- Email: [email protected]
Simulative investigation on head injuries of electric self-balancing scooter riders subject to ground impact
Accident Analysis and Prevention 89 (2016) 128–141
Contents lists available at ScienceDirect
Accident Analysis and Prevention journal homepage: www.elsevier.com/locate/aap
Simulative investigation on head injuries of electric self-balancing scooter riders subject to ground impact Jun Xu a,b,c , Shi Shang a,b , Hongsheng Qi d , Guizhen Yu e , Yunpeng Wang e , Peng Chen e,∗ a
Department of Automotive Engineering, School of Transportation Science and Engineering, Beihang University, Beijing 100191, PR China Advanced Vehicle Research Center, Beihang University, Beijing 100191, PR China Beijing Key Laboratory for High-efficient Power Transmission and System Control of New Energy Resource Vehicle, Beihang University, Beijing 100191, China d Institute of Transportation Engineering, Zhejiang University, Zhejiang 310058, PR China e Department of Transportation, School of Transportation Science and Engineering, Beihang University, Beijing 100191, PR China b c
a r t i c l e
i n f o
Article history: Received 2 November 2015 Received in revised form 24 January 2016 Accepted 25 January 2016 Keywords: Self-balancing scooter Traffic accidents Ground impact Brain injuries
a b s t r a c t The safety performance of an electric self-balancing scooter (ESS) has recently become a main concern in preventing its further wide application as a major candidate for green transportation. Scooter riders may suffer severe brain injuries in possible vehicle crash accidents not only from contact with a windshield or bonnet but also from secondary contact with the ground. In this paper, virtual vehicle–ESS crash scenarios combined with finite element (FE) car models and multi-body scooter/human models are set up. Postimpact kinematic gestures of scooter riders under various contact conditions, such as different vehicle impact speeds, ESS moving speeds, impact angles or positions, and different human sizes, are classified and analyzed. Furthermore, head–ground impact processes are reconstructed using validated FE head models, and important parameters of contusion and laceration (e.g., coup or contrecoup pressures and Von Mises stress and the maximum shear stress) are extracted and analyzed to assess the severity of regional contusion from head–ground contact. Results show that the brain injury risk increases with vehicle speeds and ESS moving speeds and may provide fundamental knowledge to popularize the use of a helmet and the vehicle-fitted safety systems, and lay a strong foundation for the reconstruction of ESS-involved accidents. There is scope to improve safety for the use of ESS in public roads according to the analysis and conclusions. © 2016 Elsevier Ltd. All rights reserved.
1. Introduction Electric self-balancing scooters (ESSs), as newly emerging pollution-free transportation tools, are gradually being popularized for short-distance traveling or the last-mile trip after traditional public transportations because of their convenient performance (Blackman and Haworth, 2013). However, the security concerns of the public about ESS are increasing simultaneously because of the many cases of accidents with serious injuries (Keith et al., 2011; Roider et al., 2015). Similar to pedestrians and bicyclists, ESS riders are generally regarded as vulnerable road users (VRUs) because riders may suffer from critical injuries during accidents (Lin and Tsai, 2009; Tsai et al., 2010). According to the World Health Organization statistics, VRUs accounted for approximately 50% of the total fatalities in 2013 (World Health Organization, 2013) worldwide. The
∗ Corresponding author. E-mail address: [email protected] (P. Chen). http://dx.doi.org/10.1016/j.aap.2016.01.013 0001-4575/© 2016 Elsevier Ltd. All rights reserved.
latest report also indicates that nearly 270,000 pedestrian deaths on the road occur every year (World Health Organization, 2015). Head injury is one of the research focuses because of its severity and lethality (Yang, 2011b), and it accounts up to 80% of all VRU fatalities in several districts (Edirisinghe et al., 2014; Hui et al., 2014). However, no investigations have been conducted on ESS safety in traffic accidents, although limited studies have focused on self-balancing and yaw control of the ESS (Lin and Tsai, 2009; Tsai et al., 2010). Recently (Xu et al., 2016), pioneered a study on the ESS safety situation in vehicle crash accidents by considering the head–vehicle contact. Compared with pedestrians, ESS riders are more likely to have head contact with higher regions of vehicles and the head–vehicle impact timing in vehicle–ESS crash accidents is tens of milliseconds later than that in vehicle–pedestrian crash accidents under the same impact conditions. This contact time difference may cause different injuries to riders. Studies on VRU head injuries are mainly focused on pedestrians and cyclists using several methods such as in-depth accident investigation (Otte et al., 2005; Yao et al., 2007; Deck and Willinger,
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
2008; Li and Yang, 2010) and numerical accident reconstruction (Xu et al., 2009; Li and Yang, 2010; Peng et al., 2012a; Nie and Yang, 2014; Peng et al., 2014). Research has shown that VRU head injuries are sensitive to various variables, such as vehicle impact speed (Anderson et al., 1997; Yao et al., 2008; Simms and Walsh, 2009; Elliott et al., 2012), vehicle type (Yang, 2003; Ballesteros et al., 2004; Lefler and Gabler, 2004; Han et al., 2012; Kerrigan et al., 2012; Crocetta et al., 2015), VRU moving speed (Crocetta et al., 2015), walking posture (Peng et al., 2012b), and impact location (Maki et al., 2003; Yang et al., 2005; Lin et al., 2007; Yao et al., 2008). The VRU head–vehicle collisions are always on the bonnet, windshield, and A-pillar areas during a crash. Then, human body falls off the car and the head contacts the ground. The coupling effects of head–vehicle contact and head–ground contact cause more difficulty in identifying the head injury mechanism. Preliminary studies have indicated that vehicle speed is the governing factor in the major head injury source (Simms and Wood, 2006; Yang et al., 2007). An effective way to reveal the main cause of the head injury is accident reconstruction based on in-depth vehicle–VRU accidents (Badea-Romero and Lenard, 2013). The finite element method (FEM) is regarded as a useful tool to analyze the injuries caused by head–vehicle contact from the biomechanics perspective (Yao et al., 2008; Peng et al., 2013). In terms of head injury caused by contact with the ground, parameterizations of simulation processes are used to analyze the ground impact effects on head injury (Gupta and Yang, 2013; Gupta, 2014; Crocetta et al., 2015). Although people have used ESS for short-distance transportation on city roads for quite a long time, relevant laws and regulations are still not published by the traffic management bureau. Consequently, people riding ESS on the roads without helmets or any other protection device may pose a potential safety risk to themselves. This study aims to evaluate the brain injuries of ESS riders during a secondary head collision, i.e., head–ground impact. First, numerical models of the traffic accident scenes are established based on the MADYMO (TASS, 2010) platform. In finite element (FE) reconstruction impact models of head–ground collisions are computed using LS-DAYNA (Hallquist, 1998). The Abbreviated Injury Scale (AIS) is used as the evaluating indicator of brain injury based on injury biomechanics and quantitative degrees of injury (AAAM, 1985). Comprehensive parametric studies involving two representative ESSs and five types of vehicles on head–ground injury are also conducted to fully investigate ESS rider safety.
2. Methods 2.1. Traffic accident scenarios The proposed research process for ESS rider safety is divided into several steps. Earlier research analyzing the head injuries caused by vehicle contact (Xu et al., 2016) has been conducted. By contrast, we mainly focus on ground contact in the present study. To evaluate the brain injury risks of single-wheel and double-wheel riders during head–ground impacts, MADYMO, the most commonly applied numerical simulation software to study crash safety during accidents in previous literature (Simms and Wood, 2006; Yao et al., 2008; Carter and Neal-Sturgess, 2009; Peng et al., 2012a; Nie and Yang, 2014), is used to model and simulate the entire impact processes of impact accidents. For example, a vehicle driving at a speed of 10 m/s brakes at a sustained rate by the time it hits the ESS rider on its side. The lateral impact is set as the baseline vehicle-ESS crash scenario because of its extremely high frequency and proportion (accounting for over 90%) in vehicle–VRU collisions (McLean et al., 1996; Yao et al., 2008; Yan et al., 2011). The intersection of facing directions between human and vehicle is /2 (shown in Fig. 1). The ESS rider hit by a vehicle with braking action falls to the ground
129
Fig. 1. Description of the baseline vehicle–ESS crash scenario.
earlier than that in cases without braking action. To avoid cases in which the human body flips over the vehicle roof and thus leading to a more complicated situation, cars are designated with a continuous brake with a 0.8 g deceleration, which indicates for the brake response of occupants, and a good friction contact is assumed between road and tire (Heinrichs et al., 2004). In addition, vehicle impact speeds, vehicle–ESS contact angles, vehicle–ESS contact positions, and human sizes are parameterized to investigate the human dynamic mechanisms and brain injuries under different impact conditions with constant deceleration. To meticulously examine the brain injuries of ESS riders, the processes of head–ground impact are reconstructed and simulated using FEM. This method is commonly accepted in impact safety research (Lei et al., 2009; Yang, 2011a; Han et al., 2012), such as in investigating the head response during head–windshield contact (Yao et al., 2008; Xu et al., 2010, 2011a,b; Peng et al., 2014) and simulating skull dropping tests (Shaoo et al., 2015).
2.2. Human model The 50th percentile male pedestrian model available in the MADYMO database (Automotive, 2001), which is a widely and the most used dummy model in the field of numerical accident reconstruction and analysis, is chosen as the ESS rider model in the baseline scenario because of its excellent performance in human body kinematics and injury analysis. The 95th percentile male pedestrian model and the 5th percentile female pedestrian model are used to represent ESS riders as the control group. These two human models with different anthropometries are also commonly applied in the VRU safety analysis (Crocetta et al., 2015). Fig. 2 shows the three human models used. More detailed information are presented in Ref. Automotive (2001).
2.3. ESS model The two most common types of ESS (a single-wheel ESS and a double-wheel ESS) are considered in this study. The numerical model of the single-wheel ESS is modeled by three ellipsoids to depict the external shape. Fig. 3(a) and Table 1 illustrate the rough profiles of the single-wheel model and the stiffness parameters, respectively. Another typical and widely used ESS is chosen as the representative of the double-wheel model. Six ellipsoids are modeled to describe the outer surface of the ESS body, including a controlling bar component and a couple of wheels. The multi-body model double-wheel ESS and its outside dimensions are presented in Fig. 3(b), with the stiffness setting shown in Table 2.
130
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
Table 1 Simplified force deflection data for the double-wheel ESS model. Wheel
Frame
Handlebar
Deflection (m)
Force (N)
Deflection (m)
Force (N)
Deflection (m)
0 0.0015 0.002
0 4000 9000
0 0.0012 0.0054 0.0103 0.0161 0.0293 0.0358 0.055
0 1500 2000 3000 4000 6500 6750 6950
0 0.04 0.07
Force (N) 0 5000 10,000
Table 2 Simplified force deflection data for the single-wheel ESS model. Wheel
Pedal
Board
Deflection (m)
Force (N)
Deflection (m)
Force (N)
Deflection (m)
0 0.0015 0.002
0 4000 9000
0 0.0012 0.0054 0.0103 0.0161 0.0293 0.0358 0.055
0 1500 2000 3000 4000 6500 6750 6950
0 0.04 0.07
Fig. 2. Illustrations of three different human models.
