TRANSPORTATION RESEARCH RECORD 1204 77

Efficient Test Setup for Determining the

Water-Permeability of Concrete
PRASIT SooNGSWANG, MANG TIA, DAVID G. BLOOMQUIST, CoNSTANTINos MELETiou,
AND LARRY M. SESSIONS

concrete does not exist (2), quite a few different methods for
The permeability of concrete is one of the most important
measuring permeability have been used in the past. Research-
factors influencing the durability of a concrete structure, and
is a property of great interest to designers of concrete struc- ers who have studied the permeability of concrete include
tures. However, very little data on this property of concrete McMillan and Lyse (3), Norton and Pletta of Wisconsin (4),
exist, and there is not any widely used standard procedure for Ruettgers, Vidal, and Wing (5), Tyler and Erlin (1), and
determining the permeability of concrete. The major problems Meulen and Dijk (6). The various apparatuses used by these
generally encountered in performing a water-permeability test researchers to measure permeability of concrete were similar
on concrete include (a) the great amount of time required for in that all of them measured the flow rate of water under
each test; (b) the possible leakage of water through the sides pressure through the concrete specimens after a steady flow
of the test specimens; and (c) the extremely low water flow condition had been reached. The sides of the concrete spec-
rate, which makes an accurate measurement difficult to obtain. imens were sealed so that water could flow through the spec-
This paper presents a test setup and procedure that have been
imens without leakage through the sides. The sealants used
developed through experimentation and that have proven
effective in dealing with the above-listed problems. The devel- included asphalt (4, 5), epoxy (6), and paraffin and resin (3).
oped test setup can test 20 concrete specimens simultaneously The major problems generally encountered in performing
by using 20 permeability apparatuses, which are connected to a permeability test on concrete include (a) the great amount
the same regulated compressed air source. The concrete spec- of time required for each test, (b) the possible leakage of
imens can be placed into and removed from the permeability water through the sides of the specimens, and (c) the extremely
apparatuses with a minimal loss of time and with no disturb- low water flow rate that makes an accurate measurement
ance to the other test specimens. The problem of the effects of difficult to obtain. This paper presents a test setup and pro-
water evaporation on the measured flow rate has also been cedure that have been developed through experimentation
effectively dealt with. Another major advantage of this test
and that have proven to be effective in dealing with the above-
setup is that it can be duplicated easily and economically. This
paper also presents the results of permeability tests on three listed problems.
different Florida concretes subjected to four different curing
conditions. The test setup has been demonstrated to be both
reliable and efficient for determining permeability on a large PERMEABILITY APPARATUS
number of concrete specimens.
In the development of a suitable permeability apparatus, the
The permeability of concrete plays a very important role in following requirements and concerns were considered:
influencing the durability of a concrete structure. First, it
controls the rate of flow of water, which can cause disruption • The apparatus should be easily duplicated.
to the concrete upon freezing. Second, it controls the rate of • The concrete specimens could be inserted and removed
flow of chemicals, such as chloride ions, which can reduce with a minimal loss of time.
the pH of the concrete and increase the rate of corrosion of • Concrete specimens cut from larger-size concrete speci-
the steel rebars in the concrete structure. The work presented mens should be used.
in this paper deals with the development of an efficient lab- • The small quantity of water flowing through the concrete
oratory test setup for measuring the permeability of concrete specimen should be measured accurately.
and is part of an ongoing research project on the study of
factors affecting the durability of concrete. Through months of experimentation, the permeability
The permeability of concrete to water is a property of inter- apparatus as shown in Figure 1 was developed. It consists of
est to nearly all designers of concrete structures (1). Very a 1-in.-thick acrylic top plate, a 1-in.-thick acrylic base plate,
little test data are available on the permeability of concrete an acrylic tube (with an inside diameter of 4.5 in., a wall
at present, especially for Florida concrete. Although a widely thickness of 0.5 in., and a height of 1 in.), four steel bolts
used standard procedure for determining the permeability of (6.5 in. long and 0.375 in. in diameter) and four matching
nuts. The top and bottom of the acrylic tube are provided
with neoprene gaskets. One gasket is placed between the
P. Soongswang, M. Tia, D. G. Bloomquist, and C. Meletiou, Department
of Civil Engineering, University of Florida, Gainesville, 32611. L. M.
bottom of the acrylic tube and the top of the specimen; the
Sessions, Bridge Inspection and Evaluation Engineer, Florida Department other gasket is placed between the top of the acrylic tube and
of Transportation, 605 Suwannee St., Tallahassee, 32301. the acrylic top plate. A steel connection tube is inserted through
78 TRANSPORTATION RESEARCH RECORD 1204

