2. Literature and Specification Review

Previous | Table of Contents | Next

LITERATURE REVIEW

One of the initial tasks associated with the project was to conduct a literature search to determine what information was currently available on the topic of QA specifications and to determine its potential relevance to the project. This task was an iterative process requiring multiple searches as the project moved forward. To encompass all of the potentially pertinent information available, a methodical approach to the literature search was adopted.

Search Methodology

Since the project team members have written extensively in the area of statistically based and QA specifications, the initial step in the literature search was to locate and review all of the information previously published by members of the research team. This yielded many published documents on such topics as sampling schemes, statistical specifications, incentive and disincentive payment schedules, acceptance procedures, statistical methods for analyzing data, and the construction of operating characteristic (OC) curves. Upon completion of this phase of the search, efforts were then directed toward outside sources of information.

The next source of information that was explored was all of the publications contained in Clemson University's Cooper Library databases. The databases were explored using various combinations of keywords, as well as subject searches. This generated numerous potential sources of information. Each possible source was reviewed and either added to the list of source documents or discarded based on the relevance of the information to the project. Upon completion of this phase, efforts were then directed toward national and international databases.

Searching these databases by keyword and subject yielded hundreds of potential titles. Among the databases searched were:

• TRIS (Transportation Research Information Service).
• Ei Compendex (engineering information).
• Dissertation Abstracts (online database of doctoral dissertations).
• ICONDA (International Construction Database).
• NTIS (National Technical Information Service).
• MathSci (American Mathematical Society).
• Engineering Materials Abstracts (engineering materials).
• Federal Research in Progress (ongoing Federal research projects).

The MathSci database was included to identify related articles on topics such as sampling plans, acceptance plans, etc., from mathematical and statistical sources such as the Journal of Quality Technology and the Journal of the American Statistical Association.

All of the titles generated from this search were examined to determine if the publications should be located and obtained. Where the titles did not provide a clear indication as to the relevance of the publication, abstracts were requested to determine if the publication was an appropriate source document.

The final phase of the literature search was to review the bibliographic contents of the collected articles to determine if publications cited in the current source documents would also be relevant to the research project. This phase yielded some additional sources.

Results

The information contained in the bibliographic sources can be separated into three major categories:

• Sampling plans.
• Price- or payment-adjustment plans.
• QA specifications and acceptance plans.

Sampling Plans: The bibliographic sources contain several different types of sampling plans. Some of the plans are currently in use for highway construction, while others were evaluated to determine whether there is an appropriate use for the plans in QA specifications. Sampling plans for which literature was reviewed included:

• Stratified random sampling.
• Random sampling.
• Zero acceptance number sampling plans.
• Variables sampling plans.
• Attributes sampling plans.
• Attributes single sampling plans.
• Attributes double sampling plans.
• Attributes multiple sampling plans.
• Attributes proportion sampling plans.
• Inverse Gaussian acceptance sampling plans.
• Conditional sampling plans.
• Bayesian acceptance sampling schemes.

Price- or Payment-Adjustment Plans: The bibliographic sources also contain a number of articles discussing price- and payment-adjustment plans. Again, some of the practices discussed are currently in use, while others were examined for potential relevance to the project. Several articles discuss adjusted payment schedules as they pertain to the quality of work and expected performance, and whether current practices are fair or can be improved. The development of price-adjustment systems is also addressed from serviceability, cost of production, value concept, and OC curve approaches. The concept of composite payment equations and their development and use is addressed in several of the bibliographic sources.

QA Specifications and Acceptance Plans: The last category of bibliographic sources is related to specifications development, acceptance procedures, and QA programs. Several articles investigate the older prescriptive highway specifications and discuss the shift toward end-result and performance-related specifications (PRS). Included is an examination of several statistical measures and how each has been used.

Summary

Not much new information was learned from the literature review. The project team members had written much of the relevant literature from the highway materials and construction industry. Most of the material from outside the highway industry was directed toward industrial applictions, where more stable processes and much larger sample sizes are involved. Some of the relevant information from the literature search is discussed in subsequent chapters related to the specific analysis topics, and some is presented in the QA specifications manual that was developed for the project.

A summary of some of the documents that were reviewed during the course of the search is presented as an annotated bibliography in appendix A.

SPECIFICATION REVIEW

At the beginning of the project, all SHAs were requested to provide copies of their current specifications for HMAC and PCC pavements. In response, 23 provided copies of their specifications for HMAC, 8 provided specifications for Superpave^® AC mixes, and 9 provided specifications for PCC (see table 3). Many States indicated that Superpave specifications were under development at that time. The reasons for collecting and reviewing the various agency specifications were to:

• Determine what properties were being evaluated for quality control (QC) purposes.
• Determine what properties were being evaluated for acceptance purposes.
• Determine what properties were being evaluated for payment determination.
• Determine what statistical methods were being used to determine a composite payment factor when there are multiple quality characteristics.

