Construction of a Precast Prestressed Concrete Pavement Demonstration Project on Interstate 57 Near Sikeston, Missouri
Chapter 6. Pavement Construction
Pre-Bid and Pre-Construction Meetings
Being the first project of its kind in Missouri, pre-bid and pre-construction meetings were an essential part of the I-57 PPCP demonstration project. Very early in the project development process, local precasters were invited to attend project meetings to discuss panel fabrication issues. Once preliminary details and specifications had been developed, a pre-bid meeting was conducted by MoDOT to reiterate the scope of the proposed project and answer questions from precast suppliers interested in bidding on the project.
After the contract that included the PPCP demonstration project had been awarded, a pre-construction meeting was held at the MoDOT offices in Sikeston to discuss the specifics of the project. Representatives from MoDOT, Gaines Construction (installation contractor), CPI Concrete Products (precast fabricator), K. Bates Steel Services (post-tensioning contractor), University of Missouri-Columbia, and The Transtec Group, Inc., were present at the meeting. The purpose of the pre-construction meeting was to open the lines of communication among all parties involved and ensure coordination of all aspects of the demonstration project. Of particular importance was coordination of the instrumentation activities by the University of Missouri with both the precast fabricator and installation contractor. This meeting was very beneficial for establishing who was responsible for each aspect of the project; such clarity was particularly important for a project which was experimental in nature.
As discussed in Chapter 3, the PPCP demonstration project was incorporated into a much larger project to reconstruct a section of jointed reinforced concrete pavement on I-57 that had begun to rapidly deteriorate in previous years. For this reconstructed section, the existing JRPC was removed, the subgrade was re-graded, and a new base constructed. Outside of the limits of the PPCP section, new edge drains were installed to replace the clogged and deteriorated existing drains, and a new jointed concrete pavement was constructed.
The reconstructed base beneath the PPCP section consisted of 100 mm (4 in.) of permeable asphalt-treated base over 100 mm (4 in.) of dense-graded granular base. To better ensure that the precast panels would be properly supported, the straightedge requirement for the permeable asphalt-treated base was specified such that a plate 150 mm (6 in.) in diameter and 3 mm (1/8 in.) thick could not be passed beneath a 6-m (20 ft) straightedge resting on the base at any location. Correction of any deviations from this requirement was completed at the contractor's expense. The contractor was also required to establish grade control for construction of the asphalt base. As such, a stringline was used for construction of the permeable asphalt-treated base. Figure 30 and Figure 31 show the construction of the permeable base and the final surface, respectively.
Figure 30. Photo. Construction of the permeable asphalt-treated base (photo from MoDOT).
Figure 31. Photo. Finished surface of the permeable asphalt-treated base.
The polyethylene sheeting used as friction-reducing material beneath the precast panels was rolled out just prior to placement of each panel, as shown in Figure 32, to minimize any damage to the sheeting that might be caused by foot traffic or construction vehicles and equipment.
Figure 32. Photo. Polyethylene sheeting rolled out prior to installation of precast panels.
Precast panel installation took place December 12-20, 2005. Because this was a first demonstration of PPCP construction in Missouri, time restrictions were not placed on the contractor for installing the panels. The intent was to evaluate the PPCP construction process to determine its viability for future rapid reconstruction and rehabilitation projects.
Transportation and Staging
The panels were transported to the job site on flat-bed tractor trailers, approximately 240 km (150 mi) from the fabrication plant in Memphis, Tennessee. Due to the weight of each panel (+/-18 Mg [20 tons]), only one panel could be transported on each truck.
The 54-Mg (60 ton) crawler crane used for panel installation was positioned on the base in front of the panel installation. Panel delivery trucks lined up on the existing pavement south of the project, then pulled onto the shoulder next to the PPCP section where the crane picked the panels off the truck and lowered them into place, as shown in Figure 33. Due to the "soft" nature of the permeable asphalt-treated base, the tracks of the crane rutted the base slightly. Sand was used to fill these ruts prior to installing the polyethylene sheeting and precast panels, as shown in Figure 34.
Figure 33. Photo. Staging of delivery trucks and the installation crane.
Figure 34. Photo. Sand was used to fill the ruts in the permeable asphalt-treated base.
