Permeability of Stressed Concrete and Role of Fiber Reinforcement

(ProQuest Information and Learning: ... denotes formulae omitted.)

INTRODUCTION

Permeability, defined as the movement of fluid through a porous medium under an applied pressure head, is the most important property of concrete governing its long-term durability.1,2 Permeability of concrete,

in turn, is influenced by two primary factors:3 1) interconnected porosity in the cement paste; and 2) microcracks in the concrete. Porosity in the cement paste and the extent of its interconnectivity are controlled, for the most part, by the water-cement ratio (w/c), degree of hydration, and the degree of compaction. Density and location of interfacial microcracks, on the other hand, are determined by the level of applied stress or deformation, external or internal, experienced by the concrete. Internal stresses and deformations occur in concrete as a result of shrinkage, thermal gradients, abrupt changes in the hygrothermal environment, and factors causing volumetric instability. External stresses and deformations arise as a result of imposed dead and live loads.

The influence of an externally applied stress on the permeability of concrete remains poorly understood. While there are techniques available to measure the permeability of concrete without stress,4-7 very few techniques exist for measuring permeability under stress.8-10 Hearn8 subjected the specimens to stress prior to carrying out the permeability tests, but in the studies by Kermani,9 permeability tests were carried out in the presence of an applied stress. In a later development, Hearn and Lok10 also carried out nitrogen permeability tests while maintaining a stress on concrete. There exists a need to develop simple test procedures for carrying out water permeability tests on concrete where equilibrium flow can be attained rapidly, stress fluctuations can be minimized, and the variability can be controlled.

The primary purpose of the study reported herein was to investigate the influence of stress on the water permeability of concrete and to evaluate the effect of fibers.

RESEARCH SIGNIFICANCE

Permeability is without a doubt the most important transport property governing the long term durability of concrete. Unfortunately, the influence of stress on the permeability of concrete is poorly understood. Also, while it is well known that fibers improve the crack-growth resistance and energy absorption capability of stressed concrete, the exact influence of fiber reinforcement on the permeability of concrete under stress remains unknown. A test program was carried out to measure the water permeability of stressed concrete with and without fiber reinforcement. It is anticipated that the data generated in this paper will be useful in carrying out a logical life cycle engineering (LCE) analysis of concrete structures.

EXPERIMENTAL PROCEDURE

Permeability apparatus

A schematic representation of the apparatus designed for carrying out water permeability tests on concrete in the presence of an applied stress is shown in Fig. 1. The apparatus consists of five major parts: a cylindrical concrete specimen with a hollow core, a permeability cell that houses the concrete specimen, a pressurized water supply unit, an outflow measurement device, and a universal testing machine (UTM) for the application of load.

Specimen preparation

Cylindrical concrete specimens 102 mm (4 in.) in diameter and 204 mm (8 in.) long were cast with a 50 mm (2 in.) diameter hollow cylindrical core. No form of release oil was used as the oil may have affected concrete permeability.

After casting, specimens were cured in lime-saturated water until tested. On the day of the test, specimen ends were ground to obtain two smooth parallel surfaces and rubber O-rings were glued to the ends (Fig. 2) using a silicone building sealant. This was done to eliminate the leakage of water during a permeability test.

Assembling permeability cell

Various components of the permeability cell are shown in Fig. 3. To assemble, the specimen was first placed on the base with an additional O-ring (Fig. 2). The main tube (made of aluminum) was then slowly inserted between the four vertical columns with threaded top ends. Next, the upper disc was placed on the specimen with an additional O-ring. Finally, the top plate was positioned on the upper disc and secured in place by tightening the four nuts on top of the columns.

Water supply unit and outflow measurement device

The water supply unit and the outflow measurement device are shown in Fig. 4(a). The driving pressure in the water supply unit was regulated by controlling the air pressure in the unit and was kept constant during a test. The outflow measurement device consisted of an electronic scale connected to a computer. The scale continuously monitored the mass of the outflow water drained from the core of the specimen with an accuracy of 0.01 g (0.00035 oz).

