Construction and Building Materials 109 (2016) 177187

Contents lists available at ScienceDirect

Construction and Building Materials

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

Compression response of normal and high strength pervious concrete

Rui Zhong, Kay Wille
Department of Civil and Environmental Engineering, University of Connecticut, 261 Glenbrook Road, Unit 3037, Storrs, CT 06269-3037, United States

h i g h l i g h t s

Investigation of compression response of pervious concrete.

Matrix mixture design up to 174 MPa.

Distinction between effective porosity and total porosity.

Evaluation of existing compressive strength predicting models.

Proposal of semi-empirical equation to predict compressive strength.

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

Article history: The last century in the construction industry has shown an increasing interest in pervious concrete, an
Received 8 January 2015 environmentally friendly material. Although emphasis has been placed on the relationship between com-
Received in revised form 28 November 2015 pressive behavior of conventional pervious concrete and total porosity, not much research has been car-
Accepted 28 January 2016
ried out to characterize the relationship between compressive behavior of pervious concrete with varied
Available online 8 February 2016
matrix strengths, aggregate to binder ratios, and aggregate sizes. In this research 27 series of pervious
concrete were mixed and experimentally tested in compressive strength using matrix strengths ranging
Keywords:
from 29 MPa to 174 MPa, aggregate to binder ratios from 2.5 to 3.5 and aggregate sizes from 1.2 mm to
Pervious concrete
Compressive strength
4.8 mm. A systematic analysis has been carried out to quantify the influence of matrix strength, amount
Effective porosity of binder, and size of aggregates on the compressive strength of pervious concrete. Four existing com-
Total porosity pressive strength prediction equations are examined and their validity for the different series investi-
gated. Based on the existing equations, an extended equation is proposed and its validity is verified by
the enhanced agreement between predicted and experimentally measured compressive strength. In this
context, the necessity to distinguish total porosity from effective porosity is highlighted and taken into
account.
2016 Elsevier Ltd. All rights reserved.

1. Introduction
Pervious concrete, also referred to as porous or permeable con-
crete, is a class of concrete characterized by a relative high volume
of connected pores, typically in the range of 1530% with pore
sizes ranging from 2 to 8 mm [1], and a water permeability of
about 26 mm/s [23]. This is achieved by intentionally incorpo-
rating continuous voids through gap grading the coarse aggregate
and eliminating or minimizing the fine aggregate. The American
Concrete Institute (ACI) defines pervious concrete as concrete
containing little, if any, fine aggregate that results in sufficient
voids to allow air and water to pass easily from the surface to
underlying layers [4]. The ability to allow water penetrating
Corresponding author.
E-mail addresses: ruz10002@engr.uconn.edu (R. Zhong), kwille@engr.uconn.edu
(K. Wille).
through its open pore structure makes it a very effective tool to
control storm water runoff. Although pervious concrete has been
used for over 30 years, the material is attracting renewed interests
due to the Federal Water Pollution Control Act [5] and the Environ-
mental Protection Agency (EPA) storm water regulations [6].
Despite the various environmental benefits, the limited bond
between the aggregates covered with cementitious matrix limits
the raveling resistance of the material and thus its structural integ-
rity. Therefore, conventional pervious concrete has typically been
applied to sidewalks, parking lots and other light traffic areas.
The bonding force between the aggregates is mainly influenced
by the contact area (aggregate size) and the matrix tensile strength
(directly correlated to the matrix compressive strength). The com-
pressive strength of pervious concrete is a quick indicator reflect-
ing the bond strengths between the gap-graded aggregates and
the effect of its porosity (aggregate to binder ratio).

http://dx.doi.org/10.1016/j.conbuildmat.2016.01.051
0950-0618/ 2016 Elsevier Ltd. All rights reserved.
178 R. Zhong, K. Wille / Construction and Building Materials 109 (2016) 177187
Typical compressive strengths of conventional pervious con-
crete with porosities between 15% and 30% range from 7 to
25 MPa [3]. The variation of compressive strength with changes
in porosity is shown in Fig. 1. Results with porosity lower than
15% or higher than 30% [1] or specimens with special treatment
[7] were not included here. Among researchers, consensus exists
that an increase of porosity will decrease the compressive strength.
Furthermore, Fig. 1 demonstrates the influence of matrix strength,
and thus bond strength between the aggregates, on the compres-
sive strength of pervious concrete. Emphasis is placed on compar-
ing the compressive strength of high strength pervious concrete
(HSPC) using a cementitious matrix with strength of 174 MPa to
the compressive strength of conventional pervious concretes from
the literature [1,813]. In most cases the matrix strength of con-
ventional pervious concrete is not provided. However, based on
the provided range of water to cement ratio (0.270.38), the
expected matrices strengths potentially range between 40 and
60 MPa based on Abrams law [14].
HSPC has been developed by Wille and Zhong [1516] through
the application of mixture design principles of ultra-high perfor-
mance concrete (UHPC) to the pervious concrete matrix.
Beside porosity and matrix strength, it is hypothesized that the
contact area between the aggregates affects the compressive
strength of pervious concrete. The contact area can be increased
by using smaller aggregates. It is worth noting that changing
aggregate size and aggregate to binder ratio could have a signifi-
cant effect on porosity. Therefore, it is necessary to isolate the
parameters to draw suitable conclusions.
In this investigation, 27 series of various pervious concretes
have been designed, characterized in their compressive strength
and porosity, and analyzed to address the following current
deficiencies:
Most publications do not provide the matrix strength of the per-
vious concrete mixes. This limits correlation between matrix
strength and pervious concrete strength.
No test results for pervious concretes with matrix strength in
excess of 150 MPa have been reported.
Fig. 1. Influence of porosity and matrix strength on the compressive strength of
pervious concrete.
Current equations to predict the compressive strength of pervi-
ous concrete have not been verified for pervious concretes using
high and ultra-high strength matrices.
The effect of aggregate size and aggregate to binder ratio has
not been taken into account to predict the compressive
strength.
The distinction between total and effective porosity and how
each affects the prediction of the compressive strength is often
overlooked.
2. Experimental study
2.1. Materials and mixture proportions of the matrix
In total 27 series of pervious concrete were investigated in their compressive
strength and porosity. The mixture designs include three different matrices, three
different sizes of quartz aggregates and three different aggregate to binder ratios.
The three matrices are designated in accordance to their strength as normal
strength matrix (NSM), high strength matrix (HSM) and ultra-high strength matrix
(UHSM). Their mixture proportions are presented in Table 1 together with their
spread values (see Fig. 2) and compressive strengths (see test set up in Fig. 3).
White cement was selected as a constituent for the design of ultra-high strength
matrix due to its high amount of tricalcium silicate (C3S) and dicalcium silicate
(C2S), as well as its low quantity of tricalcium aluminate (C3A). For simplicity, white
cement (Type I) was used for all matrices in this research meeting the ASTM C150
specification for Portland cement. For the normal strength matrix a water to cement
ratio of 0.55 was employed leading to a spread value in accordance to ASTM C230/
C230M of 220 mm (8.7 in) and a compressive strength of 29 MPa (4.2 ksi) after
28 days obtained by using a 2-inch cube in accordance to ASTM C109/C109M-13.
Reducing the water to cement ratio to 0.45 and adding a high range water reducer
(HRWR) in the amount of 0.2% to the weight of the cement yielded comparable
workability (spread = 240 mm [9.4 in]) and an enhanced compressive strength of
61 MPa (8.8 ksi). The mixture design of the ultra-high strength matrix (UHSM)
was based on prior research on the material design of ultra-high performance con-
crete [1718]. The mix proportions of the UHSM, shown in Table 1, led to a com-
pressive strength of 174 MPa (25.2 ksi). It is observed that for UHSM there is a
loss of viscosity and workability during mixing and casting of pervious concrete.
This result is because the low water to cement ratio of 0.22 accelerates the process
of surface drying. Therefore, considering the large surface area of pervious concrete
the workability of the UHSM was increased to a spread value of 340 mm (13.4 in) to
counteract the effect of viscosity loss.
2.2. Mixture proportion of pervious concretes
In addition to the variation in matrix strength (29 MPa, 61 MPa, 174 MPa), three
sizes of aggregates (1.19 mm, 2.38 mm, 4.75 mm) and three aggregate to binder
(A/B) weight ratios (2.5, 3.0, 3.5) were combined to design the 27 pervious concrete
series listed in Table 2.
2.3. Specimen preparation and curing
Prior to mixing, all aggregates were washed to remove adhered clay and impu-
rities, dried and sieved to obtain single sized aggregates. First the matrix was mixed
and then the aggregates were added. After mixing, the pervious concretes were cast
in three layers in cylindrical molds (76 mm [3 in] by 152 mm [6 in]) to full capacity
using slight vibration. Each layer was slightly compacted by a 2.4 kg (5.3 lb) ham-
mer with three blows. During casting and compaction emphasis was placed on mit-
igating discontinuity in material porosity over the specimen height. After casting,
each specimen was covered with plastic sheets and stored at room temperature
for 24 h. The next day the specimens were demolded and stored in a water tank
at 20 C for an additional 27 days. For each series, a total of five cylinders were cast,
three for compressive strength testing and two for porosity testing.
2.4. Test methods
2.4.1. Compressive strength
The compressive strength of pervious concrete was determined based on ASTM
C39. The load was applied under displacement control at a rate of 0.5 mm/min.
Prior to testing, specimens were removed from the curing tank for test preparation.
About 6 mm (1/4 in) was cut from each load surface of the cylinder. After being left
to dry in laboratory environment, the specimens were sulfur capped at both ends
ensuring smooth and parallel loading surfaces. The reported results reflect the aver-
age compressive strengths of three specimens at an age of 28 days.
 R. Zhong, K. Wille / Construction and Building Materials 109 (2016) 177187 179

(a) UHSM (b) HSM (c) NSM

Fig. 2. Spread of matrices.

