Proposal talk:Substation refinement

From OpenStreetMap Wiki
Jump to navigation Jump to search

Transformer type

This topic has been moved to here from Talk:Proposed_features/Power_generation_refinement

Unresolved: In progress --Don-vip 06:14, 12 June 2011 (BST)

I don't see the point to tag a transformer type with the values tower/indoor/outdoor:

  • It's not a transformer type but its location
  • indoor seems redundant to me with building if it's a small building not tagged as an area, and useless if a bigger building area is surrounding the transformer.

If we want to tag a transformer type, it would be rather the types described on wikipedia, but they seem far too technical and without geographic interest to me (and quite difficult to find, as the operators usually don't share this kind of technical information).

So, I'd like to delete this attribute, but wanted to talk about it with you first. --Don-vip 13:48, 4 June 2011 (BST)

I have added many local 20kV transformers in my home town:

If you check these transformers, you can see how many different transformers are hidden within residential areas.

E.g., our local hospital has two transformers which are converting 20kV power to local voltage:

If you deleted them, our hospital would not get any electricity, thus operations would become impossible ...

These transformers are located within a building.

In rural areas, transformers are usually located on towers:

This tower often supplies a whole village. bahnpirat and I agreed to tag it with power=transformer and transformer=tower (or tower=yes) in order to distinguish it from indoor transformers (like the one in the hospital). --FK270673 22:28, 4 June 2011 (BST)

Alright :) Would you mind to explain it on the main page ? Thanks :) --Don-vip 00:57, 5 June 2011 (BST)

After much more thoughts, I have change my mind. I definitively agree this information is helpful, so it is still OK to tag it, but not as it was done:

bahnpirat and I agreed to tag it with power=transformer and transformer=tower (or tower=yes)..

I don't remember saying this in this way. If there is a smaller power line power=minor_line to connect a village and the transformer is mounted to the pole, there is a conflict in OSM. This node Node is a pole and transformer at the same time. In this case I would use power=pole;transformer (it's a pole on the minor line). Additional there is a conflict with ref for pole and ref for transformer. Same rules for power=tower, just less often to be see. --Bahnpirat 09:35, 10 June 2011 (BST)

Sorry (or no sorry) for including you into our discussion ;-)

We have power=substation/transformer, location=tower, ref=64241 (transformer number beats pole number, otherwise we must introduce pole_ref, tower_ref_left, tower_ref_right). If you want to tag the transformer separately, you should mark him 20cm away from the tower ;-) It's the same problem with subway station, cable distribution cabinet, postbox, pharmacy, museum etc. located in the same building on several floors. --FK270673 15:54, 10 June 2011 (BST)

What about tagging it as power=tower;transformer + ref=tower_ref;transformer_ref ? --Don-vip 06:01, 12 June 2011 (BST)

I would like to add an option that no-one has suggested yet. power=tower + transformer=*. If you wonder why tower/pole should be primary and transformer secondary to it then try to imagine a free floating transformer in the air without pole/tower.--RM87 21:05, 12 June 2013 (UTC)

transformer=* is an attribute for transformers, hence there must be a power=transformer tag. And the transformer is not free floating, it is firmly connected to ground via the pole indicated by the location=pole tag now part of the proposal. --polderrunner (talk) 21:50, 11 June 2013 (UTC)
power=tower (or pole) should be the "first tag" when describing a vertical structure supporting the wires, and any other features of the poles and mounted on the poles should not overwrite that tag. Imagine any current rendering, that wants to show the lines, and symbols at the points where the towers or poles are. If there isn't either of the tags (they're normally rendered differently - because they are different), most maps would be missing the towers/poles with transformers. Even the existence - let alone the details - of the transformers is secondary to anyone not making a specific "power line" map. There could well be aviation marker balls represented as nodes on the way, so you can't draw tower/pole markers on every node part of a (minor_)line. Don't mix orthogonal features to one key, when they can appear together, and alone. And oh, often especially in rural areas the osm ways for major lines visible in the poorer, say, landsat imagery, the lines are drawn with too little precision, so the nodes of such ways only represent roughly course of the power line, but not the locations of individual towers. Alv (talk) 22:36, 11 June 2013 (UTC)

Switch on a 20 kV power line pole

Please note that you can have many other features than only transformers on a pole. Switches would be another example.

IMHO power=pole/tower should be added prior to power=tranformer. But since power=transformer is important too, why shouldn't we create 2 nodes linked by way power=line?
Without any relation not to mention.
Feature on the pole will be tagged as usual with a corresponding value of location=*. Hosting pole can be retrieved by following the linking way.
It won't be so wrong and you find this way of drawing/mapping on many operators's charts.
Maybe this idea needs to be refined with your comments but it would solve many issues, here and in Proposed_features/Power_transmission_refinement Comment by User icon 2.svgFanfouer (on osm, edits, contrib, heatmap, chngset com.)

I have adapted the proposal for pole mounted transformers as suggested by RM87. It is now power=pole + transformer=distribution and no location tag. This should overcome the objections. It could also be used with towers but I still have to see a transmission tower with a transformer. Concerning other objects in a pole. I think this is mapping at a too detailed level and I won't propose any mapping scheme for such devices. The proposal includes a tag for switches but it is really intended for large circuit breakers, primary at voltages above 100 kV (and you don't find such things mounted in poles or towers). --polderrunner (talk) 19:12, 6 August 2013 (UTC)
Except transformers and switch, I didn't see anything else hosted on a pole. It's a pity not to include switches in those configurations (mainly for topological reasons. I agree to over-cluttering mapping).
And what happens if we see 2 or more transformers on the same pole ? It may be uncommon but it can occur.
The problem is transformer=distribution - and all the rest - applies here to a power=pole feature and not to a power=transformer one. Does voltage=* with 2 values has a signification on a pole instead of a transformer ? Fanfouer (talk) 09:41, 7 August 2013 (UTC)
Two transformers in one pole? Still have to see that and transformer=distribution could also be interpretated as meaning "at least one transformer present here" :-) For the voltages, well they would of course apply to the implicit transformer.
I don't think there will be a great interest in creating topological grid maps at the medium voltage levels. In most places it is also very difficult to map distribution networks at a sufficient level of completeness to make such use possible (lack of published data, undergrounding, poor aerial imagery etc). So mapping switches at that level seems of little value. If you insist you could add a switch=yes tag (not part of this proposal). Transmission networks are easier to map, more documentation is available and more people have interest in the network topology at higher voltage levels. For distribution networks (minor_line, poles etc) I would say: Map what is visible in the landscape, useful for general navigation etc but probably not for analysing such networks. --polderrunner (talk) 19:43, 7 August 2013 (UTC)
According to what I sent you last night, I think the problem goes further than the singular case of transformer or switches.
By letting both poles/tower (supports) and lines/transformer/switches/whatever (hosted) in power=* tag, we'll have difficulties to bring up a consistent tagging model. Even if we can find tricks with transformer now.
Towers or poles don't have so much to do in power. They're only man made supports. Thus they can be mapped with man_made=pole or man_made=tower, not to mention there are many many many features currently mapped with power=pole or power=tower - it's only an idea.
Keeping only hosted features like lines or transformer in power=* prevent us from encoutering future difficulties and consistently solve our big problem.
Deprecating power=pole or power=tower is more a transmission case than substation. I let everyone telling me how does it feel about that. Fanfouer (talk) 13:29, 13 August 2013 (UTC)


This topic has been moved to here from Talk:Proposed_features/Power_generation_refinement

Just to let you know if you missed it, I have changed the voltages tagging as follows:

  • substation: The different voltages in input/output of this substation, ordered from highest to lowest (ex: 400000;225000;63000).
  • transformer: The two voltages used at this transformer, input first, output then. A step-down 400kV/225kV is tagged 400000;225000 whereas a step-up transformer is tagged 225000;400000.

--Don-vip 14:02, 4 June 2011 (BST)

Rural transformers usually have voltage=20000;1000 or voltage=20000;400. --FK270673 22:32, 4 June 2011 (BST)
You're right, I've added low voltage values as a step-down example :) --Don-vip 00:51, 5 June 2011 (BST)

One moment, please! A transformer is the device to transform between exact two (2) levels. So Matthew Newton once proposed voltage-high=* and voltage-low=* and I think these would work best here. Is there really a differense in step-up and step-down transformers? Aren't these the same? A power=substation and power=line can have one or more than 1 (may be 4) different electric tension for in/outputs. I agree to use it here like voltage=big;lower;lowest. --Bahnpirat 09:35, 10 June 2011 (BST)

Resolved: I agree, it's the same thing. I've changed the tagging to set it it "high;low", so we have a consistent tagging with substations. --Don-vip 04:49, 12 June 2011 (BST)

Three levels: low, medium, high

I don't know, what the differences are in detail (especially if I had to explain it in english ...) - but there is a big difference between voltages over and below 1000 V (safety precautions, safety zones etc.). So I would support having three rough levels of voltage: high, medium and low. The following levels could be possible: low: up to 1000 V, medium: 1000 V - 50000 V (or 60000 or 80000 - I don't know where there could be the upper level according to international used levels - in germany the upper level could even be something like 30000 V, because the most used levels are 20000 V, 10000 V and 15000 V) , high: more than 100000 V (or whatever the upper level of medium will be). I don't know if these three levels are common wording in english, but they are in german - and I think they are very reasonable. -- Schusch 12:12, 13 June 2011 (BST)

If we introduce these levels, we cannot refer to their voltage numerical value, because it is unknown ! This wording is used on English Wikipedia, and we have the same terms in French. I will probably create a new Voltage chapter to define these values, but the choice will be subjective, according to what the mapper estimates the voltage is. --Don-vip 15:48, 13 June 2011 (BST)
it's quite easy to see if the voltage level is low, medium or high - depending on the length of the insulator (at least in germany). But especially for medium voltage, it's not easy to know, if the medium voltage is 10, 15, 20 or maybe 30 kV. That's what I wanted to address with these levels. Greetings -- Schusch 22:22, 21 June 2011 (BST)
These voltage levels have moved to this new proposal: Proposed_features/Power_transmission_refinement. --Don-vip 23:12, 30 June 2011 (BST)

Substation type

This topic has been moved to here from Talk:Proposed_features/Power_generation_refinement


Why did you change the spelling of 2 out 3 substation types ? English is not my native language, I don't know if "generating" is better than "generation", but Wikipedia always use "-ion". If we switch to "-ing", would it be "transmitting" instead of "transmission" ? --Don-vip 14:02, 4 June 2011 (BST)

In America, they are usually abbreviated with DS and TS. Some English sources call it distributing substation, some call it distribution substation. We should try to find a solution that is not confusing for non-English (e.g. Asian) speakers, so we should avoid inflected words. Given that distributing, generating and transmitting are inflected words, we should try to avoid them, so I did a mistake. --FK270673 22:40, 4 June 2011 (BST)
Resolved: No problem :) I've switched back to "ion". --Don-vip 00:53, 5 June 2011 (BST)


Resolved: Minimalist tagging scheme described. Can be improved later, if needed. --Don-vip 13:32, 17 July 2011 (BST)

Is substation=generation needed? Surely no substation generates electricity, even one that is associated with a power plant. Lets stick with the Wikipedia definitions and use transmission substation for substations that are associated with generator plant as well and we then have the same three types as wikipedia describes.

  • Collection for gathering power from multiple locations within a single plant.
  • Transmission for conversion between voltages within the network.
  • Distribution for final supply to a residential area. It might still be difficult to distinguish between transmission and distribution in all cases.

-- PeterIto 06:23, 8 June 2011 (BST)

OK, generation deleted. Let's stick with Wikipedia :) In France, it's easy to differentiate transmission from distribution: first ones are operated by RTE (national electricity transportation company) and their voltage go from 400 kV to 63 kV, whereas the other ones are operated by ERDF (national electricity distribution company) and their voltage go from 20 kV to 230/400 V --Don-vip 04:58, 12 June 2011 (BST)

Collecting from a single source sounds a little bit crazy, but we have many crazy wordings here. What do native English speakers think about it?

In Germany, substations with 110kV are considered distribution substations whereas substations with 220kV or 380kV are considered transmission substations. Substations with 20kV are considered transformers ("Trafostation") even if there were two redundant transformers within one substation. Should we introduce a new tag "trafostation" for a little single transformer substation, e.g. on a tower?

--FK270673 07:05, 12 June 2011 (BST)

Should we consider each substation with less than 200kV as distribution substation?

  • 765/500/400/380kV substation=transmission (long-distance) + voltage=500000 (undoubted)
  • 161/138/120/110kV substation=distribution (for a whole county) + voltage=110000
    • or 161/138/120/110kV substation=transmission (from high to low voltage) + voltage=110000 (filter them manually by their voltage)
    • or 161/138/120/110kV substation=conversion (from high to low voltage) + voltage=110000
  • 20/12/10kV substation=distribution (for a small village) + voltage=20000 (undoubted)

--FK270673 07:21, 12 June 2011 (BST)

Please, FK270673, could we first discuss before you add something on the proposal ? For substation types, if Wikipedia's unclear, we could stick to the official IEC classification instead:

IEC substation types
IEC number IEC denomination OSM type Description
605-01-01 substation (of a power system) none ? the part of a power system, concentrated in a given place, including mainly the terminations of transmission or distribution lines switchgear and housing and which may also include transformers. It generally includes facilities necessary for system security and control (e.g. the protective devices)
605-01-02 switching substation switching a substation which includes switchgear and usually busbars, but no power transformers
605-01-03 transformer substation transformer a substation containing power transformers interconnecting two or more networks of different voltages
605-01-04 step-up substation transformer a transformer substation in which the outgoing power from the transformers is at a higher voltage than the incoming power
605-01-05 step-down substation transformer a transformer substation in which the outgoing power from the transformers is at a lower voltage than the incoming power
605-01-06 traction substation traction a substation, the main function of which is to supply a traction system
605-01-07 converter substation converter a substation including converters and the main function of which is to convert alternating current into direct current or vice versa
605-01-08 frequency converter substation frequency_converter a substation in which an a.c. current at a given frequency is converted into an a.c. current at another frequency

Dear Don-vip,

you may add, criticize, delete or modify any of my suggestions whenever you want if you have arguments against them as I consider them as brainstorming ;-) --FK270673 06:10, 13 June 2011 (BST)

It would just be easier to reach a consensus on the talk page :) Concerning the substation types, are you OK, Peter and you, to remove wikipedia definitions and use only IEC ones I gave in the first table ? --Don-vip 16:16, 13 June 2011 (BST)

I've written a minimalist tagging scheme for that. I want to move forward with power plants and generators with this proposal, and this substation nature discussion was going too much detailed. --Don-vip 23:14, 30 June 2011 (BST)

Transformers and substations

This topic has been moved to here from Talk:Proposed_features/Power_generation_refinement

Unresolved: Have to think about it. --Don-vip 05:12, 12 June 2011 (BST)

How does one distinguish a substation and a transformer? Do substations contain transformers or are transformers simply small substations for low voltages? I suggest that anything that converts power within the transmissions network is a substation regardless of the voltages involved. A single transformer on a pole in a residential area in the USA would therefor be a substation consisting of a single transformer. The transformer tag would be reserved for a device visible within a large substation. Lets nail this one at this point because there are different views within the community. PeterIto 06:17, 8 June 2011 (BST)

In Germany, these little substations converting 20kV into 1kV are called trafostation. --FK270673 07:10, 12 June 2011 (BST)
Lets be clear to avoid confusing the 'thing' with the 'name of the thing'. There are many places in the world where different things are known by the same name and same thing is known by different names. How can we reliably describe the difference between a substation and a transformer? Either... 'a transformer is a part of a substation, particularly a large substation', or 'a transformer is a small unit that converts voltages which is up on a pole rather than being on the ground'. Or possibly something else? If we can't describe it in a way that anyone in the world can interpret reliably then we should possibly call them all substations. PeterIto 18:26, 12 June 2011 (BST)

A substation (601-05-01) is

* the part of a power system, 
* concentrated in a given place, 
* including mainly the terminations of transmission or distribution lines 
* switchgear and housing and 
* which may also include transformers. 
* It generally includes facilities necessary for system security and control (e.g. the protective devices)

Definitely not an easy definition!

A substation is an area used for connecting a high-voltage transmission system with other power systems, e.g. a low-voltage distribution system.

is my trial to get an easy definition.

--FK270673 05:52, 13 June 2011 (BST)

I agree a substation would only concern areas. A "substation" consisting of only one transformer on a single power line should be tagged as a transformer.

I see the same philosophy with small generators and power plants: a single wind turbine should be tagged only as a generator, and not a power plant.

--Don-vip 16:05, 13 June 2011 (BST)

I agree that a substation is generally an area that may contain one or more transformer nodes (and a plant is an area that may contain one or more generator nodes; a plant could also contain a substation).

However, real life does not always match the above ideal. We may not always have enough detail available to mark the outline of an area; therefore, we must allow a substation (or a plant) to be represented by a single node.

--T99 05:35, 20 June 2011 (BST)

What's missing ?

This topic has been moved to here from Talk:Proposed_features/Power_generation_refinement

Street level

Resolved: Out of scope. This should go to Proposed_features/Power_transmission_refinement. --Don-vip 13:27, 17 July 2011 (BST)

In residential areas there are smaller transformers, may be little houses or big boxes, which transform 20 kV (or so) to level for households ([1]). I like to name them also power=transformers with voltage=* and name=* and/or ref=* if visible. These distribute the electric energy to power=cable_distribution_cabinets, which may also have ref's.


Not mentioned till now is the frequency. I know four different exist on the planet. 50 Hertz in Europe and 60 Hertz in North America and Japan. Additional there is the power grid of german-swiss-austria railways which are using 16.7 Hertz. And of cource 0 if it is DC. Are there any more? These can be (in OSM DB, not electrical) on the same power=line! I use frequency=50;16.7 in this case and would like to add this to the proposal.

Resolved: The frequency=* is not mentioned in the proposal because already described on the wiki and widely used. --Don-vip 06:08, 12 June 2011 (BST)

Details within a substation

Resolved: Too much detailed. This should go to Proposed_features/Power_transmission_refinement. --Don-vip 13:27, 17 July 2011 (BST)

Inside the surrounding fence of a substation there are transformers, switches and busbars. These should be added.

  • way power=busbar an open metal pipe
  • node power=switch There are two switches behind eachother. One separating two lines electrical and very fast ([2]), filled with SF6 gas. The other one additional disconnect optical but slow ([3]).

A power=busbar would be appropriate if you trace each of the three phases separately. If we draw a single way to represent the entire three-phase circuit (this is called a "one-line diagram"), then a power=bus would be more appropriate.

There are also other equipment in a substation, such as "current transformers" that measure the current flowing in a circuit.

--T99 05:42, 20 June 2011 (BST)

I used power=bus (one bus equals three busbars) and power=switch tags to trace a simple switching substation (only one voltage and apparently only one type of switches) here. Does this approximate what you had in mind?

--T99 06:21, 21 June 2011 (BST)

Okay. Since both switches are very close to eachother and hard to seperate on an arial image, and in the past I also use only one node to tag this, it would be okay for me to tag one node as power=switch. Using power=bus for 3 phases is also good for me. But for each parallel bus, if possible.
I forgot to mention power=portal for low metal construction with the shape of a bottom up 'U' would this be okay? --Bahnpirat 08:15, 22 June 2011 (BST)

What if ...

This topic has been copied to here from Talk:Proposed_features/Power_generation_refinement. Please only continue discussions related to substations and transformers here. Further comments concerning power lines should be added at Talk:Proposed_features/Power_generation_refinement#What_if_....

Some open questions

.. there is a power=transformer mounted at the power=pole or power=tower ?

location=tower --FK270673 16:10, 10 June 2011 (BST)

Resolved: I mark this question as resolved because it will be answered in the very first subject of this discussion page. --Don-vip 06:12, 12 June 2011 (BST)

.. a power=line and power=minor_line using the same towers ?

Tag it with power=line;minor_line ?

PeterIto and I agreed to use voltage for distinguishing power lines as mappers generally don't agree where a minor_line begins (20kV, 46kV, 69kV). We should slowly deprecate minor_line and use power=line, voltage=20000 instead! --FK270673 16:10, 10 June 2011 (BST)

Resolved: I agree. I' have added minor_line to deprecated tags, with some others (including minor_undergound_cable) --Don-vip 05:58, 12 June 2011 (BST)
I think it is sensible to deprecate minor-line. Possibly we should suggest the value of 'voltage=low' for lines where the voltage is is definitely low because the of the design of the insulators, poles, height off the ground but the exact voltage is not known. Similarly it may be useful to codify 'high' for a more serious power line where the voltage is not known. This would also allow a bot to convert 'power=minor_line' into 'power=line,voltage=low'. PeterIto 18:33, 12 June 2011 (BST)
You're right. I have added voltage=low in the deprecated chapter. --Don-vip 01:59, 13 June 2011 (BST)
Different voltage lines are not connected, even when they are supported high above by the same structure. They are not the same line, so they should be different ways with different tags, but using the same nodes. Alv 08:25, 16 June 2011 (BST)
That is not the tagging way currently in use (a single way power=line, with its number of cables=*, and the semicolon-separated values of voltage=* if the cables have different voltage). --Don-vip 11:56, 16 June 2011 (BST)

.. different power=lines using the same tower and split up later? (introducing relations)

Resolved: Beyond the scope of this proposal ? --Don-vip 21:06, 15 June 2011 (BST)
To much for just one proposal, may be seperate proposal later. --Bahnpirat 08:21, 16 June 2011 (BST)

Use relations in this case. Like bus routes they use the same streets (power=line) but are only connected at the bus stops (power=substations). The way is tagged as normal and mentioned in the proposal. I think of a relation with the main components (generator,line,substation) as member. Not included are power=towers. Example relation 941141 on map.

Hi Bahnpirat :)
Thanks for the interest you find in my proposal. After many thoughts, I don't think I will include this idea in this proposal, even if I find it very good. Why ? Because it's far beyond the scope of what I want to make of this proposal. I have created it essentially to define a consistent way of tagging power plants, not overhaul the full power scheme. Don't forget we have an entire section for power proposals, we are not forced to merge all of our ideas in a single proposal. For the same reasons, I think it's finally a bad idea to consider the deprecation of power=minor_line, power=underground_cable and power=minor_underground_cable in this proposal. I think that would need a separate proposal, focused on power transmission (mine is entitled generation, after all). If you want to create it, I assure you will have my vote :) --Don-vip 21:06, 15 June 2011 (BST)
I agree this proposal is big enough and so don't include it here. --Bahnpirat 08:21, 16 June 2011 (BST)
Two methods of circuits tagging with relations are proposed in the arcticle Power_lines. --Surly (talk) 19:00, 11 February 2013 (UTC)

.. power=substation is part of the power=plant. One or two areas ?

Many areas! Some plants have a 500kV substation, a 230kV substation, a 138kV substation and a 69kV substation!

Brunot Island is a very big plant with many substations that could be used as a sample for complex plants:

--FK270673 16:10, 10 June 2011 (BST)

.. line is deconstructed, but still visible on aerial image ?

see Talk:Proposed features/Power generation refinement#Life cycle management power=abandoned, voltage=20000 --FK270673 16:10, 10 June 2011 (BST)

No ! If the line has been deconstructed, it's probably because it has been grounded, and so tagged as it (power=cable). --Don-vip 05:30, 12 June 2011 (BST)
Clarified in the proposal. --Don-vip 16:19, 13 June 2011 (BST)
There are two good reasons NOT to delete a demolished line. First of all it prevents other mappers not aware of the current situation from remapping the line from aerial imagery. Secondly it may also be useful to keep a record of a historical line that is no more. I simply change "power=line" to "power:dismantled=line" which effectively deletes the line from the map. I also add "end_date=*" if I know when the line disappeared. --polderrunner (talk) 20:53, 11 February 2013 (UTC)

line change from air to ground

Resolved: Bahnpirat 10:33, 10 June 2011 (BST)

There are fenced substations, towers or poles where power=line become power=underground_cables! The places could be nodes or even areas! How to tag this on a node pole/tower or an area ?

There is no need to tag this! Whereever overhead and underground lines share a common node, there is a changing point! --FK270673 16:10, 10 June 2011 (BST)
I'm not sure to understand the issue ? When the power line becomes an undergound cable, you will have a power=tower shared between the power=line and the power=cable, right ? --Don-vip 05:36, 12 June 2011 (BST)
I'm tagging such areas as substations, they can be quite big, see for example here (2*380 kV + 2*150 kV transitioning to underground cables). Technically they are not real substations as there is no switching or transforming taking place but there is currently no alternative tags. If the cable terminates in the tower itself I tag it "tower=air_to_ground" --polderrunner (talk) 20:35, 11 February 2013 (UTC)
I agree with polderrunner here. We can tag the fenced area as a power=substation since it can sometimes house transformation or switching devices (often not obviously). Underground lines may be in HVDC instead of aerial ones which mostly are powered by alternative voltage. Power poles can have tower=air_to_ground tag to signify the environment change. Fanfouer (talk) 21:45, 11 February 2013 (UTC)
This question must had been solved by introducing the substation=transition class. It may be pleasant to specify, if it's not already available, the kind of transition the substation surrounds : air -> underground, underground -> air, underground -> undersea, etc. Fanfouer (talk) 15:00, 25 March 2013 (UTC)
I find it unnecessary to further specify the nature of a transition station, that is already evident from the power lines and cables entering it.--polderrunner (talk) 17:37, 2 April 2013 (UTC)
I agree, resolved question :) Fanfouer (talk) 17:33, 4 April 2013 (UTC)

Filtering rules for substations

This topic has been moved to here from Talk:Proposed_features/Power_generation_refinement

Most maps and applications needs to distinguish between low voltage substations/transformer stations and high voltage substations/switchyards.

Filtering rules for mapnik and Overpass API are easy in the existing power=station/sub_station concept but very difficult with the proposed "voltage" tag. If the only recommended "voltage" tag is missing, an application cant distinguish at all.

I recommend to continue to use different tags. --Seawolff (talk) 01:15, 6 March 2013 (UTC)

Except that the 'station' tag is not being used as 'intended' by its creator so you don't get what you would expect. If you retrieve 'power=station' objects you get a mix of big and small substations (including the 'kiosk' type transformer boxes) blended with a few percent power plants. This indicates that mappers really struggle to understand and use the 'station' tag. Mapnik currently doesn't care about substation type, they are all rendered the same shade of gray. However, it is possible to render substations based on voltage at least for renderers that understand regular expressions such as Maperitive. I have created a Maperitive ruleset demonstating that, see User:Polderrunner/Maperitive_rules/Power_rules. With Overpass API you need to query using both tags to get all substations otherwise the result will be rather useless. If you query for 'station' only you will probably get only half of the big substations (and a lot of objects you weren't looking for). In most cases you probably want to retrieve all power related objects anyway.
One possible solution to the 'difficult' voltage tag is already in the proposal: The new substation=* key with appropriate values such as 'transmission' vs 'distribution' could be used to distinguish 'large' and 'small' substations. This would provide an alternative to the voltage tag for querying (mappers should use the voltage to determine the transmission/distribution cutoff). --polderrunner (talk) 20:43, 6 March 2013 (UTC)
Parsing the "voltage" tag for classification is complex and error-prone. On your example ruleset you forgot the voltage 450000 used on the Baltic Cable for example. Any new power line with an different voltage demands to modify all rulesets. Some tools can´t evaluate complex expressions and many users can´t write such rules.
Voltage parsing is not ideal but it is possible. I didn't forget 450000 as this voltage is not used in AC grids as far as I know. The ruleset does not render HVDC lines according to voltage, just as a fixed colour. I believe that the AC voltage list is more or less complete and that new AC voltages are unlikely to be created due to current standards (HVDC is another story as each project seems to define a new voltage). --polderrunner (talk) 20:55, 11 March 2013 (UTC)
A (recommended) key with fixed values would be a better solution, but it shouldn´t mix voltage level and user (domestic, industrial, transport). --Seawolff (talk) 19:16, 10 March 2013 (UTC)
I'm ok for fixed values (in parallel of voltage=*). May we introduce two tags? The first for level and second for usage as for making rules more custom than now (or only with voltage). The commonest tag names, the better. Fanfouer (talk) 19:54, 10 March 2013 (UTC)
I have introduced the two values substation=transmission and substation=distribution, see Proposed_features/Substation_refinement#Substation_properties. My suggestion is to used the former for substations of more than 100 kV and the latter for voltages below 100 kV (this cutoff voltage seems to suit at least Belgium, Netherlands and Denmark well (>=110 kV and <=70 kV with nothing in between). The other values for substation type are more specialized. --polderrunner (talk) 20:55, 11 March 2013 (UTC)
In France, limit between transmission and distribution is lower : 400 kV >= Transmission > 50 kV while 20 kV >= Distribution > 0V. It's not a problem since such values are right here for abstracting real and numeric voltage levels.Fanfouer (talk) 15:19, 25 March 2013 (UTC)

Changes to proposal

After taking over the substation proposal the following changes have taken place:

  • added several new values to substation=* including transmission, distribution, converter, transition.
  • added location=* with several substation specific values such as indoor and outdoor.
  • expanded the transformer section to a full section on substation components. It is noted that the tagging scheme for substation components is for mappers having a special interest in power infrastructure and that the normal mappers may safely ignore this rather technical section.
  • switchyard components:
  • added transformer=* values converter, generator, auto, phase_angle_regulator. Added property tags phases=* and rating=*
  • added power=converter with various property tags.
  • compensation components such as power=shunt_reactor and power=shunt_capacitor

--polderrunner (talk) 21:33, 22 March 2013 (UTC)

A summary of later updates to the proposal:

  • added power=switchgear as feature when switchgear cannot be mapped in finer details (inside a building etc)
  • redefined the compensation tagging. Now power=compensator + compensator=shunt_reactor/shunt_capacitor/..

--polderrunner (talk) 20:38, 22 May 2013 (UTC)

Bjæverskov example

On the Bjæverskov example given at the end of proposal, I don't understand why there's only one horizontal busbar showed on chart whereas there are both on bing.
There's a switch between that two 3-phases busbars and I don't think it's correct to connect the two side of that switch to the same busbar don't we?
Moreover there are bays between those busbars and they're not connected the same for all incoming power lines. Fanfouer (talk) 21:58, 1 April 2013 (UTC)

In this case I have mapped the two busbar assemblies as a single line (cables=6). Some details are lost (like the 'self connecting' bus coupler) but the rendering gets less cluttered. It would probably be a good idea to specify that either every three-phase busbar assembly should be mapped individually or that they may be 'grouped' together using the cables tag (leaving the choice to the mappers discretion). The Bjæverskov substation is a bit unusual in that the vertical busbar for the HVDC converter is attached as a bay to the horizontal (double) busbar (I need to correct that). --polderrunner (talk) 17:56, 2 April 2013 (UTC)
Ok for the less cluttered render but we lost useful information about substation architecture. As for maintaining reliable topological information dealing with power networks, every circuit elements must be mapped separately. By using cables=6, we lose the switch utility which connect both busbars, as you said yourself. Moreover, with such cables=* value, someone may think it's a 6-phases circuit instead of 2 3-phases ones.
It will be hard to recover from that for reconstructing topology, won't you? Fanfouer (talk) 11:27, 5 April 2013 (UTC)
Ok, I have now made it a recommendation to map each busbar assembly individually. The Bjæverskov example has been updated accordingly. Ok? --polderrunner (talk) 19:47, 5 April 2013 (UTC)
Sure I'm ok :) Thanks for update. Fanfouer (talk) 12:59, 6 April 2013 (UTC)

