U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
202-366-4000


Skip to content
Facebook iconYouTube iconTwitter iconFlickr iconLinkedInInstagram

Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations

Report
This report is an archived publication and may contain dated technical, contact, and link information
Publication Number: FHWA-RD-03-094
Date: March 2005

Estimating Cumulative Traffic Loads, Volume II:
Traffic Data Assessment and Axle Load Projection for The Sites With Acceptable Axle Weight Data, Final Report for Phase 2

CHAPTER 5. PROTOTYPE LTPP PAVEMENT LOADING GUIDE

This chapter describes the purpose, design parameters, and functionality of the proposed PLG. It contains a blueprint for the development of the PLG and examples of using the PLG to obtain traffic load projections for LTPP sites without site–specific truck class or axle load data.

The need for the PLG was identified during the Phase I work and was reinforced by the traffic prediction work carried out in Phase 2. The process of estimating, selecting, or assessing truck class and axle weight distributions for LTPP sites needs to be supported by a reference database summarizing characteristic truck class and axle load distribution data. This reference database should contain not only traffic characteristics encountered on the LTPP sites, but also typical benchmark characteristics for truck class and axle load distributions. The PLG would fulfill:

  • The need to estimate missing truck class and axle load distributions for Category 3 and 4 sites.
  • The need to assess the quality of monitoring traffic data used for traffic projection.

The Need to Estimate Missing Traffic Data

The traffic data assessment and traffic projection work summarized in this report processed all 890 LTPP sites. Of the 890 sites, axle load projections were carried out for 558 sites. Of the 558 sites with axle load projections, the projections for 194 sites were assigned an acceptable projection confidence code and 364 sites were assigned a questionable projection confidence code. No traffic load estimates were provided for 332 LTPP sites for which the required site–specific traffic data (truck class distributions or axle loads) were not available. Yet many of the 332 sites without the site–specific data contain a wealth of information regarding pavement materials, environment, and pavement performance. For example, pavement materials were subjected to a battery of laboratory tests, and the pavement performance has been evaluated using a series of profile, distress, and falling weight deflectometer (FWD) measurements over the years. Without the corresponding traffic data, this wealth of information cannot be utilized for developing load–related pavement performance models.

The additional traffic data that need to be estimated depend on the traffic load projection category of the site:

  • For Category 3 sites, axle load spectra for the individual truck.
  • For Category 4 sites, truck class distribution and axle load spectra for the individual truck classes.

The estimation of missing truck class and axle weight distributions can have significant consequences on the reliability of the resulting traffic load estimates and must be done judiciously.

The estimating process should consider data collected in the vicinity of the site on the same highway (site–related data) and data collected on similar highway links in the same area or jurisdiction (regional data). In many cases, applicable site–related or regional traffic data may not be available. To facilitate the task of estimating traffic data for a variety of sites, and to make the estimating process consistent, a database containing traffic values for a variety of sites, as well as typical benchmark values, is required. Such a database will be developed as part of the proposed PLG.

There is no substitute for site–specific traffic data. However, to utilize the large amount of pavement–related data collected on the 332 LTPP sites with missing traffic data for the prediction of pavement performance, some traffic data need to be estimated. Research shows that it is possible to use estimated traffic data to carry out basic calibration and verification of pavement design models.[16] Without traffic estimates, it would not be possible to utilize many LTPP sections for the development and verification of load–dependent pavement performance models.

The Need to Assess the Quality of Monitoring Traffic Data

As discussed in chapter 2, traffic data stored in the LTPP databases exhibit considerable variation in quality. Consequently, the annual truck class and axle load distributions need to be examined and verified before they can be used for traffic projection. The assessment and verification of truck class and axle weight distributions requires the development of software that displays and compares data.

In addition, it has become a common practice to assess the quality of axle weight data by examining the GVW distribution of Class 9 vehicles (5–axle single trailer trucks).[12] However, the data necessary to plot the GVW of Class 9 trucks against the frequency of occurrence are not stored in the IMS database, and the retrieval of the necessary data from the CTDB and the development of appropriate graphical displays require considerable effort and specialized skills. Computational software that can facilitate the development of graphical displays showing the GVW distribution of Class 9 vehicles exists, but is not readily available and would need to be adapted to process annual data.

For the efficient verification and QC of traffic data used for the projection, it is desirable to compare axle load and other data obtained for a specific site with corresponding data obtained for similar sites and with typical or expected data. By facilitating the comparison of axle load spectra and other traffic data, and by providing typical values and ranges for traffic variables, the proposed PLG will facilitate the QA process.

Scope of Pavement Loading Guide: LTPP PLG or General PLG

A catalogue containing validated truck class and axle load distributions, and with additional capabilities to display and compare data graphically, would be useful not only for the projection of LTPP traffic data, but also for estimating traffic loads for general pavement design purposes. The need for such tool is highlighted by the emergence of mechanistic–based pavement design procedures that require axle load spectra. The use of axle load spectra in pavement design is relatively new. Many pavement designers may benefit from information on typical values of truck class and axle load distributions that are expected to occur on different classes of highways. Consequently, the LTPP PLG has the potential for wider applicability beyond the needs of the LTPP program.

The main purpose of the LTPP PLG is to facilitate traffic loading projections for LTPP sites without site–specific traffic loading data, and the work summarized in this chapter has been directed toward the development of the PLG that would serve this need. Only a limited amount of work has been done to prepare groundwork to enhance the LTPP PLG to serve general pavement design needs.

There are many similarities between a PLG serving LTPP's needs and a PLG serving general pavement design needs. Of course, the two versions also have several differing characteristics.

Characteristic Features of LTPP PLG
  • The LTPP PLG is concerned mainly with backcasting of traffic loads using historical and monitoring data.
  • Data input and output functions need to be compatible with the LTPP databases (IMS and CTDB).
  • The number of potential active users is limited, and the majority of users are expected to have a research background. Consequently, user–friendliness of the software and application guidelines may not be of paramount importance.
  • Traffic projections need to be carried out only in terms of axle load spectra. The LTPP traffic projection process uses ESALs for QC purposes only.
Characteristic Features of General PLG
  • The general PLG is concerned with forecasting of future traffic loads.
  • Data input functions need to be flexible to accommodate user data; data output functions and formats should match the input format requirements of common pavement design methods (e.g., 1993 AASHTO guide[17] or the not–yet–finalized National Cooperative Highway Research Program (NCHRP) 2002 guide[7].
  • The number of potential users is large, and they may have quite diverse backgrounds. User–friendly software and application guidelines are important.
  • The majority of pavement designers currently need axle loads in ESALs. Axle load spectra will be required by the NCHRP 2002 guide and by other mechanistic–based pavement design methods.

This chapter concentrates on describing the development of the LTPP PLG to facilitate the projection of traffic loads on LTPP sites. However, the concepts for the development of the LTPP PLG described in this chapter also apply to the development of a general PLG.

Objectives of the PLG and Purpose of This Chapter

It is necessary to make a distinction between the objectives of the PLG and the purpose of this chapter. The objectives of the PLG refer to the functionality and potential use of the proposed PLG, its software and guidelines, and include the following:

  • Provide tools for projecting traffic loads for LTPP sites without site–specific truck class and axle loads distribution data.
  • Facilitate the assessment and QA of LTPP traffic data.
  • Facilitate understanding of traffic load characteristics, such as truck distributions and axle load spectra, and provide educational and training opportunities.
  • Provide groundwork for a PLG that can be used for forecasting traffic loads for pavement design.
  • Promote the use of LTPP traffic data for other applications.

This chapter describes the functionality of the proposed PLG and the methodology for its development. Some specific objectives of this chapter are to:

  • Develop design criteria and functional features required to meet the objectives of the PLG.
  • Describe data, information, and data analysis needed to develop the engineering underpinnings for the PLG, such as typical characteristics of axle load spectra for different highway functional classes.
  • Describe the use of the PLG to obtain LTPP traffic projections for sites without site–specific data using practical examples.
  • Illustrate the operation of the PLG software using prototype demonstration software.

Working with axle load spectra involves working with large data sets that are best handled by a computer. To illustrate some features of the proposed PLG, particularly how axle load spectra can be compared, combined, or selected using graphical displays, prototype PLG software was developed. The operation of the software is described briefly in this chapter.

The prototype demonstration software contained several functions of the proposed PLG but lacked many of the analytical and engineering underpinnings that still need to be developed. The main purpose of the prototype software was to illustrate the overall concept and functionality of the proposed PLG.

Role of PLG in Traffic Projection

Traffic data collected in the field are typically in the form of samples that are uneven in duration and quality. The samples of collected traffic data are used to calculate cumulative traffic loads through an analytical process involving factoring and traffic modeling referred to as traffic projection. The LTPP traffic projection is done in two steps. First, the monitoring data collected during a given year are projected (or factored up) to obtain annual monitoring data (this factoring step is required because traffic monitoring equipment seldom operates all the time). Second, the annual monitoring data (available for some of the years) are combined with historical data and are projected to all years the pavement was in service.

Traffic data collection alone is not enough to obtain cumulative traffic loads. The data that were not collected in the past will remain missing, and it is not possible (nor it is necessary) to collect traffic data all the time. To obtain cumulative traffic loads for all years the pavement has been in service, it is necessary to use both the traffic data (historical and monitoring data) and the traffic projection procedures.

The PLG is not a substitute for the site–specific collection of traffic data, which is required to obtain reliable estimates of traffic loads and needs to be encouraged and promoted. However, traffic data alone, without traffic projection procedures, will not provide the required results, either. Both the data collection and the data projections activities have undeniable roles and need to be n balance. The role of the PLG is to strengthen and facilitate the traffic data projection. The use of PLG, together with guidelines for selecting missing data, can be instrumental in alleviating some uncertainty resulting from the unavoidable need to factor up and estimate traffic data.

Conceptual Outline of PLG

This section outlines how the concept of estimating ESALs using TFs can be modified to apply to estimating axle load spectra.

The AASHTO Guide for Design of Pavement Structures 1993, the most common pavement design method in North America, uses ESALs to characterize traffic loads.[17] Consequently, many pavement professionals are familiar with the concept of estimating ESALs for pavement design. In the following, the procedure of estimating traffic loads using ESALs is described to demonstrate its similarity with the procedure, utilized by the PLG, of estimating traffic loads using axle load spectra.

Estimating ESALs typically starts with the division of the average annual daily traffic volume into a car volume and a truck volume. Truck volume is then subdivided into several classes, and each truck class is assigned a representative TF defined as the number of ESALs per truck. For example, the computerized version of the 1993 AASHTO guide (the DARWin® 3.1 pavement design software) encourages the user to classify vehicles into 13 vehicle classes and to provide TFs for each vehicle class.

