Long-Term Pavement Performance Compliance With Department of Transportation Information Dissemination Quality Guidelines

CHAPTER 1. LTPP DATA SYSTEM PLANNING

The DOT IDQG indicates that the following factors should be considered for planning a data system:

• Data system should be linked to the organization's strategic plan.
• Objectives for the data system should be expressed in terms of goals.
• Data requirements should be based on achievement of goals.
• Data acquisition methods are based on data requirements and use.

Data System Objectives

In order to garner widespread support and keep relevancy over a long period of time, LTPP goals were purposely expressed in terms of general topical areas of pavement engineering needs. Although a formal assessment of the program has been performed, the objective statements still remain relevant.

The LTPP program was started as a States' initiative and implemented as part of the SHRP, operated under the auspices of the National Academy of Sciences, to meet a need for research quality data on long-term performance of pavements in North America. The strategic goals for the program were establish by a committee with representation from State and Provincial highway agencies, Federal agencies, and academia. These goals were endorsed by AASHTO.

Although the LTPP program did not start as a program managed by a Federal agency, its goals and objectives were congruent with those of FHWA in the early 1990s. This is why at the end of the 5-year SHRP effort in 1992, FHWA management agreed to provide management services for the remainder of this long-term effort. Although the LTPP program is not specifically addressed in the changing strategic objectives of FHWA, it has continued to be a line item in all of the highway bills passed by Congress to date. Due to the LTPP program's unique enabling role in the advancement of pavement technology, its data will be used as a tool far into the future to address future highway infrastructure engineering needs.

Highlights of LTPP's compliance with this portion of the guidelines include the following:

• Updates to the formal objective statements have not been necessary.
• Key questions to be answered by the data are expressed in the objective statements.
• Timeliness of data is not an important aspect to the accomplishment of the project's goal, as this is a long-term project, and therefore is not directly addressed in the objectives.
• The data system objectives are distributed with the data in the form of a database user's reference guide. This guide is also available from the LTPP Publications Web site.
• Data user information and feedback is obtained through a variety of sources. Biannual meetings of the LTPP TRB Committee are held to provide input on programmatic issues. The LTPP ETGs provide input on specific data topics. A data user satisfaction survey is distributed with the data and is available on the LTPP Web site. LTPP holds two sessions at the annual TRB meeting allowing questions and comments from Federal, State, academia, and industry. These groups represent LTPP's stakeholders.

The objective of the LTPP data system is to provide data to engineering researchers that enable the evaluation of existing pavement design methods, the development of improved pavement design methods, and the study of the effect of relevant factors that influence pavement performance.

Data Requirements

The LTPP data requirements are based on the achievement of the program's goals and objectives. The data requirements were developed by experts in each associated engineering discipline with review and critique by other stakeholders through TRB- facilitated national meetings.

The development of measurement concepts for the LTPP program followed a scientific approach methodology. Data needs were based on existing models, theories of pavement performance, and anticipated needs of future models of pavement performance. The findings from previous studies of the performance of in-service pavements were also used.

The LTPP program relies on thousands of measurement concepts. Some of the more important measurement concepts used in the LTPP program include the following:

• Important characteristics in determining the cause and effect relationship on why pavements perform as they do include the following:
  • The strength of the pavement structure.
  • The thickness and types of materials in the pavement structure.
  • The strength, elastic, and plastic properties of the pavement materials.
  • The magnitude and volume of wheel loads applied to the pavement.
  • The response of the pavement structure to load.
  • The climate characteristics at the site.
  • The condition of a pavement as a function of its roughness and features related to fracture, distortion, and disintegration of the surface materials.
• In order to maintain coherence across other databases, many existing concepts used in other databases were adopted. Some of these include the following:
  • Federal Information Processing Standards (FIPS) for geographic codes.
  • Highway Performance Monitoring System (HPMS) test section identification.
  • National Transportation Thesaurus for keywords.
  • World Geodetic System 1984 for location coordinates.
  • AASHTO classification system for soils and unbound base material.
  • AASHTO material test standards.
  • American Society for Testing and Materials (ASTM) test standards.
• While desired data accuracy and repeatability targets were developed for a vast array of measurement concepts, formal targets were not set for all data elements. Where possible, LTPP has strived to provide data users with quantitative information that permits an assessment of actual measurement error, accuracy, and repeatability.
  • An accuracy of 90 percent the confidence level being within 10 percent of the mean was a goal for traffic monitoring measurements, stated in terms of annual loading of equivalent 18-kips single axle loads in the LTPP test lane. Because many factors enter into the accuracy of field axle weight measurements, some of which are difficult to control, information on sources of error introduced by sampling and equipment calibration are stored in the database.
  • Use of industry standardized material testing protocols permits use of accuracy and repeatability studies performed in development of the protocol.
  • In some cases, where equipment calibration standards did not exist, LTPP developed calibration protocols and test procedures to assess equipment accuracy and repeatability. For example, a reference calibration procedure was developed by LTPP for FWDs.