Force (N) 0 5000 10,000
are used to represent vehicles in the simulated traffic accidents because of their profiles. Specifically, a multi-purpose vehicle (MPV), a sport utility vehicle (SUV), a pick-up truck, a passenger sedan, and a small electric vehicle (EV) are chosen because these vehicle models are major vehicles with representative vehicle profiles running on the roads. The first four models were developed by the National Crash Analysis Center (NCAC) of the George Washington University under a contract with the FHWA and NHTSA of the US DOT. More detailed information on the FE car models can be found by accessing the NCAC website. As only the outer surfaces of the vehicles are needed to obtain the human kinematic processes during the ESS–vehicle contact, only the outermost layers of the FE car models are captured and converted to the MADYMO platform on which the vehicle–ESS crash accidents are simulated. The original and modified models are shown in Fig. 4. Combined contact is chosen as the type of vehicle–human contact, vehicle–ESS contact, and human–ESS contact. The human and ESS–ground contact and the vehicle tire–ground contact are set as the slave type. A friction coefficient of 0.3 is used for the human–vehicle contact, and this coefficient was validated and used by Simms and Wood (2006) and Crocetta et al. (2015) to achieve satisfactory simulation results. 2.5. FE head-ground impact model
Fig. 3. Multi-body models and the corresponding dimensions of a single-wheel ESS model and a double-wheel ESS model.
2.4. Vehicle model To investigate the effect of vehicle type on the risks of brain injury, five FE vehicle models with diverse shapes of the front-end
An FE human body head (HBM-head) model developed by Yang et al. (2007) and Yang et al. (2008) is used in the reconstructions of head–ground collision processes. The HBM-head model (shown in Fig. 5) consisting of scalp, skull, cerebrum, cerebellum, brain stem, and other vital components was constructed on the basis of the anthropometry of a 50th percentile male. This model has been validated and widely used in the field of skull and brain injury research (Yao et al., 2008; Huang and Yang, 2010; Yang, 2011a; Peng et al., 2012c; Chen et al., 2013). Detailed information on the HBM-head model, such as parameter settings and components, can be referred to in Ref. Yang et al. (2007). A linear elasticity shell model with a square size of 500 mm × 500 mm is constructed as the FE equivalent model of road. The material characteristics (density = 2400 kg/m3 , Poisson’s ratio = 0.2, Young’s modules E = 21,100 MPa) are set based
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
131
Fig. 4. Selected FE car models and the modified versions.
contrecoup pressure PCC , equivalent effective stress (Von Mises stress) VM , and maximum shear stress obtained from FE calculations have a strong relationship with the probability of AIS 3+ brain injury (Ward et al., 1980; Zhang et al., 2004; Yao et al., 2008), which indicates serious and irreversible injuries. By equating the injury indices values with AIS 3+ injuries, the critical values of PC , PCC , VM , and are set to 256, −152, 14.8, and 7.9 kPa, respectively. The thresholds are obtained from the study in Ref. Yao et al. (2008), and they are captured by using the method of regression analysis with real-world accident data (Kong and Yang, 2010). To better illustrate the method used to analyze the brain injury from ground, a flow process diagram of the design technique is presented in Fig. 6. 3. Results
Fig. 5. Sectional view of the HBM–head model.
on a type of asphalt concrete, C30 (Zhang et al., 2008b), which is widely used in major roads in China (Wang et al., 2011). To reconstruct the head–ground impact process, the relative position of the HBM-head and ground in three coordinate planes and the linear and angular velocities before collision on the MADYMO platform are captured as the input boundary condition similar to the methods adopted by Yang et al. (2008), Peng et al. (2014) and Shaoo et al. (2015). A surface-to-surface contact is set as the type of head–ground contact with a friction coefficient of 0.58, which is similar to previous studies (Simms and Wood, 2006; Crocetta et al., 2015). 2.6. Injury evaluation index Contusion and laceration of the brain, including coup and contrecoup injuries caused by the concentrated stress, are the most common brain injuries (Sosin et al., 1996) in clinical medicine. Important physical parameters such as coup pressure PC ,
By inputting the MADYMO, the post-impact kinematic and dynamic mechanisms of ESS riders are revealed given that the VRU kinematics dominates the head–ground impact configuration (Crocetta et al., 2015). 3.1. Categories of kinematics Simulations show that a large portion of ESS riders, similar to pedestrians, would wrap onto the vehicle or fall forward after contact with the vehicle. In some cases, bodies are hit and fly airborne in a counter-clockwise rotation. Table 3 illustrates the kinematic gestures categorized by quantifying the head rotation angle in the vertical plane. The rotation angle is closely connected with the head moving apex in the vertical direction that determines the landing speed of the head. Therefore, such a criterion is used for classification in consideration of the indirect effects of rotation angle on the head–ground impact speed. For example, Category 1 (C1) is a forward projection in which the body is pushed forward to the same direction as the vehicle movements; before the head–ground contact, the body rotates less than 90◦ in the clockwise direction, which is marked −90◦ in later sections. The similarity of kinematics from
Fig. 6. Flow process diagram of the design technique.
132
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
Table 3 Description of the identified kinematic categories. Kinemaic gestures
Classification
Description
Category 1
Before impacting on ground, the body rotates away from the vehicle front
Category 2
Before impacting on ground, the head rotates less than 90◦
Category 3
Before impacting on ground, the head rotates less than 180◦ .
Category 4
Before impacting on ground, the head rotates less than 270◦
Category 5
Before impacting on ground, the head rotates less than 360◦
Category 6
Before impacting on ground, the head rotates less than 450◦
Category 7
Before impacting on ground, the head rotates less than 540◦
Category 8
Before impacting on ground, the head rotates less than 630◦
Category 9
Before impacting on ground, the head rotates less than 720◦
Category 10
Before impacting on ground, the head rotates less than 810◦
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
133
front geometry. Statistical results are presented separately at each vehicle impact speed in Fig. 8(a)–(d). In cases with a vehicle impact speed of 10 m/s, all the kinematic processes of ESS riders are classified into C2 when hit by an SUV and a pick-up truck. A small number of cases fall into C3 and C4 in the MPV, EV, and sedan cases. With the increase in vehicle impact speed, ESS riders rotate more in the air before landing. In general, the risks of riders suffering rotations in the SUV and pick-up truck cases are lower than those in the cases of other vehicle types. The effect brought by the impact speed will be discussed later. 3.3. Relationship between kinematic categories and head–ground impact speed
Fig. 7. Breakdown of the kinematic categories.
C2 to C10 is the body wrapping onto the car in a counter-clockwise direction, and it is only distinguished by the head rotation angles with an interval of 90◦ . The proportion of each impact category is demonstrated in the pie chart (Fig. 7) within a parametric study of 5 types of vehicles, 2 kinds of ESSs, 4 different vehicle speeds (10, 15, 20, and 25 m/s,), and 4 different ESS moving speeds (1, 2, 3, and 4 m/s) in combination with 160 simulation groups.
In each kinematic category, head–ground impact speeds are different because of the variations in height of arch of ESS riders and the head–ground relative impact position. The relation between average head–ground impact speeds (with standard deviation) and kinematic categories is shown in Fig. 9. From the statistical point of view, the average speed of each kinematic category clearly varies. The motion postures involve 10 different kinematic categories, and thus various kinds of landing posture are inevitable. In some cases, when a rider falls to the ground, the head hits the ground directly. Conversely, the arms, legs, or abdomen parts hit the pavement first, followed by the head part, in some cases. The combined action of these random impact modes with other factors results in various average speeds for each kinematic category.
3.2. Relationship between kinematic categories and vehicle type 3.4. Brain variations of pressures and stresses Considerable previous research (Ballesteros et al., 2004; Han et al., 2012; Peng et al., 2012b; Gupta and Yang, 2013; Crocetta et al., 2015) has confirmed that the injury mechanisms of VRUs are significantly affected by vehicle type because of its complex
In the head–ground collision process, the cerebrum suffers sustained pressures and stresses. For example, a case of a double-wheel rider at 1 m/s moving speed hit by an SUV with an impact speed
Fig. 8. Occurrences of the 10 kinematic categories for each vehicle type when an ESS rider collides with each vehicle’s impact speed.
134
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
Fig. 9. Relation of kinematic categories with ESS riders’ head–ground impact speed at all vehicle impact speeds. Fig. 11. Relation of impact speed with ESS riders’ head–ground impact speed at all vehicle impact speeds.
of 10 m/s is provided for demonstration. Fig. 10(a)–(c) presents the pressure variation, the Von Mises stress variation, and the maximum shear stress variation in brain, respectively. Notably, a pressure or stress diffusion from the collision side (the lower part of brain in Fig. 10) to the other is observed. The values of the Von Mises stress and the maximum shear stress have a large gap, but extreme similarity is found among the stress distributions of these stress indices. 4. Discussion 4.1. Vehicle impact speed effect on brain injury Fig. 8 shows that the vehicle impact speed plays a significant role in ESS kinematic categories. Consequently, this factor has a non-trivial influence on head injury caused by the counterforce of
the ground. The relation of vehicle speed VC with the head–ground impact speed VH is shown in Fig. 11. The relationship presents a general trend with a positive correlation. With increasing VC , the probability of producing higher VH increases accordingly. The uneven dispersion of the VH phenomenon is mainly caused by various body and head landing postures. To investigate the brain injury in secondary impact, the values of PC , PCC , VM , and of the brain are extracted and analyzed. Only 2 out of 160 cases are under the safety margin according to the comparison of the safety threshold values defined in Section 2.6. Apart from the relationship between VC and VH , Fig. 12 also illustrates the relation between VC and the average values of injury evaluation indices along with average VH . The average values of PC , PCC , VM , and increase with increasing VC . Compared with the critical value, all the average values of the
Fig. 10. Pressure and stress variations in the brain in the case of a double-wheel rider with a 1 m/s moving speed hit by an SUV with an impact speed of 10 m/s.
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
135
Fig. 12. Relation of vehicle impact speed with ESS riders’ head SUV ground impact speed and ESS riders’ average values of brain injury indices.
Fig. 16. Description of different impact angles and different impact positions.
Fig. 13. Relation of vehicle BLE.H with ESS riders’ head SUV ground impact speed and ESS riders’ average values of brain injury indices at all vehicle impact speeds. Fig. 17. Relation of vehicle–ESS impact angle with ESS riders’ head–ground impact speed and ESS riders’ average values of brain injury indices at 15 m/s vehicle impact speed.
brain damage criteria are beyond the marks. Therefore, most ESS riders will suffer AIS 3+ brain injury from ground impact. 4.2. Vehicle type effect on brain injury
Fig. 14. Occurrences of the 10 kinematic categories for each ESS moving speed when the ESS rider collides at all vehicle impact speeds.
Fig. 15. Relation of ESS moving speed with ESS riders’ head–ground impact speed and ESS riders’ average values of brain injury indices at 15 m/s impact speed cases.
The severity of head injuries of VRUs can be affected by various factors, such as vehicle speed, vehicle type, VRU speed, impact angle and position, and human age. Thus, we examined this issue by coupling the dominating impact factors from Sections 4.2–4.3. These factors are regarded as governing factors, and they have been validated in previous references (Simms and Wood, 2006; Zhang et al., 2008a; Li and Yang, 2010; Elliott et al., 2012; Kerrigan et al., 2012; Crocetta et al., 2015). Then, for a comprehensive analysis of influence factors on this new vehicle–VRU impact crash, coupling effects such as vehicle speed and impact angle, vehicle speed and impact position, and vehicle speed and human size are also evaluated in Sections 4.4–4.6, respectively. Three different impact speeds (10, 15, and 20 m/s) are considered. Vehicle front-end geometry directly determines the location of the head–vehicle collision point, which strongly affects the kinematics of the human body and secondary head injury. In Fig. 8, human kinematics in the SUV and pick-up truck cases is mostly centralized in the C2 to C6 categories, and this finding indicates relatively few angles of rotation. For the purpose of quantitative analysis, the height of the vehicle bonnet’s leading edge is regarded as a representative of a vehicle type. The relation between vehicle BLE.H and VH /PC /PCC / VM / is demonstrated in Fig. 13. ESS riders impacted by vehicles with a relative low BLE.H, such as EVs and sedans, are observed to produce higher VH than those impacted by vehicles with a high BLE.H. In terms of injury indices, the values of PC , PCC , VM , and are obviously larger in the EV cases than in the
136
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
Fig. 18. Coupling effect of a vehicle impact speed and impact angle on ESS riders’ head–ground impact speed and ESS riders’ average values of brain injury indices.