(a) Load frame (b) Compressive strength test set up

Fig. 3. Compressive strength test equipment and sample set up.

where /e is the effective porosity of the sample, M1 is the mass of the oven dried
sample, M2 is the mass of the container filled with water, M3 is the container with
Table 1
saturated sample filled with water at the same level, qw is the density of water,
Mixture proportion for the matrices employed in this study.
and VT is the volume of the sample. The same level of water has been achieved by
Constituents Proportions by weight using an automatic overflow system considering the water surface tension effect.
Prior to testing, one inch was cut from both ends of the specimen, taking into
UHSM HSM NSM
account the potential discontinuation of vertical porosity distributions in pervious
Cement 1 1 1 concrete [19].
Silica fume 0.25 0.00 0.00
Silica powder 0.25 0.00 0.00
Water 0.22 0.45 0.55 3. Results and discussions
High range water reducer 0.036 0.002 0.000
Spread in mm (in) 340 (13.4) 240 (9.4) 220 (8.7) 3.1. Effective porosity versus total porosity
0
Compressive strength f c0 in MPa (ksi)* 174 (25.2) 61 (8.8) 29 (4.2)

*
The aforementioned method measures the volume of accessible
28 days, conventional curing under ambient temperature, average value out of
pores (Vap). The total volume V is the sum of total pore volume Vp,
3 specimens (2-in. cube).
volume of the aggregate Vagg and volume of the matrix Vm. Here,
effective porosity (/e) is defined as the ratio of Vap to the total
volume V. The effective porosity /e is critical for the hydraulic con-
2.4.2. Effective porosity
ductivity of pervious concrete. Many researchers correlate effective
The pore system in pervious concrete comprises the spaces between the aggre-
gates and the pores in the matrix. The volumetric portion of the accessible pores is porosity with compressive strength. However, it is the total poros-
defined as effective porosity here. In light of the high void volume, fraction standard ity that influences the compressive strength. The total porosity (/t)
tests, such as ASTM C173, are not directly applicable to measure the porosity of per- is defined as the ratio of total pore volume Vp to the total volume V
vious concrete. Therefore, the porosity was calculated using the weight difference (Eq. (2)). The total porosity takes into account the volume of non-
between an oven dried sample and a water saturated sample. Eq. (1) was used to
determine the effective porosity:
accessible pores (Vnap). The volume of non-accessible pores may be

 further divided into non-connected pores left in the space between
M1 M 2  M3
/e 1   100% 1 aggregates (Vncp) and the pores in the matrix (Vmp). Fig. 4 illustrates
qw V T
the pore system in pervious concrete.
180 R. Zhong, K. Wille / Construction and Building Materials 109 (2016) 177187

Table 2
Mixture proportions of pervious concretes.

Serial No. Mixture No.a A/B Aggregate size (mm) Matrix strength (MPa)
P-UHSM UHSM-2.5-1.19 2.5 1.19 174
UHSM-3.0-1.19 3.0 1.19 174
UHSM-3.5-1.19 3.5 1.19 174
UHSM-2.5-2.38 2.5 2.38 174
UHSM-3.0-2.38 3.0 2.38 174
UHSM-3.5-2.38 3.5 2.38 174
UHSM-2.5-4.75 2.5 4.75 174
UHSM-3.0-4.75 3.0 4.75 174
UHSM-3.5-4.75 3.5 4.75 174
P-HSM HSM-2.5-1.19 2.5 1.19 61
HSM-3.0-1.19 3.0 1.19 61
HSM-3.5-1.19 3.5 1.19 61
HSM-2.5-2.38 2.5 2.38 61
HSM-3.0-2.38 3.0 2.38 61
HSM-3.5-2.38 3.5 2.38 61
HSM-2.5-4.75 2.5 4.75 61
HSM-3.0-4.75 3.0 4.75 61
HSM-3.5-4.75 3.5 4.75 61
P-NSM NSM-2.5-1.19 2.5 1.19 29
NSM-3.0-1.19 3.0 1.19 29
NSM-3.5-1.19 3.5 1.19 29
NSM-2.5-2.38 2.5 2.38 29
NSM-3.0-2.38 3.0 2.38 29
NSM-3.5-2.38 3.5 2.38 29
NSM-2.5-4.75 2.5 4.75 29
NSM-3.0-4.75 3.0 4.75 29
NSM-3.5-4.75 3.5 4.75 29
a
Mixture identifications start with the type of matrix, followed by the aggregate to binder ratio (A/B) and the aggregate size d in millimeter.

