Vertical Porosity Distributions in Pervious Concrete Pavement

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

INTRODUCTION

Researchers, designers, and builders are always looking for new and improved ways to protect the environment during development in the most cost effective ways possible. One emerging technology is the use of

porous materials such as pervious concrete as an alternative paving material for surfaces such as parking lots, residential streets, and other low traffic areas. The pervious concrete pavement is used mainly due to its ability to reduce nonpoint source (NPS) runoff as compared with impervious pavements.1 Impervious pavements block the subsurface from natural water infiltration and are associated with NPS runoff which leads to negative environmental effects on the surroundings.2 The NPS runoff from these sites carries with it a large quantity of pollutants, chemicals, and hydrocarbons that impact the receiving water systems and the surrounding wildlife and plant life in their respective ecosystems.

Pervious concrete is different from conventional portland cement concrete mainly due to the reduced amount of fine aggregates (or fines), the reduced amount of water used in the mixture, and the narrow gradation of coarse aggregate. These different mixture proportions, along with careful compaction and minimal vibratory techniques used during placements, allow it to be placed with a higher percentage of voids.3,4 There are many different variations in mixtures for both conventional and pervious concrete pavements; however, a common rule of thumb is that, by mass, conventional concrete is one part cementitious material to two parts fine aggregate to three parts coarse aggregate to one half part water, and that pervious concrete is mixed with one part cementitious material to four parts coarse aggregate to one quarter part water.

It is these intentional voids within the pervious concrete that allow stormwater to percolate through to the subbase, therefore reducing the amounts of associated runoff from construction of an impervious paved surface. These voids also allow for some additional volume storage of the stormwater that may decrease the peak rate of runoff from developed areas as compared with the predevelopment runoff rates and reduce the associated environmental impacts downstream.5 This may allow the developers of a residential, commercial, or industrial area to reduce the amount of land required for retention ponds or other stormwater best management practices (BMPs) that might otherwise be required to reduce the volume and rate of runoff, preventing it from entering the water systems of the surrounding areas.4

The porosity, or percent pore volume, of pervious concrete is an important system parameter as it is related to the permeability of the pavement system, and together with the thickness of the pavement, it impacts the storage capacity and the strength of the material.6-9 There is a need to understand how porosity may vary within a field-placed slab. Researchers in Belgium have studied a technique to evaluate a vertical porosity distribution for a special type of pervious concrete used as a surface layer on roads.10 There have been few or no further studies into the porosity distributions within a field-placed section of pervious concrete as may be used for a parking area.

Many researchers have shown that there is a relationship between strength and porosity in pervious concrete, with strength typically increasing for a similar mixture with a decrease in porosity.6,7,9 It is important to understand how the porosity of a slab might vary in the vertical direction to determine the limiting effects on this variable. Both compressive and tensile strengths of pervious concrete tend to decrease with an increase in porosity.6 Pavements tend to fail in tensile stress along the bottom of a slab in the interior area and along the top of a slab in edge regions.11 Therefore, understanding the vertical porosity distribution within the medium may allow for the development of future design criteria to avoid bottom up cracking in the middle of a slab and to reduce top down cracking at the edges from large applied forces such as trucks, if these are of concern.

In addition, the permeability of pervious concrete drainage systems will also be limited by the clogging of the system. Reduced porosities in the vertical direction have the potential to create bottleneck locations for clogging as different pore sizes allow for filtering and transport of different size materials. Clogging of the system by different materials can be a major factor in the drainage of water into the subsurface through the pavement and into the water tables and aquifers below. The locations for the greatest potential for clogging may be those with the lowest porosity and those with high traffic.9 It is expected that a vertical porosity distribution with the lower porosities near the surface will produce a positive environmental effect due to filtering. Soils from polluted runoff and erosion from neighboring areas may be carried onto pervious concrete pavement surfaces. With smaller pore sizes and lower porosities near the pavement surface, these materials may be filtered out of the runoff near the surface instead of percolating deep into the media. Standard geotechnical design protocol for soil where filtering may be preferred upstream to avoid downstream clogging is to go from less porous media to effectively reduce downstream (in this case bottom) clogging of the system by filtering in the upstream location.12 Due to easy access at the surface, topical clogging on pervious concrete can be effectively removed through periodic maintenance, whereas interior clogging may be much more difficult to counteract.