Before each panel was lifted from the delivery truck, epoxy (segmental bridge adhesive) was applied to its mating faces and the compressible foam gaskets were installed in the recess around each post-tensioning duct on the recessed keyway side of the panel, as shown in Figure 35. After the polyethylene sheeting was rolled out, the panel was installed. In general, two precast panels were installed before the crane was moved.
Figure 35. Photo. Epoxy applied to the keyways (left) and installation of the foam gaskets (right).
A laser was used to align the precast panels to the centerline of the pavement. The laser was set on the panels already in place and aligned to nail heads marking the pavement centerline, pre-surveyed into the base. An alignment guide was installed in the post-tensioning duct at the centerline of each precast panel and was used to align the laser. Figure 36 shows the alignment laser and the alignment guide installed in a panel.
Figure 36. Photo. Laser and alignment guide used to align the precast panels as they were installed.
After two consecutive panels were installed, two temporary post-tensioning strands, located approximately at the quarter points, were fed through the panels. Post-tensioning rams were then used to temporarily post-tension the panels together from the face of the newly installed panels, as shown in Figure 37. Temporary post-tensioning helped to close the joints between individual panels as much as possible prior to final post-tensioning, and also seated the panels in the epoxy before it reached final set.
Figure 37. Photo. Temporary post-tensioning was used to close the joints between panels as much as possible during panel installation.
Panel installation rate was primarily dictated by availability of delivery trucks. Because the roadway was closed to traffic during construction, panel installation was not constrained by traffic-control restrictions. In general, 20 to 30 minutes were required for the installation of each panel. This included applying the epoxy, installing the panel, and completing temporary post-tensioning. Problems with alignment of the panels to the centerline of the roadway caused delays for installation of panels for sections 2, 3, and 4, as will be discussed below.
After all of the panels for each section were installed, the final longitudinal post-tensioning strands were inserted into the ducts from the stressing pockets in the joint panels and threaded through all of the panels to the stressing pockets at the opposite end of the section. A mechanical strand pusher was used to feed the strands through the panels, as shown in Figure 38. Difficulty in feeding the strands through the ducts was experienced with several tendons. These difficulties were the result of offsetting panels to correct the pavement alignment, obstructions such as ice in the post-tensioning ducts, and possibly bowed ducts within the panels. For tendons where problems pushing the strands were experienced, it was necessary to use a smaller strand as a "fish line" to pull the strands through the ducts. In the end, all strands were successfully installed and tensioned. Issues associated with strand installation are discussed in more depth below.
Figure 38. Photo. Mechanical pusher used to feed post-tensioning strands through the panels.
Final longitudinal post-tensioning was completed from the stressing pockets in the joint panels using stressing rams with a "banana nose" attachment. The banana nose permitted the ram to protrude from the pockets in the precast panels, minimizing the required size of the pockets. Two tendons were stressed simultaneously, one on either side of the pavement centerline, to expedite the stressing process and to minimize stressing eccentricities (Figure 39). Stressing began with two tendons near the pavement centerline and progressed outward toward the outside edges of the pavement.
Tendons were stressed to 80 percent of the ultimate strength of the strand or approximately 209 kN (46.9 kips). Because of the loss of prestress over the length of the tendon due to friction and wobble in the post-tensioning ducts, tendons were stressed from both ends. This ensured that the full post-tensioning force was applied to the ends of each tendon. The majority of the elongation occurred when stressing the first end. The rams were then moved to the opposite end of the tendons, and each strand was tensioned again. Elongations were measured by first tensioning the strand to 20 percent of the total required jacking force, marking the strand, then tensioning to the full required jacking force, and measuring the movement of the mark on the strand.
Figure 39. Photo. Final longitudinal post-tensioning completed using two post-tensioning rams (photo from MoDOT).
Ideally, the mid-slab anchors should be installed prior to completion of final post-tensioning so that the ends of the section will be drawn in towards the middle of the section during post-tensioning. Due to the timing of the different construction operations, this procedure was not possible for the I-57 project, and the mid-slab anchors were installed after the completion of post-tensioning. The mid-slab anchors, shown in Figure 15 in Chapter 4, consisted of a reinforcing bar, 25 mm (1 in.) in diameter, that was drilled and grouted a minimum of 610 mm (24 in.) deep into the underlying base/subgrade using the anchor sleeves cast into the anchor panels.
The Job Special Provisions required the contractor to adjust the width of the expansion joints after completion of post-tensioning. However, because post-tensioning was not complete until all four sections of the pavement had been installed, it was necessary to adjust the joint width prior to post-tensioning, while also allowing the post-tensioning operation to pull the joints open. Unfortunately, attempts to open the joints using hydraulic rams attached to the top surface of the panels were not successful, as the two halves of the joint panels had bonded together. The solution was to leave a gap between the joint panel and adjacent base panel that would be pulled closed as the expansion joint opened during the post-tensioning operation. This worked successfully for joints 2 and 3, as shown in Figure 40, but the two halves of joint 4 were bonded together so well that the joint fractured along a plane parallel to the actual expansion joint, as shown in Figure 41. This required the concrete between the actual joint and the fracture to be removed and patched.
Figure 40. Photo. Expansion joint after being pulled open during the post-tensioning operation. (photo from MoDOT).
Figure 41. Photo. Expansion joint No. 4 after fracturing during the post-tensioning operation. (photo from MoDOT).
One of the reasons for using a header-type expansion joint was that it could be diamond ground with the rest of the pavement surface. Unfortunately, construction sequencing prevented the header material from being installed until after diamond grinding had been completed. The header material was finished flush with the existing pavement after diamond grinding, however, and did not significantly affect the overall smoothness of the pavement. All expansion joints but one have performed well, as shown in Figure 42. The exception is joint 4, where the header material was installed over the patch, and the joint has not performed well under traffic due to poor bonding to the surrounding concrete. The header material deteriorated significantly (Figure 43) under traffic, but has since stabilized. Deflection testing by MoDOT (discussed in Chapter 7), however, shows load transfer across this joint comparable to the other expansion joints.
The expansion joint seals were installed after the header material. A poured silicon joint seal with back rod, compatible with the header material, was specified for the joint seal. The sealant had the expansion and compression capacity discussed in Chapter 4. The joint seals were installed under winter climatic conditions when the expansion joints would be expected to be at their maximum width. As such, during the hottest time of the day under summer climatic conditions, the joint sealant has been observed protruding from the pavement surface, as shown in Figure 44, which makes it susceptible to damage from traffic.(18)
Figure 42. Photo. Header expansion joints performing well.(18)
Figure 43. Photo. Deterioration of header material for expansion joint No. 4.(18)
Figure 44. Photo. Joint sealant protruding from the surface of the pavement during under summer climatic conditions.(18)
After completion of final longitudinal post-tensioning, the stressing pockets and mid-slab anchor sleeves were filled and finished flush with the pavement surface using a pea-gravel concrete mixture. Subsequent diamond grinding of the pavement surface removed any roughness associated with the stressing pockets.
Grouting of the post-tensioning tendons was essentially the final step in the construction process, and was completed following post-tensioning and filling of the stressing pockets. As discussed previously, grout inlets/vents were located in front of the post-tensioning anchors and at every fifth base panel, limiting the maximum spacing between vents to 15 m (50 ft). The epoxy applied to the mating edges of the panels and the compressible foam gaskets around the post-tensioning ducts were provided to minimize grout leakage from the tendon ducts.
The grout mixture used was a prepackaged, nonbleed, "cable grout" mixture specifically formulated for post-tensioning tendons. The efflux time for checking fluidity, using ASTM C939 ("Standard Method for Flow of Grout for Preplaced-Aggregate Concrete-Flow Cone Method"),(19) was required to be between 10 and 30 seconds, unless otherwise specified by the grout manufacturer. Grout was pumped from one end of each tendon to the other. Intermediate grout vents were closed off as the grout flowed through the tendons. If a significant amount of grout was pumped into a tendon and flow of grout was not observed from the end of the tendon or from intermediate vents, grout was then pumped into the nearest intermediate port.
Significant grout leakage from the tendons was apparent during the grouting operation as the quantity of grout used was several times what should have been required. In the end, all of the tendons were fully grouted, but the leakage of grout likely filled significant portions of the underlying permeable asphalt-treated base.
