Skip to contentUnited States Department of Transportation - Federal Highway AdministrationSearch FHWAFeedback

Pavements

<< Previous ContentsNext >>

Development and Implementation of a Performance-Related Specification for SR 9a, Florida: Final Report

Chapter 4: Implementation of the Performance-Related Specification

Pre-Bid Construction Meeting

A mandatory pre-bid conference was held in Jacksonville on July 19, 2001, for the SR 9A interchange to US 1 (project 209600-1-52-01) and the I-295 / SR 9A / I-95 interchange (project 213290-1-52-01). Attendees included representatives of 41 companies. Information about the letting date (August 29, 2001), the end date (February 2005), and incentives for early completion was provided. During the concrete pavement discussions, Tim Ruelke, district materials engineer, discussed the section 400-15.2.5.4 smoothness evaluation of bridges greater than 300 ft (91.4 m) and requirement for full grinding of the concrete pavement surfaces. Next, he presented the plans for implementing PRS on portions of the project. Mike Darter of ARA, Inc., presented the key aspects of PRS; testing methods; target values for strength, thickness, and smoothness; pay factor curves; probabilities for pay increases; and several case studies. The possibility of extra testing on the PRS site using the FHWA mobile concrete laboratory was also discussed.

Pre-Construction Meeting

The SR 9A project was let on September 26, 2001, and awarded in October 2001 to AMEC Civil, LLC, who held a subcontract with McCarthy Improvement to complete the concrete paving. On June 12, 2003, a pre-construction meeting was held in the project field office in Jacksonville with representatives from FDOT, FHWA, AMEC (prime contractor), McCarthy Improvement (paving subcontractor), JEAces (construction inspection), TARMAC (cement supplier), PTGcsc (field inspection), the University of North Florida, and ERES Consultants. Nasir Gharaibeh of ARA, Inc. presented a summary of the PRS methods planned for the SR 9A PRS pavement. This included the project layout, sampling and testing plans, target, rejectable, and maximum pay values for strength, smoothness, and thickness, and pay factor computation methods. FDOT also distributed the approved Technical Special Provisions for PRS for Rigid Pavements at the SR 9A site.

No significant problems or concerns were noted with the special provisions. The paving contractor was not concerned with meeting the post-grinding smoothness specifications. Because the lot and sublot definitions in the special provision precluded PI testing of areas less than 0.05 mi (264 linear ft [80.5 m]) in length, and the sublots west of the US 1 bridge did not meet the 0.05-mi (80.5 m) threshold, it was determined that the PI of these sublots would not be tested for the purpose of determining pay. The State materials office was asked to run PI tests for informational purposes only.

It was noted that specifications for grinding methods were not included in the special provision. Plans to begin paving in September 2003 were discussed, followed by a field tour of the PRS site.

Construction

Construction of the base layer for the PRS project was completed in stages between December 2003 and June 2005. Paving of the PRS lots and sublots occurred between January 8, 2004, and April 29, 2004, for all but the sublot 1 sections of lots 1, 2, and 3. The inside lanes were paved together, and the widened outside lane was paved later. Grinding was also completed on the northbound sections in April 2004. The northbound lanes were opened to traffic on April 14, 2004, and southbound lanes were opened on July 17, 2005. Paving of sublot 1 of the southbound lanes was completed on June 6, 2005, with grinding conducted subsequently. Final lot and sublot layout, testing patterns, and paving operations are described below.

Layout of Lots and Sublots

An overview of the project site is shown in figure 11. Locations for the PRS lots and sublots remained as planned in the pre-construction meeting, according to the dimensions shown in figure 12. Each travel lane in the northbound and southbound directions was considered a lot, and the area on the east side of the bridge was divided into two approximately equal sublots. Paving lanes west of the bridge were also considered as single sublots. Note that on longer projects, all lanes included in the paving width are normally considered to be in the lot, reducing testing requirements.

Figure 11. Florida SR 9A performance-related specification project overview.

Figure 11. Photograph. Florida SR 9A performance-related specification project overview. Aerial view of the project highway, its north- and southbound bridge spans and cloverleafs.

Figure 12. SR 9A lot and sublot layout for the performance-related specification project.

Figure 12. Diagram. SR 9A lot and sublot layout for the performance-related specification project. This drawing shows the project layout with lots 1-3 (three southbound lanes) and lots 4-6 (three northbound lanes) (the fourth lane in each direction for speed change is not included), each lot comprising one bridge lane between Sta 127+00 (MP 24916), beginning of the PRS PCC project, and Sta 149+17.52 (MP 24.496), end of the PRS PCC project. The two spans of the bridges are excluded from the lots. Lot 1 lane is 14 ft wide; lots 2 and 3 lanes are 12 ft wide. Lots 4 and 5 lanes are 12 ft wide; lot 6 is 14 ft wide. The lengths of sublots 1, north of the bridge, vary between 170.93 ft (lot 1) and 248.17 ft (lot 6) in length, corresponding to the diagonal where the bridge spans and their approaches (~ 20 ft) meet the highway pavement. Sublots 2, south of the bridge, vary between 515.45 ft and 443.30 ft in length. Sublots 3 are all 528.00 ft. in length.
1 ft = 0.305 m.

Paving Operations

Due to differing phasing demands, the northbound PRS sections were completed prior to the southbound lanes. Construction of the project in both directions included preparation of the embankment and base layers; placement of stringlines and dowel baskets; and concrete paving, finishing, curing, and surface grinding.

Embankment and Base Preparation

In the northbound lanes (sublot 1 of lots 4 and 5), the base surface grading was initially completed using a Topcon global positioning system (GPS) on the grader with two control stations. The accuracy of this system proved inadequate, and a third control station was used for final grading. No other problems were reported in the base surface preparation.

Stringline Placement

Nylon stringlines were installed using supports spaced at about 15 ft (4.8 m). The stringline supports were positioned and adjusted using a GPS receiver. Then the contractor used a stringline between the longitudinal stringlines to check the base for proper grade. Additional grading was required in sublot 1 of lot 5 prior to paving. The contractor placed stringline at the bridge to guide the paving train as it drove up onto the bridge surface.

Dowel Basket Placement

The contractor installed dowel baskets at the contraction joints and dowel baskets with expansion joint material on the sleeper slab near the bridge. These baskets were staked into the base material using steel stakes approximately 12 in. (300 mm) long. Wooden stakes were placed on both sides of the paving lane near the center of each dowel basket to assist in later sawing operations.

Paving

PCC paving was accomplished using a spreader with a side loader for the PCC mix and a slip-form paving machine, as shown in figure 13. The two inside lanes were paved in one pass, and the outside lanes were paved in an additional pass. The side-dump spreader had a wheel in the center to insert 0.5-in. (12.7 mm) tie bars in the longitudinal joint. Dowel basket areas were skipped during this insertion. On the side of the paving machine was a device that pushed bent tie bars into the edge of the pavement. A double layer of burlap was used behind the paver for initial texturing. No problems were reported with the nonagitating supply truck delays, possibly because of the short hauling distance.

Figure 13. Florida performance-related specification lots 4 and 5, sublot 1, paving.

Figure 13. Photograph. Florida performance-related specification lots 4 and 5, sublot 1, paving. A spreader with a side loader for the PCC mix and a slip-form paving machine are shown.

Finishing and Curing

After about 30 minutes, the contractor used a separate device to apply a final burlap drag texture to the fresh concrete surface. That device also included a spray distribution system for evenly applying the curing compound.

Grinding

Grinding in the northbound lanes was completed by Diamond Surfaces using a Caterpillar 10-25 grinder. Southbound lanes were diamond ground by Central Atlantic Contracting using a Caterpillar 10-18 grinder. The specifications allowed for up to 30 percent unground dips, but the contractor left less than 3 percent of the surface unground.

Sampling and Testing

The general sampling and testing plan for the PRS lots is shown in figures 14 and 15. Two randomly selected 6-in. (150 mm) cylinders were filled from each sublot prior to placement for subsequent strength quality assurance testing. The engineer allowed sample collection to be done at the batch plant following mixing, because of the close proximity (approximately 4 minutes) of the plant to the construction site. Following grinding, profilograph measurements were collected in each wheel path of each lane and approved with no subsequent grinding. Two core samples, 6 in. (150 mm) in diameter, of the P-501 surface were then collected from random locations in each sublot, 4 and 8 ft (1.2 and 2.4 m) from the adjacent joint. Results of this sampling and testing for strength, smoothness, and thickness are shown in tables 13, 14, and 15.

