Dummy kinematics in frontal impact, comparison of simulations in Madymo with sled tests

The paper aims to respond to the current requirements regarding the usage of modern virtual analysis tools in accidentology research, presenting the methods of performing the virtual and experimental studies. The main objective of this paper is to determine the movement of the dummy by comparing the kinematic parameters provided by the simulations performed in Madymo with the experimental tests (sled tests). A prototype test device was used to analyse the movement of the dummy in a vehicle cabin at different impact situations of a frontal collision. The experiment was performed for two different cases: with seatbelt and without seatbelt. For the virtual study in Madymo two different setups were prepared in order to obtain the results for both testcases based on a parameterized model. The analysis of the results is performed by comparison, between experimental results and simulation results. Three different phases are taken into consideration in order to make a study of a frontal collision: pre-collision, collision and post-collision.


Introduction
In the past years, the number of road accidents has increased significantly, and the development of prevention techniques has become imminent.Whether we are talking about active safety systems, road accident reconstruction programs or virtual analyzes through computer-aided simulation tools, their continuous improvement must lead to a reduction of the number of road accidents and to have an important contribution in decreasing the severity of injuries 0.
To study the responses of occupants and injury criteria in vehicle collisions two approaches are used in this work.One approach consists of performing experimental tests of real collisions in polygon or using a sled test device and crash test dummies.This aspect has been highlighted in recent studies where a prototype dummy was developed and compared to a Hybrid III 50th percentile male dummy through sled test 11].Other recent studies present various dummies used to determine head injury criteria.For example, a prototype dummy head is used on a test-rig, recording the kinematic parameters through which the head injury criterion (HIC) will be determined 0, or when head injury index is assessed for standard Hybrid III 5th, 50th and 95th dummies using the same restraint system 0.
Other approach is to perform virtual studies conducting computerized simulations in which the occupant kinematics is predicted and assessed by software packages such as Madymo or LS Dyna.In a relevant study 0, published in 2019, the reconstruction of a frontal accident is carried out, the movement of the vehicle being simulated with the help of PC-Crash and to obtain the kinematic 1303 (2024) 012035 IOP Publishing doi:10.1088/1757-899X/1303/1/012035 2 parameters and implicitly the response of an occupant Hybrid III 50th male, Madymo is used.With the kinematic parameters obtained, an assessment of the severity of head injuries was also made.
The two methods of conducting the experimental and virtual study were compared in the study 0, where a THOR dummy restrained by a standard 3-point shoulder subjected to two frontal crash sled tests was evaluated.By comparison, numerical simulations were carried out following the same procedures as in the case of the experimental test.The dummy kinematics and the interaction forces of dummy with the sled and the belt were recorded from the experimental test and compared with the simulations.In 2021, the study 0 was published, in which a comparison is made between the capability of the Madymo to predict the response of the occupant and human volunteers in the event of a frontal collision.
The main objective of this study is to determine the dummy movement by comparing the kinematic parameters provided by the simulations performed in Madymo with the experimental tests (sled tests) 0. The aim of conducting this comparison is to calculate the probability of injury correlated with the kinematics of the dummy, using the head injury criterion in a time interval of 36 ms.

Testing procedure of the experiment
For the test environment, the parking lot from the Research and Development Institute of the Transilvania University of Brasov was used.In this location was placed a prototype test device, Figure 1, used to analyse the movement of an 84.12 kg prototype dummy in a vehicle cabin within different impact scenarios in a frontal collision.The experiment was performed for two different cases: with seatbelt and without seatbelt.

Equipment features and performance
A high speed camera which is capable of recording at 600 fps, was used to record the actual impact.To ensure the correct light regardless of the test conditions, led projectors were used, Figure 2.

3D modelling in the Madymo
The 3D modelling was performed in Madymo using a parameterized model of a frontal impact.The occupant was a Hybrid III 50% male dummy of approximately 78 kg.Two different setups were prepared in order to get the results for both test cases: with seatbelt and without seatbelt.
The pre-processor XMADgic was used for editing the .xmlcontent.For second test case (without seatbelt) the corresponding node was disabled (Figure 3).Furthermore, for the two load cases, the simulation parameters start and end time were modified (Figure 4), but the time step was used by default.The period of the simulation time, measured in seconds, was correlated with the values of the time obtained from the experimental data.To include the motion of the sled in the simulation, its velocity and the acceleration on X component were added in the .xmlcontent (Figure 5).For this type of crash, the acceleration on X component is the most significant.As output request the .CSV format was chosen to get the output signals of the dummy 0. The next step was to validate the document (Figure 6) and to ensure that the solving process can start.Following the indications from 0, two setups were created, as presented in Figure 7.For the processing part was used the default solver and the job was submitted (Figure 8).

