 |
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 
|
Technical lessons learned in material handling at a demonstration site By Robert Traver,
Andrea Welker, Clay Emerson,
This article explores the lessons learned from the construction of the porous concrete surface for the Villanova University Porous Concrete Infiltration Best Management Practice. At the time of construction, this site was the first in the region to use porous concrete. Considerable knowledge was gained in the construction and reconstruction of the surface. The initial installation occurred in August 2002 and, because of the failure of the original surface, was resurfaced in May 2003. The initial failure was caused by a number of elements, the most significant being not understanding the impact the porous concrete material properties had on construction practice. The site was redesigned, and the resurfacing was performed, incorporating the lessons learned during the initial failure. The second attempt has been successful to date. Project Overview A porous concrete infiltration BMP was constructed during the retrofit of an existing paved area in the center of the Villanova University campus in Villanova, PA, near Philadelphia. The contributing watershed is approximately 50,000 ft2 and highly impervious, consisting of pedestrian walkways, rooftops, and some grass areas. The rooftops and some adjacent paved areas are directly connected to three separate rock storage beds (4 ft deep) that underlie the porous concrete surface. The rock beds are linked through piping systems to distribute the runoff between beds and allow for overflow during major storm events. The original porous concrete surface was edged using decorative pavers. Installation involved demolishing the original site, extensive regrading, and construction of the infiltration BMP. The site was designed to capture and infiltrate the first 2 inches of runoff, thereby reducing downstream stormwater volumes, streambank erosion, and nonpoint source pollution. The project joins the Villanova University Best Management Practice Research and Demonstration Park as both a research and a demonstration site (Traver 2002). Funding for the project was provided through the Pennsylvania Section 319 Nonpoint Source Pollution Program, and the site has been designated as a national monitoring site by EPA. Further information on this and other projects can be found by accessing the Villanova Urban Stormwater Partnership (VUSP) Web site (www.villanova.edu/VUSP). Initial Pour Observations
The original design consisted of three large porous surfaces bordered with pavers, as shown in Figure 1. The original plan was for the material to be batched off-site, poured and spread, leveled using a traveling vibratory screed, hand compacted, and finally covered. An admixture to improve bonding and thus strength of the concrete would be added to the drum at the site prior to pouring. Before construction, a small test pad (130 ft2) was poured, and a visual inspection proved it to be satisfactory in porosity and strength. The chief concern from the test pad experience was the cosmetic appearance of the material. Two tamping methods (hand and vibratory) were tried. The weather during pouring was hot (95-100°F). The concrete was provided by a concrete plant approximately 45 minutes away from the construction site; however, this travel time was unpredictable because of traffic congestion. When the truck arrived at the site, the admixture was added, mixed thoroughly for approximately 15-20 minutes, and poured.
The first truckload of concrete lacked the desired consistency. The concrete resembled wet, loose gravel and was discarded. It was assumed that too much water had been added and that this prevented the concrete from curing properly. The second truckload of concrete arrived shortly thereafter and appeared to have a better consistency. It was poured and then spread by shovels and rakes. The vibratory screed was not useful because of the poor workability of the material (stiffness). After the concrete was spread, a vibratory tamper was used to smooth and lightly compact it. Originally a hand tamper was used (Figure 2). Unfortunately hand tamping proved to be time-consuming, and the laborers could not keep up with the pour. Later a vibrating push tamper was modified with a larger bottom plate and used to finish the concrete behind the screed (Figure 3). This technique was successful because the compaction was distributed over a larger area, thereby improving the final surface appearance. There was a small window of opportunity in which the concrete maintained its workability. The third truckload became unworkable before it was poured completely. The contractors and the admixture industry representative discussed the issue and decided to dispose of the remaining material. By this time, the fourth truckload already had arrived on-site and had been sitting for a while. The material had already lost its workability when it was time for it to be poured and thus could not be used. At the end of each day, the porous concrete was covered with plastic sheets, but these sheets were extremely difficult to hold in place because of the size of the pour and gusting winds. They blew off long before the recommended 48-hour cure time. Similar events unfolded in the following days. To better control and predict the workability of the concrete, it was decided that the trucks should bring the concrete mix to the site without the addition of any water. The water would be added on-site, eliminating some of the uncertainties in travel and mixing times. Occasionally the mix consistency would be acceptable for placement, but the material often would become unworkable. Portions of the material that did appear acceptable at the time of placement did not cure properly. These portions were removed and replaced the following day. After consultation with the concrete plant, it was discovered that a retardant had been added by the plant to delay setup because of the excessive heat and dry conditions. This retardant was not documented on the delivery sheets.