The LTPP data requirements are based on the concept of providing the user with baseline raw data that can be used to support a wide range of measurement concepts. This is a fundamental requirement of a research-based program. While the database contains traditional measurement concepts, data users are able to access the baseline measurement data in order to investigate their applicability for development of new measurement concepts.

Methods to Acquire Data

The methods to acquire LTPP data were developed by expert staff working in-concert with program stakeholders and other experts. The preliminary planning for the LTPP program was funded under an NCHRP-sponsored study conducted under the auspices of the LTPP TRB committee. A comprehensive process was used to examine and evaluate data acquisition methods that took into account budget, complexity, ease, and time considerations.

Due to the complexity of many LTPP data elements, an early decision in the program was to use qualified data collection contractors with experience in operation of technical data collection equipment and working with highway agencies on collection of other data. Greater resources were planned and used on the more complex data elements requiring specialty resources.

FHWA served as a cooperative partner in this work by funding an equipment evaluation project to evaluate the state of the practice and art in the critical data elements needed for the project during the preimplementation phase of the project.

Sources of Data

LTPP data sources were developed based on past experience in performing similar field studies. Findings from a research study sponsored by FHWA in the early 1980s called the Long-Term Pavement Monitoring Program⁽¹⁾ provided the basis for selection of data sources.

Since provision of data to the LTPP program was not a matter of public law, LTPP had to rely upon contractual data collection services, available data sources, and the good will of participating highway agencies.

The sources of data selected for the LTPP program included the following:

• Participating highway agencies were used for inventory and construction information that were contained only in their files.
• Pavement monitoring measurements were assigned to a team of expert contractors operating the same type of equipment in order to improve uniformity and reduce variability. Highly technical measurements which included specialized test section instrumentation were assigned to contractors.
• Climate data were obtained from the National Climate Data Center (NCDC) for test sections in the United States and the Canadian Climate Center (CCC) for test sections in Canada.
• The FHWA's advance material testing laboratory at Turner-Fairbank Highway Research Center (TFHRC) was used for measurement of the coefficient of thermal expansion on portland cement concrete (PCC) and for Superpave® tests on asphalt cement.
• The National Aggregate Institute performed material tests on the angularity of fine aggregates in asphalt concrete (AC) mixtures.
• Mixed approaches for the collection of some data items were required due to budgetary limitations. Early in the program, all material tests were performed under LTPP contract. Another set of material tests on newly constructed pavement test sections was assigned to participating highway agencies.
• Over time, shifts in data sources have been necessary due to data quality and availability concerns. Collection of traffic data on LTPP test sections has been a participating highway agency responsibility. In order to overcome the limitations of this data, the LTPP Specific Pavement Study Traffic Pooled Fund Study TRF-05 (004) was implemented to obtain needed data from high priority test sections.

Data Collection Design

The LTPP data collection design was based on a wide variety of methods, as follows:

• Due to cost considerations, the population of test sections was not selected to represent a statistical sample of the pavement types used in North America but rather was restricted to a subset of pavement types thought by experts to represent "good" engineering practice.
• Factorial sampling designs were used in which test sections were categorized by significant design features relative to the pavement type and environmental factors to direct test section selection. A large majority of LTPP experiments can be most properly described as uncontrolled fractional factorials since test section selection and sample size depended upon voluntary agency participation and the "diagonal" effect of public roads. The diagonal effect of properly engineered roads means that some of the combinations of factors in an experimental design are not common. For example, thin pavement structures on high traffic volume routes on poor subgrade are difficult to find since they are contrary to engineering design practice. The "diagonal" is best described as typical engineering practice in which thicker pavements are built on roads with higher traffic loadings and volume.
• For the Specific Pavement Studies (SPS)-1 and -2 controlled experiments, which started with new construction, blocked fractional factorial experimental designs were used to reduce the number of test sections constructed at a site and still maintain desired statistical inferences. This permitted construction of some off "diagonal" pavement structures designed to have a "limited" life.
• All experiment designs were developed and reviewed by expert statisticians with pavement research experience. These designs also received peer review by an ETG consisting of both expert statisticians and pavement research engineers.
• Testing patterns were designed to capture known seasonal and location effects; in order to capture these effects, randomized testing patterns were not used.
• Repeat measurements were used when possible to provide an estimate of measurement variances. For example, measurements of pavement stiffness using a FWD used four drops from each drop height. Measurement of longitudinal pavement profile used a minimum of five and up to nine repeat measurement passes on a test section.
• When possible, the sampling theory used for the experiment and the data collection design are contained in the program documentation. Most of this information is included within the Reference Library disseminated with the data.
• The one critical data element which LTPP relies on for statistical sampling is the collection of traffic weight, classification, and volume data. Although the stated preferred sampling practice is continuous monitoring of traffic weight and classification by a calibrated weigh-in-motion (WIM) station, since this portion of LTPP data collection is primarily dependent upon voluntary highway agency participation, the program developed a hierarchical approach to different sampling scenarios. Due to the lack of other information, an analysis study of data from sites with full-time operating WIM scales was used to evaluate traffic data sampling errors and bias. The basis of these sampling scenarios is documented in a report whose findings are published in a LTPP TechBrief.⁽²⁾