Data acquisition and processing
No accelerometers were used for measurements, instead the acquisition of the experimental data was performed using the Tracker software which is a video analysis tool performing both, measurements and calculations, on X & Y axis in a relatively short period of time for different landmarks in the videos recorded with the high speed camera.
In Madymo some CFC (Channel Frequency Class) filters were automatically introduced in postprocessing as a result of processing 0. For instance CFC 1000 for head accelerations, CFC 180 for thorax accelerations and CFC 180 for velocities.All the values were exported in .csvformat and processed in the Microsoft Excel.
For a better visualization of the velocity graphs, the values corresponding to the rebound movement for both, Madymo and sled test, were multiplied by -1.The software Origin was used to obtain smooth curves on the graphs from Microsoft Excel, this being done with the help of FFT filtering.
A measure that can be used to study and analyse the causes of a road accident, regarding the mechanism of brain injury, is the HIC (Head Injury Criterion, defined by NHTSA), which is determined with the formula: Where: a R =√a x 2 + a y 2 + a z 2 a R -the resultant acceleration of the components (a x , a y , a z ) [g] t 1 ,t 2 -the initial moment, respectively the final moment of the time interval during which the evolution of the resultant acceleration is considered, in which the HIC reaches its maximum value for 0 < t The risk of injury is evaluated by the probability of producing injuries of a certain severity 0: The severity of injuries is assessed according to the AIS scale and Table 1 shows the most common types of injuries that occur in road accidents 0. Severe (with danger of death) Critical (uncertain survival) Unsurvivable Figure 9 shows an equivalence between the HIC criterion for head injury and the AIS scale, which assesses the severity of injuries.The limit value of human tolerance is considered 1000 for the HIC criterion, a value above which human life is in danger.The limitation of HIC is that it does not take into account the angular acceleration of the head, as it is calculated based on the resulting translational acceleration on the 3 axes 0.

Results analysis
The kinematic parameters for the head and thorax of an 84.12 kg humanoid anthropometric device (dummy) were recorded through two experimental frontal impact tests conducted at a velocity of approximately 42 km/h.Two simulations were in Madymo with a Hybrid III 50% dummy of approximately 78kg, and their results were used for the comparison with the measurements results of the experimental frontal impact tests.The analysis of the results is performed for two cases when the occupant is equipped with seatbelt and without seatbelt.In order to study a frontal collision, three different phases are considered: pre-collision, collision and post-collision.
Both movements of the belted dummy during the experimental test, respectively the Madymo simulation, are illustrated in the Figure 10 describing the collision phases.
The dummy movement is analyzed in 3 different moments at 0 ms, 50 / 70 ms and 183 ms.The collision phase of the experimental test has a delay of approximately 20 ms, the aspect being visible on the plots in Figure 11 which presents a comparison of the kinematic parameters obtained to the time interval [0, 0.1883 s].It can be observed on the graphs that the velocity was approximately 42 km/h (11.7 m/s) before the collision.
For the case when the occupant is equipped with seatbelt, the peaks of the kinematic parameters are reached when the head hits the airbag, being protected from significant damage with the interior parts inside the dummy compartment.The thorax is held on the seat with the help of the safety belt, without significantly rotating 0, 0. Some regulations (ECE-R 17 in Europe and FMVSS 201 in the USA) provide as a performance criterion for the head impact the limit of deceleration in the antero-posterior direction.The head injury tolerance limit is 80 g over a maximum period of 3 ms 0.  Regarding the biomechanical limit of the thorax, the FMVSS 208 regulation provides that the limit is the acceleration value of 60 g for a maximum duration of 3 ms when the occupant is belted 0. In this case the greatest acceleration values on X axis of the head and thorax when the occupant is belted don't exceed the limits.The seatbelt can cause injuries only at relatively higher velocities or if it is not positioned correctly, which consists of breaking the clavicle on the side of the shoulder belt, tearing the liver, damage to the small intestine etc. 0.
Collision phases for the unbelted dummy are presented in Figure 12.The values obtained from the experimental test as well as from the simulation in Madymo are illustrated comparatively by representative graphs (Figure 13), observing that all tests were conducted at similar velocities before the collision: For the experimental test, the unbelted occupant hits the ceiling with his head during the impact representing the peak of the head acceleration value displayed on the graph at 0.07 s approximately.In the Madymo simulation, this aspect results from the similar acceleration values obtained, being close with the ones from the experiment 2, 0.Even if the limits according to the regulations ECE-R 17 and FMVSS 201 are exceeded, the occupant is not wearing a seatbelt as tested by these standards.According to some specialized literatures, the tolerance limits of the head are described more detailed and more classifications are available, for example for the forehead and chin 80g, respectively 60g and for the crown of the head the limit is 200 g.It is observed that the greatest acceleration values of the occupant's head on X axis when not wearing the seatbelt are between 112 g and 128 g and even if the biomechanical limit for the crown of the head is not exceeded, the impact occurs in the area of the dummy's forehead exceeding the injury tolerance limit.
The occupant can suffer external injuries such as skull fractures and internal injuries such as displacement of the brain in the skull or other injuries to the face or brain, with the risk that some of these may be irreversible, even fatal 0. The thorax injury tolerance limit for an occupant who is not wearing a belt, according to the FMVSS 208 is 60 g for a maximum duration of 3 ms.The thorax acceleration values don't exceed the tolerance limit value using both analysis methods.However, a risk can be associated if the unbelted occupant is close to the airbag at the time of its inflation, fatal injuries can occur through a strong impact to the neck and chest like an explosion, which can cause ventricular fibrillation, tearing of liver or lung injuries.Injuries can also occur in the upper cervical area of the spine because the airbag reaches under the chin 0.
It is important to mention that the study has some limitations, and the accuracy of the measurements was influenced by factors that were not taken into account, such as the position of the seat, the height of the safety belt or the position of the anchor points, the differences in the design of the dummies, the position of the markers, as well as in the generation of collision forces, amplifying the rebound movement, the test device being equipped with air cushions mounted in the frontal side.For technical reasons of visibility, the thorax markers were mounted on the upper part of the dummy's arm, the difference in weight mainly leading to different measured kinematic parameters compared to the values provided by Madymo.The values obtained for the acceleration of the dummy's head on X axis were close in the experimental test compared to Madymo when the occupant is unbelted.However, higher head acceleration values on X axis are observed in the results provided by the experimental test compared to Madymo, the difference being more noticeable when the occupant is belted.The differences are mainly caused by a different design of the dummy in the experimental test, more specifically a neck joint with a different stiffness 0. In the current study, the HIC criterion was calculated in a range of 36 ms, as can be seen in Figure 14 and Figure 15, the range in which the head injury criterion reaches the maximum value being marked according to the origin of the resultant acceleration values, namely sled tests or Madymo.
In the case of tests where the occupant is equipped with a seat belt, the calculated HIC value is 950 for the sled and for the Madymo 737.The values do not exceed the human tolerance threshold of 1000, but correlated with the AIS scale, for the sled corresponds to code 3, and for Madymo the code 2, which represents a moderate or even serious severity, but without the risk of fatal injuries.The probability of injury is 8.1% for Madymo and 15.6% for sled tests.
When the occupant is unbelted, the calculated HIC value is 1768 for the sled and for the Madymo 1483.The human tolerance limit is exceeded and the severity of the injuries is fatal for the occupant in the case of the sled test, reaching index 6 on the AIS scale and critical with uncertain survival in the case of the Madymo simulation with index 5 on the AIS scale.The probability of injury is 54.6% for Madymo and 76.6% for sled tests.During a real collision, the vehicle absorbs through its deformation the kinetic energy and protects the occupant, while for the sled test device, the kinetic energy produced during the impact is transferred to the dummy 0. The obtained results highlight the limitation of a sled device, as it is not able to reproduce the current response of a real vehicle during a collision and both the HIC values and the probability of producing injuries are higher in the case of tests performed on the sled compared to the results provided by Madymo, where the values and probabilities of injuries are lower, simulating a real impact.

