1 Introduction
Container semitrailers are extensively used for medium- and long-distance road transportation because they can transport large cargo volumes, are efficient, and are easy to load and unload. However, they also have a large loading mass and high center of gravity; moreover, interactions between tractors and semitrailers often lead to poor lateral stability when driving on curves, and these vehicles are prone to rollover accidents [1–3]. Freeway interchange loop ramps have a small radius and large front- and rear-road alignment changes; thus, they might be unsafe in nonroutine scenarios, and the rate of container semitrailer rollover accidents on these curves is high [4, 5]. Therefore, studying the mechanisms of container semitrailer rollover accidents on freeway loop ramps is critical for improving freeway operation management, safety, and service level [6, 7].
Container semitrailers comprise tractors and semitrailers, which are connected using a traction pin and saddle. In contrast to a single-unit truck, a semitrailer is influenced by the coupling effects of the two vehicle bodies, and its dynamics are more complex than are those of a single-unit truck. When driven on curves, semitrailers not only have similar instabilities to single-unit trucks but also have unique instabilities, which primarily manifest as yaw and roll instabilities [8]. Vehicle, road, and weather factors affect the occurrence of lateral instability in semitrailers. Several scholars have investigated the vehicle factors by establishing vehicle dynamics models by using simulation software [9–13]. Some researchers have also used simulations to investigate the influence of geometric road parameters and the road pavement friction coefficient on the stability of semitrailers moving on curves; these parameters include the radius of horizontal curves, superelevation, and longitudinal gradient [14–16]. Semitrailers are large; thus, some scholars have investigated the effect of crosswind on their lateral stability under severe weather conditions by conducting wind tunnel tests or using the TruckSim simulation software [17–20]. To improve the safety of semitrailers moving on curves, some academics have established models of semitrailer rollovers to predict risk to warn drivers and enable drivers to take timely corrective action by adjusting the vehicle body’s attitude, thereby preventing rollovers [21–25]. In summary, investigations of the mechanisms of container semitrailer rollover accidents while driving on curves are primarily conducted through simulations. Because real-world road tests of semitrailer rollovers are risky, are expensive, and have poor repeatability, few scholars have performed real-world semitrailer tests for verification. In studies that performed such tests [26, 27], to ensure safety, the driving speed of semitrailers did not reach the limit value for rollover. The results of simulation studies without real-world vehicle verification and real-world data have discrepancies.
In response to this challenge, an accident case in which a container semitrailer rolled over on the loop ramp of a trumpet-shaped interchange was investigated in this research. This case was simulated by establishing vehicle, road, and driving behavior models in TruckSim on the basis of data obtained from an in-depth investigation of the accident, such as road geometry parameters and the vehicle driving status, and validating the established models and evaluation indices. We then investigated the influences of vehicle and road factors on the rollover of the container semitrailer and obtained the vehicle’s speed threshold and rollover state for each factor. The findings of this study can be used to provide recommendations for the safe operation of container semitrailers on loop ramps and as references for future research in this area.
2 Materials and methods
2.1 Background of the accident case
On June 18, 2022, a rollover accident occurred while a container semitrailer was traveling along the S18 Xinkai Freeway to the Xinkai interchange loop ramp in Yueyang, Hunan Province, China; this accident resulted in the driver and passenger dying instantly. Fig 1 depicts the accident vehicle’s ultimate stopping state. The accident vehicle was a container semitrailer with a two-axle tractor and two-axle semitrailer (Fig 1). The tractor’s curb weight and outline dimensions were 6800 kg and 6.04 m × 2.495 m × 3.6 m (length, width, and height), respectively. The overall dimensions of the heavy semitrailer were 15.75 m × 3.0 m × 4.45 m (length, width, and height), and the container dimensions were 15.75 m × 3.0 m × 3.45 m (length, width, and height); the curb weights of the semitrailer chassis and container were 3400 and 3650 kg, respectively. The mass of vehicle’s cargo was 17 210 kg. The road section on which the accident occurred is a right-turn loop ramp with a speed limit of 40 km/h, a radius of 60 m, a superelevation of 7%, and a longitudinal slope of −2.45%; the road is made of asphalt and was dry at the time of the accident. Fig 2 illustrates the accident vehicle’s path and its ultimate resting position.
[Figure omitted. See PDF.]
[Figure omitted. See PDF.]
2.2 Theoretical mechanical model
When a vehicle is driven at a constant speed on a horizontal straight road, the vertical loads of the left- and right-side tires are identical; the vertical loads on the front and rear axles are not transferred between the tires, and the vehicle is stable. When a vehicle enters a curved section of road, centrifugal forces and lateral tire friction combine to create a rollover moment, which causes the weight on the wheels on the inside of the curve to be transferred to the wheels on the outside of the curve. As the centrifugal force of the vehicle increases, the wheels on the inside of the curve begin to lift off the road, and the vehicle eventually overturns. If the curve has superelevation, the vehicle’s gravity component reduces the centrifugal force, thereby improving the vehicle’s driving stability. Fig 3 illustrates the forces acting on a vehicle moving around a curve under the assumption that the vehicle is a rigid body that is unaffected by tire and suspension deformation.
[Figure omitted. See PDF.]
In Fig 3, Fzl is the support force for the left-side tire, Fzr is the support force for the right-side tire, Fφl is the lateral friction force on the left-side tire, Fφr is the lateral friction force on the right-side tire, G is the total mass of the vehicle, Fg is the centrifugal force, hg is the height of the center of gravity, and B is the wheeltrack.
A critical vehicle rollover state occurs if the vertical force exerted by the road on the inner wheel is 0 N; at this time, the moment equation for the contact point between the outer wheel and the road is applicable. This equation is expressed as follows:(1)
Eq (2) is obtained by expanding Eq (1).
(2)
Where m is the total mass of the vehicle and r is the radius of the road’s horizontal curve.
Typically, the road’s cross-slope angle α is small; thus, cos α ≈ 1, and sin α ≈ tan α ≈ ih, where ih is the superelevation. Eq (3), which is obtained by rearranging Eq (2), is used to determine the critical speed vh at which rollover occurs for a vehicle traveling on a curve.