A typical field placement technique for pervious concrete is for the practically no slump concrete to be placed in a single lift and then slightly compacted from the top down. Because the product has little or no slump, it does not flow or settle like conventional concretes, keeping a fairly uniform material distribution in the vertical direction when first placed into a form, acting more like granular material. There would be some bottom compaction due to weight loading, but the lift depth is not very high which would reduce the impact of this load. The material is then compacted from the top down, usually with a static roller. There is usually no further internal vibration compaction process. This type of compacting process usually results in higher loads and subsequent higher compaction (lower porosity) in the upper areas and lower compaction on the bottom.13 Consequently, it is expected that the vertical porosity distribution in a typical pervious concrete slab will have the lowest porosity region on top and the region of highest porosity near the bottom.

In summary, it is important to understand the typical vertical porosity distribution within a field-placed pervious concrete slab so that proper corrective maintenance can be determined to prevent clogging and to predict the possible potential for cracking. This research focuses on the vertical distributions of porosity from field-placed samples. The intention is to evaluate the average porosity of cored samples and then to make smaller vertical cores with horizontal cuts and use these to determine regional porosities within the cores.

RESEARCH SIGNIFICANCE

Very little research has been done on determining vertical porosity distributions in pervious concrete. A major reason to use pervious concrete as a pavement surface is to take advantage of its pore structure that allows water to percolate through, reducing nonpoint source pollution. It is important to know the vertical porosity distribution in pervious concrete to understand how infiltrating water may carry sediment through the media and lodge in bottleneck areas prone to clogging due to different porosity levels.

THEORY DEVELOPMENT AND EXPERIMENTAL METHOD

Theory development

Consider a column of freshly placed pervious concrete which is then compacted with a static roller from the top and cored after curing as shown in Fig. 1. The following assumptions are then made:

1. When initially placed and prior to compaction and with no internal vibratory compaction process, the porosity of the fresh pervious concrete column is even throughout and equal to the initial porosity of the column P'. This assumption is made based on the theory that pervious concrete acts more like a thin layer of granular material than conventional concrete;

2. Assuming no paste segregation to lower levels within the pavement thickness, after compaction and curing, the porosity distribution of the column is assumed to be linear with the lowest porosity at the top and the highest porosity at the bottom and the average porosity P in the center such that the porosity in the center P^sub 0.5h^ equals the average porosity. This may vary with much smaller or larger pavement thicknesses, different percent compactions, and other compaction techniques; but the simple linear model for typical 15 cm (6 in.) deep pervious sections will be assumed herein; and

P^sub 0.5h^ = P (1)

3. The porosity at the bottom (P^sub 0^) after compaction does not change and is equal to the precompaction average porosity or

P^sub 0^ = P' (2)

With these assumptions for the representative core of average porosity P and height h, the porosity distribution would be similar to the solid line shown in Fig. 2.

The theory of similar triangles can be used to readily show how the vertical porosity of the core at any height y (P^sub y^) in a compacted core of height h can be derived from the average porosities prior to compaction P' and postcompaction P. These are shown as the dashed and solid lines in Fig. 2. The vertical legs of the large and small triangles are h and 0.5h (half of the large), respectively. The horizontal leg of the large triangle would be 2(P' - P), and the horizontal leg of the small triangle would again be half that of the large triangle or (P' - P). Therefore, the porosity at the top P^sub h^ where y = h would be

P^sub h^ = P - (P' - P) = 2P - P' (3)

