Fix zero-sequence model of Yyn transformer

[TonyXiang8787]

Relates to #1162

Allow magnetization path flow of zero sequence in the yn side.

Physical modelling

More attention is needed to correctly (enough) to model the zero-sequence magnetization impedance, espacially for 3-leg core-type transformer.

OpenDSS

See refereces from OpenDSS

https://opendss.epri.com/3-PhaseTransformerModeling.html
https://opendss.epri.com/ModelingTransformerCoreEffectsin.html

pandapower

pandapower uses the following two attributes to represent zero-sequence magnetization:

https://pandapower.readthedocs.io/en/latest/elements/trafo.html

variable name	description
mag0_percent	z_mag0 / z0 ratio between magnetizing and short circuit impedance (zero sequence)
mag0_rx	zero sequence magnetizing r/x ratio

But this definition has some problem. If the user does not wish to model zero-sequence magnetization impedance at all. He/she has to give infinity to the mag0_percent. Also, it depends on the specification of leakage impedance of zero-sequence, which the user might not want to specify.

proposal in PGM

We can adopt pandapower attribute but allow user to direct specify i0 in zero sequence.

variable name	description	default
i0_zero_sequence	zero-sequence magnetizing current in per unit	i0
p0_zero_sequence	zero-sequence magnetizing core loss in W	p0 + pk * (i0_zero_sequence^2 - i0^2)

The values defaults to the positive sequence variant to allow user to leave them empty. We can add in the documentation to emphasize that if user has a 3-leg core-type transformer with Yyn winding, i0_zero_sequence should be given, and a good guess of the value is 1.0. We cannot default i0_zero_sequence to 1.0 because for tranformers which use separated iron core (usually for large transformers), the magnetization flow is the same for positive and zero sequence.

Note: The reason we name these *_zero_sequence is due to naming conflicts and the precedence of full naming like we have done with r_grounding_from.

Check list

• fix core
• fix existing test
• add new validation test
• documentation on new attributes
• documentation on explanation

[petersalemink95]

Added the do-not-merge tag, so we don't accidentally merge without positive confirmation from the user side.

[petersalemink95]

Adding a comment so we won't forget to update three-winding trafo's - can be done in a follow-up PR. This will just be updating the input - and probably tests, since it exists of 3 Transformers

[mgovers]

LGTM. Also approving given the positive feedback on real data as reported in #1162 (comment)

[sonarqubecloud]

Quality Gate passed

Issues
5 New issues
0 Accepted issues

Measures
0 Security Hotspots
92.3% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[petersalemink95]

FYI: for three winding transformers the issue still needs to be fixed. @TonyXiang8787 will make an issue for it

[furqan463]

@nitbharambe @mgovers is this change only for Yyn or all 2-W trafos. Should corresponding change be made in pandapower_converter of PGM-IO to read zero seq parameters specified in PP.

[TonyXiang8787]

@nitbharambe @mgovers is this change only for Yyn or all 2-W trafos. Should corresponding change be made in pandapower_converter of PGM-IO to read zero seq parameters specified in PP.

Hi @furqan463,

Theoretically this change affects all 2W trafos. But only Yyn will see significant improvements. You are right, the PGM-IO should be adjusted by fetch pandapower zero sequence parameters. Can you create an issue in PGM-IO issue board?

[furqan463]

@nitbharambe @mgovers is this change only for Yyn or all 2-W trafos. Should corresponding change be made in pandapower_converter of PGM-IO to read zero seq parameters specified in PP.

Hi @furqan463,

Theoretically this change affects all 2W trafos. But only Yyn will see significant improvements. You are right, the PGM-IO should be adjusted by fetch pandapower zero sequence parameters. Can you create an issue in PGM-IO issue board?

I'll create a PR.

[furqan463]

@nitbharambe @mgovers is this change only for Yyn or all 2-W trafos. Should corresponding change be made in pandapower_converter of PGM-IO to read zero seq parameters specified in PP.

Hi @furqan463,

Theoretically this change affects all 2W trafos. But only Yyn will see significant improvements. You are right, the PGM-IO should be adjusted by fetch pandapower zero sequence parameters. Can you create an issue in PGM-IO issue board?

@TonyXiang8787 just to be sure, power-grid-model is still not getting uk_zero_sequence and pk_zero_sequence from the user, as pandapower does take these inputs as well from the user as vk0_percent, vkr0_percent.

[TonyXiang8787]

@nitbharambe @mgovers is this change only for Yyn or all 2-W trafos. Should corresponding change be made in pandapower_converter of PGM-IO to read zero seq parameters specified in PP.

Hi @furqan463,
Theoretically this change affects all 2W trafos. But only Yyn will see significant improvements. You are right, the PGM-IO should be adjusted by fetch pandapower zero sequence parameters. Can you create an issue in PGM-IO issue board?

@TonyXiang8787 just to be sure, power-grid-model is still not getting uk_zero_sequence and pk_zero_sequence from the user, as pandapower does take these inputs as well from the user as vk0_percent, vkr0_percent.

@furqan463, yes, that's correct. power-grid-model still uses the same uk and pk for zero-sequence (adjusted for winding configuration).

In general, we are very careful to add new attributes, because removing them will be a breaking change. If users have a strong use-case to include uk_zero_sequence and pk_zero_sequence, we might consider it.