Incremental Costs and Performance Benefits of Various Features of Concrete Pavements

Previous | Table of Contents | Next

CHAPTER 2. DATA COLLECTION ACTIVITIES

Introduction

A significant data collection effort was required to obtain the information needed to fulfill the overall project objectives. This data collection effort consisted of:

• Literature review. A detailed literature review was conducted to identify the available information on costs and benefits of various PCC pavement design procedures, and to provide a foundation for the pavement design categories and design features to be evaluated through the questionnaire surveys.
  
  
• Questionnaire surveys of State Highway Agencies (SHAs). To provide information on the relative performance benefits of various PCC pavement design features, questionnaire surveys were submitted to targeted SHAs. These surveys solicited information on what performance benefits could be expected when adding selected PCC pavement design features to a Standard pavement design.
  
  
• Questionnaire surveys of PCC paving contractors. Questionnaire surveys were also sent to PCC paving contractors to obtain information on the costs associated with various PCC design features. These surveys solicited information on what the relative increase in cost might be for adding PCC pavement design features to a Standard pavement design.

The development, distribution, and collection of the questionnaire surveys as well as the overall processing of the data represented a major work effort. The processed and summarized performance and cost data serve as the "default" database (i.e., default cost and performance data sets) for use in evaluating the relative cost and performance benefits of each PCC pavement design feature.

This chapter describes each of the primary data collection activities mentioned above. The overall approach and methodology used in each activity are described, along with a summary of how the collected data were used in this project.

Literature Review

The detailed literature review for this project aimed to identify pertinent reference documents discussing the costs and performance benefits of different PCC pavement design features. A previous literature search conducted for a National Highway Institute (NHI) training course was the basis for the search conducted for this project.⁽⁵⁾ However, the previous search by Smith and Hall targeted pavement design and performance information for the three major PCC pavement types (jointed plain concrete pavements (JPCP), jointed reinforced concrete pavements (JRCP), and continuously reinforced concrete pavements (CRCP)), so portions were not necessarily applicable to the current study which focuses primarily on JPCP designs and on the costs and performance benefits of PCC design features. Moreover, additional information was sought on pavement cross sections, PCC strength, PCC materials, and ride specifications, as these saw limited coverage in the original literature review.

To supplement the applicable documents identified in the original literature review, additional targeted searches were completed using the Transportation Research Information Services (TRIS), National Technical Information Services (NTIS), and Engineering Index (EI Compendex) national bibliographic databases. These searches were limited to recent publications (TRIS: 1995 to 2000; NTIS: 1990 to 2000; EI Compendex: 1998 to 2000). The search identified more than 100 additional records that were deemed potentially useful to the project.

After reviewing the records identified in the new literature searches, the researchers prepared a short annotation describing the record's content. A final annotated bibliography was prepared and grouped by the following topics:

• Base/Subbase Design.
• Concrete Mix Considerations.
• Construction.
• Costs.
• Cross Section.
• Design and Performance.
• Drainage.
• Joint Design.
• Maintenance.
• Reinforcement.
• Ride Specifications.
• Shoulder Design and Performance.
• Subgrade.
• Surface Texture and Noise.

The final annotated bibliography (appendix A) is not intended to be a comprehensive coverage of every PCC pavement design and performance study; instead, it is a listing of pertinent recent reports and papers (those published within the past 10 years) that provide guidance on the costs and performance of PCC pavement design features. The key documents were reviewed and used as a foundation for the pavement design categories and design features to be evaluated in the study, and also provided insight into general cost and performance trends.

Questionnaire Survey Development

This section describes the data collection effort for the performance and cost surveys conducted for this project. First, general background information is provided on the PCC pavement design features included in the surveys and on the construction of the surveys themselves. This is followed by a description of the data collection processes utilized in both the SHA (relative performance) surveys and the PCC paving contractor (relative cost) surveys.

Design Categories and Alternative Design Features

Based on the literature review, 10 primary categories of PCC pavement design features were identified, along with a list of possible design feature alternatives within each category. Because pavement designs can vary considerably within each of the design feature categories, one feature was selected to be the "standard" feature within that category. The collection of "standard" features from all design feature categories then represents the Standard PCC pavement cross section. The Standard PCC pavement cross section is the basis for comparison and allows for the determination of the incremental increase (or decrease) in cost and performance relative to that standard design (instead of, for example, each agency's standard design).