Conclusions
The occupant movement in case of a crash is seriously analyzed by the car manufacturers so they can develop new and more performant systems that reduce the effects of the collision on the passengers.The in-car movements experienced by the occupants depend on multiple aspects such as the type of impact, vehicle type, adjustment of the interior parts and vehicle occupancy.
During the impact, the injuries of the occupants are mainly caused by the movements of the occupants and different parts of the vehicle as a result of the forces developed in the impact, also known as collision forces.
The experimental research was carried out by creating simulations of the situations in which the road accidents occur, from the point of view of the occupant's behavior in case of an impact.This work represents a research of the kinematics of the occupant, in situations where the occupant is equipped or not with a safety belt.This type of impact is very common in real-life cases of giving way at intersections, roundabouts and more.The situations presented can occur when the occupant neglects to wear the seatbelt or when the restraint device breaks down.
The novelty element of this work presents the usage of a prototype test device specially designed and executed for carrying out impact tests.A prototype dummy was also used.The obtained results are compared with the data provided by Madymo software.The advantage of using the research test devices versus testing by using the vehicles on a scale of 1:1 in the polygon is primarily the possibility of controlling the test conditions and ensuring the repeatability of the tests, and then the difference in logistics costs.
The methodology used and the data obtained through the experimental tests can be used to improve the test device, during the study being identified the main technical limits that led to the differences in the kinematic behavior of the occupant.
Once the identified limitations of this study are removed, an improved correlation of the experimental test with the model of Madymo software can be achieved.This aspect will lead to a reference Madymo simulation set-up which can be used for further simulations, resulting into a cut of the financial and time resources for the experimental set-up.On top of that, the numerical simulation method used in this research can also be applied for the verification and validation of other test devices or prototype dummies used in crash simulations 0.

Figure 5 .
Figure 5. Motion parameters of the sled.

Figure 10 .
Figure 10.Collision phases when the dummy is belted.

Figure 11 .
Figure 11.Kinematics diagrams when the dummy is belted.

Figure 12 .
Figure 12.Collision phases when the dummy is unbelted.

Figure 13 .
Figure 13.Kinematics diagrams when the dummy is unbelted.