Once the entire site was constructed, it was clear that the final finish of the porous concrete would not be acceptable. In many locations, the surface of the concrete was rutted, while other areas did not set up at all and resembled loose gravel. It was theorized that there might have been cement in the concrete that had not yet hydrated. The contractor brushed and wet the concrete in an attempt to promote hydration. Many of the areas of loose gravel were removed and patched with another batch of porous concrete. This created a tremendous variation in color and texture of the concrete, as shown in Figure 4. Due to the start of classes, there was no time left for further repairs. Speculations
Construction was restricted to the summer months when the dormitories adjacent to the site were vacant. Pouring began on July 30, 2002, and was completed on August 20. The pour occurred amid a severe heat wave with temperatures consistently in the upper 90s to low 100s. This environmental factor made the installation of the porous concrete difficult. Unpredictability in the on-site mixing process caused curing to occur much too quickly. The freshly poured concrete was covered with plastic sheets but the ill-secured covers blew off soon after being laid down (Figure 5). The large size of the pour made covering difficult. Therefore, the poured concrete was not allowed to properly hydrate, contributing to the observed surface failure. There were many inconsistencies between the different loads, including travel and mix times. Some of the loads had the concrete sitting for more than two hours before the pour began. To eliminate the influence of variable travel times on the concrete reaction time, it was decided that water should be added on-site (after the first day). This also allowed the workers to have better control of the amount of water added per truck. Because of the extreme heat and often long travel times, the plant had added a retardant to some of the trucks to slow the reaction in the mix, a standard practice with conventional concrete. The workers at the site were not made aware of the retardant, however, and its effect on porous concrete containing a bonding agent were not known. It also was found (after construction was completed) that the aggregate used was not washed as specified and contained excessive fines. These fines absorbed water, thereby increasing the amount of water required. Because the composition of each load varied (i.e., additives, fine content), it was impossible to determine the proper volume of water to add. The inability to accurately know and control the mix contributed to the surface failure. The traveling vibratory screed used to level and compact the concrete proved cumbersome and ineffective. The screed was used because the pours were 15-25 ft wide. Simply moving the concrete material side to side was difficult. The setup speed was fast compared to conventional concrete, and the end of a full drum's pour often was difficult to work, leading to ruts and irregularities in the surface appearance. Covering the concrete to promote hydration was ineffective. After the concrete had been poured and compacted, it was sprayed with the additive and water mixture, as per the specifications, and then covered with large sheets of plastic. The plastic was not put down until 15-20 minutes after compaction (or later). During the heat of the day, this delay in covering the fresh material might have played a significant role in the failure of the surface. Although this was apparently a normal construction practice for regular concrete, the large void spaces in porous concrete might have accelerated the drying process and contributed further to the problem. Once placed, the plastic sheets were held down at the edges by rocks, cinder blocks, buckets, and pieces of wood. During the 24 hours after pouring, the plastic sheets often came loose, causing the concrete to dry too quickly. Most of the problems that occurred during construction were not anticipated. As the construction team had built a satisfactory test pad, the only concern at the start of the pour was the surface appearance. Unfortunately, due to the project's relatively small construction window, there was no extra time built into the construction schedule. The failure of the first truckload of porous concrete caught the team by surprise. As the team was unaware of the plant additives and unwashed aggregate, the changes to the water mixture were made by trial and error, causing mixed results. Besides the mix's design difficulties, the inability to properly cover the material, and the short construction window, failure of the surface clearly resulted from multiple factors. While the surface was not satisfactory in appearance or durability, porosity was not a problem. The surface drained quickly and easily, without significant ponding. The porosity did not diminish during the year before replacement. Reconstruction Because of the surface failure of the first installation, the porous concrete site was redesigned and resurfaced in spring 2003. In addition to the lessons learned from the initial construction, information gathered on visits to other sites was included in the redesign and construction.