the center of the acrylic top plate to provide a connection

between the water in the acrylic tube to the pressure source
and the manometer tube, which measures the amount of water
going into the specimen. A quick connection is installed on
the top acrylic plate at 2 in. from the center. It is used for
refilling water into the manometer tube.
The 20 permeability apparatuses described above were built
and connected to 20 manometer tubes, which were held in a
vertical position against two boards. All the manometer tubes
were connected to a main pressure line, which runs along the
top of the vertical boards. The pressure is held constant by a
regulator and is indicated on a pressure gauge, which is attached
to one side of the vertical board for ease of reading. Any line
can be disconnected from the main pressure line at any time
without interfering with the operation of the other lines by
turning off the individual valves. A picture of the test setup
is shown in Figure 2.
Another regulator and pressure gauge are used to regulate
the pressure in the water reservoir. The water reservoir, which
is used for refilling water into the manometer tube, consists
of an acrylic tube 8 in . in diameter and 14 in. long, top and
bottom steel plates, two quick connections (one on each end),
eight 0.375-in. x 18-in. threaded rods, and eight matching
nuts. The water reservoir used is shown in Figure 3.

PREPARATION OF TEST SPECIMENS

The test specimens, cylindrical specimens 4 in. in diameter

and 2 in. high, are made by cutting 8-in.-thick cast cylinders
FIGURE 2 Permeability test setup.

into 2-in.-thick cylinders. A 0.0625-0.125-in .-thick layer is

removed by sawing from the bottom surface of the cast spec-
imens. The circumference of the test specimens are wire-
brushed until all loose particles have been removed. Care is
taken to ensure that the surfaces of the test specimens are
dry.
A 1-in.-thick layer of epoxy (Sikadur 32, Hi-Mod) is cast
around the side of the test specimen by means of a casting
mold. The casting mold is 6 in. in diameter and 2.25 in. high
and is bolted down to a steel plate. The inner surface of the
casting mold and the top surface of the steel plate are coated
with a very thin layer of release agent. The side of the spec-
imen is first coated with a thin layer of epoxy. The specimen
is then placed in the center of the casting mold, and the
remaining gap between the casting mold and the test specimen
is filled with epoxy. The air bubbles are eliminated from the
epoxy by using a small hand vibrator. When the epoxy has
hardened, the test specimen is taken out of the casting mold.
Quick Connection The small amount of epoxy that may adhere to the top surface
Neoprene Gasket of the test specimen is removed. Figure 4 shows a casting
mold; a picture of the prepared concrete specimens coated
1" Top Plate
with epoxy on the side is shown in Figure 5.
318" Steel Rod - - --.; Acrylic Tube
Epoxy ---~r.h'.:-::=-1~"-
Concrete ---;==:E;z:l:~_,_~;__;,:..J'.~-----, Neoprene Gasket
Specimen TESTING PROCEDURE
1" Bollom
Plate
The prepared test specimen is placed into the permeability
FIGURE 1 Permeability test apparatus. apparatus in the following steps:
Soongswang et al. 79

1. The test specimen is placed on the center of the base

plate.
2. A neoprene gasket is placed on the specimen.
3. The acrylic tube is placed on the neoprene gasket .
4. Another neoprene gasket is placed on the acrylic tube.
5. The top plate is placed on the neoprene gasket and
bolted to the base plate with four 0.375-in.-diameter rods
(Figure 1).

The permeability apparatus is then connected to the

manometer tube in the following procedure:

1. The valve for the manometer tube is turned off.

2. The manometer tube is connected to the connection tube
at the center of the top plate of the permeability apparatus.
3. Pressure is applied to the main pressure line and main-
tained at 100 psi by the regulating valve, which controls the
pressure from the air compressor.
4. The water reservoir is connected to the other pressure St~el Bottom Plate
regulator through one of the two quick connections of the
water reservoir, and a pressure of 102 psi is maintained. FIGURE 4 Casting molds.
5. The other quick connection of the water reservoir is
connected to the quick connection at the top plate of the
ability apparatus and the manometer tube by opening the
permeability apparatus, and water is injected into the perme-
manometer valve slowly until the water in manometer tube
reaches the top level. The valve is then turned off, and the
permeability apparatus is disconnected from the water
reservoir.
6. The air bubbles trapped inside the permeability appa-
ratus are released through the quick connection at the top
plate.

At this stage, the permeability test can be started by simply

opening the manometer valve to connect the 100-psi air pres-
sure line to the manometer. The water level in the manometer
tube and the time at the start of the test are recorded. Possible
leakage of water between the epoxy and the side of the test
specimen is to be checked. If leakage occurs, the specimen
must be discarded. However, leakage has occurred very infre-
quently. The test is continued if no leakage is found. When
the water level in the manometer tube reaches a low level,
additional water can be added to the manometer tube from
the water reservoir. The process of refilling the manometer
tube is the same as the initial filling as described earlier.

METHOD OF MEASURING RATE OF FLOW

The amount of water flowing into a test specimen is measured

by reading the water level of the 0.125-in. manometer tube
Acrylic Tube - -- --+--..
Steel Plate - -- ----i..YAV
Steel Rod ---~

Steel Plate

FIGURE 3 Water reservoir. FIGURES Prepared concrete specimens.

80 TRANSPORTATION RESEARCH RECORD 1204

at least once a day. A plot of the cumulative amount of water CALCULATION OF COEFFICIENT OF PERMEABILITY
flowing into the test specimen vs. time is drawn for each test
specimen to determine when a steady-state flow condition has The coefficient of permeability of a test specimen is computed
been reached. The steady-state condition is usually reached from the net rate of inflow by using the following equation
within 10- 14 days . The test is continued for about 5-7 days which is based on Darcy's Law:
beyond the steady-flow condition. The average rate of inflow
in the last 5- 7 days is used as the rate of inflow for the test HQ
K = p-- (1)
specimen. Two typical plots of cumulative amount of water PA
flow vs. time are shown in Figure 6. where K = coefficient of permeability in in./sec or cm/sec.
The evaporation of water introduces complications to the p = density of water in lb/in. 3 or kg/cm 3 •
measurement of water inflow in the following two ways: H = length of test specimen in in. or cm.
P = water pressure in psi or kg/cm2 .
• The water in the manometer may evaporate into the com- Q = net rate of inflow in in .3/sec or cm3 /sec.
pressed air line , resulting in a drop in the manometer water A = cross-sectional area of test specimen in in. 2 or cm2 .
level.
• After water has flowed through the test specimen, the
water at the bottom surface of the specimen may evaporate
and produce a slight suction which can increase the rate of RESULTS OF PERMEABILITY TESTS ON FLORIDA
flow . CONCRETES
The amount of water evaporating into the air line is esti-
mated by using a reference permeameter, which is connected This section presents the results of permeability tests on three
to a manometer tube of the same size and the same main different Florida concretes using the developed test setup and
pressure line. The reference permeameter contains no test procedure . The three concretes tested were Florida type fl ,
specimen and has a base plate with no hole, creating a con- type Ill and type IV concretes made with a type II portland
dition of zero flow . The reference permeameter is filled with ce ment and F lorida aggregate . Table 1 di plays the physical
water and connected to a manometer tube and the main pres- properties of the coarse and fine aggregates u ed. A ll the con-
sure line in a manner similar to that for the other permea- crete mixes had a target slump of 3 in., which was r~ h :~ ved
meters. The water level in the reference manometer by adding the appropriate amount of a sur~
tube is read periodically to estimate the amount of water proportions of ingredients for these th.re<"
evaporating into the air line, which is then used to adjust the the properties of the fresh concrete a1 ... in Table 2.
measured rate of inflow for the other permeameters. Four different curing condition were ~ ·~"u on each of the
The problem of the effects of evaporation of water from three concrete mixes: (a) 16-hr steam curing at 160°F, (b) 28-
the bottom surface of the test specimen is dealt with by soak- day moist-room curing, (c) 3-day moi t-room curing followed
ing the lower portion of the permeameter in water about by 25-day ai r curing, and (d) 1-day curing in form followed
4-5 days after the test is started. The permeameter is not by stripping and applying cnring co111po11 nd and 27-day cming.
soaked in water immediately after the start of test to allow Two replicate specimens were tested for each of the combi-
examination for leakage. nations of concrete type and curing condition.

60

50
~
...I
w
>
w 40
...I
IC
w
I-
w 30
:E
0
z<[
:E
20
~
IL
0
IC
c 10

0
0 200 400
TIME (hrs)

FIGURE 6 Typical plots of manometer readings in permeability tests.

 TABLE 1 PHYSICAL PROPERTIES OF AGGREGATE

Coarse Agg. Fine Agg.

