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
Back to Publication List        
Publication Number:  FHWA-HRT-13-091    Date:  November 2014
Publication Number: FHWA-HRT-13-091
Date: November 2014

 

Verification, Refinement, and Applicability of Long-Term Pavement Performance Vehicle Classification Rules

Chapter 6. Sensitivity of Pavement Design Models to Differences in Classification of Class 5 and Class 8 Vehicles

Background

From the analysis of different vehicle classification rule sets, it was found that variations in vehicle classification rules frequently affected two FHWA vehicle classes: Classes 5 and 8. All AVC-based vehicle algorithms and many WIM-based classification rule sets classify pickup trucks, sport utility vehicles (SUV), and vans (FHWA Class 3) pulling a light trailer as FHWA Class 8 based on the total number of axles. Many of the same algorithms classify large SUVs, vans, or large pickup trucks from FHWA Class 3 as FHWA Class 5 based on axle spacing tolerances. This can result in potentially large increases in estimated truck volumes.

Unlike the classification rule sets described above, the LTPP classification rule set uses GVW and axle weight in addition to the number of axles and axle spacing to classify these types of vehicles. By using weight as an additional parameter, the LTPP rule set eliminates lightweight pickups, vans, SUVs, and pickups pulling light trailers from counts for Classes 5 and 8. This results in 1) lower truck volume estimates (AADTT), 2) lower percentages of Class 5 and 8 vehicles in VCD, and 3) heavier axle load spectra for Class 5 and 8 vehicles.

For MEPDG pavement design purposes, no problem results if 1) the same classification rule set is used to obtain both volume by class and axle loading by class data and 2) both data types are collected at the same location (MEPDG Level 1 traffic input) or data are collected at different locations that have similar percentages of lightweight vehicles mixed in Class 5 and Class 8 volume counts. However, a potential error could be introduced in pavement design when truck volume and VCD collected using a non-LTPP classification rule set are combined with load spectra developed based on data collected using the LTPP classification rule set. Similar errors could be introduced when load spectra collected at a location that has a high percentage of lightweight vehicles classified as Class 5 or 8 are applied to a location that has a low percentage of lightweight Class 5 or 8 vehicles, even if the same vehicle classification rule set is used to collect both vehicle class by volume and axle load data.

Part 1 of this report stated that, for LTPP test sites that have lower annual loading rates and less significant Class 9 truck volumes, the use of the LTPP TPF load spectra with volume by class data collected using a non-LTPP rule set can result in significant errors in the estimated traffic load. In general, the greater the percentage of total traffic load coming from truck Classes 5 and 8, the greater the potential error could be. The report also suggests that it is important to determine whether the specific State classification rule set creates significantly different truck volumes in truck Classes 5 and 8 than the LTPP rule set when applied to the same traffic stream. The analysis presented in this chapter aims to quantify what "significant" means from a pavement design perspective.

Analysis Objective

In this analysis, two scenarios were investigated. In one scenario, a large number of Class 3 pickup trucks, vans, and SUVs that belong to FHWA Class 3 are misclassified as Class 5 by non-LTPP classification rule sets, resulting in an increase in Class 5 volume and, consequently, an increase in the total AADTT. The other scenario is the effect of an increase in Class 8 volumes when non-LTPP classification rule sets are used. In this case, Class 3 pickup trucks, SUVs, or vans pulling a trailer are classified as Class 8, resulting in an increase in Class 8 volumes and an increase in total AADTT. In both cases, vehicles that were not previously classified as trucks (FHWA Classes 4 through 13) are added to total truck volume.

The objective of this analysis was to check whether these classification errors would cause practical differences in pavement design outcomes using MEPDG Level 2 or 3 traffic loading inputs based on the LTPP vehicle classification rules. Both the MEPDG and AASHTO design models were considered in this analysis.