(3)
Eq (3) clearly reveals that (assuming that driver error, pavement damage, foreign objects, or other abnormal factors are absent), wheeltrack, the height of the center of gravity, superelevation, and turning radius are the main variables influencing the critical vehicle speed at which rollover occurs.
2.3 Construction of the simulation model
Mechanical Simulation Corporation created the vehicle dynamics simulation software known as TruckSim (version 2019.0, Mechanical Simulation Corporation, Ann Arbor, Michigan, USA) specifically for simulating the movements of trucks, buses, and trailers. In this study, vehicle, driver, and road models were established in TruckSim on the basis of the in-depth investigation data obtained from the studied accident.
2.3.1 Vehicle model.
In TruckSim, a model of the vehicle’s dynamics was constructed, including the entire vehicle arrangement, driving system, steering system, braking system, transmission system, and other elements. The “2A Cab Over w/2A Trl (Susp.Cab)” model was used in TruckSim as the vehicle model. Similar to the accident vehicle, this model comprises two elements, namely a heavy tractor and heavy semitrailer. Table 1 presents the key indicators of the heavy tractor.
[Figure omitted. See PDF.]
A semitrailer model was created in TruckSim by using the “15.5t Trailer, Tandem, Leaf” module; the key characteristics of this semitrailer are listed in Table 2. The semitrailer’s total mass is affected by its load; therefore, the weight was set as the actual cargo weight of the accident vehicle (17 210 kg). The longitudinal position of the cargo’s center of gravity is 8.2 m from the hitch, and its lateral position coincides with the container’s geometric center.
[Figure omitted. See PDF.]
The height of the center of gravity of a container semitrailer is primarily affected by the heights of the centers of gravity of the semitrailer chassis and container (including cargo), as presented in Eq (4) [28]. The height of the center of gravity of the semitrailer chassis is fixed, whereas the container’s center of gravity is affected by the stacking height and mass distribution of the loaded cargo. In this study, the mass of the cargo was assumed to be uniformly distributed, and the cargo stacking height was assumed to be the container’s height.(4)where h1 and h2 are the heights of the center of gravity of the semitrailer chassis center and container (including cargo), respectively. Moreover, G1 and G2 denote the masses of the semitrailer chassis and container (including cargo), respectively.
2.3.2 Road model.
The geometric characteristics of the road and the pavement friction coefficients were input to the TruckSim three-dimensional road module to complete the road model. The “U-Road, 45.72-m (150-ft) Radius” module was selected to create a roadway horizontal alignment comprising the following segments sequentially: a straight line, transition curve, circular curve, transition curve, circular curve, transition curve, and straight line. For the road longitudinal section, the “Road: Reference Line Elevation” module was selected with the sample interpolation and linear extrapolation settings to ensure that the longitudinal section was smoothly aligned. The “Road: Off-Center Elevation Map, S-L Grid” module was selected as the road cross-section setting, and two-dimensional sample bar interpolation and extrapolation methods were adopted to ensure smooth transitions for the road section’s cross-slope. The ramp studied in this paper is the dry asphalt pavement, and the existing research results show [29] that the adhesion coefficient of dry asphalt pavement is mainly affected by the roughness of the pavement and tire, the contact area between the tire and pavement, and the speed of the vehicle traveling, and the range of its value is 0.7–0.8. Therefore, the friction coefficient of the road pavement was selected from the “Road: Friction Map, S-L Grid” module and set to a minimum value of 0.7.
2.3.3 Driving behavior model.
The TruckSim driver control model primarily comprises vehicle speed control, shift control, steering control, and braking control. The vehicle was set to be driven at various constant speeds. An automatic shift control model was used to shift gears in response to changes in driving speed. The braking control module was a braking control model available in TruckSim that enables the vehicle to maintain a constant speed on downslopes. The steering control module was a closed-loop control model with the driver’s preview time and lag time set as 1.5 and 0.15 s [30], respectively. The road centerline was used as the target trajectory for the driving vehicle, and the movement of the vehicle was controlled in accordance with the target trajectory.
2.4 Rollover condition
The ground support force acting on the wheels on the inside of a curve is a key indicator of a vehicle’s tendency to roll over on the curve. The lateral load transfer ratio (LTR), which is defined as the ratio of the difference between the vertical loads of the left and right wheels of the same axle to the total vertical load, was selected as the evaluation index for determining rollover. This index can accurately reflect the tendency of the vehicle to roll over during turning and driving, as presented in Eq (5). The LTR has values in the range of [0, 1]. The LTR is 0 when the vertical loads of the wheels on both sides are equal. Vehicles begin to roll over if the wheels on the inside of a curve have a vertical load of 0 N; in this case, the LTR is 1, and the vertical weight acting on the wheels on the inside of the curve is entirely transferred to the wheels on the outside of the curve. According to relevant studies [31, 32], vehicles are in a safe state when 0 < LTR ≤ 0.2, a low-rollover-risk state when 0.2 < LTR ≤ 0.6, a greater-rollover-risk state when 0.6 < LTR ≤ 0.8, and a severe-rollover-risk state when 0.8 < LTR ≤ 1.(5)where n = 1, 2, 3, 4. Moreover, Fzln and Fzrn are the ground support forces acting on the left and right wheels of the nth axle, respectively.
3 Simulated reconstruction of the accident case
From the semitrailer chassis and container box size data in the in-depth accident report, the height of the semitrailer cargo’s center of gravity (h3) was calculated to be approximately 2.725 m. Assuming that the cargo mass was uniformly distributed, the container semitrailer in the considered accident case was fully loaded with a cargo mass of 17 210 kg. A trumpet-type interchange loop ramp model was developed from field-measured values obtained from the accident location and relevant design data; Table 3 lists the specific parameters.
[Figure omitted. See PDF.]
On the basis of the recorded driving speed and braking conditions from the accident vehicle’s traffic recorder, the container semitrailer entered the loop ramp at a speed of 77 km/h. The driver applied the brakes to reduce its speed to 55 km/h and subsequently released them. The semitrailer accelerated to 58 km/h because of gravity, at which point the driver applied the brakes until the vehicle rolled over. A brake control model based on these braking conditions was created in Simulink such that the speed changes of the simulated container semitrailer were similar to those in the accident case. Fig 4 presents the semitrailer speed and horizontal curve radius at each station (distance from the point at which the vehicle trajectory was start position). At station 349.930 m, the vehicle began to roll over at a speed of 58.9 km/h.