Table 4 Average values of brain injury indices for each vehicle–ESS impact position at a 15 m/s vehicle impact speed. Vehicle–ESS impact position (rad/s)
Left (d = 0.667)
Middle (d = 0)
Right (d = −0.667)
Avg. head-ground speed (m/s) Avg. coup pressure (kPa) Avg. contrecoup pressure (kPa) Avg. Von Mises stress (kPa) Avg. maximum shear stress (kPa)
5.50669 7801.811 −4547.767 21.17438 12.19607
6.76873 8684.761 −4552.5307 25.62121 14.79544
4.43026 5115.1178 −2653.436 17.27902 9.09344
Table 5 Average values of brain injury indices for each human model at a 15 m/s vehicle impact speed. Human size
5th percentile female
50th percentile male
95th percentile male
Avg. head-ground speed (m/s) Avg. coup pressure (kPa) Avg. contrecoup pressure (kPa) Avg. Von Mises stress (kPa) Avg. maximum shear stress (kPa)
5.30087 6730.994 −3293.208 18.95203 11.23507
6.76873 8684.761 −4552.531 25.62121 14.79544
4.47468 8702.291 −4011.083 24.59029 14.18797
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
137
other types of vehicles. The highest average VH may be responsible for this result (based on the stress diagram of the sectional head model in Fig. 13, the Von Mises stress concentrations in the areas of the collision side and the offside are obvious). 4.3. ESS moving speed effect on brain injury
Fig. 19. Occurrences of the kinematic categories for each vehicle–ESS impact position at a 15 m/s vehicle impact speed.
The impact angle between VC and ESS speed (VE ) has a significant effect on the VRU kinematics after the first impact between the VRU and the vehicle. Thus, investigating the ESS speed effect on the brain injury of riders during the secondary impact is necessary. Four different ESS moving speeds, i.e., VE = 1, 2, 3, and 4 m/s, are set as the parameter variables. The relationship between VE and the human kinematic categories is presented in Fig. 14. More C2 than C3 cases are found at low ESS moving speeds (VE = 1 m/s or 2 m/s), and C3 cases appear more frequently than C2 when VE = 3 m/s or 4 m/s. In terms of brain injury, the relationship between ESS moving
Fig. 20. Coupling effect of vehicle impact speed and impact position on ESS riders’ head–ground impact speed and ESS riders’ average values of brain injury indices.
138
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
Fig. 21. Coupling effect of vehicle impact speed and human size on ESS riders’ head–ground impact speed and ESS riders’ average values of brain injury indices.
speed and the average values of brain injury indices is illustrated in Fig. 15. The average values of PC , PCC , VM , and in high ESS speed cases (VE = 3 and 4 m/s) are significantly larger than those in low ESS speed cases (VE = 1 and 2 m/s). According to the results presented above, if an ESS rider drives the transportation at a relatively high speed, then the ESS rider will have a high risk in suffering more serious brain injury from the ground impact in vehicle crash accidents. 4.4. Vehicle-to-ESS impact angle effect on brain injury More crash scenarios at 15 m/s impact speed with other initial conditions, such as the vehicle impacting an ESS rider at different angles ( = 0, /3, 2/3, and ), position (left, middle, and right part of the vehicle front-end), and human size (5th percentile female, 50th percentile male, and 95th percentile male) are
modeled and simulated to explore the potential effect on brain injuries during head–ground collision. Fig. 16(a) and (b) demonstrates the modeling of different impact angles and positions, respectively. As the impact angles have changed the relative positions between the VRU and the vehicle, the coming contact between the ESS rider and the vehicle generates various human kinematic gestures that significantly influence the severity of the brain injury of ESS riders caused by the head–ground collision. The relation of with VH /PC /PCC / VM / is illustrated in Fig. 17. When ESS riders drive backwards ( = 0) to the car, the pressures and stresses of the brain are clearly higher than those obtained from other cases. In cases in which ESS riders drive facing ( = ) the car, most of the injury indices are quite low. Specifically, in cases in which = , the handlebar and the control rod of a double-wheel ESS hit the vehicle first before the ESS rider–vehicle contact. This situation highly mitigates
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
the contact force during human–vehicle collision. Accordingly, the secondary head impact to the ground is alleviated. Fig. 18 illustrates the relations between brain injury indices and the coupling factors of and VC . When VC = 10 m/s, the general trend is that PC /PCC / VM / are the highest in the cases in which = 0, followed by the cases in which = /3, = , and = 2/3. This trend is consistent with the relationship between VH and . The reason is that the post-impact kinematics is similar in each case (most of the kinematic categories are C2) when ESS riders suffer a low-speed car impact. When the impact speed increases, the variations of injury indices are unpredictable because of the different landing postures caused by complex human airborne moves. 4.5. Vehicle-to-ESS impact position effect on head injury The first impact point locations of head–vehicle contacts are significantly determined by the vehicle–VRU impact position (Maki et al., 2003). The moving directions of ESS riders are perpendicular to the vehicle direction, and thus the impact position effects on human gestures are even larger. As shown in Fig. 16(b), virtual crash scenarios with three different impact positions, including first contact with the left (d = 0.667), middle (d = 0), and right (d = −0.667) components of the vehicle, are constructed. In this part, d represents the ratio of d0 (the actual horizontal distance in the first human–vehicle impingement point) to W (half of the vehicle frontend length). Fig. 19 shows the proportion of human kinematic categories for each impact position in 15 m/s cases. The risk of ESS riders to rotate in more angles (during the airborne flip) increases as the collision point is located from left to right of the car. The ESS has a moving speed of 1 m/s and the direction is from the left anterior to the right anterior of vehicle, and thus the head of the ESS rider does not hit the car in a few cases. Consequently, the ESS rider has more freedom space to rotate in more angles without a barrier, thus leading to various human kinematic gestures after the ESS–vehicle contact. However, the relationship between impact position and the average values of the brain injury indices listed in Table 4 shows that the damage in the case of d = −0.667 is lighter than that in other cases. The reason is that the face contacts the ground first (revealed in Fig. 17), and thus the brain areas suffering from sharp impact stress is avoided. The coupling effect of vehicle impact speed and impact position on the brain injuries of ESS riders is also studied. As observed in Fig. 20, an obvious positive relationship between VC and VH /PC /PCC / VM / is found only in cases in which d = 0. For the other two impact positions, when VC = 10 m/s, the injuries are higher than those in cases in which d = 0. However, the results are the opposite when VC = 20 m/s. Greater vehicle speed produces irregular human movements, head landing, postures, and impact areas, which cause significant differences in the brain injury indices. 4.6. Human size effect on head injury Anthropometry absolutely affects the brain injuries of VRU in crash accidents because the post-impact kinematic of human bodies performs with a significant difference if the heights of ESS riders are different (Crocetta et al., 2015). Three different human models in different sizes (95th percentile male, 50th percentile male, and 5th percentile female) are used to study the difference in the kinematic and head–ground injuries. Previous studies (Sahoo et al., 2015; Shaoo et al., 2015) have showed that the mass of an FE head strongly affects its dynamic properties under lateral impacts, whereas the differences caused by size are trivial. Therefore, based on the 50th percentile male, head models with scaled up or down densities are used to represent the heads of the 95th percentile male and the 5th percentile female during the processes of ground
139
contact similar to the strategy adopted in Ref. (Shaoo et al., 2015). Table 5 shows the brain injury values from the head–ground contact for each type of ESS rider at a vehicle speed of 15 m/s. The average values of PC , PCC , VM , and are close in the 50th percentile male case and the 95th percentile male case, whereas the damage indices of the 5th percentile female ESS rider are small, thus leading to a relatively lighter injury. This result is verified by the additional investigation on the coupled influence of vehicle speed and human size on brain injuries induced by the ground (see Fig. 21). The injury indices of the 50th percentile male cases and the 5th percentile female cases increase with vehicle speed. The stress values of the 95th percentile male brain decrease in the case of 20 m/s vehicle speed because of the relatively small strain of the head collision region.
5. Concluding remarks As the prevailing personal transportation tool for traveling at a relatively short distance, ESS has become increasingly popular and widely used in recent years. Evaluating the safety performance of ESS in potential vehicle crash accidents is necessary as head–ground impact may be responsible for permanent or severe brain injuries. In this study, several vehicle–ESS crash scenarios using different cars, impact speeds, impact angles, impact positions, ESS moving speeds, and human models are modeled and simulated to investigate the brain injury of ESS riders during secondary impact to the ground. However, this study is limited in that the bars and wheels of the ESS models cannot be rotated, and this limitation may result in a slight difference in human kinematics during the post-impact process. Brain injuries are studied and analyzed only in the head–ground impact process without considering the influence of the head–vehicle impact. The most significant difference between injuries caused by vehicle and those caused by the ground is that the head usually contacts with the vehicle directly in a stable kinetic form, but this form of head–ground collision is varied and unpredictable. In terms of the Head Injury Criterion (HIC), a widely accepted criterion for head injury in the automotive industry (Henn, 1998), the HIC levels registered for head impacts to the ground are often higher than any other type of head impacts. However, high HIC levels can also be observed in head–vehicle impacts due to higher relative speeds and the variable kinematics of the human body at impact. This unsolved and complicated comparison will be reported in a future study. First, human kinematic gestures are classified into 10 categories based on the head rotation angles. In general, greater vehicle impact speed produces more rotation angles of the human head. ESS riders are hit by a car with low BLE. Taller VRUs will suffer relatively more serious brain injury from ground impact. Second, head–ground impact speed has a positive correlation with vehicle impact speed, and the severity of brain injury increases as the head–ground impact speed increases. ESS moving speed also has a significant effect on brain injury. An ESS with great speed has high values of coup and contrecoup pressures, Von Mises stress, and maximum shear stress. Third, in case of a 15 m/s impact speed, the average values of brain injury indices are lower than those in other cases when an individual drives an ESS facing the car. The reason for this finding is due to the protection performance of the handlebar and the control rod of the double-wheel ESS. Fourth, ESS impacting on the left components of a vehicle leads to a kinematic gesture with a lesser angle of rotation and a more serious brain injury from ground impact than in cases in which an ESS collides on the right components of the vehicle. Finally, compared with the 50th percentile male and 95th percentile male, the 5th percentile female ESS riders suffer a relatively lighter brain injury. A possible reason for this finding is the females’ shorter height and lower weight than those of the males.
140
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
Corresponding experiments will be conducted and discussed in a future work.
Acknowledgements This work is financially supported by the Fundamental Research Funds for the Central Universities, Beihang University, the startup fund for “Zhuoyue 100” titled professor, Beihang University. We’d like to appreciate helps in providing FE head model from Prof. Jikuang Yang and providing Electric Self-balancing Scooters from Beijing HSP Technology Company Limited.
References AAAM, 1985. The abbreviated injury scale revision. Injuries. Anderson, R.W.G., McLean, A.J., Farmer, M.J.B., Lee, B.H., Brooks, C.G., 1997. Vehicle travel speeds and the incidence of fatal pedestrian crashes. Accid. Anal. Prev. 29 (5), 667–674. Automotive, T., 2001. Manual: Madymo Human Body Models, Delft, The Netherlands. Badea-Romero, A., Lenard, J., 2013. Source of head injury for pedestrians and pedal cyclists: striking vehicle or road? Accid. Anal. Prev. 50 (1), 1140–1150. Ballesteros, M.F., Dischinger, P.C., Langenberg, P., 2004. Pedestrian injuries and vehicle type in Maryland, 1995–1999. Accid. Anal. Prev. 36 (1), 73–81. Blackman, R.A., Haworth, N.L., 2013. Comparison of moped, scooter and motorcycle crash risk and crash severity. Accid. Anal. Prev. 57, 1–9. Carter, E., Neal-Sturgess, C.E., 2009. Madymo reconstruction of a real-world collision between a vehicle and cyclist. Int. J. Crashworthiness 14 (4), 379–390. Chen, J., Yang, J., Zhou, X., Zhang, C., Yu, X., Li, W., Wang, M., 2013. Fea of helmet–head injury protection based on motorcycle accident reconstruction. In: Proceedings of the 2013 Fifth International Conference on Measuring Technology and Mechatronics Automation, pp. 570–573. Crocetta, G., Piantini, S., Pierini, M., Simms, C., 2015. The influence of vehicle front-end design on pedestrian ground impact. Accid. Anal. Prev. 79, 56–69. Deck, C., Willinger, R., 2008. Improved head injury criteria based on head FE model. Int. J. Crashworthiness 13 (6), 667–678. Edirisinghe, P., Kitulwatte, I., Senarathne, U., 2014. Injuries in the vulnerable road user fatalities: a study from Sri Lanka. J. Forensic Legal Med. 27, 9–12. Elliott, J., Simms, C., Wood, D.P., 2012. Pedestrian head translation, rotation and impact velocity: the influence of vehicle speed, pedestrian speed and pedestrian gait. Accid. Anal. Prev. 45, 342–353. Gupta, V., (Dissertations and theses) 2014. Pedestrian Head Protection During Car to Pedestrian Accidents: In the Event of Primary Impact with Vehicle and Secondary Impact with Ground. Gradworks. Gupta, V., Yang, K.H., 2013. Effect of vehicle front end profiles leading to pedestrian secondary head impact to ground. Stapp Car Crash J. 57 (57), 139–155. Hallquist, J.O., 1998. LS-DYNA theoretical manual. In: Proceedings of the 2006 SEM Annual Conference and Exposition on Experimental and Applied Mechanics, St. Louis, MO, June 4–7, p. s33p03. Han, Y., Yang, J., Mizuno, K., Matsui, Y., 2012. Effects of vehicle impact velocity, vehicle front-end shapes on pedestrian injury risk. Traffic Injury Prev. 13 (5), 507–518. Heinrichs, B.E., Allin, B.D., Bowler, J.J., Siegmund, G.P., 2004. Vehicle speed affects both pre-skid braking kinematics and average tire/roadway friction. Accid. Anal. Prev. 36 (5), 829–840. Henn, H.W., 1998. Crash tests and the head injury criterion. Teach. Math. Appl. 17 (4), 162–170. Huang, S., Yang, J., 2010. Optimization of a reversible hood for protecting a pedestrian’s head during car collisions. Accid. Anal. Prev. 42 (4), 1136–1143. Hui, Z., Yin, Z., Yang, G., Che, X., Xie, J., Wei, H., Wang, Z., 2014. Analysis of 121 fatal passenger car–adult pedestrian accidents in China. J. Forensic Legal Med. 27, 76–81. Keith, B., Mary Pat, M., Raymond, L., Alison, S., Neal, S., 2011. Serious injuries related to the Segway® personal transporter: a case series. Ann. Emerg. Med. 57 (4), 370–374. Kerrigan, J., Arregui-Dalmases, C., Crandall, J., 2012. Assessment of pedestrian head impact dynamics in small sedan and large SUV collisions. Int. J. Crashworthiness 17 (3), 243–258. Kong, C., Yang, J., 2010. Logistic regression analysis of pedestrian casualty risk in passenger vehicle collisions in China. Accid. Anal. Prev. 42 (4), 987–993. Lefler, D.E., Gabler, H.C., 2004. The fatality and injury risk of light truck impacts with pedestrians in the United States. Accid. Anal. Prev. 36 (2), 295–304. Lei, G., Bao, C., Liu, X., 2009. Finite Element Analysis of Head-Ground Collision. Li, F., Yang, J.K., 2010. A study of head–brain injuries in car-to-pedestrian crashes with reconstructions using in-depth accident data in China. Int. J. Crashworthiness 15 (2), 117–124. Lin, S.C., Tsai, C.C., 2009. Development of a self-balancing human transportation vehicle for the teaching of feedback control. IEEE Trans. Educ. 52 (1), 157–168. Lin, Q.A., Xu, H.T.C., University, J., Changchun, 2007. Throw distance model for vehicle/pedestrian collisions. Automot. Eng.
Maki, T., Kajzer, J., Mizuno, K., Sekine, Y., 2003. Comparative analysis of vehicle–bicyclist and vehicle–pedestrian accidents in Japan. Accid. Anal. Prev. 35 (6), 927–940. McLean, A., Kloeden, C., Anderson, R.W., Baird, R., Farmer, M., 1996. Data collection and analysis of vehicle factors in relation to pedestrian brain injury. In: Proceedings of the Proceedings: International Technical Conference on the Enhanced Safety of Vehicles, pp. 1408–1411. Nie, J., Yang, J., 2014. A study of bicyclist kinematics and injuries based on reconstruction of passenger car–bicycle accident in China. Accid. Anal. Prev. 71, 50–59. Otte, D., Pape, C., Krettek, C., 2005. Kinematics and injury pattern in rollover accidents of cars in German road traffic – an in-depth-analysis by GIDAS. Int. J. Crashworthiness 10 (1), 75–86. Peng, Y., Chen, Y., Yang, J., Otte, D., Willinger, R., 2012a. A study of pedestrian and bicyclist exposure to head injury in passenger car collisions based on accident data and simulations. Safety Sci. 50 (9), 1749–1759. Peng, Y., Deck, C., Yang, J., Willinger, R., 2012b. Effects of pedestrian gait, vehicle-front geometry and impact velocity on kinematics of adult and child pedestrian head. Int. J. Crashworthiness 17 (5), 553–561. Peng, Y., Han, Y., Chen, Y., Yang, J., Willinger, R., 2012c. Assessment of the protective performance of hood using head FE model in car-to-pedestrian collisions. Int. J. Crashworthiness 17 (4), 415–423. Peng, Y., Yang, J., Deck, C., Willinger, R., 2013. Finite element modeling of crash test behavior for windshield laminated glass. Int. J. Impact Eng. 57 (7), 27–35. Peng, Y., Yang, J., Deck, C., 2014. Development of head injury risk functions based on real-world accident reconstruction. Int. J. Crashworthiness 19 (2), 105–114. Roider, D., Busch, C., Spitaler, R., Hertz, H., 2015. Segway® related injuries in Vienna: report from the Lorenz Böhler Trauma Centre. Eur. J. Trauma Emerg. Surg., 1–3. Sahoo, D., Deck, C., Yoganandan, N., Willinger, R., 2015. Influence of head mass on temporo-parietal skull impact using finite element modeling. Med. Biol. Eng. Comput. Shaoo, D., Deck, C., Yoganandan, N., Willinger, R., 2015. Influence of stiffness and shape of contact surface on skull fractures and biomechanical metrics of the human head of different population under lateral impacts. Accid. Anal. Prev., 97–105. Simms C.K.D.P.W., Walsh, D.G., 2009. The relationship between vehicle impact speed and pedestrian and cyclist projection distance. Solid Mech. Appl. Simms, C.K., Wood, D.P., 2006. Effects of pre-impact pedestrian position and motion on kinematics and injuries from vehicle and ground contact. Int. J. Crashworthiness 11 (4), 345–355. Sosin, D.M., Sniezek, J.E., Thurman, D.J., 1996. Incidence of mild and moderate brain injury in the United States, 1991. Brain Injury 10 (1), 47–54. TASS, B., 2010. Madymo Model Manual, Release. Tsai, C.-C., Huang, H.-C., Lin, S.-C., 2010. Adaptive neural network control of a self-balancing two-wheeled scooter. IEEE Trans. Ind. Electron. 57 (4), 868–873. Wang, Y.M., Liu, K., Yang, W.W., Sun, J.Z., Wang, Y.M., Liu, K., Yang, W.W., Sun, J.Z., 2011. Properties of flexible fiber and rigid fiber reinforced cement concrete used on surface layer of wharf pavement. Adv. Mater. Res. 224, 70–78. Ward, C.P., Chan, M., Nahum, A., 1980. Intracranial pressure – a brain injury criterion. Cadavers. World Health Organization, 2013. Global Status Report on Road Safety 2013: Supporting a Decade of Action. World Health Organization. World Health Organization, 2015. Make Walking Safe: A Brief Overview of Pedestrian Safety Around the World. World Health Organization. Xu, J., Li, Y., Lu, G., Zhou, W., 2009. Reconstruction model of vehicle impact speed in pedestrian–vehicle accident. Int. J. Impact Eng. 36 (6), 783–788. Xu, J., Li, Y., Chen, X., Yan, Y., Ge, D., Zhu, M., Liu, B., 2010. Characteristics of windshield cracking upon low-speed impact: numerical simulation based on the extended finite element method. Comput. Mater. Sci. 48 (3), 582–588. Xu, J., Li, Y., Liu, B., Zhu, M., Ge, D., 2011a. Experimental study on mechanical behavior of PVB laminated glass under quasi-static and dynamic loadings. Composites B 42 (2), 302–308. Xu, J., Li, Y.B., Chen, X., Ge, D.Y., Liu, B.H., Zhu, M.Y., Park, T.H., 2011b. Automotive windshield — pedestrian head impact: energy absorption capability of interlayer material. Int. J. Automot. Technol. 12 (5), 687–695. Xu, J., Shang, S., Yu, G., Qi, H., Wang, Y., Xu, S., 2016. Are electric self-balancing scooters safe in vehicle crash accidents? Accid. Anal. Prev., 102–116. Yan, X., Ma, M., Huang, H., Abdel-Aty, M., Wu, C., 2011. Motor vehicle–bicycle crashes in Beijing: irregular maneuvers, crash patterns, and injury severity. Accid. Anal. Prev. 43 (5), 1751–1758. Yang, J., 2011a. Investigation of brain trauma biomechanics in vehicle traffic accidents using human body computational models. Comput. Biomech. Med.: Soft Tissues Musculoskelet. Syst., 5–14. Yang, J., 2011b. Overview of research on injury biomechanics in car–pedestrian collisions. Chin. J. Automot. Eng. Yang, J., 2003. Effects of vehicle front design parameters on pedestrian head–brain injury protection. In: Proceedings of the 18th International Technical Conference of the Enhanced Safety of Vehicles. Yang, J., Yao, J., Otte, D., 2005. Correlation of different impact conditions to the injury severity of pedestrians in real world accidents. In: Proceedings of the 19th Int. Technical Conf. Enhanced Safety of Vehicle. Yang, J.K., Xu, W., Xiao, Z., Li, F., 2007. A study on head injury biomechanics in car-to-pedestrian collisions using human body models. J. Biomech. 40 (07), S87.
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141 Yang, J., Wei, X.U., Dietmar, O., 2008. Brain injury biomechanics in real world vehicle accident using mathematical models. Chin. J. Mech. Eng. 21 (04), 81–86. Yao, J., Yang, J., Otte, D., 2007. Head injuries in child pedestrian accidents – in-depth case analysis and reconstructions. Traffic Injury Prev. 8 (1), 94–100. Yao, J., Yang, J., Otte, D., 2008. Investigation of head injuries by reconstructions of real-world vehicle-versus-adult-pedestrian accidents. Safety Sci. 46 (7), 1103–1114.
141
Zhang, L., Yang, K.H., King, A.I., 2004. A proposed injury threshold for mild traumatic brain injury. J. Biomech. Eng. 126 (2), 226–236. Zhang, G., Cao, L., Hu, J., Yang, K.H., 2008a. A field data analysis of risk factors affecting the injury risks in vehicle-to-pedestrian crashes. Ann. Adv. Automot. Med. 52, 199–214. Zhang, J., Wang, Q., Hu, S., Wang, C., 2008b. Parameters verification of concrete damaged plastic model of ABAQUS. Build. Struct. 38 (8), 127–130.
