Results 1 to 3 of 3

Thread: Modeling Two Tanks

  1. #1

    Join Date
    May 2014

    Modeling Two Tanks

    Is there any recommended best practice for modeling two water tanks that are next to each other? They should basically operate as one tank with their levels rising and falling together. However my simulation always shows crazy oscillations in level. I tried adding a short large dummy pipe between them but that did little to help. Anything else to try? Or is it recommended to model them as a single tank?


  2. #2
    Forum Moderator

    Innovyze Employee

    Innovyze Employee

    Join Date
    May 2015

    This has been a known issue in EAPNET since it's initial creation. While there is some newer code that will hopefully address this, that code has not been formally added to the latest "official" version of EPANET used by our software, so the issue persists.

    What happens is that small differences in water flow into each tank cause a head differences between the two tanks, even if small. To mathematically account for this a flow that will cause that much headloss must occur between the tanks is needed so that the heads match. Since they are near to one another and usually have large pipes connecting them, this often requires a large flow to induce the headloss. At the next iteration, the situation is now reversed and the water flows back. The result is large oscillations in the tank levels of both tanks due to the water sloshing back and forth between them (I think it makes tank levels look like they are having a heart attack). Mathematically this is often unavoidable for two tanks very close to one another in a hydraulic model.

    The easiest solution by far to this is to combine the tanks into an single equivalent tank in the model. This will act hydraulically just like two tanks, but will not have the potential for oscillation.

    Recall Area of a tank (ft^2) = D_ft^2 * pi/4 If you sum the area of each tank, you can solve for the equivalent diameter needed to model both tanks as a single tank.

    If the tanks are of variable are or non-identical shape use a variable area tank to represent both tanks (curve is cumulative volume (ft^3 per ft of tank height) in a curve applied to the tank.

    This is by far, what most modelers find resolves the issue easily and what most modelers do to account for it.

    Without an official change to the EPANET code, this is currently the best and easiest way to address this within our software.

    Once an equivalent tank is used, there will be no oscillations due to the water flow between the two or more tanks.

    Patrick Moore

  3. #3

    Join Date
    May 2014
    I figured this was some hydraulic limitation to be solved by combining the tanks, but your detailed explanation is very much appreciated! Now I know what's actually happening, very interesting. Thanks for the reply, Patrick.

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts