False positive for "Connection between different voltages"

[wolfy1339]

Whenever a single phase line connects to a 3 phase line, osmose complains about a connection between different voltages without a transformer.

Ex: node 11927030446 where a single phase line (14.4kV) branches off of a 3 phase (25kV) line

https://osmose.openstreetmap.fr/en/issue/4408c4ad-0518-233d-4bca-795f3dbf383b

Item 7040 Class 3

[frodrigo]

cc @flacombe

[flacombe]

That's a good point, thank you @wolfy1339

I add this one to the to do list of improvements already set for this analyser.
We should add a sqrt(3) term in the voltage calculation.

By the way, and despite it has no sense from electrotechnical perspective, we use to map single phase lines with the same voltage as the poly phase ones, particularly when they are supported by mono poles.
This one won't get 289 kV instead of 500 kV

[wolfy1339]

Your example isn't the same situation.

Mine is actually a single phase like that branches off, not simply that the 3 phases split onto different poles and rejoin afterwards.

[flacombe]

It's true we shouldn't compare a phase split and a branch, on the exact point they occur.
Sometimes, phases that split off never join again.

However, they both lead to the same situation of a power=line with cables=1, on which we have different practices about the voltage expression. Hopefully, we can do math to infer the correct value. I'll set it in the next PR about this analysis.

[flacombe]

Hello

I'm currently working on this issue and I manage to properly compare voltages depending on cables=*.
But the expected single phase line voltage is 14434, not 14400.

It will still drop a warning as voltages aren't exactly the same.

I think I miss an implicit rounding in this computation.
I could round on 100s, to get 14400 in this particular situation, but why on 100s and not on 1000s?

[wolfy1339]

These seem to be the voltages in the specifications for Hydro-Quebec 1

tension nominale 25 kV (14 400/24 940 V)

[flacombe]

Yes, 14400 works for 3-phase voltage 24941, but main line is tagged with voltage=25000, not 24941.
voltage=25000 leads to 14434 (25000 / sqrt(3))

[flacombe]

25 / sqrt(3) rounds to 14.4.
I may adapt the rounding to the scale, I'm looking forward something generic