Journal of Advanced Transportation

Journal of Advanced Transportation / 2021 / Article

Research Article | Open Access

Volume 2021 |Article ID 8857458 | https://doi.org/10.1155/2021/8857458

Jaeyoung Lee, Suyi Mao, Mohamed Abdel-Aty, Yanqi Lian, Lishengsha Yue, Ilsoo Yun, "Association between Truck Crashes due to Mechanical Failure and Truck Age", Journal of Advanced Transportation, vol. 2021, Article ID 8857458, 7 pages, 2021. https://doi.org/10.1155/2021/8857458

Association between Truck Crashes due to Mechanical Failure and Truck Age

Academic Editor: Dilum Dissanayake
Received02 Aug 2020
Revised31 Mar 2021
Accepted07 Apr 2021
Published15 Apr 2021

Abstract

There have been efforts to restrict older trucks in many jurisdictions all over the world. The primary goal of the restrictions is to minimize greenhouse gas emissions. In addition to the environmental benefits, it is also possible that the truck age restriction could contribute to the enhancement of traffic safety. Older trucks are subject to longer travel-miles than newer trucks and tend to have higher mechanical failure rates. Extremely few studies have been done to explore the impact of trucks’ age on their crash occurrence due to mechanical problems. This study aims to investigate the association between the truck crashes due to mechanical issues and the truck age. Two approaches are adopted to achieve the objective. First, a chi-square test reveals that the proportions of the mechanical failures among older trucks are higher than those among newer ones ( = 256.199, ). Second, the modeling results indicate that the number of truck crashes due to mechanical failures is significantly increased by the truck age. The findings suggest that policies restricting older trucks should consider not only environmental effects but also traffic safety benefits.

1. Introduction

In the recent decades, a number of governments have started restricting older trucks on highways. The main goal of the restriction is to minimize environmental impacts from the older trucks. According to the California Air Resources Board (CARB), nearly all trucks and buses will be required to be equipped with 2010 or newer model year engines to reduce the particulate matter and oxides of nitrogen by January 1, 2023 [1]. Some governments encourage truck drivers to replace old trucks with new ones by providing incentives [2]. Incentive-based, voluntary vehicle replacement programs were found effective to reduce harmful air emissions by about 4%. Nevertheless, such voluntary programs are less effective compared to mandatory programs [3]. In addition, low emission zones (LEZs) programs restrict access by the most polluting vehicles. LEZs have been implemented in many cities worldwide, particularly in European Union countries [4]. Several studies have proven that LEZs are effective to decrease air pollutants’ concentrations in urban areas by restricting old trucks [5, 6].

As shown above, policies and research studies related to the old truck restrictions have focused on the environmental impacts. In addition to the environmental benefits, the old truck ban policies might lead to the improvement of traffic safety. Extremely few studies have explored traffic safety of the old trucks. Conrad [7] asserted that older trucks are subject to longer travel-miles of wear than newer ones and eventually become unsafe, without a proper maintenance. Along with the wear, older trucks are less likely to be equipped with safety devices to prevent mechanical failures. In the United States, Antilock braking systems (ABS) have been mandated on tractors since 1997, and then ABS and air brakes have been required on semitrailers since 1998. Moreover, electronic stability control (ESC) was recently mandated for three-axle truck tractors manufactured after 2017 (49 CFR § 571). These mandatory safety devices for new trucks are expected to considerably reduce traffic crashes related to mechanical failure. In other words, older trucks without such equipment might be more exposed to danger. However, no study has examined the effect of truck’s age on the number of traffic crashes occurring because of mechanical problem.

This study consists of two parts: (1) comparing the proportions of truck-involved crashes due to mechanical failure and truck-involved crashes due to nonmechanical issue and (2) estimating a crash prediction model to identify the effect of truck age on the number of crashes occurring due to mechanical failure. The rest of the paper is organized as follows. First, the data collection and processing procedure are described. Subsequently, the methodologies used in this analysis are explained, followed by the results and discussions. The final section summarizes and concludes the manuscript.

