Dummy Crash Simulation problems
Altair Employeeand crashes into an obstacle with a 90° angle towards the obstacle (straight front crash).
The person is represented by a dummy (50% Male), the seat structure was modeled by altair, as well as the dummy.
I created two seatbelts with the Hypercrash tool Belt Generator (Hypercrash --> Safety --> Belt Generator) and deformed the seat with the seat deformer (old)
(Hypercrash --> Safety --> Seat Deformer --> Geometrical displacement (old)).
Due to simplification I removed the complete lower sector of the seat group and replaced it with a rigid body, which is linked to
the floor and the seat.
I built the model as follows:
-Initial velocity: applied to all parts of the model (13.88 m/s)
-Imposed velocity: A function, which decreases the speed from 13.88m/s to 0 within 30 ms. Applied to all parts but the dummy itself.
-Gravity load: To maintain physical correctness of the dummys motion.
-Added Mass: Applied to all nodes of the dummy to be representative for the human weight.
-BC for the seat: All degrees of freedom are locked except for the translation in X-axis (which is the direction of motion).
This is applied only for the floor and the ground part of the seat.
-Interior contact: Contact between the two seat layers (was set by altair when I imported the model, I did not change any values)
-Seat to floor: Contact seat to floor (TYPE7)
-Feet to floor: Contact feet to floor (TYPE10)
-Seatbelt to dummy: Contact upper/lower seatbelt to dummy/seat (TYPE7)
-Dummy to seat: Contact Dummy to seat (TYPE7)
-The Runtime for a single simulation is about 9500 seconds. Can this be regarded as normal, considering the model complexicity?
-Results of the simulation differ too much from expactations/reality
-->Normally I would have expected the backrest to at least do a little motion, but it remains in its position.
-Because of the constraints of the seatbelts, the head of the dummy experiences a very high acceleration.
-->Is there any chance, that the Belts ease a little bit (in the beginning) and afterwards tension up, just like in reality?
-All elements of the part 'LINK', ID:6 delete on the beginning of every simulation. I could not figure out the reason for that, neither in the starter file, nor in the engine file.
Thanks in advance.
Answers
-
Hi,
We have a similar tutorial in our Help Menu, RD-T: 3150 Seat Model with Dummy and I recommend you to go through the same.
At the seat belt anchorage point you can add a pretension spring property (/PROP/TYPE32) so that you can simulate it like in reality. A time sensor is created for this and and is recalled in the sensor of this property card. A Stif0 value is added to the spring and it will be the stiffness before sensor activation. As soon as the time specified in the Tdelay is reached the spring will act based on the the force vs displacement function the user specify. In the Help tutorial we have used this option.
Your model has many intersections and I think it's causing the issue of node shooting. Clear all the intersections and penetrations in the model.
0 -
Hello,
thank you very much for your hints.
There are still some unanswered questions:
I have a Simulation, where the Dummy and the airbag are merged. The airbag ignition and inflation did work very well, when they were not merged. When the models are merged into one model and get an initial velocity and a decceleration via an imposed velocity, the airbag can not infalte well. There are not any contact problems, I did check out on this.
Within another simulation I just placed the airbag into the final Dummy position and it did ignite and inflate well, so I assume, that the problem is with the velocity and decceleration (see video)
Is there a solution, that I can solve this problem, that the airbag can ignite and inflate normally, while moving and decellerating?
Also I noticed, that the dummy does not 'lean' forward enough, when deccelerated. Thus he can not fall deep enough into the inflated airbag. I also checked the belt force limiter and also simulated once without an upper belt, but the dummy just does not lean forward (see video)
Attached you can find videos of the simulation and the model itself.
Thanks in advance.
Greetings,
Mohammad
0 -
Hi Mohammad,
Reviewd your model.
Please try running with these recommendations:
1). For the imposed loading condition please remove all the nodes from the airbag surfaces.
2). For the contact between airbag and dummy create a symmetry interface. And for this contact between airbag and dummy use the recommended parameters:
Istf=4
Igap=2
Fscale_Gap=0.8
Stmin=1
INACTI=6
Gap_min=
Iform=2.
Please update the model with these changes and try running the model.
0 -
Hi George,
thank you very much for your immedeate answer.
Everything worked fine!
Now there are other problems, that occur, namely:
The dummy just does not move Forward into the direction of Motion. The upper Body does not rotate and thus does not really dive into the Airbag.
I have already checked following Points:
-Center of gravity of the dummy. It is identical with the COG of the root model, which I got from the Altair Folder
--> Can this dummy even be used for this Kind of Crash Simulation or do I Need to take another one? The Name of the Dummy is: 'HYBRID III 50% DUMMY FINE MESH V_1.2'
-The inertia of the dummy's rigid Bodys. Same here
-The torque and Tension stiffnesses of the spring Elements, which refer as Joints within the dummy
-The belt force limiter function aswell as the belt pretensioner
-I simulated the whole model without an upper belt attached to the dummy --> still no Rotation of the upper body
-The contact Interfaces were updated as advised by you, so there are no frozen contacts at all.
Attached you can find the latest Video of the Simulation results.
Note: The contact Interfaces between the Airbag and the steering wheel Need to be refined, I am Aware of that.
Greetings, Mohammad
0 -
Hi,
Can you share the latest model file?.
0 -
Hi Mo,
in case if you still have this problem, I solved it by adding 'added mass' to the dummy.
Taha
0