Head impact simulation

INTRODUCTION:

Pedestrian safety simulation is a part of vehicle safety engineering. This study focuses on car and pedestrian interaction. The main focus of this is to reduce the pedestrian injury in case of an accident (head injury, leg injury etc). Usually, these tests are performed in the lower velocity range from 20-40KmpH. HIC(Head injury criteria/coefficient) is a measure of the likelihood of the head injury caused due to the impact after an accident. This is basically a measure of average acceleration during a small interval of the crash.

OBJECTIVE:

To perform head impact simulation and calculate the HIC(Head injury criteria/coefficient) value in the LS-PrePost.

In this study, the simple head impact on the rigid wall is studied at first. Then the impact of given headform model with the rigid body is studied further followed by headform impact on the hood. Important results such as stress, strain, system energy, HIC values are derived in post-processing.

PROCEDURE:

The whole analysis is divided into three cases and each case setup is discussed below.

1) Case 1- Simple head impact on the rigid wall:

A simple head geometry is given with the problem statement. This geometry is meshed with only shell elements. To replicate the impact with a rigid wall, a planar rigid wall is created from create entity>Rigidwall>planar utility. The created rigid wall is normal to the Z-axis and placed 10mm away from the simple head geometry. The planar rigid wall is generated by picking two nodes which are along the Z-axis. Figure 1.1 shows the simple head geometry and the planar rigid wall at the bottom.

Kg/mm/ms unit system is used throughout the analysis.

Fig 1.1: Simple head geometry with a planar rigid wall at the bottom

The section properties for the simple head are given by using *SECTION_SHELL card. All the properties are set to default and a thickness of 1.5mm is assigned to the shell element. Simple elastic material property is assigned to it by using *MAT_ELASTIC card. Figure 1.2 shows the material card assignment.

Fig 1.2: Elastic material properties assigned to simple head geometry

The head geometry is assumed to have 40KmpH velocity along the travelling path which is 45degree along Z-axis. Being circular in shape, rotating and translating operation will not make any changes in impact simulation. So, to assume the impact of simple head and rigid wall at 45degree, a velocity of 7.8567 mm/ms is given to the geometry along negative X and negative Z direction. The resultant of these velocity vectors replicate 40KmpH velocity which is 45dgree about Z-axis. Automatic single surface contact is defined for this simulation, only the slave segment definition is required for this contact type. The simple head geometry is assigned as the slave segment and the master segment column is kept empty.

The HIC value is extracted through NODOUT card in *DATABASE_ASCII_options. But for extracting a node statistics, the desired node should be selected. Here the bottom-most node is selected for extracting results. Figure 1.3 shows the *DATABASE_HISTORY_NODE card used to select the bottom-most node. Binary D3Plot is activated in the database section to get the final impact animation.

The termination time for the problem is defined by *CONTROL_TERMINATION card. The distance between the head geometry and the rigid wall is calculated and the approximate time taken to hit the rigid wall and some additional time after the impact is calculated. This cumulative time (3ms) is defined as termination time in the card.

Fig 1.3: Node selected by history node card for extracting HIC values

2) Case 2- Headform impact on the rigid wall:

Another geometry resembling a child’s head is given with the problem statement. In this study, it is considered to impact of head and rigid wall at an angle of 45degree. But the given geometry is not perfectly circular and hence the transformation of the original geometry is necessary to replicate the real impact situation. *DEFINE_TRANSFORMATION card is used to rotate the original body about Y-axis. *INCLUDE_TRANSFORM card is necessary to complete the rotation procedure. In this card, the original geometry is called through filename and defined transformation is carried out on that geometry. Transformation ID(TRANID) is set to 1 for this card. Figure 1.4 and 1.5 shows the transformation cards used to rotate the original head geometry. Figure 1.6 shows the final transformation. A separate keyword file is created with these keywords. The original input file is not changed but only called for transformation. Further required keywords are added in this new input file. Both the keyword files the original input file and the new (transformed) input file are kept in the same folder.

Fig 1.4: *DEFINE_TRANSFORMATION card assignment

Fig 1.5: *INCLUDE_TRANSFORM card assignment

a) Original geometry b) Transformed geometry

Fig 1.6: Transformation of original geometry by rotation

Once the head geometry is transformed in a desired orientation rigid wall is generated normal to Z-axis as discussed in the above section. This head geometry is made up of solid elements. So section properties of a solid element are assigned to this geometry. Basically the whole geometry is made up of four different parts skull, skin, accel block and backplate and all these subparts are modelled with solid elements. Only skin which is outermost part is modelled with an Ogden rubber, hyper-elastic material model and rest of the parts are modelled as rigid bodies. Automatic single surface contact is defined with the complete head as a slave segment. Automatic surface to surface contact is pre-defined between skull and skin.

The rigid wall is generated normal to Z-axis as discussed in the previous section. The case set-up is similar to the previous study. The initial velocity is defined along X and Z direction and a node is selected to extract the post-processing results. Figure 1.7 shows the headform with the rigid wall at the bottom.

