adjustable vents with the module 30 m3/h
at 20 Pa. In damp rooms, the air extraction is done with four vertical
ducts of 160 mm diameter each. The ducts reached the roof with a
static extractor at the exhaust. The depression coefcient of the
extractor is 0.4. The effective section of the extraction grids is
100 cm2. The input data of the simulations are shown in the
following Table 6.
To estimate the impact of ventilation systems on the operation
of the SDS, the indoor concentration of the pollutant from the soil is
evaluated. Fig. 15 presents the evolution of the indoor concentration of the pollutant in zone 4 when the different systems are
operating. The reference concentration C0 to estimate the SDS efciency is that obtained in the building without the operation of
the SDS and with natural ventilation. Cint is the indoor concentration obtained when the SDS is operating. Fig. 15 shows the evolution of the indoor concentration when the SDS is operating with
different ventilation systems. Fig. 16 shows the depression in the
gravel layer when the SDS is operating.
With natural ventilation, the SDS is able to effectively dilute the
concentration of the pollutant in the building. This is due to the fact
that during the simulated period, the SDS maintains a sufcient
depression under the slab (Fig. 16).
With supply ventilation, when the SDS is operating, the pollutant
concentration in the building becomes zero. This is explained by
the fact that the supply ventilation causes an overpressure in the
building above the slab which limits blocks the entry of pollutants
into the building (Fig. 12). It should be noticed that generally supply
ventilation alone without SDS could be sufcient enough, because
of the overpressure generated in the building.
When SDS is operating with mechanical extraction, concentration
in the building is not attenuated compared to reference case. Fluctuation of the concentration is explained by the observed competition of the two systems. SDS tends to block the ow of pollutant to the
interior of the building and conversely the mechanical extractor
tends to draw the pollutant into the building. When the mechanical
extractor dominates, the concentration in the building increases, and
when the SDS dominates it is reduced. However and in the case
simulated, in the majority of the time, the SDS is not effective with
mechanical extraction. This is explained by the fact that the mechanical extraction exacerbates the depression of the building and
therefore tends to favor the entry of the pollutant in the building
(Fig. 16). The SDS remains sometimes effective in operation with

Fig. 15. Evolution of the indoor concentration.

mechanical extraction. It can be explained by the paradoxical effect of

ventilation that sometimes dilutes the pollutant in the building. It
should be noticed that with this ventilation system and without the
SDS, the resulting indoor concentration would be much higher. Fig.12
shows that for balanced ventilation system the SDS is effective. The
balanced ventilation systems causes approximately an equivalent
gravel layer depression as the SDS operates with natural ventilation of
the building (Fig. 16). This means that the system does not promote
the entry of pollutants into the building. Fig.17 shows that the exhaust
ventilation system increases the depression of the building and
therefore decreases the depression generated by the SDS. Fig. 17
presents the efciency of ventilation systems when the SDS is operating. This efciency is expressed as follow:

Efficiency

1

Cind
C0

Where Cind is the indoor concentration when the SDS is operating

and C0 is the reference concentration when the SDS is not operating. The efciency of the SDS with supply ventilation is 100%. For
balanced and natural ventilation it's 90% and for exhaust ventilation the average efciency is 0.67%. The negative efciency means
that the exhaust ventilation system exacerbates the entry of the
pollutant in the building, compared to reference situation.
7. Conclusion
In this paper, an analytical method is used to develop airow
models to design Soil Depressurization Systems (SDS): passive
and active systems. The develop airow models is integrated in a

Table 6
Inputs of simulations.
Inputs of simulations
Weather
Simulation time (1 month)
Building permeability
Pressure coefcient of natural extractor
Effective soil permeability (ksoil)
Gravel layer permeability kgrav (m2)
Effective slab permeability (kslab)
(measured value for Maria dwelling)
(Abdelouhab, 2011)
Source concentration (Radon)
Source depth
SDS duct diameter

Paris
July
1.2 m3/h/m2
0.4
1011 m2
107 m2
1.69$1010 m2

5$104 Bq/m3
5m
0.2 m

(35)

Fig. 16. Evolution of the pressure under the slab.

340

T.M.O. Diallo et al. / Building and Environment 87 (2015) 327e341

Fig. 17. Evolution of the efciency.

multizone airow building code to consider the inuence of

meteorological conditions (stack effect, wind), building characteristics (height, diameter of duct for SDS, airtightness of building,
type of static extractor) and ventilation systems. The comparison
between numerical and experimental extracted airows from the
SDS shows that the airow models are accurate to approach the
realty. The rst application of the models shows that the well
running of the passive SDS depends strongly on the local meteorological conditions and building characteristics. The second
application illustrates the impact of four ventilation systems in
the operation of SDS: natural ventilation, exhaust ventilation,
supply ventilation and balanced ventilation. This study shows that
supply ventilation and balanced ventilation systems are optimal
for the functioning of natural SDS. The supply ventilation reverse
the pollutant ux into the building and thus promotes the proper
functioning of natural SDS. Meanwhile, the balanced ventilation
and the natural ventilation do not interfere with the proper
functioning of natural SDS. The exhaust ventilation system
emerges as the least optimal to promote the well-functioning of
SDS because it exacerbates the depression of the building and
then, can counterbalance the impact of SDS to block the convective ux from the soil. These illustrations show the ability of the
developed airow models to study and design properly the SDS.

Acknowledgment
This study was conducted in the framework of a Ph.D at CSTB
(Scientic and Technical Center for Building) in collaboration with
LaSIE (Laboratory of Engineering Sciences for Environment) at
University of La Rochelle and partly supported by the ADEME
(French Environment and Energy Management Agency) through
the AGIR-QAI project. Complementary experiments needed for this
study to determine soil permeability were conducted by IRSN
(Institute for Radiological Protection and Nuclear Safety Institut).

m
m
VP  cu2  u
k
k

(A1)

with m (Pa$s) the dynamic viscosity of air, k (m2) the air permeability of the gravel, c the Forchheimer coefcient and u (m/s) the
velocity of air.
On this basis, we assume that the driving pressure difference
between two interfaces in gravel layer can be written as followed:

Pg  Pe a1 Q 2 b1 Q

(A2)

The analogy with DarcyeForchheimer equation (A1) implies

that a1 et b1 coefcients are respectively proportional to 
a kc/m et
k/m.

a1

m
m
c a1 ; b1 b 1
k
k

(A3)

Equation (A2) becomes:

Pg  Pe

m
m
c a1 Q 2 b1 Q
k
k

(A4)

In this equation, m=kc a1 and m=kb1 need to have the dimension

of a resistance (Pa/m3/s). Using dimensional analysis, it can be
concluded that:


a1

M3


and b1

 
1
M

(A5)

It is assumed that a1 and b1 include a shape factor S (m)

depending on geometry considered between the two interfaces. So
that:
1
a1 A1
and b1 S1
c S

(A6)

with Ac (m2) a surface depending on shape of uid interface.

Replacing eq. (A6) in eq. (A4) and considering a resistance of gravel
layer as Rg m/(k S), eq. (A4) becomes:

Annex. Pressure loss in gravel layer

2
Pg  Pe Rg c A1
c Q R g Q

Air ow in a gravel layer can be approached by the nonlinear

equation of DarcyeForchheimer [10]:

For model presented in this paper, it is needed to dene Rg and

Ac for the two congurations as presented below:

(A7)

T.M.O. Diallo et al. / Building and Environment 87 (2015) 327e341

For conguration 1, it is based on the denition a thermal ux

between a semi-innite surface and a hemisphere [24], with D (m)
the diameter of the hemisphere and Ah (m2) the surface of the
hemisphere.
For conguration 2, it is based on the denition of thermal ux
between an innite surface and a cylinder [25], with Lg (m), the
depth of gravel layer and Aduct (m) the surface of cylinder entrance.
Please refer to the main conclusions of the work.

References
[1] USEPA. Radon reduction techniques for existing detached houses. Technical
guidance (third edition) for active soil depressurization systems. 1993. EPA/
625/R-93/011.
[2] Scivyer C. Surveying dwellings with high indoor radon levels: a BRE guide to
radon remedial measures in existing dwellings. London: Construction
Research Publications; 1993. 085125-582-5.
[3] Scivyer CR. Radon protection for new buildings: a practical solution from the
UK. Sci Total Environ 2001;272:91e6.
[4] Scivyer C. Radon: guidance on protective measures for new buildings (BR211).
Garston: BRE Press; 2007. ISBN 978-1-84806-013-5.
[5] Collignan B, O'Kelly P. Dimensioning of soil depressurization system for radon
remediation in existing building. In: Proceedings of healthy building conference, 7the11th Dec. 2003, Singapore; 2003. p. 517e23.
[6] Rydock James P, Skaret Eimund. A case study of sub-slab depressurization for
a building located over VOC-contaminated ground. Build Environ 2002;37:
1343e7.
[7] Angell WJ. Radon control in new homes: a meta-analysis of 25 years of
research. In: AARST proceedings 2012; 2012.
[8] Abdelouhab M, Collignan B, Allard F. Experimental study on passive soil
depressurization system to prevent soil gaseous pollutants into building. Build
Environ 2010;45:2400e6.
[9] Reddy TA, Gadsby KJ, Black HE, Harrje DT, Sextro RG. Modeling air ow dynamics in radon mitigation systems: a simplied approach. J Air Waste Manag
Assoc 1991;41(11):1476e82.
[10] Cripps AJ. Air modeling and measurement of soil gas ow. Construction
Research communication Ltd; 1998. AIVC 11619.
[11] Gadgil Ashok J, Bonnefous Yves C, Fisk William J. Relative effectiveness of subslab pressurization and depressurization systems for indoor radon mitigation:
studies with an experimentally veried numerical model. Indoor Air J
1994;4(4):265e75.



rique des syste
mes de ventilation du sol pour
[12] Bonnefous Y. Etude
nume
se de doctorat en Ge

nie
diminuer la concentration en radon dans l'habitat. The

riaux, Structures et Physique du b^
Civil: sols, mate
atiment. Institut National

es de Lyon; 1992. p. 256.
des Sciences Applique

341

[13] Martin Jiranek, Svoboda Zbynek. Numerical modelling as a tool for optimisation of sub-slab depressurisation systems design. Build Environ 2007;42:
1994e2003.
[14] Holford DJ, Freeman HD. Effectiveness of a passive subslab ventilation system
in reducing radon concentrations in a home. Environ Sci Technol 1996;30:
2914e20.
[15] Al-Ahmady KK, Hintenlang DE. Modeling of the sub-slab depressurization
(SSD) radon mitigation systems for large structures. In: International radon
symposium VP e 1.1; 1994.
[16] Turk BH, Harrison J, Sextro RG. Performance of radon control systems. Energy
Build 1991;17:157e75.
[17] Collignan B, Abdelouhab M, Allard F. Experimental study on passive sub-slab
depressurisation system. In: Proceedings of the American Association of
Radon Scientists and Technologists 2008. International symposium,
September 14e17, 2008, Las Vegas USA. AARST 2008; 2008.

de l'air
[18] Diallo Thierno MO. Impact des polluants gazeux du sol sur la qualite

rieur des ba
^timents. 2013. The
se Universite

de La Rochelle, soutenue le 10
inte
octobre 2013.
[19] Diallo Thierno MO, Collignan B, Allard F. Analytical quantication of airows
from soil through building substructures. Build Simul J March 2013;6(1):
81e94.
[20] I.E IDELCIK. Memento des pertes de charge e coefcients de pertes de charge
res et de pertes de charge par frottement. Traduction du Russe par
singulie
Madame M. Meury. Collection n 13 du Centre de Recherches et d'essais du
Chatou. Paris: Eyrolles; 1969.

lisation des transferts d'air dans les b^
[21] Mounajed MR. La mode
atiments appli


tude de la ventilation. PhD the
se. LEcole
cation ^
a l'e
Nationale des Ponts et


Chaussees; 1989.
[22] Elenbaas W. Heat dissipation of parallel plates by free convection. Physica
1942;IX(1):1e28.
re des strate

gies de ventilation en maisons indi[23] Kof J. Analyse multicrite
se Universite

de la Rochelle].
viduelles. 2009 [The
[24] Sunderland JE, Kenneth Johnson R. Shape factors for heat conduction through
bodies with isothermal or convective boundary conditions. In: ASHRAE 71st
annual meeting in Cleveland, Ohio; 1964.
[25] Comsol mutiphysics. www.comsol.com.
[26] Robinson AL, Sextro RG. Radon entry into buildings driven by atmospheric
pressure uctuations. Environ Sci Technol 1997;31:1742e8.
[27] Waitz MFW, Freijer JI, Kreule P, Swartjes FA. The Volasoil risk assessment
model based on Csoil for soils contaminated with volatile compounds. Bilthoven, the Netherlands: National Intitute for public Health and the environment (RIVM); 1996. RIVM report no.715810014.
[28] Allard F, Rodriguez EA. Coupling COMIS airow model with other transfer
phenomena. Energy Build 1992;18:147e57.
 l'e

tude du transfert des polluants gazeux entre
[29] Abdelouhab M. Contribution a

rieurs des ba
^timents. 2011 [The
se Universite


le sol et les environnements inte
de La Rochelle].

abilite

effective du sol aux gaz autour de la maison
[30] IRSN. Mesure de la perme

rimentale MARIA du CSTB. 2012. RT/PRP-DGE/2012e00019.
expe