Wrong handling of two-dimensional arrays inside arrays of models

[casella]

Consider the attached test case. The model TestArrays.TubeArray instantiates an array of Tube models. Each Tube models contains a Flow1DFV model, which in turn contains a two dimensional array dddX.

The following error is generated:

Error: Model is structurally singular, error found sorting equations 

If dumpSimCode is switched on, the following variables show up:

56: tube[1].flow1DFV1.dddX[1,1,1] = 0.0;
60: tube[1].flow1DFV1.dddX[1,2,1] = 0.0;
64: tube[1].flow1DFV1.dddX[1,3,1] = 0.0;
68: tube[1].flow1DFV1.dddX[1,4,1] = 0.0;
130: tube[2].flow1DFV1.dddX[2,1,1] = 0.0;
134: tube[2].flow1DFV1.dddX[2,2,1] = 0.0;
138: tube[2].flow1DFV1.dddX[2,3,1] = 0.0;
142: tube[2].flow1DFV1.dddX[2,4,1] = 0.0;

The array dddX should be two-dimensional, not three-dimensional, and these equations should instead look like this:

56: tube[1].flow1DFV1.dddX[1,1] = 0.0;
60: tube[1].flow1DFV1.dddX[2,1] = 0.0;
64: tube[1].flow1DFV1.dddX[3,1] = 0.0;
68: tube[1].flow1DFV1.dddX[4,1] = 0.0;
130: tube[2].flow1DFV1.dddX[1,1] = 0.0;
134: tube[2].flow1DFV1.dddX[2,1] = 0.0;
138: tube[2].flow1DFV1.dddX[3,1] = 0.0;
142: tube[2].flow1DFV1.dddX[4,1] = 0.0;

I guess this gives a hint why the matching fails.

[lochel]

This seems to be an issue in pre-optimization module inlineArrayEqn. Unfortunately, it cannot just get disabled, since other modules depend on this one.

[lochel]

I think I've a ​fix for this issue. However, I will ask for feedback before merging it.

[lochel]

This is fixed. However, the model is still not working due to #3305.

[casella]

Please close this once the problem is actually solved.

[lochel]

Replying to casella:

Please close this once the problem is actually solved.

Well, actually this issue is already solved. Also #3305 is now solved. However, the attached test case does still not simulate:

Simulation execution failed for model: TestArrays.TubeArray
assert            | debug   | Solving non-linear system 1370 failed at time=0.
|                 | |       | For more information please use -lv LOG_NLS.
assert            | info    | simulation terminated by an assertion at initialization

Francesco, can you have a look for the new failure?

[casella]

This is a different story, we can take care of that. Thanks for the quick fix!

[sjoelund.se]

Milestone renamed

[sjoelund.se]

Milestone renamed