OpenSWMM · Laboratory comparison · 21 October 2026
One opening. Two directions of flow.
Compare the published steady experiments with reconstructed model results, then follow a synthetic surcharge event across the 8 × 4 m flume.
Experimental qualification is incomplete. The steady matrix uses a 0.5 s routing step with unresolved sensitivity. The animation uses synthetic forcing and contains no measured transient series.
01 · Measured comparison
How close are the exchange rates?
Every point pairs one published Table 2 observation with a calculated discharge magnitude. The dashed line is exact agreement. Horizontal bars show ξf, the paper’s instrument/balance variation.
● Physical shaft△ Junction× Paper equation
measured rows
shaft MAE · L/s
within illustrative band
Residual = shaft − experiment. The comparison band is ±(ξf + 2% of measured flow); it is not a confidence interval. Across all regimes only 8/36 shaft rates meet it. These are calibration observations, not an independent validation dataset.
02 · Synthetic T9 demonstration
Watch surcharge become drainback
Actual saved cell depths, a fixed 0–25 mm colour scale, and the native coupling hydrograph. Each frame is a saved result; fields are not interpolated. The orange ring marks the 0.24 m manhole.
to surface · L/s
gross spill so far · m³
gross drain so far · m³
Positive exchange moves water from sewer to surface; negative exchange drains it back. The hydrograph cursor tracks the displayed depth frame. Surface inflow continues throughout the run, so remaining surface water is not all surcharge water.
Final gross spill: 1.415 m³ · drainback: 0.201 m³ · net to surface: 1.214 m³.
■ Sewer → surface■ Surface → sewer
60 saved frames at approximately 10–600 s. All-step peak cell depth is 24.803 mm; the largest depth in the saved frames is 20.909 mm. Brief between-frame peaks are not shown. The Q readout is the nearest native routing-step sample; cumulative values use the same sample.
Data and provenance
Measurements: Rubinato et al. (2017), Journal of Hydrology 552, 421–432, Table 2 (CC BY 4.0). Model results: the archived OpenSWMM CPU comparison of 21 October 2026.
Recorded sources and binary identifiers
The source checkout was dirty. Binary and input hashes identify these results; they do not certify later builds. This offline viewer has no external libraries, analytics or network data requests.