We use cookies to make your experience of our website better. You can delete or block cookies, but some parts of this website won't work without them. By using this website you accept our use of cookies.

adambroadhead

01/02/2011 13:39:11

I am trying to build a simple model of a city stream, and am getting a fatal flow error which I don't know how to fix. The stream flows under a bridge, reappears, and then goes into a culvert for a while. I have attached spills to connect flow bypassing the bridge and culvert. Error E1514 suggests I don't have enough river sections between the inlet/outlet, yet it seems I have this. Is somebody able to see where I have gone wrong please? I think it is something simple, but the schematic looks right. I have attached the diagnostic file plus a few others. ### Not letting reply to my own post with the .DAT, sorry :S

tmashby

01/02/2011 13:43:49

Do you have a junction downstream of the culvert outlet? If so then you will need to specify a remote downstream node at the culvert outlet. Typically that would be the section downstream of the junction.

adambroadhead

01/02/2011 13:49:42

Yes, I do have a downstream Junction. I've attached the .DAT - not sure how to use the remote node as you suggest, but this would seem to be a sensible solution.

tmashby

01/02/2011 13:53:16

If you open up the node labelling bit for the culvert outlet you will see 4 fields: Upstream label Downstream label Upstream Remote label Downstream Remote label The first two will be connected as usual to the upstream conduit section and to the junction. The upstream remote label can be left blank as it is connected directly to the conduit section. However, because you have the junction connected to the downstream label, you should put the label for the river section downstream of the junction into the Downstream Remote label field. In your case, Downstream Remote should be Port3_ds. I believe this is so that the culvert outlet has a downstream velocity to use for calculations. Hope that is clear enough.

adambroadhead

01/02/2011 15:33:29

Still struggling I'm afraid. Remote connections now make sense, the reason you gave about needing a velocity term for the calcs. I've added remote connections in for all the applicable ones, as far as I understand. There is still something wrong with it - some instability that I am not understanding. I variously get errors E1017, E1100, E1066. Is there something obviously wrong with the model set up, or is it something deeper that I am not understanding? I've attached the latest .DAT if anyone is able to provide some advice. Many thanks.

tomgent

01/02/2011 15:55:47

Hi Adam, Had a quick look at your .DAT file. Looked at improving the initial conditions. Try using the attached file. I managed to get it to run OK with these. Cheers Tom

tmashby

01/02/2011 15:58:42

Was just about to suggest the same thing. I'd probably remove the structures, build up some initial conditions without them in, then put them back in again.

adambroadhead

02/02/2011 09:50:07

Thank you both, this works well now, and with a smaller timestep I can achieve an acceptable mass balance with no nasty oscillations. For reference, I was wondering whether you might be able to briefly explain what you did to "improve" the initial conditions? I'd really appreciate this. Thanks Adam

tomgent

02/02/2011 11:19:25

Hi Adam, When I downloaded the .DAT file there were no initial conditions associated with the file. I had attempted to run the model in steady state to determine initial conditions, but with no success. You can edit the initial conditions manually in the bottom half of the ISIS interface. A quick look at the reach long section can provide a rough idea of what should be sensible. In this case i initially only looked at the stage values, and took a blanket value of 60mAD for all nodes. Re-running the model in unsteady then worked. It may also have been useful to use an initial flow value of 6m3/s for all nodes. To further improve on the initial conditions you can extract values from any results files you have using Tools, Tabular CSV. Tom

tmashby

02/02/2011 11:22:18

My approach was to save a copy of the model and remove the structures which allowed it to run in steady state creating its own initial conditions. The imported the structures back in and ran using the initial conditions created without the structures.

Sign in to create or reply to forum posts.

Privacy & Terms

Copyright