In a similar manner based on the theory of similar triangles at any height y, the ratio of the height y to the total height h after compaction is equal to the following ratio

... (4)

When Eq. (3) is combined with Eq. (4), this ratio becomes

... (5)

Equation (5) can be rearranged to solve for P^sub y^ such that the theoretical vertical porosity distribution is

... (6)

This relates the porosity at any point to the initial (precompaction) average and to the final (postcompaction) average porosities. The percent porosity before compaction, however, is not a readily measured variable and can be removed from the equation with the percent volumetric compaction of the column C^sub c^, a variable based on the degree of compaction, assuming one type of compaction technique as previously stated, using the following definitions and series of equations.

By definition, the total volume is equal to the volume of the solids plus the void volume in the columns such that for the total volumes postcompaction V^sub T^ and precompaction V'^sub T^

V^sub T^ = V^sub sol^ + V^sub voids^ (7)

and

V'^sub T^ = V'^sub sol^ + V'^sub voids^ (8)

It is assumed that the volume of the solids does not change during the compaction process from the initial column volume of the solids. Therefore, after setting the volume of the solids equal and rearranging

... (9)

By definition, the percent volumetric compaction of the column C^sub c^ is

... (10)

and the porosities precompaction P' and postcompaction P are

... (11)

and

... (12)

If Eq. (10) and (11) are substituted into Eq. (9), the result is

... (13)

Equation (10) can also be rearranged to solve for V'^sub T^ such that

... (14)

Substituting Eq. (14) into Eq. (13) yields

... (15)

When the definition of the average porosity of the core (Eq. (12)) is included, this yields

... (16)

When Eq. (16) is rearranged and solved for P ', it becomes

... (17)

Finally, by substituting the relationship between the two average porosities given in Eq. (17) into the theoretical vertical porosity distribution equation already developed (Eq. (6)), the final theoretical vertical porosity distribution equation becomes

... (18)

Therefore, based on the assumptions made, the porosity can be predicted at any height y in a pervious concrete core based on the average porosity of the core and the percent compaction used in the field placement process.

However, average porosities in vertical regions of the core may be much more useful. This research focused on the top quarter and bottom quarter volumes of a pervious concrete column, where the extremes of lower and higher porosity are predicted to be located, and compared these to the center half average porosity. As previously stated, the top porosity is particularly important for environmental reasons relating to clogging from surface infiltration; therefore, it is important to be able to predict the average vertical porosity of the top quarter of the core P^sub top^. Likewise, to determine possible modes of failure due to tensile strength in the center of a slab, it is useful to estimate the average vertical porosity in the bottom quarter of the core P^sub bo^t. The average vertical porosity in the middle half of the core P^sub mid^ can be assumed equal to the overall average porosity based on the aforementioned assumptions.

Equation (18) can be solved for the top quarter of the core by using a y-to-h ratio of approximately 0.87, for the middle half portion using a y-to-h ratio of 0.50, and for the bottom quarter of the core by using a y-to-h ratio of 0.13. Inputting these values yields the following equations for predicting the volume averaged vertical porosities in the various vertical regions of a core

P^sub top^ = P + 0.007PC^sub c^ - 0.7C^sub c^ (19)

P^sub mid^ = P (20)

P^sub bot^ = P - 0.007PC^sub c^ + 0.7C^sub c^ (21)

The research method includes measuring the porosity of a core P, slicing this core into the three vertical regions (top quarter, middle half, and bottom quarter) and measuring the porosities in these regions. There is always the potential for a change in regional porosities due to damage or knockout of aggregate or cement during the slicing process. To avoid these potential errors in the analysis, an additional variable is defined as the volume weighted average porosity of the core, where

... (22)

Assuming that there was relatively no knockout between the volume weighted average porosity P^sub w^ and the cored porosity P, the relationship would be linear or 1:1 ratio between the two such that

P^sub w^ = P (23)

Therefore, it is recommended that P^sub w^ be used for P in the experimental analyses, because this may better represent the porosities after some experimental handling in the laboratory may have affected the results.