[Figure omitted. See PDF.]
As displayed in Fig 5, the vertical force acting on each semitrailer wheel changed during the simulation. Each wheel is denoted with a code in which L or R indicates the left or right side, respectively; a number between 1 and 4 indicates one of the four axles of the tractor and semitrailer; and i or o indicates an inner or outer wheel, respectively. For example, the inner wheel on the right side of the third axle is denoted as R3i. First, at station 349.930 m, the vertical force acting on R4o first reached 0 N. Second, the vertical force on R3o decreased to 0 N at station 356.120 m. Finally, the vertical force between R2o and R4i became 0 N at station 361.886 m. In summary, the right-side wheels of the fourth axle of the semitrailer lifted off the ground first. The tractor’s right wheels on the other axles continued to lift off the ground as the rollover worsened; eventually, the container semitrailer rolled over on a curve with a radius of 60 m. Fig 6 reveals that at station 361.878 m, the LTR of the semitrailer’s fourth axle reached the critical value of 1. These results indicate that LTR can accurately describe semitrailer rollovers. When the vertical force of R4o reached 0 N in the simulation, the semitrailer’s speed was 58.9 km/h. During the rollover, the recorder captured a maximal speed of 59 km/h. On the basis of Eq (3), the critical speed for a semitrailer to overturn is 59.1 km/h. Therefore, the critical rollover speed obtained from the simulation was similar to the theoretical and actual rollover speeds, which indicates that the simulation model is valid. In addition, Fig 5 reveals that when the container semitrailer was driven on the loop ramp, the vertical load on the left-side wheels of each axle increased sharply as the radius of the horizontal curve gradually decreased, whereas the vertical load on the right-side wheels of each axle decreased rapidly to 0 N. The vertical load on the right-side wheels of the container semitrailer’s fourth axle reached 0 N before entering a 60-m circular curve section, which indicated that the semitrailer had begun to roll over in the transition curve section (Figs 7 and 8), and the simulated rollover process was almost identical to the video captured by the accident vehicle’s traffic recorder.
[Figure omitted. See PDF.]
[Figure omitted. See PDF.]
[Figure omitted. See PDF.]
[Figure omitted. See PDF.]
Fig 7 displays the attitude of the accident vehicle body early in the overturning process. The accident vehicle’s overturning process is consistent with the visualization of the simulation results in Fig 8; in both cases, the semitrailer overturned first, followed by the tractor. Moreover, the simulated rollover behavior is the expected behavior for the actual change in the vertical force of each wheel.
Fig 9 presents an image of the simulated semitrailer rollover. A comparison of Figs 9 and 10 reveals that in both the simulation and actual accident, the left side of the vehicle body was in contact with the ground in the final resting position of the vehicle.
[Figure omitted. See PDF.]
[Figure omitted. See PDF.]
Figs 11 and 12 present top views of the final resting positions of the container semitrailer after overturning on the loop ramp in the simulated accident and actual accident, respectively; these positions are consistent. Moreover, on-site data collected at each station were compared with the corresponding simulated data; the difference between the stations of the final simulated and actual stopping positions of the overturned container semitrailer was less than 10 m, and all the corresponding radii of the circular curves were 60 m, which indicated that the simulation accurately reproduced the accident.
[Figure omitted. See PDF.]
(Simulation).
[Figure omitted. See PDF.]
(Accident case).
4 Results and discussion
The theoretical model of rollover mechanics, simulated rollovers for the actual accident, relevant parameters of the accident vehicle, and relevant parameters in Chinese standards were analyzed; these aspects indicate that the primary factors influencing the rollover risk for container semitrailers moving on curves include travel speed, the height of the cargo’s center of gravity, interactions between the tractor and the semitrailer, and the radius of the ramp curve. The produced TruckSim model was used for further investigating how these factors affect the risk of rollover for container semitrailers moving on loop ramps.
4.1 Travel speed
To analyze the effect of travel speed, simulations were performed with the road model established in Section 2.3.2. To ensure that other factors do not affect the rollover risk, the load model in the aforementioned simulations was a standard box model in which the load’s center of gravity coincided with the container’s geometric center at a height of 2.75 m, and the load weight was set as 17 210kg. The speed limit of the road section in which the accident occurred is 40 km/h; thus, speeds of 30, 35, 40, 45, and 50 km/h were selected for the simulation tests. The maximum LTR of the container semitrailer in the aforementioned section at each speed was determined through simulation (Table 4).
[Figure omitted. See PDF.]
Fig 13 depicts a plot and best-fit line of the simulation results for the speed and LTR. Eq (6), which was obtained through quadratic regression, represents the relationship between vehicle speed and LTR.
[Figure omitted. See PDF.]
(6)where v is the travel speed of the container semitrailer.
For Eq (6), R2 = 0.992, which indicates a perfect fit. Therefore, Fig 13 reveals that vehicle speed and LTR are significantly and positively correlated, with the correlation coefficient being 0.996 and the significance level being 0.01.
Table 5 suggests that the container semitrailer could successfully traverse the accident section by maintaining a speed below 43.5 km/h. If its speed reached 54.9 km/h, the LTR of its fourth axle reached the rollover limit value of 1, at which point the semitrailer lost stability and rolled over. After the semitrailer began to roll over, the vehicle continued to roll over even if the steering angle was reduced rapidly by turning the steering wheel to increase the semitrailer’s steering radius. In the accident case, the container semitrailer traveled at a maximum speed of 59 km/h while passing through the loop ramp during the rollover. Therefore, the vehicle was traveling at a speed considerably greater than the safe speed of 43.5 km/h for the aforementioned curve and greater than the critical rollover speed of 54.9 km/h. The generated lateral centrifugal force for a container semitrailer travelling at this critical speed is sufficient to cause the wheels on the inside of the curve to detach from the ground, which in turn causes the entire vehicle to roll over. As the container semitrailer passes through the transition curve and enters a road section with a radius smaller than 60 m, the lateral centrifugal force generated by the container semitrailer increases as the radius of the horizontal curve gradually decreases; therefore, the vehicle begins to roll over as it approaches the road section with a curve radius of 60 m. When the container semitrailer enters the circular curve section with the smallest radius from the transition curve section, the container semitrailer overturns completely.
[Figure omitted. See PDF.]
4.2 Influence of the height of the cargo’s center of gravity
The established road model was used for simulations to investigate the effect of the height of the cargo’s center of gravity on the lateral stability of a semitrailer traveling on a curve. The simulated test road had a curve with a radius of 60 m. A vehicle’s speed on a curve is positively correlated with its lateral stability. In the tests conducted using the aforementioned model, the semitrailer’s speed was set as 40 km/h (the speed limit of the road section) to control this effect. The Chinese standard “GB 1589–2016: Limits of Dimensions, Axle Load and Masses for Motor Vehicles, Trailers and Combination Vehicle” states that the maximum outer height of container semitrailers should be 4 m [33]. Assuming that the cargo mass in the container was uniformly distributed and that the cargo’s stack height and container height were the same, the height of the container floor for the accident vehicle was 1 m from the ground. The height of the cargo’s center of gravity is the height of the container floor above the ground plus half the container height. For container heights of 2.4, 2.7, 3.0, 3.3, and 3.6 m, the heights of the cargo’s center of gravity were 2.2, 2.35, 2.5, 2.65, and 2.8 m, respectively. Simulations were conducted to obtain the maximum values of the container semitrailer’s LTR for different heights of the cargo’s center of gravity when the semitrailer passed through a curve at a speed of 40 km/h (Table 6).
[Figure omitted. See PDF.]
Fig 14 displays a scatter diagram of the simulated results presented in Table 6. Linear regression was performed to determine the relationship between LTR and the height of the cargo’s center of gravity; this result is expressed in Eq (7).
[Figure omitted. See PDF.]
(7)where h is the height of the cargo’s center of gravity.
For the model expressed in Eq (7), the R2 was 0.999, which indicated a favorable fit. Correlation analysis of the best-fit line (Fig 14) revealed that the correlation coefficient between the LTR and the height of the cargo’s center of gravity for the turning container semitrailer was 1.00 at a significance level of 0.01, which indicated that a significant and positive correlation between these factors. In summary, for a container semitrailer traveling at a constant speed, an increase in the height of the cargo’s center of gravity causes the LTR to increase marginally.
If the total height of a container semitrailer is 4 m, Table 6 reveals that the container height is 3 m and that the height of the cargo’s center of gravity is 0.225 m less than that of the accident vehicle. The maximum LTR for a standards-compliant container semitrailer with a 2.5-m center of gravity that is traveling through a curve with a radius of 60 m at 40 km/h is 0.429. The standards-compliant container semitrailer must be traveling at 45.3 or 57.2 km/h (the critical speeds) to reach the LTR thresholds of 0.6 and 1, respectively. The accident vehicle was initially traveling at 77 km/h. A simulation was performed for a standards-compliant container semitrailer initially traveling at this speed. The simulation revealed that the LTR of both semitrailer axles reached 1, and the wheels on the inside of the curve left the ground. However, the LTR of the tractor axles had not yet reached this critical rollover value. Continuous braking then reduced the vehicle speed, and the vehicle passed through the accident section’s curve without rolling over. However, the actual accident vehicle was excessively tall; its container height was 3.45 m, which corresponds to a height of 2.725 m for the cargo’s center of gravity. In the simulation of this vehicle, the maximum LTR through the curve at a speed of 40 km/h was 0.467, and the LTR was 0.6 at a speed of 43.5 km/h. Moreover, the critical speed of the vehicle (LTR = 1) was 54.9 km/h. A comparison of the aforementioned cases revealed that under the same curve radius, the safe speed for the vehicle with a height of 2.5 m for the cargo’s center of gravity was 4.14% greater than that for the accident vehicle with a height of 2.725 m for the cargo’s center of gravity. This result indicates that ensuring that the height of a container does not exceed the maximum value in the relevant standards and avoiding the high stacking of heavy objects in the container to achieve a low cargo center of gravity can increase the safe speed of a semitrailer through a bend. When the horizontal curve radius is 60 m, Table 7 displays the speed thresholds for the rollover of container semitrailers under different cargo center gravity heights. The theoretical rollover speed thresholds are determined by Eq (3).
[Figure omitted. See PDF.]
(r = 60 m).
4.3 Tractor–semitrailer interactions
Semitrailers and tractors are hitched together through a saddle; each of these parts has freedom of lateral yaw and roll at the hitch such that the roll movement of the tractor during turning and driving is transferred to the semitrailer through the saddle. In addition, the large distance between the saddle and the rear axle of a semitrailer as well as the vehicle body’s flexibility can amplify the roll effects at the rear of the semitrailer. When a tractor and semitrailer are subjected to the same lateral acceleration but with the aforementioned rear amplification, the semitrailer can easily roll excessively, thereby resulting in the inner wheels of the trailer axle set leaving the ground because of an excessive transfer of lateral load. This phenomenon causes the inner wheels of the tractor drive axle and steering axle to leave the ground, and the entire vehicle eventually rolls over. To obtain the LTR for each axle of the semitrailer (Fig 15), the semitrailer model established in Section 2.3.1 was again applied in simulation experiments with a vehicle with a cargo center of gravity at 2.75 m and a cargo mass of 17 210 kg that passed through the road model (Section 2.3.2) at 40 km/h. During turning, the fourth axle exhibited the highest LTR among the axles, whereas the first axle exhibited the smallest LTR. Therefore, during turning, the lateral load transfer to the rear axle was most likely to be excessive, thereby resulting in the load on one side being completely transmitted to the other side and the wheels lifting off the ground followed by rollover.
[Figure omitted. See PDF.]
Fig 16 was obtained through the simulated reconstruction of the accident case. This figure shows the curves of the changes in the roll angles of the tractor and semitrailer. Before entering the segment with the minimum curve radius, the roll angle values of the tractor and semitrailer were similar; as the radius of the horizontal curve decreased, the roll angle values of the tractor and semitrailer increased. However, during rollover, the semitrailer’s roll angle increased considerably before the tractor’s roll angle; thus, the semitrailer tended to roll over first and subsequently caused the tractor to roll over. At a given station, the roll angle of the tractor was less than that of the semitrailer during rollover. The roll angle of the tractor was transmitted to the semitrailer through the saddle and then amplified in the semitrailer, thereby further increasing the roll angle of the semitrailer. Data from the accident vehicle’s recorder indicates that when the container semitrailer rolled over, the semitrailer’s roll angle was greater than the tractor’s roll angle (Fig 17). Therefore, in the accident, the semitrailer overturned first, and the tractor overturned as the roll angle of the semitrailer increased. A comparison of the accident case and simulation results revealed that both rollover processes were nearly identical; both processes clearly demonstrated that the tractor’s rollover angle was amplified in the semitrailer after being transmitted through the saddle.
[Figure omitted. See PDF.]
[Figure omitted. See PDF.]
4.4 Effect of horizontal curve radius
To investigate the effect of bend radius on the rollover of a container moving at 40 km/h, simulations were performed for circular curves with radii of 60, 70, 80, 90, and 100 m. These curves had a design speed of 40 km/h and a superelevation of 7% in accordance with the standard “Design Specification for Alignment (JTG D20-2017)” [34]. The simulated vehicle moved through each bend at 40 km/h. The height of the cargo’s center of gravity and its weight were set to those of the accident vehicle. The maximum LTR for each bend radius was obtained from the simulation (Table 8).
[Figure omitted. See PDF.]
The results presented in Table 8 were plotted as a scatter plot (Fig 18), and curvilinear regression was performed to obtain the best-fit line, which is expressed as follows:(8)
[Figure omitted. See PDF.]
For Eq (8), R2 = 0.999, which indicates a favorable fit. The correlation coefficient between the LTR and the bend radius was −0.985 at a significance level of 0.01; therefore, the LTR and bend radius were significantly and negatively correlated with each other. In the accident case, the loop ramp’s minimum circular curve radius was 60 m, which is the minimum circular curve radius for the design speed of 40 km/h. The critical speed is the minimum speed for a container semitrailer to overturn on a curve with a radius of 60 m. In the accident case, the excessive height of the vehicle resulted in a greater stacked cargo height and therefore higher center of gravity; the results presented in Section 4.2 reveal that this phenomenon decreased the critical rollover speed. If the overall height of the container semitrailer is 4 m (the maximum height limit of the relevant standard) and the container height is 3 m, driving through the circular curve section at a speed less than or equal to the design speed is safe because the critical speed is 57.2 km/h. However, the accident vehicle’s speed before the accident reached 59 km/h, and the container was excessively tall, which caused the vehicle to begin to roll over in the transition curve before entering the 60-m-radius segment of the circular curve. The simulation results indicate that as the radius of the circular curve increases, the critical rollover speed also increases. Therefore, in road design, extreme bend radii should be avoided; if the minimum bend radius must be used, stepped speed limit signs should be posted to ensure that vehicle speeds are substantially lower than the critical rollover speed. Table 9 shows the rollover speed thresholds of container semitrailers under different horizontal curve radii when the cargo center gravity height of the semitrailer is 2.725 m, where the theoretical rollover speed thresholds are obtained from Eq (3).
[Figure omitted. See PDF.]
(h3 = 2.725 m).
5 Conclusions
The main conclusions of this study are as follows:
* When a container semitrailer is traveling on a curve, speed significantly affects rollover risk. The speed limit for interchange loop ramps (such as in the accident case) where freeways connect with other freeways should be reduced in steps, and drivers should be reminded of the speed limit by providing speed measurements, reminders to drive cautiously, and warnings regarding the sharp curve such that container semitrailers can reduce their speed to a safe cornering speed before entering the curve section with the minimum radius.
* Taller container semitrailers have a higher center of gravity, which reduces the speed at which they can safely pass through a loop ramp and therefore significantly increases the risk of rollover when traveling at excessive speed. To decrease or completely prevent the risk of rollover for container semitrailers on circular ramps, relevant departments should increase their supervision of container heights.
* Container semitrailers are affected by interactions between the tractor and the semitrailer, and the increase in the height of the entire vehicle’s center of gravity when the container semitrailer is loaded causes the tractor’s roll to be amplified on the semitrailer during turning. When container semitrailer rollover occurs, the semitrailer first rolls over and then causes the tractor to roll over. To avoid rollover for turning container semitrailers, the change in the LTR of each axis of the semitrailer should be considered, and the vehicle’s steering angle should be adjusted, or the driving speed should be reduced in a timely manner.
* The minimum bend radius for a circular ramp affects the safe speed of a container semitrailer; therefore, the extreme bend radius in the specification should be avoided if possible in road design. If the minimum bend radius is used, speed limits and sharp curves should be clearly marked.
Supporting information
S1 Dataset.
https://doi.org/10.1371/journal.pone.0309139.s001
(XLSX)
S2 Dataset.
https://doi.org/10.1371/journal.pone.0309139.s002
(XLSX)
S3 Dataset.
https://doi.org/10.1371/journal.pone.0309139.s003
(XLSX)
S4 Dataset.
https://doi.org/10.1371/journal.pone.0309139.s004
(XLSX)
S5 Dataset.
https://doi.org/10.1371/journal.pone.0309139.s005
(XLSX)
References
1. 1. Beglini M, Belvedere T, Lanari L, Oriolo G. An intrinsically stable MPC approach for anti-jackknifing control of tractor-trailer vehicles[J]. IEEE/ASME Transactions on Mechatronics. 2022; 27 (6): 4417–4428.
* View Article
* Google Scholar
2. 2. Cai H, Xu X. Lateral stability control of a tractor-semitrailer at high speed[J]. Machines. 2022; 10 (8): 716.
* View Article
* Google Scholar
3. 3. Hou Y, Xu X. High-speed lateral stability and trajectory tracking performance for a tractor-semitrailer with active trailer steering[J]. PLOS ONE. 2022; 17(11): e0277358.
