Incorporating coarse DEMs

[Shreyas Rama]

Thought I would give some written thoughts on this feature I'm working on for future reference - the idea behind using coarse digital elevation models (greater than 1m) is that we can get estimates for large areas more quickly.

The downside is you lose roof characteristics - i.e. you can no longer do things like filter out parts of roofs that have a certain slope (for example there could be a chimney on the roof), or understand shading from nearby trees. So we're really only using terrain data and the shading that might come with that. It'll help regional and larger scale analysis, but not at the individual roof level.

So at some point, especially to obtain data that is useful for consumers interested in solar panels, we'd need to go back over with a finer 1m digital surface model and/or incorporate the buffering technique Cameron was talking about.

My two cents (this depends on Rewiring's goals)

While using a 1m DSM is slow computationally, we could run it for parts of NZ on an ad-hoc basis - perhaps there's an Electrify event happening in Lower Hutt, we could run the entire area beforehand for an entire year, upload the GeoPackage to Felt and have an interactive map people could use to look at their houses. We don't have modelling for different kW systems/inverters though - although we could probably do some rough calculations based on certain fixed parameters. But aside from that, all of that in theory is ready to go with our current pipeline, and to me seems like a way to get something more practical and immediate out of this project.

In saying that though I haven't been to an Electrify event and I'm not sure if there are panel installers that attend that do this sort of thing anyway!

[Jenny Sahng]

This is good to have written down, thanks Shreyas!

I thought the coarser DEMs were essentially pre-calculating the shading from the bigger environmental features from a bigger study area (hills etc.), but then we would "overlay" (I wasn't sure how, I just assumed there's a fancy geospatial way to do this) the finer 1m DSM (with a smaller study area around the building being analysed) to also take into account nearer obstructions and features like chimneys, trees, nearby tall buildings etc. Is that still possible?

The main output we need for Rewiring is an estimate of rooftop solar potential for all of NZ that takes into account:

• all shading: hills via the coarse DEMs (assuming the accuracy validation we talked about this morning turns out all right), but also trees/buildings/chimneys via the finer DSMs.
• weather: some estimate of weather, although this doesn't have to be super sophisticated because most existing solutions don't take weather into account, so any inclusion of weather along with the shading & topology would be an improvement on what's currently out there.
• topology of the roofs (angle, aspect): this includes filtering out roofs with unfavourable aspects (like totally south facing) or too-steep slopes (although the cost of solar is coming down so much that people are installing vertical panels, like solar fences!). These can just be parameters in the pipeline, so we can tweak based on assumptions (e.g. let's say NZ goes hundy on solar, and they just cover every roof including traditionally "bad" angles, what's the potential then?)

Anything around serving an individual user to estimate solar potential for their specific rooftop is secondary (like your event example) is secondary, because there are already existing tools that do this (homes.co.nz SolarEstimate, Raff's solar savings calculator), and solar installers do it via OpenSolar when giving out free quotes. While Electrify Waiheke's dashboard is super cool and feels like a satisfying tangible output, not knowing whether your roof can generate well isn't actually the main barrier to people getting solar - it's the lack of supportive policies in place. This is why we're focussing on getting the evidence we need to drive systemic change, i.e. the real nationwide potential of rooftop solar, which we suspect has been heavily underestimated to date and therefore deprioritised in policymaking.

--
You received this message because you are subscribed to the Google Groups "solarmap-nz-tech" group.
To unsubscribe from this group and stop receiving emails from it, send an email to solarmap-nz-te...@googlegroups.com.
To view this discussion, visit https://groups.google.com/d/msgid/solarmap-nz-tech/3eda0858-59a1-407e-9146-878c1da7fb27n%40googlegroups.com.

[Cameron Shorter]

Hey Shreyas,

Sorry for the delay,

Good to know about the down-scaling option.

Like Jenny says, I think this is something we may selectively apply to rural regions, but we wouldn't apply at the building level.

So maybe we wouldn't incorporate directly into the pipeline.py script, but rather call in after preprocessing.

Cheers, Cameron

To view this discussion, visit https://groups.google.com/d/msgid/solarmap-nz-tech/CALGcMYOim-SnHdW472c350R-7_HpZra49hFCKtcLvjFbXKQAhw%40mail.gmail.com.