Transformation losses collector with unbalanced loads

  • Nuria Cuartero

    Nuria Cuartero - 2014-04-04

    Hello GridLAB-D team,

    I am working with GridLAB-D 3.0. I am simulating a grid with some consumers and generators connected only to one phase and I have obtained unexpected results in the transformation losses during the simulations.

    When loads are balanced in the three phases everything seems to be OK. However, when the flow of current in the three phases is different, the transformation losses recorded by the collector object are too high.

    I attach an example which shows the problem. The network has a transformer and two load objects, one of them connected to three phases and the another one connected only to one phase. At the beginning of the simulation, only the first load is consuming. But at 03:00 am the second load starts to consume reading a player file, and an important increase of transformation losses can be observed in the Trans_Losses.csv collector file. What is more, the increase of losses matches the consumption of the load.

    Could you tell me if the results obtained can make sense or if the collector is not calculating the losses properly?

    Thank you very much for your help.

    Last edit: Nuria Cuartero 2014-04-04
  • Frank Tuffner

    Frank Tuffner - 2014-04-04

    Hello Nuria,
    The results are being thrown off due to the Delta-Ground-Wye (Delta-Star) transformer on the system. When it computes the power_in and power_out values, there is a small translation issue with how it compares them for losses (since one is delta connected and one is wye connected). I thought this had been fixed, but obviously not. We'll look into it further, but this is a bug with how the losses are being computed on that particular transformer type. The powerflow solution beyond that is valid, but we clearly have a sign error or similar problem with the delta-gwye transformer loss calculations.

    Unfortunately, the easiest way to compute losses in this scenario would be to put meters on both sides of the transformer and manually take the difference in their measured_power values. I'll try to post back here once we have it fixed and where it is posted.


  • Jason Fuller

    Jason Fuller - 2014-04-04


    Sorry for your inconvenience. This is a known bug that was discovered shortly after the release of v3.0:

    We are planning to have it corrected for version 3.1 release this summer.

    NOTE: the power loss calculations are "secondary" calculations, as in they do not occur until after all of the powerflow voltage and current solutions are determined, so it does not invalidate the base line solution. It is just that the power losses through that style of transformer (particularly DELTA connected transformers) are not properly calculated from the determined voltages and currents.

    One less satisfying way of seeing this is to change node1 and add a recorder to it, similar to what is being used in the swing meter. Subtract the difference, and you'll see more appropriate loss numbers. My quick run shows 0.06 to 0.24 W losses.

  • Jason Fuller

    Jason Fuller - 2014-04-04

    Hmm, sorry about the double post. Seems Frank and I should communicate before we respond!!

  • Nuria Cuartero

    Nuria Cuartero - 2014-04-07

    Thank you very much Frank and Jason for your responses. I will take into account your advices and will follow the ticket status.

    Best regards


Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:

No, thanks