To obtain the daily number of ESALs for a base year, it is necessary to sum the product of the daily truck volumes in different classes and their corresponding TFs. To obtain the annual number of ESALs, the average daily number of ESALs is multiplied by 365. Mathematically, the process of calculating ESALs can be expressed by equation 3.

The total number of equivalent single axle loads for a base year is equal to the sum of average daily volume from 1 to the number of truck classes times the truck factor for truck class i times the truck factor for truck class i times 365.
(3)

Where:

ESALs year
= Total annual number of ESALs for a base year.
i = Truck class number.
n = Number of truck classes.
ADTVi = Average daily volume of truck class i.
TFi = Truck factor for truck class i.
365 = Constant to convert daily traffic to annual traffic.

To obtain the number of ESALs for an entire pavement design period, the annual number of ESALs estimated for a base year is factored to account for traffic that will occur during the entire design pavement service period.

For LTPP traffic projection purposes, the projection is not done in terms of ESALs but in terms of axle load spectra. Thus, instead of the representative TFs for different truck classes used in equation 3, we need to use the representative axle load spectra for different truck classes. The overall concept of classifying trucks into classes, assigning a representative load–related factor to each truck class (i.e., axle load spectra or TFs), and combining the result remains the same.

Using the axle load spectra instead of TFs, equation 3 can be written as follows.

Annual combined axle load spectra for a base year is equal to the sum of average daily volume of truck class I from 1 to the number of truck classes times representative axle load spectrum for truck class i times 365.
(4)

Where:

ALS year
= (Annual) combined axle load spectra (for single, tandem, triple axles) for a base year.
i = Truck class number.
n = Number of truck classes.
ADTVi = Average daily volume of truck class i.
RTSi = Representative axle load spectrum for truck class i.
365 = Constant to convert daily traffic to annual traffic.

To obtain annual axle load spectra for the entire pavement design period, the annual axle load spectra estimated for the base year are factored to account for traffic that will occur during the entire pavement design period.

Equation 4 is an abbreviated expression intended to demonstrate the similarity between the use of ESALs and TFs on the one hand, and the use of axle load spectra on the other. The variable ALSyear represents three separate arrays for single, tandem, and triple axle spectra. (There is no array for quadruple axles\ spectrum because the LTPP traffic data do not contain any data for quadruple axles.) RTSi represents three arrays for each truck class i, as shown in equation 5.

Representative axle load spectrum for truck class i is equal to adjusted normalized axle load spectrum for single axles, or tandem axles, or triple axles.
(5)

The Si, Di, and Ti are arrays for single, tandem, and triple axle spectra, respectively. They are defined for each truck class i according to equations 6, 7, and 8.

 
Si = siai
(6)
 
Di = dibi
(7)
 
Ti = tici
(8)

Where:

Si

= Adjusted normalized axle load spectrum for single axles (an array containing the number of single axles belonging to each of the pre–defined load categories for single axles).

si

= Array containing normalized single axle load spectra for vehicles class i.
ai
= Single–axle coefficient (number of single axles per vehicle) for vehicle class i.

Di, Ti

= Adjusted normalized axle load spectra for tandem and triple axles, respectively.
di, ti
= Arrays containing normalized tandem and triple axle load spectra, respectively.

bi, ci

= Tandem and triple axle coefficients, respectively.

The adjusted normalized load spectra used in arrays Si, Di, and Ti are expressed as the product of normalized spectra and axle–per–class coefficients. The normalized spectra contain proportions of axle loads that occur within designated load ranges. The axle–per–class coefficients are required to obtain actual axle load spectra from the normalized spectra. The calculation procedure embodied in equations 3 through 8 corresponds to the calculation procedure developed for Category 3 and 4 sites in Phase 1.[1]

The use of axle load spectra as the basic traffic characteristics for pavement design is relatively new and presents several challenges:

  • Axle load spectra are large. They consist of many values representing axle load distributions for single, tandem, triple, and quadruple axles for different vehicle classes. Because of the voluminous nature of axle load spectra, their manipulation requires the development of computerized procedures.
  • There is little information available on the characteristics of axle load spectra. A novice user of axle load spectra would have difficulties judging whether the spectra reported for different vehicle classes or for the total traffic flow are reasonable. Guidelines on the typical spectra for different vehicle classes on different highway classes are required for QA purposes and for promoting confidence in the use of axle load spectra. The guidelines may include information on the location of typical peaks and valleys of the axle load distribution, and on the temporal and spatial variability of spectra.
  • The easiest way to compare spectra rapidly and to evaluate their reasonableness and interpret their meaning is by using graphical displays of spectra and summary characteristics such as ESALs. This also requires the development of computerized procedures.

The PLG is intended to overcome these challenges to computation and data comparison. The PLG will also contain a catalogue of typical benchmark values and characteristics of axle load spectra, as well as software to display, compare, and combine axle load spectra, and to calculate annual cumulative axle load spectra.

Functionality of PLG

This section contains a detailed description of the main functions and features of the proposed PLG. The functionality of the PLG and its design parameters have been formulated to meet the objectives of the PLG presented previously. Briefly, the main purpose of the PLG is to facilitate traffic loading projections for LTPP sites without site–specific traffic data. To meet the objectives of the PLG, the PLG will support the following main functions:

  • Database management, including:
    • Data storage.
    • Selection, sorting, and retrieval.
    • Importing and exporting data.
  • Data comparison and assessment.
  • Development of pavement loading estimates for LTPP sites in Categories 3 and 4.
  • Development of combined axle load spectra and cumulative traffic estimates.

