Office of Operations Freight Management and Operations

Comprehensive Truck Size and Weight Limits Study: Linkage between the Revised Desk Scans and Project Plans Report

Chapter 4: Safety Comparative Analysis

4.1 Purpose

The purpose of this section is to document how the information and understanding gained through the Desk Scan informed the technical methodology undertaken in the Safety Comparative Analysis. The technical methodology was initially provided in the Project Plan and subsequently revised throughout the course of the project as details about available data emerged and analytical approaches were refined. As such, the linkages in this report reference the Project Plan as presented in the Highway Safety and Truck Crash Comparative Analysis technical report (USDOT, 2014).

This report establishes linkages in two principal areas:

  • Linkages regarding the general technical approach: The report establishes the linkage between literature findings on approaches to assess the safety implications of control and alternative vehicle configurations, and
  • Linkages regarding available data/analysis methods: Based on the assessment of research and data needs summarized in the revised Desk Scan, this report links literature findings with the use and integration of the data sources used in the 2014 CTSW Study.

4.2 Linkages Regarding General Technical Approach

4.2.1 Background

The general technical approach to the safety analysis is based on the USDOT 2000 CTSW Study (USDOT, 2000). In that study, separate analyses were attempted for crash analysis and vehicle stability and control. The 2000 study crash analysis was largely unsuccessful, owing to data problems in identifying crashes and matching exposure for the control and alternative truck configurations. A similar problem occurred during the 2004 Western Governor's Association Study (USDOT (2004). In 2014, the safety team attempted 3 different crash analyses, described in the following paragraph, in an effort to obtain more meaningful and useful crash analysis results. The team retained and enhanced the use of vehicle stability and control simulations to obtain insights on configuration performance that could not be obtained from the crash-based studies. Finally, inspection and violation data are compared to better understand the violations associated with the control and alternative configurations.

In considering these challenges, the safety team structured the analysis, where possible, to respond to the original Congressional request for the study which sought to explore differences in safety risk and truck crash frequency between truck configurations currently operating on the nation's roadways at and below current federal limits compared to those operating above such limits. This wording led to a focus on existing operations where truck configurations of interest were allowed to operate through the complex series of state exemptions and other special provisions. The safety team worked with others within the Study team to identify states where control and alternative configurations were in current operation. Identifying these states help direct the data collection and modeling efforts described in further detail in section 4.3.

4.2.2 Crash Analysis

The crash analysis portion of the safety plan sought to overcome the problems in the 2000 CTSW Study by pursuing 3 crash analysis approaches (see summary in Table 4.1):

  1. Segment-level crash comparisons using crash and exposure data from states allowing travel of control and alternative vehicles. This approach initially required that control and alternative configurations be identified in crash data (including the weight of involved configurations). This is similar to the approach unsuccessfully taken in the 2000 CTSW Study; it was also unsuccessful in the 2014 effort, again due to lack of configuration data in weight records. In response to this challenge, the team developed an approach in which states were identified that allowed both control and comparison configurations that could be identified in crash records by their axle counts and/or number of trailers. While this approach did not allow the identification of weight-specific individual crash records, it identified groups of configurations that were allowed to be operating above federal limits. This was interpreted as consistent with the intent of the federal legislation. It resulted in the comparison of control vehicles operating at a range of weights with alternative vehicles also operating at a range of weights. The team believes this type of comparison reflects how groups of vehicles may respond to changes in size and weight regulation: all trucks do not operate at the maximum allowable weight at all times; so a more realistic comparison may be one that includes a range of crashes (of unknown weight) involving vehicle that are allowed to carry maximum loads in excess of current federal limits.
  2. Route-level crash comparisons using WIM data and exposure data from states to identify routes in which only control vehicles operated and could be compared to routes in which primarily alternative configurations operated. This is another method that does not require crash-level configuration weight data, but is built on the assumption that alternative configurations are limited in the routes of their operations. Information received from candidate states for this method, however, revealed that travel by alternative configurations of interest are ubiquitous in most states, so the underlying assumption of this method was incorrect. As a result, the method was not used.
  3. Fleet-based analyses were attempted to obtain crash details from the carriers involved. Investigations led to an understanding that only few carriers used tractor triple-trailer in their regular operations. It was felt that this knowledge could be used to increase the sample size of triples crashes in a way that would facilitate the comparison with tractor double-trailers. This was a new approach not previously attempted in the 2000 CTSW Study. Jovanis had used a similar approach in a study from the 1980s (Jovanis et al., 1989). Crash data were successfully assembled from carriers, but consistent data on exposure could not be obtained; crash rates could thus not be computed and compared. The crash data were useful in a set of severity analyses, however.