Villanova University representatives and the contractor visited three existing sites before beginning the redesign. The first, to the Centre County Visitor Center adjacent to Pennsylvania State University, was encouraging. Porous concrete (no admixture) had been used to construct a sidewalk, which had been rolled during construction. The surface was smooth and durable, but the color variation of each pour was evident (Figure 6). The other two sites were near Raleigh, NC. The first was an older site where the driveways and a cul-de-sac were all porous concrete. Again the surface was durable. The second site was a very large parking area at the University of North Carolina that had included an admixture. The surface appeared durable, but the cosmetic appearance was inconsistent (Figure 7). While this was acceptable for a parking area, it was not acceptable for a central pedestrian mall that was to be a showcase on Villanova's campus. Based on the site visits, it was decided to construct several concrete test pads to give the contractor more experience and to check durability and surfacing techniques. The first 200 ft2 pad (built in fall 2002) was acceptable for durability, but the surface color varied excessively (partially because the contractor tried several compaction techniques during the pour). To provide a larger test, four 100 ft2 porous concrete dumpster pads were installed at various locations on campus before the start of reconstruction, as shown in Figure 8. These installations were used to compare two mix designs and some compaction methods. Several new quality control measures were imposed on the batch plant. Two of the pads were constructed using a bonding agent, and two were constructed without that admixture. A vibrating plate tamper was used to compact two of the test pads, and a 48-inch, 50-gallon plastic drum roller, filled with water, was used on the other two. Following construction of the test pads, each was inspected and core samples were taken for compaction strength testing. Each test pad's porosity was verified by pouring water on the pad and observing how quickly it infiltrated: It was determined that the mix containing a binding agent and compacted with the roller provided the best results. The roller was filled with water to exert a 10 psi load. The mix containing bonding additive was more uniform in color and produced higher-strength concrete. An important advantage of the admixture was that the manufacturer provided experienced and effective field supervision to train the contractor's crew.
Initial data and observations of the site showed that the original design had more than enough porous concrete surface, and this area could be decreased without affecting the site's performance. Therefore, a new layout was designed to include narrow strips of porous concrete around the perimeter of each bed, with conventional concrete replacing the porous concrete in the middle. Pavers remained along the perimeter. The impervious concrete was crowned along the center strips to promote drainage toward the porous strips on the perimeter. These narrow strips allowed the concrete to be compacted using the 50 gal. roller instead of the vibratory screed. The roller covered the entire width of the surface and limited the distance the concrete had to be spread (Figures 9 and 11). Several Villanovans had complained that the original surface was too rough. The aggregate used in the original mix was relatively sharp, crushed rock. A smoother river pebble was used for the reconstruction. Review of the batch plant's practices resulted in the new aggregate being fully washed and dried before being introduced to the drum. The mix would be delivered dry with no admixtures introduced. The reconstruction began on May 19, 2003, with the first day of pouring occurring two days later. The spring weather was much more favorable than the hot summer conditions during the initial construction. During the reconstruction, the temperatures remained cool with highs in the mid-70s. Many days were overcast with occasional drizzle. The high humidity and cool temperatures were ideal for concrete hydration. Applying the lessons learned from the first pour, no water was added to the concrete mix until it reached the site. This gave the workers total control over what was added to the mix and the mixing times. Consistency of the mix was ensured through pre-inspection of the concrete plant. The bonding admixture's representative was in charge of onsite concrete mixing and pouring. From the moment the concrete truck arrived, the foreman was in charge of what, how much, and when to add the water and admixture. He carefully inspected the concrete to ensure it had the consistency to properly consolidate. When a problem arose, he decided the proper course of action. The representative trained one of the contractor's foremen to recognize and produce the proper mix consistency for later pours. Each truckload contained 6-7 yd3 of material. Typically, 10 gallons of water per yard of concrete were used, but this varied from 9 to 11 gallons, depending on weather conditions. The foreman remained at the top of the truck so he could see into the drum and examine the mix's consistency (Figure 10). Batching typically lasted between 5 and 10 minutes. After the water was added, the admixture was slowly introduced. The specifications call for 3.3 gallons of admixture per yard of concrete. After this was added, the concrete was thoroughly mixed again for approximately 5 minutes before the drum was reversed for release into the chute. The batches were often hand-sampled to visually determine slump. It was at this point that the foreman would decide if the consistency was correct or if it needed more water or mixing. He would periodically feel the hydration rate (and resulting heat) inside the drum by placing the palm of his hand on the rotating drum's exterior. If it was satisfactory, the material was poured. If it wasn't, steps were taken to remedy the situation until the desired consistency was achieved. The entire process of mixing took an average of 20 minutes from the introduction of water into the drum to release for pouring. This relatively quick mixing process might have played a key role in maintaining the concrete's workability during the pour. Although the workability greatly improved, it also became difficult at the end of each load as the concrete began to hydrate. This is why the drum's maximum 9 yd3 capacity was not utilized. Rakes were used to move the concrete horizontally no more than 2 ft. Excessive movement within the pour resulted in a blotchy surface. Troweling or movement with shovels was not permitted because they left spots of glazed surface. The rolling compaction technique proved superior to the vibratory screed and plate tamper used in the initial construction. The narrow pour widths allowed the crew to quickly prepare the concrete for rolling and cover. It was found that wetting the 50-gallon roller prior to and during the compaction process helped keep the concrete from sticking to its face. Immediately after rolling, the concrete was covered with plastic sheets and these were properly secured. This process was maintained throughout the pour duration. The plastic was not removed until at least 48 hours later. Current State The reconstruction was completed on May 30, 2003, and went much more smoothly than the initial effort. No areas needed to be patched, removed, or repaired. One truckload was rejected, however, because of poor consistency. It was speculated that moisture left in the truck from a previous job caused early hydration. Overall, the surface of the porous concrete met expectations. The appearance of the site is greatly improved as compared to the original design, and much of the surface remains durable to date. The new river-pebble aggregate provides a smoother and more aesthetically pleasing surface. The site has been observed during multiple rainfall events and is porous.