Contents lists available at ScienceDirect
Accident Analysis and Prevention journal homepage: www.elsevier.com/locate/aap
Simulative investigation on head injuries of electric self-balancing scooter riders subject to ground impact Jun Xu a,b,c , Shi Shang a,b , Hongsheng Qi d , Guizhen Yu e , Yunpeng Wang e , Peng Chen e,∗ a
Department of Automotive Engineering, School of Transportation Science and Engineering, Beihang University, Beijing 100191, PR China Advanced Vehicle Research Center, Beihang University, Beijing 100191, PR China Beijing Key Laboratory for High-efficient Power Transmission and System Control of New Energy Resource Vehicle, Beihang University, Beijing 100191, China d Institute of Transportation Engineering, Zhejiang University, Zhejiang 310058, PR China e Department of Transportation, School of Transportation Science and Engineering, Beihang University, Beijing 100191, PR China b c
a r t i c l e
i n f o
Article history: Received 2 November 2015 Received in revised form 24 January 2016 Accepted 25 January 2016 Keywords: Self-balancing scooter Traffic accidents Ground impact Brain injuries
a b s t r a c t The safety performance of an electric self-balancing scooter (ESS) has recently become a main concern in preventing its further wide application as a major candidate for green transportation. Scooter riders may suffer severe brain injuries in possible vehicle crash accidents not only from contact with a windshield or bonnet but also from secondary contact with the ground. In this paper, virtual vehicle–ESS crash scenarios combined with finite element (FE) car models and multi-body scooter/human models are set up. Postimpact kinematic gestures of scooter riders under various contact conditions, such as different vehicle impact speeds, ESS moving speeds, impact angles or positions, and different human sizes, are classified and analyzed. Furthermore, head–ground impact processes are reconstructed using validated FE head models, and important parameters of contusion and laceration (e.g., coup or contrecoup pressures and Von Mises stress and the maximum shear stress) are extracted and analyzed to assess the severity of regional contusion from head–ground contact. Results show that the brain injury risk increases with vehicle speeds and ESS moving speeds and may provide fundamental knowledge to popularize the use of a helmet and the vehicle-fitted safety systems, and lay a strong foundation for the reconstruction of ESS-involved accidents. There is scope to improve safety for the use of ESS in public roads according to the analysis and conclusions. © 2016 Elsevier Ltd. All rights reserved.
1. Introduction Electric self-balancing scooters (ESSs), as newly emerging pollution-free transportation tools, are gradually being popularized for short-distance traveling or the last-mile trip after traditional public transportations because of their convenient performance (Blackman and Haworth, 2013). However, the security concerns of the public about ESS are increasing simultaneously because of the many cases of accidents with serious injuries (Keith et al., 2011; Roider et al., 2015). Similar to pedestrians and bicyclists, ESS riders are generally regarded as vulnerable road users (VRUs) because riders may suffer from critical injuries during accidents (Lin and Tsai, 2009; Tsai et al., 2010). According to the World Health Organization statistics, VRUs accounted for approximately 50% of the total fatalities in 2013 (World Health Organization, 2013) worldwide. The
∗ Corresponding author. E-mail address: [email protected] (P. Chen). http://dx.doi.org/10.1016/j.aap.2016.01.013 0001-4575/© 2016 Elsevier Ltd. All rights reserved.
latest report also indicates that nearly 270,000 pedestrian deaths on the road occur every year (World Health Organization, 2015). Head injury is one of the research focuses because of its severity and lethality (Yang, 2011b), and it accounts up to 80% of all VRU fatalities in several districts (Edirisinghe et al., 2014; Hui et al., 2014). However, no investigations have been conducted on ESS safety in traffic accidents, although limited studies have focused on self-balancing and yaw control of the ESS (Lin and Tsai, 2009; Tsai et al., 2010). Recently (Xu et al., 2016), pioneered a study on the ESS safety situation in vehicle crash accidents by considering the head–vehicle contact. Compared with pedestrians, ESS riders are more likely to have head contact with higher regions of vehicles and the head–vehicle impact timing in vehicle–ESS crash accidents is tens of milliseconds later than that in vehicle–pedestrian crash accidents under the same impact conditions. This contact time difference may cause different injuries to riders. Studies on VRU head injuries are mainly focused on pedestrians and cyclists using several methods such as in-depth accident investigation (Otte et al., 2005; Yao et al., 2007; Deck and Willinger,
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
2008; Li and Yang, 2010) and numerical accident reconstruction (Xu et al., 2009; Li and Yang, 2010; Peng et al., 2012a; Nie and Yang, 2014; Peng et al., 2014). Research has shown that VRU head injuries are sensitive to various variables, such as vehicle impact speed (Anderson et al., 1997; Yao et al., 2008; Simms and Walsh, 2009; Elliott et al., 2012), vehicle type (Yang, 2003; Ballesteros et al., 2004; Lefler and Gabler, 2004; Han et al., 2012; Kerrigan et al., 2012; Crocetta et al., 2015), VRU moving speed (Crocetta et al., 2015), walking posture (Peng et al., 2012b), and impact location (Maki et al., 2003; Yang et al., 2005; Lin et al., 2007; Yao et al., 2008). The VRU head–vehicle collisions are always on the bonnet, windshield, and A-pillar areas during a crash. Then, human body falls off the car and the head contacts the ground. The coupling effects of head–vehicle contact and head–ground contact cause more difficulty in identifying the head injury mechanism. Preliminary studies have indicated that vehicle speed is the governing factor in the major head injury source (Simms and Wood, 2006; Yang et al., 2007). An effective way to reveal the main cause of the head injury is accident reconstruction based on in-depth vehicle–VRU accidents (Badea-Romero and Lenard, 2013). The finite element method (FEM) is regarded as a useful tool to analyze the injuries caused by head–vehicle contact from the biomechanics perspective (Yao et al., 2008; Peng et al., 2013). In terms of head injury caused by contact with the ground, parameterizations of simulation processes are used to analyze the ground impact effects on head injury (Gupta and Yang, 2013; Gupta, 2014; Crocetta et al., 2015). Although people have used ESS for short-distance transportation on city roads for quite a long time, relevant laws and regulations are still not published by the traffic management bureau. Consequently, people riding ESS on the roads without helmets or any other protection device may pose a potential safety risk to themselves. This study aims to evaluate the brain injuries of ESS riders during a secondary head collision, i.e., head–ground impact. First, numerical models of the traffic accident scenes are established based on the MADYMO (TASS, 2010) platform. In finite element (FE) reconstruction impact models of head–ground collisions are computed using LS-DAYNA (Hallquist, 1998). The Abbreviated Injury Scale (AIS) is used as the evaluating indicator of brain injury based on injury biomechanics and quantitative degrees of injury (AAAM, 1985). Comprehensive parametric studies involving two representative ESSs and five types of vehicles on head–ground injury are also conducted to fully investigate ESS rider safety.
2. Methods 2.1. Traffic accident scenarios The proposed research process for ESS rider safety is divided into several steps. Earlier research analyzing the head injuries caused by vehicle contact (Xu et al., 2016) has been conducted. By contrast, we mainly focus on ground contact in the present study. To evaluate the brain injury risks of single-wheel and double-wheel riders during head–ground impacts, MADYMO, the most commonly applied numerical simulation software to study crash safety during accidents in previous literature (Simms and Wood, 2006; Yao et al., 2008; Carter and Neal-Sturgess, 2009; Peng et al., 2012a; Nie and Yang, 2014), is used to model and simulate the entire impact processes of impact accidents. For example, a vehicle driving at a speed of 10 m/s brakes at a sustained rate by the time it hits the ESS rider on its side. The lateral impact is set as the baseline vehicle-ESS crash scenario because of its extremely high frequency and proportion (accounting for over 90%) in vehicle–VRU collisions (McLean et al., 1996; Yao et al., 2008; Yan et al., 2011). The intersection of facing directions between human and vehicle is /2 (shown in Fig. 1). The ESS rider hit by a vehicle with braking action falls to the ground
129
Fig. 1. Description of the baseline vehicle–ESS crash scenario.
earlier than that in cases without braking action. To avoid cases in which the human body flips over the vehicle roof and thus leading to a more complicated situation, cars are designated with a continuous brake with a 0.8 g deceleration, which indicates for the brake response of occupants, and a good friction contact is assumed between road and tire (Heinrichs et al., 2004). In addition, vehicle impact speeds, vehicle–ESS contact angles, vehicle–ESS contact positions, and human sizes are parameterized to investigate the human dynamic mechanisms and brain injuries under different impact conditions with constant deceleration. To meticulously examine the brain injuries of ESS riders, the processes of head–ground impact are reconstructed and simulated using FEM. This method is commonly accepted in impact safety research (Lei et al., 2009; Yang, 2011a; Han et al., 2012), such as in investigating the head response during head–windshield contact (Yao et al., 2008; Xu et al., 2010, 2011a,b; Peng et al., 2014) and simulating skull dropping tests (Shaoo et al., 2015).
2.2. Human model The 50th percentile male pedestrian model available in the MADYMO database (Automotive, 2001), which is a widely and the most used dummy model in the field of numerical accident reconstruction and analysis, is chosen as the ESS rider model in the baseline scenario because of its excellent performance in human body kinematics and injury analysis. The 95th percentile male pedestrian model and the 5th percentile female pedestrian model are used to represent ESS riders as the control group. These two human models with different anthropometries are also commonly applied in the VRU safety analysis (Crocetta et al., 2015). Fig. 2 shows the three human models used. More detailed information are presented in Ref. Automotive (2001).
2.3. ESS model The two most common types of ESS (a single-wheel ESS and a double-wheel ESS) are considered in this study. The numerical model of the single-wheel ESS is modeled by three ellipsoids to depict the external shape. Fig. 3(a) and Table 1 illustrate the rough profiles of the single-wheel model and the stiffness parameters, respectively. Another typical and widely used ESS is chosen as the representative of the double-wheel model. Six ellipsoids are modeled to describe the outer surface of the ESS body, including a controlling bar component and a couple of wheels. The multi-body model double-wheel ESS and its outside dimensions are presented in Fig. 3(b), with the stiffness setting shown in Table 2.
130
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
Table 1 Simplified force deflection data for the double-wheel ESS model. Wheel
Frame
Handlebar
Deflection (m)
Force (N)
Deflection (m)
Force (N)
Deflection (m)
0 0.0015 0.002
0 4000 9000
0 0.0012 0.0054 0.0103 0.0161 0.0293 0.0358 0.055
0 1500 2000 3000 4000 6500 6750 6950
0 0.04 0.07
Force (N) 0 5000 10,000
Table 2 Simplified force deflection data for the single-wheel ESS model. Wheel
Pedal
Board
Deflection (m)
Force (N)
Deflection (m)
Force (N)
Deflection (m)
0 0.0015 0.002
0 4000 9000
0 0.0012 0.0054 0.0103 0.0161 0.0293 0.0358 0.055
0 1500 2000 3000 4000 6500 6750 6950
0 0.04 0.07
Fig. 2. Illustrations of three different human models.
Force (N) 0 5000 10,000
are used to represent vehicles in the simulated traffic accidents because of their profiles. Specifically, a multi-purpose vehicle (MPV), a sport utility vehicle (SUV), a pick-up truck, a passenger sedan, and a small electric vehicle (EV) are chosen because these vehicle models are major vehicles with representative vehicle profiles running on the roads. The first four models were developed by the National Crash Analysis Center (NCAC) of the George Washington University under a contract with the FHWA and NHTSA of the US DOT. More detailed information on the FE car models can be found by accessing the NCAC website. As only the outer surfaces of the vehicles are needed to obtain the human kinematic processes during the ESS–vehicle contact, only the outermost layers of the FE car models are captured and converted to the MADYMO platform on which the vehicle–ESS crash accidents are simulated. The original and modified models are shown in Fig. 4. Combined contact is chosen as the type of vehicle–human contact, vehicle–ESS contact, and human–ESS contact. The human and ESS–ground contact and the vehicle tire–ground contact are set as the slave type. A friction coefficient of 0.3 is used for the human–vehicle contact, and this coefficient was validated and used by Simms and Wood (2006) and Crocetta et al. (2015) to achieve satisfactory simulation results. 2.5. FE head-ground impact model
Fig. 3. Multi-body models and the corresponding dimensions of a single-wheel ESS model and a double-wheel ESS model.