Fig. 4. Pore system of pervious concrete.

Eq. (2) is defined as follows: where magg, mm, mw, mc, mSF and mGP are the weight of aggregate,
matrix, water, cement, silica fume and glass powder respectively,
/t V p =V 1  q=qt
2 Vagg and Vm are the dry volume of the aggregate and matrix, respec-
q M=V tively, and DVm is the matrix volume reduction after hydration, qagg,
where Vp is the total volume of the pore system, V and q are the qw, qc, qSF, qGP are the density of aggregate, water, cement, silica
bulk volume and bulk density respectively and qt is the density of fume and glass powder, respectively. Research has shown that after
the void free pervious concrete which can be calculated by Eq. (3). hydration matrix volume reduction is 25% of the volume of the non-
evaporable water (Vnew). Neville [20] reported that the weight of
magg mm M 1 A=B  mm
qt non-evaporable water is about 23% of anhydrous cement based on
V agg V m  DV m V agg V m  DV m V agg V m  DV m the assumption that cement is fully hydrated. This leads to Eq. (4)
1 A=B  mm to calculate DVm:

V agg 1  DV m =V m  V m DV m 25%  V new 25%  23%  mc-hydrated =qw 4
V agg magg =qagg
mm where Vnew is the volume of non-evaporable water, mc-hydrated is the
Vm mw =qw mc =qc mSF =qSF mGP =qGP mass of anhydrous cement which will be fully hydrated, qw is the
qm
density of water. Table 3 shows the calculated relative volume
3 reduction after cement hydration.
 R. Zhong, K. Wille / Construction and Building Materials 109 (2016) 177187 181

Table 3 Table 4
Relative volume reduction. Porosity and compressive strength of pervious concrete mixtures.

Matrix A/B DV m =V m (%) DV m =V agg V m (%) Mix /e (%) /t (%) fc0 * (MPa) /t-adj** (%) fc0 -adj* (MPa)

2.5 2.53 UHSM-2.5-1.19 9.51 19.84 65.8 27 47.8

NSM 3.0 6.62 2.24 UHSM-3.0-1.19 15.72 24.65 52.9 27 47.0
3.5 2.01 UHSM-3.5-1.19 20.56 29.18 42.3 27 47.8
UHSM-2.5-2.38 11.62 20.93 47.5 27 31.9
2.5 2.69
UHSM-3.0-2.38 19.60 27.08 31.7 27 31.9
HSM 3.0 7.20 2.38
UHSM-3.5-2.38 25.58 31.04 21.5 27 31.9
3.5 2.13
UHSM-2.5-4.75 13.95 22.46 34.9 27 23.1
2.5 1.51 UHSM-3.0-4.75 23.60 26.97 23.4 27 23.3
UHSM 3.0 4.74 1.32 UHSM-3.5-4.75 26.70 30.22 14.6 27 22.9
3.5 1.18 HSM-2.5-1.19 16.28 23.81 23.4 27 18.4
HSM-2.0-1.19 24.16 29.62 14.4 27 18.5
HSM-3.5-1.19 28.37 32.03 10.6 27 18.6
HSM-2.5-2.38 17.05 23.46 22.9 27 17.3
The low w/c ratio of UHSM does not provide sufficient water for HSM-2.0-2.38 24.56 29.74 12.9 27 17.2
full cement hydration, which is considered to be at w/c = 0.36 [20]. HSM-3.5-2.38 29.41 32.75 8.6 27 17.6
Therefore for P-UHSM, cement is separated into two parts: a HSM-2.5-4.75 20.40 25.88 18.1 27 16.4
hydrated part (amount equivalent to w/c = 0.36) and an un- HSM-2.0-4.75 25.82 30.25 11.3 27 16.4
HSM-3.5-4.75 29.57 31.63 9.1 27 16.5
hydrated part (Fig. 5). The potential volume change through the NSM-2.5-1.19 13.63 17.02 23.2 27 12.5
pozzolanic reaction of silica fume has not been considered here. NSM-3.0-1.19 23.88 27.06 12.4 27 12.5
Based on Eqs. (2)(4), the total porosity can be calculated. NSM-3.5-1.19 29.11 30.94 8.4 27 12.6
Table 4 summarizes the effective porosity, total porosity and the NSM-2.5-2.38 18.75 22.10 17.6 27 12.4
NSM-3.0-2.38 24.90 27.85 11.5 27 12.4
measured compressive strength. Additionally, Table 4 lists the
NSM-3.5-2.38 29.59 31.59 7.5 27 12.4
adjusted compressive strengths related to the reference total NSM-2.5-4.75 20.49 23.35 16.0 27 12.1
porosity of 27%. The compressive strengths and adjusted compres- NSM-3.0-4.75 25.89 28.59 10.5 27 12.2
sive strengths will be used in Section 3.2 and the following NSM-3.5-4.75 28.27 30.18 8.8 27 12.2
sections. * 0 0
f c and f cadj are measured and adjusted strength.
Finally, the measured effective porosity /e is plotted against the **
/t-adj is the reference total porosity (average value) used to calculate adjusted
calculated total porosity /t in Fig. 6. Their relationship over all the strength.
series can be expressed by Eq. (5)

/t 0:67/e 12
A similar relationship is reported by Lian and Zhuge [13]
(Fig. 6).
3.2. Effect of matrix type on compressive strength and porosity of
pervious concrete
Fig. 7 shows the influence of total porosity and matrix type on
the compressive strength of the investigated pervious concretes.
As expected a decrease in total porosity and an increase in the
matrix strength, increases the compressive strength of pervious
concrete. It is worth noting that the relationship between porosity
and concrete strength is well represented by a linear correlation for
P-NSM and P-HSM (expressed by R2). Further improvement, espe-
cially for P-UHSM, might potentially be achieved by considering
5
the effect of contact area varied through different sizes of aggre-
gates. In order to isolate the effect of total porosity on the compres-
sive strength the linear equations for P-HSM and P-NSM (Fig. 7)
and for P-UHSM (Fig. 8) will be used to adjust the compressive
strength to a reference total porosity of 27%. This value represents
the average total porosity over all the series.
The effect of matrix strength on compressive strength is illus-
trated in Fig. 9 for all the series with A/B = 2.5. Once the total poros-
ity is set to the reference porosity of 27%, the linear increase of the
adjusted compressive strength with the increase of matrix strength
becomes apparent (Fig. 9a). The same trend is observed for all ser-
ies with A/B = 3.0 (Fig. 10a) and A/B = 3.5 (Fig. 11a).
The increase of the compressive strength with the increase of
matrix strength is attributed to the better bond strength between
matrix and aggregate. Enhanced particle packing and improved

Fig. 5. Volume reduction of matrix hydration product.

182 R. Zhong, K. Wille / Construction and Building Materials 109 (2016) 177187
Fig. 6. Effective porosity versus total porosity.
Fig. 7. Influence of /t and matrix strength on compressive strength of pervious
concrete.
Fig. 8. Enhanced correlation between total porosity and strength of P-UHSM.
fine particle distribution densifies the ITZ for P-UHSM in
comparison to P-HSM and P-NSM. Consequently, better bond
strength and ultimately higher compressive strength is achieved.
Matrix-aggregate bond improvement for pervious concrete by
the addition of silica fume was also reported by Chen et al. [21].
It is notable that for the different matrices at the same A/B ratios,
different strengths were observed. This implies that aggregate size
influences the strength of pervious concrete in addition to total
porosity and matrix strength.
3.3. Effect of aggregate size on compressive strength
The effect of different aggregate sizes in the mixture design of
pervious on the compressive strength and total porosity of
pervious concrete is plotted in Figs. 1214 for P-UHSM, P-HSM
and P-NSM series, respectively. The general trend shows the
decrease in compressive strength with the increase of aggregate
size. Yang and Jiang [7], Bhutta et al. [22] and Huang et al. [23]
reported a similar trend. From Figs. 12b, 13b and 14b it is observed
that the influence of aggregate size on the total porosity is
marginal. Due to gap grading of pervious concrete a larger
aggregate size increases the pore size. Assuming unchanged total
porosity, an increase in pore size will lead to a decrease in the
number of pores per unit volume. This results in a reduced contact
area between the aggregates covered by the cementitious matrix. It
is hypothesized that the reduction in contact area per unit volume
reduces the compressive strength of pervious concrete.
Once the total porosity is set to the reference porosity of 27%, it
becomes apparent that a linear relationship exists between the
adjusted compressive strength and aggregate size. It is observed
that the influence of aggregate size on compressive strength
increases with matrix strength. The absolute value of the average
slope of the linear regression relationship between aggregate size
and pervious concrete strength increases from 0.09 for P-NSM to
0.56 and 6.4 for P-HSM and P-UHSM, respectively.
Furthermore, the adjusted compressive strengths also indicate
that the influence of the A/B ratio on the compressive strength is
marginal. This can be seen by the almost overlapping linear regres-
sion relationships for different A/B ratios. It is worth noting that
changing the A/B ratio is only one influential parameter to control
the porosity, and thus the compressive strength. The results sug-
gest that the A/B ratio does not affect the compressive strength
beyond the effect of porosity (Figs. 1214).
Additionally, visual inspection of the failed specimens confirms
that the failure of pervious concrete is initiated by microcrack
development between aggregate and matrix followed by microc-
rack coalescence (see Fig. 15).
3.4. Effect of aggregate to binder ratio (A/B) on compressive strength
Figs. 1618 show the effect of aggregate to binder ratio (A/B) on
the compressive strength and total porosity for mixtures with the
aggregate size of 1.19 mm, 2.38 mm and 4.75 mm, respectively.
The trend for the unadjusted values shows an increase in porosity
with increase of A/B ratio. This is attributed to the fact that less
matrix is available to fill the pores between aggregates leading to
a higher total porosity. Increased porosity results in decreased
compressive strength. A similar observation is reported by Park
and Tia [24]. The negligible direct influence of A/B ratio on the
strength of pervious concrete becomes apparent when adjusted
strength is plotted against A/B ratio (Figs. 1618) resulting in
nearly horizontal trend lines. This confirms that the influence of
A/B ratio on the strength of pervious concrete is solely through
its effect on total porosity in the range investigated here.
 R. Zhong, K. Wille / Construction and Building Materials 109 (2016) 177187 183