* View Article
* Google Scholar
4. 4. Hu JB, He LC, Wang RH. Review of safety evaluation of freeway interchange[J]. China Journal of Highway and Transport. 2020; 33(7): 17–28.
* View Article
* Google Scholar
5. 5. Liang ZL. Investigation on the Safety Evaluation of Expressway Interchange[D]. Guangzhou: South China University of Technology, 2018.
6. 6. Jiang C, Yin S, Yao Z, et al. Safety evaluation of mixed traffic flow with truck platoons equipped with (cooperative) adaptive cruise control, stochastic human-driven cars and trucks on port freeways[J]. Physica A: Statistical Mechanics and its Applications. 2024; 643: 129802.
* View Article
* Google Scholar
7. 7. Yao Z, Li L, Liao W, et al. Optimal lane management policy for connected automated vehicles in mixed traffic flow[J]. Physica A: Statistical Mechanics and its Applications. 2024; 637: 129520.
* View Article
* Google Scholar
8. 8. Sahin H, Akalin O. Articulated vehicle lateral stability management via active rear-wheel steering of tractor using fuzzy logic and model predictive control[J]. SAE International Journal of Commercial Vehicles. 2020; 13(02-13-02-0008): 115–128.
* View Article
* Google Scholar
9. 9. Xu M, Huang X, Zhang C, Shao DJ. Application of fuzzy synthesis evaluation to driving safety analysis of sharp curves on mountain expressways[J]. China Journal of Highway and Transport. 2016; 29(6): 186–197.
* View Article
* Google Scholar
10. 10. Kamel Salaani M, Grygier P A, Heydinger G J. Evaluation of heavy tractor-trailer model used in the National Advanced Driving Simulator [J]. SAE transactions. 2003; 1440–1454.
* View Article
* Google Scholar
11. 11. Law E H, Janajreh I. Effects of tire and vehicle design characteristics on rollover of tractor semi-trailers[J]. SAE transactions. 2004; 38–53.
* View Article
* Google Scholar
12. 12. Barbieri F A A, de Almeida Lima V, Garbin L, et al. Rollover study of a heavy truck combination with two different semi-trailer suspension configurations[R]. SAE Technical Paper, 2014.
* View Article
* Google Scholar
13. 13. Wang CL, Hu YQ, Li P. Safe driving speed threshold on curve section of road under different loading conditions based on TruckSim[J]. Journal of Beijing University of Aeronautics and Astronautics. 2018; 44(6): 1337–1346.
* View Article
* Google Scholar
14. 14. Jiang K, Zhang MY, Chen YK. Driving safety analysis of semi-trailer train at circular curve section in mountain area[J]. Journal of Traffic and Transportation Engineering. 2015; 15(03): 109–117.
* View Article
* Google Scholar
15. 15. Qu GX, He YL, Sun XD, Feng BB. Lateral Stability of Tractor-semitrailer on Freeway Horizontal Curve Section[J]. Journal of Beijing University of Technology. 2017; 43(12): 1907–1915.
* View Article
* Google Scholar
16. 16. Zhang S, Liu Y, Li P, et al. Study of instability mechanisms of trucks turning right at long downhill T-junctions based on Trucksim simulation[J]. PLOS ONE. 2023; 18(3): e0282779.
* View Article
* Google Scholar
17. 17. Zhang YP, Ding NG. Simulation of driving stability of a tractor semi-trailer in cross wind[J]. Journal of Automotive Safety and Energy. 2012; 3(2): 136–141.
* View Article
* Google Scholar
18. 18. Cheli F, Ripamonti F, Sabbioni E, et al. Wind tunnel tests on heavy road vehicles: Cross wind induced loads—Part 2[J]. Journal of Wind Engineering and Industrial Aerodynamics. 2011; 99(10): 1011–1024.
* View Article
* Google Scholar
19. 19. Salati L, Schito P, Cheli F. Strategies to reduce the risk of side wind induced accident on heavy truck[J]. Journal of Fluids and Structures. 2019; 88: 331–351.
* View Article
* Google Scholar
20. 20. Lin Q, Dou Z, Zhou G, et al. Influence analysis of crosswind on handling stability of tractor semi-trailer in double lane change situation[M]//ICTIS 2013: Improving Multimodal Transportation Systems-Information, Safety, and Integration. 2013; 1715–1721.
* View Article
* Google Scholar
21. 21. Hyun D, Langari R. Modeling to predict rollover threat of tractor-semitrailers[J]. Vehicle System Dynamics. 2003; 39(6): 401–414.
* View Article
* Google Scholar
22. 22. Chou T, Chu T W. An improvement in rollover detection of articulated vehicles using the grey system theory[J]. Vehicle System Dynamics. 2014; 52(5): 679–703.
* View Article
* Google Scholar
23. 23. Harun M H, Hudha K, Samin P M, et al. A modified odenthal rollover index algorithm for tractor-semitrailer using steering and vehicle speed inputs[J]. International Journal of Heavy Vehicle Systems. 2021; 28(4): 563–584.
* View Article
* Google Scholar
24. 24. Sanchez E N, Ricalde L J, Langari R, Shahmirzadi D. Rollover prediction and control in heavy vehicles via recurrent high order neural networks[J]. Intelligent Automation and Soft Computing. 2011;17(1): 95–107.
* View Article
* Google Scholar
25. 25. Wang D, Chen S, Zhang W, et al. The roll stability analysis of semi-trailer based on the wheel force[J]. CMC-COMPUTERS MATERIALS & CONTINUA. 2022; 71(1): 1837–1848.
* View Article
* Google Scholar
26. 26. Qu G, He Y, Sun X, et al. Modeling of Lateral Stability of Tractor‐Semitrailer on Combined Alignments of Freeway[J]. Discrete Dynamics in nature and Society. 2018; 2018(1): 8438921.
* View Article
* Google Scholar
27. 27. Gertsch J, Eichelhard O. Simulation of dynamic rollover threshold for heavy trucks[R]. SAE Technical Paper. 2003.