Analysis Approach

Traffic Loading Scenarios

Two road functional classes were considered in the analysis: RIs and ROPAs. The AADTT values and representative vehicle class distributions for these road types were selected based on averages computed from all SPS and GPS sites. For design of flexible and rigid pavements representing RIs, an AADTT of 2,000 trucks and a VCD equivalent to MEPDG Truck Traffic Classification (TTC) 1 default were used. For design of flexible and rigid pavements representing ROPAs, two scenarios were selected, one for flexible pavements with an AADTT of 500 trucks and a VCD equivalent to MEPDG TTC 6, and one for rigid pavements with an AADTT of 700 and a VCD equivalent to MEPDG TTC 2. Table 21 shows the AADTT values and the normalized VCDs selected for analysis. These values were used to develop base pavement design scenarios for the analysis.

Table 21. AADTT and VCDs used for MEPDG sensitivity analysis.

Class

RI Flexible and Rigid

ROPA Flexible

ROPA Rigid

AADTT = 2,000, TTC 1

(percent)

AADTT = 500, TTC 6

(percent)

AADTT = 700, TTC 2

(percent)

Class 4

1.3

2.80

2.40

Class 5

8.5

31

14.10

Class 6

2.8

7.30

4.50

Class 7

0.3

0.80

0.70

Class 8

7.6

9.30

7.90

Class 9

74.0

44.80

66.30

Class 10

1.2

2.30

1.40

Class 11

3.4

1

2.20

Class 12

0.6

0.40

0.30

Class 13

0.3

0.30

0.20

RI = Rural Interstate

ROPA = Rural Other Principal Arterial

AADTT = Annual Average Daily Truck Traffic

The NALS for this study were developed based on the average of values obtained from representative NALS computed for the 26SPS TPF sites. Load spectra development was done as part of another LTPP traffic data analysis study currently underway, where the objective is to develop new default traffic datasets for the MEPDG using the SPS traffic data from the TPF. These load spectra are provided in the Task Order 12 Interim Report database appendix.(3)

Representative Pavement Structures

The pavement structures were designed following the MEPDG(1) recommendations for selected road functional class and traffic volume. Table 22 summarizes critical pavement structure and material properties for flexible pavements, while table 23 provides a summary for rigid pavements.

Table 22. Flexible pavement reference designs.

Property

RI

All Climates

ROPA

Wet-Freeze

ROPA

All Remaining Climates

AC Thickness (inches)

10.5

8.5

8

Superpave Binder Grade

76-22

70-22

70-22

Base Material, Thickness

Crushed stone, 12 inches

Crushed stone,
12 inches

Crushed stone,
12 inches

Subgrade

A-1-b

A-1-b

A-1-b

AC = Asphalt Concrete

RI = Rural Interstate

ROPA = Rural Other Principal Arterial

Table 23. Rigid pavement reference designs.

Property

RI

All Climates

ROPA

All Climates

PCC Thickness (inches)

10

8

Base Material,

Thickness

Cement stabilized, 6 inches

Soil cement, 6 inches

Layer 3

A-6, 12 inches

A-6, 12 inches

Layer 4

A-6, semi-infinite

A-6, semi-infinite

Erodibility Index

Extremely Resistant (1)

Erosion Resistant (3)

Dowel Diameter, Spacing (inches)

1.25, 12

1.25, 12

PCC = Portland Cement Concrete

RI = Rural Interstate

ROPA = Rural Other Principal Arterial

Analysis Execution

To conduct sensitivity analysis of pavement design outcomes to changes in Class 5 volume, additional Class 5 vehicles were added in incremental steps, as discussed below. For each reference design in table 22 and table 23, 1,000 Class 5 trucks were added for RIs and 500 for ROPAs. AADTT and VCD values were then recomputed to account for additional Class 5 volumes. The pavement structure from the base scenario was redesigned for these new traffic inputs using the MEPDG. For simplicity, only the surface layer thickness was changed, and the new thickness value was compared with the reference base design. If the change in thickness was less than 0.5 inches, a new increment of 1,000 or 500 trucks in Class 5 was introduced, and the process was repeated until the change threshold of 0.5 inches in thickness of the top pavement layer was obtained.

The same process was applied to both road classes and pavement types in all four LTPP climate zones. Later, the same approach was applied to study MEPDG sensitivity to changes in Class 8 volumes.

Discussion of Findings From MEPDG Analysis

Table 24 and table 25 provide the results from the MEPDG sensitivity analysis for Class 5 and Class 8, respectively. Both tables provide the original volume of Class 5 and Class 8 and the new volume (simulating potential misclassification) needed to generate the need for a 0.5-inch increase in design thickness of the surface layer. As can be seen, the additional volumes of Class 5 or 8 vehicles needed to require significantly different pavement designs for RIs are higher than those needed for a ROPA. There are minor but consistent variations in additional class volume for both classes resulting from different climatic conditions for all pavement types and functional classes. However, these minor variations are not practically significant.

Table 24. Original and new Class 5 volume leading to a 0.5-inch difference in design thickness using the MEPDG.

Pavement Type

Functional Class

Original Class 5 Volume

New Class 5 Volume

WF

WNF

DF

DNF

Rigid

RI

170

> 10,000

> 10,000

> 10,000

> 10,000

ROPA

99

1,620

1,690

1,300

1,180

Flexible

RI

170

7,050

8,070

7,400

7,130

ROPA

155

2460

2,480

2,370

2,250

RI = Rural Interstate

ROPA = Rural Other Principal Arterial

WF = Wet-Freeze

WNF = Wet-No Freeze

DF = Dry-Freeze

DNF = Dry-No Freeze

Table 25. Original and new Class 8 Volume leading to a 0.5-inch difference in design thickness using the MEPDG.

Pavement Type

Functional Class

Original Class 8 Volume

New Class 8 Volume

WF

WNF

DF

DNF

Rigid

RI

152

1,300

1,120

890

1,020

ROPA

40

440

450

440

350

Flexible

RI

152

2,600

2,680

2,920

2,550

ROPA

47

510

460

320

300

RI = Rural Interstate

ROPA = Rural Other Principal Arterial

WF = Wet-Freeze

WNF = Wet-No Freeze

DF = Dry-Freeze

DNF = Dry-No Freeze

Findings From the Class 5 Sensitivity Analysis

For the case of Class 5 sensitivity for RI designs, it was found that an increase in Class 5 volume of more than 7,000 vehicles (flexible design) or more than 10,000 vehicles (rigid design) per day from the original 170 vehicles is required to result in any significant difference in the design thickness. These additional Class 5 volumes would result in AADTT per lane values of 9,000 to 12,000, which are unlikely to be observed in the field. It also would mean that the percentage of Class 5 vehicles in the VCD would be about 80 percent, which also is unlikely for RIs. Therefore, the conclusion from the analysis is that misclassification of Class 3 vehicles as Class 5 does not have any practical impact for MEPDG designs for RIs.

For the case of Class 5 sensitivity for ROPA designs, it was found that an increase in Class 5 volume of more than 2,250 vehicles (flexible design) or 1,180 (rigid design) per day from the original 155 (flexible) or 99 (rigid) vehicles is required to produce any significant difference in the design thickness. The resulting AADTT per lane values (2,651 for flexible and 1,725 vehicles for rigid) are possible but unusual for ROPAs. Therefore, the conclusion from the analysis is that misclassification of Class 3 vehicles as Class 5 is not likely to have any practical impact for MEPDG designs for ROPAs if 1) AADTT per lane is less than 2,600 vehicles for flexible design or 1,700 vehicles for rigid design and 2) the percentage of Class 5 vehicles is less than 68 percent for rigid pavements and 78 percent for flexible pavements. For roads with AADTT per lane above these values, the load spectra developed based on the LTPP rule set should be used only if site-specific analysis concludes that the portion of misclassified Class 3 vehicles among all Class 5 vehicles at the site is less than 96 percent for flexible pavements and 87 percent for rigid pavements. From a practical standpoint, the percentage of misclassified Class 3 vehicles among all Class 5 vehicles always will be less than these percentages.

Based on the assessment of the LTPP database (standard data release 24), none of the LTPP sites had AADTT and percentages of Class 5 vehicles at or above these levels.