HMAC Specifications

A total of 23 agencies (see table 3) provided copies of their current HMAC specifications. A summary of each agency's HMAC specifications is provided in appendix B. While current at the time they were obtained, these specifications may no longer be current as of the writing of this report. Each of the agency's specifications was carefully reviewed to determine the similarities and differences in the properties evaluated for QC, acceptance, and payment determination. Additionally, the statistical measures for determining compliance and the calculations for payment factors were reviewed. This task was complicated and required many iterations because of the differing terminologies used by the various agencies. For example, some agencies use assurance testing for the same purpose as tests that are referred to as acceptance testing by other agencies.

Table 3. Agencies that provided copies of their specifications.

Examination of the specifications revealed that the majority of the agencies use the Marshall mix design and therefore the quality characteristics evaluated for QC and acceptance are similar. The most significant difference is in the number of quality characteristics that the contractor is responsible for controlling. Some agencies require the contractor to control a few common characteristics, such as gradation, asphalt content, density, voids in the mineral aggregate (VMA), voids filled with asphalt (VFA), and total air voids. Almost all of the agencies use these characteristics for QC or acceptance. However, some agencies require control over many more characteristics, including Hveem stability, Marshall stability, Marshall flow, dust-to-asphalt ratio, maximum specific gravity (MSG), bulk specific gravity (BSG), moisture content, binder temperature, liquid limit, plastic index, fractured faces, absorption, indirect tensile strength (ITS), and tensile strength ratio (TSR). Additionally, differences in the lot sizes for testing varied widely from agency to agency. Testing frequencies are also significantly different for the various agencies. This review indicated that with the exception of a few commonly measured characteristics, the QC and acceptance procedures varied widely among the responding agencies.

The methods for determining acceptance were also investigated. The method for acceptance of material varies from agency to agency, but can be grouped into four general categories:

• Acceptance testing by the department.
• Verification of the contractor's QC tests by the department's assurance tests.
• Acceptance testing by the contractor under departmental supervision.
• Some combination of contractor and departmental testing.

The final aspects of the specifications examined were the properties evaluated and the methods for determining payment factors. Most of the responding agencies evaluate only a few properties for determining payment factors. The most common properties used are gradation, in-place density, asphalt content, VMA, and air voids. Additional properties evaluated by some agencies include Marshall stability, crushed particle count, thickness, moisture content, theoretical maximum density (TMD), laboratory-molded density, and smoothness.

Superpave Specifications

The eight States that responded with copies of their Superpave specifications are indicated in table 3. The response for this specification was low at the time because of its recent introduction as a method of mix design. A summary of each agency's Superpave specifications is provided in appendix C. The Superpave specifications were reviewed for QC, acceptance, and payment factor information.

Examination of the specifications indicated that verification of the mix design is similar for all of the agencies. Additionally, the quality characteristics evaluated for QC and acceptance do not differ substantially from agency to agency. However, there is a significant difference in the quality characteristics evaluated for acceptance. Most of the agencies evaluate the following characteristics: asphalt content, gradation, air voids, VMA, and in-place density. In addition to these commonly evaluated characteristics, three agencies evaluate mix moisture, VFA, and BSG, and two agencies evaluate TMD, dust-to-asphalt ratio, and G_mb @ N_des. At least one agency evaluates a number of other quality characteristics, such as TSR, sand equivalent, percent crushed aggregate, N_ini, N_des, and N_max. Four of the eight agencies use smoothness as an acceptance quality characteristic.

The methods used to determine acceptance can be grouped into two categories: acceptance testing by the agency and verification of the contractor's tests by the agency's verification tests. The most common quality characteristics used in determining payment factors are gradation, asphalt content, air voids, in-place density, and smoothness.

PCC Specifications

The nine agencies that provided copies of their PCC specifications are indicated in table 3. A summary of each agency's PCC specifications is provided in appendix D. The review of the PCC specifications revealed similarities in the quality characteristics that the contractor is responsible for controlling. Each of the agencies requires the contractor to conduct QC tests for aggregate gradation, air content, slump, unit weight/yield, and compressive or flexural strength. A majority of the agencies also require the contractor to control thickness, temperature, and smoothness. Additional characteristics evaluated by at least one agency include water-cement ratio, percent passing the 75-micrometer (mm) sieve, moisture content of the aggregate, fineness modulus, sand equivalent, fine aggregate organic impurity, and admixture dosage. The QC testing frequency requirements vary widely from agency to agency.

The quality characteristics evaluated for the acceptance of material vary only slightly from agency to agency. Most of the agencies evaluate aggregate gradation, slump, temperature, smoothness, unit weight/yield, thickness, air content, and compressive or flexural strength. The method for determining the acceptance of material varies from agency to agency, with either agency verification testing of the contractor's tests or acceptance testing by the agency. The responding agencies assigned payment factors for one or more of the following quality characteristics: smoothness, thickness, air content, compressive strength, and flexural strength.

Summary

While the review of agency specifications provided background information, not a great deal of new and useful information resulted from the review. The project team members were already familiar with many agency specifications. Also, at the panel meeting to approve the QA specification process flowchart (see next chapter), the panel decided that the QA manual would not be prescriptive with regard to which quality characteristics should be measured for QC, acceptance, or payment purposes.