Simulation wrongly reported as successful.

[massimo ceraolo]

The simulation of the enclosed file InvSVPWM.mo wrongly reports to be successful. Indeed all plot variables seem to be empty.
Note that InvSVPWM.mo requires loading power Systems Library.

[Martin Sjölund]

I get:

marsj@marsj-Precision-5510:~/tmp$ ./InvSVPWM 
LOG_SUCCESS       | info    | The initialization finished successfully without homotopy method.
LOG_SUCCESS       | info    | The simulation finished successfully.
marsj@marsj-Precision-5510:~/tmp$ echo $?
0
marsj@marsj-Precision-5510:~/tmp$ ls -lh InvSVPWM_res.mat 
-rw-rw-r-- 1 marsj marsj 37M maj 22 07:50 InvSVPWM_res.mat

Did you run the simulation with any special flags?

[massimo ceraolo]

My mistake, sorry: final time was inadvertently set to 0.

I'm going to close this.
However, maybe it is not a bad idea to emit a warning when final time equals initial time.

[Francesco Casella]

Replying to ceraolo:

My mistake, sorry: final time was inadvertently set to 0:

experiment(StopTime = 0, Interval = 2e-05),

OpenModelica_commandLineOptions = "");

I'm going to close this.
However, maybe it is not a bad idea to emit a warning when final time equals initial time.

This could be useful.

Or maybe in cases like this, in which Interval is larger than (StopTime-StartTime) it is more appropriate to generate an error?

I don't think so. If one has issues with initialization, a typical strategy is to turn on all LOG_NLS_V and at the same time set StopTime=0, so that when the initialization problem is finally fixed, the simulation result window is not clogged by a ton of (irrelevant) debug info during simulation, which may actuall hang OMEdit for good. This situation should not lead to an error.

[Adeel Asghar]

Replying to casella:

Replying to ceraolo:

My mistake, sorry: final time was inadvertently set to 0:

experiment(StopTime = 0, Interval = 2e-05),

OpenModelica_commandLineOptions = "");

I'm going to close this.
However, maybe it is not a bad idea to emit a warning when final time equals initial time.

This could be useful.

OMEdit already have this check,

if (starttime > stoptime) then error;

I will update it to,

if (starttime >= stoptime) then error;

Or maybe in cases like this, in which Interval is larger than (StopTime-StartTime) it is more appropriate to generate an error?

I don't think so. If one has issues with initialization, a typical strategy is to turn on all LOG_NLS_V and at the same time set StopTime=0, so that when the initialization problem is finally fixed, the simulation result window is not clogged by a ton of (irrelevant) debug info during simulation, which may actuall hang OMEdit for good. This situation should not lead to an error.

[Martin Sjölund]

No, OMEdit should not make starttime == stoptime an error. Better would be to make the simulation output something like:

LOG_SUCCESS       | info    | Skipped time integration since starttime == stoptime

Instead of:

LOG_SUCCESS       | info    | The simulation finished successfully.

[Lennart Ochel]

Replying to sjoelund.se:

No, OMEdit should not make starttime == stoptime an error. Better would be to make the simulation output something like: [...]

+1