Fig 1.7: Headform geometry with a rigid wall

*CONTROL_TIMESTEP card is pre-defined for this analysis. LS-DYNA calculates the time step automatically for all the elements and uses the minimum value for the analysis. The calculation is done through the following equations,

`∆t=L/C` Critical time step (1)

`C= √(E/ρ)` Beam element (2)

`C= √(E/(ρ(1-v^2)))` Shell element (3)

`C= √((E(1-v))/(ρ(1+v)(1-2v)))` Solid element (4)

`L`= characteristic element length

`C`= Speed of sound through an element

`E`= Young’s Modulus

`ρ`= Density

`v`= Poisson’s ratio

But there might be very small elements which need a very small time step for satisfying the Courant-Friedrich-Levy (CFL) criteria. This increases the computational time and even requires a higher computation cost. There are a few ways to resolve this issue. One way to resolve this problem is to remesh those regions with a coarse mesh and another solution is mass-scaling. In mass-scaling, the mass of elements, having time step size below the target time step size is increased artificially thus increasing the density. Time-step size is directly proportional to the density of the material. Thus the minimum time step size required for that particular element increases. But this artificial mass also increases the kinetic energy of the system which is undesirable. For dynamic analysis, the mass scaling should be restricted up to 5%. Time step size for mass scaled solutions is defined by DT2MS value. This is minimum time step time permitted and mass scaling is only done if it is necessary to meet the CFL criteria. Negative values are used for dynamic problems where inertial effects are significant, the actual value considered is TSSFAC*|DT2MS|.

The final calculated time step size is also scaled with appropriate scaling. For stability reasons, this value is kept less than 1 and the time step scale factor (TSSFAC) is defined as 0.9 for this analysis. The maximum time step size is defined by the load curve. Figure 1.8 shows the card assignment for the discussed properties.

Fig 1.8: *CONTROL_TIMESTEP card assignment

Control energy card is also pre-defined to get all energy statistics in post-processing. All other details in the input keyword file are not changed and kept as default. The termination time is given as 5ms to replicate the impact and consider the aftereffect of the impact. This input keyword file is saved as a new .k file similar to the previous case.

3) Case 3- Headform impact on hood:

The third part of the study is performing and analyzing the headform impact on the hood of a car. The meshed hood geometry is given with the problem statement. The impact direction and velocity of a head on the hood is kept identical to the previous cases. For this, the transformed headform geometry from the previous case is used as it is. This head geometry is included with *INCLUDE card to the given hood input file. This new input file is saved as a different .k file which will be solved by the solver.

Open the meshed hood input file in LS-PrePost. This file has only mesh and geometry data. So, a piecewise elastic material is assigned for hood and shell section property is assigned along with that. The material properties of aluminium are used to define hood. Fig 1.9 shows the material properties assigned for the hood. Automatic surface to surface contact is defined between the hood and head geometry. The velocity of 7.8567 mm/ms is given to the geometry along negative X and negative Z direction. For selecting head as a complete one part, a new part set is created through create entity utility. All the other properties of head geometry and hood are kept to default values. Most of the properties necessary for simulation have been pre-defined.

Fig 1.9: Material definition for the hood

Termination time of 20ms is given to consider the impact and after effect. GLSTAT, RCFORC, SLEOUT etc cards are defined to get post-processing results. *DATABASE_EXTEND_BINARY card is used which writes the strain tensor to d3plot. STRFLG is set to 1. Figure 1.10 shows the case set up for the headform to hood impact simulation. This input file is called by the solver and the results are extracted in LS-PrePost.

Fig 1.10: Headform geometry with the hood

RESULTS AND DISCUSSION:

Various output parameters are evaluated and are discussed in the following section.

Effective stress values:

Figure 1.10 shows the stress behaviour of simple head after the impact. The stress values are higher at the instant of impact and reduce with fluctuations after the impact as the material is purely elastic. There is no significant deformation in the simple head geometry. Figure 1.11 depicts stress behaviour in headform geometry after the impact. The material properties for this geometry are pre-defined and different from the simple head so, it shows different behaviour. There is significant deformation in the geometry after the impact. Figure 1.12 shows that the hood deforms after the impact and there is less deformation in the head compared to the second case(figure 1.11). This is because of the elastoplastic material which has been defined to the hood, this material tends to deform elastically till the yield point and then deforms plastically after that.

Fig1.10: Effective stress (v-m) contour for simple head impact on a rigid wall

Fig1.11: Effective stress (v-m) contour for headform impact on a rigid wall

Fig 1.12: Effective stress (v-m) contour for headform impact on the hood

Strain values (elastic + plastic):

The strain values for all the three cases are extracted from head geometry. A bottom node of head geometry is considered for strain value extraction. The effective plastic strain value is plotted from History>Element>Effective Plastic Strain. We have defined STRFLG as 1 in *DATABASE_EXTEND_BINARY this actually gives total strain (elastic + plastic) state of deformation. So the plotted effective plastic strain is the total strain and not just the effective plastic strain.

Fig 1.13 shows the total strain for simple head impact on a rigid wall. The total strain value is zero throughout. This is because of the high young’s modulus assigned to the simple head geometry (210GPa). This was a trail case and hence the material parameters were chosen randomly. It is also visible from fig 1.10 that there is no deformation of simple head geometry after the impact this relates to the very low strain values.

