Unsteady flow Simulation error

[Jack smith ASA]

when I run an unsteady flow model with 8 pumps , I keep getting the following error messages in RS near pump stations:

I have tried to increase the number of iterations,decrease computation interval, turn on mixed flow regime,but that doesn’t help.
Any help would be appreciated!

[cameron]

Those are common error messages in Unsteady HEC-RAS which just states it couldn’t converge to a solution within the iterations so it picked the best value and moved on. Depending on what type of study you are doing, the error amount may or may not matter. For instance, with dam breach studies I try to keep the error value below 1.0 and for other studies below 0.5 or less depending on how complex the model is.

In your case the two largest errors occur right at the beginning of the simulation and so the error would have something to do with your initial conditions. Not sure the purpose of the study, but I would probably not worry about the error as it is not that large and goes away eventually.

[Jack smith ASA]

Thank you so much for your suggestion. I am doing a pumping station optimization project in a river network, so water surface elevation needs to satisfy a certain accuracy. The errors in the picture is just a small part of errors in the simulation and the error grows as the pump station scale increases which can reach as much as several meters. In this example , the 8 pump stations are set to be instantaneous opening at the water surface elevation of 1.0 m and closing at 0.4 m. Since the errors happened in RS near pump station, the error may be induced by the instantaneous opening and closing of pump stations. I have tried to increase the number of iterations,decrease computation interval, turn on mixed flow regime,but that doesn’t help. Is there any other way to solve it? Looking forward for your reply, thank you.

[cameron]

could you do a gradual open and gradual close of the pumps? maybe only turn on a couple at 1.0m and a couple more at 1.1m and so forth? Not sure how big the pumps are, but 8 starting or stopping at the same time might be causing a shock to the models.

[Jack smith ASA]

Yes,I have tried to close and open the pumps in 1 minute or 2 minute ,but error still happened. May be that is because the scale of pump is too large for the river network , every pump is set 32 m3/s to make the max water surface meet demand. How can I decrease the error in such a pump flow?

[cameron]

Based on the plot you provided, it appears that turning on all 8 at the same time is overkill until about hour 8. It also appears that your starting stage is above 0.4m which is when the pumps turn off. Have you tried a warm up period or a run where the pumps do not start out on? This is possibly why you are getting an error at the beginning of your run. Also it is probably not good for the pumps to be turning on, off, and back on again so quickly. That just tells me that you have too many pumps turned on or you have the wrong size of pumps.

When do the largest errors occur? Does it occur during the highest stages at the pump when all the pumps are running full (hour 8 to 14) or before and after? If it is before or after, it is not going to impact the max wse plot (if that is what you care about). If it is during the peak stage, that may impact the max wse which depending on the error may be ok.

It is very difficult to get a complex unsteady model that has no error and is in some cases impossible.

[Author]

Posts