2. Data

In order to achieve the main objective of this study, data from two sources were collected and processed. The first data is obtained from the Fatality Analysis Reporting System (FARS) database, which is archived and managed by the US National Highway Traffic Safety Administration (NHTSA). The FARS is a nationwide census providing public yearly data regarding fatal injuries suffered in motor vehicle traffic crashes. The NHTSA archives data from 1975 to 2018. The data of truck-involved crashes that occurred between 2014 and 2018 were collected. Light trucks less than 10,000 lbs. GVWR (gross vehicle weight rating), such as pickup trucks and sport utility vehicles, were removed from the data. Overall, 21,726 truck-involved crashes occurred over the five years. The types of trucks and their counts are summarized in Table 1. As shown in the table, the most frequently observed type is truck-tractor (64.5%), and 25.6% are single-unit straight truck or Cab-Chassis (regardless of gross vehicle weight rating or GVWR).


Truck typeCountPercentage

Truck-tractor (cab only, or with any number of trailing unit)14,01564.51
Single-unit straight truck or cab-chassis (26,000 lbs + GVWR)3,02413.92
Single-unit straight truck or cab-chassis (19,500–26,000 lbs. GVWR)1,1675.37
Single-unit straight truck or cab-chassis (10,000–19,500 lbs. GVWR)1,3646.28
Medium/heavy pickup (10,000 lbs. + GVWR)1,2655.82
Single-unit straight truck or cab-chassis (GVWR unknown)4392.02
Unknown medium/heavy truck type890.41
Camper or motorhome, unknown truck type850.39
Medium/heavy truck based motorhome1200.55
Step van (10,000 lbs. GVWR)740.34
Unknown if single unit or combination medium truck heavy truck (26,000 lbs. + GVWR)470.22
Unknown if single unit or combination medium truck (10,000–26,000 lbs. GVWR)280.13
Unknown truck type (light/medium/heavy)90.04
Total21,726100.00

Approximately 3.13% of the truck-involved crashes happened due to the mechanical failure of the trucks. Although the mechanical failure has a small proportion in the truck crashes, they are preventable with regular maintenance and management. Mechanical problem information was processed from the data file: contributing circumstances of the FARS. The contributing circumstances are defined as the vehicle’s possible preexisting defects or maintenance conditions that may have contributed to the crash. The detailed mechanical problems and distribution are presented in Table 2. It is shown that tires and brake system are the majority of the mechanical problems (43.1% and 29.6%, respectively), followed by the problems of power train or exhaust system (6.0%).


Mechanical problemCountPercentage

Tires29343.1
Brake system20129.6
Power train or exhaust system416.0
Combination of two or more problems253.7
Head, signal, or other light324.7
Truck coupling/trailer hitch/safety chains202.9
Wheels91.3
Suspension71.0
Steering91.3
Body, doors50.7
Safety systems40.6
Wipers/windows/windshield20.3
Vehicle contributing factors-no details324.7
Total680100.0

The FARS provides the model year of the trucks that were involved in crashes. Based on the model year, truck age was calculated by subtracting the truck model year from the calendar year at the date of the crash. The distribution of the truck age is shown in Figure 1. About 62% of the trucks are less than ten years old, while 38% are ten years and older.

Furthermore, highway statistics of the Federal Highway Administration (FHWA) provides the number of registered trucks in the United States [8]. As of 2018, there were 152,701,631 registered trucks in the country. Nevertheless, the FHWA does not provide the registered trucks by age. Thus, the number of registered trucks by age was estimated using the “quasi-induced exposure” method. More details are provided in the Methodology section.

3. Methodology

As described in the Introduction, the current analysis has two parts. The first one is to compare the proportions of truck-involved crashes due to mechanical failure and nonfault (both not-at-fault driver and nonmechanical failure). The second one is to estimate a crash prediction model to identify the safety effects of truck age. For the first part, a chi-square test with the contingency table was conducted if the difference of the proportions is statistically significant. For the second part, a crash prediction model is developed for estimating the effects of truck age on safety. From the modeling result, a crash modification factor (CMF) that shows the safety effects of truck aging is estimated.