4.2.3 Analysis of Vehicle Stability and Control

An alternative to relying on crash data analysis is to conduct detailed simulations of vehicle performance using available computationally intensive computer software. Both the 2000 CTSW Study and 2004 Western Uniformity Analysis used this method to gain insight into the potential safety performance of a range of vehicle configurations. The advantage of the simulation of vehicle stability and control simulations is that one has experimental control over the vehicle configuration and the test protocols used to assess vehicle configuration performance.

Table 4.1: Summary of Crash Analysis Methods Used in 2014 CTSW Study
Method Critical Assumptions Comments
State-based WIM data allows classification of travel exposure at segment level for all roadway functional classes. Identify control and alternative vehicle configurations using axle count and number of trailers in crash data. Similar to 2000 CTSW approach and study by Abdel-Rahim, but chose states to compare groups of control and alternative vehicles in crash reports. Results still affected by small crash sample size for some configurations. Only able to compute rates for interstates due to WIM limitations
Route-based Alternative configurations operate on subset of all routes that can be identified by WIM and/or state data. Safety estimated by comparing these routes, without knowing configuration weight from individual crashes Control and alternative vehicles operate ubiquitously in states; unable to compare routes as planned
Fleet-based Collecting data from fleets would result in an increased number of alternative configuration crashes for analysis Obtained crash data but unable to obtain reliable fleet-based exposure data. Crash data used for severity analysis but not crash rate comparisons.

Many details of the vehicle configurations are specified in standard test protocols (e.g., vehicle speed, weight and distribution of weights within trailer units, brake condition). So all vehicle configurations can be compared under controlled, nearly identical, conditions. The weakness is that the simulations are unable to replicate the range of real-world conditions experienced in the field, including variation in traffic, weather, roadway and driver attributes. What is gained in experimental control is lost in the ability to encapsulate real-world operating conditions.  

4.2.4 Analysis of Inspections and Violations

The third component of the safety analysis was a study of inspections and violations of control and alternative configurations. This component of the safety study explored potential connections between truck configuration (e.g., tractors pulling two trailers and tractors pulling three trailers) and their record of operating violations (other than over-weight). This approach compared the violation record of control vehicle configurations from a set of states identified as allowing legal operations of alternative configurations in excess of current federal limits for weight and configuration. There was no study identified in the literature that is comparable to the one completed concerning inspections and violations in the 2014 CTSW Study.

4.3 Linkages Concerning Available Data and Methods

4.3.1 Crash Analysis

The state-level data analysis was similar to the study conducted by Abdel-Rahim (Abdel-Rahim, 2006) and discussed within the desk scan with one important difference: the 2014 CTSW Study team specifically chose states for inclusion that allowed a comparison of groups of control and alternative vehicle configurations. Abdel-Rahim used a roughly similar approach in choosing states that operated LCV's and using WIM data to estimate configuration exposure. The 2014 CTSW Study team sought a more precise comparison of crash rates by selecting states that allowed legal operations of truck configurations exceeding federal limits for configuration and weight (as described in the 2014 CTSW Study safety report, USDOT, 2014). The team selected states where, for example, tractor semi-trailers with 6 axles were allowed at weights beyond 80,000 lb. and identified as six-axle trailers in crash reports. This allowed the team to assemble crash records for both the control and alternative vehicles and conduct a comparison of rates using number of trailers and axles alone; without vehicle weight. The state-level analysis is thus associated with the methods of Abdel-Rahim, but sought a more precise crash rate comparison. Exposure data and crash data were assembled on a segment-by-segment basis; rates computed and compared.

