Hey there, I need some help with this error. Would anyone be willing to provide guidance?

How do I reduce the fixed step size or tighten the error tolerances?

Answer 1

[Sam Chak]

HI @Danikha Shyken

For a fixed-step solver, you can modify the settings in the Solver tab of the Configuration Parameters dialog box (see image below). However, when it indicates that the derivative of the state at a certain time is not finite, the issue may be caused by a singularity in the solution. Singularities typically occur when the model contains division-by-zero terms.

If this is under the direct control of a model designer, you can alter the mathematical expression that leads to the zero in the denominator. If the model is constructed in accordance with the governing laws of physics, where you cannot alter the mathematical principles (e.g., black holes or kinematic singularities of robot arms), then you can attempt a coordinate transformation or apply an externally-manipulated input signal to the system to actively avoid the singularity.

[Danikha Shyken]

It makes sense now, thanks a bunch!

[Danikha Shyken]

How can I confirm that this is the case, and what steps should I take to resolve it?: If this is under the direct control of a model designer, you can alter the mathematical expression that leads to the zero in the denominator. If the model is constructed in accordance with the governing laws of physics, where you cannot alter the mathematical principles (e.g., black holes or kinematic singularities of robot arms), then you can attempt a coordinate transformation or apply an externally-manipulated input signal to the system to actively avoid the singularity.

[Sam Chak]

I suggest that you show what's under the state-space block mask so that we can analyze the dynamics.