A Poisson model has been widely used for crash frequency modeling; however, it is unsuitable if the data is overdispersed. A negative binomial model is an alternative to Poisson model, which overcomes possible overdispersion in the data. A negative binomial model is specified as follows [9]:where is the observed number of truck crashes occurring due to mechanical failure for each truck age i, is the Poisson mean of , is the intercept term, are the coefficient estimates of model covariates , and is a gamma-distributed error component with mean 1 and variance . The relationship between the variances from the mean in the negative binomial model is as follows:

The parameter is referred to as the overdispersion parameter. If the overdispersion parameter is zero, the negative binomial converges to a Poisson model. In other words, the crash data is not dispersed (i.e., ).

A CMF from the estimated model is calculated by exponentiating the coefficient of the variable of interest (i.e., truck age) [1012]. Along with other count modeling approaches for crashes, an exposure variable is required in the modeling estimation. Nevertheless, the number of trucks by each truck age is not obtainable. For this reason, a quasi-induced exposure method is employed to estimate the number of trucks by age.

3.1. Quasi-Induced Exposure Approach

The quasi-induced exposure method is applied to estimate the increase of being involved in a crash associated with driver-related or vehicle-related characteristics when there is no direct way to measure the intensity of exposure of the characteristics [13]. The basic idea of the quasi-induced method is that nonfault vehicles (e.g., no mechanical failure, no violation, and no alcohol/drugs) involved in traffic crashes are an approximately random sample of the vehicle population [1215].

Some might be concerned of assuming that non-at-fault trucks are a random sample of trucks because, in fatal multivehicle crashes involving a truck, the truck driver is likely to survive and provide a narrative that might blame that the truck had mechanical faults. Thus, in this study, “nonfault trucks” include not only not-at-fault drivers’ trucks (e.g., with a traffic violation), but also nonmechanical failure trucks. In this paper, the percentage of trucks of each age in nonfault trucks is estimated using this assumption. The number of trucks for each age is calculated by multiplying the percentage of each age and the total number of registered trucks (as of 2018) from the FHWA statistics [8].where is the number of trucks for age i, is the number of nonfault trucks in crashes for age i, and is the total number of registered trucks.

4. Results

For the first analysis, a chi-square test with the contingency table was conducted to check if there is a significant difference between the proportions of truck-involved crashes due to mechanical failure and nonmechanical failure by four age groups (i.e., 0–9, 10–19, 20–29, and 30–40 years old). Table 3 presents the contingency table and the chi-square test result.


Classification0–9 years10–19 years20–29 years30–40 yearsSum

Nonfault12,2456,0081,09913219,484
Mechanical failures25529810421678
Sum12,5006,3061,20315320,162
Percentage of the mechanical failures2.0%4.7%8.6%13.7%3.4%
Chi-square test256.19906 ( < 0.0001)

The contingency table shows that only 2.0% of the truck-involved crashes occurred due to mechanical failure for the trucks aged 0–9 years. On the other hand, the percentages went up to 4.7% to 8.6% for the trucks 10–19 years and 20–29 years old, respectively. The percentage is the highest among the trucks aged 30–40 years (13.7%). The chi-square test ( = 256.19906,  < 0.0001) confirmed that the proportions are significantly different by the age group.

For the second analysis, the number of trucks for each age was estimated using the quasi-induced exposure method before developing the crash model (Table 4).


AgeNo. of nonfault trucks% of nonfault trucksEstimated number of trucksMechanical failures

0226511.6217,751,44722
117789.1313,934,69017
216168.2912,665,05035
312516.429,804,44215
410405.348,150,77525
57984.106,254,15226
67513.855,885,80015
78474.356,638,17937
89224.737,225,97532
99775.017,657,02631
109574.917,500,28040
119234.747,233,81337
126993.595,478,26128
135662.904,435,90221
145502.824,310,50630
155832.994,569,13628
165142.644,028,36428
174832.483,785,40832
184212.163,299,49630
193121.602,445,23224
202481.271,943,64626
212021.041,583,13113
221480.761,159,9186
231240.64971,82310
24930.48728,8677
25730.37572,12210
26560.29438,8889
27630.32493,7498
28430.22337,0036
29490.25384,0279
30240.12188,0954
31280.14219,4442
3280.0462,6980
33120.0694,0472
3470.0454,8613
3590.0570,5360
3680.0462,6982
3790.0570,5363
38130.07101,8853
3950.0339,1861
4090.0570,5361
Sum19,484100.0152,701,631678