The final design feature categories and alternative design features considered within each category are summarized below. Those design features included as part of the Standard pavement design are indicated by "STD." A two-lane roadway (two lanes in one direction and part of a four-lane divided highway) is assumed as the Standard construction section.

• Subgrade
  -Prepared subgrade (natural soil, no other treatment) (STD).
  -300-mm (12-inch) lime-treated subgrade (and elimination of 150-mm (6-inch) dense-graded aggregate base).
  
  
• Base/Subbase
  -150-mm (6-inch) dense-graded aggregate base on prepared subgrade (STD).
  -No base (PCC pavement placed directly on prepared subgrade).
  -150-mm (6-inch) asphalt-treated base (ATB).
  -150-mm (6-inch) cement-treated base (CTB).
  
  
• Drainage
  -No positive drainage system (STD).
  -Open-graded nonstabilized aggregate layer with longitudinal underdrain system. (Note: A longitudinal drainage system consists of an open-graded trench wrapped with a geotextile material and containing 150-mm (6-inch) flexible longitudinal drainage pipe and rigid pipe outlets at 152-m (500-ft) spacings).
  -Asphalt-treated permeable base (ATPB) with longitudinal underdrain system.
  -ATPB with daylighted drainage into drainage ditch or swale (no longitudinal underdrain system).
  -Cement-treated permeable base (CTPB) with longitudinal underdrain system.
  -CTPB with daylighted drainage into drainage ditch or swale (no longitudinal underdrain system).
  
  
• Thickness/Slab Size
  -250-mm (10-inch) PCC pavement with 4.6-m (15-ft) transverse joint spacing (STD).
  -200-mm (8-inch) PCC pavement with 3.7-m (12-ft) transverse joint spacing.
  -300-mm (12-inch) PCC pavement with 5.5-m (18-ft) transverse joint spacing.
  -250-mm (10-inch) JRCP with 150 x 300 mm-MW 80 x MW 35 (6 x 12 inch-W 12 x W 5.5) welded wire fabric mesh placed on steel chairs at midslab depth. Transverse joint spacing is 9.1 m (30 ft) with 32-mm (1.25-inch) diameter epoxy-coated dowel bars that are 450-mm (18-inches) long
  and spaced at 300 mm (12 inches) across the transverse joints.
  -240-mm (9.5-inch) CRCP with 19-mm (0.75-inch) epoxy-coated deformed bars, 200 mm (8 inches) on-center longitudinally and 900 mm (36 inches) on-center transversely.
  -240-mm (9.5-inch) CRCP with 19-mm (0.75-inch) noncoated deformed bars, 200 mm (8 inches) on-center longitudinally and 900 mm (36 inches) on-center transversely.
  
  
• Pavement Cross Section
  -250-mm (10-inch) thick uniform cross section (STD).
  -Trapezoidal cross section, with the PCC pavement 275 mm (11 inches) thick at the outside edge of the truck lane, tapering to 200 mm (8 inches) thick at the outside edge of the passing lane.
  -Thickened edge design, with the outer edges of both the truck and the passing lane constructed 275 mm (11 inches) thick and tapering to a thickness of 200 mm (8 inches) at the centerline of the pavement.
  
  
• Joints/Load Transfer
  -250-mm (10-inch) JPCP, 4.6-m (15-ft) perpendicular transverse joint spacing, 32-mm (1.25-inch) diameter epoxy-coated dowel bars that are 450-mm (18-inches) long and spaced at 300 mm (12 inches) across the transverse joints (STD).
  -250-mm (10-inch) JPCP, 4.6-m (15-ft) perpendicular transverse joint spacing, 32-mm (1.25-inch) diameter noncoated dowel bars that are 450-mm (18-inches) long and spaced at 300 mm (12 inches) across the transverse joints (no epoxy on the dowel bars).
  -250-mm (10-inch) JPCP, 4.6-m (15-ft) skewed transverse joint spacing, 32-mm (1.25-inch) diameter epoxy-coated dowel bars that are 450-mm (18-inches) long and spaced at 300 mm (12 inches) across the transverse joints. The joints are skewed counter-clockwise at 0.6 m (2 ft) per 3.7-m (12-ft) lane.
  - 250-mm (10-inch) JPCP, 4.6-m (15-ft) perpendicular transverse joint spacing, and no dowel bars.
  -250-mm (10-inch) JPCP, 4.6-m (15-ft) perpendicular transverse joint spacing, with three 32-mm (1.25-inch) diameter epoxy-coated dowel bars that are 450 mm (18 inches) 450-mm (18-in) long located in the center of each wheelpath on 300-mm (12-inch) spacings.
  
