New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Questions regarding FluxTubes, HysteresisModelComparison #2248

Open
chria opened this Issue May 2, 2017 · 3 comments

Comments

Projects
None yet
5 participants
@chria
Contributor

chria commented May 2, 2017

Hi,

Simulating the model Modelica.Magnetic.FluxTubes.Examples.Hysteresis.HysteresisModelComparison using JModelica result in that the variables:

TellinenTable.Phi
TellinenSoft.Phi
PreisachEverett.Phi

are chosen as states. These are poor choices for states as their magnitude are around 1e-4 and because of that they are missing nominal values. Would it be possible to include nominal values for these?

A better choice, in this model, for states would be

TellinenTable.B
TellinenSoft.B
PreisachEverett.B

is this true in general? If so, would it be possible to set state select on these?

Thanks!

Best
/Christian

@beutlich

This comment has been minimized.

Show comment
Hide comment
@beutlich

beutlich Jun 2, 2017

Member

@ThomasBoedrich Did you have a look at this?

Member

beutlich commented Jun 2, 2017

@ThomasBoedrich Did you have a look at this?

@ThomasBoedrich

This comment has been minimized.

Show comment
Hide comment
@ThomasBoedrich

ThomasBoedrich Jun 6, 2017

We did have a look at this ticket, but unfortunately at the moment we don't have capacities to implement and test other state variables. However, selecting flux density instead of magnetic flux as state variable should be possible in principle.
During model development we've experimented with different state variables and start values, each of them resulting in certain benefits and drawbacks (e. g. numerical problems). The present hysteresis models have been sucessfully tested with different simulation environments prior to inclusion into the Modelica Standard Library, including OpenModelica. However, they have not been tested with JModelica at that time. During the tests we did not encounter problems with small values of the automatically chosen states or with missing start values.

ThomasBoedrich commented Jun 6, 2017

We did have a look at this ticket, but unfortunately at the moment we don't have capacities to implement and test other state variables. However, selecting flux density instead of magnetic flux as state variable should be possible in principle.
During model development we've experimented with different state variables and start values, each of them resulting in certain benefits and drawbacks (e. g. numerical problems). The present hysteresis models have been sucessfully tested with different simulation environments prior to inclusion into the Modelica Standard Library, including OpenModelica. However, they have not been tested with JModelica at that time. During the tests we did not encounter problems with small values of the automatically chosen states or with missing start values.

@hubertus65

This comment has been minimized.

Show comment
Hide comment
@hubertus65

hubertus65 May 24, 2018

Contributor

From a numerical point of view, it is very obvious that the flux densities are better choices of states in the library example and will lead to more accurate results in any simulation environment.

In addition, all of the 3 hystereses models are actually written as states by the modeler (with a der() on the flux density B ), so it would basically be a confirmation of what I see as the intent of the modeler to add stateSelect.prefer. Phi as a variable is structurally equivalent, scaled by a constant, but a worse scale basically always. Alo Dymola gets faster with B as a preferred state in all 3 models.

Contributor

hubertus65 commented May 24, 2018

From a numerical point of view, it is very obvious that the flux densities are better choices of states in the library example and will lead to more accurate results in any simulation environment.

In addition, all of the 3 hystereses models are actually written as states by the modeler (with a der() on the flux density B ), so it would basically be a confirmation of what I see as the intent of the modeler to add stateSelect.prefer. Phi as a variable is structurally equivalent, scaled by a constant, but a worse scale basically always. Alo Dymola gets faster with B as a preferred state in all 3 models.

@hubertus65 hubertus65 added enhancement and removed question labels May 24, 2018

@beutlich beutlich changed the title from Questions regarding Fluxtubes, HysteresisModelComparison to Questions regarding FluxTubes, HysteresisModelComparison May 30, 2018

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment