Proposal talk:Power storage

From OpenStreetMap Wiki
Jump to navigation Jump to search

This is the discussion page for the Power Storage Proposal.

Storage isn't a source

Resolved: Tagging as been reworked accordingly

Hi and thank for this comprehensive proposal, which extends the current generator tagging. Despite I agree to tag such devices and it's interesting to have them in OSM as they are more and more used everywhere, storage isn't a source. OSM understands power generators as kind of power converter from a source to an output. It's particularly the case for power to gas processes, where hydrogen could be used for numerous other uses than producing back electricity.

It's not a reproach about proposing something that shouldn't be mapped, but storage capability should get its own generator=* subkey I think.

Furthermore, I think power=generator should get a refinement to introduce a stronger input definition, instead of source that confuses actual inputs and sort of power family. This proposal could be an opportunity to make some experiments, how do you feel about that? Fanfouer (talk) 22:25, 22 December 2020 (UTC)

I agree with you. My first intention was to put it under a separate tag power=storage, I was just worried to deviate that far from what is currently in use.
The question that would still remain, would be how to distinguish between the individual modules (battery modules for example) and the whole storage facility. If you have suggestions on that, I would see how I can change up things. --Hedaja (talk) 20:20, 24 December 2020 (UTC)
Thanks for writing this proposal (I am very bad at tagging proposals)! I would argue that a power storage facility is still a power plant/generator, and should not be tagged separately as `power=storage`. I agree that source should not be "storage" however.
One important example of this is pumped-storage hydro, which in some cases can act both as a traditional generator (making use of surplus water flow), or as a storage method. It makes sense here to keep the source tagged as plant:source=hydro, and to use an additional tag to indicate that it's also capable of storage.
Personally, I don't see what's wrong with simply listing battery storage as plant:source=battery or generator:source=battery.
Incidentally, I have been using plant:storage=* to record the storage capacity of a plant (in MWh or whatever), and I documented that briefly on the wiki page. Russss (talk)
I think you are right Russss I'll restructure the whole thing. :source would be electrochemical/electrostatic/heat/mechanical/magnetic/.
I made an updated example under the topic tagging (V2). I would appreciate if you could have a look at it and let me know what you think about it.--Hedaja (talk) 15:55, 28 December 2020 (UTC)
I rebuild the whole table after some external input --Hedaja (talk) 10:11, 29 December 2020 (UTC)


Resolved: Tagging as been reworked accordingly

"Sabatier process" value likely should "sabatier_process" Mateusz Konieczny (talk) 08:58, 23 December 2020 (UTC)

I checked the tag and it already says sabatier_process I only kept the methode names shown in the table like they are for other power generators. Or am I missing something. Feel free to edit it. Thanks --Hedaja (talk) 20:24, 24 December 2020 (UTC)

Is supercapacitor a battery?

It is a bit confusing to have supercapacitors here.

As I understand it is 100% theoretical and but it is a subset of batteries, especially from functional viewpoint (I may be wrong with both).

Is it really necessary to have separate plant:method? Is it really useful to list theoretical power storage methods? Mateusz Konieczny (talk) 09:04, 23 December 2020 (UTC)

Super capacitors are a electro-statical energy storages while batteries are elector-chemical storages. They are not theoretical but they have niche functions. Some universities for example use super capacitors to store energy to release them in a very short time for experiments or for example National Ignition Facility (NIF) fusion experiment uses capacitors to power its lasers (
There are many types still in a development phase but I wanted to be future proof. --Hedaja (talk) 20:30, 24 December 2020 (UTC)

Hydro storage

What about pumped water storage? That is currently the most prevalent mode of energy storage. Right now those are tagged with plant:source=hydro + plant:method=water-pumped-storage. With your proposal, this would be changed to plant:source=hydro;storage + plant:method=water-pumped-storage. -- Janko (talk) 11:52, 23 December 2020 (UTC)

Yes I thought about that. But Since it is already so well established, I was a bit hesitant to change it. I know the trench fights that often breakout around retagging established things (see contact: prefix)
I'll add a note that this needs further discussions. --Hedaja (talk) 20:58, 24 December 2020 (UTC)
However, proposal isn't clear enough to state if storage:*=* should be added to hydro pumped storage plants. Could you improve this please? Fanfouer (talk) 13:20, 30 December 2020 (UTC)
I still have my worries about changing pumped hydro. I therefore so far left it out but I now added a section about what would change for it. --Hedaja (talk) 15:27, 30 December 2020 (UTC)
I agree to your suggestion written on the proposal page (section "Changes to existing tags that would need further discussion"). Pumped storage facilities have to loose plant:method=water-pumped-storage for storage:type=pumped_hydro (or similar) to be consistent. As a consequence, plant:method=* can be set to plant:method=water-storage in case that the upper basin of the facility is also fed by rivers/streams (i.e. the upper basis has a natural catchment area). -- BlueG (talk) 12:06, 2 January 2021 (UTC)

Prevent redundancy in tag values

Resolved: storage no longer uses prefix for plant or generator

Currently, plant:storage:*=* and generator:storage:*=* are proposed here. As introducing plant:output=* and generator:output=* was an actual mistake and output=* would have been enough, I suggest to only define storage:*=* to be used independently on power=plant and power=generator Fanfouer (talk) 08:57, 29 December 2020 (UTC)

I united it under storage:* but we still have the discussion about that further down. --Hedaja (talk) 15:05, 30 December 2020 (UTC)


As power=generator definition is a device that converts a form of energy to another, it may be relevant to define power=battery. From an external point of view, a battery doesn't convert anything, with electricity as input and output. This new feature would solve issues we currently have in defining generators capabilities they don't actually have. Fanfouer (talk) 09:03, 29 December 2020 (UTC)

But that would only allow for battery based power storage. How would we include other types of storage? I don't think it would be feasible to have an individual power=* tag for each type of storage. Battery might be most prominent (at the moment) but the other technologies are in use too.
After all a battery does convert energy. It converts it from electricity into an electrochemical potential and back.But I agree that it doesn't convert it from input to output. Having a dedicated power=storage sounded good to me at first, before I stumbled across the problem of distinguishing between individual modules (generators) and the whole facility(plant). What would you think about power=storage + storage=facility/module with all the :source/:method/:type tags? --Hedaja (talk) 10:09, 29 December 2020 (UTC)
I also think that using "generator" is confusing. But I like your idea of using/defining power=storage for the individual storage devices within a power plant. But I would not invent a key storage=facility/module. This is already covered/implied by your subkeys storage:source=*/storage:method=*/... -- BlueG (talk) 11:48, 29 December 2020 (UTC)
Sorry, I think I misread your idea of storage=facility/module. You suggest using storage=facility to map an facility which exclusively stores energy and has no other sources of energy, right? And you suggest using storage=module for an individually mapped storage device within a power plant, right? And finally power=plant with plant:storage:*=* if the storage device is not mapped individually within a power plant? -- BlueG (talk) 12:04, 29 December 2020 (UTC)
Yes, you summed up the idea of distinguishing between individual storage and a whole facility nicely. Hornsdale Power Reserve for example would be a dedicated power=storage +storage=facility (with lots of storage=module) while for example Kraftwerk Fenne ( would be power=plant with individual power=storage modules for the battery storage ( --Hedaja (talk) 12:17, 29 December 2020 (UTC)
See my mail in Tagging@. A battery as an assembly of cells (492-01-01) is actually a particular kind of generator (151-13-35). Then I agree to define a battery a generator for chemical energy only. A flywheel, compressed air systems aren't batteries at all.
Storage is a capability of some generators but power=storage won't be consistent with other power=* values which are devices or sites and storage isn't. Fanfouer (talk) 12:30, 29 December 2020 (UTC)
Technically, the most correct would probably be power=energy_storage_device but I don't see the harm in shortening that to power=energy_storage with the device part mentioned in the definition. - Luke (talk) 09:46, 30 December 2020 (UTC)
power=energy_storage_device semantically shows storage is a capability of an undetermined device. I'm not happy with that as power=* currently got values referring to precise devices: transformer, generator, pole... It's way different than respective electricity_transferring_device, energy_converting _device and power_line_support and any new value should be consistent with that Fanfouer (talk) 13:23, 30 December 2020 (UTC)
I for my part don't feel like generator or plant are not any more precise or loose than storage. All of them describe what this thing can do without defining how. For example if I only have power=generator I don't know whether I have a huge wind turbin:e in front of me or a small flat solar panel . Only additional tags make them more precise.--Hedaja (talk) 15:10, 30 December 2020 (UTC)

