MEPDG Traffic Loading Defaults Derived From Traffic Pooled Fund Study

CHAPTER 10-DESCRIPTION OF THE NEW MEPDG TRAFFIC LOADING DEFAULTS FOR LTPP SITES

OVERVIEW OF DEFAULTS

All the defaults developed in this study are based on SPS TPF WIM data. These defaults include the following:

• NALS
  
  
• APC coefficients.
  
  
• Average wheelbase and axle spacing.

The following two sets of NALS defaults were computed:

• Tier 1: Global defaults based on all applicable SPS TPF data.
  
  
• Tier 2: Supplemental defaults based on subsets of SPS TPF data representing a specific loading condition.

Tier 1 global traffic loading defaults were developed based on averaging of MEPDG traffic loading inputs computed for individual SPS TPF sites and designed to represent global loading condition based on all the averaging of NALS for all SPS TPF sites. This approach is similar to the one used in NCHRP Project 1-37A to develop NALS defaults provided in the NCHRP 1-37A MEPDG software and in DARWin-ME^TM software.⁽³⁾

Tier 2 supplemental traffic loading defaults were developed based on averaging of MEPDG traffic loading inputs computed for SPS TPF sites that had similar axle loading conditions. This computation was done separately for each vehicle class and axle group type. Thus, these defaults provide users with multiple choices with regard to selection of traffic loading conditions for different truck classes and axle group types.

GLOBAL AXLE LOADING DEFAULTS BASED ON LTPP SPS TPF SITES (TIER 1)

NALS defaults were computed based on a simple averaging of RANALS for 26 SPS TPF sites. Appendix D contains global NALS defaults for each vehicle class and axle group type and can be used to generate MEPDG and DARWin-ME^TM axle load spectra input files.

Computed global NALS were compared to the NALS defaults included in the MEPDG software version 1.1. Comparisons of NALS are provided in figure 33 to figure 36. NALS and number of axles per truck were used in combination with TTC1 VCD to compute combined axle load spectra for each axle group type. These combined spectra were normalized and percentile distributions plotted for each axle group type.

The differences in tandem axle load spectra plot are of a particular interest, as tandem axle loads carry over 50 percent of all loads under typical highway loading conditions. As can be seen from the tandem plot (see figure 47), NCHRP 1-37A defaults had a larger number of overloads but lower percentage of heavy legal loads compared to the SPS TPF defaults. For the quad spectra plot, only the new spectra are displayed, as the number of axles per truck coefficients default is set to zero in the original MEPDG defaults (there were not enough quads in the original study to compute these values).

Figure 46. Graph. Comparison of NCHRP 1-37A and SPS TPF default NALS for single axles, all classes combined using TTC1.

Figure 47. Graph. Comparison of NCHRP 1-37A and SPS TPF default NALS for tandem axles, all classes combined using TTC1.

Figure 48. Graph. Comparison of NCHRP 1-37A and SPS TPF default NALS for tridem axles, all classes combined using TTC1.

Figure 49. Graph. Comparison of NCHRP 1-37A and SPS TPF default NALS for quad axles, all classes combined using TTC1.

Conclusions from Comparison of Tier 1 and Original NALS Defaults

Visual comparison of SPS TPF-based tier 1 and original NALS defaults revealed that the newly computed defaults had fewer very light and fewer very heavy loads. This is most likely due to the fact that the new defaults were collected with more consistently calibrated WIM equipment. The better calibration of the WIM scales used to develop the new defaults means that fewer very light loads (caused by undercalibrated scales observing light loads) and fewer very heavy loads (caused by overcalibrated scales observing heavy loads) are observed in the new default database. Because the new NALS have smaller percentages of overloads^[1], pavement life predicted using the new defaults is likely to be longer than using the old defaults. However, these differences are not expected to be dramatic because the new tier 1 defaults also have a higher percentage of legally loaded heavy axles. Assuming that the new defaults are more accurate and representative of typical loading conditions, a conclusion could be made that pavement
designs using the old MEPDG defaults are more conservative, compared to the new defaults. However, from the practical perspective, the differences in the design thickness are not likely to be significant.

