Construction and Building Materials 187 (2018) 803–818

Contents lists available at ScienceDirect

Construction and Building Materials

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

Experimental simulation study on pore clogging mechanism of porous

pavement
Jiong Zhang ⇑, Rui She, Zhaoxia Dai, Ruiping Ming, Guodong Ma, Xinzhuang Cui, Li Li
School of Civil Engineering, Shandong University, Jingshi Road, Jinan 250061, China

h i g h l i g h t s

Visualized method is developed to study the mechanism of pervious pavement clogging.

Clogging process especially the internal movement of sediments is studied.

The effects of porosity, horizontal runoff, seepage and sediment size are studied.

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

Article history: Permeability reduction caused by clogging is a major issue with porous concrete pavements. Particulate
Received 21 March 2018 matter infiltration owing to rainwater, movement, and accumulation/deposition in the pores is the main
Received in revised form 12 July 2018 cause of clogging in porous pavements. In this study, in order to observe the internal movement of par-
Accepted 25 July 2018
ticles in pores and the clogging evolution directly, we focus on the accumulation of transparent sodium
polyacrylate beads instead of pervious concrete aggregate. The clogging process caused by particles in
surface runoff entering the internal pores of porous concrete was studied, and a visualization method
Keywords:
was applied to analyze the clogging process in porous media. The effects of the porous concrete porosity,
Clogging development process
Porosity
horizontal runoff velocity, seepage velocity, and particle size distribution on the clogging development
Horizontal runoff velocity with time, clogging development rate, and particle distribution in the clogging state were studied using
Seepage velocity the control variable method. The results demonstrate that the clogging process exhibits four stages,
Particle size distribution namely rapid clogging, partial recovery, slow clogging, and clogging stability. The effects of porosity
and clogging particle size on the clogging development rate are significant. Porous concrete porosity,
seepage velocity, and clogging particle size all have impacts on the distributions of clogging particle in
the final clogging process, while the horizontal runoff effect can be neglected.
Ó 2018 Elsevier Ltd. All rights reserved.

1. Introduction velocity and pressure gradient can be fitted by the nonlinear

Portland cement pervious concrete contains a large number of
interconnected pores within the aggregate skeleton. Generally,
the pervious concrete porosity is between 15% and 35%, and the
initial permeability is between 2 and 6 mm/s, but can be as high
as 10 mm/s [7,13]. The internal structure properties and hydraulics
of concrete can be analyzed using X-ray CT images [14,15]. Zhang
et al. [17] studied the pore characteristics and seepage flow in
pervious concrete using computed tomography (CT) imaging and
numerical simulation. The relationship between the seepage
⇑ Corresponding author.
E-mail address: jiongzhang@sdu.edu.cn (J. Zhang).
Forchheimer formula.
However, solid matter suspended in storm water runoff can
enter the pervious surface, leading to a gradual reduction in
permeability [16], which decreases the service life of pervious
concrete pavements [12]. Porous concrete is susceptible to clog-
ging within the first three years [11]. The permeability reduction
of porous systems may be caused by two types of siltation pro-
cesses: a slow siltation process with continuing or cyclic deposi-
tion of small sediment quantities, and a rapid siltation process
triggered by a sudden slump or landslide [10]. Studies have been
conducted on the slow siltation process; that is, clogging owing
to long-term particle deposition or clogging phenomena based on
cycle tests [9,5,8,4,6]. Kayhanian et al. [6] assessed the clogging
of pervious concrete pavements in parking lots using the falling

https://doi.org/10.1016/j.conbuildmat.2018.07.199
0950-0618/Ó 2018 Elsevier Ltd. All rights reserved.
804 J. Zhang et al. / Construction and Building Materials 187 (2018) 803–818

head method, with a National Center for Asphalt Technology
(NCAT) device in the field. In order to evaluate the contribution
of pores towards clogging susceptibility, Deo et al. [4] developed
a falling head permeability cell to measure the clogging cycling
simulation. Cui et al. [3] designed a continuous permeability mea-
surement system to record the rapid clogging process of a pervious
concrete pavement caused by storm water runoff. The tests indi-
cated that the pore clogging process generally includes three
phases: quick clogging, temporary clogging mitigation, and pro-
gressive clogging.
At present, although certain researches exist on clogging caused
by the infiltration of particulate matter with rainwater, the meso-
Therefore, in this study, a comprehensive analysis of the meso-
mechanism of internal clogging is carried out through visualization
of the porous concrete test method.
2. Experimental materials and processes
2.1. Test system design
The following experimental system was developed in order to observe the clog-
ging process, which cannot be observed in the pores of the opaque permeable pave-
ment structure layer. The main equipment includes a flume, circulating water tank,
water pump, constant water supply tank, and HD camera, among others, as illus-
trated in Fig. 1.

scopic mechanism of clogging inside porous concrete has not yet

been intuitively established. In particular, there has been almost
no direct observation of particle movements inside the pores and
clogging evolution. Direct observation of the clogging process is
crucial for explaining clogging causes and studying its prevention.
2.1.1. Permeable pavement simulation device design
The main purpose of this study is to observe and record the clogging process of
sediment particles in the permeable pavement structure layer pores directly, while
the actual permeable pavement structure layer is not transparent. In order to study
the internal pore clogging process, transparent sodium polyacrylate beads with a

Constant Permeable pavement

head water
Base course with
tank
open-graded aggregate
Overflow

HD camera Valve
Filter
Sub-base with
open-graded
aggregate

Computer
Circulating Circulating
water tank pump
(a) Schematic of test device (b) photo of test device

Fig. 1. Test device.

(a) Without water (b) Within water

Fig. 2. Porous concrete pavement experimental simulation with sodium polyacrylate beads.
 J. Zhang et al. / Construction and Building Materials 187 (2018) 803–818 805

