 |
OpenSWMM Engine
6.0.0-alpha.4
Data-oriented, plugin-extensible SWMM Engine (6.0.0-alpha.4)
|
|
OpenSWMM Engine
6.0.0-alpha.4
Data-oriented, plugin-extensible SWMM Engine (6.0.0-alpha.4)
|
Phase A3 — water age on subcatchment surfaces. More...
Go to the source code of this file.
Namespaces | |
| namespace | openswmm |
| namespace | openswmm::runoff |
| namespace | openswmm::transport |
Functions | |
| void | openswmm::transport::routeSubcatchmentAge (SimulationContext &ctx, const runoff::RunoffSoA &soa, double dt) |
| void | openswmm::transport::addRunonAge (SimulationContext &ctx, int donor_sc, int receiver_sc, double q) |
Phase A3 — water age on subcatchment surfaces.
Each subcatchment carries THREE ages, one per ponded subarea (IMPERV0, IMPERV1, PERV), mirroring the RunoffSolver's own depth_imperv0/1/perv. Per-subarea rather than lumped is a user decision (2026-08-17): impervious water is systematically younger than pervious, and the depths are already separate, so lumping would discard information the state already has.
Per step, per subarea: age +dt, then mix with what arrived, volume-weighted — the same order the node/link mirror uses. Water leaving as runoff carries the volume-weighted mean of the contributing subareas, and THAT is what reaches the outlet node and what run-on hands to a downstream subcatchment.
v_new == v_old, which is the ordinary state of an impervious surface during a storm. A mixing volume taken as the NET gain, max(0, v_new − v_old), is therefore zero exactly when rain is pouring through the surface, and the age stops being a residence time and becomes the elapsed time since the surface first wetted. Measured on a 100 % impervious, zero-depression deck under sustained 2 in/h rain (V = 5104.5 ft³, Q = 10.08 cfs, so V/Q = 0.14062 h): a net-gain mixing volume reports 0.88592 h, 6.3x too old.Only the INFLOW is needed to avoid this. Complete-mix means outflow leaves at the subarea's own age, so the outflow never enters the update — and the inflow the solver applied is already published: ctx.subcatches.rainfall (ft/s, written by Runoff.cpp:295) plus run-on, which the solver spreads over the whole area as extra precipitation (Runoff.cpp:331-333). So v_in = (rain + runon/area) · frac · area · dt, and the same deck returns 0.14022 h against the analytic 0.14062 h — 0.3 %.
snow_net_imperv/perv for rainfall per subarea (Runoff.cpp:543-548), and inter-subarea routing (RouteTo IMPERV/PERV) moves water between subareas without appearing in either term. Both leave the arriving volume mis-stated; neither is exercised by a RouteTo OUTLET, snow-free deck. Runoff also leaves at the stored-volume weighted mean of the subareas rather than at their outflow-weighted mean, because per-subarea outflow is genuinely not published — a depression-storage subarea therefore counts toward the departing age in proportion to what it holds, not what it sheds.
1.13.2