Fig. 9. Effect of matrix strength on pervious concrete properties (A/B = 2.5).

Fig. 10. Effect of matrix strength on pervious concrete properties (A/B = 3.0).

Fig. 11. Effect of matrix strength on pervious concrete properties (A/B = 3.5).

3.5. Compressive strength prediction model
Since porosity is one of the most important and easily obtained
microstructure features, extensive research has been conducted to
investigate its effect on mechanical properties. Three models
proposed for different porous materials are presented in Table 5
(model 13).
Despite progress made in the development of compressive
strength prediction models, few of them are proposed exclusively
for pervious concrete (Table 5, model 46) [1,13,28]. According
184 R. Zhong, K. Wille / Construction and Building Materials 109 (2016) 177187

Fig. 12. Effect of aggregate size on pervious concrete properties (P-UHPM).

Fig. 13. Effect of aggregate size on pervious concrete properties (P-HSM).

Fig. 14. Effect of aggregate size on pervious concrete properties (P-NSM).

to the limited literature available, most of the models are devel- high and ultra-high strength matrix and the necessity to distin-
oped using effective porosity for normal strength pervious con- guish between effective and total porosity, their validity is
crete with compressive strength below 25 MPa. Due to the use of examined here.
 R. Zhong, K. Wille / Construction and Building Materials 109 (2016) 177187 185

The linear model (model 3) is exemplified to show the correla-

tion of measured strength to predicted strength using effective and
total porosity (Fig. 19).
Less scattered data along the fitting line and improved R2 by
using total porosity confirms the necessity to distinguish total
porosity from effective porosity in predicting the compressive
strength. The performances of the aforementioned models are
compared and summarized in Table 6.
The linear model (model 3) outperforms the power (model 1),
exponential (model 2) and mixed model (model 6), which is con-
firmed by the strong linear relation between strength and porosity
in Fig. 7. The linear relationship between matrix strength (r0) and
compressive strength defined in the existing models is strongly
expressed and confirmed in Figs. 911. However, none of the mod-
els mentioned above take into consideration the effects of aggregate
size, which increases with matrix strength. Therefore, an equation
employing total porosity is proposed (Eq. (6)) which is applicable
to matrix strengths ranging from 29 to 174 MPa and able to account
for the impact of aggregate size. Since the A/B ratio affects the
Fig. 15. Failed specimen. strength only through change in total porosity, this parameter is
not explicitly incorporated into the proposed equation.

n
d
r r0 1  m/ 6
d0

Fig. 16. Effect of A/B on pervious concrete properties (d = 1.19 mm).

Fig. 17. Effect of A/B on pervious concrete properties (d = 2.38 mm).

186 R. Zhong, K. Wille / Construction and Building Materials 109 (2016) 177187

Fig. 18. Effect of A/B on pervious concrete properties (d = 4.75 mm).

Table 5
Summary of compressive strength prediction models.

No. Equation Originally for Author Year

1 r = r0(1  /) n*
Metal Balshin [25] 1949
2 r = r0 exp (n/)* Ceramic Ryshkevitch [26] 1953
3 r = r0(1  n/)* Glass Hasselmann [27] 1964
4 r = r0 exp [(m  nn)/]* Pervious concrete Chindaprasirt [28] 2009
   1
5 Pervious concrete Deo et al. [1] 2010
r a0 a1 lndMFS
lndn
a2 /A
Sp a3 lnC3D **
q
6 Pervious concrete Lian [13] 2011
r r0 1  /m en/
*
m and n are empirical constants, r0 is the matrix strength, n is the fineness modulus of aggregate.
**
dMFS is the mean free spacing of pores which is defined as the average value of uninterrupted surface-to-surface distances between all the neighboring pores, /A area
fraction of pores from 2D images, Sp is specific area of pores, C3D is three dimensional pore distribution density, dn is the number averaged pore size and can be calculated
P
Nd
from dn P Ni i , in which Ni is the number of pores with an average diameter of di.
i

Table 6
Coefficient of determination R2 for different models.

Linear Power Exponential Mixed Proposed

/e 0.87 0.16 0.18 0.24 0.88
/t 0.97 0.40 0.38 0.45 0.99

Table 7
Performance comparison of linear model and proposed equation.
 n
Mix r = r0(1  m/t) r r0 1  m/t d
d0

