Load summation during reversed flows

[joelnc]

Related to #71 (comment)

In reviewing simulated SCM removal in a larger model with ~90 SCMs, I noted some discrepancies in pollutant removal when taken from the report file (i.e., aggregate loads in the links US and DS of treatment node) vs. when compared to the same values manually calculated from simulated flow and pollutant time series in those links. Digging in, flow reversals appear to be the source of the discrepancies.

ReverseFlow_TestModel.txt

E.g., looking at link l1850 in the attached stripped down model, run at a 1 sec time step, the Link Pollutant Load for link l1850 is 12.223 lbs. For this 3 day sim, there is flow reversal.

• If I sum the load from the full time series, subtracting the load when negative, adding when positive, I get 9.58 lbs.
• If I sum the load from only the positive flows I get 10.768 lbs
• If I sum the loads but using absolute value of all flows, I get 11.96 lbs, closest to the reported summary load of 12.223.

Similarly, a GUI plot of Flow and Volume for l1850 or l1626 shows link volume accumulating across negative flows, which is then used in the load aggregation/summary.

I can just work with the time series rather than summary loads now that I understand what is going on. However, as is, the Link Pollutant Load summaries seems to be over counting when flow is reversing which may not be obvious to other users.