Table 1
The physical indexes of aggregates used in pervious concrete mixtures.
Particle size Apparent density Stacking density Porosity
(mm) (g/m3) (g/m3) (%)
4.75–9.5 2.665 1.665 37.89
refractive index close to water were filled in a transparent glass flume to form a
water-permeable pavement structure model with different porosity levels; that
is, an artificial porous medium was used for experimental simulation. This superab-
sorbent polymer has the ability to absorb up to 200–300 times its mass in water;
therefore, its refractive index is very close to that of water (Fig. 2), and the material
offers the advantage of reusability.
In most practical project, pervious concrete aggregate size range is 4.75–9.5
mm. The physical indexes of aggregates used in pervious concrete mixtures are
listed in Table 1.
The gradation of the aggregate has an important influence on porosity of pervi-
ous concrete. But there is no sieve has the pore size between 4.75 and 9.5 mm, so
particle size distribution of aggregate cannot be applied by sieve analysis. In this
research, an image analysis technique using ImageJ is proposed to evaluate particle
size distribution of gravels. Fig. 3(a) shows the photo of aggregates. Fig. 3(b) shows
the photo of aggregates processed by binaryzation in imageJ software. And the mea-
sured aggregate size is in the range of 4–12 mm shown in Fig. 3(c).
Three kinds of transparent sodium polyacrylate beads with diameter of 4 mm,
8 mm, and 12 mm are mixed to replace aggregates in real porous concrete based
on the gradation curve. Then a slight compaction by a plate is performed to gener-
ate three different the porosities. In order to match actual construction cases, the
thickness of the porous concrete pavement uppermost surface with 4–12 mm beads
used in the test was selected as 100 mm, the base course with 12 mm beads was 75
mm, and the sub-base with 14 mm beads was 75 mm. Metal mesh was placed
between each layer in order to separate them, and was placed on the side wall to
prevent sediment particles from entering the drain. In order to fix the porous
medium layer particles so as not to be driven by the water flow, a metal panel
Crush was placed on the porous medium and holes array with a sufficient pore size to
value allow sand particles to enter the porous medium was arranged in the middle of
8.6 the metal panel (see Fig. 4).
The end of the channel was free-flowing and flowed into a container with a filter
to collect grit and inject fresh water into the recirculation tank. Owing to sudden
changes in the cross-section, a small device for reducing turbulence and bubbles
was installed at the channel entrance. The device consisted of a straw arrangement
set. A valve was provided at the channel to regulate the influent water flow rate, as
well as at the channel outlet to adjust the seepage velocity, as illustrated in Fig. 4.
2.1.2. Water supply
The water flow in the channel was the gravity flow provided by a constant head
tank. The device ensured a constant inlet pressure and inflow to the funnel. The
design values were obtained through the control valve at the funnel inlet. Once
the water in the channel flowed out into the circulating water tank, the water in
the tank could be transported to the constant water tank again by the water pump,
so that the water in the device formed a circulation loop. Under gravity, the flow
rate in the water supply system could reach 7 L/min, corresponding to an average
maximum velocity of 0.46 m/s. This flow rate is basically the same as the average
surface runoff velocity during rainfall.
2.1.3. Flow rate measurement method
In this study, a stopwatch and measuring cylinder were used to measure the flow
in the channel. During the measurement, the flow rate before and after the test was
measured for the channel inlet and the outlet. By measuring the water volume
received over a given period, the volumetric flow Q per unit time was calculated:
Q = V/t.

(a) (b)

100
experiment data

80
Percent passing %

60

40

20

0
4 5 6 7 8 9 10 11 12
Aggregate size mm

(c)
Fig. 3. Aggregates size gradation analysis.
806 J. Zhang et al. / Construction and Building Materials 187 (2018) 803–818

Metal panel Overflow Table 2

with hole Blocked particle grading table.

Permeable pavement Sand size Sand Fine sand Coarse sand Well-graded sand
Base course with open-graded aggregate (mm) color (%) (%) (%)
Metal mesh
Sub-base with open-graded aggregate 2.36–4.75 Pink 0 44.06 30.83
1.18–2.36 Yellow 0 28.20 19.74
Gauze 0.6–1.18 Green 0 27.74 19.42
0.3–0.6 Black 42.70 0 12.81
Fig. 4. Porous media structure distribution diagram. 0.15–0.3 Red 27.31 0 8.20
<0.15 Blue 29.99 0 9.00

Table 3
Permeable pavement pore clogging test program.

Cases Porous media Horizontal runoff Seepage flow Sand

porosity (%) velocity (m/s) rate (mm/s) grading cases
1 25 0.21 1.3 Coarse and fine
2 30 0.21 1.8 Coarse and fine
3 35 0.21 2.2 Coarse and fine
4 30 0.35 2.4 Coarse and fine
5 30 0.48 3.2 Coarse and fine
6 30 0.21 2.6 Coarse and fine
7 30 0.21 3.7 Coarse and fine
8 30 0.21 4.8 Coarse and fine
Fig. 5. Particle addition device schematic. 9 30 0.21 1.8 Fine
10 30 0.21 1.8 Coarse
2.1.4. Particle addition equipment
In order to add sediment particles into the flow in the same position, with the
initial energy as equal and minimal as possible, the device illustrated in Fig. 5 was 2.3.2. Test cases
used. In this study, the effects of the horizontal runoff velocity, seepage velocity, per-
meable pavement structure porosity, and sand grading were studied. Table 3 dis-
plays the different test cases.
2.1.5. Particle velocity measurement
In the experiment, a Nikon D7100 camera was used to capture the particle
motion, which provides a resolution of 6000  4000 pixels and a camera burst 3. Experimental results and analysis
speed of 2 fps. The trigger time is shown in Fig. 6. The actual shooting area of the
camera was 400 mm * 267 mm. The smallest particle fraction was less than 0.15
mm, the device length was 400 mm, the resolution in the length direction was
3.1. Study on clogging development in pervious concrete
6000 pix, and the size of one pixel was 0.067 mm. Therefore, the camera resolution
satisfied the requirement of recording the smallest particle fraction. High-definition The research on the pore clogging development process
images captured during the testing process were processed using ImageJ software. includes three aspects. Firstly, the clogging process of clogging par-
ticles in a certain position of the pervious concrete is studied. A ser-
2.2. Clogging particles
ies of images about the detailed areas when clogging process occur
In this study, sieved sand of different colors were used as a clogging material in dramatically are analyzed. Thereafter, clogging process classified
order to carry out simulations. Table 2 provides the blocked particle grading table. by the participation of large particles (ds > 1.18 mm) in the same
The different sand colors for the screening method are divided into fine, coarse, and depth of porous concrete is studied. Finally, the entire movement
well-graded sand. The fine sand particle size range is 0–0.6 mm, the coarse sand of clogging particles in the porous concrete is compared.
particle size range is 0.6–4.75 mm, and the well-graded sand size range is 0–4.75
mm.
3.1.1. Clogging particles at location in porous concrete clogging
2.3. Test process and parameter selection development and evolution
The entire pore clogging development process was obtained by
2.3.1. Test process means of analyzing the images captured during the experiment.
(1) Install and connect the test device, as shown in Fig. 1. The permeable pave-
Taking case 2 as an example, a certain area at the most serious
ment simulation material is layered into the device from bottom to top clogging location was intercepted, and a series of images were ana-
according to the experimental design, and the material is compacted to lyzed during the period with the most obvious clogging change.
the target porosity by a perforated metal panel. Turn on the pump, adjust Fig. 7 illustrates the clogging process at a certain location of the
the pump flow, and fill the tank with water to maintain a constant water
porous concrete. It shows that there are almost no clogging parti-
level in the head tank.
(2) Adjust the inlet flow to obtain a specific flow rate level, and adjust the cles exist in the pores at t = 3.165 s, and the obvious change in the
bottom-right outlet valve to vary the seepage flow rate in the porous media. particle clogging process mainly occurs from t = 3.165 s to 6.165 s.
Measure the water flow at this time and record the valve position. When the clogging tends to be stable, the clogged area no longer
(3) Turn on the external lighting source, start adding jammed material, and changes significantly. During the clogging process, a large particle
simultaneously initiate automatic shooting of the camera until the blocked
particles reach the equilibrium state and no longer migrate. Measure the
colored yellow (ds = 1.18–2.36 mm) firstly deposited in the porous
water flow at this time and record. media pores, which plays a bridging role. Several medium-sized
(4) In order to change the parameters, repeat steps 1–3. particles such as the green and black (ds = 0.6–1.18 mm;

0 0.165 1 1.165 2 2.165 Trigger time t/s

Fig. 6. Camera trigger time diagram.

 J. Zhang et al. / Construction and Building Materials 187 (2018) 803–818 807

5 mm

t=3.165 s t=4 s t=4.165 s t=5 s t=5.165 s t=6 s t=7.165 s

Fig. 7. Clogging particles at a location in the porous concrete clogging development process.

5 mm

t=3.165 s t=4 s t=4.165 s t=5 s t=5.165 s t=6 s t=7.165 s t=8 s

t=9.165 s t=10 s t=10.165 s t=11 s t=11.165 s t=12 s t=13.165 s t=14 s

Fig. 8. Clogging process at a certain depth of porous concrete (no large particles involved).

ds = 0.3–0.6 mm) then rapidly deposited around them and the 50

small-sized clogging particles began to fill the pores. 45 No large particles involved
Proportion of clogging particle area/%

Large particles involved

40
3.1.2. Effect of large particle size on pore clogging process
35
Comparison and analysis of the clogging evolution process with
or without large particle (ds > 1.18 mm) in pores were carried out 30
in the same working case. In the photos captured under the same 25
experimental cases, two clogged areas at the same depth were
20
selected, one is shown in Fig. 6 in which large-sized particles (yel-
low, ds = 1.18–2.36 mm) involves in the clogging process. The other 15
group shown in Fig. 7 does not include large-sized particles. 10
By means of comparison, it concluded that during the migration
5
and deposition of clogging particles in the porous concrete pores,
parts of the large particles hardly move again once deposited in a 0
certain position of the porous concrete owing to their large size. -5
They evolve into new framework particles leading to the narrow 0 10 20 30 40 50
of the pore channels, so that smaller particles rapidly accumulate time/s
around them (Fig. 7). However, the clogging scenario caused by Fig. 9. Proportion of clogging particle area.
small particles is not sufficiently stable and exhibits a high mobil-
ity owing to its smaller particle size. Particles previously deposited
in the pores may restart and move downwards owing to the seep-
involving large particles is relatively short, because when the large
age force, resulting in mitigation of the clogging of this area (Fig. 8).
particles are trapped in the pores, the pores are almost filled and
The development processes of clogging with or without large
the clogging area approaches the peak. For large-size particles,
particles as illustrated in Figs. 7 and 8, were compared, and the
the clogging process no longer changing significantly requires only
results are shown in Fig. 9.
approximately 3–4 s. For small particles, the stabilization time can
It can be observed from Fig. 9 that whether large particles
reach approximately 10 s, which is a great longer than that of large
involved in the pore clogging, the trend of the proportion of clog-
particles.
ging particle area is basically identical. The clogging area increased
substantially over time, and eventually reached a relatively stable
level, which indicates that clogging areas accumulates rapidly in a 3.1.3. Analysis of pore clogging at different layers in porous concrete
short period and finally attains stabilization. The porous concrete surface layer was divided into four layers:
However, the clogging area containing large particles was basi- 0–10 mm, 10–30 mm, 30–60 mm, and 60–100 mm. The relation-
cally stable after adding particles for 10 s, while the clogging area ship between proportion of clogging particles area and time in dif-
without large particles was not steady until the clogging particles ferent layers was analyzed as shown in Figs. 10–13 clogging
were placed for 20 s. This means that the pore clogging period particle
808 J. Zhang et al. / Construction and Building Materials 187 (2018) 803–818

7 7

Proportion of clogging particle area /%

Proportion of clogging particle area /%
0-10 mm 0-10 mm
10-30 mm 10-30 mm
6 30-60 mm 6 30-60 mm
60-100 mm 60-100 mm
5 5

4 4

3 3

2 2

1 1

0 0
0 10 20 30 40 50 0 10 20 30 40 50
Time /s Time /s
Case 1 Case 2

7
Proportion of clogging particle area /%

0-10 mm
10-30 mm
6 30-60 mm
60-100 mm
5

0
0 10 20 30 40 50
Time /s
Case 3
Fig. 10. Clogging development process at different layers with different porosity.

(1) Impact of porosity that of the second layer. As the depth increases, the clogging parti-
cle area proportion gradually decreases, and the area in the 60–
The effects of different porosities (cases 1, 2, and 3) on the clog- 100 mm layer is almost zero. It can be observed that as the depth
ging of porous concrete at different layers are illustrated in Fig. 10. increases, the time required for the clogging particle area attains to
By means of comparison, it concludes that as for the porous con- peak increases. Therefore, with the increase of horizontal runoff
crete of 25% porosity, the area of clogging particles is the largest velocity, the peak of the clogging process area postponed.
in the layer of 0–10 mm, whose value is twice of that in the second
layer. As the increase of depth, the clogging particle area propor- (3) Impact of seepage velocity
tion gradually decreases, thus the proportion of the 60–100 mm
The effects of different seepage velocity (cases 6, 7, and 8) on
layer is smallest. That is to say, the clogging degree in 0–10 mm
the clogging of porous concrete at different layers are illustrated
layer is serious, while the clogging degree in lower layers is less
in Fig. 12. When the seepage velocity is vs = 2.6 mm/s (Case 6),
serious, and the bottom exhibits almost no clogging. It is also found
the clogging particle area proportion in the uppermost layer
that the clogging reaches the peak in the uppermost layer for the
rapidly decreases (t = 6 s). The reduced clogging particles move to
shortest time, and takes longer time to reach the peak in lower lay-
the lower layer (10–30 mm), and it has the highest clogging parti-
ers. With increased porosity, the number of particles decreases in
cle area percentage. However, when the depth exceeds 30 mm, the
the uppermost layer and increases gradually in lower layers. As
clogging particle area proportion gradually decreases with the
the upper particles continue to move downward, they continue
increasing depth, and almost no clogging occurs in the 60–100
to accumulate and the areas in lower layers reach the peak. When
mm layer.
the pores in lower layers are mostly filled, the area in upper layer
When the seepage velocity vs = 4.8 mm/s (Case 8), the upper-
then attains to its peak.
most clogging particle area proportion firstly increases and then
decreases rapidly to the value of approximately 0.5%. Because as
(2) Impact of horizontal runoff velocity
the increase in seepage velocity the upper clogging particles will
move downwards, resulting in increased clogging particles in the
The effects of different horizontal runoff velocity (cases 2, 4, and
lower layers (10–30 mm and 30–60 mm). The second layer (10–
5) on the clogging of porous concrete at different layers are illus-
30 mm) exhibits the largest clogging particle area proportion, the
trated in Fig. 11. When the horizontal runoff velocity v = 0.35 m/s
value is at approximately 10 times that of the top layer. In the third
(Case 4), the clogging particles area in the uppermost layer (0–
layer (30–60 mm), the value of the proportion of the clogging par-
10 mm) occupies the largest proportion, which is three times than
ticle area is only lower than that of the second layer, at approxi-
 J. Zhang et al. / Construction and Building Materials 187 (2018) 803–818 809

7 7
0-10 mm

Proportion of clogging particle area /%

Proportion of clogging particle area /%
0-10 mm
10-30 mm 10-30 mm
6 30-60 mm 6 30-60 mm
60-100 mm 60-100 mm
5 5

4 4

3 3

2 2

1 1

0 0

0 10 20 30 40 50 0 10 20 30 40 50
Time /s Time /s
Case 2 Case 4

7
Proportion of clogging particle area /%

0-10 mm
6 10-30 mm
30-60 mm
60-100 mm
5

0
0 10 20 30 40 50
Time /s
Case 5
Fig. 11. Clogging development process at different layers with different horizontal runoff velocity.

mately eight times than that of the top. Furthermore, with the and small size of the clogging particles can pass through to lower
increase in seepage velocity, clogging particles were captured at layers. The third layer (30–60 mm) exhibits the largest clogging par-
the bottom (60–100 mm) after approximately 2 s and up to 1.2% ticle area proportion, the value is at approximately twice than that of
when time is 12 s. Moreover, the area proportion of the clogging the top layer. The clogging particle area proportion of the second
particles in the bottom layer is slightly higher than that in the layer (10–30 mm) is only lower than the third layer, at approxi-
uppermost layer. With an increase in seepage velocity, the parti- mately 1.7 times than that of the top layer. In the bottom layer
cles in the surface gradually decrease and the particles in the deep (60–100 mm), the occupied clogging particle area increases with
regions gradually increase. time, while the bottom layer clogging particle area is 1.6 times than
that of the uppermost layer clogging particles. It can conclude the
(4) Impact of clogging particle size distribution time that clogging particles attains to their peak value is longer
when use coarse sand rather than fine sand. This is because fine sand
The effects of particles with different gradations (cases 9 and 10)
can pass through pores easily, while coarse sand is more likely to
on the clogging of porous concrete at different depths are illustrated
become stuck in the surface layer pores and moves difficultly.
in Fig. 13. When all the clogging particles were coarse (ds = 0.6–4.75
mm, Case 9), the clogging particle area in the uppermost layer (0–10
mm) accounts for the highest proportion. The most serious conges- 3.1.4. Curve fitting of clogging process at different layers in porous
tion occurs in the first layer (0–10 mm) and second layer (10–30 concrete
mm). Moreover, the third layer (30–60 mm) and bottom layer The time and proportion of the clogging particle area were nor-
(60–100 mm) almost don’t exist clogging particles. It indicates that malized, thus the peak values of clogging area for each layer and
the clogging in the uppermost layer slowly decreases after reaching time were selected as 1. Taking the layer of p = 35% in the depth
the peak value, and the clogging particle area proportion in the lower of 3–6 mm as an example, the normalized clogging process was
layer gradually increases. When all of the clogging particles are fine obtained as illustrated in Fig. 14.
sand (ds < 0.6 mm, Case 10), the clogging particle area proportion in Through observation of the clogging tendency, it was found that
the uppermost layer increases then decreases rapidly. Furthermore, the clogging area first increases rapidly, then decreases slightly,
it’s proportion of clogging particle area accounts for the lowest then increases for a while, and finally stabilizes. Therefore, the
value, but proportion of clogging particle area in the lower layers clogging process is divided into the following four stages,
(10–30 mm, 30–60 mm, and 60–100 mm) increase. That is because according to the clogging area development: rapid clogging
no large-sized particles involved in clogging process in upper layer, section, clogging recovery section, slow clogging section, and
810 J. Zhang et al. / Construction and Building Materials 187 (2018) 803–818

7 18
Proportion of clogging particle area/% 0-10 mm

Proportion of clogging particle area/%

0-10 mm
10-30 mm 16 10-30 mm
6 30-60 mm 30-60 mm
60-100 mm 14 60-100 mm
5
12
4 10
3 8

2 6
4
1
2
0
0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Time /s Time /s
Case 6 Case 7

16
Proportion of clogging particle area/%

0-10 mm
14 10-30 mm
30-60 mm
12 60-100 mm

10

0
0 10 20 30 40 50 60
Time /s
Case 8
Fig. 12. Clogging development process at different layers with different seepage velocity.

12 6
0-10 mm
Proportion of clogging particle area/%
Proportion of clogging particle area/%

11 10-30 mm 0-10 mm
30-60 mm 5 10-30 mm
10 60-100 mm 30-60 mm
9 60-100 mm
8 4
7
6 3
5
4 2
3
1
2
1
0
0
0 10 20 30 40 50 60 0 10 20 30 40 50
Time /s Time /s
Case 9 Case 10
Fig. 13. Clogging development process at different layers with different clogging particle size distribution.

stable clogging section. This result is consistent with the conclu- The rapid clogging section function is y ¼ 1 

1 p ;
1 þ t
sion provided by Cui et al. [3]. The first phase is fitted by a logistic x0

function, the second by a linear function, and the fourth by a hor- The clogging recovery section function is y = ct + d;
izontal line. Thereafter, each experimental porous concrete layer The slow clogging section function is y = et + f; and
group was divided into four layers and analyzed: 0–10 mm, 10– The stable clogging section function is y = h.
30 mm, 30–60 mm, and 60–100 mm, respectively. Among these, x0, p, c, d, e, f, and h are parameters.
 J. Zhang et al. / Construction and Building Materials 187 (2018) 803–818 811

1.2 3.2. Study on clogging development velocity in pervious concrete

Under the action of seepage, the clogging speed of porous con-

Normalized blocked area proportion

1.0
crete can be analyzed by means of the following method. First, the
entire porous concrete test section can be divided into several lay-
0.8 ers every 2 mm along the depth direction. According to the analy-
sis results of Section 3.1, the images of the two adjacent moments
0.6 when the clogging area exhibits the fastest growth rate is selected.
As illustrated in Fig. 15, it can be observed that some of the parti-
cles clogged in the upper layer will migrate to lower layer pores
0.4 under the seepage action. Therefore, it can be considered that the
added particles in the divided porous concrete layer originate from
0.2 the upper layer. The particle movement speed can be calculated
according to the position changes of the particles during the period
of t2-t1 clogging particle.
0
0 0.5 1.0 1.5 2.0 2.5 3.0 In the case of inconsistencies in the numbers of particles
Normalized time decreasing from the upper layer and being added to the adjacent
lower layer, the moving clogging particles need to be closely
Fig. 14. Normalized clogging process. matched to the adjacent positions, according to the nearest match-
ing method. The excess quantity is dispensed in turn downwards,
until all matches are completed. Based on this process, it is possible
The coefficients obtained by the fitting function under different to obtain the displacement of particles and ratio of moved particles
cases are listed in Table 4. to the total amount clogging particle. Then, the particle moving
As can be seen from the above table, the x0 range is 0–1. speed is calculated according to a weighted average algorithm, as
indicated in Eq. (1):

Table 4
Fitting curve coefficient table.

Case Depth (cm) Rapid clogging Clogging recovery Slow clogging section Stable clogging section
section section
y = 1–1/(1 + (t/x0)p) y = ct + d y = et + f y=h
x0 p c d e f h
1 0–1 0.40 4.19 0.53 1.53 0.29 0.60 0.97
1–3 0.29 4.03 2.17 3.17 0.87 0.02 0.98
3–6 0.33 4.51 2.05 3.05 0.50 0.41 0.95
6–10 0.37 3.76 2.56 3.36 1.07 0.20 0.89
2 0–1 0.26 7.31 1.44 2.44 1.28 0.23 0.94
1–3 0.22 2.08 4.21 5.21 0.37 0.41 0.93
3–6 0.30 1.85 6.13 7.13 0.54 0.22 0.89
6–10 0.36 1.85 7.16 8.16 0.59 0.13 0.87
3 0–1 0.85 7.57 0.26 1.25 0.20 0.69 0.91
1–3 0.90 37.34 0.30 1.29 0.22 0.65 0.97
3–6 0.61 4.30 0.51 1.51 0.40 0.48 1.02
6–10 0.39 4.50 0.69 1.69 1.19 0.32 1.02
4 0–1 0.52 4.06 1.77 2.77 1.49 1.30 0.68
1–3 0.48 3.41 1.44 2.44 0.85 0.17 0.89
3–6 0.44 3.39 1.60 2.58 0.42 0.28 0.71
6–10 0.57 3.96 3.65 4.65 0.93 0.14 0.73
5 0–1 0.19 1.07 1.11 2.00 0.42 0.48 0.22
1–3 0.26 0.99 0.56 1.55 0.20 0.30 0.63
3–6 0.45 0.93 0.40 1.40 0.31 0.44 0.78
6 0–1 0.86 38.05 1.92 2.88 0.92 0.40 0.69
1–3 0.86 42.54 0.32 1.32 0.46 0.02 0.79
3–6 0.77 13.47 0.82 1.82 0.65 0.13 0.81
6–10 0.61 6.12 1.53 2.55 1.74 1.18 0.53
7 0–1 0.70 9.43 1.77 2.77 1.49 1.30 0.73
1–3 0.37 2.19 0.47 1.46 3.03 2.62 0.80
3–6 0.50 2.50 0.58 1.58 0.32 0.63 0.82
6–10 0.54 2.80 1.92 2.92 8.73 8.26 0.84
8 0–1 0.48 3.07 0.46 1.46 1.12 0.51 0.41
1–3 0.74 4.50 0.35 1.26 0.20 0.30 0.64
3–6 0.60 6.91 0.40 1.40 0.31 0.44 0.79
6–10 0.59 5.64 1.92 2.88 0.45 0.16 0.68
9 0–1 0.56 43.34 2.41 3.41 0.48 0.25 0.72
1–3 0.83 32.43 0.31 1.31 1.81 2.12 1.69
3–6 0.83 40.83 2.44 3.44 0.26 0.20 0.44
6–10 0.80 8.89 2.44 3.44 0.26 0.20 0.54
10 0–1 0.55 6.68 0.08 1.08 0.15 0.76 0.92
1–3 0.37 3.81 3.30 4.30 0.77 0.12 0.29
812 J. Zhang et al. / Construction and Building Materials 187 (2018) 803–818

Clogging
particles Porous media
particles

Depth direction

Fig. 15. Schematic of clogging particles at different times in porous concrete migration.

Table 5
Particle movement matching table for particle sizes ranging from ds = 0.6–1.18 mm.
 J. Zhang et al. / Construction and Building Materials 187 (2018) 803–818 813

200 200

180 Case 1 180 Case (2)

Case 2 Case (4)
Particle transport speed (mm/s)

Particle transport speed (mm/s)

160 Case 3 160 Case (5)

140 140

120 120

100 100

80 80

60 60

40 40

20 20

0 0
<0.15 0.15-0.3 0.3-0.6 0.6-1.18 1.18-2.36 2.36-4.75 <0.15 0.15-0.3 0.3-0.6 0.6-1.18 1.18-2.36 2.36-4.75
Particle size (mm) Particle size (mm)

Fig. 16. Particle clogging development rate curve with change in particle size in Fig. 17. Particle clogging development rate curve with change in particle size in
different porosity. different horizontal velocity.

Pn
i¼1 Dyi f i 200
v¼ P ð1Þ
ðt 2  t 1 Þ ni¼1 f i 180 case (2)
Particle transport speed (mm/s) case (6)
where v is the maximum speed at which particle clogging develops 160 case (7)
at a certain size, Dyi is the particle displacement, fi is the ratio of the 140
case (9)
particle area moved by Dyi to the total particle area, t1 is the initial
time, and t2 is the end time. 120
For example, when the porosity p = 25%, Table 5 displays the 100
results of the particle movement matching with particle size ds =
80
0.6–1.18 mm (green). The blue values in the table indicate the
reduction in a layer in the porous concrete; that is, the proportion 60
of particles exiting the porous media layer. The red values in the
40
table indicate the addition proportion of particles corresponding
to a layer. 20
It can be seen that, from t1 to t2, 6% of the clogging particles
0
move from 1 mm to 11 mm, 3% of the particles move from 6 mm <0.15 0.15-0.3 0.3-0.6 0.6-1.18 1.18-2.36 2.36-4.75
to 25 mm, 2% of the particles move from 9 mm to 27 mm, and 4% Particle size (mm£©
of the particles move from 21 mm to 35 mm. The time interval
between t1 and t2 is 0.165 s. As a result, the fastest particle (ds = Fig. 18. Particle clogging development rate curve with change in particle size in
0.6–1.18 mm, green) clogging development rate is v = 84.44 mm/s. different seepage velocity.

The velocities of particles with different diameters are calcu-

lated according to the above method. The influence of different fac-
tors on the clogging development rate was analyzed. Clogging
particle Clogging particle
(1) Effect of porosity on particle clogging development rate
It can be observed in Fig. 16 that large particles with a size of
2.36–4.75 mm exhibit the lowest migration velocity in the porous
media. Medium particles with sizes in the range of 0.6–2.36 mm
migrate the most rapidly, while particle sizes of ds < 0.15 mm,
0.15–0.3 mm, and 0.3–0.6 mm exhibit migration speeds between
the maximum and minimum values. Furthermore, it can be seen
from the figure the clogging growth rate increases with increasing
porosity for the most time.
(2) Effect of horizontal velocity on particle clogging develop-
ment rate
It can be observed in Fig. 17 that clogging particle of sizes of
2.36–4.75 mm and <0.15 mm have the minimum migration veloc-
ity in porous media. Particles with size ranging from 0.6 to 1.18
mm exhibit the greatest transport velocity, while the transport
speeds for small-sized particles in the ranges of 1.18–2.36 mm,
0.3–0.6 mm, and 0.15–0.3 mm are between the maximum and
minimum speed values. A secondary speed peak appeared when
particle size is 0.3–0.6 mm. It can also be seen from the figure that
the horizontal flow rate has no obvious effect on the clogging
development speed.
(3) Effect of seepage velocity on particle clogging development
rate
As can be seen from Fig. 18, small particles with sizes of 0.15–
0.3 mm and 0.3–0.6 mm migrate relatively quickly, and the migra-
tion speed of small particles increases with increasing seepage
velocity. Large particles with diameters larger than 1.18 mm are
relatively close to each other, and the influence of seepage velocity
on these is not obvious.
(4) Effect of clogging particle size distribution on particle
clogging development rate
It can be seen that when only coarse sand is used as clogging
particle, the clogging development rate is close to that of
well-graded sand. However, when only fine sand is used as
814 J. Zhang et al. / Construction and Building Materials 187 (2018) 803–818

200 mm) exhibit the lowest migration speed. This is owing to the fact
case (2) that large-sized particles enter the porous medium and are easily
180
case(9) trapped in the surface layer pores, and subsequently difficult to
Particle transport speed (mm/s)

