Simscape's Slider-crank doesn't move between a min/max, instead diverges.

Hello, I am having problems to understand why the simscape's slider-crank mechanism drifts like in the following picture

Before the slider-crank there is only a Ideal Angular Velocity Source and an inertia.

Slider-crank's parameters are:

the translational sensor is working in absolute measurement reference.

Why is it drifiting up? How do i Constraint it to work between a min and a maximum?

Thank you for your time.

GM

Answer 1

[Steve Miller]

Can you share the model or let us see your solver settings?

I believe your model is drifting because the solver choice or solver tolerances are not appropriate for the frequency at which your system is oscillating. Try manually setting to ode23t or daessc and limit the maximum step size to make sure there are enough steps per oscillation to capture the movement.

--Steve

[Gianluca Mammone]

This model is part of a bigger project too complex and not interesting for our problem. I tried to replicate it in a different file, with same parameters etc, but it doesn't replicate this behavior.

Since this model is in a Simulink file with a Stateflow Chart and i left variable-size with Automatic solver, I think you're right and it's a problem of such nature.

I tried ode23t but it still doesn't work, with ode45 instead it seems to work fine!

I'm leaving this thread open for a while to be sure this was the issue, then if everything is all right i'll mark your answer as correct.

For the time being, I thank you!

[Steve Miller]

ode45 is normally not recommended for Simscape models - it is an explicit solver, not great for systems of implicit equations. I am guessing the main reason it is working better than ode23t is because ode45 tends to choose a very small step size, and small steps always help when there is a numerical issue.

I am guessing there is something else in your model outside of Simscape that is causing this behavior. I am glad you have a way forward for the moment, but I would normally not recommend using ode45 with Simscape models.

--Steve

[Alexandre]

Hi Steve, I encounter the same issue that Gianluca mentionned, for a simple slider-crank model, and in a more complex multi-physic model.

I tried to reduce the max step size of the solver, to change the solver, but my slider-crank doesn't behave as expected. The oscillations representing the stroke of the slider are moving upward, reaching the maximum absolute stroke I specified (200 mm in my case) :

Is there any other solution than modifying the solver configurations ?