Existing capacity=*

It's currently proposed to use new storage:capacity=* to state which amount of energy a plant can store. Can you consider using capacity:electricity=* instead please?
Experience shows that current generator:output:*=* and plant:output:*=* should be a single output:*=*, part of a scheme completed by similar input=* and capacity=* without prefixing it usage as to get as generic concepts as possible Fanfouer (talk) 13:28, 30 December 2020 (UTC)

I think it might not be ideal to tag it as capacity:electricity=* since the energy within isn't really stored as electricity. Especially with things like thermal storage, it might confuse people. I could see that just using capacity could work but so far capacity itself has only been used to mainly been used to describe an amount of people and not materials/electricity. But I agree that output for plants and generators could be short without loosing much (besides the tags being listed underneath each other more nicely in the tag list). What advantages do you see of creating the new tag capacity:electricity=* instead of the new tag storage:capacity=*? --Hedaja (talk) 15:03, 30 December 2020 (UTC)


Resolved: Tagging as been reworked accordingly

I just add this as a new section, since it is only mentioned as a side note in the section above: storage:input=* for the peak input power (in kW, MW, ...) is an important characteristic of a storage device and is (mostly?) not equal to storage:output=*. Therefore, this tag should be added to the proposal. -- BlueG (talk) 12:16, 2 January 2021 (UTC)

thanks for reminding me. I had this on my mind one day but forgot to add it. I added it now. --Hedaja (talk) 12:50, 2 January 2021 (UTC)

molten_salt vs. molten-salt

These seem too similar, to the extent that I thought there was one duplicated tag at first (the former is proposed for thermal storage by melting salts, the latter for electrochemical storage using batteries with electrolytes made from liquid salts). Recalling which is which would be next to impossible. I wonder if a more generic hot_liquid would be better for an always-liquid thermal store, whether salt or not. --ChrisHodgesUK (talk) 11:40, 3 January 2021 (UTC)

didn't even realize that. Since the electrochemic batteries are often called "liquid metal batteries" too, I would say this would be a good alternative --Hedaja (talk) 13:49, 3 January 2021 (UTC)

How can a hot water tank be tagged?

How can a hot water tank be tagged? Hot water storage tanks are energy storage systems with a boiler from wind power or solar collector (power=generator + generator:source=solar + generator:method=thermal) to be charged.

Currently, they are tagged with man_made=storage_tank + content=hot_water.

I think they also belong to the power storages and should be taken into account. The current tagging has not yet been finalized and could be adjusted again.--geozeisig (talk) 10:23, 4 January 2021 (UTC)

I would leave them as they are right now. because they usually need a combination of storage (the tank itself) and a generator that transforms the energy back. Therefore I would only tag the facility, as shown in the tagging table. I can make that more clear in the description. For example for the compressed air storage I would only tag the facility and in case for example you have an underwater storage I would tag them as storage tank ( and have the compressor/turbine separate. --Hedaja (talk) 20:48, 4 January 2021 (UTC)