An overview of the PLG functions is presented in figure 31.

 

Click to view alternative text

Figure 31. Overview of PLG functions.

Data Storage

The LTPP databases (IMS and CTDB) contain all LTPP traffic data submitted by participating agencies that have passed basic data QA checks. During the course of this project, large amounts of traffic data stored in the IMS were identified by the project team and by the participating agencies as being of dubious quality. Consequently, not all IMS and CTDB data should be part of the PLG, as the PLG data will be used for the estimation of traffic loads at other LTPP sites and need to be of high quality. All traffic data stored in the LTPP databases should be verified using a comprehensive QA process that would identify and remove dubious and nonsensical data. In the absence of such QA process, the only data that should be included in the PLG database at present (that is, if the development of the PLG were to commence prior to the proposed comprehensive traffic data QA process) are those that were used to produce traffic load projections assigned the acceptable projection confidence code. The inclusion of only verified or "acceptable" data in the PLG is one of the main distinctions between the PLG and the DataPave software.[18]

The data stored in the PLG will be of two origins:

  • LTPP data obtained or derived from IMS and CTDB databases.
  • User–supplied data.

Regardless of the origin, the data stored in the proposed PLG will be of four types: supporting data, monitoring traffic data, projected traffic data, and generic or typical traffic data.

Variables that will be stored in the PLG to support its functionality are described below and summarized in table 17.

Supporting Data

Supporting data are non–traffic data required for the identification and retrieval of traffic data. Supporting data stored in the PLG will include:

  • Site identification data.
  • Data describing the physical properties of the site: pavement structure, number of lanes, direction of travel, etc.
  • Highway functional class.
Monitoring Traffic Data

Monitoring traffic data are data derived from AVC and WIM measurements, and are typically annual data reported for a specific year. Monitoring traffic data stored in the PLG will include:

  • Annual truck volume distribution into the 10 classes.
  • Annual axle load spectra for all trucks combined (normalized spectra plus axle counts).
  • Annual axle load spectra for individual truck classes (normalized spectra plus axle counts).
  • Annual TFs for all trucks combined.
  • Annual TFs for individual truck classes.
  • Annual axles–per–truck coefficients for single, tandem, and triple axle types, and for individual truck classes.
  • Truck percentage.
  • Annual AADT truck volumes.

Table 17. List of main variables stored in PLG.

FHWA Highway Functional Type Truck Traffic Classification (TTC) of Highways FHWA Vehicle Classes 4 through 13 Classes 4 through 13 Combined
Class 4 Classes 5, 6, 7–13

Rural Interstates

TTC 1

  • Axle load spectra for single and tandem axles (normalized).
  • Axle per vehicle class coefficients.
  • Load spectra coefficients.
  • ESAL/truck.

Same for other vehicle classes (including triple and quad axle groups, if appropriate)

  • Truck volume distribution (normalized).
  • Annual load spectra for single and tandem axles (normalized).
  • AADT volume.
  • Truck percentage.
  • Load spectra coefficients.
  • ESAL/truck.

TTC 2 through TTC 17

Repeated for other TTC categories

Repeated for other highway functional types

Urban Collector

TTC 1 through TTC 17

[As Above]

[As Above]

[As Above]

Projected Traffic Data

Projected traffic data are data derived from monitoring data. Projected traffic data stored in the PLG will include:

  • Base annual truck distribution.
  • Base annual axle load spectra for all trucks (normalized spectra plus axle counts).
  • Base annual axle load spectra for individual trucks (normalized spectra plus axle counts).
  • Base annual axles–per–truck coefficients for single, tandem, and triple axle types and for individual truck classes.
  • TF for base axle load spectra for all trucks.
  • TF for base axle load spectra for individual trucks.
  • Truck growth rate, typical for the past 5 years.
Generic Traffic Data

Generic traffic data are typical or default traffic data. Unlike monitoring and projected traffic data that are tied to the specific LTPP and other traffic sites, generic data will be provided only for highway functional classes. For the purposes of the PLG, the 11 highway functional classes (6 rural and 5 urban) currently used by the LTPP may need to be further subdivided.

Generic traffic data items stored in the PLG will be similar to the projected traffic data items listed above. However, generic traffic data will not be "base" data but "typical" data.

Notes on Data Storage

The following notes apply to monitoring, projected, and generic traffic data stored in the proposed PLG:

  • To facilitate comparisons and mathematical operations, all truck class distributions and axle load spectra will be stored as normalized distributions or spectra. Normalized axle load spectra will be accompanied by axle–per–class coefficients.
  • TFs should be calculated using the AASHTO load equivalency factors.[17]
  • Data on monthly variation in traffic loads may also be stored in the PLG. However, substantial research effort will be required to identify monthly variation in traffic loads using LTPP traffic data.
Selection, Sorting, and Retrieval

In many respects, the selection, sorting, and retrieval functions of the PLG will be similar to corresponding functions in DataPave.[18] In view of the popularity of DataPave software, and of similar needs to select, sort, and retrieve LTPP data, the user interface of the PLG software and DataPave will be similar. This will make it easier for users to become proficient in using either software package.

Basic site and data selection, sorting, and display functions will include the following modules:

  • Section Selection Module. Selection of any number of LTPP traffic sites from the PLG database, based on user–selected filtering and sorting criteria such as participating agency, highway functional class, pavement type, and experiment type.
  • Select by Map Module. Selection of any number of LTPP traffic sites from the database using an interactive map option.
  • View by Map Module. Selection of traffic data from the PLG database using interactive selection of sites from a map display combined with additional filtering/sorting criteria.
  • Presentation Module. Presentation of detailed traffic data (e.g., annual axle load spectra for the individual vehicle types) for single sites in tabular and graphical formats.
  • Data Extraction and Retrieval Module. Retrieval of traffic data from the PLG database using a variety of filtering and sorting criteria.
Importing and Exporting Data

To supplement the resident PLG data with the user–supplied data, the user will have the option to import additional data into the PLG. User–supplied data may be particularly useful if site–specific LTPP data are missing. For example, some highway agencies have reported monitoring axle weight data that were collected on the LTPP sites during time periods shorter than 24 consecutive hours. Such data have not been included in the LTPP database and, consequently, will not be transferred from the LTPP to the PLG. Yet the use of this type of site–specific data is probably preferable to the use of site–related or regional data.

The storage of user–supplied data may be temporary or permanent. The permanent storage of user–supplied data has the advantage of customizing the PLG with local data that can be used in the future. The PLG database will distinguish between the LTPP and user–supplied data.

Appropriate functions will be developed to export data stored in the PLG, as well as traffic projections developed through the PLG, for subsequent use.

Data Comparison and Assessment

The comparison and assessment function facilitates cross–comparison and evaluation of axle load spectra, and other traffic data. It builds on the PLG's data–storage, selection, sorting, and retrieval functions. Whereas the retrieval function will typically display data for one section at the time, the comparison and assessment function will display data for several sections simultaneously. The comparison and assessment function is unique to the PLG, and will enhance the traffic data assessment and projection by helping the user to:

  • Compare the measured or projected site–specific truck class and axle load distributions with expected or typical distributions or spectra.
  • Select surrogate truck class or axle load distributions for sites without the site–specific distributions (Category 3 and 4 sites).
  • Obtain a better understanding of the spatial and temporal variability of traffic data characterized by truck class and axle load distributions.

The comparison and assessment function should include the following capabilities:

  • The option to provide graphical and tabular displays of several data sets at the same time. The typical data sets to be displayed will include monitoring and projected truck class and axle load distributions.
  • The ability to display multiple data sets of different origins (LTPP, generic, and user–supplied) and of different types (monitoring, projected, and generic). For example, one could compare axle load spectra for vehicle Class 9 measured on an LTPP site in 1994 with the generic axle load spectra for Class 9 vehicles that are characteristic of rural interstates.
  • The option to display statistical measures such as means, ranges, and standard deviations for selected traffic data sets.
  • The option of displaying not only multiple data sets, but also multiple data types. For example, in addition to displaying axle load spectra in a graphical format, the screen would also display, at the same time, the corresponding TFs.

Many features of the data–selection, sorting, retrieval, and comparison and assessment functions were implemented in the prototype PLG demonstration software.

Development of Pavement Loading Estimates for LTPP Sites in Categories 3 and 4

The PLG will contain guidelines for developing missing truck class distributions (required for Category 4 sites) and axle load spectra for individual truck classes (required for Category 3 and 4 sites), and a mechanism for calculating base annual and cumulative axle load spectra.

Developing Truck Class Distribution

The user will have several options to develop truck class distributions required for Category 4 sites: direct input of truck class distributions, use of truck class distributions for the selected LTPP sites, and use of generic truck class distributions.

Direct Input of Truck Class Distribution –For some LTPP sites, additional truck class distribution data (i.e., data not included in the LTPP traffic database) may be available. Also, in some situations, it may be preferable to modify truck distributions developed by analyzing truck distributions on similar sites or suggested by the generic truck class distribution data. Table 18 provides an example of truck class distribution and contains a provision to accommodate an additional truck type designated as "special."

Table 18. Example of truck class distribution.

FHWA Vehicle Class Number

4

5

6

7

8

9

10

11

12

13

Special

Fraction of Commercial Vehicles

0.01

0.11

0.06

0.00

0.10

0.65

0.01

0.03

0.02

0.01

0.00

Truck Class Distributions for Selected Sites–The estimation of a missing truck class distribution would typically utilize site–related data or data on similar sites in the same agency. The user will be able to employ the comparison and assessment function to display the data for the selected sites graphically, and to calculate the means of the selected truck class distributions.

Generic Truck Class Distribution–The use of generic truck class distributions is an option if there are no suitable surrogate truck class distribution data from other (nearby or regional) sites. Generic truck class distributions will be provided for all highway functional classes.

Development of Axle Load Spectra for Individual Truck Classes

The user will have three options to develop axle load spectra for the individual truck classes required for Category 3 and 4 sites: direct input of axle load spectra, use of axle load spectra for selected sites, and use of generic axle load spectra. The direct input of axle load spectra and the use of axle load spectra for selected sites would use similar procedures to select and modify surrogate data as outlined for the selection of truck class distributions. Development of the axle load spectra will be also facilitated by the comparison and assessment function of the PLG.