Figure 14. Southbound SR 9A sampling locations.

Figure 14. Diagram. Southbound SR 9A sampling locations. Sublots 1 in lots 1, 2, and 3 range from 170.93 ft to 184.79 ft in length; sublots 2 in lots 1, 2, and 3 range from 503.99 ft to 515.45 ft in length; sublots 3 in lots 1, 2, and 3 are 528 ft long. Each sublot contains a profile trace in its two wheelpaths; two thickness core samples, one replicate per sample; and one sample strength cylinder, with two replicates per sample. The thickness and strength sampling sites are at varying locations within the sublots. (The "left bridge" span is shown between sublots 1 and 2.)

Figure 15. Northbound SR 9A sampling locations.

Figure 15. Diagram. Northbound SR 9A sampling locations. Sublots 4 range from 234.26 ft to 248.17 ft in length; sublots 5 from 443.30 ft to 454.82 ft in length; sublots 6 are 528 ft long. Each sublot contains a profile trace in its two wheelpaths; two thickness core samples, one replicate per sample; and one sample strength cylinder, with two replicates per sample. The thickness and strength sampling sites are at varying locations within the sublots. (The "right bridge" span is between sublots 1 and 2.)

Average compressive strength results from all sublots except sublot 1 of lot 3 exceeded the target average 28-day compressive strength of 4,500 lbf/in2 (31.0 MPa), as noted in table 13. Sublots 1 of lots 1 and 3 and sublot 2 of lot 5 were significantly lower than the averages of the other sublots. This is unusual because sublots 1 of lots 2 and 3 were reportedly paved simultaneously. No explanation was found for the lower values.

Average concrete pavement core thickness measurements, shown in table 14, were all at or above the 12.5-in. (320 mm) target average. Individual core thicknesses ranged from 12.0 to 14.6 in. (304.8 to 370.8 mm), with an average thickness of 13.3 in. (337.8 mm).

Smoothness measurement results, shown in table 15, indicate that the grinding contractor did an excellent job of achieving good surface smoothness. Sublot PI values ranged from 0 to 5.2 in./mi (0 to 81 mm/km). Sublots 1 of lots 1, 2, 3, and 4 were each less than 250 ft (76.2 m) long and were combined with respective sublots 2 for smoothness computations, as required by the PRS specification.

