Anda di halaman 1dari 15

Construction and Building Materials 176 (2018) 213–227

Contents lists available at ScienceDirect

Construction and Building Materials


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

Characterisation of specified granular fill materials for radon mitigation


by soil depressurisation systems
Le Chi Hung a,b, Jamie Goggins b,c,d,⇑, Marta Fuente a,b, Mark Foley a
a
School of Physics, National University of Ireland Galway, Ireland
b
Civil Engineering, College of Engineering & Informatics, National University of Ireland Galway, Ireland
c
Ryan Institute, National University of Ireland Galway, Galway, Ireland
d
Centre for Marine and Renewable Energy (MaREI), Galway, Ireland

h i g h l i g h t s

 A test apparatus was developed to measure air permeability of granular materials.


 Air permeability is strongly influenced by the compaction and the change of moisture.
 Poorly-graded granular material extracts air much better than well-graded granular material.
 Empirical equations have been proposed.

a r t i c l e i n f o a b s t r a c t

Article history: A series of experimental laboratory tests were conducted to quantitatively investigate the characterisa-
Received 17 November 2017 tion of the T1 Struc and T2 Perm specified granular fill materials for soil depressurisation systems for
Received in revised form 19 April 2018 radon reduction under buildings. The characterisation included determination of grading curves, mea-
Accepted 25 April 2018
surement of air permeability, porosity, and the effective particle diameter of the stone. A test apparatus
Available online 10 May 2018
was developed to measure the air permeability of the granular fill materials under different compaction
degrees. Test results showed that the T1 Struc and T2 Perm specified granular fill materials could be clas-
Keywords:
sified as well-graded and poorly-graded granular materials, respectively. The air permeability and poros-
Granular fill materials
Radon
ity of T1 Struc decrease with the increase in compaction degree and are strongly affected by the change of
Soil depressurisation system moisture content. However, the air permeability of T2 Perm was found to be independent of the com-
CFD paction degree and variation of moisture content. Computational Fluid Dynamic (CFD) simulations were
Compaction validated to simulate the flow behaviour of the T1 Struc and T2 Perm granular fill materials. The primary
parameters for simulating the flow behaviour of the materials were confirmed to be the air permeability,
porosity, and effective particle diameter. Based on the CFD simulation results, the effective particle diam-
eter was found to vary with the compaction degree for T1 Struc, whereas that of T2 Perm was constant at
16.2 mm.
Ó 2018 Elsevier Ltd. All rights reserved.

1. Introduction centration globally is between 5 Bq/m315 Bq/m3, and does not


pose a health risk [2,3]. However, indoor concentrations can be sig-
Radon (222Rn) is a product from the decay chain of uranium nificantly higher, which has been shown to cause lung cancer
(238U) present in soils and rocks. It is a colourless, odourless, taste- through the decay of its short-lived daughter products resulting
less gas which has been identified as a human carcinogen by the pulmonary cell DNA damage. It is estimated that indoor radon
World Health Organisation (WHO), the International Agency for levels account for 9% of deaths from lung cancer and about 2% of
Research on Cancer (IARC) and the U.S. Environmental Protection all deaths from cancer in Europe [4].
Agency (EPA) [1]. Studies have shown that the outdoor radon con- Reduction of the indoor radon concentration is an important
issue for buildings [5]. There are several methods for radon preven-
tion and mitigation, such as the active and passive soil depressuri-
⇑ Corresponding author at: Civil Engineering, College of Engineering & Informat- sation (SD), sealing of surfaces, barriers and membranes,
ics, National University of Ireland Galway, Ireland. ventilation of unoccupied spaces, and ventilation of occupied
E-mail address: jamie.goggins@nuigalway.ie (J. Goggins).

https://doi.org/10.1016/j.conbuildmat.2018.04.210
0950-0618/Ó 2018 Elsevier Ltd. All rights reserved.
214 L.C. Hung et al. / Construction and Building Materials 176 (2018) 213–227

DP Pressure drop [Pa] P2 Pressure recorded by pressure sensor P2 of the test


A Cross section area of air flow [m2] apparatus [Pa]
c Forchheimer constant P2CFD Absolute pressure at outlet chamber in CFD [Pa]
CFD Computational fluid dynamics PCFD Absolute pressure in CFD [Pa]
def, def (Erg) Effective particle size of the granular fill material PB Atmospheric pressure [Pa]
[mm, m] PI Inlet gas pressure [Pa]
DF Darcy-Forchheimer Preading Pressure sensor reading [Pa]
G Mass flow rate [kg/m2.s] Ps Standard reference pressure[Pa]
Gs Specific gravity of material Ptheoretical Theoretical pressure with the water column height [Pa]
k Turbulent kinetic energy [m2/s2] QCFD Inlet air flow rate in CFD simulations [m3/s]
kah Air permeability of the granular fill material [m2] Qin Inlet air flow rate in CFD [m3/s]
kah(ASTM) Air permeability obtained from ASTM (2013) equation Qm Measured air flow rate [m3/s]
[m2] Qav Average air flow rate [m3/s]
kah(wn) Air permeability of the granular fill material at varying SD Soil depressurisation
moisture content [m2] T Kelvin room temperature [K]
kah(wnc) Air permeability of the granular fill material at com- TS Reference Kelvin temperature [K]
paction moisture content [m2] R2 Correlation coefficient
kah(D–F) Air permeability obtained from Forchheimer (1901) v Air flow velocity (m/s)
equation [m2] Vsolid Volume of the solid phase [m3]
kah(Ergun) Air permeability obtained from Ergun (1952) equation Vtotal Total volume of material [m3]
[m2] wnc Compaction moisture content [%]
kas Air permeability of soil [m2] wnc(opt) Optimal compaction moisture content [%]
kp Darcy air permeability [Darcy] wn Test sample moisture content [%]
L Length of the compacted granular fill material (m)
m Correction factor Greek letters
nh Porosity of the granular fill material 3
cdry (95%) 95% maximum dry unit weight [kN/m ]
Nk Normalised air permeability cdry Dry unit weight [kN/m ] 3

Nw Normalised moisture content cw Unit weight of water [kN/m3]


P1 Pressure recorded by pressure sensor P1 of the test l Dynamic viscosity [Pa.s]
apparatus [Pa] x Specific dissipation rate [1/s]

spaces. However, the active and passive SD methods have proven tent, fan speed, and atmospheric conditions [10–16]. Studies have
to be the best options for indoor radon prevention and mitigation shown that the kah of the granular fill material layer significantly
[5–9]. The principle of a SD system is to decrease the pressure in affects the effectiveness of the SD system [7,8,11,13,18,19]. In addi-
the granular fill material layer beneath the floor, thus reversing tion, the granular fill material layer beneath the floor should satisfy
entry path of radon from soil into the building, as well as extracting the bearing capacity and serviceability criteria and should be per-
gas trapped in the granular fill material layer (Fig. 1). meable to ensure the effectiveness of the SD system.
There are several different factors which can influence the effec- Since July 1998, all new buildings constructed in the Republic of
tiveness of the SD system. These factors include the air permeabil- Ireland have been required to include a radon mitigation system,
ity (kah) of granular fill material and native soil (kas) beneath the regardless of the level of the radon concentration [20]. The T1 Struc
floor, cracks in the floor, radon sump size, sealing, moisture con- and T2 Perm granular fill materials are specified as permeable
granular layers beneath the concrete floors and foundations of
buildings in Ireland [21]. These materials are used to increase the
bearing capacity of the foundation, as well as to avoid the expan-
sion of the underfloor caused by the existence of pyrite bearing
material. However, a comprehensive characterisation of these
materials has not been published.
In buildings, the bearing layer beneath the foundation and floor
should be compacted before any superstructure is built. This pro-
cess is to increase the density of the bearing layer, thus, ensuring
Exhaust pipe

the bearing capacity and serviceability criteria. Therefore, under-


standing the degree of permeability of compacted granular fill
material layers beneath the floor and foundation is fundamental.
The degree of permeability of a granular fill material layer is pri-
marily influenced by its kah and porosity (nh) values
Sump
[7,11,13,17,18,19]. In addition, the granular fill material might
Concrete floor Fan box experience a range of moisture contents (wn) after compaction
Ground level during its service life. The moisture content of the granular fill
Hardcore material could be dry, partially saturated or fully saturated
Fill
depending on the environmental conditions, water table, and geo-
Foundation logical conditions.
Native soil
This paper presents results of characterisation of the Irish T1
Fig. 1. A schematic diagram of a house with an SD system. Struc and T2 Perm granular fill materials. The characterisation
L.C. Hung et al. / Construction and Building Materials 176 (2018) 213–227 215

includes the determination of the grading curves, compaction 2.4. Development of air permeability test apparatus for granular fill
degree, kah and nh of the materials. A test apparatus was developed material
to measure the kah value of the materials under different com-
paction levels and variation of moisture content. In addition, Com- A test apparatus was developed to measure the kah and nh val-
putational Fluid Dynamic (CFD) simulations were adopted and ues of T1 Struc and T2 Perm (Fig. 3). A typical test setup is shown in
validated to confirm their reliability for simulating the flow beha- Fig. 4. The working principle of the test apparatus was adopted fol-
viour of T1 Struc and T2 Perm. This validation is important as CFD lowing the standard ASTM-D6539 [24], and two previous works of
simulation can be used as a valuable tool in design of SD systems. Gadgil et al. [6] and BRE [11]. However, the main difference being
that the test apparatus in the current study enables the measure-
2. Materials and methods ment of kah and nh values of the granular fill materials under differ-
ent compaction degrees, as well as varying of wn. The test
2.1. Materials apparatus was made of a plastic pipe with a 150 mm inner diame-
ter, 3250 mm length, and 5 mm thick wall. The inner diameter of
The degree of permeability of T1 Struc and T2 Perm granular fill 150 mm was intentionally chosen as it is similar to the BS standard
materials was investigated in the current study. The materials are compaction mould used in the compaction tests. The pipe was split
formed by crushed stone at two different grades, i.e. well  graded into different segments, including an inlet (1.5 m length) and out-
for T1 Struc (Fig. 2a) and poorly  graded for T2 Perm (Fig. 2b). let (0.25 m length) chambers and three sample segments (0.5 m
Details of the particle size distribution (grading) curves of the gran- length/segment). These segments are joined by four flexible cou-
ular fill material will be discussed in Section 3. plings during testing to prevent air leakage.
Two air pressure sensors with model 26PCBFA6G (Honeywell,
Golden Valley, MN, USA) were used and placed at the inlet and out-
2.2. Particle size distribution test
let chambers to measure the gas pressure difference in the test
sample. Calibration of the pressure sensors was conducted by mea-
From initial CFD simulations, it is found that the grading curve
suring the pressures derived from different fixed water column
is essential to the success of simulating SD systems using CFD. For
heights e.g., a 1 cm water column height corresponds to 98.1 Pa.
this reason, sieve analysis was performed to determine the grading
The water column heights were varied from 0.5 cm to 150 cm.
curves for the T1 Struc and T2 Perm granular fill materials. The test
Based on the calibration results, the following correlation was
was conducted in accordance with the standard BS 1377  Part 2
introduced to obtain the theoretical pressure measured in the tests,
[22].
as follows:

2.3. Compaction test Ptheoretical ¼ m  Preading ð1Þ

The field compaction degree for a material can be determined where Ptheoretical is the theoretical pressure associated with the
by referring to an optimum compaction degree in the laboratory. water column heights, Preading is the pressure recorded by the pres-
The compaction degree is defined by the relationship between sure sensors, m is the correction factor and m = 1.304 and 1.1073 for
the dry unit weight (cdry) of compacted material and its initial the sensors P1 and P2, respectively. These m factors were obtained
compaction moisture content (wnc). In Ireland, the field com- with a correlation factor between Ptheoretical and Preading of R2 =
paction for T1 Struc and T2 Perm is described in the Irish Standard 0.999.
I.S.888 [21]. To measure the flow rate (Qm) through the test samples, three
Laboratory compaction tests for the T1 Struc and T2 Perm were air flowmeters were used to measure different flow rates i.e., from
performed following the standard BS 1377  Part 4 [23] using the 4.68  106 m3/s to 3.34  103 m3/s. These air flowmeters are
BS standard compactor. At the beginning of the test, the material VFA-3 model (Dwyer, Indiana, USA), PRM-1 (Cole-Parmer, Hanwell
was mixed with water using a 0.5 m3 revolving drum mixer. The London, UK) and RF1MHAI46 (Influx Measurements, Hamshire,
compaction moisture content wnc was controlled intentionally by UK). In addition, a velocity sensor model F900 (Degree Controls,
adjusting a fixed amount of water. The material was then com- Milford, NH, USA) was used to measure higher Qm (i.e., Qm > 3.34
pacted with the compactor. In addition, three different small sam-  103 m3/s). A compressed air regulator was used to control Qm
ples of the material were taken to determine wnc for each test by of the pressurised air at the air inlet chamber.
drying these samples using an oven dryer for 24 h. The optimum Air leakage in the test apparatus was carefully examined by
wnc was determined by increasing gradually the amount of water introducing pressurised (20 kPa) air into the test apparatus when
in the materials. it was empty. After pressurising the air into the test apparatus,

Fig. 2. Photos of (a) T1 Struc and (b) T2 Perm granular fill materials. T1 Struc and T2 Perm are formed by the crushed stone. T1 Struc is standardised with grade varied from
032 mm. T2 Perm is standardised with grade of 4/40 mm according to NSAI [21].
216 L.C. Hung et al. / Construction and Building Materials 176 (2018) 213–227

Fig. 3. Diagram of the air permeability test apparatus used in this study (not to scale).

Experimental tests were performed as follows. A fixed air flow


rate Qm is pressurised into the air inlet chamber using an auto-
mated compressed air generator. This fixed Qm is controlled by a
compressed air regulator. The Qm is obtained by reading the air
flow rate values from an air flow meter or air velocity meter. After
a fixed Qm is specified, the regulator is locked to ensure a constant
air follow rate is maintained during testing.
Each test is recorded for approximately 350 s. The P1 and P2 will
increase gradually and an equilibrium condition is observed. At
this stage, the P1 and P2 are constant (Fig. 5), and constant values
of P1 and P2 are averaged. The average values of P1 and P2 are used
to calculate the pressure difference as DP = P1P2. This DP is
related to the degree of permeability of the material; the lower
the degree of permeability, the higher the DP and vice versa. Each
test sample is tested at least three times to minimize uncertainty.

2.5. Calculation of air permeability

Fig. 4. Typical test setup of the air permeability test apparatus. Four well  known equations were adopted to determine the
kRahR of the granular fill materials. These equations are the Darcy
all valves were closed and the pressure trapped inside the appara- [25], Ergun [26], ASTM [24] and Darcy–Forchheimer [27] equa-
tus was monitored for 5 min. No pressure drop was observed, tions. Each equation is described briefly below.
which means that the test apparatus is entirely sealed although Darcy [25] proposed a simple linear equation to present the
at high pressure. However, tests were conducted at a maximum flow of fluid through a porous medium using his experimental
pressure of 12 kPa ± 1 kPa at the inlet gas chamber. Measurements tests. The Darcy equation is written as follows:
were interpreted and recorded with the National Instruments NI
9237 data logger and the LabVIEW program version 2016 (National A  kah DP
Qm ¼  ð2Þ
Instruments, Austin, Texas, USA). l L

Fig. 5. Example of experiment time for test sample T1#4-1 at Qm = 1.67  104 m3/s.
L.C. Hung et al. / Construction and Building Materials 176 (2018) 213–227 217

or step 1; the material was mixed with water using a 0.5 m3 revolv-
ing drum mixer. The compaction moisture content wnc
DP l
¼ v ð3Þ was intentionally controlled by adding a fixed volume of
L kah water.
step 2; the material derived from Step 1 was compacted into the
where Qm is the measured air flow rate (m3/s), kah is the air perme-
pipe segment following the standard BS 1377  Part 4
ability of the granular fill material (m2), l is the dynamic viscosity
[23]. In addition, three different samples of the material
of the air (Pa.s), A is the cross section area of flow (m2), DPis the
obtained from Step 1 were taken to determine wnc for each
pressure drop and DP = P1  P2 (Pa), P1 and P2 are the measured
test by drying these samples using an oven dryer for 24 h.
pressures at the air inlet chamber and air outlet chamber respec-
step 3; sample segments derived from Step 2 were connected to
tively, L is the length of the compacted granular fill material (m),
each other using flexible couplings. Primary tests showed
v is the air flow velocity (m/s).
that there is no difference in the kah values if the material
Ergun [26] reviewed the flow of fluid through the granular por-
is compacted separately (i.e., sample pipe segments are
ous material. He stated that pressure loss is a result of the kinetic
connected after compaction) or compacted connected
and viscous energy losses. Ergun suggested the following
(i.e., pipe segments are connected before compaction).
equations.

DP ð1  nh Þ2 lV ð1  nh ÞGV Each test is defined by a test ID, as shown in Table A 1. For exam-
¼ 150 þ 1:75 3 ð4Þ ple, the test ID T1#8–1 is the 1st test of sample #8 for the T1 Struc
L 3 2 nh def ðErgÞ
n def ðErgÞ
h material with compaction moisture content of wnc = 3.94%. The test
ID T1#8–2 is the 2nd test of sample #8 derived by drying the sample
1 ð1  nh Þ2 T1#8–1 using an oven dryer for 24 h to have a moisture content of
¼ 150 ð5Þ
kah 2
n3 def ðErgÞ wn = 1.72%. The test ID T1#8-3 has much higher moisture content at
h
wn = 4.17%. This high moisture content was obtained by filling a
where def(Erg) is the Ergun effective diameter (m) that can be calcu- fixed volume of water (e.g., 300 ml–400 ml) into sample T1#8-2.
lated as def(Erg) = def/1000, def is the effective diameter which can be After filling, the sample was kept standing vertically for 4 h. This
determined from the grading curve (mm), G is the mass flow rate step is to ensure that water inside the sample has enough time to
(kg/m2s). distribute by either capillary action or gravitational force. These
ASTM [24] gives equations to determine kah of porous materials procedures were repeated identically for other samples to investi-
when kah is between 1.0  1015 m2 and 1.0  1010 m2. The data gate the effect of the change of moisture content on kah.
required to calculate kah are calculated average flow rate (Qav)
and DP, and should lie in between ±25%. The equation is valid only 2.7. Determination of nh values
with relatively small flow rate, i.e., Qav = 1.67  105 m3/s. The
ASTM equations are as follows: Theoretically, the porosity of a porous material influences its
degree of permeability. The porosity can be determined as the ratio
Ps T of total volume of pore space within the media divided by the total
Q av ¼ Q m   ð6Þ
ðPI þ P B  D2PÞ T S volume of material, as shown in Eq. (10).
V total  V solid cdry
Q L nh ¼ ¼1 ð10Þ
kP ¼ AV   l  1:013  1012 ð7Þ V total Gs c w
DP A
where Vsolid is the volume of the solid phase (m3), Vtotal is the total
kah ¼ 9:869  10 13
 kp ð8Þ volume of material (m3), cdry is the dry unit weight of the material
(kN/m3), Gs is the specific gravity of material, cw is the unit weight
where Qav is the average air flow rate (m3/s), PB barometric pressure of water (cw = 9.81 kN/m3).
or atmospheric pressure (Pa), PB was taken to be 101325 Pa, PI is the In this study, the nh values of the T1 Struc and T2 Perm were
air inlet pressure which is equal to the P1 in this study, Ps is the obtained by either measuring the volume of pore space of the com-
standard reference pressure (Ps = 101300 Pa), TS is the reference pacted granular fill material or using Eq. (10). To measure the vol-
Kelvin and TS = 293 K, T is the Kelvin room temperature, kP is the ume of pore space, the material is compacted into the pipe
air Darcy permeability (Darcy). segment of 500 mm, which is the same as that shown in Fig. 4.
Forchheimer [27] suggested an equation that considers high Sample preparation for this measurement is the same as that dis-
flow rate in the porous material. The equation is also known as cussed in Section 2.6. After sample preparation had been done,
the DarcyForchheimer (DF) equation. Forchheimer observed the material was taken out and poured into a known volume tank.
that the relationship between pressure drop and flow rate in the Thereafter, water was immediately filled into this tank. By measur-
specimen is not linear. It is because there is an inertial effect to ing the amount of this filled water, the volume of the solid part of
the kinetic energy of the flow at high flow rate. Forchheimer added material was obtained by subtracting the total volume of the tank
the internal effect term into the Darcy [25] equation, as follows: to that of fill water. However, this measurement was done only
with the T1 Struc at wnc  3% and T2 Perm. The reason is that a sig-
DP l l
¼ v þc  v2 ð9Þ nificant amount of fine content of T1 Struc at dryer condition will
L kah kah be a cause of water absorption. Thus, measurement of the solid vol-
ume is incorrect. Therefore, a value of approximately Gs = 2.69 was
where c is the Forchheimer constant, which depends on the
obtained from back-calculation to determine the nh values for
material.
dryer compacted material.

2.6. Test sample preparation 2.8. CFD simulation

Test samples were prepared at different wnc and wn values. The The CFD simulations were performed in this study to confirm
preparation procedure is as follows: their reliability for the design of the SD system. Several studies
218 L.C. Hung et al. / Construction and Building Materials 176 (2018) 213–227

have successfully applied numerical models to simulate the flow 3.3. Effect of wnc and wn on the kah values
behaviour of material, as well as the SD system [6,7,11]. However,
these models were developed in-house and are not freely available. Figs. 8 and 9 show examples of the test results for the samples
Thus, access to these models is limited. In addition, a widely used T1#4-1 and T2#3–1, respectively. The results plotted in these fig-
numerical program is preferred in design. Therefore, a well-known ures were selected randomly from 55 test specimens. Test results
CFD package Ansys Fluent version 15.0 [28] was used in this study. for other samples show a similar trend, thus are not presented
Initial CFD results of this study have shown that kah, nh, and def are here. It can be seen from the figures that the D-F equation [27]
the three key parameters of the material required for simulations. shows excellent agreement with the test results inducing a corre-
These parameters can be obtained experimentally in the laboratory lation factor R2 of 0.99 for both materials, whereas the Ergun equa-
and will be discussed in the next sections. All simulations were tion [26] gives R2 of 0.86 and 0.89 for the T1 Struc and T2 Perm,
performed under transient analysis obeying the Reynolds- respectively. The ASTM [24] and Darcy [25] equations are applica-
Averaged Navier-Stokes (RANS) Realizable kx turbulence model. ble for low flow rate only, i.e., laminar air flow. Therefore, the kah
The RANS Realizable kx model provides more accuracy in com- values derived from the D-F equation [27] is recommended in
parison with the other RANS models [29]. Definition of numerical practical design, as this accurately represents a wider range of
mesh and the boundary conditions will be discussed in Section 4. the flow rate that might be experienced in a SD system during its
service life.
The kah values derived from all equations are presented in
3. Results of experimental tests and discussion
Table A. 1. Fig. 10 shows a variation of kah values obtained from
the D-F equation [27] measured at different wnc and is denoted
3.1. Particle size distribution
as kahðwnc Þ in the figure. For the T1 Struc, the kah values decrease
with the increase in wnc, whereas those of the T2 Perm do not vary
Fig. 6 shows the particle size distribution curves for each gran-
significantly with the variation of wnc. The reason is that the T1
ular fill material. Based on the test results, the T1 Struc and T2
Struc is a well-compacted material. When it is compacted, the pore
Perm can be classified as well-graded and poorly-graded granular
space tends to reduce, thus reducing its kah. In contrast, the T2
fill material, respectively. The grading curves of three British gran-
Perm is not well-compacted with the small-scale laboratory com-
ular fill material types (i.e., MOT type 1 sub-base, 20 mm, and 40
paction test apparatus. Therefore, the kah values do not show a
mm granular fill materials) published in the work of the BRE [11]
clear trend with the variation of wnc. An average value of kahðwnc Þ
were also plotted to compare with those of the T1 Struc and T2
Perm. It can be seen from the figure that the particle size distribu- of approximately (22.5 ± 7.2)  109 m2 can be used for the T2
tion curve of the MOT-1 is similar to that of T1 Struc. A similar Perm, whereas that of T1 Struc can be interpolated using Eq. (11).
observation was found between British 20 mm, 40 mm materials
and T2 Perm. This finding suggests that characterisation of the Irish kahðwnc Þ ¼ 20  109 e1:25wnc ð11Þ
T1 Struc, T2 Perm granular fill materials could apply to the British
MOT-1, 20 mm, 40 mm granular fill materials and vice versa.
for 1%  wnc  4.5%.
Fig. 11 presents the effect of wn on kah of the T1 Struc material in
3.2. Compaction test results k
terms of normalisation N k ¼ k ahðwn Þ and N w ¼ wwncn . Nk of 1 represents
ahðwnc Þ

Compaction test results for the T1 Struc and T2 Perm are shown the air permeability of the granular fill material measured at wnc,
in Fig. 7. It is observed that the optimum wnc(opt) for compaction of i.e., Nw = 1. ‘‘Dry side” means that the moisture content of the
the T1 Struc is approximately 5.7% inducing a maximum dry unit material was reduced by drying the sample with the oven dryer
weight cdry(max) of about 23 kN/m3 (Fig. 7a). However, optimum in 24 h, whereas ‘‘wet side” presents the increase of moisture con-
compaction degree for the T2 Perm is relatively independent of tent by adding water into the sample. As shown in the figure, wn
moisture content (Fig. 7b). The reason is that the T2 Perm is a strongly influences kah, as expected. For example, if wn reduces
poorly  graded material. Therefore, it is relatively difficult to com- below wnc, then kah generally increases as the moisture content
pact this type of material with the small  scale laboratory com- reduces, up to 2.5 times kahðwnc Þ on the dry side. On the other hand,
paction test apparatus. This assumption will be confirmed using adding moisture to the granular fill material after compaction
the large scale experimental test in the next phase of this project. decreases the permeability of the fill material with an average

Fig. 6. Particle size distribution of Irish T1 Struc, T2 Perm, and British granular fill materials.
L.C. Hung et al. / Construction and Building Materials 176 (2018) 213–227 219

Fig. 7. Compaction test results for T1 Struc and T2 Perm granular fill materials.

Fig. 8. Relationship between DP and air flow rate (Sample: T1#4-1). These test results were selected randomly from 49 measurements for T1 Struc. Test results for other
samples show similar trend thus are not presented here.

value of Nk being 0.37 for specimens tested up to 1.8 times wnc (i.e. moisture content. For material that is susceptible to increased
‘‘wet side”). moisture content during the service life of the building, e.g. due
The above finding shows that the degree of permeability during to presence of water table within depth of material, it is suggested
its service life of the T1 Struc is dependent on the change of its to assume the permeability of the granular fill material for the
220 L.C. Hung et al. / Construction and Building Materials 176 (2018) 213–227

Fig. 9. Relationship between DP and air flow rate (Sample: T2#3-1). These test results were selected randomly from 6 measurements for T2 Perm. Test results for other
samples show similar trend thus are not presented here.

Fig. 10. Variation of kah(wnc) with wnc for T1 Struc and T2 Perm.

an SD system, it is recommended to use the permeability of the


material at compaction.

3.4. Test results of porosity

As discussed previously, the nh values of the T1 Struc and T2


Perm were obtained by either measuring the volume of pore space
of the compacted granular fill material or using Eq. (10). Variations
of the nh values of T1 Struc and T2 Perm with wnc and cdry are pre-
sented in Fig. 12, Fig. 13 and Table A.1. For T1 Struc (Fig. 12a), it is
observed that the nh values are constant at approximately 0.23
when wnc  2. At 2% <wnc  6%, the nh values decrease, and slightly
increase when wnc > 6%. These trends are expected, as cdry
increases up to the optimum moisture content (wnc(opt)) of approx-
imately 6%, while it decreases when the moisture content is larger
than the optimum moisture content (Fig. 7). As compaction mois-
ture content continuously increases, water will fill the void space
Fig. 11. Effect of wn on kah for the T1 Struc. within the compacted material which causes a low increment of
nh at wnc > 6%. On the other hand, the nh values of T2 Perm are rel-
atively constant at approximately 0.37. It is because the T2 Perm is
design of an SD system is 0.37 times the initial permeability of the relatively difficult to compact with the standard compaction ham-
compacted material (bottom right quadrant of Fig. 11). On the mer in the laboratory (Fig. 12b).
other hand, if the moisture content of the material reduces after Fig. 13 shows a relationship between nh and cdry obtained from
compaction, e.g. as a result of drying, the permeability of the mate- the test results. As can be seen, the nh values decrease linearly with
rial is expected to increase. However, for the conservative design of the increase in the cdry as expected. Porosity values calculated
L.C. Hung et al. / Construction and Building Materials 176 (2018) 213–227 221

Fig. 12. Variation of nh and wnc for T1 Struc and T2 Perm.

Alternatively, based on Fig. 13, nh values can be estimated from


the dry unit weight of (cFidry) and specific gravity (Gs) of the T1
Struct material as follows:
 
cdry
nh ¼ 0:95 1  when 2% < wnc  8% ð12cÞ
Gs cw
For T2 Perm, the nh value was found to be constant for com-
pacted moisture contents less than 1.4% and can be expressed as
(Fig. 12b):

nh ¼ 0:37 when wnc  1:4% ð13Þ

4. Results of CFD simulation

4.1. Numerical model and boundary conditions

Fig. 13. Relationship between variation of nh and cdry for T1 Struc. Fig. 14 shows a typical numerical mesh of the domain for the air
permeability test apparatus used in this study. The boundary con-
using Eq. (10) are plotted to compare with those from physical ditions were specified based on the test setup shown in Figs. 3 and
measurements. It is found that the porosity calculated from the fit- 4. An inlet boundary was set at the left end of the model, whereas
ting curve (dash line in Fig. 13) is approximately 0.95 times Eq. the outlet is at the right end. The wall boundary was defined along
(10). the external wall of the apparatus. Hexahedral element type was
Based on the test results presented in Fig. 12a, equations are applied for meshing. The optimum mesh size of 0.01 m was deter-
proposed to evaluate the nh values of T1 Struc at different wnc, as mined so that the numerical mesh size does not affect the results.
follows. Residual convergence criterion for velocity was set to be 104, and
For T1 Struc: the scaled energy residual was 106. These criteria were to ensure
nh ¼ 0:23 when wnc  2% ð12aÞ there were enough iterations to achieve a numerical converged
solution.
nh ¼ 0:28  0:03wnc when 2% < wnc  wncopt ð12bÞ As discussed, the kah, nah and def parameters are required for
CFD simulations. The def value can be determined from the grading

Fig. 14. Typical numerical mesh for the test apparatus.


222 L.C. Hung et al. / Construction and Building Materials 176 (2018) 213–227

curve. The term def is adopted to represent the inertial loss of flow 4.2. Validation of the CFD simulation
through a granular specimen [26]. This inertial loss can be calcu-
lated corresponding to the second term in the Ergun equation The accuracy of the CFD simulations in this study was validated
[Eq. (3)]. The air used in the CFD simulations has a density of by comparing the CFD results with experimental test results from
1.225 kg/m3 with a viscosity of 1.7894  105 kg/(m.s). published works of Gadgil et al. [6] and BRE [11].

Fig. 15. Comparison of CFD simulation and experimental test for poorly-graded granular fill materials (data adopted from Gadgil et al. [6] and BRE [11]).
L.C. Hung et al. / Construction and Building Materials 176 (2018) 213–227 223

4.2.1. Gadgil et al. (1991) laboratory test and nh values were 1.3  107 m2 and 0.4, respectively. The air
Gadgil et al. [6] measured the air permeability of the 3/400 (ca. pressure drop across the material versus the flow velocity was
19 mm) rounded granular fill materials. A test apparatus was made numerically simulated and compared with the test results. It is
of wood with dimensions of 0.222 m high  0.254 m wide  2.43 found that the CFD simulations simulate the test results correctly
m long. An average def value was assumed to be 11 mm. The kah with an error of approximately ± 2.5% (Fig. 15a).

Fig. 16. Comparison of CFD simulation and experimental test for well-graded MOT Type 1 (data adopted from BRE [11]). (a) at def = 12 mm, the CFD simulation does not
simulate well the test results. (b) at def = 6.5 mm, the CFD simulation yields well the test results with an error of approximately ± 2.5%.

Fig. 17. CFD simulation example of pressure difference (PCFD-P2(CFD)) contour developed along the length of the test apparatus with the compacted T1 Struc granular fill
material (T1#1-1) for Qin = 2.51  104 m3/s. The pressure difference is relative to the absolute pressure at outlet air chamber P2(CFD).

Fig. 18. CFD simulation example of pressure difference PCFD-P2(CFD) developed along the length of the test apparatus containing compacted T1 Struc and T2 Perm materials
(Specimens T1#1-1 and T2#1-1, respectively). At the same inlet air flow rate Qin, the T1 Struc induces a much higher pressure difference in comparison with that of the T2
Perm, suggesting that the ability to extract air from the T2 Perm is much better than from the T1 Struc.
224 L.C. Hung et al. / Construction and Building Materials 176 (2018) 213–227

4.2.2. BRE (1998) laboratory test graded material can be taken directly as d50 in the grading curve.
BRE [11] reported a series of laboratory tests to measure kah of The d50 value taken from the grading curve does not reflect the
different types of well-graded and poorly-graded granular fill def value for well-graded material. This issue will be discussed fur-
materials used for SD systems in the UK. Poorly-graded materials ther in Section 4.3.
were 10 mm, 20 mm single sized, 20 mm graded, 40 mm, and a
well-graded material was MOT type 1 sub-base. The test apparatus 4.3. Numerical simulation of experimental tests of this study
was the same as that reported in Gadgil et al. [6]. Measurements
were performed for the uncompacted material only. The kah values As examined in Section 4.2, the def value of the well-graded
were obtained to be between 2.1  108 m2 and 2.7  107 m2, granular fill material could not be obtained as d50 from the grading
whereas the nh values varied between 0.32 and 0.46 according to curve. Therefore, the def value of T1 Struc needs to be confirmed.
material type. The def was taken as d50 on grading curves. The def value of T2 Perm is obtained as d50 of 16.2 mm from its
Values of kah, nh and def are presented in Fig. 15 (be) and 16 (a grading curve (Fig. 6b). A total of circa 300 CFD simulations were
and b). The CFD results are in good agreement with the test results performed to match the CFD results with the measurements of
for 10 mm, 20 mm single sized, 20 mm graded, 40 mm with an T1 Struc and T2 Perm. This step will enable the def values to be
error of between 2% and 4% (Fig. 15 (be)). However, the CFD sim- obtained for T1 Struc. The kah and nh values are taken from
ulations for the MOT type 1 sub-base (Fig. 16a) failed to yield the Table A.1.
test results accurately. It is because the d50 value on the grading Fig. 17 shows an example of the results from the CFD simulation
curve of the MOT type 1 does not well reflect def of this material. for the pressure difference contour developed along the length of
Therefore, additional simulations were performed to find the cor- the test apparatus for T1 Struc granular fill material (T1#1-1) with
rect def value for MOT type 1. It was found that def of 6.5 mm gives an inlet air flow rate of Qin = 2.51  104 m3/s. The inlet air flow
excellent agreement between the CFD and the test results with an rate was numerically pressurised with air entering the air inlet
error of approximately 2.5% (Fig. 16b). chamber (Fig. 17). The pressure difference was defined as the abso-
From the validation, it can be concluded that the CFD simula- lute pressure P(CFD) at the point of interest along the length of the
tion is reliable and could be used to analyse the airflow behaviour sample relative to the absolute pressure at the outlet air chamber
of the granular fill materials. In addition, the def value for poorly- P2(CFD). Fig. 18 shows examples of this pressure difference along the

Fig. 19. Comparison of CFD simulation and test results for T1 Struc and T2 Perm. The test results were selected randomly. Test results for other samples show similar trends,
thus are not presented here.
L.C. Hung et al. / Construction and Building Materials 176 (2018) 213–227 225

length of the test apparatus for two samples T1#1-1 and T2#1-1, at for 1%  wnc  4:5%:
inlet flow rates of Qin = 2.51  104 m3/s and Qin = 4.18  104 m3/
s. As shown in the figure, the pressure difference decreases linearly kah ¼ 6:51def ð15Þ
along the length of the test apparatus. In addition, the pressure
for 1%  wnc  4.5%.
developed in the T2 Perm is significantly lower than that in the
T1 Struc material. This is because the T2 Perm material has much
higher kah value in comparison with that of the T1 Struc material. 5. Implication of practice
It suggests that the ability to extract air from the T2 Perm is much
better than from the T1 Struc. An important aspect of this study is a method to determine kah
Fig. 19 compares the CFD simulations and test results for T1 values of granulated fill material under compaction and the deter-
Struc and T2 Perm. As can be seen, the CFD simulation agrees well mination of such values for specified granulated fill material in Ire-
with the experimental test results. The def values were determined land. No standard or regulation describes detailed design of the SD
by backcalculating the test results using the CFD simulations. systems in Ireland. Thus, selection of the permeable layer for a SD
Variation of the def values with wnc is shown in Fig. 20. It is found system is ideally based on practical experience of a designer rather
that the def values decrease significantly with the increase in wnc than a scientific view. This study will enable engineers to select an
until wnc  3.5%. The effect of def on kah of the T1 Struc material appropriate wnc value, thus kah value, for a compacted fill layer
is evident, as the kah values increase linearly with the increase in under footings or concrete floors.
def (Fig. 21). In the field, the compaction degree of the granulated fill mate-
Based on the CFD results, a simple equation, which can be used rial could be higher than those reported in this study. This higher
to evaluate the def value, has been proposed (Eq. (14)). Alterna- compaction degree might cause lower kah and nah in comparison
tively, def values can be used to interpolate kah as shown in (Eq. with those presented in Table A.1. Therefore, further investigation
(15)). of the compaction degree of T1 Struc and T2 Perm is required with
The equation is expressed as the function of the compaction large scale compactors, e.g., field compaction or large-scale tests.
moisture content wnc. The CFD simulations have been validated to be reliable for sim-
ulating the air flow behaviour of the granulated fill material. CFD
def ¼ 1:56e0:83wnc ð14Þ results reported in this study were focused only on input parame-
ters for the granulated fill material. A more comprehensive simula-
tion of a SD system should include other factors, such as cracks, the
permeability of concrete, the thickness of granular fill material
layer, the thickness of the slab, radon sump size, and pipe size [30].

6. Conclusion

A series of laboratory tests and CFD simulations were conducted


in this study to extensively examine the characteristics of the spec-
ified granular fill material for SD systems in Ireland, namely T1
Struc and T2 Perm granular fill materials. The characterisation
included determination of grading curves, compaction test and
measurement of the kah, nh and def values of the granular fill mate-
rials, also known as hardcores.
From sieve analysis results, the T1 Struc and T2 Perm materials
can be classified as well-graded and poorly-graded granular fill
materials, respectively. The effective particle diameter def required
for CFD simulations can be taken from grading curves as d50 for T2
Perm. However, def of T1 Struc varies according to compaction
Fig. 20. Variation of def with wnc for T1 Struc. degree. In addition, compaction test results showed that T1 Struc
is relatively easier to compact in comparison with T2 Perm. The
optimum wnc value of T1 Struc is approximately 5.75%, whereas
that of T2 Perm is unclear.
A test apparatus has been developed to measure the kah and nh
values of granular fill materials under compaction. Based on the
test results, it was observed that the kah and nh of T1 Struc decrease
with the increase in wnc, whereas those of T2 Perm are relatively
independent.
The kah, nh and def were confirmed to be the three key parame-
ters of the granular fill material for the CFD simulations. CFD sim-
ulations with transient analysis obeying the RANS Realizable k-x
turbulence model are reliable in design of SD systems with the
T1 Struc and T2 Perm granular fill materials. These parameters
can be determined using the graphs, proposed equations or
Table A.1 reported in this study.
The investigation procedures developed in this study can be
adopted to characterise other granular fill materials. The results
presented in this study apply to the conditions considered only.
However, the results might apply to other granular fill materials
Fig. 21. Relationship between def and kahðwnc Þ for T1 Struc. which have similar characteristics to T1 Struc and T2 Perm.
226 L.C. Hung et al. / Construction and Building Materials 176 (2018) 213–227

Author declaration direct communications with the office). He/she is responsible for
communicating with the other authors about progress, submis-
We wish to confirm that there are no known conflicts of interest sions of revisions and final approval of proofs. We confirm that
associated with this publication and there has been no significant we have provided a current, correct email address which is acces-
financial support for this work that could have influenced its out- sible by the Corresponding Author and which has been configured
come. We confirm that the manuscript has been read and approved to accept email from Jamie.goggins@nuigalway.ie.
by all named authors and that there are no other persons who sat-
isfied the criteria for authorship but are not listed. We further con-
firm that the order of authors listed in the manuscript has been Acknowledgments
approved by all of us.
We confirm that we have given due consideration to the protec- This study is a part of the OPTI-SDS project, which is funded by
tion of intellectual property associated with this work and that the Irish Environmental Protection Agency through Grant No.
there are no impediments to publication, including the timing of 2015  HWMS5. The authors also wish to acknowledge the
publication, with respect to intellectual property. In so doing we DJEI/DES/SFI/HEA Irish Centre for High-End Computing (ICHEC)
confirm that we have followed the regulations of our institutions for the provision of computational facilities and support required
concerning intellectual property. for the CFD simulations. The second author would like to acknowl-
We understand that the Corresponding Author is the sole con- edge the support of Science Foundation Ireland through the Career
tact for the Editorial process (including Editorial Manager and Development Award programme (Grant No. 13/CDA/2200).

Appendix Table A 1. Summary of the test results for T1 Struc and T2 Perm.

No. Test ID wn (%) cdry (kN/m3) nh kah  1010 (m2)


Darcy (1856) Ergun (1952) ASTM (2013) D-F (1901)
Compacted T1 Struc granular fill material
1 T1#1-1⁄ 4.26 22.46 0.140 1.69 ± 0.41 0.72 ± 0.04 1.78 ± 0.22 1.78 ± 0.14
2 T1#1-2 3.81 22.46 0.140 1.86 ± 0.25 0.70 ± 0.02 1.95 ± 0.17 1.81 ± 0.42
3 T1#2-1⁄ 5.58 23.38 0.100 Difficult to flow due to extremely low kah
4 T1#3-1⁄ 3.25 21.02 0.200 5.25 ± 0.32 3.33 ± 0.16 5.64 ± 0.38 6.16 ± 0.45
5 T1#3-2 2.68 21.02 0.200 6.51 ± 0.23 3.89 ± 0.10 6.88 ± 0.22 6.15 ± 0.19
6 T1#4-1⁄ 3.73 22.40 0.142 2.27 ± 0.18 0.90 ± 0.03 2.40 ± 0.12 3.78 ± 0.30
7 T1#4-2 0.87 22.40 0.142 2.96 ± 0.22 1.26 ± 0.12 3.13 ± 0.32 8.47 ± 0.15)
8 T1#4-3 4.25 22.40 0.142 0.68 ± 0.022 0.35 ± 0.03 7.05 ± 0.17 0.41 ± 0.03
9 T1#4-4 3.81 22.40 0.142 2.22 ± 0.31 0.73 ± 0.02 2.27 ± 0.26 1.64 ± 0.13
10 T1#5-1⁄ 4.34 21.70 0.175 0.97 ± 0.042 0.58 ± 0.04 1.08 ± 0.41 0.75 ± 0.06
11 T1#5-2 2.63 21.70 0.175 1.65 ± 0.12 1.03 ± 0.21 7.92 ± 0.24 1.53 ± 0.35
12 T1#6-1⁄ 3.40 21.81 0.164 1.64 ± 0.35 0.86 ± 0.05 1.73 ± 0.33 1.76 ± 0.27
13 T1#6-2 2.55 21.81 0.164 2.57 ± 0.21 1.62 ± 0.13 2.73 ± 0.32 2.51 ± 0.16
14 T1#6-3 4.8 21.81 0.164 0.90 ± 0.042 0.49 ± 0.03 0.95 ± 0.03 0.70 ± 0.02
15 T1#7-1⁄ 3.17 21.74 0.167 2.26 ± 0.37 1.09 ± 0.30 2.31 ± 0.28 1.83 ± 0.45
16 T1#7-2 2.1 21.74 0.167 5.63 ± 0.15 3.21 ± 0.12 5.13 ± 0.23 4.06 ± 0.35
17 T1#7-3 4.70 21.74 0.167 0.35 ± 0.033 0.35 ± 0.02 0.35 ± 0.03 0.35 ± 0.04
18 T1#8-1⁄ 3.94 22.53 0.136 1.07 ± 0.35 0.38 ± 0.04 1.12 ± 0.28 0.98 ± 0.06
19 T1#8-2 1.72 22.53 0.136 2.79 ± 0.55 1.79 ± 0.37 2.99 ± 0.16 2.76 ± 0.15
20 T1#8-3 4.17 22.53 0.136 0.29 ± 0.044 0.12 ± 0.03 0.24 ± 0.01 0.36 ± 0.02
21 T1#9-1⁄ 3.66 22.46 0.140 0.90 ± 0.011 0.34 ± 0.02 0.96 ± 0.01 0.84 ± 0.02
22 T1#9-2 2.52 22.46 0.140 2.08 ± 0.22 1.02 ± 0.32 2.18 ± 0.13 2.84 ± 0.27
23 T1#9-3 4.00 22.46 0.140 0.23 ± 0.015 0.19 ± 0.05 0.22 ± 0.03 0.20 ± 0.03
24 T1#10-1⁄ 3.61 21.87 0.162 2.00 ± 0.19 0.99 ± 0.03 2.15 ± 0.17 1.69 ± 0.15
25 T1#10-2 2.17 21.87 0.162 3.90 ± 0.25 1.64 ± 0.15 4.14 ± 0.18 3.34 ± 0.23
26 T1#11-1⁄⁄ 3.12 17.20 0.353 183 ± 20.0 374 ± 10.0 NA 248 ± 18.0
27 T1#12-1⁄ 2.97 20.3 0.23 6.06 ± 1.4 4.87 ± 0.10 6.42 ± 1.3 10.5 ± 1.0
28 T1#13-1⁄ 2.99 20.04 0.23 3.91 ± 0.6 3.37 ± 0.10 4.18 ± 0.7 3.59 ± 0.10
29 T1#14-1⁄⁄ 2.70 17.40 0.35 301 ± 30.0 238 ± 40 NA 238 ± 20.0
30 T1#15-1⁄ 2.70 21.33 0.20 2.04 ± 0.4 1.51 ± 0.10 2.12 ± 0.4 1.92 ± 0.20
31 T1#16-1⁄ 2.87 22.18 0.18 1.12 ± 0.14 9.74 ± 0.17 NA 15.5 ± 30.0
32 T1#17-1⁄ 2.60 21.73 0.19 130 ± 32 108 ± 22.0 NA 87.6 ± 1.70
33 T1#18-1⁄ 0.97 21.04 0.23 56.4 ± 1.1 93.7 ± 2.80 NA 80.8 ± 1.90
34 T1#19-1⁄ 5.99 23.68 0.13 Difficult to flow due to extremely low kah
35 T1#20-1⁄ 3.66 22.19 0.18 2.86 ± 0.24 1.80 ± 0.11 3.04 ± 0.25 2.74 ± 0.06
36 T1#20-2 2.40 22.19 0.18 4.17 ± 0.55 2.62 ± 0.22 4.43 ± 0.13 3.99 ± 0.34
37 T1#20-3 3.75 22.19 0.18 2.52 ± 0.35 1.59 ± 0.22 2.68 ± 0.31 2.42 ± 0.23
L.C. Hung et al. / Construction and Building Materials 176 (2018) 213–227 227

Appendix (continued)

No. Test ID wn (%) cdry (kN/m3) nh kah  1010 (m2)


Darcy (1856) Ergun (1952) ASTM (2013) D-F (1901)
38 T1#21-1⁄ 1.32 21.05 0.22 38.5 ± 4.7 38.4 ± 10.0 NA 34.0 ± 5.0
39 T1#22-1⁄ 2.54 21.36 0.21 17.5 ± 2.3 13.7 ± 2.0 NA 24.3 ± 4.4
40 T1#22-2 3.11 21.36 0.21 7.56 ± 0.45 5.92 ± 0.33 NA 10.5 ± 2.5
41 T1#23-1⁄ 2.14 21.14 0.22 11.6 ± 4.0 12.3 ± 5.0 NA 25.0 ± 3.9
42 T1#24-1⁄ 2.83 21.44 0.21 22.2 ± 3.2 22.9 ± 4.2 NA 20.2 ± 3.4
43 T1#24-2 3.87 21.44 0.21 2.27 ± 0.17 2.34 ± 0.21 NA 2.07 ± 0.45
44 T1#25-1⁄ 2.41 21.35 0.21 12.4 ± 2.4 12.9 ± 3.7 NA 18.4 ± 4.7
45 T1#26-1⁄ 1.45 21.01 0.23 14.3 ± 1.0 12.7 ± 3.3 NA 20.9 ± 1.7
46 T1#27-1⁄ 3.51 21.92 0.19 4.68 ± 0.32 3.17 ± 0.24 4.96 ± 0.35 4.80 ± 0.54
47 T1#27-2 1.24 21.92 0.19 9.2 ± 0.21 6.29 ± 0.32 9.84 ± 0.43 9.53 ± 0.22
48 T1#27-3 1.24 21.92 0.19 4.13 ± 0.18 2.81 ± 0.20 4.38 ± 0.20 4.23 ± 0.15
49 T1#28-1⁄ 1.64 20.95 0.23 31.0 ± 3.40 36.4 ± 4.90 NA 35.1 ± 3.60
Compacted T2 Perm granular fill material
50 T2#1-1⁄ 0.184 16.45 0.373 317 ± 10 641 ± 20 N/A 312 ± 15.00
51 T2#2-1⁄ 0.77 16.35 0.373 191 ± 23 526 ± 16 N/A 190 ± 24.0
52 T2#3-1⁄ 0.3 16.4 0.371 248 ± 26 487 ± 20 N/A 322 ± 25.0
53 T2#4-1⁄ 0.155 16.63 0.363 102 ± 15 576 ± 25 N/A 138 ± 30.0
54 T2#5-1⁄ 1.17 16.7 0.360 323 ± 33 275 ± 15 N/A 196 ± 12.0
55 T2#6-1⁄ 0.61 16.5 0.373 171 ± 11 449 ± 17 N/A 190 ± 33.0

Note: ‘‘⁄” symbol is for compacted. ‘‘⁄⁄” symbol is for uncompacted material.

References based on underground depressurization, tested in prototype housing built in a


high radon area in Spain, J. Environ. Radio. 102 (2011) 378–385.
[15] C.R. Boardman, S.V. Glass, Basement radon entry and stack driven moisture
[1] ICRP, Lung cancer risk from radon and progeny and statement on radon, ICRP
infiltration reduced by active soil depressurization, Build. Environ. 85 (2015)
Publication 115, Ann. ICRP 40 (1) (2010) 64
220–232.
[2] G.A. Gunning, D. Pollard, E.C. Finch, An outdoor radon survey and minimizing
[16] T.M.O. Diallo, B. Collignan, F. Allard, Air flow models for sub-slab
the uncertainties in low-level measurements using CR39 detectors, J Rad.
depressurization systems design, Build. Environ. 87 (2015) 327–341.
Prot. 34 (2014) 457–467.
[17] E.P.A., Us Handbook Sub-Slab Depressurization for Low- Permeability Fill
[3] WHO, http://www.who.int/mediacentre/factsheets/fs291/en/ (accessed
Material, US EPA, Washington, DC20460,USA, 1991.
04.09.17).
[18] DETR, Passive venting of soil gases beneath buildings, Research report 1 (1997)
[4] BMJ, Radon in homes and risk of lung cancer: collaborative analysis of
pp. 61.
individual data from 13 European case-control studies, (2004) 16, 10.1136/
[19] DETR, Passive venting of soil gases beneath buildings, Research report 2 (1997)
bmj.38308.477650.63.
pp. 56
[5] WHO, Handbook on Indoor radon, WHO Press, World Health Organisation,
[20] EHLG, Technical guidance document C: Site preparation and resistance to
Geneva, Switzerland, 2009. pp. 94.
moisture, Buil. Regula. (1997) pp. 23
[6] A.J. Gadgil, Y.C. Bonnefous, W.J. Fisk, R.J. Prill, A. Nematollahi, Influence of sub-
[21] NSAI, Code of Practice for the procurement and use of unbound granular fill
slab aggregate permeability on SSV performance, Lawrence Berkeley
hardcore material for use under concrete floors. I.S. 888: 2016, NSAI, 1 Swift
Laboratory, Report: LBL31160 (1991).
Square, Northwood, Santry, Dublin 9, Ireland (2016).
[7] Y.Y. Bonnefous, A.J. Gadgil, W.J. Fisk, R.J. Prill, A. Nematollahi, Field study and
[22] BS, BS 1377  Methods of Test for Soils for Civil engineering purposes. Part 2:
numerical simulation of sub  lab ventilation systems, Environ. Sci. Technol.
Classification tests (1990).
26 (1992) 1752–1759.
[23] BS, BS 1377  Methods of Test for Soils for Civil engineering purposes. Part 4:
[8] A.J. Gadgil, Y.C. Bonnefous, W.J. Fisk, Relative effectiveness of sub-slab
Compactionrelated tests (1990).
pressurization and depressurization systems for Indoor Radon Mitigation:
[24] ASTM, D6539Standard Test Method for Measurement of the Permeability of
studies with an experimentally verified numerical model, Indoor Air 4 (1994)
Unsaturated Porous Materials by Flowing Air, ASTM Inter. 100 Barr Harbor
265–275.
Drive, PO Box C700, West Conshohocken, PA, 194282959 USA (2013).
[9] M. Abdelouhab, B. Collignan, F. Allard, Experimental study on passive Soil
[25] H. Darcy, Les Fontaines Publiques de la Ville de Dijon, Dalmont, Paris, 1856.
Depressurization System to prevent soil gaseous pollutants into the building,
[26] S. Ergun, Fluid flow through packed columns, Chem. Eng. Pro. 48 (2) (1952)
Build. Environ. 45 (2010) 2400–2406.
89–94.
[10] A.L. Robinson, R.G. Sextro, Radon entry into buildings driven by atmospheric
[27] P. Forchheimer, Wasserbewegung durch Boden. Zeitschrift des Vereines
pressure fluctuations, Environ. Sci. Technol. 31 (1997) 1742–1748.
Deutscher. Ingenieuer, (45 ed.) (1901).
[11] BRE, Modeling and Measurement of Soil Gas flow, BRE report pp. 274, IHS BRE
[28] Ansys Inc, Ansys User’s Manual, Ansys Inc., Canonsburg, Pennsylvania, United
Press, The United Kingdom, 1998.
States, 2013.
[12] A.M. Odeh, W. Abu-El-Sha’r, R. Al-Ruzouq, Gas transport through concrete
[29] Ansys Inc, Introduction to Ansys Fluent, Ansys Inc, Canonsburg, Pennsylvania,
slabs, Build. Environ. 41 (2006) 492–500.
United States, 2014.
[13] M. Jiranek, Z. Svoboda, Numerical modeling as a tool for optimisation
[30] L.C. Hung, J. Goggins, M. Fuente, M. Foley, Investigation of sub-slab pressure
of sub-slab depressurisation systems design, Build. Environ. 42 (2007) 1994–
field extension in specified granular fill materials incorporating a sump-based
2003.
soil depressurisation system for radon mitigation, Sci. Total Environ. (2018),
[14] B.F. Vazquez, M.O. Adan, L.S.Q. Poncela, C.S. Fernandez, I.F. Merino,
https://doi.org/10.1016/j.scitotenv.2018.04.401.
Experimental study of the effectiveness of four radon mitigation solutions,