* View Article
* Google Scholar
28. 28. Xiao JY. Gonglu Dajian Yunshu Jishu. Beijing: China Communications Press; 2014.
29. 29. Shao JJ, Zhang HJ, Tan DR, et al. Research of safety speeds model on curve road by using Trucksim[J]. Journal of Guangxi University (Natural Science Edition). 2017; 42(03): 961–969.
* View Article
* Google Scholar
30. 30. Hu CB, Lin M, Tong SH. STUDY ON FORMATION MECHANISM AND CHARACTERISTICS OF INERTIA LOAD OF TRUCK IN CURVED ROAD SECTIONS[J]. Engineering Mechanics. 2022; 18(02): 198–205.
* View Article
* Google Scholar
31. 31. Zhang XB. Study on formation mechanism and safety improvement of accident-prone ramp on “T” shape hub interchange[J]. Journal of Safety Science and Technology. 2022; 18(02): 198–205.
* View Article
* Google Scholar
32. 32. Li L, Lu Y, Wang R, Chen J. A three-dimensional dynamics control framework of vehicle lateral stability and rollover prevention via active braking with MPC[J]. IEEE Transactions on Industrial Electronics. 2016; 64(4): 3389–3401.
* View Article
* Google Scholar
33. 33. National Public service platform for standards information, China. GB 1589–2016, limits of dimensions, axle load and masses for motor vehicles, trailers and combination vehicle[S]. 2016; Available from: https://std.samr.gov.cn/gb/search/gbDetailed?id=71F772D8118AD3A7E05397BE0A0AB82A.
34. 34. Ministry of Transport of the People’s Republic of China. Design Specification for Highway Alignment: JTG D20–2017[S]. Beijing: China Communications Press; 2017.
Citation: Zhang S, Yang X, Liu Y, Wang L, Li P, Li H, et al. (2024) TruckSim-based study of the rollover accident mechanisms of container semitrailers on freeway interchange loop ramps. PLoS ONE 19(9): e0309139. https://doi.org/10.1371/journal.pone.0309139
About the Authors:
Shibo Zhang
Roles: Funding acquisition, Investigation
E-mail: [email protected]
Affiliations: School of Automobile and Transportation, Xihua University, Chengdu, China, Sichuan Xihua Jiaotong Forensic Center, Chengdu, China
ORICD: https://orcid.org/0000-0003-3820-1601
Xuefeng Yang
Roles: Methodology, Software, Writing – original draft
Affiliations: School of Automobile and Transportation, Xihua University, Chengdu, China, Sichuan Xihua Jiaotong Forensic Center, Chengdu, China
Yishan Liu
Roles: Formal analysis, Software
Affiliation: School of Automobile and Transportation, Xihua University, Chengdu, China
Li Wang
Roles: Data curation, Methodology
Affiliation: School of Automobile and Transportation, Xihua University, Chengdu, China
Pingfei Li
Roles: Investigation, Validation
Affiliations: School of Automobile and Transportation, Xihua University, Chengdu, China, Sichuan Xihua Jiaotong Forensic Center, Chengdu, China
Hui Li
Roles: Investigation, Resources
Affiliations: School of Automobile and Transportation, Xihua University, Chengdu, China, Sichuan Xihua Jiaotong Forensic Center, Chengdu, China
Yong Luo
Roles: Funding acquisition, Visualization
Affiliation: Sichuan Vocational and Technical College of Communications, Chengdu, China
Yi Li
Roles: Data curation, Formal analysis
Affiliation: Traffic Police Headquarters of Sichuan Public Security Department, Chengdu, China
Qiaoyuan Liu
Roles: Validation, Visualization
Affiliation: Traffic Police Headquarters of Sichuan Public Security Department, Chengdu, China
[/RAW_REF_TEXT]
[/RAW_REF_TEXT]
[/RAW_REF_TEXT]
[/RAW_REF_TEXT]
1. Beglini M, Belvedere T, Lanari L, Oriolo G. An intrinsically stable MPC approach for anti-jackknifing control of tractor-trailer vehicles[J]. IEEE/ASME Transactions on Mechatronics. 2022; 27 (6): 4417–4428.
2. Cai H, Xu X. Lateral stability control of a tractor-semitrailer at high speed[J]. Machines. 2022; 10 (8): 716.
3. Hou Y, Xu X. High-speed lateral stability and trajectory tracking performance for a tractor-semitrailer with active trailer steering[J]. PLOS ONE. 2022; 17(11): e0277358.
4. Hu JB, He LC, Wang RH. Review of safety evaluation of freeway interchange[J]. China Journal of Highway and Transport. 2020; 33(7): 17–28.
5. Liang ZL. Investigation on the Safety Evaluation of Expressway Interchange[D]. Guangzhou: South China University of Technology, 2018.
6. Jiang C, Yin S, Yao Z, et al. Safety evaluation of mixed traffic flow with truck platoons equipped with (cooperative) adaptive cruise control, stochastic human-driven cars and trucks on port freeways[J]. Physica A: Statistical Mechanics and its Applications. 2024; 643: 129802.
7. Yao Z, Li L, Liao W, et al. Optimal lane management policy for connected automated vehicles in mixed traffic flow[J]. Physica A: Statistical Mechanics and its Applications. 2024; 637: 129520.
8. Sahin H, Akalin O. Articulated vehicle lateral stability management via active rear-wheel steering of tractor using fuzzy logic and model predictive control[J]. SAE International Journal of Commercial Vehicles. 2020; 13(02-13-02-0008): 115–128.
9. Xu M, Huang X, Zhang C, Shao DJ. Application of fuzzy synthesis evaluation to driving safety analysis of sharp curves on mountain expressways[J]. China Journal of Highway and Transport. 2016; 29(6): 186–197.
10. Kamel Salaani M, Grygier P A, Heydinger G J. Evaluation of heavy tractor-trailer model used in the National Advanced Driving Simulator [J]. SAE transactions. 2003; 1440–1454.
11. Law E H, Janajreh I. Effects of tire and vehicle design characteristics on rollover of tractor semi-trailers[J]. SAE transactions. 2004; 38–53.
