HEC-RAS modeling excessively long culverts

[jleiben]

Has anyone had any experience with modeling a structure on the order of a triple 7' x 7' box culvert in HEC-RAS? I seem to have run across an issue with the program when trying to model a proposed 2,400'+ long version of the aforementioned structure for a project. We were tasked with needing to show the backwater effects for blocking one of the three, as crazy as that seems, and HEC-RAS was giving very unstable results, including a lower WSEL for the 100-yr than the 25-yr. It was defaulting to the Critical WSEL since it wasn't able to calculate a WSEL for the given circumstances; however, some troubleshooting revealed that upon shortening the culvert to approximately 1500' it gave reasonably accurate results. Anyone ever seen this issue previously?

Thanks.

 

[pinkpig]

is the model under steady state or unsteady?

 

[jleiben]

Steady State. Originally we didn't have the time to develop an unsteady flow model, but if someone believes this might alleviate the issue I'm willing to develop it.

 

[gbam]

Quick question is the long box sealed or free surface during the blockage and regular condition? if unsealed you used to be able to model a lidded structure in HECRAS, I haven't done that in years, however. Can you perform a calculation externally using a storm cad type model to develop a rating curve at the inlet to compare your results?