|FHWA > Engineering > Pavements > Research > FHWA-IF-06-010 > Chapter 8|
Construction of the California Precast Concrete Pavement Demonstration Project
Chapter 8. Project Evaluation and Recommendations for Future Projects
The design procedures and details for the California demonstration project were based primarily on the recommendations from the Texas pilot project.(2) An evaluation of the design procedures and details and recommendations for future projects are discussed below.
The design procedure for the California demonstration project was originally developed during the FHWA precast pavement feasibility study.(1) Essentially, the precast pavement section is designed to have a fatigue life equivalent to a much thicker conventional pavement. While the original design only called for a 250-mm (10 in.) jointed concrete pavement with a 30-year design life, the prestressed precast pavement, which varies in thickness from 250 mm (10 in.) to 330 mm (13.1 in.) will have a significantly longer fatigue life, as shown in chapter 4.
The design procedure is mechanistic in nature, incorporating both layered elastic analysis for fatigue evaluation and separate analysis using the PSCP2 program for behavior specific to prestressed pavement. While the PSCP2 design program has been calibrated using data from actual prestressed pavements, it was not originally intended to be used for precast pavements and should be modified for this purpose. The program should also account for both bonded and unbonded post-tensioning tendons. A more user-friendly, Windows-based interface that incorporates the layered elastic analysis and the PSCP2 analysis would greatly simplify the design process.
Other aspects of the design procedure, such as slab-support restraint and long-term performance, have also not been well established. Slab-support restraint is dependent upon the material on which the precast panels are placed. While push-off tests were conducted on the hot-mix asphalt leveling course for the Texas pilot project,(2) frictional characteristics for other materials, such as the LCB used for this project, have not been established, and therefore values must be assumed. Likewise, long-term performance has not been established due to the lack of precast prestressed pavements constructed thus far. Long-term performance data will help with refinement of the design procedure for a more accurate prediction of pavement life.
The design details for this demonstration project did not cause any problems with fabrication or assembly of the precast panels. Several changes were made to the design details from the original feasibility study based on the recommendations from the Texas pilot project, however. These changes greatly improved both the fabrication and construction of the precast pavement section.
Cross Slope—Perhaps the most significant change from the previous project in Texas was the change in cross slope (from 1.5 percent to 5 percent) cast into the top surface of the panels. While this change presented a challenge to the fabrication process, requiring special forms and casting procedures, the finished product demonstrated that this detail can easily be incorporated into a precast pavement.
Expansion Joints—The expansion joints were designed specifically for this project. Because Caltrans was anticipating diamond grinding the finished pavement, plain doweled joints were specified in lieu of the armored joints used for the Texas pilot project.(2) Although plain doweled joints may not be as durable under repeated wheel loads, they should provide similar performance to typical concrete pavement joints. This plain doweled joint design is not recommended, however, when excessive expansion joint opening is expected. Based on the climate, thermal expansion characteristics of the coarse aggregate, and the selected slab length, a maximum expansion joint width of 25 mm (1 in.) was anticipated for this project. It is recommended that plain doweled expansion joints only be used when the anticipated maximum expansion joint width is 25 mm (1 in.) or less.
Post-Tensioning Anchors—One of the main issues with post-tensioning during the Texas project was the self-locking spring-loaded post-tensioning anchors. Due to problems with these anchors, they were replaced with standard dead-end chuck anchors for this project. Steel plates with circular collars to receive the post-tensioning ducts were cast into the joint panels, with the face of the plate open to the anchor access pockets, where standard post-tensioning chucks (for epoxy-coated strands) were fitted over the post-tensioning strands and seated flush against the steel plates. This detail eliminated problems with anchoring the post-tensioning strands and improved the efficiency of the post-tensioning process.
Keyways—The dimensions of the keyways for the Texas pilot project occasionally resulted in a wedging action as the panels were assembled, preventing some joints from closing completely and causing minor corner cracking from stress concentrations.(2) As shown in chapter 5, the dimensions of the keyways for the California project were relaxed slightly, resulting in a better fitting keyway. For future projects, the keyway dimensions should always permit the top and bottom vertical faces of the keyways to come together, while still providing load transfer across the joint. Additionally, when casting panels with varying thickness (due to changes in cross slope) the keyways should always be straight, or parallel to the bottom of the panel.
Another keyway detail that was incorporated into the panels was a chamfer along the bottom edge of the panels, as shown in figures 15 and 16. This chamfer prevented corner breaks when the panels were removed from the forms, which was a common occurrence during the Texas pilot project.(2) A 6-mm (1/4 in.) minimum (13 mm [1/2 in.] preferred) chamfer should be cast along the bottom edge of all panels for future projects.
Lifting Anchors—While the lifting anchors used for the Texas pilot project allowed workers to attach and remove the lifting lines from the panels faster, they left a recess 100 mm (4 in.) in diameter to be cleaned and patched.(2) For this project, screw-type lifting anchors were used, greatly reducing the size of the hole to be patched. The holes from the lifting anchors were small enough that they would not have any effect on traffic prior to being patched if it were necessary to open the pavement to traffic immediately. For future projects, screw-type lifting anchors are recommended unless alternative anchors that leave an equally small or smaller hole can be utilized.
Grout Channels—Another detail added to the panels was underslab grout channels. These channels and corresponding inlets/vents were cast into the panels to permit grouting beneath the precast panels after installation. Although only minimal underslab grouting was required for the Texas project,(2) these channels allowed the contractor to perform underslab grouting as needed without having to drill holes into the finished pavement. Grout channels are highly recommended for future projects.
Stressing Pockets—The dimensions of the stressing pockets proved to be slightly too small for the stressing ram used for longitudinal post-tensioning. For future projects where stressing pockets are used, it is recommended that the pocket dimensions be adjusted according to the dimensions of the stressing ram to be used. Alternatively, if the tendons could be tensioned at their ends from smaller pockets, this would better facilitate opening the pavement to traffic sooner. This option should be explored further for future projects.
Overall, no major problems were encountered during the fabrication process. However, there were several differences between the fabrication of the panels for this project and the previous project in Texas. Below are some of the key aspects of fabrication, including changes made after the Texas project, and recommendations for future projects.
Tolerances—While the actual tolerances did not change for this project, additional tolerances were added. Most notably, a “squareness” tolerance, or the measured distance from corner to diagonal corner over the top surface of the panels, was added. The Texas pilot project revealed this to be a key tolerance to maintain classment of the panels. For future projects, additional tolerances should be specified, including tolerances for vertical classment (camber), horizontal skew, vertical batter, dowel location and classment, and expansion joint straightness and width.
Casting Bed—The “long line” casting bed, where the panels are cast end-to-end with the pretensioning strands extending through all of the panels, has proven to be an efficient and effective technique for panel fabrication. The process is much faster than match-casting, and no problems have been encountered to date with assembling non-match-cast panels. A key element of the fabrication process is casting panels with a flat bottom. Any variations in cross slope should be cast into the surface of the panel only. The flat bottom will better ensure full support when the panels are installed over a flat base rather than trying to match a grade break in the base with the panels.
Finishing—The finish of the top surface of the panel has a major impact on the smoothness of the finished pavement. The panels for this project were finished by hand due to the grout vents protruding from the surface of the panels. While the final result was still a very smooth and flat finish, the panels should ideally be screeded with a vibratory screed followed by minimal hand finishing. A vibratory screed will provide a more uniform and flat surface. Unfortunately, a vibratory screed cannot pass over grout vents or lifting anchor supports protruding from the surface. If possible, the grout vents should be capped just below the finished surface until screeding and carpet drag texturing have been applied. The vents can then be carefully uncovered by removing the thin layer of concrete above them.
An issue that was encountered during casting was slight sloughing of the fresh concrete at the edges of the panels. Because the sideforms varied in depth from 250 mm (10 in.) at each end to a “peak” of 330 mm (13.1 in.), the fresh concrete mixture wanted to flow “downhill” away from the peak, resulting in slight sloughing at the peak. Although this did not significantly affect the finished pavement surface, it could be minimized on future projects by using a “drier” concrete mix that will not flow as readily, and by vibrating the mixture as little as necessary.
Steam Curing—While the panels for the Texas pilot project received only two coats of curing compound after casting, the panels for this project were steam cured overnight. Steam curing is a preferred option for precast pavement panels because it is a “moist cure” technique that helps to ensure the concrete has sufficient moisture for hydration during the critical hours after casting. Although not all fabrication plants are set up for steam curing, any “moist cure” technique, such as fogging, wet mats, or steam curing is preferable to curing compound, particularly in hot or arid climates. Steam curing also promotes strength gain, permitting the use of non-high-early-strength cements such as Type II cement rather than Type III cement.
As discussed in chapter 6, pavement construction was very successful, with only minor problems encountered during panel installation and post-tensioning. Evaluation of the construction process and recommendations for future construction are presented below.
Base Preparation (Lean Concrete Base)
LCB was not an ideal material for supporting precast pavement panels. While it is a high-strength material, providing excellent structural support, it is a rigid material that will not conform to the bottom of the precast panels. Because it was difficult to finish LCB to strict tolerances, voids were present beneath the panels after they were installed. Although underslab grouting filled these voids, minimizing these voids before grouting is always preferred. If LCB is to be used for the leveling course, the use of a precision laser screed is recommended to achieve the smoothness tolerances described below. Under stringent time constraints, such as 6-to-10-hour construction windows, LCB may not be the best material since it requires time to cure before the precast panels can be installed over it.
Tolerances—A tolerance for the LCB was not specified in the original plans for the precast pavement. For future projects, however, it is recommended that the tolerances given in table 9 be specified for the leveling course.(2)
Alternative Materials—While hot-mix asphalt and LCB have proven to work for the leveling course material, neither are ideal materials for construction under very short construction windows, such as for overnight construction. Some alternatives should be considered for future projects:(2)
Precision-graded base and screeded grout bed have proven successful for other precast pavement projects.(20, 21)
Several modifications to the panel placement process from the Texas pilot project greatly improved the process. These modification and further recommendations are discussed below.