It is important to allow the user to make adjustments to the truck class and axle load distributions using data from other sites or using generic data. Users may have various bits of information about the composition of traffic stream that could improve the estimates (for example, the proportion of short and long trucks, or the number of buses using an urban highway).

Development of Combined Axle Load Spectra and Cumulative Traffic Load Estimates

The calculation of the projected traffic data, such as the annual base spectra or cumulative axle load estimates, can be accomplished by using the existing LTPP traffic projection procedures. However, before exporting the projected traffic data from the PLG for subsequent use, the user may want to see the overall results of the traffic projection process expressed in terms of ESALs, and have the option of carrying out sensitivity analysis by quantifying the consequences of selecting different truck class and axle load distributions.

In addition, the knowledge of the overall traffic projection results (for example the average TF or the annual number of ESALs) will provide valuable feedback to the process of selecting truck class and axle load distributions, and provide additional assurance to the analyst that the traffic projection results are sound. Consequently, the option to calculate the projected traffic data within the PLG is required and should be implemented.

Development of Generic Truck Characteristics and Data–Selection Guidelines

Extensive additional statistical and engineering analyses are needed to develop generic truck characteristics (in terms of truck class and axle load distributions), and guidelines for the use of the generic and other non–site specific truck characteristics in the projection procedure. These analyses will use LTPP and other traffic data.

The main features of this effort may include the development of: a new highway classification system; generic truck class distributions and generic axle load spectra for individual truck classes; and guidelines for judicious selection of missing truck class and axle load distributions for Category 3 and 4 sites.

Highway Classification System

Generic traffic data will be defined for typical highway functional classes. At present, the LTPP sites are classified into the six rural highway functional classes and five urban functional classes listed below:

  • Rural Principal Arterial–Interstate.
  • Rural Principal Arterial–Other.
  • Rural Minor Arterial.
  • Rural Major Collector.
  • Rural Minor Collector.
  • Rural Local Collector.
  • Urban Principal Arterial–Interstate.
  • Urban Principal Arterial–Other Freeways or Expressways.
  • Urban Other Principal Arterial.
  • Urban Minor Arterial.
  • Urban Collector.

These classes are broad and have a universal applicability. However, they may not be specific enough for the projection of traffic loads. The development of LTPP functional classes for the projection of traffic should be based on the assessment of commonalities in truck class and axle load distributions among LTPP sites. Consideration should also be given to the development of regional highway functional classes.

The number of functional classes for which the generic traffic data will need to be developed may be different from the existing LTPP functional highway classes. For example, the 2002 Pavement Design Guide is expected to use 17 distinct highway classification groups, called Truck Traffic Classifications (TTC):

  • TTC 1: Major Single–Trailer Truck Route (Type I).
  • TTC 2: Major Single–Trailer Truck Route (Type II).
  • TTC 3: Major Single– and Multi–Trailer Truck Route (Type I).
  • TTC 4: Major Single–Trailer Truck Route (Type III).
  • TTC 5: Major Single– and Multi–Trailer Truck Route (Type II).
  • TTC 6: Intermediate Light and Single–Trailer Truck Route (I).
  • TTC 7: Major Mixed Trailer Truck Route (Type I).
  • TTC 8: Major Multi–Trailer Truck Route (Type I).
  • TTC 9: Intermediate Light and Single–Trailer Truck Route (II).
  • TTC 10: Major Mixed Trailer Truck Route (Type II).
  • TTC 11: Major Multi–Trailer Truck Route (Type II).
  • TTC 12: Intermediate Light and Single–Trailer Truck Route (III).
  • TTC 13: Major Mixed Trailer Truck Route (Type III).
  • TTC 14: Major Light Truck Route (Type I).
  • TTC 15: Major Light Truck Route (Type II).
  • TTC 16: Major Light and Multi–Trailer Truck Route.
  • TTC 17: Major Bus Route.
Generic Truck Class and Axle Load Distributions

The development of generic truck class distributions should start with the evaluation of the truck class distributions on all LTPP sections to identify commonalities in the distribution for various highway functional categories.

The development of generic axle load spectra may draw on the parallel between TFs (number of ESALs per truck) and the axle load spectra. The relationship between TFs and axle load spectra for individual truck classes was discussed in the section titled "Conceptual outline of PLG."

There is a body of knowledge dealing with truck axle loads in terms of TFs. Several highway agencies have evaluated the temporal and spatial variation of TFs, and many highway agencies use generic and other TFs in the pavement design process.[19,20] The variation in TFs is to some degree indicative of the variation in the corresponding axle load spectra. However, our current understanding of the expected axle load spectra on different class highways is limited and will need to be developed by an in–depth analysis of LTPP and other traffic data.

The understanding of the variation of truck class and axle load distributions will affect the division of the highway network into highway functional classes. Consequently, the enhancement of the existing LTPP highway functional classes, or the development of new classes, and the development of generic truck class and axle load distributions will need to be interactive.

Guidelines for Data Selection

Guidelines for selecting surrogate truck class and axle load distributions for Category 3 and 4 sites will help a user who is estimating missing data. The guidelines will use new information on the characteristics of truck flows (expected or generic truck class and axle load distributions on characteristic highway links) and the PLG comparison and assessment functions.

Prototype Demonstration Software

To cope with the large amount of traffic data, the proposed PLG will function as a stand–alone software product operating as a relational database with many additional built–in computational and reporting features. Prototype software was developed to illustrate the operation of the PLG.