160 case(10) transport. The medium-sized particles (0.6–2.36 mm) will not be
stuck on the surface and move to the deep quickly, under the
140
action of gravity and seepage.
120 It should also be noted that, although the analysis results
indicate that particles with the smallest size migrate at a small
100
speed, these small particles are also easily to become suspended
80 substances and affected by gravity, buoyancy, drag force, and so
on. Therefore, the distribution of smallest particles is uniform in
60
all layers.
40 The analysis method proposed in this study is used to match the
input and output locations manually; thus, the particle clogging
20
development rate of this component may not be accurate.
0
<0.15 0.15-0.3 0.3-0.6 0.6-1.18 1.18-2.36 2.36-4.75
3.3. Study on stabilized sediment distribution in clogged pervious
Particle size (mm)
concrete
Fig. 19. Particle clogging development rate curve with change in particle size in
clogging particle size distribution. The test clogging images of the final stability state are selected
for analysis. Each different color corresponding to different particle
size coordinates is extracted, and each 2 mm is divided into one
clogging particles, the particle migration speed is also significantly layer in the depth direction for statistical analysis. The ratio of
faster than the well-grade sand (see Fig. 19). the clogging particles number in each layer to the total number
Based on analysis of the above experimental results, a universal of clogging particles is obtained. The proportion of particles of var-
law can be established. Particles with the largest size (2.36–4.75 ious sizes in the same layer is stacked, as illustrated in Figs. 20–23.

100 100
90 ds<0.15 mm 90 ds<0.15 mm
ds=0.15-0.3 mm ds=0.15-0.3 mm
80 80
Clogging particle area (%)

Clogging particle area (%)

ds=0.3-0.6 mm ds=0.3-0.6 mm
ds=0.6-1.18 mm ds=0.6-1.18 mm
70 ds=1.18-2.36 mm 70
ds=1.18-2.36 mm
60 ds=2.36-4.75 mm 60 ds=2.36-4.75 mm

50 50
40 40
30 30
20 20
10 10
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Porous dielectric layer thickness (mm) Porous dielectric layer thickness (mm)
Case 1 Case 2

100
90 ds<0.15 mm
ds=0.15-0.3 mm
80
Clogging particle area (%)

ds=0.3-0.6 mm
ds=0.6-1.18 mm
70
ds=1.18-2.36 mm
60 ds=2.36-4.75 mm

50
40
30
20
10
0
0 20 40 60 80 100
Porous dielectric layer thickness (mm)
Case 3
Fig. 20. Effects of porosity on distribution frequency of clogging particles in porous concrete.
 J. Zhang et al. / Construction and Building Materials 187 (2018) 803–818 815

100 100
ds<0.15 mm
90 90 ds<0.15 mm
ds=0.15-0.3 mm ds=0.15-0.3 mm
Clogging particle area (%)

Clogging particle area (%)

80 ds=0.3-0.6 mm 80 ds=0.3-0.6 mm
ds=0.6-1.18 mm ds=0.6-1.18 mm
70 70
ds=1.18-2.36 mm ds=1.18-2.36 mm
60 60 ds=2.36-4.75 mm
ds=2.36-4.75 mm
50 50
40 40
30 30
20 20
10 10
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Porous dielectric layer thickness (mm) Porous dielectric layer thickness (mm)
Case 2 Case 4

100
90 ds<0.15 mm
ds=0.15-0.3 mm
80 ds=0.3-0.6 mm
ds=0.6-1.18 mm
Clogging particle area (%)

70 ds=1.18-2.36 mm
60 ds=2.36-4.75 mm

50
40
30
20
10
0
0 10 20 30 40 50 60 70 80 90 100
Porous dielectric layer thickness (mm)
Case 5
Fig. 21. Effects of horizontal runoff velocity on distribution frequency of clogging particles in porous concrete.

3.3.1. Impact of porous concrete porosity on clogging The graph of Case 2 provides a bar graph illustrating the change
The graph of Case 1 provides a bar graph illustrating the change in distribution frequency of clogging particles in different layers
in distribution frequency of the clogging particles in different lay- when the particles enter the porous medium with p = 33%. It can
ers with the seepage flow when the particles enter the p = 30% por- be seen from the figure that the particles (pink) with ds = 2.36–
ous media. It can be seen from that particles (pink) with ds = 2.36– 4.75 mm are most highly distributed at the 0–2 mm position in
4.75 mm are the most highly distributed at the 0–6 mm position in the porous medium layer. The rest of pink particles in the 14–22
the porous medium layer, and mainly concentrated in the porous mm and 28–32 mm locations also exhibit significant distribution.
concrete surface (<6 mm) position. Particles (yellow) with ds = The particles (yellow) with ds = 1.18–2.36 mm exhibit a peak fre-
1.18–2.36 mm are the most highly distributed at 0–6 mm in the quency value at 4–6 mm in the porous medium, and the majority
porous media layer. Furthermore, the 10–16 mm and 30–40 mm of particles are concentrated in the porous medium layer (<20
locations exhibit increased particle distribution. Particles will occa- mm) position. The other particles mostly invade up to 40 mm
sionally distribute randomly in other locations. The particles and exhibit a small amount of distribution at 20–40 mm. Particles
(green) with ds = 0.6–1.18 mm are the most highly distributed at (green) with ds = 0.6–1.18 mm are the most distributed at 0–2 mm
0–6 mm in the porous medium layer. The majority of particles in the porous medium layer. The distribution of the remaining
are concentrated in the porous medium layer (<20 mm) position. green particles in other layers is more uniform, and invades 90
The distribution frequency of the rest of green particles decreases mm at the deepest level. This is because the particles of ds = 0.3–
with increasing depth. The particles (black) with ds = 0.3–0.6 mm 0.6 mm (black) are most highly distributed at the 24–26 mm posi-
are the most highly distributed at 0–6 mm in the porous medium tions in the porous medium layer, while the remainder are uni-
layer. The distribution of particles (red) with ds = 0.15–0.3 mm is formly distributed. The distribution of particles (red) with ds =
mostly concentrated at <40 mm. The distribution frequency of par- 0.15–0.3 mm is mostly concentrated at 24–26 mm. The remaining
ticles (blue) with ds < 0.15 mm reaches the peak at the 4–6 mm red particles are distributed evenly at <60 mm. The particles with
position in the porous medium layer, and the distribution is mostly ds < 0.15 mm (blue) are most highly distributed at the 0–2 mm
concentrated at <40 mm. The distribution frequency of particles position in the porous medium layer, while the blue particles in
with sizes of <0.6 mm decreases with an increase in depth. In gen- other parts are evenly distributed.
eral, as the clogging particle size decreases, they can invade a lar- The graph of Case 3 provides a bar graph indicating the change in
ger intrusion range and their concentrated locations move deeper. the distribution frequency of clogging particles in different layers
816 J. Zhang et al. / Construction and Building Materials 187 (2018) 803–818

100 100
90 ds<0.15 mm 90 ds<0.15 mm
ds=0.15-0.3 mm ds=0.15-0.3 mm
80 80
Clogging particle area (%)