Joint Treatment/Epoxy—Segmental bridge epoxy proved to be beneficial not only for sealing the joints between panels, but also as a lubricant during assembly of the Texas pilot project.(2) Due to material supplier issues, however, a different epoxy was used for this project. The epoxy was much thinner in consistency, and therefore did not provide the same lubrication or sealing qualities. The epoxy was only effective in sealing the joints where the joint was tighter than 3 mm (1/8 in.), and was only applied to the top and bottom vertical faces of each keyway. This did demonstrate, however, that the panels could still be assembled without the epoxy as a lubricant. Although a thick consistency epoxy is recommended for future projects to both seal the joint and act as a lubricant, it is not absolutely essential for panel installation.
Temporary Post-Tensioning—Temporary post-tensioning consisted of two post-tensioning strands, each 15-mm (0.6 in.) in diameter, inserted into elongated access pockets in the joint panels and incrementally fed into each new precast panel as it was installed. After each panel was set as close as possible to its adjoining panel, the two strands were tensioned to pull the panels together as tightly as possible. This technique proved to be very efficient, increasing installation time only marginally, and very effective in ensuring the joints between panels were as tight as possible. Provisions for temporary post-tensioning should be incorporated into all future projects.
Panel/Duct classment—One issue encountered during the Texas pilot project was misclassment of post-tensioning ducts caused by using the ends of the panels as a reference point.(2) For this project, a chalk line was projected onto the surface of each panel exactly above the center of a given post-tensioning duct. This helped to ensure the post-tensioning ducts were consistently classed by using one of the ducts as a reference. This technique for panel classment is recommended for all future projects.
Ducts—Corrugated galvanized steel ducts, which are standard materials for Caltrans, were used for the longitudinal post-tensioning ducts. The primary benefit of using this material (vs. plastic duct) was that it remained straight in the forms during fabrication, eliminating the need for bar stiffeners. While these ducts normally are less susceptible to crushing, one incidence of a pinched duct was encountered when the strands were fed through the panels, as discussed in chapter 6. This material, or comparable rigid plastic duct, is recommended for use on future projects. A minimum inside diameter of 25 mm (1 in.) for 15-mm strand is recommended. Although the larger duct requires more grout, it provides a larger tolerance for panel misclassment.
Strand—Fine-grit-impregnated epoxy-coated strand was used for longitudinal post-tensioning. Although the grit increases friction and makes strand insertion slightly more difficult, the workers did not have any problems feeding the strand through 18 to 21 m (60 to 70 ft) of duct. The corrosion protection offered by the epoxy coating is a significant benefit over bare strand. As mentioned in chapter 6, however, a stressing ram with a long enough stroke to apply the full prestress force should be used to prevent multiple “bites” through the epoxy coating during post-tensioning.
Stressing Pockets—As discussed above, elimination of the central stressing pockets would greatly benefit the construction process by eliminating the need to patch the pockets before opening to traffic, although a fast-setting mixture could always be used for patching. For future projects, alternative stressing schemes should be examined to eliminate or reduce the stressing pockets.
Joint Seal/Gasket—The most significant improvement to the grouting process, based on recommendations from the Texas pilot project, was the use of a foam gasket around each of the post-tensioning ducts between each of the panels. The gaskets were fitted into a recess cast around each duct along the female keyway when the panels arrived on site. A foam gasket, 25 mm (1 in.) thick and compressible to at least 13 mm (1/2 in.), was used at each duct. While grout leakage was not completely eliminated, it was significantly reduced over the Texas pilot project. A foam or neoprene gasket is recommended for all future applications where the tendons are to be grouted.
Grout Vents—For applications where the precast pavement may need to be opened to traffic prior to tendon grouting, it will not be possible to leave grout vent tubing protruding from the surface of the pavement. Alternative grout vents or vent locations should be examined for future projects to eliminate this potential conflict. Eliminating grout vents protruding from the surface will also improve finishing and pavement smoothness. Grout vents were spaced approximately every 9 to 12 m (30 to 40 ft) along each tendon. Intermediate vents were important to the grouting operation, and therefore it is recommended that grout vents be spaced not more than 15 m (50 ft) apart for future projects.
The final cost for the California Demonstration Project was approximately $228,000. This cost includes all fabrication and installation costs, including traffic control. The total translates to a unit cost of approximately $268 per m² ($224 per yd2 ) of finished pavement. This unit cost is significantly higher than that of the Texas pilot project, which was approximately $194 per m² ($162 per yd2 ). However, the amount of pavement constructed for this project was only 853 m² (1,020 yd2 ), whereas the Texas pilot project was 7,690 m² (9,200 yd2 ). As with any project, there are economies of scale that decrease unit costs as quantity increases.
As might be expected, the largest portion of the cost (about 83 percent) was for panel fabrication and post-tensioning. Panel fabrication cost included significant capital investment for the panel forms. On future projects, it should be possible to reuse sideforms, significantly reducing the cost of fabrication. It should also be noted that this was an experimental project with many unknowns. As contractors and fabricators become more familiar with precast paving technology, unit costs will decrease. Also, it is anticipated that unit costs for competitively bid projects (rather than for change orders such as this project was) will be significantly less.