Chapter 10 Analysis of Human Body Dynamics in Simulated Rear-End Impacts and whiplash phenomena
10.1Analysis of Human Body Dynamics in Simulated Rear-End Impacts
Recently, numerical simulation of human body motion has been receiving an increasingly wider range of application. Computer simulations have greatly contributed to the understanding of the kinematics and dynamics of human body motion. Such computer simulations are being used in ergonomics to better understand and thus improve the interaction of workers with machines and with workplace configurations. In sports, the computer simulations are being used to attempt to optimize performance and to develop equipment for injury prevention, and to speed and enhance recovery from injury. But of all the applications of human body motion simulation, the most significant is the simulation of accident victim dynamics--particularly in motor vehicle accidents.
In recent years, there has been considerable discussion about the influence of seatback and seat belts on occupant dynamics in rear-end impacts. In 1987, Collision Safety Engineering presented a report that indicates that seat belts play an important role in rear-end impacts. The report also showed that the belts decrease the incidence of any injury by 12 percent and reduce the incidence of serious injury by over 57 percent [1]. Warner et al. [2] also described how the controversy over seats can be classified into two types: One suggests that rigid seats are better because they provide a better constraint for occupant. The other argument maintains that a yielding seat is better than rigid one because it can absorb energy to protect the occupant. In 1992, Viano [3] indicated that when the seatback angle increases rapidly the probability of injury also increase. Most studies of the factors associated with injury in rear-end impacts indicate that the following factors significantly affect the incidence of neck injury: (1) head displacement, rotation and acceleration; (2) the relative velocity between head and torso; (3) the occupant ramping up the seatback; and (4) the rebounding phenomenon.
In 1993, Svensson [4] indicated that the angular displacement between the head and the chest is generally larger for rearseats than frontseats. McConnell [5] showed that although the traditional whiplash neck response to rear-end impacts and the widely accepted hyperextension or hyperflection cervical injury mechanism have been extensively written about, the reason why whiplash occurs is still unknown. Compared to other experiments, this research, which used real people as subjects, found that the subjects have light neck injury after relatively low velocity rear-end impact.
There have been a number of noteworthy efforts to obtain numerical simulation of vehicle accident victim dynamics. Among these are the works of McHerny [6], McHenry and Naab [7], Young [8], Robbins, Bennett, and Bowman [9], Huston, Hessel, and Passerello [10], Karnes, Tocher, and Twigg [11], Fleck [12], Huston, Hessel, and Winget [13], Wismans, Hoen, and Wittebrood [14], Prasad [15], Wismans and Hermans [16] and Huang [17]. However, these works do not primarily discuss the rear-end impact environment. Here, an attempt is made to use computer simulation to investigate the dynamic response of the occupant during rear-end impacts. The procedure uses the finite segment human body model seated in an accelerating or decelerating vehicle. The model consists of a finite-segment representation of the occupant, together with the restraint system (seat and belts). The occupant model consists of 15 rigid bodies representing the torso and limbs of the human frame. Nonlinear springs and dampers are used at the connection joints to represent human anatomical characteristics and limits imposed by muscles, ligaments, and soft tissue. Equations of motion are written for this model using Kane’s equation [18]and the multibody dynamics analysis procedures developed by Huston[19].
10.1.1 The Biodynamic Model
The model developed in this chapter is to analyze rear-end impact situations. Seatback, and headrest constraints are applied to the model to simulate the rear-end impact dynamic environment. This model consists of three submodels. They are the human body model, the contact model and the restraint belts model.
(1) Human Body Model
The human body model consists of 15 rigid bodies representing the torso, the head, the neck, the feet and the upper and lower arms and legs. It is a three-dimensioal model with ball-and-socket joints except for the elbows and knees which are modeled as hinge joints (Table 10.2.1). The ball-and-socket joint is described by three relative orientation variables, and the hinge joint by one rotational variable. These joints are provided with damping restraints to limit joint angle rotations. The motion of each joint is simulated by a spring and damper to simulate the behavior of the soft tissue (muscle and tendon). A representation of this model is given in Figure 10.1.1.