2.4. Vehicle model To investigate the effect of vehicle type on the risks of brain injury, five FE vehicle models with diverse shapes of the front-end
An FE human body head (HBM-head) model developed by Yang et al. (2007) and Yang et al. (2008) is used in the reconstructions of head–ground collision processes. The HBM-head model (shown in Fig. 5) consisting of scalp, skull, cerebrum, cerebellum, brain stem, and other vital components was constructed on the basis of the anthropometry of a 50th percentile male. This model has been validated and widely used in the field of skull and brain injury research (Yao et al., 2008; Huang and Yang, 2010; Yang, 2011a; Peng et al., 2012c; Chen et al., 2013). Detailed information on the HBM-head model, such as parameter settings and components, can be referred to in Ref. Yang et al. (2007). A linear elasticity shell model with a square size of 500 mm × 500 mm is constructed as the FE equivalent model of road. The material characteristics (density = 2400 kg/m3 , Poisson’s ratio = 0.2, Young’s modules E = 21,100 MPa) are set based
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
131
Fig. 4. Selected FE car models and the modified versions.
contrecoup pressure PCC , equivalent effective stress (Von Mises stress) VM , and maximum shear stress obtained from FE calculations have a strong relationship with the probability of AIS 3+ brain injury (Ward et al., 1980; Zhang et al., 2004; Yao et al., 2008), which indicates serious and irreversible injuries. By equating the injury indices values with AIS 3+ injuries, the critical values of PC , PCC , VM , and are set to 256, −152, 14.8, and 7.9 kPa, respectively. The thresholds are obtained from the study in Ref. Yao et al. (2008), and they are captured by using the method of regression analysis with real-world accident data (Kong and Yang, 2010). To better illustrate the method used to analyze the brain injury from ground, a flow process diagram of the design technique is presented in Fig. 6. 3. Results
Fig. 5. Sectional view of the HBM–head model.
on a type of asphalt concrete, C30 (Zhang et al., 2008b), which is widely used in major roads in China (Wang et al., 2011). To reconstruct the head–ground impact process, the relative position of the HBM-head and ground in three coordinate planes and the linear and angular velocities before collision on the MADYMO platform are captured as the input boundary condition similar to the methods adopted by Yang et al. (2008), Peng et al. (2014) and Shaoo et al. (2015). A surface-to-surface contact is set as the type of head–ground contact with a friction coefficient of 0.58, which is similar to previous studies (Simms and Wood, 2006; Crocetta et al., 2015). 2.6. Injury evaluation index Contusion and laceration of the brain, including coup and contrecoup injuries caused by the concentrated stress, are the most common brain injuries (Sosin et al., 1996) in clinical medicine. Important physical parameters such as coup pressure PC ,
By inputting the MADYMO, the post-impact kinematic and dynamic mechanisms of ESS riders are revealed given that the VRU kinematics dominates the head–ground impact configuration (Crocetta et al., 2015). 3.1. Categories of kinematics Simulations show that a large portion of ESS riders, similar to pedestrians, would wrap onto the vehicle or fall forward after contact with the vehicle. In some cases, bodies are hit and fly airborne in a counter-clockwise rotation. Table 3 illustrates the kinematic gestures categorized by quantifying the head rotation angle in the vertical plane. The rotation angle is closely connected with the head moving apex in the vertical direction that determines the landing speed of the head. Therefore, such a criterion is used for classification in consideration of the indirect effects of rotation angle on the head–ground impact speed. For example, Category 1 (C1) is a forward projection in which the body is pushed forward to the same direction as the vehicle movements; before the head–ground contact, the body rotates less than 90◦ in the clockwise direction, which is marked −90◦ in later sections. The similarity of kinematics from
Fig. 6. Flow process diagram of the design technique.
132
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
Table 3 Description of the identified kinematic categories. Kinemaic gestures
Classification
Description
Category 1
Before impacting on ground, the body rotates away from the vehicle front
Category 2
Before impacting on ground, the head rotates less than 90◦
Category 3
Before impacting on ground, the head rotates less than 180◦ .
Category 4
Before impacting on ground, the head rotates less than 270◦
Category 5
Before impacting on ground, the head rotates less than 360◦
Category 6
Before impacting on ground, the head rotates less than 450◦
Category 7
Before impacting on ground, the head rotates less than 540◦
Category 8
Before impacting on ground, the head rotates less than 630◦
Category 9
Before impacting on ground, the head rotates less than 720◦
Category 10
Before impacting on ground, the head rotates less than 810◦
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
133
front geometry. Statistical results are presented separately at each vehicle impact speed in Fig. 8(a)–(d). In cases with a vehicle impact speed of 10 m/s, all the kinematic processes of ESS riders are classified into C2 when hit by an SUV and a pick-up truck. A small number of cases fall into C3 and C4 in the MPV, EV, and sedan cases. With the increase in vehicle impact speed, ESS riders rotate more in the air before landing. In general, the risks of riders suffering rotations in the SUV and pick-up truck cases are lower than those in the cases of other vehicle types. The effect brought by the impact speed will be discussed later. 3.3. Relationship between kinematic categories and head–ground impact speed
Fig. 7. Breakdown of the kinematic categories.
C2 to C10 is the body wrapping onto the car in a counter-clockwise direction, and it is only distinguished by the head rotation angles with an interval of 90◦ . The proportion of each impact category is demonstrated in the pie chart (Fig. 7) within a parametric study of 5 types of vehicles, 2 kinds of ESSs, 4 different vehicle speeds (10, 15, 20, and 25 m/s,), and 4 different ESS moving speeds (1, 2, 3, and 4 m/s) in combination with 160 simulation groups.
In each kinematic category, head–ground impact speeds are different because of the variations in height of arch of ESS riders and the head–ground relative impact position. The relation between average head–ground impact speeds (with standard deviation) and kinematic categories is shown in Fig. 9. From the statistical point of view, the average speed of each kinematic category clearly varies. The motion postures involve 10 different kinematic categories, and thus various kinds of landing posture are inevitable. In some cases, when a rider falls to the ground, the head hits the ground directly. Conversely, the arms, legs, or abdomen parts hit the pavement first, followed by the head part, in some cases. The combined action of these random impact modes with other factors results in various average speeds for each kinematic category.
3.2. Relationship between kinematic categories and vehicle type 3.4. Brain variations of pressures and stresses Considerable previous research (Ballesteros et al., 2004; Han et al., 2012; Peng et al., 2012b; Gupta and Yang, 2013; Crocetta et al., 2015) has confirmed that the injury mechanisms of VRUs are significantly affected by vehicle type because of its complex
In the head–ground collision process, the cerebrum suffers sustained pressures and stresses. For example, a case of a double-wheel rider at 1 m/s moving speed hit by an SUV with an impact speed
Fig. 8. Occurrences of the 10 kinematic categories for each vehicle type when an ESS rider collides with each vehicle’s impact speed.
134
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
Fig. 9. Relation of kinematic categories with ESS riders’ head–ground impact speed at all vehicle impact speeds. Fig. 11. Relation of impact speed with ESS riders’ head–ground impact speed at all vehicle impact speeds.
of 10 m/s is provided for demonstration. Fig. 10(a)–(c) presents the pressure variation, the Von Mises stress variation, and the maximum shear stress variation in brain, respectively. Notably, a pressure or stress diffusion from the collision side (the lower part of brain in Fig. 10) to the other is observed. The values of the Von Mises stress and the maximum shear stress have a large gap, but extreme similarity is found among the stress distributions of these stress indices. 4. Discussion 4.1. Vehicle impact speed effect on brain injury Fig. 8 shows that the vehicle impact speed plays a significant role in ESS kinematic categories. Consequently, this factor has a non-trivial influence on head injury caused by the counterforce of
the ground. The relation of vehicle speed VC with the head–ground impact speed VH is shown in Fig. 11. The relationship presents a general trend with a positive correlation. With increasing VC , the probability of producing higher VH increases accordingly. The uneven dispersion of the VH phenomenon is mainly caused by various body and head landing postures. To investigate the brain injury in secondary impact, the values of PC , PCC , VM , and of the brain are extracted and analyzed. Only 2 out of 160 cases are under the safety margin according to the comparison of the safety threshold values defined in Section 2.6. Apart from the relationship between VC and VH , Fig. 12 also illustrates the relation between VC and the average values of injury evaluation indices along with average VH . The average values of PC , PCC , VM , and increase with increasing VC . Compared with the critical value, all the average values of the
Fig. 10. Pressure and stress variations in the brain in the case of a double-wheel rider with a 1 m/s moving speed hit by an SUV with an impact speed of 10 m/s.
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
135
Fig. 12. Relation of vehicle impact speed with ESS riders’ head SUV ground impact speed and ESS riders’ average values of brain injury indices.
Fig. 16. Description of different impact angles and different impact positions.
Fig. 13. Relation of vehicle BLE.H with ESS riders’ head SUV ground impact speed and ESS riders’ average values of brain injury indices at all vehicle impact speeds. Fig. 17. Relation of vehicle–ESS impact angle with ESS riders’ head–ground impact speed and ESS riders’ average values of brain injury indices at 15 m/s vehicle impact speed.
brain damage criteria are beyond the marks. Therefore, most ESS riders will suffer AIS 3+ brain injury from ground impact. 4.2. Vehicle type effect on brain injury
Fig. 14. Occurrences of the 10 kinematic categories for each ESS moving speed when the ESS rider collides at all vehicle impact speeds.
Fig. 15. Relation of ESS moving speed with ESS riders’ head–ground impact speed and ESS riders’ average values of brain injury indices at 15 m/s impact speed cases.
The severity of head injuries of VRUs can be affected by various factors, such as vehicle speed, vehicle type, VRU speed, impact angle and position, and human age. Thus, we examined this issue by coupling the dominating impact factors from Sections 4.2–4.3. These factors are regarded as governing factors, and they have been validated in previous references (Simms and Wood, 2006; Zhang et al., 2008a; Li and Yang, 2010; Elliott et al., 2012; Kerrigan et al., 2012; Crocetta et al., 2015). Then, for a comprehensive analysis of influence factors on this new vehicle–VRU impact crash, coupling effects such as vehicle speed and impact angle, vehicle speed and impact position, and vehicle speed and human size are also evaluated in Sections 4.4–4.6, respectively. Three different impact speeds (10, 15, and 20 m/s) are considered. Vehicle front-end geometry directly determines the location of the head–vehicle collision point, which strongly affects the kinematics of the human body and secondary head injury. In Fig. 8, human kinematics in the SUV and pick-up truck cases is mostly centralized in the C2 to C6 categories, and this finding indicates relatively few angles of rotation. For the purpose of quantitative analysis, the height of the vehicle bonnet’s leading edge is regarded as a representative of a vehicle type. The relation between vehicle BLE.H and VH /PC /PCC / VM / is demonstrated in Fig. 13. ESS riders impacted by vehicles with a relative low BLE.H, such as EVs and sedans, are observed to produce higher VH than those impacted by vehicles with a high BLE.H. In terms of injury indices, the values of PC , PCC , VM , and are obviously larger in the EV cases than in the
136
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
Fig. 18. Coupling effect of a vehicle impact speed and impact angle on ESS riders’ head–ground impact speed and ESS riders’ average values of brain injury indices.