Busbars mapped as assembly, not single conductors

I noticed that user Surly has added the tag phase=* to the proposal suggesting that busbars could be mapped as single conductors, each with a name or ref. First of all single busbar conductors are normally not having a name or ref as far as I know (or this is strictly internal information of the substation operator). Mapping busbars to this level of detail provides serves no practical purpose in my opinion (and mostly not possible to map due to limited resolution of available imagery). Therefore I have removed this tag again. --polderrunner (talk) 20:06, 5 April 2013 (UTC)

I don't know about the other countries, but in Russia each phase of busbars, switches and transformers are marked by a label. A label has the letter and is coloured. The phase L1 (A) is marked by yellow colour and letter "Ж" (Желтый=yellow), the phase L2 (B) is marked by green and letter "З" (Зеленый=green), and the phase L3 (C) is marked by red and letter "К" (Красный=red). Markers are hanged onto the cable and are well-visible. This is a country-wide standard and it is always applied.
So, as I think, it is up to a mapper, whether he/she wants to map a busbar as a whole or as an individual phases. Sometimes resolution of an imagery, mapper's knowledge, instrumental or eye-sight measurements make possible to map phases, sometimes not.
I have never seen anybody mapping busbars per phase (but I haven't looked at Russian power infrastructure either). There are exceptional cases where power lines are mapped per phase when monopolar towers are used. My example (Bjæverskov) also shows use of single phase lines due to the employment of single-phase transformers. Otherwise I would stick to the practice to map per circuit.--polderrunner (talk) 18:32, 6 April 2013 (UTC)
I see at least one practical case. If you know that the substation is on your way, and the yellow phase must be on the left and the red phase must be on the right, but you see them vice-versa -- then you instantly understand that you have gone astray.
That seems a bit far fetched to me.
The other practical case is that we needn't restrict OSM to "common" wanderer. If OSM can help to create a detailed map for power engineers then we mustn't throw away this possibility.
Even power engineers would normally use one-line diagrams when analyzing power networks. Detailed per phase diagrams are most likely only used during substation construction and to make sure you get the phases of incoming feeders connected correctly.
You CAN map busbars (single conductors) individually using the cables tag (cables=1) and you can attach any tag you like. But if it provides any useful added value I really doubt. It just clutters the map and makes it difficult to get a good overview. It may actually hamper the making of power grid maps as they usually expect circuits, not individual phases to be mapped.--polderrunner (talk) 18:32, 6 April 2013 (UTC)
That's not to mention local lore needs, that are cognitive and interesting.
--Surly (talk) 21:24, 5 April 2013 (UTC)
Let's say it the same question than in Bjæverskov. We had opportunity to map 2 different busbars as a single way but for topological reasons it was better to distinguish both busbars with two different ways.
Can't we find a solution that would allows both detailed and global approach? Using cables=* for instance with correct values (cables=3;3 is better for two 3-phases busbars with the same topology than cables=6 but it would contradict current definition of this tag where lists are for distinguish voltages).
Hard but interesting question.Fanfouer (talk) 13:22, 6 April 2013 (UTC)

Gas Insulated Substations (GIS)

Sometimes, substations may be insulated with gas when not enough place is available to do it with air (used as usual). This proposal doesn't introduce anything to map it properly. Have a look :
I propose to add a tag gas_insulated=SF6, gas_insulated=dinitrogen (experimental) or even gas_insulated=yes if unknown.
An example on GMaps : A wole 400 kV French substation is insulated with SF6. Fanfouer (talk) 10:55, 4 September 2013 (UTC)

Good idea. I had actually thought about introducing such a tag. I will only propose gas_insulated=yes as the gas composition is probably not known to mappers (it is either pure SF6 or a mixture of SF6 and nitrogen in current installations). By the way, there are also such things as gas insulated lines (GIL)! They can be overground or underground. Might be interesting for your transmission proposal. --polderrunner (talk) 18:12, 4 September 2013 (UTC)
Ok. You should add the tag on busbars and bay too because sometimes only part of the whole substation or the whole switchgear is gas insulated (and when the switchgear is mapped in details, we shouldn't have any area corresponding to this switchgear as described).
I'll think on how can I add it to the transmission proposal indeed. Fanfouer (talk) 08:38, 5 September 2013 (UTC)