Using the estimated trucks by age and the crash data, the rates of mechanical failures per estimated trucks can be calculated (Table 5). The mechanical failure rate of 0–9 years is 2.657, and that of 10–19 years is 6.329. When the truck age is between 20 and 29 years, the crash rate went up to 12.075. The highest crash rate is observed in the age 30–40 years old group, which is 20.299. The crash rates indicate that the older trucks are more likely to have mechanical problems.


Age groupNo. of mechanical failuresEstimated number of trucksMechanical failure rates per million trucks

0–9 years25595,967,5362.65715
10–19 years29847,086,3996.32879
20–29 years1048,613,17512.07453
30–40 years211,034,52120.29924
Sum678152,701,6314.44003

In the modeling process, the authors firstly attempted a negative binomial model; however, the overdispersion parameter was quite small (α = 0.068). Thus, a Poisson model was developed and compared with the negative binomial model using likelihood ratio test. The test indicates no significant difference between the two models’ performance (χ2 = 0.7, df = 1,  = 0.403). Also, the AIC of the Poisson model (229.5) was slightly smaller than that of the negative binomial model (230.8). Therefore, the Poisson model was chosen as the final model in this study.

The Poisson modeling results are summarized in Table 6. The model was estimated for the number of truck crashes occurring due to the mechanical failure. The estimated exposure variable: the natural logarithm of estimated trucks well explains the variations in the mechanical failure truck-involved crash data. The truck age variable has a significant positive association with the number of truck’s mechanical failure.


VariableCoefficientStandard error

Intercept−15.82561.77368<0.0001
Log (estimated number of trucks)1.160420.10521<0.0001
Truck age0.099020.01413<0.0001

Goodness-of-FitStatisticEquation
Log-likelihood (full)−111.37728
Log-likelihood (null)−310.88424
McFadden’s R20.64059
MAE (median absolute error)3.41879
RMSE (root mean squared error)4.93716
MAPE (mean absolute percentage error)0.20674

is the observed crash count at truck age i, and is the predicted crash count at truck age i. Significant at 99% confidence level.

5. Discussion of the Results

The most important finding from the analysis is that older trucks are more likely to cause a crash due to mechanical failure. From the coefficient and the standard deviation, a CMF and its 95% confidence interval were estimated (Table 7). The estimated CMF for one-year age increase is 1.10409, which suggests that an older truck is 10.41% more likely to have a mechanical failure. The 95% confidence interval of the CMF indicates the probability to face a mechanical failure for an older truck by one year is increased by 7.39%–13.51%.


FactorCMF95% confidence interval
Lower limitUpper limit

Truck age1.104091.073921.13510

Figure 2 describes the relationship between the rate of mechanical failure (multiplied by million) and the truck age. The predicted rates are quite accurate when the truck age is younger than 20 years old, while the errors are larger after 20 years old. It is because the number of mechanical failures for the trucks younger than 20 years old is sufficient, whereas the number for trucks older than 20 years old is extremely small, particularly after 30 years old. The small number of mechanical failures resulted in large fluctuations in the older trucks (see Table 4).

6. Summary and Conclusions

The study investigated the relationship between the number of truck crashes due to mechanical failure and truck age. The first analysis compared the proportion of mechanical failure crashes by age group. The result showed that only 2.0% of the truck-involved crashes occurred due to mechanical failure for the trucks aged 0–9 years. On the other hand, the percentages went up to 4.7% and 8.6% for 10–19 years and 20–29 years old trucks, respectively. The percentage is the highest among the trucks aged 30–40 years (13.7%). The chi-square test indicated that the difference in the proportions is statistically significant. In the second analysis, a Poisson model was developed to find the association between the number of mechanical failure truck crashes and truck age. The modeling results identified that the probability of crash occurrence due to mechanical failure is increased by 10.41% for a one-year older truck. From those findings, it is concluded that the number of truck crashes involving mechanical failure is significantly increased by truck age.

