How can I model a flow restriction at an outfall, to limit the maximum flow that the outfall can receive? For example this may represent a pump station or treatment plant that can only receive a certain maximum flow. Flow above that will cause a backup and will be stored in pipes and potentially overflow upstream manholes.
One possible approach to modeling a flow restriction is to use a Start Control Structure on the last conduit before the outfall, with the "Depth-Flow Curve" Structure Type. Configure the table of Depth and flow to model a fixed flow at all depths. For stability purposes, include a zero flow at zero head entry for the first row. For the second row, use a small depth, with the flow equal to the desired flow restriction. Add a few additional rows to cover all possible depths, with a very slight increase in the associated flow rate. For example 3801 L/s, 3802 L/2, 3803 L/s, and so on. Ensure that the "Crest Elevation" in the conduit control structure is set equal to the invert of the flow restriction, and ensure that you select this conduit control structure in the properties of the conduit.
Note that this approach will only work with the Implicit and Explicit numerical solvers. The Explicit (SWMM) solver is recommended in this case for best stability, with the following advanced calculation options set, if the model experiences significant surcharging and/or flooding as a result of the backup from the restriction:
Use Bentley transition equation = TrueMinimum Surface Area - If the default of zero does not work well, try 30 square meters and adjust to a lower value if stable results can still be achieved.Routing Time Step - try 5 seconds and adjust as needed for stability. Very small routing steps such as 1 second or below may not work well in this type of situation. In some cases, a much larger routing step may work well, even up to 30 seconds+. Try a range of values to test the sensitivity. Head Convergence Tolerance - Try a smaller value, such as ten times smaller (add an extra zero such as 0.006 instead of 0.06). The "max trials per timestep" may need to be increased in order to achieve convergence - check the calculation summary and adjust as needed. If a smaller value for the head convergence tolerance improves results, try an even smaller value (add another zero).
Note that you may still experience instability at the very start of the simulation, which you may be able to ignore and focus on the results after the initially dry pipes first fill up and a more steady condition is reached.
Troubleshooting unstable SewerGEMS and CivilStorm model results using the Explicit SWMM Solver