Field placement

The placement techniques of pervious concrete differ from that of conventional portland cement concrete. A common field placement technique is as follows: For a 15 cm (6 in.) thick slab, usually a 15 to 18 cm (6 to 7 in.) form is placed with a riser (typically 1.6 cm [0.6 in.]) on top of the forms. After the pervious concrete is poured into these forms in the field, a vibrating screed is used for initial strike-off of the concrete surface. The riser is removed, and then to ensure that the surface level of the concrete is further smoothed, a static steel roller is used for final compaction. Normally there are no other finishing procedures needed. The concrete is then covered for 7 to 10 days using a minimum 6 mil polyethylene sheet for curing.4,9

Two different field-placed slabs were used in this research. The first slab was poured in Charleston, S.C., on April 1, 2004. This slab was placed on top of an existing portion of an asphalt parking lot with a form height of 16.5 cm (6.5 in.). The second slab used was placed in Spartanburg, S.C., on January 6, 2004. The subgrade for this pour was a sand and top soil mixture with a form height of 17.8 cm (7 in.).5 Both of these slabs were placed for educational, training, or research purposes with no intention of remaining available for use; therefore, the subbase was of little or no significance for environmental reasons. They were both sawcut into sections and removed for testing purposes within a few weeks of placement. Table 1 gives a listing of the materials used and weather conditions existing on the day of placement for both slabs.

Field removal technique

The slabs were allowed ample time for curing and then samples were sawcut out of the slab. Most of the samples were sawcut into approximately 30 x 30 cm (1 x 1 ft) blocks, additional samples were also sawcut into 15 x 60 cm (0.5 x 2 ft) and approximately 15 cm (0.5 ft) in depth beams. These samples were brought back into the laboratory at the University of South Carolina (USC) for testing. The Charleston slab (high porosity) and Spartanburg slab (low porosity) were labeled C and B, respectively.

Coring technique

The samples were then cored using a heavy-duty coring machine. The smaller cores were cored with a 76 mm (3 in.) cylindrical coring bit and the larger cores were cored with a 102 mm (4 in.) cylindrical coring bit. In total, there were four large cores and twenty-one small cores. All the samples were cored with the top surface facing upward and were placed on the same geometric plane with the coring machine to ensure that the outside walls of the cores remained vertical.

Some of the samples cored had very uneven bottom surfaces due to the fragmentation of the samples during coring or their placement on uneven surfaces in the field. It would be difficult to accurately find the porosity of these samples due to the difficulty in accurately measuring their volumes so the bottom of these cores were planed down to achieve better volumetric measurements on the samples. The planed cored samples are so denoted in the data Table 2, 3, and 4.

Slicing technique

After testing the cores for their porosity P, they were later sawcut horizontally in two places to create three smaller vertical cores to evaluate regional porosity distributions. The cuts were made in locations to create vertical cores approximately representing the top quarter, middle half, and bottom quarter of the cores. It had been noted in previous experiments when sawcutting and coring pervious concrete with typical field equipment, sometimes aggregate and some other pieces of the pervious concrete were knocked out of the sample by the cutting process. Therefore, to achieve the smallest level of knockout and loss in height between the vertical cores when compared with the original cores, there was a special diamond-tipped saw blade with a width of 1.14 mm (0.045 in.) used for the horizontal slicing of the cores. The small amount of aggregate or cement paste knockout lost between the vertical cores from slicing appeared to be negligible. This allowed the researchers to examine the porosity of smaller regions of the samples, especially at the top and bottom, because this is where potential clogging of the system is mostly likely to occur and where the strength characteristics are at the greatest level of concern. An example of the labeling system used between the cores and their associated vertical cores is shown in Fig. 3. The top of the first vertical core is the top surface of the original core.

Porosity testing method

The porosities were measured first on the cores P and then, after slicing, on the vertical regional cores designated as the top P^sub top^, middle P^sub mid^, and bottom P^sub bot^. The porosities of the cores were obtained through the use of Archimedes Principle to determine the volume of voids over the total core volume. This was carried out according to the standard porosity test developed by the USC Civil and Environmental Engineering Department.3 The dry mass was taken after a minimal 24 hours of drying time.

EXPERIMENTAL RESULTS AND DISCUSSION

There were a total of twenty-one 76 mm (3 in.) (smaller) and four 102 mm (4 in.) (larger) cores tested for porosity according to the standard porosity test developed at the USC Civil and Environmental Engineering laboratories.3 After the initial porosity tests were carried out, all the cores were then sliced into the three vertical core regions-the top quarter, the middle half, and the bottom quarter-and subsequently tested for porosity. These vertical core regions received no other special treatment or coating. The weighted (by volume) average porosity P^sub w^ was also calculated for each core. The results are listed in Table 2, 3, and 4. There is clearly a relationship and pattern between porosity and depth within the pervious concrete samples. The data for both sized cores (small and large) show an average regional increase in porosity with depth within the slab.

Application of theory to experimental data

The Spartanburg slab was placed with a form height of 17.8 cm (7 in.) and a riser height of 1.9 cm (0.75 in.). The Charleston slab was placed with a form height of 16.5 cm (6.5 in.) and a riser height of 1.6 cm (0.625 in.). Therefore, the respective percent compactions C^sub c^ are 10 and 9%. For Cc in the range of 9 to 10%, Eq. (19), (20), and (21) then simplify to

P^sub top(R = 9-10%)^ = 1.07P - 7 (24)

P^sub mid^ = P (25)

P^sub bot(R = 9-10%)^ = 0.93P + 7 (26)

The regional vertical porosities of the top P^sub top^, middle P^sub mid^, and bottom P^sub bot^ are plotted against the weighted average porosity P^sub w^ in Fig. 4. (As previously explained, P^sub w^, which is the weighted average porosity of the core after slicing, may better represent the samples after handling and is used for the comparisons instead of the average porosity P measured prior to slicing.) A linear regression has been performed on the three comparative sets of data (regional porosity versus weighted average porosity) with the trendlines forced through -7, 0, and 7 to fit Eq. (24) through (26), respectively. The resulting linear equations and coefficient of determinations R^sup 2^ are also depicted in Fig. 4.14 The slopes of the trendlines are within 2% of the theoretical slopes with relatively high coefficients of determination for all three of the lines. A paired Student t-test was also conducted on the paired theoretical porosities versus the experimental for all three vertical regions. The resulting t values were 0.44, 1.97, and 0.57 for the top, middle, and bottom regions, respectively, which are significant for the top and bottom, but do not show as good a fit in the middle. This variation in the middle may partially be due to the assumption of linearity because variations due to nonlinearity may be cumulative in the larger middle regions, whereas the top and bottom regions were closer to the limiting boundary conditions. In addition, previous research by USC into the method used for porosity determinations for similar sized cores has determined that the expected uncertainty in porosity due to experimental techniques is 2.2% porosity, which is significantly less than the differences between the regional porosities for each core.3 Therefore, these theoretical equations seem to fit the experimental data well, especially for the extreme boundary conditions and are fairly significant.

Although little or no knockout of aggregate or cement paste was observed during the slicing process when producing the vertical regional cores from each core, a comparison was still made between the measured porosity of each core P and the weighted average porosity of the core sections after slicing P^sub w^ to confirm that this slicing technique did not significantly alter the results. It is assumed that P^sub w^ is equal to P. These values are plotted in Fig. 5, and a linear regression performed with the trendline forced through the origin. The fit is excellent with an R^sub 2^ of 0.96 and a slope no more than 1% different from predicted. Almost all of the data points are within the expected uncertainty of 2.2% porosity. Therefore, the horizontal slicing technique appears to have had negligible impact on the experimental porosity results.