Based on the previous demonstration projects in Texas and California, diamond grinding to meet interstate highway smoothness requirements was expected. While the finished pavement surface was smooth enough to open to traffic if necessary, it did not meet MoDOT profilograph smoothness specifications for concrete pavement. Diamond grinding was used to bring the pavement surface back into specification. Only the traffic lanes were diamond ground to minimize the cost of diamond grinding. Figure 45 shows the final surface of the pavement after diamond grinding.
No major surface repairs were required for the finished pavement. While a number of minor spalls were observed at the joints between panels, diamond grinding removed many of these spalls. Deeper spalls will be monitored over time so that repairs can be made if needed.
Figure 45. Photo. Final pavement surface after diamond grinding.
Tying to Existing Pavement
As discussed previously, the pavement to the north and south of the PPCP section was also replaced with a conventional jointed concrete pavement. The adjacent pavement was not constructed until after the PPCP section was in place, so there was no need to develop a method for tying the PPCP panels to an existing pavement during panel installation. It was, however, still necessary to tie the PPCP section to the cast-in-place pavement to be constructed. To accomplish this, two-piece tie bars were cast into the joint panels abutting the cast-in-place pavement, as shown in the project plans in Appendix A. After the joint panels were installed, the second half of the tie bars were screwed into the half cast into the joint panels, and the cast-in-place pavement was constructed up to the PPCP section (Figure 46).
Figure 46. Photo. Adjacent cast-in-place pavement constructed at the end of the PPCP section. (photo from MoDOT)
Construction Issues and Challenges
This project was one of the first PPCP projects constructed in the United States, and as such, construction challenges were anticipated. This section will discuss some of the more critical construction issues and the solutions developed or recommended for future projects.
Rutting of Base-As discussed in this chapter, the weight of the crawler crane on the permeable asphalt-treated base left ruts or track indentations. While these ruts were filled with sand to help level the surface, this situation should be avoided if possible in future projects. The sand may clog the permeable base (depending on the fineness of the sand), and it was also observed that when the sand "patches" were disturbed by foot traffic it created high points that the precast panels rested on, which could have potentially led to stress concentrations in the panels and problems with fitting the panels together. Whenever "soft" bases such as this are used, the crane should be located off of the base that will be supporting the panels. If traffic constraints do not allow this, the contact pressure of the crane on the base should be determined to see if additional measures are needed to distribute the weight of the crane (tracks or outriggers) over the base to prevent indentations.
Panel Alignment-Maintaining the alignment of the pavement, or keeping the centerline of the precast panel on the centerline of the roadway, was another issue that was encountered. Section 1 was installed without any problem, but sections 2-4 were difficult to keep in alignment. One of the reasons for this was an uneven gap left between the final base panel of section 1 and the joint panel between sections 1 and 2. This gap was provided to allow the expansion joint at the end of section 1 to open during post-tensioning. Unfortunately, this uneven gap caused the alignment of subsequent panels for section 2 to creep away from the centerline of the roadway by approximately 100 mm (4 in.) at the end of section 2.
To correct the alignment, shims up to 13 mm (1/2 in.) thick were inserted in the joints between panels at the outside edge (Figure 47), and the panels were offset from one another. While offsetting helped to bring the panels back on line, it adversely affected feeding the post-tensioning strands through the ducts, requiring several of the strands to be pulled through the ducts using a "fish line" welded to the end of the post-tensioning strand. Shimming also helped bring the panels back on line, but prevented the joints between panels from closing completely, allowing incompressible material to fall into the joint, and also causing uneven distribution of the post-tensioning force across the width of the pavement (discussed in Cshapter 7).
Figure 47. Photo. Shims at the outside edge of the pavement used to correct centerline alignment.
Because it is critical that the panel joints are closed and sealed as tightly as possible, offsetting the panels may be the best solution for correcting alignment of the centerline. However, provision should be made for offsetting in the design process. A maximum permissible offset should be established based on the size of ducts and strands used for the project. Additionally, multistrand, flat ducts should be used to accommodate offsetting of up to 25-50 mm (1-2 in.).
Faulting-Several instances of faulted joints in the shoulder regions of the pavement were observed during the installation process, as shown in Figure 48. This faulting was the result of butt joints used in the shoulders rather than keyway joints, as discussed in Chapter 4. Fortunately, this faulting was relatively minor and could be removed with diamond grinding if needed. Whenever possible, however, keyways should extend across the full width of the roadway to prevent this, particularly in the traffic lanes.
