Bug in RWS-HWS 2024.1.0 model

[d2hydro]

Running the Ribasim HWS 2024.1.0 model (https://deltares.thegood.cloud/f/61936) in the latest CLI , results in the following Error:

@visr, can you see what's wrong?

[visr]

Hmm this error happens after the simulation when writing results. Could you first add initial Basin levels in the 'Basin / state' table? That will be required starting from the next release. Then I can debug on the main branch.

[d2hydro]

Hmm this error happens after the simulation when writing results. Could you first add initial Basin levels in the 'Basin / state' table? That will be required starting from the next release. Then I can debug on the main branch.

I,ve added states to the model following this example: https://deltares.github.io/Ribasim/python/examples.html. 0m, but more than 1m above minimum profile level:

state = model.basin.profile.df.groupby("node_id").min()["level"].reset_index()
state.loc[:, ["level"]] = state["level"].apply(lambda x: max(x + 1, 0))
model.basin.state.df = state

Model, with still the same issue, uploaded here: https://deltares.thegood.cloud/f/61973

[visr]

Thanks. The issue is there are many duplicate edges in the model, all these are the same edge:

I'll make an issue to enforce better validation for this, but it doesn't seem like a bug to me.

[d2hydro]

Thanks. The issue is there are many duplicate edges in the model, all these are the same edge:

I'll make an issue to enforce better validation for this, but it doesn't seem like a bug to me.

That's really odd. I suspect this happened while removing Nijkerkersluis using the QGIS plugin; the Ribasim-NL modelbuilder writes 260/258 width/without Nijkerkersluis. Maybe something to reproduce later.

Revised model is passed the previous error, but hanging on probably on small timesteps: https://deltares.thegood.cloud/f/61998. Any way to debug this, change settings and/or network?

[visr]

For debugging convergence issues we still need better support: https://github.com/Deltares/Ribasim/issues/452

Though often selecting an explicit solver with fixed small timesteps tends to help.

[solver]
algorithm = "RK4"
adaptive = false
dt = 10.0

[d2hydro]

Check, that works (26mins for 1 year period). I've uploaded a working model with revision 2024.1.3 (https://deltares.thegood.cloud/f/62026) and close this issue.