Stress application

The permeability cell was mounted directly in a 200 kN (44.95 kip) hydraulic UTM (Fig. 4(a)) such that a uniform compressive stress can be applied directly on the concrete specimen housed in the cell. As expected, load relaxation occurred in the machine with time. To maintain a fixed load, the stroke of the machine was therefore adjusted during a test. A companion specimen without stress was tested using an identical setup outside the UTM (Fig. 4(b)).

Test procedure and data analysis

In a typical test, two permeability cells were assembled with identical concrete specimens in them. One of the cells was placed in the UTM where a constant stress was applied. The other cell remained outside the UTM under the conditions of no stress. Water was allowed to permeate through both the cells under identical flow conditions. The mass of the water permeated through the two cells was collected in separate collection reservoirs and its mass measured as a function of time. The water collected was related to the coefficient of water permeability (K^sub w^) by applying Darcy's law

... (1)

where K^sub w^ equals the coefficient of water permeability (m/second); Q equals the rate of water flow (m^sup 3^/second); L equals the thickness of the specimen wall (m); A equals the permeation area (m^sup 2^); and ?h equals the pressure head (m).

All tests were conducted at an age of 7 days using a constant inflow water pressure of 0.48 MPa (70 psi). Typically, it took approximately 30 hours after the start of the test to achieve conditions of full flow equilibrium. Data were recorded only after equilibrium was established, and they were used in the permeability analysis. Extreme caution was exercised to detect any leakage in the system.11 If leakage occurred, the test was rejected.

TEST PROGRAM

The mixture design of the concrete used is given in Table 1. CSA Type 10 (ASTM Type I) normal portland cement, fly ash, coarse aggregate with a maximum size of 9.5 mm (0.37 in.), saturated surface dry (SSD) clean river sand with a fineness modulus of approximately 2.5, and potable water were used. The same mixture proportion was used for plain and fiberreinforced concrete.

Two broad series of tests were performed (Table 2). In Series I, permeability tests were carried out on an unstressed fiber-reinforced concrete specimen and on a companion unstressed plain concrete specimen. In Series II, tests were conducted on stressed specimens of plain and fiber-reinforced concrete. For Series II, two applied stress levels of 0.3f^sub u^ and 0.5f^sub u^ were investigated, where f^sub u^ represents the strength of concrete in compression at the time of the test. Two replicas were tested for each fiber volume fraction and stress level unless otherwise specified. The fiber used was a virgin, fully purified plantation softwood fiber. These fibers, with an average length of approximately 2.3 mm (0.09 in.) were collated in the form of a chip and surface treated to enhance their alkali tolerance and bond with concrete. The fiber had a specific gravity of 1.1, a tensile strength of 750 MPa 108.75 ksi), and an elastic modulus of 8.3 GPa (1.20 x 10^sup 3^ ksi).

RESULTS AND DISCUSSION

Influence of fiber reinforcement in unstressed specimens (Series I)

Representative permeability plots for fiber-reinforced concrete (FRC) with 0.1, 0.3, and 0.5% fiber by volume are compared with plain concrete in Fig. 5(a), (b), and (c), respectively. Notice that fiber reinforcement was significantly effective in reducing the permeability of concrete. Notice also the stable nature of the permeability plots indicating a definite attainment of equilibrium conditions in these tests.

The permeability coefficients calculated over a 12-hour period were averaged to obtain a representative value of permeability for the specimen. Due to the large variability normally associated with water permeability results, it is often advisable to compare permeability data based on ratios.12 To quantify the effect of fiber reinforcement in unstressed conditions, ratios of the permeability R^sub I^ (I indicating Series I) between FRC and plain concrete were calculated using the following equation.

... (2)

where K^sub wFRC^ equals average water permeability coefficient for fiber-reinforced concrete without stress; and K^sub wPC^ equals average water permeability coefficient for plain concrete without stress.

The average of R^sub I^ from two separate tests was taken and was termed as a constant F, quantifying the effect of fiber reinforcement on water permeability of unstressed concrete.

The values of F are given in Table 3. Notice that the average permeability of FRC with 0.1, 0.3, and 0.5% fiber was, respectively, 0.57, 0.36, and 0.18 times that of plain concrete under conditions of no stress. These results are also plotted in Fig. 6.

The reduction in the water permeability of unstressed concrete due to fiber reinforcement, as seen in Fig. 6, is in agreement with the results of Sanjuan et al.13 but contrary to the findings of Al-Tayyib and Al-Zahrani.14 A reduction in permeability due to fiber reinforcement can be related to the two known mechanisms. First, fibers produce mixture stiffening, reduce the settlement of aggregates, and decrease bleeding. This, in turn, is expected to reduce the formation of bleed channels and decrease the ease with which flow can occur through the material.15 Second, hydrophilic fibers such as cellulose are expected to better engage water in the mixture and reduce the overall early-age shrinkage. This is expected to produce a more intact material with less internal cracking. It follows then that the apparent ability of a fiber to reduce the permeability of unstressed concrete will depend on the mixture design, fiber type and dimensions, hydrophilic/hydrophobic nature of the fiber, specimen conditioning, casting details, and specimen geometry.

Influence of fiber reinforcement in stressed specimens (Series II)

Representative permeability plots for plain concrete; FRC with 0.1% fiber, FRC with 0.3% fiber, and FRC with 0.5% fiber under various levels of applied stress are given in Fig. 7(a), (b), (c), and (d), respectively. As indicated before and shown in these plots, tests under a certain stress level were always conducted in parallel with an identical specimen under no stress. Also, as in Series I, permeability coefficients calculated over a 12-hour period after the establishment of flow equilibrium were averaged to obtain a representative permeability coefficient for the specimen.

Notice in Fig. 7(a) through (d) that the stress had a decisive influence on the permeability of both plain and fiber-reinforced concrete. When the stress was first increased to 0.3f^sub u^, both plain and FRC showed a decrease in the permeability. When the stress was increased to 0.5f^sub u^, however, plain and FRC showed very different trends. At 0.5f^sub u^, the permeability of plain concrete increased substantially over that of the unstressed specimen, but for FRC, while there was an increase in the permeability over 0.3f^sub u^, the permeability still stayed below that of the unstressed specimen. These effects are discussed in the following.

Due to the large variability normally associated with water permeability results, in order to facilitate comparison, the permeability data were normalized and compared as ratios. To quantify the effect of stress, ratios of the permeability R^sub II^ (II indicating Series II) for a specific material (plain or FRC) were calculated using the following equation

... (3)

where K^sub wstressed^ equals average water permeability coefficient under stress, and K^sub wunstressed^ equals average water permeability coefficient without stress.

The average of R^sub II^ from two separate tests was taken and was termed as a constant S, quantifying the effect of stress on water permeability of concrete. The value of S is given in Table 4 and plotted in Fig. 8. Notice that highly consistent relative permeability values (R^sub II^) were obtained for each of the materials with a low coefficient of variation.

In order to obtain a holistic view of the aforementioned test results, that is, to compare permeability values for plain and fiber-reinforced concrete with and without stress using a common and reliable datum, normalized permeability coefficients were calculated as follows.

The permeability of plain concrete under zero stress condition was chosen as the common reference point against which all other permeability coefficients were normalized. To obtain a high statistical confidence in this reference datum, 10 replicates of plain concrete under no stress were analyzed to obtain the most statistically significant value of its permeability coefficient. These results are given in Table 5.

To compare the permeability coefficients obtained from other series with K^sub wplain-unstressed^, an overall correlation constant R^sub I, II^ was calculated as

R^sub I, II^ = FxS (4)

The normalized permeability coefficients were then calculated by taking a product of the overall correlation constant R^sub I,II^ and K^sub wplain-unstressed^.

K^sub normalized^ = FxSxK^sub wplain-unstressed^ (5)

The rationale for Eq. (5) is as follows. R^sub I^ (Eq. (2)) expresses the influence of fibers in the unstressed state and R^sub II^ (Eq. (3)) expresses the influence of stress on permeability of a given materials (plain or FRC). Collectively, therefore, R^sub I,II^ expresses the relationship between permeability of a given material (plain or FRC) and that of plain concrete under no stress. This calculation is based on the assumption that the average value of water permeability coefficient for FRC under zero stress condition should be the same for Series I and II experiments. The normalized permeability coefficients for plain and FRC under various stress levels are plotted in Fig. 9.

The normalized permeability data indicate that stress has a significant influence on the permeability of concrete. An initial stress increase to 0.3f^sub u^ reduced the permeability of both plain and FRC. This can be directly attributed to pore compression that would occur under stress. At stresses greater than 0.3f^sub u^, however, while both plain and FRC showed an increase in the permeability, the increases in plain concrete were significantly greater than those in the FRC. In the case of FRC, at 0.5f^sub u^, the permeability increased, but still stayed at a level below the unstressed level.

The aforementioned observations can be related to cracking. At 0.3f^sub u^, it is conceivable that in both plain and FRC, there is no discernible cracking that can affect the flow of water. At 0.5f^sub u^, however, the stress-strain response for both plain and FRC would become nonlinear indicating the presence of cracking. As given by the Poiseuille Law,16 the flow of water through cracks is proportional to cube of the crack width

... (6)

where q^sub 0^ equals the water flow through idealized smooth cracks (m^sup 3^/second); ?^sub p^ equals the differential water pressure (N/m^sup 2^); w equals the length of the crack (m); w equals the width of the crack (m); ? equals the absolute viscosity (Ns/m^sup 2^); and d equals the flow path length of the crack (m).

In the case of FRC, one can expect the fibers to suppress cracking and hence maintain the rate of flow similar to an unstressed specimen. When combined with the phenomenon of pore compression, this implies that the permeability of FRC under stress can in fact be lower than that of an unstressed specimen. This unfortunately remains a relatively poorly understood phenomenon. Rapoport et al.17 reported that for a given imposed global strain on the material in the post-elastic range, plain concrete can demonstrate a significant increase in its permeability, but FRC will maintain its permeability closer to the unstressed state. A reduced permeability for FRC beyond cracking was also reported by Lawler et al.,18 Lepech and Li,19 and Shah et al.20 These observations are highly encouraging and are of significant importance in our global quest for durable concrete materials.21

CONCLUSIONS

Based on the research, the following conclusions can be made.

1. Fibers reduce the permeability of unstressed concrete, and the reduction appears to be proportional to the fiber volume fraction;

2. When stress is applied to concrete, up to a certain threshold value of stress, permeability is expected to decrease due to pore compression. For both plain concrete and fiber-reinforced concrete, this threshold appears to be approximately 0.3f^sub u^ where f^sub u^ is the strength in compression. Beyond this level of stress, a significant increase in permeability of plain concrete can be expected. For fiber-reinforced concrete, on the other hand, an increase in the permeability does occur beyond 0.3f^sub u^ but up to 0.5f^sub u^, the increases remain minimal. At a stress level of 0.5f^sub u^ the permeability of fiberreinforced concrete regardless of the fiber volume fraction was lower than that in the unstressed state; and

3. Based on the previous two observations, one can infer that fiber reinforcement may be expected to improve the overall durability of concrete in service.

ACKNOWLEDGMENTS

The authors wish to thank the Natural Science and Engineering Research Council of Canada for their continued financial support. Thanks are also due to Buckeye Corp. for providing the cellulose fibers for the project.