Table 4 Average values of brain injury indices for each vehicle–ESS impact position at a 15 m/s vehicle impact speed. Vehicle–ESS impact position (rad/s)
Left (d = 0.667)
Middle (d = 0)
Right (d = −0.667)
Avg. head-ground speed (m/s) Avg. coup pressure (kPa) Avg. contrecoup pressure (kPa) Avg. Von Mises stress (kPa) Avg. maximum shear stress (kPa)
5.50669 7801.811 −4547.767 21.17438 12.19607
6.76873 8684.761 −4552.5307 25.62121 14.79544
4.43026 5115.1178 −2653.436 17.27902 9.09344
Table 5 Average values of brain injury indices for each human model at a 15 m/s vehicle impact speed. Human size
5th percentile female
50th percentile male
95th percentile male
Avg. head-ground speed (m/s) Avg. coup pressure (kPa) Avg. contrecoup pressure (kPa) Avg. Von Mises stress (kPa) Avg. maximum shear stress (kPa)
5.30087 6730.994 −3293.208 18.95203 11.23507
6.76873 8684.761 −4552.531 25.62121 14.79544
4.47468 8702.291 −4011.083 24.59029 14.18797
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
137
other types of vehicles. The highest average VH may be responsible for this result (based on the stress diagram of the sectional head model in Fig. 13, the Von Mises stress concentrations in the areas of the collision side and the offside are obvious). 4.3. ESS moving speed effect on brain injury
Fig. 19. Occurrences of the kinematic categories for each vehicle–ESS impact position at a 15 m/s vehicle impact speed.
The impact angle between VC and ESS speed (VE ) has a significant effect on the VRU kinematics after the first impact between the VRU and the vehicle. Thus, investigating the ESS speed effect on the brain injury of riders during the secondary impact is necessary. Four different ESS moving speeds, i.e., VE = 1, 2, 3, and 4 m/s, are set as the parameter variables. The relationship between VE and the human kinematic categories is presented in Fig. 14. More C2 than C3 cases are found at low ESS moving speeds (VE = 1 m/s or 2 m/s), and C3 cases appear more frequently than C2 when VE = 3 m/s or 4 m/s. In terms of brain injury, the relationship between ESS moving
Fig. 20. Coupling effect of vehicle impact speed and impact position on ESS riders’ head–ground impact speed and ESS riders’ average values of brain injury indices.
138
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
Fig. 21. Coupling effect of vehicle impact speed and human size on ESS riders’ head–ground impact speed and ESS riders’ average values of brain injury indices.
speed and the average values of brain injury indices is illustrated in Fig. 15. The average values of PC , PCC , VM , and in high ESS speed cases (VE = 3 and 4 m/s) are significantly larger than those in low ESS speed cases (VE = 1 and 2 m/s). According to the results presented above, if an ESS rider drives the transportation at a relatively high speed, then the ESS rider will have a high risk in suffering more serious brain injury from the ground impact in vehicle crash accidents. 4.4. Vehicle-to-ESS impact angle effect on brain injury More crash scenarios at 15 m/s impact speed with other initial conditions, such as the vehicle impacting an ESS rider at different angles ( = 0, /3, 2/3, and ), position (left, middle, and right part of the vehicle front-end), and human size (5th percentile female, 50th percentile male, and 95th percentile male) are
modeled and simulated to explore the potential effect on brain injuries during head–ground collision. Fig. 16(a) and (b) demonstrates the modeling of different impact angles and positions, respectively. As the impact angles have changed the relative positions between the VRU and the vehicle, the coming contact between the ESS rider and the vehicle generates various human kinematic gestures that significantly influence the severity of the brain injury of ESS riders caused by the head–ground collision. The relation of with VH /PC /PCC / VM / is illustrated in Fig. 17. When ESS riders drive backwards ( = 0) to the car, the pressures and stresses of the brain are clearly higher than those obtained from other cases. In cases in which ESS riders drive facing ( = ) the car, most of the injury indices are quite low. Specifically, in cases in which = , the handlebar and the control rod of a double-wheel ESS hit the vehicle first before the ESS rider–vehicle contact. This situation highly mitigates
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
the contact force during human–vehicle collision. Accordingly, the secondary head impact to the ground is alleviated. Fig. 18 illustrates the relations between brain injury indices and the coupling factors of and VC . When VC = 10 m/s, the general trend is that PC /PCC / VM / are the highest in the cases in which = 0, followed by the cases in which = /3, = , and = 2/3. This trend is consistent with the relationship between VH and . The reason is that the post-impact kinematics is similar in each case (most of the kinematic categories are C2) when ESS riders suffer a low-speed car impact. When the impact speed increases, the variations of injury indices are unpredictable because of the different landing postures caused by complex human airborne moves. 4.5. Vehicle-to-ESS impact position effect on head injury The first impact point locations of head–vehicle contacts are significantly determined by the vehicle–VRU impact position (Maki et al., 2003). The moving directions of ESS riders are perpendicular to the vehicle direction, and thus the impact position effects on human gestures are even larger. As shown in Fig. 16(b), virtual crash scenarios with three different impact positions, including first contact with the left (d = 0.667), middle (d = 0), and right (d = −0.667) components of the vehicle, are constructed. In this part, d represents the ratio of d0 (the actual horizontal distance in the first human–vehicle impingement point) to W (half of the vehicle frontend length). Fig. 19 shows the proportion of human kinematic categories for each impact position in 15 m/s cases. The risk of ESS riders to rotate in more angles (during the airborne flip) increases as the collision point is located from left to right of the car. The ESS has a moving speed of 1 m/s and the direction is from the left anterior to the right anterior of vehicle, and thus the head of the ESS rider does not hit the car in a few cases. Consequently, the ESS rider has more freedom space to rotate in more angles without a barrier, thus leading to various human kinematic gestures after the ESS–vehicle contact. However, the relationship between impact position and the average values of the brain injury indices listed in Table 4 shows that the damage in the case of d = −0.667 is lighter than that in other cases. The reason is that the face contacts the ground first (revealed in Fig. 17), and thus the brain areas suffering from sharp impact stress is avoided. The coupling effect of vehicle impact speed and impact position on the brain injuries of ESS riders is also studied. As observed in Fig. 20, an obvious positive relationship between VC and VH /PC /PCC / VM / is found only in cases in which d = 0. For the other two impact positions, when VC = 10 m/s, the injuries are higher than those in cases in which d = 0. However, the results are the opposite when VC = 20 m/s. Greater vehicle speed produces irregular human movements, head landing, postures, and impact areas, which cause significant differences in the brain injury indices. 4.6. Human size effect on head injury Anthropometry absolutely affects the brain injuries of VRU in crash accidents because the post-impact kinematic of human bodies performs with a significant difference if the heights of ESS riders are different (Crocetta et al., 2015). Three different human models in different sizes (95th percentile male, 50th percentile male, and 5th percentile female) are used to study the difference in the kinematic and head–ground injuries. Previous studies (Sahoo et al., 2015; Shaoo et al., 2015) have showed that the mass of an FE head strongly affects its dynamic properties under lateral impacts, whereas the differences caused by size are trivial. Therefore, based on the 50th percentile male, head models with scaled up or down densities are used to represent the heads of the 95th percentile male and the 5th percentile female during the processes of ground
139
contact similar to the strategy adopted in Ref. (Shaoo et al., 2015). Table 5 shows the brain injury values from the head–ground contact for each type of ESS rider at a vehicle speed of 15 m/s. The average values of PC , PCC , VM , and are close in the 50th percentile male case and the 95th percentile male case, whereas the damage indices of the 5th percentile female ESS rider are small, thus leading to a relatively lighter injury. This result is verified by the additional investigation on the coupled influence of vehicle speed and human size on brain injuries induced by the ground (see Fig. 21). The injury indices of the 50th percentile male cases and the 5th percentile female cases increase with vehicle speed. The stress values of the 95th percentile male brain decrease in the case of 20 m/s vehicle speed because of the relatively small strain of the head collision region.
5. Concluding remarks As the prevailing personal transportation tool for traveling at a relatively short distance, ESS has become increasingly popular and widely used in recent years. Evaluating the safety performance of ESS in potential vehicle crash accidents is necessary as head–ground impact may be responsible for permanent or severe brain injuries. In this study, several vehicle–ESS crash scenarios using different cars, impact speeds, impact angles, impact positions, ESS moving speeds, and human models are modeled and simulated to investigate the brain injury of ESS riders during secondary impact to the ground. However, this study is limited in that the bars and wheels of the ESS models cannot be rotated, and this limitation may result in a slight difference in human kinematics during the post-impact process. Brain injuries are studied and analyzed only in the head–ground impact process without considering the influence of the head–vehicle impact. The most significant difference between injuries caused by vehicle and those caused by the ground is that the head usually contacts with the vehicle directly in a stable kinetic form, but this form of head–ground collision is varied and unpredictable. In terms of the Head Injury Criterion (HIC), a widely accepted criterion for head injury in the automotive industry (Henn, 1998), the HIC levels registered for head impacts to the ground are often higher than any other type of head impacts. However, high HIC levels can also be observed in head–vehicle impacts due to higher relative speeds and the variable kinematics of the human body at impact. This unsolved and complicated comparison will be reported in a future study. First, human kinematic gestures are classified into 10 categories based on the head rotation angles. In general, greater vehicle impact speed produces more rotation angles of the human head. ESS riders are hit by a car with low BLE. Taller VRUs will suffer relatively more serious brain injury from ground impact. Second, head–ground impact speed has a positive correlation with vehicle impact speed, and the severity of brain injury increases as the head–ground impact speed increases. ESS moving speed also has a significant effect on brain injury. An ESS with great speed has high values of coup and contrecoup pressures, Von Mises stress, and maximum shear stress. Third, in case of a 15 m/s impact speed, the average values of brain injury indices are lower than those in other cases when an individual drives an ESS facing the car. The reason for this finding is due to the protection performance of the handlebar and the control rod of the double-wheel ESS. Fourth, ESS impacting on the left components of a vehicle leads to a kinematic gesture with a lesser angle of rotation and a more serious brain injury from ground impact than in cases in which an ESS collides on the right components of the vehicle. Finally, compared with the 50th percentile male and 95th percentile male, the 5th percentile female ESS riders suffer a relatively lighter brain injury. A possible reason for this finding is the females’ shorter height and lower weight than those of the males.
140
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141
Corresponding experiments will be conducted and discussed in a future work.
Acknowledgements This work is financially supported by the Fundamental Research Funds for the Central Universities, Beihang University, the startup fund for “Zhuoyue 100” titled professor, Beihang University. We’d like to appreciate helps in providing FE head model from Prof. Jikuang Yang and providing Electric Self-balancing Scooters from Beijing HSP Technology Company Limited.