Figure 48. Photo. Faulted joint observed in the shoulder region.
Expansion Joints-One of the main problems with the expansion joints was that they did not open easily during post-tensioning. This was likely caused by either a strong bond between the two halves of the joint panels or by dowel bar misalignment. Because of the strict tolerances on the dowel bars during panel fabrication, and the rigid forms used to hold the dowel bars in place, dowel misalignment was likely not the problem. It is critical to ensure that the two halves of the joint panels do not bond together during fabrication. A heavy application of bond-breaking material, such as grease or paint, should be applied, or alternatively a positive bond breaker such as plastic sheeting or Styrofoam should be included between the two halves.
The long-term performance of the header-type expansion joint will determine whether it is truly viable for PPCP expansion joints. Based on its usage for bridge joints, there should not be any problems with long-term durability. Armored joints are likely a more durable alternative for expansion joints, but require consideration of diamond grinding the finished surface and corrosion protection for steel components.
Strand Installation-The main issue with the post-tensioning operation was installing the strands in the panels. Offsetting of the panels to correct centerline alignment caused significant problems with feeding the tendons through the ducts and likely also resulted in frictional losses in the tendons as they were stressed. If panel offsetting (which is preferable to shims) is used for future projects, larger diameter or flat, multistrand ducts should be used.
Ice in the post-tensioning ducts also inhibited strand installation. Water may have accumulated in the ducts during the steam-curing operation at the fabrication plant and froze under the unusually cold conditions at the fabrication plant and project site. Based on this experience, it is recommended that compressed air be used to blow any water out of the ducts at the fabrication plant and, if possible, at the project site.
Timing of Post-Tensioning-The initial intent was for each section of precast panels to be installed and post-tensioned prior to installing subsequent sections. Unfortunately, workers were constrained to installing the precast panels as they arrived, and were not available for completing the post-tensioning. As a result, the epoxy between the precast panels set prior to applying final post-tensioning. Hardened epoxy in the joints between panels during final post-tensioning was likely the cause of some of the spalling observed at these joints. Additionally, if the epoxy bonded the panels together well enough so that they acted as a continuous concrete slab prior to final post-tensioning, the stresses in the pavement slab from daily expansion and contraction could have exacerbated the transverse cracking that was observed at the fabrication plant. Project planning is essential to ensuring that all of the different construction operations are completed in the correct sequence and that enough workers are available for all processes to be completed when needed.
Grout Leakage-The primary issue with tendon grouting was leakage of grout from the tendons between panels. Even though foam gaskets and epoxy were used in the joints between panels, significant grout leakage occurred. While epoxy and gaskets are still recommended, a positive connection between tendons across the panel joints may need to be developed to prevent grout leakage altogether.
To help familiarize other MoDOT offices, other regional State highway agencies, and the precast and concrete pavement industries with PPCP technology, MoDOT and FHWA sponsored a workshop to showcase the completed project. The workshop, entitled "National Rollout of Precast Prestressed Concrete Pavement Technology," attracted more than 60 participants from numerous State highway agencies and industry representatives from throughout the United States. The workshop was held August 22-23, 2006, approximately 8 months after the project was constructed.
The workshop, conducted on 2 half-days, featured presentations by the different parties involved with the project on the first day, including MoDOT, FHWA, CPI Concrete Products, University of Missouri-Columbia, and Transtec (see Appendix B). The first day also included a roundtable discussion which allowed participants to ask questions of those involved in the project. The second day featured a demonstration of the panel installation and a visit to the project site. For the installation demonstration, FHWA funded the fabrication of five additional precast panels, which were shipped to a MoDOT maintenance yard in Sikeston and installed by MoDOT personnel as the participants watched. For the site visit, MoDOT provided a lane closure on I-57 so that the workshop participants could walk along the actual PPCP section. Figure 49, Figure 50, and Figure 51 show photos from the workshop.
Figure 49. Photo. Presentations were provided by all parties involved with the project.
Figure 50. Photo. The workshop participants saw a live demonstration of the panel installation process at a MoDOT maintenance yard in Sikeston.
Figure 51. Photo. Workshop participants visited the actual PPCP section on I-57.