The development of final PLG software will require, in addition to the programming effort, considerable engineering and analytical effort to develop generic truck class and axle load distributions and provide guidelines for their use. For this reason, the prototype PLG software demonstrated only data–management, comparison, and assessment functions.

Example Use of PLG

This section contains two examples of estimating annual axle loads for all in–service years for LTPP sites using the proposed PLG. One is for Category 3 sites that lack axle load data, the other for Category 4 sites that lack axle load data and truck class distribution data. These examples:

  • Describe the traffic projection procedure for sites lacking site–specific data (Category 3 and 4 sites) using realistic examples.
  • Illustrate how the proposed PLG can facilitate and improve the projection process and the functionality of the PLG.
  • Estimate the consequences of using surrogate traffic data instead of site–specific data.

These two examples use the LTPP traffic projection procedure outlined in chapter 2 of this report (and documented in the Phase 1 report), and they assume the existence of the proposed PLG.[1] To estimate the consequences of using surrogate data, the sites selected for the examples actually had site–specific monitoring truck class and axle load distribution data; however, it was assumed that the site–specific data were missing and had to be estimated using surrogate data. Thus, it was possible to compare traffic loads estimated using surrogate data with traffic loads based on site–specific monitoring data. The comparison was done in terms of the annual axle load spectra and the cumulative number of ESALs.

Use of PLG for Category 3 Sites

The example of traffic projection for Category 3 sites was based on California LTPP site 068150, located on a two–lane rural minor arterial highway east of Los Angeles, CA. It was assumed that the site had truck class distribution data (annual number of trucks that belong to different vehicle classes) but lacked axle load data. The task was to develop (select) surrogate axle load spectra for individual truck classes and combine them with the known (site–specific) number and type of trucks to calculate annual axle load spectra.

The PLG database can be queried to facilitate the selection of axle load data. This can be done by searching the database for similar sites in California (for example, all sites on rural minor or principal arterial highways) or in other jurisdictions. The search of the prototype PLG database identified two similar California sites with monitoring axle load data (site 062040 located on a two–lane rural principal arterial highway east of San Francisco, CA, and site 066044 located on a four–lane principal arterial highway south of Eureka, CA). The truck class distributions on the three California sites are compared in figure 32, which was produced by the prototype PLG software; it indicates that the truck distributions on the three sites are similar. While the similarity of truck class distributions does not mean the similarity in axle loads, it does provide assurance that the sites serve similar traffic flows and are on highways with similar functional classification.

Click to view alternative text

Figure 32. Comparison of truck class distributions for sites 062040, 066044, and 068150.

The required surrogate axle load data are needed as axle load distributions (spectra) for individual truck types. The prototype PLG database was used to compare axle load spectra on nearby sites. An example of the spectra comparison for sites 062040 and 066044 is provided in figures 33 and 34 for single and tandem axles for Class 9 vehicles (5–axle single–trailer trucks), respectively. Also shown in figures 33 and 34 is the surrogate (computed) axle load spectrum for site 068150 obtained as the mean spectrum for sites 062040 and 066044. It should be noted that figures 33 and 34 show the example of axle load spectra for only 1 of the 10 truck types for which the surrogate spectra are required.

The process of assessing available data and selecting surrogate data would benefit from the comparison of the selected spectra (shown in figures 33 and 34) with generic or typical spectra. However, the generic spectra are not available in the prototype PLG.

Click to view alternative text

Figure 33. Comparison of single axle load distributions for vehicle Class 9
for sites 062040 and 066044 with computed mean distribution.

Click to view alternative text

Figure 34. Comparison of tandem axle load distributions for vehicle
Class 9 for sites 062040 and 066044 with computed mean distribution.

The annual axle load spectrum for site 068150 was obtained by combining the surrogate axle load spectra for the individual truck classes (expressed as normalized spectra) and the site–specific (classified) annual truck volumes. The resulting estimated annual spectrum for site 068150 is presented in figure 35. The spectrum was obtained by calculations done outside the prototype PLG. Such calculation should be done by the PLG, and the PLG can then be used to compare the calculated annual spectra using surrogate data with annual spectra obtained for nearby sites or with typical spectra. It should be noted that the site–specific spectrum for site 068150 would not exist for Category 3 sites. This is the spectrum that is to be estimated using surrogate data. It has been included in figure 35 for comparison purposes.

The comparison of annual axle load spectra for site 068150 in figure 35 indicates that the actual monitoring spectrum and the estimated surrogate spectrum are quite similar. For example, the three peaks of the single axle load distribution are duplicated quite well. The annual axle load spectrum for site 068150 in figure 35 is for all vehicle classes combined and represents, using the LTPP projection terminology, the base annual spectrum, or surrogate base annual spectrum.

To quantify the consequences of using the surrogate base annual spectrum, the surrogate base annual spectrum was used to calculate the number of cumulative ESALs. The annual base spectrum was combined with the projected growth factor (established previously as part of traffic load projection for site 068150) to obtain annual axle load spectra, and then the annual spectra were expressed in terms of ESALs. The result of this calculation is presented in figure 36, which also shows the corresponding results obtained for monitoring site–specific axle load spectra (labelled Site–Specific Category 2). Figure 36 include a "Projected Annual ESALs" sheet (which was part of site–specific reports described in chapter 2) for site 068150 adapted to also display ESALs estimated using surrogate data.