The route-based method was conceived from the same foundation as the state-level analysis but was used to provide a contingency in the event that no useful vehicle configuration information could be obtained from the state crash records. In this event, the team proposed to compute truck crash rates for road segments with different levels of configuration flows (as measured by WIM stations). This approach has its foundation in the state-level analyses described above, but was novel in its use of WIM data. The assumption underlying the approach was that there were specific route that could be identified that had only control vehicle exposure and little or no alternative vehicle exposure. A comparison of crash rates could, hypothetically, provide some information about the crash risk of alternative configurations. Unfortunately, discussions with state DOT personnel in the selected states revealed that alternative configuration travel was widespread across all route and road segments with virtually no sites with zero exposure of either controls or alternative configurations. As a result, the method was infeasible in all selected states.

The fleet analysis was based on a study publish by TRB in the 1980s (Jovanis, et al., 1989). The study method was a matched pair approach where crash rates for tractors with two semi-trailers were compared to crash rates for single combination tractor semi-trailers on identical routes for the same firm.  This matched pair design controlled for route geometric characteristics, type of operation, driver management, safety culture and other factors except for driver age and experience. The matched-pair design was unfortunately not feasible for the study of tractors pulling 3 semi-trailers because companies that operated these vehicles in western states operated virtually no doubles on the same routes because triples were more economical. When the matched - pair approach became infeasible, the team sought to compare crash rates for doubles and triples throughout the fleet's national network. This proved difficult because exposure data were more difficult to obtain than expected. As a result, the primary value of the fleet-based analysis was to compare severity of crash outcome, given a crash, for double and triple combination vehicles with each fleet. These results were reported in the safety technical report (USDOT, 2014).

4.3.2 Analysis of Vehicle Stability and Control

Table 4.2 summarizes the analysis of vehicle stability and control for the 2000 CTSW Study (USDOT 2000a, b) and the 2014 CTSW Study. The simulations in 2014 were defined in conjunction with USDOT subject matter experts. In vehicle stability and control simulations, researchers define input data based on the vehicle configurations and performance characteristics of interest. The 2000 CTSW Study was concerned with rollover events and off-tracking; the simulations reflected these concerns. Performance during both high-speed and low-speed turns was simulated, along with an evasive maneuver. Metrics derived from the simulations are listed in the last column. The 2014 CTSW Study included similar off-tracking simulations (e.g., high and low speed turns and an evasive maneuver), updated reflecting more contemporary models. In addition, much more attention was paid to braking comparisons and performance during straight-line and curve traversals, including simulations with brake failures on some axles. These more extensive tests better illustrated the effects of brake failures on both control and alternative vehicle configurations. There is a clear connection between the data and models used in the 2014 CTSW Study and its predecessor study in 2000.

Table 4.2: Summary of Vehicle Stability and Control Test for 2000 CTSW and 2014 CTSW
2000 CTSW Study
Maneuver Comment Metric
Steady-state turn-induced rollover Represents roll propensity during turn Minimum lateral acceleration to result in wheel lift off ground (static roll stability)
Evasive maneuver induced rollover (SAE J 2179) Represents roll propensity during high speed evasive maneuver Rear trailer lateral motion relative to tractor (rearward amplification)
Shift in load during maneuver Lateral Load Transfer Ratio - proportion of total axle load carried on one wheel compared to the other
Low-speed off-tracking Represents difference in tracking of wheels of steering axle and rear axle of last trailer during low speed turn Offtracking
Swept Path
Encroachment to inside of track
2014 CTSW Study
Maneuver Comment Metric
Low-speed off-tracking Represents an intersection turn Off-tracking (intersections)
High-speed off-tracking Represents a curve on a highway Off-tracking (highway curves)
Straight-line braking (S5.3.1.1 of FMVSS 121, 60mph) Conducted with fully functioning brakes and with two brake malfunctions

Stopping Distance

Maximum Path Deviation

Brake in a curve (S5.3.6.1 of FMVSS No. 121. 30 mph) Conducted with fully functioning brakes and with two brake malfunctions

Stopping Distance

Maximum Path Deviation

Lateral Load Transfer Ratio

Avoidance maneuver (similar to ISO 14791, lateral stability test methods. 50 mph) Run under multiple conditions

Transient off-tracking

Rearward amplification

Lateral Load Transfer Ratio

4.3.3 Analysis of Inspections and Violations

The 2014 CTSW Study used data from the Motor Carrier Management Information System (MCMIS) for specific states to compare the pattern of inspections and violations for control and alternative configurations. The analysis was conducted to be as consistent as possible with the goal expressed in the enabling federal legislation which sought to explore differences in safety risk and truck crash frequency between truck configurations currently operating on the nation's roadways at and below current federal limits compared to those operating above such limits. As a result MCMIS data were obtained from states allowing legal operation of larger and heavier truck configurations due to state exemptions and special provisions. Mean violation and citation rates were compared for different configurations using a range of statistical approaches. No comparable analysis was identified in the desk scan.

4.4 Summary

Two comparisons have been conducted of the safety analyses undertaken for the 2014 CTSW Study. Concerning general technical approaches, the 2014 CTSW Study borrowed heavily from both the 2000 CTSW Study and the 2004 Western Uniformity Scenario Study. The crash analyses developed with a clear understanding of the difficulties encountered in both these studies. The 2014 CTSW Study team developed 3 alternative crash analysis approaches in the hope that they would yield differing perspectives on safety associated with truck size and weight. In each of the three crash-based approaches the desk scan led the team to expect challenges; and the team experienced them. A more vexing circumstance continued to be the lack of configuration weight data in crash reports, even for the carrier-supplied crash data. While the three crash-based approaches were distinct and used crash data in different ways, they all suffered from the general poor availability of information about truck weight and configuration within crash data.

Vehicle stability and control simulations were, in many ways, much easier to conduct. Once the simulation parameters were defined in consultation with USDOT subject matter experts, the models could be exercised to produce outputs comparing the control and alternative vehicle configurations of interest. The limitation of this approach is, unfortunately, that it does not capture the details and variability of actual crash events as they occur in the field; only analysis of actual crash data can capture those nuances.

The analysis of inspection and violation data from MCMIS was developed within the safety team, again with discussions among the USDOT subject matter experts. While there was discussion of several data challenges, the analysis did not draw on any specific references from the safety desk scan. 

4.5 References

Abdel-Rahim, A., S. G. Berrio-Gonzales, et al. (2006b). Longer Combinations Vehicles: A Comparative Crash Rate Analysis. Final Report Part B. University of Idaho. , National Institute for Advanced Transportation Technology.

Jovanis, P., H.-L. Chang, et al. (1989). "Comparison of Accident Rates for Two Truck Configurations." Transportation Research Record 1249.

USDOT (2014), Highway Safety and Truck Crash Comparative Analysis: Comprehensive Truck Size and Weight Limits Study.

USDOT (2000a) Comprehensive Truck Size and Weight Study. Washington DC, US Department of Transportation. Publication Number: FHWA-PL-00-029 (Volume II Chapters 5 and 6) HPTS, August. Available at https://www.fhwa.dot.gov/reports/tswstudy/Vol2-Chapter5 and Chapter6.pdf as of May 30, 2014.

FHWA (2000b). Comprehensive Truck Size and Weight Study. Washington DC, US Department of Transportation. Publication Number: FHWA-PL-00-029 (Volume III Scenario Analysis, Chapters 7 Roadway Geometry and 8 Safety) HPTS, August. Available at https://www.fhwa.dot.gov/reports/tswstudy/Vol3-Chapter7 and Chapter8.pdf as of May 30, 2014.

USDOT (2004). Western Uniformity Scenario Analysis, Washington, DC, US Department of Transportation. Available at https://www.fhwa.dot.gov/policy/otps/truck/wusr/wusr.pdf as of May 30, 2014.

previous | next
Office of Operations