Table 13. SR 9A Performance-Related Specification Compressive Strength Testing Results
Lot Sublot Sample No. Placement Date Compressive Strength, lbf/in2
Cylinder #1 Cylinder #2 Average
1 1 A1063 4/5/04 4,552 4,974 4,763
1 2 A1022 2/6/04 6,580 6,366 6,473
1 3 A1021 2/6/04 5,979 5,868 5,924
2 1 A1059 3/24/04 5,272 5,250 5,261
2 2 A1009 1/13/04 5,547 5,829 5,688
2 3 A1009 1/12/04 6,015 5,989 6,002
3 1 A1062 4/3/04 4,487 4,409 4,448
3 2 A1010 1/13/04 5,536 5,482 5,509
3 3 A1007 1/12/04 5,750 5,826 5,788
4 1 A1002 1/8/04 5,960 5,948 5,954
4 2 A1003 1/9/04 5,964 5,818 5,891
4 3 A1006 1/9/04 5,477 5,573 5,525
5 1 A1001 1/8/04 5,240 5,296 5,268
5 2 A1004 1/9/04 4,483 4,745 4,614
5 3 A1005 1/9/04 5,770 5,703 5,737
6 1 A1023 2/7/04 6,356 6,263 6,310
6 2 A1024 2/10/04 6,217 6,175 6,196
6 3 A1025 2/10/04 6,247 6,198 6,223
Note: Target average = 4,500 lbf/in2 (31.0 MPa); target standard deviation = 610 lbf/in2 (4,206 kPa); rejectable quality level = 2,700 lbf/in2 (18.62 MPa); acceptance quality level = 5,500 lbf/in2 (37.92 MPa); 1 lbf/in2 = 6.89 kPa.
Table 14. SR 9A Performance-Related Specification Thickness Testing Results
Lot Sublot Core Date Station, ft Thickness, in.
Core 1 Core 2 Core 1 Core 2 Average
1 1 3/30/04 127+40 L-3 128+61 L-3 13.8 13.0 13.4
1 2 3/30/04 142+00 L-3 142+56 L-3 12.9 12.3 12.6
1 3 3/30/04 145+11 L-3 148+22 L-3 12.6 12.6 12.6
2 1 3/30/04 128+01 L-2 128+05 L-2 13.5 14.5 14.0
2 2 3/30/04 139+41 L-2 143+80 L-2 13.3 13.5 13.4
2 3 3/30/04 144+12 L-2 149+00 L-2 13.1 14.3 13.7
3 1 3/30/04 127+25 L-1 129+00 L-1 13.7 12.7 13.2
3 2 3/30/04 139+25 L-1 143+80 L-1 13.1 13.2 13.2
3 3 3/30/04 145+00 L-1 148+85 L-1 13.3 14.6 14.0
4 1 3/30/04 128+00 R-1 128+99 R-1 12.9 12.0 12.5
4 2 3/30/04 141+21 R-1 142+00 R-1 13.2 13.1 13.2
4 3 3/30/04 145+11 R-1 147+00 R-1 12.5 13.4 13.0
5 1 3/30/04 128+89 R-2 129+49 R-2 13.2 13.2 13.2
5 2 3/30/04 141+75 R-2 143+80 R-2 13.5 13.7 13.6
5 3 3/30/04 144+13 R-2 148+91 R-2 13.4 13.5 13.5
6 1 3/30/04 127+13 R-3 129+54 R-3 12.8 13.8 13.3
6 2 3/30/04 141+60 R-3 143+82 R-3 13.7 13.0 13.4
6 3 3/30/04 145+22 R-3 147+36 R-3 13.3 13.0 13.2
Note: Target average = 12.5 in.; target standard deviation = 0.5 in.; rejectable quality level = 12 in.; acceptance quality level = 13.5 in.
1 ft = 0.305 m; 1 in. = 25.4 mm
Table 15. SR 9A Performance-Related Specification Smoothness Testing Results After Diamond Grinding
Lot Sublot Length, ft Sample date Profile Index, in./mi, Wheelpaths
Right Left Average
1 1, 2 695 5/9/05 0.99 0.00 0.50
1 3 528 5/9/05 5.20 2.30 3.75
2 1, 2 692 5/9/05 0.00 0.92 0.46
2 3 528 5/9/05 2.30 0.50 1.40
3 1, 2 690 5/9/05 1.38 1.15 1.25
3 3 528 5/9/05 0.60 0.00 0.30
4 1, 2 770 3/25/04 1.44 0.34 0.89
4 3 483 3/25/04 0.00 0.00 0.00
5 1 250 3/25/04 4.44 0.00 2.22
5 2 528 3/25/04 0.00 0.00 0.00
5 3 478 3/25/04 1.00 0.00 0.50
6 1 258 3/25/04 0.00 0.00 0.00
6 2 528 3/25/04 1.40 1.10 1.25
6 3 473 3/25/04 0.00 0.00 0.00
Note: Target average = 3.0 in./mi; target standard deviation = 1.0 in./mi; ); rejectable quality level = 7.0 in./mi; acceptance quality level = 0 in./mi.
1 ft = 0.305 m; 1 in./mi = 16 mm/km.

