This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

WaterGEMS pump switch on and off

Hi All,

I am experiencing a difficulty to fix a non-convengence model while the pumps are switching on and off. I'm using the operation data so during certain point one pump turns on while another turns off. I wonder why this causes unbalance network, and how to fix it in a general way?

Thanks,

Meng

  • More specific to my question:

    The pumps are mannual operated and I've got the pump status for each hour. Three are fixed pumps, while the other two are VSPs. I used the pattern to reflect the pump status as below.

     

         

    While I'm adjusting the pump pattern to fix the sharp curve at time=6:00 (left picture below) match the observed pressure data, the error of unbalanced network occurs even if I'm just changing from 1 to 0.95 at time=6:00 for pump #2, the model became unbalanced (right picture below).

        

    Hope anyone may help.

    Thanks,

    Meng

  • Meng, is this the same (duplicate) issue as the other post here? - WaterGEMS Connection Run Fails

    If not, I suspect the problem is likely similar to the one that was discussed this past December: Weird Simulatoin Results with Pump SCADA Data . See the answer provided in that thread regarding the timing of the patterns.

    If this does not help, please provide a copy of the model for review: Sharing Hydraulic Model Files on the OpenFlows Forum


    Regards,

    Jesse Dringoli
    Technical Support Manager, OpenFlows
    Bentley Communities Site Administrator
    Bentley Systems, Inc.

  • Hi Jesse,

    Yes that's a similiar question we've discussed month ago. As I'm changing different pump status and fixing the sparks for pressure, the same scenario happens. If I changed the pattern value for just a little, the network got unbalanced. I'm guessing that could still due to the pump switching, but after several trials still got no luck. 

    What I'm trying to do is to match the observed data (there is one pressure graph set up) and at time=6:00 and 21:00, not able to make it. 

    More important, it will be very helpful if I may understand what caused the problem and how to avoid that kind of unsteability in a general way. So I won't keep brothering you the next time ;) 

    I've shared the copy through the provided link. Thanks for the help!

    Best,

    Meng

  • Ok. I would refer to the previous explanations provided - unstable conditions can be caused when many things are changing drastically at the same time. 

    Have you considered modeling the pump operation with time-based controls and/or HGL/flow-controlled VSPs instead? This may prove to be less error prone.


    Regards,

    Jesse Dringoli
    Technical Support Manager, OpenFlows
    Bentley Communities Site Administrator
    Bentley Systems, Inc.

  • Thanks for the suggestions, Jesse. 

    I've considered the time-based controls before, but since I've got five pumps plus two of which are VSPs, it probably could need a hundred rules to reflect the pumps status. 

    For the HGL control, the two VSPs are not the same pump curve, so cannot use the same control point. 

    I will try something with the rules, and your insights are always welcome. 

    Best,

    Meng