In the latest decade, old truck restriction policies have been initiated in many jurisdictions. For example, California will require all trucks to have 2010 model year engines or newer by 2023 [1]. Such policies mainly focus to reduce air pollution. However, it is strongly recommended that policy makers consider restricting older trucks with more diverse aspects. Even in areas without serious air pollution, the old truck restrictions should be considered from the viewpoint of traffic safety. The results from this study can be supporting such policies in the safety aspects as it was shown that trucks older than 10 years are significantly likely to have mechanical failures. Also, not only restriction, but also providing incentives would be effective to encourage truck owners to replace their old trucks with new ones. If such restrictions or incentive programs are difficult to be immediately employed, proper and frequent maintenance and stricter mandatory inspections for old trucks would be required. Moreover, truck manufacturers can understand the relationship between age and mechanical failure from the findings from this study, and they might want to review their maintenance checklist by truck age for drivers, which might be helpful to prevent mechanical failure truck crashes.

Currently, it is hard to think that old trucks have intelligent connected vehicles (ICV) equipment. However, in the future, old trucks at that time will be probably equipped with ICV devices. If this is the case, real-time safety assessment will be available, and when a higher risk is anticipated due to the mechanical failure in specific situations (e.g., poor pavement condition with relatively old tire, slippery surface with dated steering), adjacent connected vehicles will be warned to keep a sufficient distance from the truck (V2V). If a truck is identified with a higher risk of mechanical failure, and the current traffic/road environment is dangerous, the infrastructure will be able to send an advisory message (e.g., slow down) to the truck (I2V) [16, 17]. These types of proactive ICV countermeasures might decrease the number of crashes due to truck’s mechanical failure.

Although this study identified key findings, it is not without limitations. Those limitations should be addressed in follow-up studies. First, the estimated number of trucks by age was included in the model using the quasi-induced exposure method. It is necessary to validate whether the estimated trucks are accurate and reliable. Second, the data used in this study was collected from the FARS, which only provides crashes involving a fatality. Since a fatal crash has a tendency to be investigated with a reliance on narratives of survivors in the absence of other evidence, the recorded contributing factors might not be very trustworthy. For these reasons, it is possible that the results are biased. Lastly, the study was conducted only using crash data. The findings from this study would be more reliable if the results are substantiated by actual truck inspections.

Data Availability

The data are available from Fatality Analysis Reporting System (FARS).

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Authors’ Contributions

Study conception and design were conducted by Jaeyoung Lee, Mohamed Abdel-Aty, and Ilsoo Yun; data collection was carried out by Jaeyoung Lee; analysis and interpretation of results were carried out by Jaeyoung Lee, Suyi Mao, Mohamed Abdel-Aty, Yanqi Lian, and Lishengsha Yue; draft manuscript preparation was carried out by Jaeyoung Lee, Suyi Mao, Yanqi Lian, and Lishengsha Yue. All authors reviewed the results and approved the final version of the manuscript.

Acknowledgments

This study was funded by the National Key R&D Program of China (2020YFB1600400) and Innovation-Driven Project of Central South University (2020CX013).

References

  1. California Air Resources Board (CARB), Truck and Bus Regulation, CARB, Sacramento, CA, USA, 2021.
  2. F. Posada, D. V. Wagner, G. Bansal, and R. Fernandez, Survey of Best Practices in Reducing Emissions through Vehicle Replacement programs ICCT White Paper, The International Council on Clean Transportation, Washington, DC, USA, 2015.
  3. M. Norsworthy and E. Craft, “Emissions reduction analysis of voluntary clean truck programs at US ports,” Transportation Research Part D: Transport and Environment, vol. 22, pp. 23–27, 2013. View at: Publisher Site | Google Scholar
  4. H. Wolff and L. Perry, “Policy monitor,” Review of Environmental Economics and Policy, vol. 4, no. 2, pp. 293–308, 2010. View at: Publisher Site | Google Scholar
  5. D. Dias, O. Tchepel, and A. P. Antunes, “Integrated modelling approach for the evaluation of low emission zones,” Journal of Environmental Management, vol. 177, pp. 253–263, 2016. View at: Publisher Site | Google Scholar
  6. C. Ezeah, K. Finney, and C. Nnajide, “A critical review of the effectiveness of low emission zones (lez) as a strategy for the management of air quality in major European cities,” Journal of Multidisciplinary Engineering Science and Technology, vol. 2, no. 7, pp. 1860–1868, 2015. View at: Google Scholar
  7. I. J. ., L. Conrad, Log Truck Liability Insurance in Georgia: Costs, Trends, and Solutions, Georgia Forestry Association Center, Forsyth, GA, USA, 2017.
  8. Federal Highway Administration (FHWA), Highway Statistics 2018, Federal Highway Administration (FHWA), Washington, DC, USA, 2019.
  9. D. Lord and F. Mannering, “The statistical analysis of crash-frequency data: a review and assessment of methodological alternatives,” Transportation Research Part A: Policy and Practice, vol. 44, no. 5, pp. 291–305, 2010. View at: Publisher Site | Google Scholar
  10. F. Gross, B. N. Persaud, and C. Lyon, A Guide to Developing Quality Crash modification Factors (No. FHWA-SA-10-032), Federal Highway Administration, Washington, DC, USA, 2010.
  11. J. Lee, M. Abdel-Aty, and E. Nyame-Baafi, “Investigating the effects of pavement roughness on freeway safety using data from five states,” Transportation Research Record: Journal of the Transportation Research Board, vol. 2674, no. 2, pp. 127–134, 2020. View at: Publisher Site | Google Scholar
  12. L. Yue, M. Abdel-Aty, J. Lee, and A. Farid, “Effects of signalization at rural intersections considering the elderly driving population,” Transportation Research Record: Journal of the Transportation Research Board, vol. 2673, no. 2, pp. 743–757, 2019. View at: Publisher Site | Google Scholar
  13. V. Martínez-Ruiz, P. Lardelli-Claret, E. Jiménez-Mejías, C. Amezcua-Prieto, J. J. Jiménez-Moleón, and J. d. D. Luna del Castillo, “Risk factors for causing road crashes involving cyclists: an application of a quasi-induced exposure method,” Accident Analysis & Prevention, vol. 51, pp. 228–237, 2013. View at: Publisher Site | Google Scholar
  14. M. Keall and S. Newstead, “Induced exposure estimates of rollover risk for different types of passenger vehicles,” Traffic Injury Prevention, vol. 10, no. 1, pp. 30–36, 2009. View at: Publisher Site | Google Scholar
  15. N. Stamatiadis and J. A. Deacon, “Quasi-induced exposure: methodology and insight,” Accident Analysis & Prevention, vol. 29, no. 1, pp. 37–52, 1997. View at: Publisher Site | Google Scholar
  16. S. Mao, G. Xiao, J. Lee, L. Wang, Z. Wang, and H. Huang, “Safety effects of work zone advisory systems under the intelligent connected vehicle environment: a microsimulation approach,” Journal of Intelligent and Connected Vehicles, vol. 18, 2021. View at: Publisher Site | Google Scholar
  17. Z. Wang and J. Lee, “Enhancing construction truck safety at work zones: a microscopic traffic simulation study,” IEEE Access, vol. 9, p. 1, 2021. View at: Publisher Site | Google Scholar

Copyright © 2021 Jaeyoung Lee et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.


More related articles

 PDF Download Citation Citation
 Download other formatsMore
 Order printed copiesOrder
Views159
Downloads162
Citations

Related articles

Article of the Year Award: Outstanding research contributions of 2020, as selected by our Chief Editors. Read the winning articles.