Clogging particle area (%)

ds=0.3-0.6 mm ds=0.3-0.6 mm
ds=0.6-1.18 mm ds=0.6-1.18 mm
70 70
ds=1.18-2.36 mm ds=1.18-2.36 mm
60 ds=2.36-4.75 mm 60 ds=2.36-4.75 mm

50 50
40 40
30 30
20 20
10 10
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Porous dielectric layer thickness (mm) Porous dielectric layer thickness (mm)

Case 2 Case 6

100 100
ds<0.15 mm ds<0.15 mm
ds=0.15-0.3 mm ds=0.15-0.3 mm
80 80
Clogging particle area (%)

Clogging particle area (%)

ds=0.3-0.6 mm ds=0.3-0.6 mm
ds=0.6-1.18 mm ds=0.6-1.18 mm
ds=1.18-2.36 mm ds=1.18-2.36 mm
60 ds=2.36-4.75 mm 60 ds=2.36-4.75 mm

40 40

20 20

0 0
0 20 40 60 80 100 0 20 40 60 80 100
Porous dielectric layer thickness (mm) Porous dielectric layer thickness (mm)
Case 7 Case 8
Fig. 22. Effects of seepage velocity on distribution frequency of clogging particles in porous concrete.

when the particles enter the porous medium with p = 36%. As can
be seen from the figure, the particles (pink) with ds = 2.36–4.75
mm exhibit no distribution at a depth of 0–6 mm, with the majority
distributed in the 6–10 mm position. The remaining pink particle
layers are more evenly distributed. Particles (yellow) with ds =
1.18–2.36 mm are the most highly distributed at 12–14 mm in
the porous medium layer. Furthermore, the majority of particles
are concentrated in the porous media layer at the 10–36 mm posi-
tion, with the deepest invasion reaching 40 mm. Particles (green)
with ds = 0.6–1.18 mm are most distributed at 28–20 mm in the
porous medium layer, the majority of which are concentrated at
<40 mm, while the distribution of green particles in the 40–90
mm layer is more uniform, with the deepest invasion reaching 90
mm. Particles (black) with ds = 0.3–0.6 mm are most highly dis-
tributed in the 6–8 mm position in the porous medium layer, while
the majority of particles are distributed at <60 mm and uniform dis-
tribution is identified in the 60–90 mm portion. The distribution of
ds = 0.15–0.3 mm particles (red) is mostly concentrated in the 18–
20 mm position, while the remainder of the red particles are more
evenly distributed. Particles (blue) with ds < 0.15 mm are most
highly distributed at 6–8 mm in the porous medium layer, and
overall, the distribution of blue particles at each depth is more
uniform.
Comparing Case 1 and Case 2, it can be observed that in the por-
ous medium with a porosity of p = 30% compared to p = 25%, the par-
ticles in the deeper layer increased significantly and the position
with the highest particle distributions are shifted downwards, being
evenly distributed within the porous concrete at a depth of <30 mm.
Comparing these three cases, and it can be seen that, when the
porosity reaches 35%, the distribution of large particles (ds = 1.18–
4.75 mm) has gone deep below the surface 6 mm. Moreover, particle
clogging is distributed more uniformly with depth, particularly for
the distribution of small-sized particles in each layer. According to
the clogging particle stacking results in all layers, when the porosity
p = 25%, the position with the highest distribution frequency is
located at close to 0–6 mm. Clogging particles above 12 mm in depth
account for over 50% of the total number of particles. In the case of
porosity p = 30%, the highest distribution frequency peak appears
at a depth of 0–2 mm. However, this value is reduced significantly
compared to the case of porosity p = 25%, and at a depth of 22 mm,
the percentage of clogging particles over the total number of parti-
cles can reach over 50%. In the case of porosity p = 30%, the highest
distribution frequency peak appears in the 8–10 mm layer. In com-
parison with the previous two cases, this peak value is very small
and at the depth of the 26 mm position, the percentage of clogging
particles over the total number of particles can reach over 50%.
The results indicate that, with an increase in porosity, the particles
move more freely in the pores, invade more widely in the depth
direction, and are distributed more evenly.
3.3.2. Impact of horizontal run-off velocity on porous concrete clogging
Comparing these three graphs in Fig. 21, it can be seen that the
surface runoff velocity has no effect on the distribution frequency
of clogging particles in the porous concrete.
 J. Zhang et al. / Construction and Building Materials 187 (2018) 803–818 817

100 100
90 ds<0.15 mm ds<0.15 mm
ds=0.15-0.3 mm ds=0.15-0.3 mm
80 80
Clogging particle area (%)

Clogging particle area (%)

ds=0.3-0.6 mm ds=0.3-0.6 mm
70 ds=0.6-1.18 mm
ds=1.18-2.36 mm
60 ds=2.36-4.75 mm 60
50
40 40
30
20 20
10
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Porous dielectric layer thickness (mm) Porous dielectric layer thickness (mm)
Case 2 Case 9

100
ds=0.6-1.18 mm
ds=1.18-2.36 mm
80
Clogging particle area (%)

ds=2.36-4.75 mm

60

40

20

0
0 20 40 60 80 100
Porous dielectric layer thickness (mm)
Case 10
Fig. 23. Effects of sand grading on distribution frequency of clogging particles in porous concrete.

3.3.3. Impact of seepage velocity on porous concrete clogging entire layer. When clogging particles with only fine sand are
Comparing these four graphs in Fig. 22, it is found that, with an present, the fine sand distributes more evenly in depth. There is
increase in seepage velocity, the position of peak value of particle no obvious concentrated distribution area, because coarse sand
distribution frequency obviously shifts to a deeper layer. When prevents the sand from clogging downwards. However, the com-
the seepage velocity reaches 4.8 mm/s, almost no clogging parti- parison of the well-graded sand and coarse sand indicates that
cles exist at 0–10 mm, and all of the clogging particles move to there is no obvious difference between the concentrated spots
the lower layer. Moreover, the distribution of particles with a small and intrusion depth; thus, the effect of sand grading on the particle
size (ds < 0.6 mm) tends to evenly in all layers at various seepage distribution frequency is not obvious.
velocities. According to the superimposed results of the clogging In conclusion, the most of well-graded sands could clogging in
particle frequency distribution in each layer, when the seepage the depth of 30 mm; coarse sands accumulate mostly within the
velocity vs = 1.8 mm/s, the position with the highest frequency of range of 20 mm, fine sands accumulate mostly within the range of
distribution is located close to 0–6 mm. Clogging particles above 60 mm and could reach to the depth of 100 mm. It shows that most
12 mm in depth account for over 50% of the total number of parti- clogging depth is limited to the upper 30 mm but sometimes can
cles. In the case of seepage velocity vs = 2.6 mm s, clogging parti- reach up to 100 mm under the surface. The findings can be sup-
cles with a depth of 14 mm or more account for over 50% of the ported by the previous studies. For example, Baladès et al. [1] sug-
total number of particles. In the case of seepage velocity vs = 3.7 gested that the depth of clogging is limited to the first centimeters
mm/s, the clogging particles with a depth above 18 mm account of the permeable floor. Borgwardt [2] concluded the infiltration per-
for over 50% of the total number of particles. When the seepage formance is highly affected by the fines clogged in the pores within
velocity vs = 4.8 mm/s, the clogging particles with a depth of 20 the upper 20 mm of porous paving blocks. Kayhanian et al. [6] ana-
mm or more account for over 50% of the total number of particles. lyzed the CT scanning images and concluded that generally clogging
The results demonstrate that, with an increase in seepage velocity, takes place at the top zone of the surface of open graded mixes, espe-
the particles move more deeply into the pores. cially surface layer mixes with thickness of about 25 mm.

3.3.4. Impact of sand grading on porous concrete clogging 4. Conclusion

Comparing the three cases of Fig. 23, it can be observed that it is
almost impossible for coarse sand to intrude into the middle part This study investigates the clogging problem caused by clogging
of the porous medium layer, while fine sand can penetrate the particles in the internal pores of porous concrete under the action
818 J. Zhang et al. / Construction and Building Materials 187 (2018) 803–818
of surface runoff. It is revealed that the permeable pavement struc-
ture porosity, horizontal runoff velocity, seepage velocity, and sand
grading play a role, among other factors. The study employs the
innovative use of transparent sodium polyacrylate beads instead
of opaque porous concrete, in order to observe the real-time
clogging process within the ‘‘porous concrete”. A series of indoor
comparison tests are conducted and the test results compared,
and the following conclusions can be drawn:
(1) The most obvious changes in the process of the porous con-
crete clogging are concentrated within 10 s, and the fine
sand is active in the pores for longer than the coarse sand.
(2) The most of well-graded sands could clogging in the depth of
30 mm; coarse sands accumulate mostly within the range of
20 mm, fine sands accumulate mostly within the range of
60 mm and could reach to the depth of 100 mm. It shows
that most clogging depth is limited to the upper 30 mm
but sometimes can reach up to 100 mm under the surface.
(3) With increased porosity, the particles in the uppermost layer
gradually decrease and the deep particles gradually increase.
The majority of particle sizes leading to congestion develop-
ment also gradually increase. Most of the particles move
more freely within the pore and they are more evenly dis-
tributed. Intrusion in the depth direction is more extensive.
(4) As the horizontal runoff speed increases, the peaked
clogging area process is delayed. However, this has no effect
on the clogging speed and distribution frequency of clogging
particles in the porous concrete.
(5) With increased seepage velocity, the surface particles
decrease and the deep particles increase gradually. The par-
ticle sizes in the pores move towards deeper levels. How-
ever, the impact on congestion development is not obvious.
(6) When clogging particles are only coarse sand, the time
reaching the peak value is less than that of fine sand. This
is because fine sand can pass through more pores, while
coarse sand is more likely to become stuck in the surface
pores and be difficult to move. Coarse sand hinders the fine
sand clogging, but the effect of fine sand on particles distri-
bution is not obvious, and the particles of the smallest size
migrate less rapidly.
(7) By means of observation of the clogging process trend, it is
found that the clogging area first increases rapidly, decreases
slightly, then increases for a while, and finally stabilizes.
Therefore, the clogging process is divided into the following
four stages according to the clogging area development:
rapid clogging section, clogging recovery section, slow
clogging section, and stable clogging section.
In this study, all aggregates are assumed as spheres for
simplification, while in reality, they are irregular-shaped. Despite
that, a good approximation of the experimental results is obtained,
especially considering that the input parameters used are the same
as real porous concrete. Therefore, the results obtained from this
study can be helpful to maintenance of pervious pavement.
Declarations of interest
None.
Acknowledgements
The authors would like to thank the National Program on Key
Basic Research Project of China (973 Program) (No.
2015CB058101), the Natural Science Foundations of China (Nos.
51479105 and 51379115), and Shandong Provincial Natural Science
Foundation, China (Nos. 2017GSF22101 and ZR2018MEE046).
References
[1] J.D. Baladès, M. Legret, H. Madiec, Permeable pavements: pollution
management tools, Water Sci. Technol. 32 (1) (1995) 49–56.
[2] S. Borgwardt, Long-term in-situ infiltration performance of permeable
concrete block pavement, Proceedings, 2006.
[3] X. Cui, J. Zhang, D. Huang, W. Tang, L. Wang, F. Hou, Experimental simulation of
rapid clogging process of pervious concrete pavement caused by storm water
runoff, Int. J. Pavement Eng. (2016) 1–9.
[4] O. Deo, M. Sumanasooriya, N. Neithalath, Permeability reduction in pervious
concretes due to clogging: experiments and modeling, J. Mater. Civ. Eng. 22 (7)
(2010) 741–751.
[5] L. Haselbach, S. Valavala, F. Montes, Permeability predictions for sand-clogged
Portland cement pervious concrete pavement systems, J. Environ. Manage. 81
(1) (2006) 42–49.
[6] M. Kayhanian, D. Anderson, J.T. Harvey, D. Jones, B. Muhunthan, Permeability
measurement and scan imaging to assess clogging of pervious concrete
pavements in parking lots, J. Environ. Manage. 95 (1) (2012) 114–123.
[7] F. Montes, S. Valavala, L.M. Haselbach, A new test method for porosity
measurements of Portland cement pervious concrete, J. Astm Int. 2 (1) (2005)
13.
[8] N. Neithalath, J. Weiss, J. Olek, Characterizing enhanced porosity concrete
using electrical impedance to predict acoustic and hydraulic performance,
Cem. Concr. Res. 36 (11) (2006) 2074–2085.
[9] L.N. Reddi, M. Xiao, M.G. Hajra, I.M. Lee, Physical clogging of soil filters under
constant flow rate versus cons, Can. Geotech. J. 42 (3) (2005) 804–811.
[10] D.W. Reiser, M.P. Ramey, T.A. Wesche, Chapter 4: Flushing flows, in: J.A. Gore,
G.E. Petts (Eds.), Alternatives in Regulated River Management, CRC Press, Boca
Raton, FL, 1989, pp. 91–135.
[11] M. Scholz, P. Grabowiecki, Review of permeable pavement systems, Build.
Environ. 42 (11) (2007) 3830–3836.
[12] S.A. Tan, T.F. Fwa, C.T. Han, Clogging evaluation of permeable bases, J. Transp.
Eng. 129 (3) (2003) 309–315.
[13] P.D. Tennis, M.L. Leming, D.J. Akers, Pervious concrete pavements, EB302,
Portland Cement Association, Skokie, Illinois, and National Ready Mixed
Concrete Association, Silver Spring, MD, 2004.
[14] H.N. Xu, W. Guo, Y.Q. Tan, Internal structure evolution of asphalt mixtures
during freeze-thaw cycles, Mater. Des. 86 (12) (2015) 436–446.
[15] H.N. Xu, Y.Q. Tan, X.A. Yao, X-ray computed tomography in hydraulics of
asphalt mixtures: procedure, accuracy, and application, Constr. Build. Mater.
108 (4) (2016) 10–21.
[16] C.F. Yong, A. Deletic, T.D. Fletcher, M.R. Grace, The clogging behaviour and
treatment efficiency of a range of porous pavements, 2008.
[17] J. Zhang, G. Ma, R. Ming, X. Cui, L. Li, H. Xu, Numerical study on seepage flow in
pervious concrete based on 3d CT imaging, Constr. Build. Mater. 161 (2018).