(Brooksvi lle Limestone) (Go 1dhead Sand)
Dry Bulk Specific Gravity 2.39 2. 54
Saturated-Surface-Dry Specific Gravity 2.47 2. 55
Absorption 3.16% o. 75%
Unit Weight (lb/ft3) 92
Fineness Modulus 2.15

TABLE 2 MIX PROPORTIONS AND PROPERTIES OF FRESH CONCRETE

Cone rete Type II III IV

w/c 0.45 0.38 0.33

cement (lb/yd3) 564 658 752
water (lb/yd3) 253 250 248
Coarse Aggregate (lb/yd3) 1702 1702 1702
Fine Aggregate (lb/yd 3 ) 1302 1238 1169
Superplasticizer (lb) 4.0 3.88 4.26
Unit Weight (lb/ft3) 145.1 146.l 147.9
Slump (inches) 2 3/4 2 3/4 3 1/4

Air Content 3% 2 1/4% 2 1/4%

TABLE 3 RESULTS OF TESTS ON FLORIDA CONCRETE

Concrete Type II III IV
w/c = 0.45 w/c = 0.38 w/c = 0.33

Curing Condition* A B c 0 A B c 0 A B c D

Coefficient of
Perme~~il ity 21 6.7 7.5 7.7 8.3 4.4 4.8 5.9 3.6 2.8 3.8 3.9
(xio- in/sec) 18 7.1 8.5 7.5 8.9 4.6 5.0 5.0 3.4 3.2 4.0 3.4

Mean Compressive
Strength 4661 7114 6574 5982 5209 7761 7209 7225 6734 8582 7800 8196
(psi)

Mean Sp 1itt i ng
Tensile Strength 252 438 483 528 385 618 477 586 507 637 562 612
(psi)

Mean Modulus
of Rupture 599 543 590 602 681 666 496 613 638 709 637 643
(psi)

Note: * Curing Conditions :

A = 16-hour steam curing c = 3-day moist & 25-day air curing
B = 28-day moist curing D = 1-day in form & 27-day with curing compound.
82 TRANSPORTATION RESEARCH RECORD 1204

u
"'c
~
.:.- 8 c
~
D
'o
.,... B
~
>
I- 6
::i
iii
<(
w
::i:
a: 4
w
Q. A 16-hour steam curing
LL B 28-dey moist curing
0 C 3-day moist & 25-day air curing
I- D 1-dey In form & 27-day with
z 2 curing compound
w
u
~
LL
w
0
()
0.33 0.36 0.45

WATER-CEMENT RATIO

FIGURE 7 Effects of water-cement ratio on the coefficient of

permeability of concrete.

The results of the permeability tests on these test specimens ACKNOWLEDGMENTS

are displayed in Table 3. For references, the mean compres-
sive, splitting tensile , and flexural strengths of these mixes The work presented in this paper is part of a research project
are also displayed in Table 3. From the permeability test supported by the Florida Department of Transportation. The
results, it can be noted that the coefficients of permeability authors also wish to extend their appreciation to Ed Dobson
of the replicate specimens (of the same mix and curing con- and Danny Richardson for their technical support and to Pat
dition) are very close to one another, while the effects of Cribbs for typing and correcting the manuscript.
concrete types (or w/c) and curing conditions on concrete
permeability are fairly significant. This indicates that the test
setup and procedure were able to produce precise and con- REFERENCES
sistent results. Figure 7 shows plots of the mean coefficient
of permeability as functions of water-cement ratio for three 1. I. L. Tyler and B. Erlin. A Proposed Simple Test Method for
curing conditions. It can be seen clearly that the permeability Determining the Permeability of Concrete . Journal of the PCA
Research and Development Laboratories , Sept. 1961 , pp . 2-7 .
of concrete decreases as the water-cement ratio decreases. Of 2. A . M. Neville and J . J . Brooks. Concrete Technology. Longman
the four curing conditions studied, the 28-day moist-room Scientific & Technical, 1987.
curing condition gave the best result. 3. F. R. McMillan and I. Lyse . Some Permeability Studies of Con-
crete. Journal of the American Concrete Institute, Vol. 25, Dec.
1929, pp. 101-142.
4. P. T. Norton and D. H . Pletta. The Permeability of Gravel Con-
CONCLUSION crete. Journal of the American Concrete Institute, Vol. 27, May
1931, pp. 1093- 1131.
The test setup and procedure as presented in this paper have 5. A. Ruettgers, E. N. Vidal, and S. P. Wing. An Investigation of
been developed through experimentation and have been dem- the Permeability of Mass Concrete with Particular Reference to
Boulder Dam. Journal of the American Concrete Institute, Vol.
onstrated to be both efficient and reliable for determining the 31, March-April 1935, pp. 382-416.
permeability of concrete, especially when a large number of 6. G . J . R. Von der Meulen and J . van Dijk. A Permeability-Testing
specimens have to be tested. One major advantage of this Apparatus for Concrete. Magazine of Concrete Research, Vol. 21 ,
test setup is that it can be duplicated easily and economically. No. 67 , June 1969, pp. 121- 123.
Water-cement ratio had a significant effect on the water
permeabilities of all the concretes in this study, regardless of
the curing conditions. The lower water-cement ratio mixes Publication of this paper sponsored by Committee on Mechanical
yielded lower permeabilities. Properties of Concrete.