  
• Joint Sealing
  - 4.6-m (15-ft) joint spacing with 3-mm (0.12-inch) control cut, 9.5-mm (0.38-inch) widening cut, backer rod, and hot-poured asphaltic sealant (STD).
  - 4.6-m (15-ft) joint spacing with hot-poured sealant and a single 3-mm (0.12-inch) saw cut for crack control.
  -4.6-m (15-ft) joint spacing with 3-mm (0.12-inch) control cut, 9.5-mm (0.38-inch) widening cut, backer rod, and silicone sealant.
  - 4.6-m (15-ft) joint spacing with silicone sealant and a single 3-mm (0.12-inch) saw cut for crack control.
  - 4.6-m (15-ft) joint spacing with 3-mm (0.12-inch) control cut, 9.5-mm (0.38-inch) widening cut, and preformed compression sealant.
  - 4.6-m (15-ft) joint spacing with no sealant and a single 3-mm (0.12-inch) saw cut for crack control.
  
  
• Shoulders
  - 150-mm (6-inch) hot-mix asphalt (HMA) shoulders over 250-mm (10-inch) dense graded aggregate base (STD).
  - 400-mm (16-inch) gravel shoulders.
  - 150-mm (6-inch) partial-depth, tied PCC shoulder over a 250-mm (10-inch) aggregate base (outer lane only, standard HMA shoulder on passing lane).
  -250-mm (10-inch) full-depth, tied PCC shoulder over a 150-mm (6-inch) aggregate base (outer lane only, standard HMA shoulder on passing lane).
  -Standard pavement section with a 0.6-m (2-ft) widened PCC slab and a 2.4-m (8-ft) HMA shoulder.
  
  
• Concrete Strength/Materials
  - 4.5-MPa (650 lb/in²) flexural strength (28-day value tested under third-point loading, American Society for Testing and Materials (ASTM) C78) (STD).
  - 5.2-MPa (750 lb/in²) flexural strength (28-day value tested under third-point loading (ASTM C78)). Higher strength is obtained by using additional Type I cement.
  - High-early strength, opening to traffic within 48 hours (high-early strength is obtained by using Type III cement).
  - PCC mix using a well-graded aggregate gradation.
  
  
• Initial Smoothness (as measured by a California Profilograph with a 5-mm (0.20-inch) blanking band)
  - 110 to 142 mm/km (7 to 9 inches/mi) (STD).
  - 79 to 110 mm/km (5 to 7 inches/mi).
  - 47 to 79 mm/km (3 to 5 inches/mi).
  - 16 to 47 mm/km (1 to 3 inches/mi).
  - Less than 16 mm/km (1 inches/mi).

Clearly, a virtually unlimited number of design feature alternatives could have been selected for this project. However, the number of alternatives was limited to not only facilitate the questionnaire survey process but also to represent more established design practices (although a few unique design features were included to reflect new or innovative practices).

Questionnaire Survey Forms

After the design feature categories were identified and the various design feature alternatives selected, questionnaire surveys were developed: one targeted at highway agencies (to solicit relative performance data) and one targeted at PCC paving contractors (to solicit relative cost data). Although these were separate surveys, the pavement design variables presented in each questionnaire are identical. This allows the results from each data collection effort to be directly paired for analysis. Additional data (including a short summary of the agency's PCC pavement maintenance activities and the ranking of the design categories in terms of their impact on pavement performance) were also a part of the relative performance questionnaire surveys.