CONCLUSIONS

Pervious concrete slabs were cast with a typical placement process that includes an approximately 10% surface compaction technique with a static roller after removal of a riser, which represented an approximately 10% height change and cored samples from these slabs were tested for vertical porosity distributions. It was shown that these slabs have a fairly linear vertical porosity distribution with the lowest porosities in the top quarter, average porosities in the center half, and the higher porosities near the bottom. These regional porosities can be predicted from formulas using average core porosities and the percent compaction. The results were statistically significant for the top and bottom regions where the variations from the average were the most extreme and the differences were greater than the expected error in the measuring techniques. It is recommended that further studies also include different pervious concrete mixtures, different substrates, and different compaction methods to determine the applicability of these formulae for other circumstances. In addition, it is recommended that future studies also see how the vertical porosity distribution may vary from linear and the impact that this distribution has on strength. It is also expected that the vertical porosity distribution formulas will be altered if different lift thicknesses are used.

These regional porosities may be important for determining characteristics of pervious concrete slabs such as the potential locations for clogging. This knowledge can be further used for research into, and design and specification of, improvements such as recommended maintenance procedures. The presence of reduced porosity near the top can be a benefit for environmental reasons as clogging may tend to accumulate near the surface from solids in runoff and surface debris and this clogged area may be readily accessible for removal through periodic maintenance such as vacuuming. Higher porosities near the bottom of the slab tend to result in lower tensile strengths in this region, which leads to the recommendation for future research into the impact of higher porosities on the strength of the bottom quarter of the slab and possibly design enhancements if applicable. This information is beneficial for the application of pervious concrete as a sustainable alternative to other pavement surfaces with respect to nonpoint source pollution and other stormwater runoff concerns.

ACKNOWLEDGMENTS

The authors would like to thank Chapman Concrete of Spartanburg, S.C., Gordon Singletary of S & W Ready Mix, Van-Smith Concrete of Charleston, S.C., and members of the Carolina Ready Mixed Concrete Association (CRMCA) for their contributions to this project. They are also grateful for the support and funding for this research made available through the Center for Manufacturing and Technology (CMAT) at the University of South Carolina. The authors would also like to recognize C. Pierce, F. Montes, and A. Fox for their assistance.

NOTATION

Theoretical

C^sub c^ = percent volumetric compaction of column. When assuming constant cross-sectional area of column, this is also percent reduction in height of column (Fig. 1).

P = average porosity of cored column after compaction, %

P' = average porosity of freshly placed column prior to compaction, %

P^sub bot^ = average porosity of bottom quarter of cored column after compaction, %

P^sub h^ = porosity at depth h in cored column after compaction, %

P^sub mid^ = average porosity of middle half of cored column after compaction, %

P^sub top^ = average porosity of top quarter of cored column after compaction, %

P^sub y^ = porosity at depth y in cored column after compaction, %

P^sub 0^ = porosity at depth 0 in cored column after compaction, %

P^sub 0.5h^ = porosity at depth 0.5h in cored column after compaction, %

V^sub sol^ = volume of solids in cored column after compaction

V'^sub sol^ = volume of solids in freshly placed column prior to compaction

V^sub T^ = total volume of cored column after compaction

V'^sub T^ = total volume of freshly placed column prior to compaction

V^sub voids^ = void volume of cored column after compaction

V'^sub voids^ = void volume of freshly placed column prior to compaction

y = depth of column such that: y = 0 at bottom; y = h at top after compaction; and y = h + ?h at top prior to compaction

Experimental

P^sub w^ = experimental weighted (by volume) average porosity of vertical cores

V^sub bot^ = experimental total volume of bottom vertical core

V^sub mid^ = experimental total volume of middle vertical core

V^sub top^ = experimental total volume of top vertical core