I am using Bentley SewerGEMS V8i to model two ponds emptying out into an outfall structure. There is piping coming from each pond and then each pipe joins together to a discharge pipe that gradually reaches an outfall. The pond takes over 200 hours to get to 90% empty and will never truly empty, no matter how long I run the simulation. I will securely upload my model as well.
Everything appears to be working as designed. If you want the ponds to drain quicker, you will need to make some adjustments to the model.
Take a look at the conduit sizes.
Before the branches join, and a little after, there are circular conduits with 36 inch diameters. After the branches join, just before the outfall, there are ellipse conduits with 36inch rise and 30inch span, which creates a bottleneck.After the branches join, typically larger conduits would be used to accommodate the increased/combined flow. If you increase the size of the conduits after the branches join, then the pond will drain quicker.
Also note that the ponds don't have any infiltration or evaporation applied. When the depth of the ponds are near 0ft the flow will become very small. If you apply infiltration and/or evaporation the remaining volume will drop quicker.If for some reason you don't want to model evaporation and/or infiltration, you might need to make a judgement call to determine the depth/volume that you can consider negligible. Otherwise, it will take a long time to drain the last bit of volume from the pond.
Keep in mind that the Explicit solver uses Manning's equation. When the pond depth is near 0ft, the low slopes and small flow areas will result in low flows leaving the ponds.
Bentley Technical Support
Answer Verified By: Sushma Choure
I forgot to mention that when I computed the model there was 14.8% "not converging".
As recommended in the article below, I first took a look at the user notifications.
Troubleshooting unstable SewerGEMS and CivilStorm model results using the Explicit SWMM Solver
CO-4: "The crown of the link adjacent to the Stop Node is above the top of the node."CO-9: "The crown of the link adjacent to the Start Node is above the top of the node."To correct these, I raised the "Elevation (Top)" value for transition "T-6" from 185.21ft to 186.21ft (to match the adjacent ellipse conduits that have a 3ft rise).After computing again, the model computed, but there was still 14.8% not converging.
Moving to the next step in the article, I reduced the routing time step from 1sec to 0.1 sec. It took a bit longer to compute, but resulted in 0.1% not converging.