Both surveys were structured so that only one design feature (from the Standard design) was changed at a time, and the survey participants were then asked to assess what effect that change might have in terms of the relative performance (agency questionnaire) or costs (contractor questionnaire). In this way, the relative effects of the change in that one design feature could be determined. For example, one scenario is to change the base from the 150-mm (6-inch) aggregate base in the Standard design to a 150-mm (6-inch) CTB. Considering this single change, the highway agencies were asked to assess what effect this would have on the relative performance of the modified pavement design in comparison to the Standard design. Assigning the relative performance of the Standard design as 1.0, if the modified pavement (with the CTB) is believed to be capable of carrying 5 percent more 80-kN (18-kip) equivalent single-axle load (ESAL) applications than the Standard pavement, the relative performance rating is 1.05. On the other hand, if the design feature change is believed to result in a 5-percent decrease in the number of 80-kN (18-kip) ESAL applications that the pavement can carry, the relative performance rating is 0.95. Again, all performance ratings are made relative to the Standard design performance rating of 1.0.

A similar approach is employed for the cost surveys. Using the same example (changing the aggregate base of the Standard section to a CTB), PCC paving contractors were asked to estimate what the change in relative cost of the modified pavement design might be, assuming a relative cost of 1.0 for the Standard section. If the cost for the modified pavement (containing the CTB) is believed to be 15 percent more than the Standard section, the relative cost rating is then 1.15.

Survey respondents provided these relative ratings for the entire group of design feature alternative changes listed above. Respondents were asked to not enter a rating if they had no experience with a particular design feature.

As part of the survey development, several design and construction assumptions were established to provide a common foundation for all respondents and to help to maintain overall consistency in the responses. These assumptions include the following:

• The pavement is assumed to be located on a rural divided highway.
• The average daily traffic (ADT) is assumed to be 20,000 vehicles per day in each direction with 15 percent trucks. This represents approximately 700,000 to 800,000 ESALs per year in the design lane (no growth in annual ESALs is assumed during the life of the pavement).
• The project is located within 80 km (50 mi) of the contractor's mobilization center.
• The project is 8 km (5 mi) long and has four lanes of traffic (two lanes in each direction).
• The project is constructed at grade with no earthwork and all paving is on longitudinal slopes less than 5 percent.
• The pavement is constructed during normal daytime business hours.
• Standard slipform paving equipment will be used for construction.
• A portable PCC batch plant will be utilized.
• The PCC for this project is a typical local mix.
• Local environmental regulations will govern the project.
• Local specifications for construction apply to the highway work.
• Local load restrictions will be used.

Appendix B contains the final questionnaires used in conducting both surveys.

Desired Number of Survey Responses

Prior to sending out the questionnaire surveys, the researchers performed an initial statistical evaluation to determine the desired number of questionnaire responses required to have a reasonable estimate of the relative performance and cost. In determining the desired sample size, it is assumed that the total population has a normal distribution. The purpose of the questionnaires is to predict the average of the population (for example, the average PCC pavement life or the average PCC pavement relative construction cost). Thus the following equation is applicable to estimate the desired sample size:

(1)

where:

n = desired number of survey samples.

s = estimated standard deviation.

z = number of standard error units (based on the desired confidence level and obtained from a normal probability table).

T = required precision or tolerance.

With this approach, the standard deviation of the to-be-determined average value is yet unknown. Furthermore, the desired confidence levels and the desired precision levels must also be selected. However, by running a range of values with an initially assumed, reasonable average, the effects of these inputs on the resulting number of samples can be determined and reasonable target values can be selected.

For the purposes of estimating the number of samples, the analyses for the performance and cost questionnaires are broken out separately, as described below.

• The relative performance questionnaires essentially ask, "What is the performance life of PCC pavements?" Assuming the average result is approximately 40 years, standard deviations of 4, 6, and 10 years (roughly equivalent to coefficients of variation of 10, 15, and 25 percent) were selected as estimates of the true standard deviation. Next, three levels of desired precision (the desired closeness of the life estimates) were selected for consideration: 2, 5, and 10 years. Based on these assumptions and using equation 1, table 1 was generated.
• The relative cost questionnaires consider the basic question, "What is the cost of constructing a PCC pavement?" Assuming the average result is approximately $35.90/m² ($30/yd²), standard deviations of $3.60, $7.20, and $10.80/m² ($3.00, $6.00, and $9.00/yd²) (roughly equivalent to coefficients of variations of 10, 15, and 20 percent) were selected as estimates of the true standard deviation. Again, three levels of required precision were selected for consideration: $2.40, $4.80, and $7.20/m² ($2.00, $4.00, and $6.00/yd²). Based on these assumptions and again using equation 1, table 2 was produced.