2 2
R m R m n
P-UHSM 0.33 3.0 0.82 2.7 0.50
P-HSM 0.97 2.6 0.99 2.6 0.13
P-NSM 0.82 2.1 0.82 2.1 0.00

and 0.82 (Table 7) for P-NSM, P-HSM and P-UHSM, respectively,

Fig. 19. Effect of /e and /t on strength prediction of linear model.
where d0 is 1.19 mm, which is the smallest aggregate size used in
this study, and m, n are empirical constants. Its usefulness is vali-
dated by the higher coefficient of determination R2 of 0.82, 0.99
in comparison to the linear model without considering the effect
of aggregate size. It is worth pointing out that the value of n
decreases as the matrix strength increases. This trend is in agree-
ment with the experimental result that the influence of aggregate
size on the strength of pervious concrete is greater for higher
strength matrices.
The measured and predicted compressive strength using the
proposed equation is plotted in Fig. 20.
 R. Zhong, K. Wille / Construction and Building Materials 109 (2016) 177187
Fig. 20. Measured compressive strength versus predicted compressive strength.
4. Conclusions
Based on the experimental tests and the parameters investi-
gated in this study, the following conclusions can be drawn:
1. It is necessary to distinguish total porosity from effective poros-
ity for predicting the compressive strength of pervious concrete.
The measured effective porosity, useful for predicting the
hydraulic conductivity, is smaller than total porosity.
2. Porosity is a necessary, but insufficient parameter, for predict-
ing the compressive strength of pervious concrete with matrix
strengths ranging from 29 MPa to 174 MPa. Matrix strength,
aggregate size, and aggregate to binder ratio significantly affect
strength.
3. An increase in matrix strength linearly increases the compres-
sive strength of pervious concrete within the range investigated
here.
4. A decrease in aggregate size increases the strength of pervious
concrete. This is attributed to the increased total bond area
between neighboring aggregates. The influence becomes more
pronounced with an increase of matrix strength.
5. The aggregate to binder ratio affects the strength of pervious
concrete only through altering the total porosity. The results
suggest no further influence on strength.
6. Based on the analyzed correlation of matrix strength, porosity,
aggregate size, and aggregate to binder ratio, a semi-empirical
model is proposed predicting the compressive strength of per-
vious concrete using matrix strengths from 29 MPa to
174 MPa, total porosities from 17% to 32%, aggregate sizes from
1.17 mm to 4.75 mm, and aggregate to binder ratios from 2.5 to
3.5. Emphasis has been placed on providing a suitable equation
for practical use with reasonable accuracy, incorporating read-
ily accessible parameter such as porosity, aggregate size and
matrix strength.
It is notable that pore size and its distribution influence the fail-
ure of pervious concrete. In this research this is indirectly
accounted for through the variation of aggregate size and aggre-
gate to binder ratio. Research is carried out to provide further
and more detailed information on the influence of pore size and
pore size distribution on the strength of pervious concrete.
187
Acknowledgements
This research has been supported by a fellowship from the State
Scholarship Fund of China, the China Scholarship Council (CSC),
and the University of Connecticut. The authors express their great
gratitude for the support. Additionally, the authors would also like
to acknowledge the support from the following companies: Elkem
Materials and Lehigh Cement Company.
References
[1] O. Deo, N. Neithalath, Compressive behavior of pervious concretes and a
quantification of the influence of random pore structure features, Mater. Sci.
Eng., A 528 (2010) 402412.
[2] P.D. Tennis, M.L. Leming, D.J. Akers, Pervious concrete pavements, EB302,
Portland Cement Association, Skokie, Illinois, 2004.
[3] V.R. Schaefer, K. Wang, M.T. Suleiman, J.T. Kevern, Mix design development for
pervious concrete in cold weather climates, Final Report. National Concrete
Pavement Technology Center, Iowa State University, Ames, IA, 2006.
[4] American Concrete Institute. Concrete Technology, ACI CT-13, 2013, pp. 78.
[5] Federal Water Pollution Control Act (33 U.S.C. 1251 et seq.) November 27th,
2002, pp. 234.
[6] EPA, Storm water technology fact sheet. Porous pavement, EPA 832-F-99-023,
Office of Water, Washington, DC, 1999.
[7] J. Yang, G. Jiang, Experimental study on properties of pervious concrete
pavement materials, Cem. Concr. Res. 33 (2003) 381386.
[8] C. Lian, Y. Zhuge, Optimum mix design of enhanced permeable concrete an
experimental investigation, Constr. Build. Mater. 24 (2010) 26642671.
[9] O. Deo, N. Neithalath, Compressive response of pervious concretes
proportioned for desired porosities, Constr. Build. Mater. 25 (2011) 4181
4189.
[10] R.C. Meininger, No-fine pervious concrete for paving, Concr. Int. 10 (8) (1998)
2027.
[11] M.T. Suleiman, J. Kevern, V.R. Schaefer, K. Wang, Effect of compaction energy
on pervious concrete properties, submitted to Concrete Technology Forum-
Focus on Pervious Concrete, National Ready Mixed Concrete Association,
Nashville, TN, May 2325, 2006.
[12] V.R. Schaefer, J.T. Kevern, An integrated study of pervious concrete mixture
design for wearing courses, Final Report DTFH61-06H-00011, National
Concrete Pavement Technology Center, Iowa State University, October 2001.
[13] C. Lian, Y. Zhuge, S. Beecham, The relationship between porosity and strength
for porous concrete, Constr. Build. Mater. 25 (2011) 42944298.
[14] S. Walker, D.L. Bloem, Effects of aggregate size on properties of concrete, ACI J.
57 (1961) 283298.
[15] Wille, K., Zhong, R., High strength pervious concrete, Proceedings of
Hipermat 2012, in: M. Schmidt, E. Fehling (Eds.), 3rd International
Symposium on UHPC and Nanotechnology for High Performance
Construction Materials, Kassel, Germany, March 79, 2012, pp. 825832.
[16] R. Zhong, K. Wille, Equal arc segment method for averaging data plots
exemplified for average stress versus strain curves of pervious concrete, ASCE
J. Mater. Civil Eng., published online May 2015. DOI: http://dx.doi.org/10.
1061/(ASCE)MT.1943-5533.000134 5, 04015071.
[17] K. Wille, A.E. Naaman, G.J. Parra-Montesinos, Ultra high performance concrete
with compressive strength exceeding 150 Mpa (22 ksi): a simple way, ACI
Mater. J. 108 (1) (2011) 4654.
[18] K. Wille, A.E. Naaman, S. El-Tawil, G.J. Parra-Montesinos, Ultra-high
performance concrete and fiber reinforced concrete: achieving strength and
ductility with no heat curing, Mater. Struct. 45 (2012) 309324.
[19] L.M. Haselbach, R.M. Freeman, Vertical porosity distributions in pervious
concrete pavement, ACI Mater. J. 103 (6) (2006) 452458.
[20] A.M. Neville, Properties of Concrete, Longman Group Ltd, England, 1995.
[21] Y. Chen, K. Wang, X. Wang, W. Zhou, Strength, fracture and fatigue of pervious
concrete, Constr. Build. Mater. 42 (2013) 97104.
[22] M.A.R. Bhutta, K. Tsuruta, J. Mirza, Evaluation of high-performance porous
concrete properties, Constr. Build. Mater. 31 (2012) 6773.
[23] B. Huang, H. Wu, X. Shu, E.G. Burdette, Laboratory evaluation of permeability
and strength of polymer-modified pervious concrete, Constr. Build. Mater. 24
(2010) 818823.
[24] S. Park, M. Tia, An experimental study on the water purification properties of
porous concrete, Cem. Concr. Res. 34 (2004) 177184.
[25] M.Y. Balshin, Relation of mechanical properties of powder metals and their
porosity and the ultimate properties of porous metal-ceramic materials, Dokl
Akad SSSR 67 (5) (1949) 831834.
[26] R. Ryshkevitch, Compression strength of porous sintered alumina and zirconia,
Cem. Concr. Res. 33 (2003) 155164.
[27] D.P.H. Hasselman, R.M. Fulrath, Effect of small fraction of spherical porosity on
elastic moduli of glass, J. Am. Ceram. Soc. 46 (1964) 5253.
[28] P. Chindaprasirt, S. Hatanaka, T. Chareerat, N. Mishima, Y. Yuasa, Effect of
binder strength and aggregate size on the compressive strength and void ratio
of porous concrete, Int. J. Miner. Metal. Mater. 16 (8) (2009) 714719.