References AAAM, 1985. The abbreviated injury scale revision. Injuries. Anderson, R.W.G., McLean, A.J., Farmer, M.J.B., Lee, B.H., Brooks, C.G., 1997. Vehicle travel speeds and the incidence of fatal pedestrian crashes. Accid. Anal. Prev. 29 (5), 667–674. Automotive, T., 2001. Manual: Madymo Human Body Models, Delft, The Netherlands. Badea-Romero, A., Lenard, J., 2013. Source of head injury for pedestrians and pedal cyclists: striking vehicle or road? Accid. Anal. Prev. 50 (1), 1140–1150. Ballesteros, M.F., Dischinger, P.C., Langenberg, P., 2004. Pedestrian injuries and vehicle type in Maryland, 1995–1999. Accid. Anal. Prev. 36 (1), 73–81. Blackman, R.A., Haworth, N.L., 2013. Comparison of moped, scooter and motorcycle crash risk and crash severity. Accid. Anal. Prev. 57, 1–9. Carter, E., Neal-Sturgess, C.E., 2009. Madymo reconstruction of a real-world collision between a vehicle and cyclist. Int. J. Crashworthiness 14 (4), 379–390. Chen, J., Yang, J., Zhou, X., Zhang, C., Yu, X., Li, W., Wang, M., 2013. Fea of helmet–head injury protection based on motorcycle accident reconstruction. In: Proceedings of the 2013 Fifth International Conference on Measuring Technology and Mechatronics Automation, pp. 570–573. Crocetta, G., Piantini, S., Pierini, M., Simms, C., 2015. The influence of vehicle front-end design on pedestrian ground impact. Accid. Anal. Prev. 79, 56–69. Deck, C., Willinger, R., 2008. Improved head injury criteria based on head FE model. Int. J. Crashworthiness 13 (6), 667–678. Edirisinghe, P., Kitulwatte, I., Senarathne, U., 2014. Injuries in the vulnerable road user fatalities: a study from Sri Lanka. J. Forensic Legal Med. 27, 9–12. Elliott, J., Simms, C., Wood, D.P., 2012. Pedestrian head translation, rotation and impact velocity: the influence of vehicle speed, pedestrian speed and pedestrian gait. Accid. Anal. Prev. 45, 342–353. Gupta, V., (Dissertations and theses) 2014. Pedestrian Head Protection During Car to Pedestrian Accidents: In the Event of Primary Impact with Vehicle and Secondary Impact with Ground. Gradworks. Gupta, V., Yang, K.H., 2013. Effect of vehicle front end profiles leading to pedestrian secondary head impact to ground. Stapp Car Crash J. 57 (57), 139–155. Hallquist, J.O., 1998. LS-DYNA theoretical manual. In: Proceedings of the 2006 SEM Annual Conference and Exposition on Experimental and Applied Mechanics, St. Louis, MO, June 4–7, p. s33p03. Han, Y., Yang, J., Mizuno, K., Matsui, Y., 2012. Effects of vehicle impact velocity, vehicle front-end shapes on pedestrian injury risk. Traffic Injury Prev. 13 (5), 507–518. Heinrichs, B.E., Allin, B.D., Bowler, J.J., Siegmund, G.P., 2004. Vehicle speed affects both pre-skid braking kinematics and average tire/roadway friction. Accid. Anal. Prev. 36 (5), 829–840. Henn, H.W., 1998. Crash tests and the head injury criterion. Teach. Math. Appl. 17 (4), 162–170. Huang, S., Yang, J., 2010. Optimization of a reversible hood for protecting a pedestrian’s head during car collisions. Accid. Anal. Prev. 42 (4), 1136–1143. Hui, Z., Yin, Z., Yang, G., Che, X., Xie, J., Wei, H., Wang, Z., 2014. Analysis of 121 fatal passenger car–adult pedestrian accidents in China. J. Forensic Legal Med. 27, 76–81. Keith, B., Mary Pat, M., Raymond, L., Alison, S., Neal, S., 2011. Serious injuries related to the Segway® personal transporter: a case series. Ann. Emerg. Med. 57 (4), 370–374. Kerrigan, J., Arregui-Dalmases, C., Crandall, J., 2012. Assessment of pedestrian head impact dynamics in small sedan and large SUV collisions. Int. J. Crashworthiness 17 (3), 243–258. Kong, C., Yang, J., 2010. Logistic regression analysis of pedestrian casualty risk in passenger vehicle collisions in China. Accid. Anal. Prev. 42 (4), 987–993. Lefler, D.E., Gabler, H.C., 2004. The fatality and injury risk of light truck impacts with pedestrians in the United States. Accid. Anal. Prev. 36 (2), 295–304. Lei, G., Bao, C., Liu, X., 2009. Finite Element Analysis of Head-Ground Collision. Li, F., Yang, J.K., 2010. A study of head–brain injuries in car-to-pedestrian crashes with reconstructions using in-depth accident data in China. Int. J. Crashworthiness 15 (2), 117–124. Lin, S.C., Tsai, C.C., 2009. Development of a self-balancing human transportation vehicle for the teaching of feedback control. IEEE Trans. Educ. 52 (1), 157–168. Lin, Q.A., Xu, H.T.C., University, J., Changchun, 2007. Throw distance model for vehicle/pedestrian collisions. Automot. Eng.
Maki, T., Kajzer, J., Mizuno, K., Sekine, Y., 2003. Comparative analysis of vehicle–bicyclist and vehicle–pedestrian accidents in Japan. Accid. Anal. Prev. 35 (6), 927–940. McLean, A., Kloeden, C., Anderson, R.W., Baird, R., Farmer, M., 1996. Data collection and analysis of vehicle factors in relation to pedestrian brain injury. In: Proceedings of the Proceedings: International Technical Conference on the Enhanced Safety of Vehicles, pp. 1408–1411. Nie, J., Yang, J., 2014. A study of bicyclist kinematics and injuries based on reconstruction of passenger car–bicycle accident in China. Accid. Anal. Prev. 71, 50–59. Otte, D., Pape, C., Krettek, C., 2005. Kinematics and injury pattern in rollover accidents of cars in German road traffic – an in-depth-analysis by GIDAS. Int. J. Crashworthiness 10 (1), 75–86. Peng, Y., Chen, Y., Yang, J., Otte, D., Willinger, R., 2012a. A study of pedestrian and bicyclist exposure to head injury in passenger car collisions based on accident data and simulations. Safety Sci. 50 (9), 1749–1759. Peng, Y., Deck, C., Yang, J., Willinger, R., 2012b. Effects of pedestrian gait, vehicle-front geometry and impact velocity on kinematics of adult and child pedestrian head. Int. J. Crashworthiness 17 (5), 553–561. Peng, Y., Han, Y., Chen, Y., Yang, J., Willinger, R., 2012c. Assessment of the protective performance of hood using head FE model in car-to-pedestrian collisions. Int. J. Crashworthiness 17 (4), 415–423. Peng, Y., Yang, J., Deck, C., Willinger, R., 2013. Finite element modeling of crash test behavior for windshield laminated glass. Int. J. Impact Eng. 57 (7), 27–35. Peng, Y., Yang, J., Deck, C., 2014. Development of head injury risk functions based on real-world accident reconstruction. Int. J. Crashworthiness 19 (2), 105–114. Roider, D., Busch, C., Spitaler, R., Hertz, H., 2015. Segway® related injuries in Vienna: report from the Lorenz Böhler Trauma Centre. Eur. J. Trauma Emerg. Surg., 1–3. Sahoo, D., Deck, C., Yoganandan, N., Willinger, R., 2015. Influence of head mass on temporo-parietal skull impact using finite element modeling. Med. Biol. Eng. Comput. Shaoo, D., Deck, C., Yoganandan, N., Willinger, R., 2015. Influence of stiffness and shape of contact surface on skull fractures and biomechanical metrics of the human head of different population under lateral impacts. Accid. Anal. Prev., 97–105. Simms C.K.D.P.W., Walsh, D.G., 2009. The relationship between vehicle impact speed and pedestrian and cyclist projection distance. Solid Mech. Appl. Simms, C.K., Wood, D.P., 2006. Effects of pre-impact pedestrian position and motion on kinematics and injuries from vehicle and ground contact. Int. J. Crashworthiness 11 (4), 345–355. Sosin, D.M., Sniezek, J.E., Thurman, D.J., 1996. Incidence of mild and moderate brain injury in the United States, 1991. Brain Injury 10 (1), 47–54. TASS, B., 2010. Madymo Model Manual, Release. Tsai, C.-C., Huang, H.-C., Lin, S.-C., 2010. Adaptive neural network control of a self-balancing two-wheeled scooter. IEEE Trans. Ind. Electron. 57 (4), 868–873. Wang, Y.M., Liu, K., Yang, W.W., Sun, J.Z., Wang, Y.M., Liu, K., Yang, W.W., Sun, J.Z., 2011. Properties of flexible fiber and rigid fiber reinforced cement concrete used on surface layer of wharf pavement. Adv. Mater. Res. 224, 70–78. Ward, C.P., Chan, M., Nahum, A., 1980. Intracranial pressure – a brain injury criterion. Cadavers. World Health Organization, 2013. Global Status Report on Road Safety 2013: Supporting a Decade of Action. World Health Organization. World Health Organization, 2015. Make Walking Safe: A Brief Overview of Pedestrian Safety Around the World. World Health Organization. Xu, J., Li, Y., Lu, G., Zhou, W., 2009. Reconstruction model of vehicle impact speed in pedestrian–vehicle accident. Int. J. Impact Eng. 36 (6), 783–788. Xu, J., Li, Y., Chen, X., Yan, Y., Ge, D., Zhu, M., Liu, B., 2010. Characteristics of windshield cracking upon low-speed impact: numerical simulation based on the extended finite element method. Comput. Mater. Sci. 48 (3), 582–588. Xu, J., Li, Y., Liu, B., Zhu, M., Ge, D., 2011a. Experimental study on mechanical behavior of PVB laminated glass under quasi-static and dynamic loadings. Composites B 42 (2), 302–308. Xu, J., Li, Y.B., Chen, X., Ge, D.Y., Liu, B.H., Zhu, M.Y., Park, T.H., 2011b. Automotive windshield — pedestrian head impact: energy absorption capability of interlayer material. Int. J. Automot. Technol. 12 (5), 687–695. Xu, J., Shang, S., Yu, G., Qi, H., Wang, Y., Xu, S., 2016. Are electric self-balancing scooters safe in vehicle crash accidents? Accid. Anal. Prev., 102–116. Yan, X., Ma, M., Huang, H., Abdel-Aty, M., Wu, C., 2011. Motor vehicle–bicycle crashes in Beijing: irregular maneuvers, crash patterns, and injury severity. Accid. Anal. Prev. 43 (5), 1751–1758. Yang, J., 2011a. Investigation of brain trauma biomechanics in vehicle traffic accidents using human body computational models. Comput. Biomech. Med.: Soft Tissues Musculoskelet. Syst., 5–14. Yang, J., 2011b. Overview of research on injury biomechanics in car–pedestrian collisions. Chin. J. Automot. Eng. Yang, J., 2003. Effects of vehicle front design parameters on pedestrian head–brain injury protection. In: Proceedings of the 18th International Technical Conference of the Enhanced Safety of Vehicles. Yang, J., Yao, J., Otte, D., 2005. Correlation of different impact conditions to the injury severity of pedestrians in real world accidents. In: Proceedings of the 19th Int. Technical Conf. Enhanced Safety of Vehicle. Yang, J.K., Xu, W., Xiao, Z., Li, F., 2007. A study on head injury biomechanics in car-to-pedestrian collisions using human body models. J. Biomech. 40 (07), S87.
J. Xu et al. / Accident Analysis and Prevention 89 (2016) 128–141 Yang, J., Wei, X.U., Dietmar, O., 2008. Brain injury biomechanics in real world vehicle accident using mathematical models. Chin. J. Mech. Eng. 21 (04), 81–86. Yao, J., Yang, J., Otte, D., 2007. Head injuries in child pedestrian accidents – in-depth case analysis and reconstructions. Traffic Injury Prev. 8 (1), 94–100. Yao, J., Yang, J., Otte, D., 2008. Investigation of head injuries by reconstructions of real-world vehicle-versus-adult-pedestrian accidents. Safety Sci. 46 (7), 1103–1114.
141
Zhang, L., Yang, K.H., King, A.I., 2004. A proposed injury threshold for mild traumatic brain injury. J. Biomech. Eng. 126 (2), 226–236. Zhang, G., Cao, L., Hu, J., Yang, K.H., 2008a. A field data analysis of risk factors affecting the injury risks in vehicle-to-pedestrian crashes. Ann. Adv. Automot. Med. 52, 199–214. Zhang, J., Wang, Q., Hu, S., Wang, C., 2008b. Parameters verification of concrete damaged plastic model of ABAQUS. Build. Struct. 38 (8), 127–130.