As expected, table 1 shows that for higher levels of precision, higher standard deviation values, and higher confidence levels, greater numbers of relative performance responses are needed. For this project, a 95 percent confidence level and a precision level of 5 years are considered appropriate. Assuming a high variability in the life estimates (10 years), a minimum of 16 surveys (rounded up from the corresponding value of 15.4 in table 1) is desirable for the performance surveys.

Similarly, table 2 shows that for higher levels of precision, higher standard deviation values, and higher confidence levels, more relative cost responses are needed. For this project, a 95 percent confidence level and a precision level of $4.80/m² ($4.00/yd²) are probably appropriate. Assuming a high variability in the responses ($10.80/m² ($9.00/yd²)), a minimum of 20 surveys (rounded up from the corresponding value of 19.4 in table 2) is desirable for the cost surveys.

The above analysis is very approximate, but it was used as an initial goal for the number of questionnaires to be obtained for the performance and cost surveys.

SHA Performance Questionnaire Survey

A project summary and request for participation was faxed to 43 SHAs; 25 agencies ultimately agreed to participate. Once appropriate contacts within each agency had been identified, the relative performance questionnaire forms were faxed or mailed to the participants; 45 days were allotted to them to submit completed forms. During this time, typically three deadline reminders were faxed to the volunteers. These reminders were sent during the third and fourth weeks and on the survey completion date.

A total of 12 SHA responses were received, despite several concentrated efforts to increase that number. This falls slightly lower than the original target of 16, but is believed to still be sufficiently high to provide meaningful results. Agencies providing responses to the questionnaire surveys were:

• Florida.
• Illinois.
• Indiana.
• Kansas.
• Minnesota.
• New Jersey.
• Ohio.
• Oklahoma.
• South Dakota.
• Tennessee.
• Washington.
• West Virginia.

A summary of the raw SHA relative performance ratings is provided in appendix C; a concise summary of the responses is shown in table 3. This table shows several different columns of information related to the relative performance ratings. The column titled Average Survey Results contains the direct averages of the performance ratings provided by all SHAs. However, because of the considerable variability associated with many of these responses, perceived outlier data points were removed and the resultant performance ratings presented under the column titled Modified Survey Results. Outlying data points were identified as those performance values that 1) grossly contradicted the expected performance trends or 2) were greatly different in magnitude from the reasonable performance range.

The column titled Average Model Results in table 3 presents the average performance ratings as computed by available performance models; it is used as a basis for checking the reasonableness of the survey results (see Performance Model Evaluation section below). Based on the consideration of the survey results and the performance models, the recommended ranges of performance ratings for each design feature category are presented in the column titled Recommended Values.

A quick review of the recommended values shown in table 3 indicates that slab thickness has the largest effect on relative pavement performance, as might be expected. Other factors noted to have a major effect on relative pavement performance include the type of base (including the absence of a base course), drainage, dowel bars (the absence of dowel bars considerably decreased performance), and widened slabs.

Performance Model Evaluation

Due to the variability observed in the performance-related survey responses, available PCC pavement performance models were used to develop relative performance ratings for certain design features that may be added to or deleted from the Standard pavement section. Unfortunately, not all design features could be evaluated using these models because many features are not direct input variables in the models. Table 4 summarizes the models that were used in this study and the corresponding design features that could be evaluated with each model.

The average performance ratios coming from an investigation of all of the aforementioned models are presented in the Average Model Results column of table 3. As previously mentioned, using a combination of the results from the modified performance ratings and the results from the available performance models, recommended ranges for each design feature category were developed and are presented in table 3. The recommended ranges are used to define the default performance changes (i.e., the default performance data set) used in the analytical software tool developed for this project.

Key: STD = Standard; ATB = asphalt-treated base; CTB = cement-treated base; ATPB = asphalt-treated permeable base; CTPB = cement-treated permeable base; o.c. = on center; n/a = not available.

All ratings are relative to the performance of the Standard pavement section.

Key: STD = Standard; n/a = not available.