Findings From Class 8 Sensitivity Analysis

For the case of sensitivity of RI designs to Class 8, it was found that an increase in Class 8 volume of more than 890 (rigid design) or 2,250 vehicles (flexible design) per day from the original 152 vehicles is required to result in any significant difference in the design thickness. These additional Class 8 volumes would result in AADTT per lane values of 2,738 to 4,398. These AADTT per lane values frequently are observed in the field for RIs. It also would mean that the percentage of Class 8 vehicles in the VCD would be about 33 percent for rigid pavements and 58 percent for flexible pavements. These Class 8 percentages are unlikely but possible for RIs, especially in the case where 33 percent is required to produce changes in pavement design for rigid pavements. Therefore, the conclusion from the analysis is that the misclassification of Class 3 vehicles pulling trailers as Class 8 may have practical impacts for MEPDG designs for those RI pavements with AADTT volumes of more than 2,700 for rigid pavements and 4,400 for flexible pavements, but only when the resulting percentage of Class 8 vehicles is more than 33 percent for rigid pavements and 58 percent for flexible pavements. A very small fraction of RI pavements is likely to have these traffic characteristics. For these roads, load spectra developed based on the LTPP rule set should be used only if site-specific analysis concludes that the proportion of lightweight vehicles among all Class 8 vehicles at the site is less than 83 percent for rigid pavements and 94 percent for flexible pavements. Based on an assessment of the LTPP database (standard data release 24), none of the LTPP sites had AADTT and percentages of Class 8 vehicles at or above these levels.

For the case of sensitivity of ROPA designs to Class 8, it was found that an increase in Class 8 volume of more than 300 (flexible design) or 350 vehicles (rigid design) per day from the original 40 to 47 vehicles is required to result in a significant difference in the design thickness. These additional Class 8 volumes would result in AADTT per lane values of 753 to 1,010. These AADTT per lane values frequently are observed in the field. Also, it would mean that the percentage of Class 8 vehicles in the VCD would be about 40 percent for flexible pavements and 35 percent for rigid pavements. These Class 8 percentages are not typical but are possible for ROPAs. Therefore, the conclusion from the analysis is that misclassification of Class 3 vehicles pulling light trailers as Class 8 may have practical impacts for MEPDG designs for ROPAs with AADTT volume of more than 700 for flexible pavements and 1,000 for rigid pavements along with percentages of Class 8 vehicles above 40 percent for flexible pavements and 35 percent for rigid pavements. A very small fraction of ROPAs could possibly have these traffic characteristics. For these roads, load spectra developed based on the LTPP rule set should be used only if site-specific analysis concludes that the portion of lightweight vehicles among all Class 8 vehicles at the site is less than 84 percent for flexible pavement and 89 percent for rigid pavements.

Based on an assessment of the LTPP database (standard data release 24), only three GPS sites had values of AADTT and percentages of Class 8 vehicles at or above these levels and only for one of the monitoring years: flexible and rigid sections 16-1009 and 16-3023 in Idaho and rigid pavement section 6-3010 in California. Of these three sections, only 16-1009 had axle loading data for the same year available in the LTPP database. These data were extracted and analyzed to determine what percentage of Class 8 vehicles is lightweight. The analysis results indicated that only 9 percent of single-axle loads and 10 percent of tandem-axle loads were lightweight (5,000 lb or less for single and 10,000 lb or less for tandem). Moreover, axle load distributions of Class 8 trucks had a shape characteristic of Class 9 vehicles. Based on this data analysis, the following two conclusions were drawn: 1) Class 8 trucks observed at GPS section 16-1009 are not lightweight, and there is no evidence of lightweight vehicles classified as Class 8; and 2) based on the weight and distribution of single and tandem axles, it is possible that Class 9 vehicles are misclassified as Class 8, because unusually low Class 9 volumes are also reported for this site. No additional analysis was possible for GPS sections 16-3023 and 6-3010; however, both of these sections have unusually low percentages of Class 9 vehicles and may have the same potential Class 9-to-8 misclassification issue.

The main finding from these analyses is that virtually no LTPP sites have combinations of AADTT and Class 5 or 8 percentages that would result in practical differences in pavement design outcomes when vehicle classification data from a non-LTPP rule set are combined with weight data based on the LTPP classification rule set.

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