The redesigned site lets the central (impervious) concrete areas drain toward the porous concrete border. In some locations, the runoff from the conventional concrete is concentrated by the slopes and decorative grooves. This concentrated flow sometimes exceeds the capacity of the porous concrete strip's width during periods of intense rainfall. Some minor inconsistencies remain. In a few areas, the concrete appears almost glazed over with excess cement (Figure 12). While these areas are probably the most durable and strong, they have little or no porosity because the voids have been clogged. This could be a result of the mix being too wet. Observations made during the construction seem to support this theory. In some areas where the porous concrete meets either conventional concrete or the decorative pavers, there are slight differences in elevation (Figure 13). This difference is likely a result of excess material deposited prior to compaction. Before the concrete was compacted, the material intentionally was elevated approximately 0.5 inches such that the material, when sufficiently compacted, ideally would rest at the same level as its border. With too much excess material, the drum roller would compact the material but still leave the material above its intended finished elevation. This protruding edge could create a vulnerable spot where the concrete could be chipped. Material Testing
On May 29, 2003, four sample cylinders were molded to conduct compressive strength tests. Because there is no American Society for Testing and Materials standard procedure for molding porous concrete cylinders, a method developed by Eco-Creto of Texas Inc. was used. The method involved filling a standard 6-inch diameter cylinder with concrete in three lifts. Each lift was compacted using a 6-inch diameter tamp struck with six firm blows (Figure 14). The cylinders then were covered and allowed to cure. Two of the cylinders were broken on July 3, 2003. The results of the two breaks were 3,460 and 3,412 psi. The samples taken from that truck appeared to be too wet, however, and therefore the concrete did not appear to have sufficient void space. This might have accounted for the high compressive strengths. The remaining two cylinders are being kept for future testing. In conclusion, the reconstruction of the porous surface was successful: porous, durable, smooth, and attractive (Figure 15). Material problems have occurred on-site over the winter. Although much of the pour remains in excellent condition, areas of the porous concrete have deteriorated. This is considered a material problem, and the contractor and material supplier are working to correct it. Design Recommendations The following recommendations are made. Design the porous surface as a series of narrow strips to promote surface uniformity. Narrowing the width of the pour reduces movement during placement of the material, allows for rolling the complete width with each pass, facilitates immediate coverage to promote even hydration, and minimizes visual ridges resulting from uneven rolling. Use a series of perimeter strips, and design the nonporous surface to drain as sheet flow into the porous concrete surface. The infiltration capacity of porous concrete makes it unnecessary for the whole area to be porous, but concentration of runoff can overload the capacity of the porous concrete surface. Runoff from the regular concrete follows the expansion joints and can, during severe weather, concentrate and overload the porous concrete. A better design to capture more runoff from intense rainfall periods would anticipate the channeling effect of the expansion joints and have the runoff cross multiple porous strips, providing redundancy to protect against areas that clog over time. This type of BMP should be located in areas with a low risk of hazardous spills. Villanova's site is a pedestrian mall, located in the center of campus, with little risk of catastrophic chemical spills or other catastrophes. The only recourse for such spills would be total site reconstruction. Generic concrete specifications will not be adequate for inexperienced contractors to perform a satisfactory porous concrete installation. Either the consulting engineer or the contractor should enlist the assistance of industry manufacturers' reps to specify batching and placement techniques and formulation. Mistakes are irreversible and thus costly because this concrete system has low initial slump and fast setup times. The material specification must be enforced. Mix proportions, additives, and composition are critical. Do not assume your concrete plant operator understands the importance of washed aggregates, additives, and so on. Water should be added at the site. Carefully controlling the water volume and the contact time is critical. Owners should enter into a general construction type of contract rather than attempt to perform the work on a multiple-prime basis. Having single-point accountability makes the general contractor responsible for all tolerances, coordination of excavation and bed preparation, and concrete mixing/placement activities. Construction Recommendations Consultants and contractors must recognize that batching and placement practices for porous concrete vary greatly compared to those for conventional (impervious) structural concrete. Regarding batching, concrete should be in 5-7 yd3 batches, not the 9 yd3 capability of most drums. Delivery of dry