SHA Design Category Ranking Survey

In addition to the relative performance ratings, SHA respondents were asked to rank the relative importance of each design feature category to PCC pavement performance. That is, of the 10 design feature categories (see below) respondents were asked to rank each factor on an integer scale of 1 to 10, with 10 representing the most important and 1 the least important. No two design features were allowed to share the same ranking, so the result was a "forced ranking" of the importance of each design feature category. These rankings were incorporated into the analysis approach as a way of accounting for the effects of multiple design feature changes on pavement performance (see discussion in chapter 3).

The final recommended category rankings are presented in table 5 and are based on the collected survey results. It is readily acknowledged that a designer or agency could have different opinions about the ranking factors. Therefore, although the category ranking factor set displayed in table 5 is used as the default ranking factor set in the software, an agency may create their own ranking factor set that reflects their expected relationships between design categories and concrete pavement performance.

Table 5. Details of the recommended category ranking factor set.

Design Category	Ranking Factor	Most Important Least Important
Joints/Load Transfer	10
Thickness/Slab Size	9
Base/Subbase	8
Drainage	7
Strength/Materials	6
Subgrade	5
Initial Smoothness	4
Joint Sealing	3
Cross Section	2
Shoulders	1

Contractor Cost Questionnaire Survey

A project summary and request for participation was faxed to 216 contracting companies provided by the American Concrete Pavement Association (ACPA). A total of 53 firms responded to the request; 38 volunteered to participate in the study. The relative cost questionnaire forms were faxed or mailed to each volunteer. Forty-five days were allotted to complete the questionnaire surveys. As with the relative performance questionnaire surveys, typically three deadline reminders were faxed to the volunteers. These reminders were sent during the third and fourth weeks and on the survey completion date.

Sixteen responses were received from contractors. As with the SHA surveys, this value falls slightly lower than the original target of 20, but it is believed to be sufficiently large to provide meaningful results. Paving contractors responding were:

• Baker Concrete Construction, Monroe, OH.
• Berns Construction Company, Indianapolis, IN.
• Cedar Valley Corporation, Waterloo, IA.
• Duit Construction Company, Edmond, OK.
• Flynn Company, Dubuque IA.
• Gale Tschuor Company, Muncie, IN.
• The Harper Company, Hebron, KY.
• Irving F. Jenson Company, Sioux City, IA.
• K-Five Construction, Lemont, IL (estimator A).
• K-Five Construction, Lemont, IL (estimator B).
• K-Five Construction, Lemont, IL (estimator C).
• Koss Construction, Topeka, KS.
• The Sundt Companies, Tucson, AZ.
• Weaver-Bailey Contractors, El Paso, AR.
• Western Plains Construction, Lincoln, NE.
• Wittwer Paving Inc., Wichita, KS.

A summary of the raw relative cost data collection responses is provided in appendix C. The averages of all survey results for each alternative design feature are displayed in table 6. An initial review of the data indicates that comprehensive drainage systems result in a substantial increase in pavement costs. Other design features that contribute significantly to pavement costs are PCC shoulders and high-early strength PCC mixtures. On the other hand, factors such as pavement cross sections (trapezoid or thickened edge), widened slabs, and joint sealing appear to have very little effect on pavement construction costs. Note that the cost associated with achieving different levels of initial smoothness is dependent on the selected base type.

Summary

This chapter summarizes the data collection activities for this project. A description of the literature search is first presented, which provided many useful resource documents that served as the basis for the development of the questionnaire surveys. The approach used in developing the questionnaire surveys is also provided, which produced two separate questionnaire surveys: one targeted to state highway agencies and intended to collect relative performance ratings for changes in design features, and one targeted to PCC paving contractors and intended to collect relative cost ratings for changes in design features. Each questionnaire survey was structured in the same manner so that the performance and cost data could be matched up for each design feature.

The results of the performance and cost surveys are summarized in this chapter, with raw survey results presented in appendix C. The performance ratings that were received from the highway agencies were evaluated in conjunction with available performance models to check their validity and reasonableness. Overall, these data serve as one "data set" for use in the computer program developed under this project (see chapter 3).

Key: STD = Standard; ATB = asphalt-treated base; CTB = cement-treated base; ATPB = asphalt-treated permeable base; CTPB = cement-treated permeable base; o.c. = on center.

All ratings are relative to the cost of the Standard pavement section.

Key: STD = Standard; ATB = asphalt-treated base; CTB = cement-treated base.