Usage of plant:method=*

I'm not sure, if plant:method=* should be further advertised. The tag itself was never approved and kind of slipped into usage as a duplicate of generator:method=* and contrary to plant:source=*, which has a similar history, it is in my opinion unnecessary and in some situations might even create confusion. Especially with hybrid power plants that use multiple forms of power generation you get long semicolon separated values that provide no real value. Instead you could also directly analyse all generators and get even better information.--TOGA (talk) 21:54, 10 January 2021 (UTC)

So you say we should just go forth and also include pumped hydro in the scheme? Was just worried to face more opposition with changing this established tag. --Hedaja (talk) 10:54, 17 January 2021 (UTC)
No, my concern is just about usage of plant:method=* vs. generator:method=*. Granted, this isn't directly related to your proposal, but in the rationale you mention that plant:method=water-pumped-storage has some usage, which is certainly correct, however you missed the more common and also approved tagging on the generator level with generator:method=water-pumped-storage. I found this a little strange, especially because I have some concerns regarding the usefulness of plant:method=* as mentioned above.
Regarding the question, if pumped hydro should be included in this proposal: With these facilities we usually have two distinct, spatially divided areas, first the generator/pump house and second the reservoir, right? So why not just tag the reservoir with power=storage and keep the already established tagging with power=generator on the generator house (or as a node inside the building)?--TOGA (talk) 06:28, 25 January 2021 (UTC)

Construction of tagging scheme

Currently you use storage:type=x + x=a, which looks a little weird. I think you should either go with storage=x + x=a or keep the parallel to power=generator and go with storage:method=* and storage:type=*. Additionally if we compare the values listed under storage:type=* with the generator tagging scheme, I think storage:method=* would be more appropriate for those and storage:type=* would be the key for the more specific values like lithium_ion, lead-acid, etc.--TOGA (talk) 23:10, 10 January 2021 (UTC)

I initially had written the proposal to be similar to generator but people opposed using this kind of scheme again because they didn't like it for generator either. Leaving out :type might be an option. I'll probably check different channels for more opinions. --Hedaja (talk) 10:52, 17 January 2021 (UTC)
Yes, I can see pros and cons with both options, but I definitely think they shouldn't be mixed.--TOGA (talk) 06:45, 25 January 2021 (UTC)

Some additional thoughts concerning the structure of the scheme. I'm wondering if the differentiation between sensible and latent heat is really necessary or if using thermal as the value would be enough. Especially for average mappers the information that a facility provides thermal storage might be easier obtained than the specific system used, which would be needed to categorise it as sensible_heat or latent_heat. As far as I can see there is no overlap between the values of the subtags, so from this viewpoint there wouldn't be a problem.
A similar shift could be done in the category mechanical. Flywheels currently have no further subtags and the distinction between cavern and underwater for compressed air systems could also be achieved with location=underground/underwater.--TOGA (talk) 07:45, 9 February 2021 (UTC)


Not directly related to your proposal, but I stumbled over your usage of brand=*. This key is normally used to describe the franchise to which for example a fuel station or fast food restaurant belongs. What you want to record is the manufacturer, this is done with manufacturer=*.--TOGA (talk) 07:57, 9 February 2021 (UTC)

Any chance you have an example of what your talking about? --Adamant1 (talk) 08:50, 9 February 2021 (UTC)
See the Retagged Examples section, the first one lists brand=Tesla for the individual modules and the third also uses this key.--TOGA (talk) 05:32, 10 February 2021 (UTC)
Resolved: marking as resolved as unrelated to proposal - comment on user talk page, or even better using changeset comments Mateusz Konieczny (talk) 10:27, 9 February 2021 (UTC)
Roger, I was unsure because the problem is clearly on the proposal page however not part of the proposed new tags. It's noted for the next time.--TOGA (talk) 05:32, 10 February 2021 (UTC)

THank you anyways. I changed it in the proposal and in the OSM data. Also added wikidata tag at the same time. --Hedaja (talk) 08:39, 12 February 2021 (UTC)