A large number of event iterations is not necessarily an error

[Francesco Casella]

Consider the model ScalableTestSuite.Electrical.BreakerCircuits.ScaledExperiments.BreakerNetwork_N_320. This model describes a cascade of events where the opening of a circuit breaker triggers the opening of the next one, iteratively, until all breakers are open. This is a very common situation in algebraic network models, e.g. when simulating voltage collapse scenarios. If the network is large, there can be very many such events.

Unfortunately, after 200 iterations the runtime gives up with the following error message:

ERROR: Too many event iterations. System is inconsistent. Simulation terminate.

Process crashed.

First of all, the process shouldn't crash, but rather exit gracefully :)

Besides that, this system should go though 320 event iterations, which are perfectly legitimate and intended in this case.

I understand that in most cases 200 iterations means that the process is never going to converge. In fact, maybe this is even a too large default (Dymola gives up after just 20). However, there might be useful cases where this limit has to be increased substantially.

Please add a flag to do so.

[anonymous]

furthermore the English: Simulation terminated or Simulation terminates, not simulation terminate.

[Lennart Ochel]

This will be improved with ​PR572.

[Martin Sjölund]

Milestone deleted