12. Barbieri F A A, de Almeida Lima V, Garbin L, et al. Rollover study of a heavy truck combination with two different semi-trailer suspension configurations[R]. SAE Technical Paper, 2014.
13. Wang CL, Hu YQ, Li P. Safe driving speed threshold on curve section of road under different loading conditions based on TruckSim[J]. Journal of Beijing University of Aeronautics and Astronautics. 2018; 44(6): 1337–1346.
14. Jiang K, Zhang MY, Chen YK. Driving safety analysis of semi-trailer train at circular curve section in mountain area[J]. Journal of Traffic and Transportation Engineering. 2015; 15(03): 109–117.
15. Qu GX, He YL, Sun XD, Feng BB. Lateral Stability of Tractor-semitrailer on Freeway Horizontal Curve Section[J]. Journal of Beijing University of Technology. 2017; 43(12): 1907–1915.
16. Zhang S, Liu Y, Li P, et al. Study of instability mechanisms of trucks turning right at long downhill T-junctions based on Trucksim simulation[J]. PLOS ONE. 2023; 18(3): e0282779.
17. Zhang YP, Ding NG. Simulation of driving stability of a tractor semi-trailer in cross wind[J]. Journal of Automotive Safety and Energy. 2012; 3(2): 136–141.
18. Cheli F, Ripamonti F, Sabbioni E, et al. Wind tunnel tests on heavy road vehicles: Cross wind induced loads—Part 2[J]. Journal of Wind Engineering and Industrial Aerodynamics. 2011; 99(10): 1011–1024.
19. Salati L, Schito P, Cheli F. Strategies to reduce the risk of side wind induced accident on heavy truck[J]. Journal of Fluids and Structures. 2019; 88: 331–351.
20. Lin Q, Dou Z, Zhou G, et al. Influence analysis of crosswind on handling stability of tractor semi-trailer in double lane change situation[M]//ICTIS 2013: Improving Multimodal Transportation Systems-Information, Safety, and Integration. 2013; 1715–1721.
21. Hyun D, Langari R. Modeling to predict rollover threat of tractor-semitrailers[J]. Vehicle System Dynamics. 2003; 39(6): 401–414.
22. Chou T, Chu T W. An improvement in rollover detection of articulated vehicles using the grey system theory[J]. Vehicle System Dynamics. 2014; 52(5): 679–703.
23. Harun M H, Hudha K, Samin P M, et al. A modified odenthal rollover index algorithm for tractor-semitrailer using steering and vehicle speed inputs[J]. International Journal of Heavy Vehicle Systems. 2021; 28(4): 563–584.
24. Sanchez E N, Ricalde L J, Langari R, Shahmirzadi D. Rollover prediction and control in heavy vehicles via recurrent high order neural networks[J]. Intelligent Automation and Soft Computing. 2011;17(1): 95–107.
25. Wang D, Chen S, Zhang W, et al. The roll stability analysis of semi-trailer based on the wheel force[J]. CMC-COMPUTERS MATERIALS & CONTINUA. 2022; 71(1): 1837–1848.
26. Qu G, He Y, Sun X, et al. Modeling of Lateral Stability of Tractor‐Semitrailer on Combined Alignments of Freeway[J]. Discrete Dynamics in nature and Society. 2018; 2018(1): 8438921.
27. Gertsch J, Eichelhard O. Simulation of dynamic rollover threshold for heavy trucks[R]. SAE Technical Paper. 2003.
28. Xiao JY. Gonglu Dajian Yunshu Jishu. Beijing: China Communications Press; 2014.
29. Shao JJ, Zhang HJ, Tan DR, et al. Research of safety speeds model on curve road by using Trucksim[J]. Journal of Guangxi University (Natural Science Edition). 2017; 42(03): 961–969.
30. Hu CB, Lin M, Tong SH. STUDY ON FORMATION MECHANISM AND CHARACTERISTICS OF INERTIA LOAD OF TRUCK IN CURVED ROAD SECTIONS[J]. Engineering Mechanics. 2022; 18(02): 198–205.
31. Zhang XB. Study on formation mechanism and safety improvement of accident-prone ramp on “T” shape hub interchange[J]. Journal of Safety Science and Technology. 2022; 18(02): 198–205.
32. Li L, Lu Y, Wang R, Chen J. A three-dimensional dynamics control framework of vehicle lateral stability and rollover prevention via active braking with MPC[J]. IEEE Transactions on Industrial Electronics. 2016; 64(4): 3389–3401.
33. National Public service platform for standards information, China. GB 1589–2016, limits of dimensions, axle load and masses for motor vehicles, trailers and combination vehicle[S]. 2016; Available from: https://std.samr.gov.cn/gb/search/gbDetailed?id=71F772D8118AD3A7E05397BE0A0AB82A.
34. Ministry of Transport of the People’s Republic of China. Design Specification for Highway Alignment: JTG D20–2017[S]. Beijing: China Communications Press; 2017.
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
© 2024 Zhang et al. This is an open access article distributed under the terms of the Creative Commons Attribution License: http://creativecommons.org/licenses/by/4.0/ (the “License”), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Abstract
In this study, to investigate the factors affecting container semitrailer rollovers while driving on interchange loop ramps, a simulation model was established in TruckSim on the basis of data collected from an in-depth investigation of the vehicle, road, and driving behaviors in a real accident. The established model was validated by reproducing the accident in a simulation. The effects of vehicle speed, the height of the cargo’s center of gravity, tractor-semitrailer interactions, and the radius of the circular curve on container semitrailer rollover were investigated using the established model. The results reveal that if the height of the container semitrailer is within the limits in Chinese standards and the container semitrailer is moving at a speed of less than 45.3 km/h, it can safely travel through a loop ramp with a circular curve radius of 60 m. The height of the cargo’s center of gravity and the lateral load transfer ratio have a negative relationship, and a higher center of gravity reduces the safe speed for a container semitrailer traveling through a loop ramp. During rollover, the rear axle of the semitrailer first begins to roll over and then drives the tractor to overturn.
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
Details
; Yang, Xuefeng; Liu, Yishan; Wang, Li; Li, Pingfei; Li, Hui; Luo, Yong; Li, Yi; Liu, Qiaoyuan