Unsteady flow simulation 2d error

[sam1]

I have a simple application of a flow area having Downstream BC and Upstream BC as flow hydrograph.
The computation is completed for a 1 days time step but with big errors and most of cells not converging.
Im trying to simulate for a 1h time step but the computation is not completed de to the following error:

Unsteady flow encountered an error.

HDF5-DIAG: Error detected in HDF5 (1.8.11) thread 0:
#000: ..\..\src\H5Dio.c line 234 in H5Dwrite(): can’t prepare for writing data
major: Dataset

1D Post Process Skipped (simulation is all 2D)

Stored maps were not written because the simulation did not complete.

[cameron]

So both the upstream and downstream boundary conditions are flow hydrographs and add flow to the model? How does the water escape?

Also it sounds like you have an 1 hour time step which is way too high for a 2D model. You need to follow the guidance in HEC-RAS manual and look at the courant number.

[sam1]

You can add Flow hydrograph as a downstream BC with negative values, this means that the flow is flowing out of the flow area.
I dont think this error is related to the time steps, since it does not appear for a 1 days time step but only for 1h.

[Anonymous]

I think you are misunderstanding how DS boundary conditions work as well as how the time step controls a models convergence to a solution. It doesnt matter if there is only data for every hour. Like cameron said, a bit of reading will definately help you.

[sam1]

No one said that it matters if there is only data for every hour.
And second, i do understand the role of the boundary condition, but what is the problem of having the upstream BC as flow hydrograph as input to the flow area and the Downstream BC as Flow ( with negative values) as outflow of the flow area.

The HEC-RAS 5.0 Reference Manual explains the possibility to have a flow hydrograph as a Downstream or Upstream BC.

[Anonymous]

“I dont think this error is related to the time steps, since it does not appear for a 1 days time step but only for 1h.”

Your statement not mine. It makes it seem like you do not know how unsteady state dt works relative to the model input as a whole. Regardless, the software has endless “possibilities” but having correct or calibrated results is the best answer. Just saying again, if youre having instability issues or absurd results, start with that DS BC.

[Author]

Posts