New feature: Tolerances about gridcal HOT 12 CLOSED

This would be a nice feature. But we'd need to think the use case very well.
I'd rather have these values change with temperature, Is there a simple model we can follow?

from gridcal.

Setting default behaviors would be fairly simple for load flow and short circuit studies; not so much of other studies. I haven't checked how you handle simulation parameters in the GUI but it could be left at the nominal value by default for any other case.

These tolerances are not related to temperature, i.e. they are not actual variation of the physical values, but imprecisions on the value itself. By the relevant IEEE standards (C57.12 or one of its cousins), when a power transformer is specified, the manufacturer is allowed a 7.5% error on the impedance value (unless another error is specified). If the impedance is not measure by a routine test, the error will remain and the nameplate will show the specified impedance, not the measured (real) impedance.

So preliminary studies (i.e. before the transformer is constructed and tested) should always somehow account for this tolerance. We could add the tolerance attribute to the transformer type class, set it to 0% by default, and think of the best way to determine whether and how (i.e. -/+) it should be applied for each simulation.

It's not urgent and I plan on integrating the additional tests and the temperature resistance adjustement first, but the feature could be added later.

from gridcal.

What's the status of this one?

from gridcal.

No work done yet, so depending on how you wish to track potential new features, you may close it or tag it accordingly.

It's up to you, if you close it I'll simply reopen it when I'm ready to discuss it further.

from gridcal.

Honestly I need to understand the use case better.
I guess that the best is to create a python test script "researching" the implementation.
Do you have anything that is worth reading to understand this?

from gridcal.

Hi Santiago,

The official reference would be the IEEE C57.12 series, starting with C57.12.00. It indicates what I mentioned in my first comment: before being tested, there's typically a +/-7.5% tolerance on theoretical transformer impedance.

After that it's mostly common sense, and up to the designer's preferences. I do find some articles that make mention of it but it comes down to: a theoretical power transformer's impedance has a +/-7.5% precision. The designer can choose to ignore it or apply it to be more or less conservative.

If there are sufficient conservative factors elsewhere, the designer may choose to ignore this factor entirely, even during the preliminary stage (i.e. before the transformer is built).

For example, we are contractually bound to meet total system losses below a predetermined value for our wind farm designs, based on our preliminary calculations. In this case, I definitely want to use the most conservative value for the transformers' impedance. For short circuit calculations for equipment duties, I don't care because we already use extremely conservative contributions from the grid. Etc.

Michel

from gridcal.

So, would it suffice with adding two properties low_tolerance and high_tolerance to the branch device, and then have some option telling which one is to use?, or is there some more advanced capability required?

from gridcal.

The commercial softwares I use also know which one to use (either low or high), depending on the study (lower impedance for short circuit, higher impedance for power flow), if it's allowed to (user option).

But yes I think it's a very simple enhancement, and I would add a precision property to the Branch class instead of the 2 properties you suggested (AFAIK it's always ±X%).

It should be an optional property, with a 7.5% default value for transformers, and 0% for other types of branch. The option to consider this precision should obviously be False by default.

from gridcal.

Indeed it looks simple enough, I'll try to implement it as soon as I get some time, or would you rather do it?

from gridcal.

I'd like to finish what I've started with the documentation before starting something else, so I'd be happy either way. Thanks for asking!

from gridcal.

I implemented the feature, did some playing around and it seems to be doing something, so it is ready for a formal test

from gridcal.

I like how you implemented it; it lets the user select the upper / lower band according to his wishes. I'll close this issue, I've added 2 basic tests in #69.

from gridcal.