High magnitude of vertical acceleration

Hi, guys

I would like to use AnyBody to simulate a seated human experiencing vertical acceleration with an amplitude up to 15G. Is it possible for AnyBody to deal with that?

Hi @zhao19920409

It depends a lot on the scenario. It is possible to do quasi-static analysis of a subject during the 15gs, but from your message I sense that your motion is more dynamic i.e. the load is applied fast as in a catapult seat or similar. These scenarios are currently not possible to model inside the AnyBody Modeling System.

You also have to consider which human body structures are going to absorb the forces of such a motion. The AMS only recruits muscles, but other tissues and bones might take up a substantial amount of the force and the model cant handle that.

Best regards,
Bjørn
AnyBody Technology

Hi, Bjorn,

Thanks a lot for your explanation.

I am currently using the seated human model to test the muscle response under vertical sinusoidal acceleration, and the frequency is about 3Hz. I think it should not be a fast motion that AnyBody can't deal with. I have tried to add the motion in AnyBody and got some positive reflections about the muscle activity and force.

May I ask another question about the AnyKinEqFourierDriver? I found that the function constrains the displacement, is there any driver that can control the acceleration directly, for example, I can add sinusoidal function of time vs acceleration?

Regards,

Zhenkai

Hi @zhao19920409

The 3Hz should be possible indeed.
As for the drivers, no we do not have a driver to constrain the acceleration as you need. The AnyKinEqFourierDriver do only accept position data so you need to integrate your acceleration data to get the positions and use those in the driver instead.

Best regards,
Bjørn
AnyBody Technology

Hi, Bjorn,

Thanks for your suggestions. After integrating the acceleration data, the amplitude of position function is 0.146, but when I run the inverse dynamics, an error occurs after few time steps "solver aborted due to singular KKT matrix". When I reduce the amplitude value, the solver can run, but it is not realistic. Is there any solution for the error?

Best Regards,

Zhenkai

Hi @zhao19920409

The KKT error typically indicates that you have a problem with balance in the system. For example, one of the degrees of freedom is not supported by muscles or reaction forces or environment, or supporting muscles are too weak, or accelerations of the relevant segments are too high for the muscles to generate such motion.

Try and look at the other aspects of your model to see if any constraints or reaction forces are missing or working unintended

Best regards,
Bjørn
AnyBody Technology

Hi, Bjorn,

I have checked the number of Dof and constraints, and they were equal.

The acceleration function I would like to use is 74sin(2pi*3.58t) and after integration, in AnyKinEqFourierDriver,
Freq = 3.58;
A = {{0, 0.146}};

I have tried to lower the frequency or the amplitude, i.e. lowering the final amplitude of acceleration and the error didn't occur, so I guess it might be caused by the weak muscle or high acceleration of the seat, is that right?

If so, is it possible to introduce a buffer region of acceleration before it reaches the desired acceleration value to stop the error, or is there any other method to deal with the problem?

Thanks a lot.

Regards,

Zhenkai

Hi, Bjorn,

Just an update for the problem. I tried to focus on half period of the sine function in the AnyKinEqFourierDriver and there is no error. After the half period, the error appeared again.
If you need more information, please don't be hesitate to tell me.

Regards,

Zhenkai

Hi @zhao19920409

How did you model the seat and seat-human interaction?
And what degrees of freedom are you using for the fourier driver?
feel free to include snapshots of your drivers/constraints if that is easier.

Best regards,
Bjørn
AnyBody Technology