Average values for each field performance factor were computed for each as-constructed lot using the following method:

Equation 3, p. 31. Equation. Mean of n random samples of the AQCs under consideration for the lot. The mean of n random sample measurements is equal to the summation of the sample measurements divided by the number of sample measurements. For smoothness and strength, sample measurement is a mean of multiple replicates, and for thickness, the mean of all cores). The sample size per lot is as follows: Strength, 1 sample per sublot (each sample is a mean of two cylinder measurements, or core measurements for partial lots); Thickness, 2 samples (cores) per sublot; Smoothness, the number of required profile sections per lot (each is represented by the mean profile passes in the wheelpaths).      Eq. 3

where:

  • x bar = Mean of n random samples of the AQC under consideration for the lot.
  • Xi = Sample measurement (for smoothness and strength, Xi is a mean of multiple replicates, and for thickness the mean of all cores).
  • n = Sample size per lot, n for each AQC is as follows:
    • Strength: 1 sample per sublot (each is a mean of two cylinder measurements, or core measurements for partial lots).
    • Thickness: 2 samples (cores) per sublot.
    • Smoothness: number of required profile sections per lot (each is represented by the mean profile passes in the wheel paths).

Thickness, strength, and smoothness lot standard deviation was computed as follows:

Equation 4, p. 32. Lot standard deviation. Lot standard deviation equals the square root of a quantity having, as the numerator, the summation of the squared differences between each sample measurement and the mean of the sample measurements and, as the denominator, the number of sample measurements minus one. To correct for the bias inherent when estimating actual lot standard deviation, the authors divide by a correction factor (CSD, obtained from table 16) that is dependent on the number of sample measurements and acts to increase the magnitude of the lot standard deviation.      Eq. 4

where:

CSD = Correction factor (based on the total sample size, n) used to obtain unbiased estimate of the actual lot sample standard deviation. Appropriate CSD values are determined using table 16.

Table 16. Correction Factors Used to Obtain Unbiased Estimates
of the Lot Standard Deviation
No. of Samples, n Correction Factor, CSD
2 0.7979
3 0.8862
4 0.9213
5 0.9399
6 0.9515
7 0.9594
8 0.9650
9 0.9693
10 0.9726
30 0.9915

Results of the computations produced the summary of critical performance factors shown in table 17. These indicate that the lot compressive strength target AQL of 4,500 lbf/in2 (31.03 MPa) was exceeded by all lots. The maximum lot standard deviation value was 372 lbf/in2 (2,505 kPa), well below the target value of 610 lbf/in2 (4,206 MPa). The range in standard deviations from 47 to 697 lbf/in2 (0.32 to 4,806 MPa) indicates that the contractor is capable of significantly reducing the variability in strength properties.

All thickness lot averages exceeded the target value, with lot 2 exceeding the MQL of 13.5 in. (342.9 mm). The average lot thickness of 13.26 in. (336.8 mm) was well above the target level of 12.5 in. (317.5 mm). Variability (using standard deviation as a standard) within each lot ranged from 0.20 to 0.69 in. (5.1 to 17.5 mm), with the average lot standard deviation being exactly the target value of 0.5 in. (12.7 mm).

The average lot smoothness for the PRS sections was 0.96 in./mi (15 mm/km), well below the target of 3.00 in./mi (48 mm/mi). Variability in smoothness levels was lower than the target 1.00 in./mi (16 mm/km), achieving an average lot standard deviation of 0.86 in./mi (14 mm/km).

Table 17. Peformance-Related Specification Lot Quality Results
Lot No. Strength, lbf/in2 Thickness, in. PI0.2-in, in./mi
Mean Std Dev Mean Std Dev Mean Std Dev
1 5,720 697 12.87 0.55 2.12 2.04
2 5,650 297 13.70 0.59 0.93 0.59
3 5,248 564 13.43 0.69 0.78 0.60
4 5,790 175 12.85 0.54 0.59 0.41
5 5,206 450 13.42 0.20 0.91 0.93
6 6,242 47 13.27 0.43 0.42 0.58
Average 5,642 372 13.26 0.50 0.96 0.86
1 lbf/in2 = 6.89 kPa; 1 in. = 25.4 mm; 1 in./mi = 16 mm/km.

Computation of Pay Factors

Using the average lot performance information from table 17, pay factors for each lot were determined using the spreadsheet shown in figure 16 following the methods provided in the supplemental specification in appendix B. This method required that the lot composite (overall) pay factor be computed as follows:

PFcomposite = (PFsmoothness * PFstrength * PFthickness) / 10,000      Eq. 5

where:

  • PFcomposite = Composite (overall) pay factor, percent (with maximum set at 110 percent).
  • PFstrength = Strength pay factor (obtain from table 10), percent.
  • PFthickness = Slab thickness pay factor (obtain from table 11), percent.
  • PFsmoothness = Initial smoothness pay factor (obtain from table 12), percent.

The actual pay adjustment for the as-constructed lot was computed using the lot composite pay factor as follows:

PAYADJLot = BID * AREALot * (PFcomposite - 100) / 100      Eq. 6

where:

  • PAYADJLot = Pay increase (+) or decrease (-), $.
  • BID = Contractor bid price for pay item 350-1-19, $/yd2.
  • AREALot = Measured actual area of the as-constructed lot, yd2.
  • PFcomposite = Composite pay factor (from equation 5), percent (e.g., 101 percent is expressed as 101.0).

PAYLot = BID * AREALot + PAYADJLot      Eq. 7

where:

  • PAYLot = Adjusted payment for the as-constructed lot, $.

Computed pay factors for strength, thickness, and smoothness are shown in table 18. Strength pay factors were exceptionally high, indicating the ease with which the contractor was able to achieve strengths greater than the AQL. This is not surprising because the average of the three projects evaluated in PRS preparation was 5,500 lbf/in2 (37.92 MPa), much higher than the 4,500  lbf/in2 (31.03 MPa) target. Thickness and smoothness pay factors were also higher than those associated with the target levels, providing incentives of about 3.5 and 2.5 percent, respectively. When the strength, thickness, and smoothness pay factors are multiplied together to determine the overall pay factor, the average incentive level is nearly 15 percent. However, the specification places an upper limit on incentives of 10 percent, so the overall project pay factor is held to 110 percent. These results indicate that all AQCs were achieved or exceeded in construction. Therefore, better than expected pavement performance should be anticipated.

Table 18. SR 9A Performance-Related Specification Computed and Maximum Pay Factors (PF)
Lot Strength PF Thickness PF Smoothness PF Overall Project Unlimited PF Overall Project Limited PF
1 107.88 102.29 100.81 111.24 110.00
2 108.78 104.17 102.61 116.28 110.00
3 107.21 104 102.78 114.6 110.00
4 108.98 102.21 103.15 114.97 110.00
5 107.32 104.13 102.54 114.59 110.00
6 109.13 103.9 103.24 117.06 110.00
Avg. 108.22 103.45 102.52 114.79 110.00

Figure 16. Example pay factor computation worksheet.

Figure 16. Illustration. Example pay factor computation worksheet. Completed form contains four sections: Lot Information, Thickness, Strength, Smoothness. The Lot Information section includes lot number, bid price ($/sq yd), lot length (ft), lot width (ft), resulting lot area (sq yd), project no., begin station, end station, lane no., and paving dates. The Thickness section, which covers a full lot, includes for sublots 1, 2, and 3 thickness core 1, thickness core 2, and sublot thickness mean (in.). It also includes the number of resulting samples per lot and the mean lot thickness (in.), lot thickness mean acceptable (yes/no), notes on lot thickness mean, standard deviation correction factor, lot thickness standard deviation (in.), and the resulting pay factor. The Strength section, which covers the same full lot, includes for sublots 1, 2, and 3, strength cylinder 1, strength cylinder 2, and sublot strength (psi); resulting samples per lot; lot mean strength (psi); lot strength mean acceptable (yes/no); notes on lot strength mean; notes on sublot strength mean; standard deviation correction factor; lot strength standard deviation (psi); and resulting pay factor. The Smoothness section for the full lot includes for sublots 1, 2, and 3, the profile index (PI) for pass 1 and pass 2 and the sublot mean PI (in./mi). It also includes the resulting samples per lot (n), lot smoothness mean (in./mi) and lot smoothness mean acceptable (yes/no), notes on lot smoothness mean, notes on sublot smoothness mean, standard deviation correction factor, lot smoothness standard deviation (in./mi).
<< Previous ContentsNext >>
Updated: 04/07/2011

FHWA
United States Department of Transportation - Federal Highway Administration