(2) Contact Model
The human body’s interaction with the seat is, modeled using a series planes that represent headrest and seatback (Figure 10.1.2), is sensed by body segments. The reaction forces on the bodysegments are generated as a function of the penetration of the segments into the contact planes. In the model, the generated load of this contact can be expressed by a contraint force and an equivalent moment acting on the mass center of the contacting body segment. The contact model has two basic routines: They are damping forces and spring forces.
a) damping forces
The relative velocity vector between the contacting segment and the contact point can be resolved into tangential component and normal component as shown inFigure 10.1.3. The damping force is defined as
(10.1.1)
where c is the damping coefficient.
When the body crosses an intrusion plane that is in loading condition (increasing penetration ), damping force is employed in the direction opposite to the velocity of the body relative to the plane. In the case of unloading condition (decreasing penetration) no damping force is acted to the contacting body.
b) spring forces
Spring routine using the force deflection characteristics computes the spring forceas a function of the penetration displacement and the stiffness characteristics and apply the corresponding spring force and equivalent moment to the contacting segment. Spring force is divided into two parts, normal constraint force ( normal to intrusionsurface ) and friction force (parallel to intrusion surface in the direction opposite tothe relative velocity component ) as shown in Figure 10.1.4. The equivalent torque isgenerated by the cross product of the distance (from mass center of the contacting segment to contact point) and the spring force. They can be expressed as:
(10.1.2)
(10.1.3)
whererepresents the normal unit vector of the intrusion plane, is the unit vector of velocity component which is parallel to the intrusion plane, and μis the coefficient ofcoulomb friction. Fn is the constraint force normal to the intrusion plane which incorporates a nonlinear spring force and hysteretic effect. Fn can be modeeled as:
(i) Loading ( Xp > 0 and )
(10.1.4)
(ii) Unloading ( Xp > 0 and )
(10.1.5)
and
H = 1 - E (10.1.6)
where Xp is the penetration displacement of the intrusion surface from collision point. H is the ratio of unloading Fn to loading Fn in linear region. E is the energy dissipation index. S is the separation point for the linear and nonlinear stiffness. K0, K1 and K2 are the stiffness parameters. Here the contact stiffnesses K0 and K1 , listed in Table 10.1.2, are based on the data provided in reference [20]. K2can be expressed as:
(10.1.7)
where Smax is the maximum penetration.
(3) Restraint Belts Model
The restraint belts (Figure 10.1.5) are modeled as springs that can be attached between specified points in the vehicle frame and the bodies of the human model. Each belt length from the point in the body to the point in the vehicle frame is calculated separately. In the current version belts can be used to simulate lap and shoulder belts.
The human body model of Figure 10.1.1 is a multibody system. To simulate an accident victim, the model is regarded as being seated relative to a vehicle frame, which in turn is regarded as moving relative to an inertial frame. As such it may be studied using Kane' s equations [18] and Huston’s method [19] for the analysis as shown in above chapter.
10.1.2 Verification and Examples
In the first stage of this study, a three-dimensional multiody system simulator for rear-end crash environment was developed. There is little experimental data available to date which can be used to check or verify the mathematical model and others like it. However, an attempt at verification was made for some data gathered by Viano [3]: The rear-end impact tests were conducted on the General Motors Research Laboratories Hyge Sled using a range of impact severities from 4.2-9.6 m/s change in velocity simulating rear-end impact. An automotive bucket seat modified for an initial angle of 9, 22 or 35 from the vertical. In all cases a loosely attached lap-belt was used to prevent the human model from entirely leaving the seat structure should rideup of the seatback occur. The chest and neck extension angles were measured. These angles then were calculated using the model with approximate, seatback structure and restraint belt configurations. A comparison of the results is shown in Figures 10.1.1-10.1.6. From the comparisons we can see that both the peak values and the general shapes of the simulation curves agree with the experimental data.
In the examples we examine the response of occupant dynamics during a rear-end crash environment. The constraint planes used in these examples were the headrest, seatback. Two different examples were considered:
(1) the effects of velocity upon head and chest accelerations were computed
The input data were modeled by using an initial velocity equal to 4.2, 6.4 and 9.6 m/s
with initial seatback set at angle 22 from vertical.
(2) the effects of seatback angle upon the head rotation relative to chest were computed (case1: with headrest; case2: without headrest)
The input data were modeled by using an initial seatback angle of 9, 22 and 35 from
vertical with initial velocity equaled 4.2 m/s.
In the first example, the results are shown in Figure 10.1.7 and 10.1.8. We can see that with fixed seatback angle, the accelerations of head and chest increase with the increase in velocity. This increased acceleration significantly aggravates a dynamic force of head and chest. In the second example, for case 1 (with headrest), the results are shown in Figure 10.1.9. From the figure, we can see that with fixed velocity, the relative rotations of head to chest increase with the decreasing of the angle of the seatback. In this case the maximum relative rotations of head to chest in 35, 22 and 9 are separately 16,30.2 and 39.8. For case 2 (without headrest ), in Figure 10.1.10, the maximum relative rotations of head to chest are 35, 22 and 9 are separately 51.9,51.6 and 40.6.
10.1.3 Conclusion
The results of this analysis are consistent with those develpoed experimentally, numerically, and empirically. They also show that the velocity and the seatback angle are two important factors that affect human injuries during vehicle rear-end impact. Next, the results clearly illustrate a whiplash effect of seat angle and headrest (Figure 10.1.9 and 10.1.10). When such impacts occur, lap belts efficiently prevent the occupant sliding upward along the seatback. Although the whiplash injury is not currently well understood now, it is nevertheless major cause of motor vehicle occupant injuries.
The results also show that, it is possible to build a reliable multibody system simulation model of rear-end impact crash vehicle occupants and the advantages of using multibody dynamics procedures in studying the kinematics and dynamics of vehicle rear-end impact crash victims. Nevertheless, the unrealistic conditions of anthropomorphic structure must be considered before these simulations can be projected into real life situtations. It appears that future work will include refinements of the model. Application of the model in the design of vehicle interior and safety devices is clear but is yet to be developed.
Table 10.1.1 Type of Joint Used for the Model
Joint / Type of JointB2 – B3 / Ball and Socket
B3 – B4 / Ball and Socket
B4 – B5 / Ball and Socket
B5 – B6 / Hinge
B4 – B7 / Ball and Socket
B7 – B8 / Hinge
B4 – B9 / Ball and Socket
B9 – B10 / Ball and Socket
B2 – B11 / Ball and Socket
B11 – B12 / Hinge
B12 – B13 / Hinge
B2 – B14 / Ball and Socket
B14 – B15 / Hinge
B15 – B16 / Hinge
Table 10.1.2 Contact Parameters Between Body Segments and Planes
Plane / K0 (N/M) / K1 (N/M2) / S (M) / E / μSeat Pan / 48570 / 32246220 / 0.02 / 0.56 / 0.65
Seat Back / 48570 / 32246220 / 0.06 / 0.56 / 0.65
B2:Lower Torso
B3:Back B10
B4:Chest
B5:Left Upper Arm
B6:Left Lower Arm B9
B7:Right Upper Arm B6
B8:Right Lower Arm B4 B8
B9:Neck B5
B10:Head B7
B11:Left Upper Leg
B12:Left Lower Leg
B13:Left Foot B3
B14:Right Upper Leg
B15:Right Lower Leg B2 B11
B16:Right Foot
z B14 B12
B15
B13
B1 (Vehicle Frame) x
Z B16
B0 (Inertia Frame) X
Figure 10.1.1 Human Body Model
Headrest B10
B9
B8 (B6)
(B5)
B4 B7
Seatback
B3
B2 B14
B15 (B12)
B16 (B13)
Z z
x
B1 (Vehicle Frame)
B0 (Inertia Frame) X
Figure 10.1.2 The Configuration of the Seatback and Headrest Model
p
Figure 10.1.3 The relative velocity resolved into two components
(p:contact point,g:mass center)
P μ
g
Figure 10.1.4 The spring force resolved into two components
(p:contact point, g:mass center)
B10
B9
ith Belt B8 (B6)
(B5)
B4 B7
qi
pi
B3
B2 B14
z
B15 (B12)
B1 (Vehicle Frame) x B16 (B13)
Z
B0 (Inertia Frame) X
Figure 10.1.5 The Configuration of the Restraint Belt
Figure 10.1.1 Comparison of simulative and experimental rotation-time histories for the chest (seatback angle--9,10g)
