Time Weight > 0 results in tank levels LOWER than MINIMUM tank levels

[mariocastrogama]

Hello all,
I have been checking the use of the TIME_WEIGHT option.
Based on the example by [2], it seems to work .
Now when the example of [2] is modified in [1], after simulation, the minimum Level (in the tanks) is lower than the elevation of nodes contiguous to the tank. This means that tanks continuously empty and show pressures of less than 0.

This has been evidenced in [1] where this problem is named mLBC or minimum Level Boundary Condition. (There is also a MLBC, corresponding to Maximum Level #34 ). The solution proposed in [1] does not seem easy to implement (see below).

Specifically, because if there are multiple pipes connecting two tanks.

I'm attaching two INP files of models studied in [1] and [2]. In this example networks, the 2 tanks must empty until reaching the minimum level as in [1] (not yet implemented in epanet-dev), and the case where TIME_WEIGHT works with the example of [2]. (tank_instabilities.zip)

Best regards.

[1] Avesani, D.; Righetti, M.; Righetti, D. & Bertola, P. The extension of EPANET source code to simulate unsteady flow in water distribution networks with variable head tanks. Journal of Hydroinformatics, , 2012, 14, 960
[2] Todini, E. Extending the global gradient algorithm to unsteady flow extended period simulations of water distribution systems Journal of Hydroinformatics, 2010, 13, 167-180