Back in 2013 (or even more earlier), several contributors started to look for a solution to state what actual paths power can follow over the physical lines we are used to map in OSM.
Because of some relevant abstractions for an efficient mapping of physical objects we see on ground, it’s not possible to deduct power paths easily from power lines. Several practices raised and are now established about power routes/circuits. Most of them use relations involving different segments of power lines.
These are important information to get a proper model suitable for power planning in many countries, researchers already demonstrated they really like OpenStreetMap over official data.
This topic has already been discussed in previous proposals that couldn’t be voted yet. It doesn’t mean that previous attempts failed but it’s a complex topic that needs to iterate on. David and I are part of the Oh my grid! initiative that has been announced earlier yesterday.
We are pleased to take a supplementary step on this topic with the Power circuits proposal. It aims to unify tagging and mapping practices on the matter.
We invite anyone interested to express their concerns about it.
This chart from the proposal rationale summaries the best what it is all about. Most of us add physical segments as power=line between substations (in red) and the proposal covers the two supplementary levels:
This proposal is the first I contribute on as part of a team like Oh my grid! initiative.
Although we are not set yet, we are aware of concerns that may rise and plan to be cautious about the proposal process, and particularly about the voting.
As a proposal intended to improve things and reinforce available tagging, we first of all seek the comments and then the approval of the community.
Feel free to discuss it here or on the wiki Talk page!
I think you will have to demonstrate how the circuitry information can
be verified on the ground (i.e. without referring to priviliged
information from the network operator).
OpenStreetMap generally maps what is verifiable on the ground, and makes
exceptions from this rule in a few important cases where there is
general agreement that the information is important for the map - like
administrative boundaries.
Power circuitry is a niche endeavour. It is important for your project
but it is not important for OpenStreetMap. Therefore it cannot sail
under the same exception that we make for administrative boundaries; it
can only be mapped in so far as it is visible on the ground.
Only data that can be verified on the ground is data that can ever
“belong” to OpenStreetMap in the sense that we mappers can improve it.
Data that is not verifiable on the ground is just third-party data for
which OpenStreetMap is used as a collection and (re)publication vehicle.
If “Oh my grid” is looking for a platform where people can collaborate
to collect publications from energy companies and build large data sets
from many small pieces they found on various websites, then you must
build such a platform outside of OpenStreetMap. I can see how it would
be convenient to piggyback this on OpenStreetMap with its existing
infrastructure, but it would be abusing OpenStreetMap and the data would
be liable to deletion for lack of verifiability.
Therefore, if you want your proposal to succeed you must explain in the
proposal how each of the things you wish to map can be deducted by
visual inspection on the ground. Anything that requires looking at plans
(or being employed by the network operator) is not suitable for OSM.
The fact that a handful of mappers have been toying with this for many
years cannot serve as a justification to roll it out as an “official”
feature in OpenStreetMap.
Could you explain how you came to the conclusion that this is generally not important for OpenStreetMap? I see a massive use and relevance of the power data around the world in academia and industry. We a currently working on a impact website that shows the relevance of the power data in general.
The fact that a handful of mappers have been toying with this for many
years cannot serve as a justification to roll it out as an “official”
feature in OpenStreetMap.
I’m sorry to say this, but “toying” is simply the wrong tone in this respect. This is unfortunately a pretty serious matter and the people who work with it are professionals and what is done with it is certainly far from “toying”.
The proposed data is some of the most important data to support the expansion and modernisation of the electricity grid, one of the biggest bottlenecks of our time in the fight against climate change. More information about this can be found here: From Taking Stocto Taking Action: How to implement the COP28 energy goals
I think there are two important points that require early clarification as to not get stuck in misunderstanding:
Yes, sections and circuits relations can be fully set from ground inspection and relies on visible knowledge first. That point isn’t elaborated enough in the proposal and we will fix it in the rationale to explain it better.
It’s true Oh-my-grid aims to gather data from various sources. We will always case by case study what is possible to send in OpenStreetMap (license, relevancy and ability) and be open to discussion. We currently focus on ending major transmission power lines in substations and find missing ones with suitable imagery. We started to put some documentation on wiki, particularly on https://wiki.openstreetmap.org/wiki/Power_networks
I do not dispute that information about the power grid is valuable and that many people want to have and use this information, including your organisation (assuming your “we” indicates that you are working for “oh my grid” too).
What I do dispute is the relevancy of this data for OpenStreetMap. OpenStreetMap’s core is not that we are a data distribution platform; we are a platform where people on the ground record information that they collect because it either hasn’t been collected by anyone yet, or maybe it has been collected but is not freely available.
The number of people who are able to survey power structure above the “there is a power pole and a line here and what looks like a substation over there” level is extremely small, and certainly will never reach a point where data about power circuitry (i.e. not “there is a line here” but “this line has this voltage and connects these elements in these substations”) can meaningfully be “owned” by OSM in the sense that we mappers can be a first-class source of this information.
My fear is that there’s going to be a handful of people in OSM who are interested in and capable of surveying these things and they will be able to cover a tiny fraction of the planet. For everything else, you will have to rely on importing third-party data and it will sit there in OSM, untouched, until someone else replaces the import with another import. That is not what OSM is for; data that cannot be meaningfully improved by us is worthless to OSM and those who wish to collect, use and distribute such data should find a different platform to do that rather than abusing OSM’s platform, convenient as it may be.
I think you underestimate both the number of people who map power feature in OSM and the effort of these people to map these features. Open Infrastructure Map is just based on OSM-data and has a much larger coverage of power lines with details than just ‘there is a line here’.
Taginfo won’t say how many user accounts are involved in such relations creation and I may find out with attic data, we’ll see (I’m involved myself in France mainly).
The curve will continue to go up, with or without this proposal. The question is the same as always: isn’t it better at the end to document, review and provide consistent tagging to anyone involved?
This proposal is not responsible at all in the start of mapping such things, they are already widely mapped and this data is less valuable with less reconcilable tagging.
It seems to me that you have a very strong opinion without being able to back it up with figures. Could you please start to disprove what we are saying in quantitative and professional way @woodpeck.
If you want we can also have a call to discuss this things. I’m also from Germany.
I requested from overpass all the power=route relation with meta data. There are 1154 different users associated to this tag, 194 users with at least 10 routes, 34 users with at least 100 routes
I’d like also to share a specific example of the benefits of routing : This power line supports 2 circuits : Way: 107167009 | OpenStreetMap
If we didn’t have the routing information, we couldn’t know what circuits is linked at this end to which substation. Thanks to the 2 associated relations, this is possible. This is primary information when considering power network, and since this is deductible for ground (even if it is a bit tricky, this is still feasible) and sometimes from satellite view, in my opinion, this has a legitimate place in OSM.
At last (but not least), there are currently two widely used tag for such configuration : route=power and power=circuit and no satisfying documentation for both of them. This proposal is welcome to harmonize and give guideline for those who are concerned by these considerations.
Here is a bit of power line. Presumably this proposal affects that, and what is at each end of it, in some way. How should I survey that power line in order to update it with that data that this proposal talks about?
Local Knowledge
OpenStreetMap emphasizes local knowledge. Contributors use aerial imagery, GPS devices, and low-tech field maps to verify that OSM is accurate and up to date.
You’re right, but I don’t see how to get the creator and not the last modifier
By the way, around 40% of the 23000 power=route relations haven’t been modified since their creation (version=1), so I would say that even if my statistics are not so accurate, the conclusion (many people are using the tag) would remains. If I look only these 40%, there are already 125 differents users that created them.
Which doesn’t say that everyone must be able to verify it, but rather that it must be verifiable.
To be clear, I agree that it is Good Practice (using the OSM definition) to have instructions on how to verify the power data being proposed. Which the proposal seems to at least attempt in the “How will we get such information?” sections.
But I don’t see where we say that a substantial amount of local mappers should be able to verify this with no further instrumentation or learning, as Frederik asks for when referring to the extremely small amount of people “who are able to survey power structure”. So I asked if there is someplace that we do say that?
An ordinary person on the street may be unable to accurately provide the height of a tree or a building, or the gauge of a rail line, but it can be measured using an instrument and a process.
But how can you verify this data? Substantial amount of learning or instrumentation is fine, as long as it does not require changing job to be let into substation.
Proposal basically shrugs and tells people to import data. If data cannot be verified at all then it should not be put into OSM.
Thank you for asking.
First of all, no the proposal doesn’t impact this power line, everything is fine regarding tagging on it.
Regarding the survey, you may go to both ends, on ground or on aerial imagery to look at how it ends in both substations, Lincoln Main and West Burton. Two circuits lines usually splits near of substations, here for instance.
Along the main section, mapper can check if the conductors go along without any connection to other lines, branch or split mainly.
Reading name plates on towers could also help to find particular points, to know which substations are actually involved.
Near of Lincoln Main, another split will certainly allows two circuits to reach different anchors/portals.
Once direct connectivity is known between two substations, it’s possible to build two relations following the conductors along the lines segments.
In this process, no public data has been used and it is observable from ground.