Figure 10.1.2 Comparison of simulative and experimental rotation-time histories for the chest (seatback angle--22,10g)
Figure 10.1.3 Comparison of simulative and experimental rotation-time histories for the chest (seatback angle--35,10g)
Figure 10.1.4 Comparison of simulative and experimental rotation-time histories for the neck (seatback angle--9,10g)
Figure 10.1.5 Comparison of simulative and experimental rotation-time
histories for the neck (seatback angle--22,10g)
Figure 10.1.6 Comparison of simulative and experimental rotation-time
histories for the neck (seatback angle--35,10g)
Figure 10.1.7 Head acceleration for increasing velocity with fix seatback angle (22)
Figure 10.1.8 Chest acceleration for increasing velocity with fix seatback angle (22)
Figure 10.1.9 Head rotation relative to chest for increasing seatback angle with fix velocity(4.2m/s,with headrest)
Figure 10.1.10 Head rotation relative to chest for increasing seatback angle with fix velocity(4.2m/s,without headrest)
10.2 Influence of the head restraint position on dynamic response of the human head-neck system under whiplash loading
Whiplash injuries result in clinically complex symptoms. The mechanism of this damage to human head/neck system is clinically important [21-26]. This type of injuries account for more than half of automobile accident insurance claims in Japan and Canada [27,28]. In the U.S., rear impacts account for more than 25% of all automobile-related injuries. Some of those injuries are caused to the head and neck by violent forward vehicle acceleration. The symptoms associated with whiplash have been described, our understanding of the injury mechanism remains insufficient. Understanding the mechanism of injury is important.
Support given by head restraints is an obvious aid towards prevention of this type of injury, but for it to be effective, a front passenger would effectively always need to travel with the back of the head against the restraint. Pro-tech gets over this problem by allowing the headrest to be in any position the seat occupant chooses for regular use, but in the event of a rear collision it moves forwards and upwards almost instantly, to “catch” the head and neck. In a rear accident it is important to minimize the distance between a person’s head and the headrest.
For the objective of this study, we will define a whiplash as a relative rotation motion between the head and torso and the head dynamical response that occurs in rear-end automobile impacts. Many studies have documented the head response to accelerations simulating rear-end impact using volunteers [29-31]. However, volunteers are tested by limit velocity changes to avoid injury, therefore data derived from these tests have limited application. Consequently, computer model of head/neck system is very useful to simulate head/neck system dynamical response during rear-end impact.
Many mechanical systems such as robots, machine tools, chain, human body model can be effectively models as systems of rigid and flexible bodies [32-40]. In this paper these systems and models are called multibody systems. During recent years there have been many attempts to develop efficient methods for obtaining equations of motion for multibody systems [41-44]. In the derivation of these equations, some researchers use Lagrange's equations, some use Newton's laws, and some use other modified theories. Each of these approaches, has the objective of efficient derivation of computer-oriented equations. The relative advantages (or disadvantages) of these approaches depend on the dynamical method which is used, and the method of organizing the complex geometry. In this paper a method of obtaining equations of motion is presented which avoids each of these difficulties. The methods are based upon Kane's equations [45-48]. The methods combine Kane's equations with geometric and accounting procedures developed by Huston [49-50].
10.2.1 Configuration of The Human Body Model
The occupant model developed in this paper can be used to analyze rear-end collision environments. The seatback, and headrest constraints were built in the model to simulate rear-end collision dynamic environment. There are two submodels in the model. They are the human body model and the contact model.A representation of this model is given in Figure 10.1.2.