how to create or structure a big model?

Hi there,

I am doing a big model of an engine with matlab-simulink. I have tried to create some sub-systems in simulink with my differential ecuations. However, I have had some errors and warnings such us this one:

"Derivative of state '1' in block 'GeometriaPiston/Geometría del pistón/Integrator' at time 0.057359477517074493 is not finite. The simulation will be stopped. There may be a singularity in the solution. If not, try reducing the step size (either by reducing the fixed step size or by tightening the error tolerances"

or some errors assosiated to infinite or NaN values.

The question is, what is the best way to make a big model in matlab-simulink?

Cheers,

Carl

Answer 1

[John D'Errico]

Your question (structuring a big model) has nothing at all to do with the error you describe, or the other problems you hint at. In fact there are multiple possible things you may have tripped over, based on the few things you have hinted.

The errors come about because you are not cognizant of numerical issues in what you are doing. Big model or small, numerical problems exist. You will need to learn to deal with them. That may require taking classes on things like numerical analysis, numerical linear algebra, differential equations, etc.

And I'm sorry, but there is NO single simple best way to build any big model, using ANY tool.

Would you ask a carpenter about the "best way" to build a house? A BIG house? You might ask them that, and if they ever stop laughing...

The "best way" to build any model is to learn the capabilities of the tools you will use. Learn to use those tools. Learn as much as you can about the mathematics involved. Understand the system/process you will try to model. The better is your understanding of the physical system, the better will be your model of that system. Learn to formulate the problem in terms of mathematics, then learn to translate that knowledge into MATLAB.

But sorry, there is no single best way. The best way is simply that which works, providing you with valid predictions that accurately emulate the physical system you will try to model.