pendulum model

[pauline neerpelt]

Hi everyone,
I'm new to OpenSim and I have a few questions about this simple pendulum model:

figure1

Fig 1.png (5.91 KiB) Viewed 484 times

I'm using SO and CMC to determine the muscle forces necessary to keep the pendulum in a static neutral position (see figure 1). The muscle placement is completely symmetrical and each muscle has the properties displayed in Figure 2. I get these results for SO and CMC:

figure2

Fig 2.png (48.99 KiB) Viewed 484 times

Fig 4.JPG (18.27 KiB) Viewed 484 times

I used the output reporter to obtain the multiplier values and the formula from the Millard 2012 muscle model to make sense of these muscle forces.
Based on the results, I'm left with the following questions:
1) Why are the muscle forces and the muscle activations not equal to zero when using SO?
2) Is it correct that the passive-force-length multiplier is assumed to be zero during SO?
If so, why is this the case?
If so, would it make sense to add passive muscle stiffness (based on literature) to an OpenSim model in the
form of a coordinate limit force to include passive muscle stiffness when performing SO?
3) Should I have added an external load to represent the ground reaction force during SO and CMC?

Thanks in advance for any help

Kind regards,
Pauline

[Carmichael Ong]

Thanks for providing a smaller example and details. Some comments below:

1) Why are the muscle forces and the muscle activations not equal to zero when using SO?

Hard to say for sure, but my guess is that since this is an optimization, it's possible that the tolerances of the optimization are being satisfied at this low (but not quite 0 level), so it considers the problem converged. I was wondering if it was possible that for the muscle to go to exactly 0 activation, but with a rigid tendon, it does seem possible given the code here?

2) Is it correct that the passive-force-length multiplier is assumed to be zero during SO?
If so, why is this the case?
If so, would it make sense to add passive muscle stiffness (based on literature) to an OpenSim model in the
form of a coordinate limit force to include passive muscle stiffness when performing SO?

That is correct for SO as implemented by default in OpenSim. My guess is that this happened because these models are often too tight, especially for a motion such as walking, so it was easier to remove the passive forces and still get a reasonable answer. Whether adding passive forces back in the model is useful will depend on your research question. Other researchers have created their own custom static optimization methods to include passive muscle stiffness with the muscle model (so you don't have to add extra coordinate limit forces).

3) Should I have added an external load to represent the ground reaction force during SO and CMC?

This depends on what you are trying to model. As long as all constraints and external forces are accounted for in your system, then there is no need to add extra loads.

[pauline neerpelt]

Thank you very much for your reply! I still have a few questions about what you mentioned. Could you clarify what you mean by the models being "often too tight"?
Also, I'm currently working with the ARMS Lab Hand and Wrist model and I'm curious about the resulting force on the radius when the wrist is in static extended position. I'm using static optimization and joint reaction analysis for this.
In this context, would it be advisable to incorporate passive muscle stiffness into the model?
Additionally, am I correct in assuming that no ground reaction forces should be integrated into this model?
Thanks in advance for any help.

Kind regards,
Pauline