Fig 1.13: Total strain after impact of simple head geometry

Figure 1.14 depicts the total strain in headform geometry before and after the impact. Here the headform has been designed to replicate the child’s head. So this has similar properties which resemble an actual head. On the other hand, it is hit by a rigid wall which is basically a non-deformable body. So, it can be seen that after the impact the strain value increases (plastic + elastic) and the drop in strain value after 2.5ms is due to the recovery of elastic strain. After the impact, the headform bounces back and we can see the elastic strain recovers fully but the plastic strain remains as a permanent deformation till the end. This same element is used to calculate the strain for the next case.

Fig 1.14: Total strain after the impact of headform and rigid wall

Figure 1.15 shows the total strain of headform impact with a hood. In this case, the hood is assigned with the aluminium properties. So, it can deform and is not undeformable like a rigid wall. The headform geometry is same as case 2 with material properties resembling the child’s head. From the figure, we can see the total strain increases gradually during impact. During impact, the hood deforms significantly and hence this reduces the impact on headform relatively. From figure 1.12 we can get a rough picture that there is not much deformation in headform compared to case 2 that is figure 1.11. Figure 1.15 shows fewer strain values compared to figure 1.14 this justifies the above explanation. The strain value reduces at the end this is due to the elastic recovery of the headform material.

Fig 1.15: Total strain after the impact of headform and the hood

Energy plots of the system:

The energy statistics are extracted through glstat in LS-PrePost. Figure 1.13 depicts the energy plot for a simple head impact on rigid wall simulation. It is visible that the kinetic energy decreases and internal energy increases at the instant of impact. This is due to the reduction of the velocity during impact. The simple head geometry being elastic material bounces back without major loss of energy and attains velocity in other direction. This causes again increase in kinetic energy and a decrease in internal energy further. The internal energy remains constant and the hourglass energy is also under desired limits.

Fig 1.13: Energy plot for simple head impact on the rigid wall

Figure 1.14 depicts the energy plot for headform impact on a rigid wall. The kinetic and internal energy show similar behaviour during impact. But as the headform is modelled with hyper-elastic material and rigid body there is a gradual change in the energies after impact compared to simple head impact case (fig 1.11). Some of the energy is gone in deformation caused due to impact. Hourglass energy is under desired limits for this case too. We can see the percentage increase in mass is also very low and hence there is no addition of artificial mass which may cause non-physical behaviour.

Fig 1.14: Energy plot for headform impact on the rigid wall

Figure 1.15 depicts the system energy for head and hood impact. We can see the kinetic energy decreases and internal energy increases gradually compared to other cases. The first two cases has an impact with a rigid wall so there was an abrupt change in the energies during impact whereas hood is made up of aluminium material defined with piecewise linear plasticity model. The hood deforms gradually and slows down the head velocity. Hourglass energy is again under desirable limits and the percentage added mass as well is in limits, almost zero.

Fig 1.15: Energy plot for headform impact on the hood

HIC-36 plots:

HIC value is generally derived from the accelerometer mounted at the centre of mass of the crash test dummy. For the simple head impact test, a node on the outer surface was manually selected for result extraction. HIC-36 value is calculated in LS-PrePost and is basically measured for a time interval less than 36ms duration. From figure 1.16, it can be seen that the HIC-36 value for simple head impact is 2.083e+07. Whereas, from figure 1.17 it can be seen, the HIC-36 value for headform impact is 2.574e+4. The HIC-36 value for headform impact is measured from an accelerometer place in the geometry. From both the figures it can be seen, the resultant acceleration is increased and then decreased after the impact because the body comes to halt and again bounces back. Figure 1.18, shows the HIC-36 value for head and hood impact. The calculated value from LS-PrePost is 280.8. If the time interval used to calculate the HIC value is shorter than 15ms, both HIC15 and HIC36 have the same value. If the time interval is larger than 15ms then HIC36 is higher.

Fig 1.16: HIC-36 for simple head impact calculated in LS-PrePost

Fig 1.17: HIC-36 for headform impact calculated in LS-PrePost

Fig 1.18: HIC-36 for headform-hood impact calculated in LS-PrePost

CONCLUSION:

The head impact simulation is performed with all relevant post-processing results.

Case 1, gives an idea of a general analysis setup. This analysis is performed with random material properties and case set up is made to perform further analysis. The simple head geometry does not deform and has negligible strain values after impact. This case has the highest HIC value due to the harder head and rigid wall.

Case 2, gives a brief idea about head impact with the rigid wall. This impact with a rigid wall is severe and it deforms headform geometry. Higher plastic deformation is seen after the impact. The HIC values are lower than case 1 but higher than case 3. This is because of different headform material properties compared to case 1.

Case 3, gives an idea of the actual head impact on car hood. The hood being a deformable body with elastoplastic material properties deforms during impact. The effect is less severe comparatively and HIC values are low for this case compared to case 1 and case 2. This signifies that the probability of head injury in this particular case is low as compared to case 1 and case2.