mix is necessary, and it must be batched in dry (not washed-out) drums. If a binder admixture is specified, no other conventional concrete admixture should be introduced without assent of the bonding agent's manufacturer. Aggregates must be fully washed and dried before being batched because excessive fines diminish the function of porous concrete and affect field introduction of water and the bonding agent, which should be performed only by experienced porous concrete personnel. Due to the product's very low slump, its workability must be determined visually before the mix is run down the chute. Adjustments in the water quantity, the number of drum rotations, and admixture introduction must be performed quickly to ensure workability. Hydration heat monitoring in the drum is important. Regarding placement practices, for multiple test pours, using the same foreman and crew as will be used for actual construction is strongly recommended. Test pads should be at least 200 ft2 to detect surface unevenness. A properly batched mix cannot be moved horizontally more than a couple of feet from the point of placement. This is a major contrast to conventional concrete exhibiting typical 3- to 5-inch slumps. Use rakes instead of flat devices, such as shovels, screed boards, and trowels, which should not be used to shift or level concrete. Using such devices will produce splotchy surfaces. Installers must plan and execute placement to effect the items above. In conventional concrete practice, concrete leaves the chute and forms puddles, which first are worked by laborers several feet horizontally, next are screeded, and then are troweled smooth by cement finishers. Concrete's lack of slump and fast setup characteristics require steady movement of porous concrete chutes to replace manual working so placement is not in pools but rather in continuous bands that need minimal working before rolling. Such placement requires laborers to develop a cement finisher's eye for level and an understanding of the drum's discharge rate down the chute. Critical is the degree of communication between the laborers and the person (foreman) responsible for field mixing at the truck's drum. Rolling is the preferred compaction method. Vibratory tampers create ridges that visually are difficult to detect until after the concrete has hydrated. Careful hand tamping might be necessary in areas where rollers cannot reach. Rollers also will leave small ridges that are difficult to detect in unhydrated concrete. Concrete must be tightly covered for a minimum of 48 hours after placement. Existing concrete material specifications and testing standards are inadequate both for batch plants not experienced with porous concrete and when proprietary binding admixtures are used. The only people capable of ensuring proper batching and placement are the bonding admixture manufacturers or contractors with proven porous concrete installation track records. Flexibility must be included in the construction schedule. Porous concrete is less forgiving than regular concrete. You cannot take shortcuts, and pouring in extreme weather conditions is not an option. Some loads might have to be discarded. Extra time must be built into the construction schedule to accommodate material and environmental contingencies. Smaller pour volumes are better. The material should be rolled and covered immediately. The fast setup requires timely placement, rolling, and covering. At the end of the pours, the material is more difficult to move, and the appearance is not as good. Villanova Urban Stormwater Partnership The mission of VUSP is to advance the evolving comprehensive stormwater management field and foster the development of public and private partnerships through research on innovative stormwater BMPs, directed studies, technology transfer, and education. Firms interested in joining the partnership can access the VUSP Web site www.villanova.edu/VUSP. Acknowledgments Funding for the project was through the Pennsylvania Section 319 Nonpoint Source Pollution Program. This support does not imply endorsement of this project by US EPA or the Pennsylvania Department of Environmental Protection. The design and construction were overseen by Robert Morro, executive director of facilities; James Zaleski, director of engineering and systems; and Leo Kob of Villanova's Facility Management Office. The design firm for the infiltration portions of the project was Cahill Associates of West Chester, PA. The site contractor was N. Abbonizio Contractors. Eco-Creto of Texas Inc. was the admixture supplier. Reference Traver, Robert. 2000. "Development of a BMP Research and Demonstration Park," 9th International Conference on Urban Storm Drainage. Environmental and Water Research Institute, American Society of Civil Engineers. Robert Traver and Andrea Welker are professors in the Department of Civil and Environmental Engineering at Villlanova University and direct the Villanova Urban Stormwater Partnership (VUSP). Clay Emerson is the VUSP research associate. Michael Kwiatkowski and Tyler Ladd are graduate assistants. Leo Kob is with the Villanova University Facility Management Office. SW July/August 2004
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
Home
+ About + Subscribe
+ News + Calendar
+ Glossary Distributed
Energy | Erosion Control Magazine
| MSW Management Magazine © 2000 - 2004 FORESTER MEDIA, INC.
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