AXLE LOADING DEFAULTS REPRESENTING ALTERNATIVE AXLE LOADING CONDITIONS

As shown in chapter 9, MEPDG pavement performance estimates for many pavement design scenarios are sensitive to traffic load inputs. In addition, by analyzing the loading conditions observed in the LTPP SPS TPF study, it is obvious that truck loading characteristics (axle weights) can vary considerably from location to location. It was therefore determined that defaults should be developed that allowed users of the MEPDG and DARWin-ME^TM software to select loading conditions for their analyses that could account for these observed differences in traffic loading per vehicle.

Chapter 8 describes the methodology used by the project team to develop NALS groups representing alternative axle loading conditions for each of FHWA's truck classifications based on the SPS TPF data. This section provides description of the defaults developed using this methodology.

Description of NALS Clusters Representing Different Axle Loading Conditions

Table 40 shows the number and categories by weight of the NALS cluster groups representing different axle loading conditions developed based on the SPS TPF data. Table 24 provides a quantitative definition of the descriptive loading conditions used in table 40, developed based on analysis of SPS TPF data. Cluster groups presented in table 40 are based on the clustering analysis of axle loading characteristics presented by mean RPPIF values and the sensitivity tests of the MEPDG described in chapter 9. NALS for each cluster group are available in appendix D. The following paragraphs provide explanations for the information presented in table 40.

Default and Alternative Loading Conditions Observed Nationally

In table 40, the cluster group that contains the largest number of SPS TPF sites is identified as the default for each vehicle class and axle group type. In this case, "default" means that NALS describes the most frequently observed loading condition for a given vehicle class and axle group type among all SPS TPF sites. NALS representing default load spectra groups are recommended for MEPDG use when no information is available that would indicate that a particular vehicle class that is expected to use that roadway will have a loading condition that is either lighter or heavier than what was typically observed for this vehicle class or truck type based on sites included in SPS TPF study. The default condition is not always a moderate condition. It may be the lightest or heaviest condition observed for a particular class of vehicles and type of axle.

In addition to the default loading cluster, groups representing alternative loading conditions were identified to represent NALS that are either lighter or heavier than the default NALS. The decision whether to develop alternative NALS clusters and the number of alternative NALS clusters to create was made individually for each vehicle class and axle group type based upon the following considerations:

• The likelihood that a given axle group type and vehicle class will represent a significant percentage of total loading based on truck and axle group types and their frequency observed in United States.
  
  
• The potential for the alternative NALS, when used in place of the default load spectra, to produce significantly different MEPDG outcomes.
  
  
• The size of the difference (measured in either RPPIF or in the percentage of heavy axles, with heavy defined as 75 percent of the Federal legal limit or above) in load observed at SPS TPF sites for each type of axle and class of vehicle.
  
  
• The availability of data (the number of sites) representing specific axle loading conditions.

Some types of axles carried by some vehicle classes (e.g., single axles within class 6) have fairly consistent loading patterns (relative to pavement performance) and do not need to be represented by multiple load spectra. Other types of axles within specific classes of vehicles exhibit highly variable loading patterns (e.g., class 7 tandem axles); thus, users of the MEPDG software need to be able to choose between multiple loading conditions to represent expected traffic conditions.

* NALS has very heavy overloads.

**NALS identified as outlier based on classification issue (high percentage of very light weight axles).

Note: Blank cells indicate that no instances were identified.

Loading Conditions Considered Special Cases

In some cases, the analysis of site-specific RANALS also indicated that some States have unique loading conditions that do not cluster with RANALS from other SPS TPF sites. Possible reasons for these unique RANALS include the following:

• Specific commodity hauls and State/local trucking regulations are present that cause unusual truck types and loads unique to that location or State to dominate the loading characteristics of a given vehicle class at the site.
  
  
• Vehicle classification algorithm differences implemented on State-installed SPS TPF WIM sites can create some differences in classification outcomes when compared to the LTPP classification scheme, thus causing some differences in the load spectra for those affected classes at those sites.
  
  
• Axle weight limits implemented by the State are different from Federal limits.
  
  
• The road containing the SPS TPF WIM site may serve commodity movements that result in a larger than usual percentage of trucks operating by permit and thus have axles weighing more than at other sites.
  
  
• State trucking regulations differ. This means that some States encourage specific categories and actively discourage others. Consequently, States with unusual truck types (e.g., a few western States allow triple trailer trucks) can exhibit unusual loading conditions for the FHWA categories which contain those specific truck types. (Triple trailer trucks will be classified as class 13, where their presence will generally increase the number of single axles found on class 13 trucks, and will change the load distribution of class 13 single axles.)

To help illustrate the effects of these site, State, or even regional loading conditions, table 40 identifies sites and/or States where loading conditions were either much lighter or much heavier than observed at the remaining SPS TPF sites, as well as States that had significantly different RANALS most likely due to differences in the vehicle classification system being used by the WIM scale collecting the data. These RANALS, identified as special cases are summarized in columns the under "Clusters Observed in a Single State..." heading in table 40. These RANALS are either lighter or heavier than the default loading condition, with a large number of instances occurring where the unique loading condition was much lighter or much heavier than the default. Some of the RANALS were identified as significantly different based on classification. These RANALS had a very high percentage of light axles in the first (lightest) load bin (States: Ohio, Washington, Minnesota, and Maine). This generally means that passenger vehicles pulling trailers have been classified as trucks.

These load spectra are not recommended for general use within the MEPDG procedure unless they represent known loading conditions or a desired hypothetical condition at the site being evaluated. States that had the largest number of RANALS identified as special cases are Arizona, Florida, Ohio, Tennessee, and Washington.

NALS Clusters for Class 9 Tandems

While for many vehicle classes and axle group types there was a clearly defined default loading condition (i.e., the majority of SPS TPF sites had similar axle loading distribution), no single loading condition dominated the class 9 tandem data. The loading condition of class 9 tandems varied between the sites, ranging from moderately loaded conditions (less than 30 percent of tandem axles over 26 kip) to very heavily loaded (up to 70 percent of tandem axles over 26 kip). All of these conditions can be found routinely on U.S. roads. In addition, MEPDG analysis indicated that pavement design outcomes are very sensitive to selection of class 9 tandem NALS primarily because this heavy axle is the most frequently observed on U.S. primary roads. As a result, it has a large impact on pavement design. Therefore, instead of one default loading condition, all three NALS cluster groups develop for class 9 tandems (one moderate and two heavy) are recommended as defaults.

If no site-specific loading information is available, the moderate loading condition is recommended for roads dominated by urban delivery trucking patterns. The heavy #1 condition is recommended for roads where long haul trucking or heavy directional hauls overlap with urban delivery movements. The heavy #2 condition is recommended for rural roads where almost all class 9 truck traffic is oriented toward long haul traffic.

Characteristics of NALS Clusters Representing Different Axle Loading Conditions

To help characterize the NALS representing default and alternative axle loading conditions so that pavement designers can more easily choose between alternative loading conditions, several statistical parameters were computed. Table 41 shows average RPPIF values computed for the default and alternative loading conditions. Table 42 shows the average percentage of heavy axles (i.e., the percentage of axles that are at or above 75 percent of the Federal legal weight limit) computed for different NALS clusters.

* NALS has very heavy overloads; **NALS identified as outlier based on classification issue (high percentage of very lightweight axles). Blank cells indicate that no instances were identified.

* NALS has very heavy overloads; **NALS identified as outlier based on classification issue (high percentage of very light weight axles). ***Outlier values based on rules in table 24. Blank cells indicate that no instances were identified.

Use of Loading Cluster NALS Defaults at the Global and Local Levels

Understanding of Local Loading Conditions

Before selection of the loading cluster NALS defaults, it is recommended that every effort is made to understand the expected traffic loading pattern at the site for which default NALS selections are being made. Dominant heavy vehicle classes (FHWA classes 4 and 6 through 13) should be identified and descriptive traffic loading conditions (see table 24) for the dominant heavy vehicle classes should be established.

For the majority of the LTPP sites (and a majority of U.S. primary roads), class 9 is the dominant heavy vehicle type. Therefore, special effort should be made for identifying loading conditions for class 9 vehicles. Class 5 vehicles frequently are more prevalent but are not heavy enough to make significant contribution to total traffic loading and may be excluded from determination of the dominant heavy vehicle classes, unless local knowledge exists of heavier-than-usual class 5 vehicles at the site.

If a State's truck sizes and weight laws allow loads exceeding Federal regulations 23 U.S.C. 127 and 23 CFR 658, and the site has a low percentage of interstate traffic, there is a higher likelihood of heavier-than-typical NALS, especially if there are heavy commodities being transported by certain vehicle classes.^(19,20)

In some instances, historical WIM data or recently collected portable WIM data may be available for the site. These data may not be accurate enough to compute NALS for the site but may be useful in establishing a descriptive loading condition.

Table 40 provides information that could be used for evaluating the relative importance of NALS associated with different vehicle classes and axle group types for pavement design. For example, the combination of a frequently observed vehicle class with a heavy or very heavy axle group type is likely to produce the most impact on pavement design outcome, and vice versa. Accurate selection of NALS for these classes and axles is more important compared to light and/or infrequent vehicles classes.

Decision Tree for Selecting Default NALS Clusters for Pavement Design

Once descriptive loading categories are determined for the dominant vehicle classes, default NALS corresponding to the identified loading categories could be selected from the LTPP database. If no local knowledge exists, either tier 1 NALS defaults or tier 2 NALS defaults representing typical conditions could be used. For a more conservative analysis, tier 2 NALS defaults representing heavy conditions could be used. MEPDG analysis using different selection of default NALS clusters could be used to assess the effect of different NALS defaults selection on pavement design and analysis outcomes. The following decision tree could be used for selecting default NALS for a given pavement design site (default NALS clusters should be assigned by vehicle class and axle group type):

1. Identify one or more dominant heavy vehicle class by weight at the site using the following guidance:
   
   
   • If this is FHWA vehicle class 4 or 6-13 and represents 30 percent or more of all vehicle classes 4-13.
     
     
   • If this vehicle class carries 25 percent or more of total load.
     
     
   • Class 5 vehicles are usually too light and should be excluded from evaluation unless this vehicle carries 25 percent or more of total load.
     
     
2. Identify if the dominant truck class is likely to have unusually heavy loads if one or more of the following applies:
   
   
   • High percentages of loads close to the Federal legal load limit (30,000 to 34,000 lb for tandems).
     
     
   • High percentages of loads above the Federal legal load limit due to lack of enforcement (> 34,000 lb for tandems).
     
     
   • High percentage of loads with permits above the Federal legal load limit (>34,000 lb for tandems).
     
     
   • Overloads due to illegal activity.
     
     
   • Legal load limits on this road are above Federal legal load limits.
     
     
   • If one or more of these applies, assign the heaviest default NALS cluster available for this class.
     
     
3. If no unusually heavy conditions are identified for the dominant class, use other means to identify the descriptive loading category defined in table 40 and table 24. This could include the following:
   
   
   • Review of historical loading data, such as axle load spectra or historical ESAL/ truck values.
     
     
   • Short-term portable WIM data collection to identify rough percentage of heavy loads per table 24.
     
     
   • Interview personnel from transportation planning and freight movement departments.
     
     
   • For class 9, the moderate loading condition is recommended for roads dominated by urban delivery trucking patterns. The heavy #1 loading condition is recommended for roads where long haul trucking or heavy directional hauls overlap with urban delivery movements. The heavy #2 loading condition is recommended for rural roads where almost all Class 9 truck traffic is oriented toward long haul traffic.
     
     
4. Based on findings of the expected loading condition for dominant vehicle classes, assign default NALS cluster code using values in table 40 and table 24 as guidance.
   
   
5. In addition to dominant heavy vehicle classes, identify any other classes that are likely to routinely carry loads above Federal legal load limits. Assign the heaviest available default NALS cluster for identified classes.
   
   
6. Assign typical default NALS cluster for all other classes and axle group types.

Limitations of Loading Cluster NALS Defaults

Axle loading cluster defaults were developed based on the load spectra data from 26 WIM sites. The small number of SPS TPF WIM sites limits the geographic scope of the weight data, which in turn means that this table is not statistically representative of all roads in the country. It is not representative of specific roads in specific States where loading conditions are dominated by State regulations or specific commodity movements that differ from national norms. When State or site-specific trucking movements dominate the use of a road, those roadways may experience loading conditions that are different than those observed in the SPS TPF data.

SPS TPF Site Membership in Different NALS Loading Clusters

Table 43 provides information about different axle loading conditions observed at individual SPS TPF sites by vehicle class and axle group type. Cells in table 43 are populated with an abbreviated NALS cluster name to indicate SPS TPF site cluster membership for each vehicle class and axle group type. Information included in this table may be useful in identifying axle loading conditions for locations and roads related to the SPS TPF sites. In addition, this table could be used to identify locations where specific axle loading conditions were observed (very light, light, moderate, heavy, and very heavy). For example, site 4-0100 in Arizona experienced much more frequent heavy or very heavy loading compared to site 4-0200 in the same State.

VL = Very light, L = Light, M = Moderate, H = Heavy, VH = Very Heavy, and T = Typical.

N/A = Not applicable.

MEPDG COMPARISON OF NCHRP 1-37A AND LTPP SPS TPF AXLE LOADING DEFAULTS

Analysis Approach

An analysis was used to compare MEPDG outputs using the original defaults and the new alternate defaults. A set of rigid and flexible pavement designs for RI and ROPA roads was developed, and pavement performance was analyzed using the NCHRP 1-37A defaults and the new SPS TPF defaults, as follows:

• NCHRP 1-37A.⁽³⁾
  
  
• SPS TPF tier 1 NALS.
  
  
• SPS TPF tier 2 lightest NALS (the lightest available tier 2 NALS cluster was selected for each vehicle class and axle group type, excluding special cases NALS).
  
  
• SPS TPF tier 2 typical (the tier 2 NALS cluster identified as "typical" (T) was selected for each vehicle class and axle group type).
  
  
• SPS TPF tier 2 heaviest (the heaviest available tier 2 NALS cluster was selected for each vehicle class and axle group type, excluding special cases NALS).

VCDs and AADTT values characteristic to RI and ROPA roads were used for the analyses. MEPDG VCD defaults for TTC1 were used for RI conditions, and TTC6 was used for ROPA. All four LTPP climatic regions were included in the analysis.

Three flexible pavement failure modes were investigated for flexible pavements: rutting, bottom-up cracking, and top-down cracking. For JPCP rigid pavements, slab cracking slab cracking and joint faulting failure modes were investigated. These failure modes were found most critical using MEPDG pavement performance prediction models calibrated to global conditions.

A summary of pavement design inputs is provided in table 44.

To compare MEPDG outcomes using different loading defaults, a pavement was first designed using NCHRP 1-37A NALS and APC coefficients, and then that design was reanalyzed using the same inputs, except for NALS and APC coefficients, which were substituted with the SPS TPF default values. Both 15- and 20-year designs were used for flexible and rigid pavements, respectively, based on observed average pavement service life for LTPP sections. MEPDG results computed for 90 percent design reliability (default MEPDG option) were obtained for each default and analyzed.

A summary of MEPDG outcomes based on the most critical design for each failure mode is provided in table 45. The table shows differences in predicted pavement service life and structural thickness using new defaults, compared to values predicted using old defaults. For the majority of defaults and pavement designs, new defaults resulted in longer lives and thinner pavements. The following sections describe in more detail a comparison of MEPDG results using NCHRP 1-37A and results when using LTPP SPS TPF defaults for NALS and APC coefficient for different road types and pavement failure modes.

Comparison for Flexible Pavements Designed for RIs

NCHRP 1-37A Versus LTPP SPS TPF Tier 1 NALS Defaults

The results of analyses for flexible pavements designed for RI indicate that an increase in service life prediction close to 3 years (20 percent of service life) is expected if the new tier 1 loading defaults are used instead of the original MEPDG loading defaults. In the case of cracking, this translates up to a 0.3-inch difference in the thickness of the hot mix asphalt (HMA) layer, and in the case of rutting, this could lead to a 1.1-inch thickness difference. The tier 1 loading defaults result in thinner pavements. From a pavement thickness design perspective, HMA thickness differences less than 0.5 inch are not significant. Based on observed thickness differences from MEPDG analysis, use of NCHRP 1-37A or LTPP SPS TPF tier 1 defaults is not likely to result in significant design thickness differences for pavements that are likely to fail in cracking mode but could produce significantly different design outcomes for pavements that are likely to fail in rutting mode. However, rutting isn't likely to be mitigated just by the increase in AC thickness. As a result, the significance of the different defaults may not be as critical. This conclusion applies to all four climatic zones.

NCHRP 1-37A Versus LTPP SPS TPF Tier 2 NALS Defaults

All tier 2 NALS defaults resulted in longer service life predictions than the NCHRP 1-37A NALS defaults. Service life prediction for the tier 2 Lightest NALS was up to 52 percent longer compared to the NCHRP 1-37A NALS defaults. This difference is considered of practical significance. Service life prediction for the tier 2 typical group was up to 32 percent longer compared to the NCHRP 1-37A NALS defaults. This difference also is considered of practical significance. Service life prediction for the tier 2 heaviest group was just 6 percent longer compared to the NCHRP 1-37A NALS defaults and is not considered of practical significance. All tier 2 loading defaults resulted in thinner pavements. For the tier 2 lightest NALS, AC thickness difference was over 0.5 inch for all distress modes: 1.7 inch for rutting, 0.5 inch for bottom-up cracking, and 0.6 inch for top-down cracking. These differences are considered of practical significance. For the tier 2 typical NALS, AC thickness difference was over 0.5 inch for the rutting distress mode only (1.2 inches); the other distress modes resulted in insignificant differences (0.3 inch or less). For the tier 2 heaviest NALS, all AC thickness differences were between zero and 0.3 inch and are considered insignificant. MEPDG predictions were very similar using the NCHRP 1-37A and SPS TPF tier 2 heaviest NALS defaults.

Comparison for Flexible Pavements Designed for ROPA Roads

NCHRP 1-37A Versus LTPP SPS TPF Tier 1 NALS Defaults

The results of analyses for flexible pavement designed for ROPA roads indicate that an increase in service life prediction close to 3 years (20 percent of service life) is expected if the new tier 1 loading defaults are used instead of the original MEPDG loading defaults. In the case of cracking modes, this translates up to a 0.3-inch HMA thickness difference. The tier 1 loading defaults result in thinner pavements. From a pavement thickness design perspective, HMA thickness differences less than 0.5 inch are not significant. Based on observed thickness differences from MEPDG analysis, use of the NCHRP 1-37A or LTPP SPS TPF tier 1 defaults is not likely to result in significant design thickness differences for pavements that are likely to fail in cracking mode. This conclusion applies to all four climatic zones.

NCHRP 1-37A Versus LTPP SPS TPF Tier 2 NALS Defaults

All tier 2 NALS defaults resulted in longer service life prediction than the NCHRP 1-37A NALS defaults. Service life prediction for the tier 2 lightest NALS was up to 51 percent longer compared to the NCHRP 1-37A NALS defaults. This difference is considered of practical significance. Service life prediction for the tier 2 typical group was up to 30 percent longer compared to the NCHRP 1-37A NALS defaults. This difference also is considered of practical significance. Service life prediction for the tier 2 heaviest group was just 6 percent longer compared to NCHRP 1-37A NALS defaults, which is not considered of practical significance. Tier 2 lightest and typical loading defaults resulted in thinner pavements (up to 0.8 and 0.4 inch difference, respectively). Tier 2 heaviest loading defaults resulted in the same thickness. Only for the tier 2 lightest NALS the AC thickness difference was over 0.5 inch and considered significant from a practical perspective. MEPDG predictions were found very similar using the NCHRP 1-37A and SPS TPF tier 2 heaviest NALS defaults.

Comparison for Rigid Pavements Designed for RIs

NCHRP 1-37A Versus LTPP SPS TPF Tier 1 NALS Defaults

The results of analyses for rigid pavements designed for RI indicate that an increase in service life prediction up to 2.4 years (12 percent) is expected if SPS TPF tier 1 loading defaults are used instead of the original MEPDG loading defaults. This translates into a 0.2-inch PCC slab thickness difference. The tier 1 loading defaults result in thinner pavements. From a pavement thickness design perspective, PCC slab thickness differences less than 0.5 inch are not significant. Based on observed thickness differences from MEPDG analysis, using either the NCHRP 1-37A or LTPP SPS TPF tier 1 defaults is not likely to result in significant design thickness differences for JPCP designed for typical RI conditions. This conclusion applies to all four climatic zones.

NCHRP 1-37A Versus LTPP SPS TPF Tier 2 NALS Defaults

The tier 2 lightest and typical NALS defaults resulted in longer service life predictions, while the tier 2 heaviest NALS defaults resulted in shorter service life prediction compared to the NCHRP 1-37A NALS defaults. Service life prediction for the tier 2 lightest NALS was up to 41 percent longer than the NCHRP 1-37A NALS defaults. This difference is considered of practical significance. Service life prediction for the tier 2 typical group was up to 20 percent longer than the NCHRP 1-37A NALS defaults. This difference also is considered of practical significance. Service life prediction for the tier 2 heaviest group was up to 9 percent shorter than the NCHRP 1-37A NALS defaults, but this difference is not considered of practical significance. The tier 2 lightest and typical loading defaults resulted in thinner pavements (up to 0.4- and 0.1-inch difference, respectively). The tier 2 heaviest loading defaults resulted in a slightly thicker PCC slab (0.1 inch). All PCC thickness differences were less than 0.5 inch and considered insignificant from a practical perspective; however, for the tier 2 lightest NALS defaults, the difference in thickness (0.4 inch) was close to 0.5 inch, and for some designs significant differences are possible. MEPDG predictions were found very similar when using the NCHRP 1-37A and SPS TPF tier 2 heaviest NALS defaults.

Comparison for Rigid Pavements Designed for ROPA Roads

NCHRP 1-37A Versus LTPP SPS TPF Tier 1 NALS Defaults

For designs that fail in faulting mode, the results of analyses indicate a slight decrease in service life prediction if the new tier 1 loading defaults are used instead of the original MEPDG loading defaults-up to 6 months for 20-year design life (2.5 percent). Based on observed differences, using either the NCHRP 1-37A or LTPP SPS TPF tier 1 defaults is not likely to result in significant design differences for pavements that are likely to fail in faulting mode under typical ROPA traffic conditions. This conclusion applies to all four climatic zones.

For designs that fail in cracking mode, the analysis results indicate an increase in service life prediction up to 4.6 years (over 23 percent of service life) is expected if the tier 1 loading defaults are used instead of the original MEPDG loading defaults. This difference in pavement design life could be considered significant from practical perspective. The tier 1 loading defaults result in thinner pavements. However, the PCC slab thickness difference is only 0.3 inch when designed using NCHRP 1-37A and LTPP SPS TPF defaults. From a pavement thickness design perspective, PCC slab thickness differences less than 0.5 inch are not significant. Based on observed thickness differences from MEPDG analysis, using either the NCHRP 1-37A or LTPP SPS TPF tier 1 defaults is not likely to result in significant design thickness differences for pavements that are likely to fail in cracking mode. This conclusion applies to all four climatic zones.

NCHRP 1-37A Versus LTPP SPS TPF Tier 2 NALS Defaults

For slab cracking failure mode, tier 2 lightest and typical NALS defaults resulted in longer service life prediction, while tier 2 heaviest NALS defaults resulted mostly in shorter service life prediction compared to the NCHRP 1-37A NALS defaults. Service life prediction for the tier 2 lightest NALS was up to 49 percent longer than the NCHRP 1-37A NALS defaults. This difference is considered of practical significance. Service life prediction for the tier 2 typical group was up to 28 percent longer than the NCHRP 1-37A NALS defaults. This difference also is considered of practical significance. Service life prediction for the tier 2 heaviest group ranged from 7.5 percent longer up to 11 percent shorter than the NCHRP 1-37A NALS defaults. These differences are not considered of practical significance (20 percent was used as a threshold). The tier 2 lightest and typical loading defaults resulted in thinner pavements (up to 0.5- and 0.2-inch thickness difference, respectively). The tier 2 heaviest loading defaults resulted in a slightly thicker PCC slab (0.2 inch). PCC thickness differences for the tier 2 heaviest and typical loading defaults were less than 0.5 inch and considered insignificant from the practical perspective. However, for the tier 2 lightest NALS defaults, the thickness difference was 0.5 inch and is considered significant. MEPDG predictions were found to be very similar between the NCHRP 1-37A and SPS TPF tier 2 heaviest NALS defaults.

For designs that fail in faulting mode, the tier 2 lightest NALS defaults resulted in longer service life prediction (up to 6 percent), while the tier 2 heaviest and typical NALS defaults resulted in shorter service life prediction (up to 9 percent) compared to the NCHRP 1-37A NALS defaults. These differences are not considered of practical significance (20 percent was used as a threshold). MEPDG predictions were found to be very similar between the NCHRP 1-37A and SPS TPF tier 2 typical NALS defaults.

Comparison of Axles per Class Coefficients

The researchers also ran a scenario in which the new load spectra defaults were used in combination with NCHRP 1-37A and LTPP SPS TPF APC coefficient default values. The results from this analysis did not show significant difference in pavement design life or thickness predictions between the two sets of APC default values.

DEFAULT APC NUMBERS BASED ON SPS TPF DATA

To compute the APC defaults, values were computed for each of the SPS TPF sites and then averaged across the sites using the methodology described in chapter 9. Appendix D includes a database CD that contains these results. APC values are provided for each site by vehicle class and axle group type. Averaging the APC values for the 26 sites resulted in one set of default values, as shown in table 46.

Comparison between NCHRP 1-37A and SPS TPF Default APC Numbers

Figure 50 compares the average number of axles within an axle group type and the total number of axles for each truck class between default values used in the MEPDG version 1.1 software and values computed using SPS TPF data. As shown, there are some differences, but they are relatively small and unlikely to result in any significant difference in terms of required layer thickness or predicted distress for flexible or rigid pavements.

Figure 50. Graph. Comparison of the APC values from all SPS TPF sites to the MEPDG default values.

The comparison of data presented in table 47 indicates that major differences in APC values occurred for classes 11 and 13. However, because these classes are relatively infrequent, improved APC values alone will not have a significant impact on predicted pavement performance.

AXLE SPACING AND WHEELBASE DEFAULTS

Average Axle Spacing and Wheelbase

SPS TPF data were used to investigate the distribution of axle spacing (wheelbase) of the tractor unit of tractor-semitrailer combination trucks for FHWA vehicle classes 8 and above. Table 48 shows the results of the axle spacing distribution analysis. Based on these results, users could define their own categories of short, medium, and long axle spacing based on selected slab joint spacing and compute corresponding percentages of axles in the short, medium, and long categories by aggregating the values presented in table 48.

Table 48. Distribution of axle spacing on tractor unit for FHWA vehicle classes 8-13.

Since 15 ft is the most frequently used joint spacing for JPCP joint design, the following axle spacing distribution of the tractor unit of tractor-semitrailer combination trucks for classes 8 and above is recommended based on SPS TPF data:

• Short (≤ 12 ft): 4.3 percent.
  
  
• Medium (>12 and ≤ 15 ft): 16.2 percent.
  
  
• Long (≥ 15 ft): 79.5 percent.

In addition, the MEPDG states that if other vehicles in the traffic stream also have the axle spacing in the range of the short, medium, and long spacing defined above, the frequency of those vehicles could be added to the axle spacing distribution of truck tractors.⁽¹⁾ For example, if 10 percent of truck traffic is from multiple trailers (class 11 and above) that have a trailer-to-trailer axle spacing in the short range, 10 percent should be added to the percentage of truck tractors for short axles. Thus, the sum of percent trucks in the short, medium, and long categories can be greater than 100.

A sample of axle spacing data from SPS TPF WIM sites was used to estimate percentages of axle spacing that fall in different length categories. The results of the axle spacing distribution analysis are shown in table 49 for vehicle classes 4 through 13 and provide additional insights into what vehicle classes are likely to have axle spacing that could contribute to the development of top-down cracking in JPCP.

Axle Spacing for Multi-Axle Groups

SPS TPF axle spacing data were used to compute average axle spacing for tandem, tridem, and quad axle groups. These averages were then compared with the current MEPDG defaults. The results are presented in table 50.

As can be seen from the table, the values are very close. SPS TPF-based averages are slightly lower for tandems and higher for tridems and quads compared to the current MEPDG default values.

The values based on SPS TPF WIM sites reported in table 50 are recommended to be used as new MEPDG defaults, as these values are obtained from accurately calibrated WIM sites and are based on the analysis of 4.7 million records from SPS TPF sites.