As already indicated by the similarity of the base annual spectra for Class 9 vehicles (figures 33 and 34), ESALs are also similar. The cumulative number of ESALs using surrogate spectra was 2.06 million, while the corresponding number of ESALs for site–specific spectra was 2.2 million (figure 36).

Use of PLG for Category 4 Sites

Arizona site 041017, located in the northbound direction on a four–lane rural interstate south of Tucson, AZ, was used to illustrate the use of the PLG to obtain traffic projection for Category 4 sites. Even though site–specific axle load data were available for this site, it was assumed that these data were not available and that both truck class and axle load distributions had to be estimated. It was assumed that the only available information about truck loads on this site were AADT truck volumes (assumed to be 540 in 1998).

The site selected to obtain surrogate data for the subject site was another Arizona site, 041007, located in the westbound direction of a rural interstate west of Phoenix, AZ. The 1998 annual AADT truck volume on this site, 3525, was approximately 7 times larger than the corresponding truck volume on the subject site. Thus, the two sites are located on different interstates and carry different truck volumes. While it is possible to use data for other sites, or to use a combination of data obtained for several sites, this example demonstrates the use of a single site.

Click to view alternative text
1 lb. = 2.202 kg

Figure 35. Comparison of site–specific and surrogate base annual spectra for site 068150.


Click to view alternative text

Year Annual ESALs
Historical Monitoring Projected
Site–Specific Spectra Category 2 Surrogate Spectra Category 3

1984

108,789

102,577

1985

63,000

113,118

106,714

1986

67,000

117,675

111,000

1987

67,000

122,388

115,443

1988

96,000

127,274

120,124

1989

81,000

132,356

124,930

1990

83,000

137,657

129,895

1991

83,000

143,154

135,134

1992

171,000

148,921

140,544

1993

473,000

154,887

146,211

1994

389,000

161,068

152,068

1995

176,297

167,491

158,168

1996

177,884

174,487

164,528

1997

169,527

181,354

171,104

1998

188,601

177,975

Cumulative ESALs

2,179,220

2,056,415

Figure 36. Comparison of projected, historical, and monitoring annual ESALs for site 068150.

The truck class distribution for the two sites is compared in figure 37. The data for the subject site shown in figure 37 would not exist for a Category 4 site and are included for comparison purposes only. To determine whether the selected surrogate truck distribution is typical, the PLG will contain not only truck class distributions for other Arizona sites, but also generic or typical truck class distributions for typical highway functional classes.

Click to view alternative text

Figure 37. Comparison of truck class distributions for sites 041007 and 041017.

The base annual spectrum for site 041017 was obtained by combining the site–specific annual truck volumes with surrogate truck classification and axle load spectra (for individual vehicle types) obtained from site 041007. The resulting surrogate base annual axle load spectrum is compared with the spectrum obtained using site–specific monitoring axle load data in figure 38.

The projected annual spectra for all in–service years are compared, in terms of ESALs, in figure 39. Figure 39 is a "Projected Annual ESALs" sheet for site 041017 also adapted to display ESALs estimated using surrogate data. The results in figures 38 and 39 show a very good agreement between traffic loads estimated using surrogate data and site–specific data. For example, the number of ESALs estimated using site–specific axle load data was 2.1 million, while the number of ESALs estimated using surrogate data was 2.9 million.

Example Summary

The two examples provided in this section indicate that reasonable traffic loading estimates can be obtained by judiciously selecting surrogate data (in the absence of site–specific truck class and axle load data). The relatively close agreement between the site–specific and surrogate projections may leave a false impression that the site–specific axle load data are not important because they can be estimated using surrogate data. While the judicious selection of surrogate data is important, and the proposed PLG can facilitate and guide the selection, surrogate data can never replace site–specific data. Also, the greater the amount of site–specific data, the easier it is to develop appropriate surrogate data.

Click to view alternative text

1 lb = 2.202 kg

Figure 38. Comparison of site–specific and surrogate base annual spectra for 041017.

Click to view alternative text

Year Annual ESAL
Historical Monitoring Projected
Site–Specific Spectra Category 2 Surrogate Spectra Category 4

1976

948,000

58,254

80,142

1977

1,089,000

60,593

83,384

1978

806,000

63,011

86,725

1979

792,000

65,514

90,192

1980

849,000

68,209

93,802

1981

821,000

70,925

97,594

1982

792,000

73,740

101,483

1983

821,000

76,717

105,564

1984

877,000

79,785

109,804

1985

962,000

82,964

114,213

1986

1,075,000

86,276

118,801

1987

1,160,000

89,785

123,561

1988

1,556,000

93,361

128,539

1989

1,556,000

97,104

133,668

1990

1,273,000

100,986

139,034

1991

230,000

105,026

144,581

1992

240,000

109,220

150,414

1993

102,230

113,585

156,450

1994

150,000

118,142

162,729

1995

150,000

122,857

169,267

1996

150,000

127,755

176,025

1997

132,850

133,030

183,091

1998

138,279

190,454

Cumulative
2,135,118
2,939,517

Figure 39. Comparison of projected, historical, and monitoring annual ESALs for site 041017.

Previous | Table of Contents | Next

 

Federal Highway Administration | 1200 New Jersey Avenue, SE | Washington, DC 20590 | 202-366-4000
Turner-Fairbank Highway Research Center | 6300 Georgetown Pike | McLean, VA | 22101