Is purple capacity and blue demand?
Datadog dashboard for the Texas power grid
21–30 of 137 posts
Re: Datadog dashboard for the Texas power grid
#22What are we supposed to glean from this as energy laymen? How do you provide data in a way that provides real meaning? All this data but no real insight.
For one, the peak price of electricity was ten times the usual low point (~$36) I'm not sure what to make of the frequency drift; federally operators are supposed to balance it out for timekeeping, among other reasons. Taking a week to get on-frequency is...not good.
The Texas interconnect is not federally regulated. There is also no such thing as a federal operator. There are three interconnects in the US, east, west and Texas. The other interconnects are not run by one central authority (let alone a federal agency). Each one has a mixture of authorities. The eastern interconnect, for instance, is run by several interstate operators (MISO, PJM and ISONE), intrastate orgs (NYISO), and vertically integrated utilities (Southern, Duke). Every one of these is a private enterprise. Each does have federal and state regulators. Ercot only has state regulators (funny enough, the Texas Railroad Commission is the Ercot regulator).
Re: Datadog dashboard for the Texas power grid
#23I was looking at the source code linked from dashboard, and surprised a bit that it's a gist, rather than a repository. Is it just to facilitate discussions on the code? Otherwise it feels a bit ad hoc that way, wouldn't consider a gist with this many files more ergonomic compared to a repository. Also, trying to follow to the data source of the Ercot website https://www.ercot.com/ I get blocked with things like this…
> Otherwise it feels a bit ad hoc that way, wouldn't consider a gist with this many files more ergonomic compared to a repository.
All gists are actually GitHub repositories — you can clone them, branch, and commit changes. There isn’t PRs or an issue tracker, though.
(Same with GitHub wikis.)
Re: Datadog dashboard for the Texas power grid
#24Odd to present temperature in Celsius for a US state, but otherwise very cool! I wish it were easier to see the moments when demand and capacity were at their closest, as those are the "interesting" moments.
Made in Germany! (I’m not the creator)
Re: Datadog dashboard for the Texas power grid
#25The surprising inference from me, as an energy layman, is that solar output coincides with periods of peak demand. Obviously this is seasonally adjusted, but in an era when the Saudis have solar contracts being supplied for $10USD/MWh [0], the potential profits for a solar provider at periods of peak demand seem more than sufficient to encourage investment, especially with interest rates still at a historic low for b…
Solar capacity in Texas tripled since last summer . https://www.washingtonpost.com/opinions/2022/07/21/heat-ener...
Thanks! Makes sense if it's being built as fast as possible given the number of workers available for installation.
Re: Datadog dashboard for the Texas power grid
#26The surprising inference from me, as an energy layman, is that solar output coincides with periods of peak demand. Obviously this is seasonally adjusted, but in an era when the Saudis have solar contracts being supplied for $10USD/MWh [0], the potential profits for a solar provider at periods of peak demand seem more than sufficient to encourage investment, especially with interest rates still at a historic low for b…
> So, what gives? Where's the solar?
Being installed as fast as the fabs can churn it out? Production capacity isn't infinitely elastic, factories take time to build. Solar is growing very rapidly, for reasons exactly like this, but it's not magic.
Re: Datadog dashboard for the Texas power grid
#27Earlier quoted context omitted.
Solar capacity in Texas tripled since last summer . https://www.washingtonpost.com/opinions/2022/07/21/heat-ener...
Thanks! Makes sense if it's being built as fast as possible given the number of workers available for installation.
Solar is actually super super hard to install right now, because tariffs are in an uncertain space, so nobody knows if, should they import, if they are are going to get a bill in the next 12 months for past sales that's 1x-2x the cost of the import (I forget the exact factor).
This sort of protectionism isn't even buying us a bunch of local solar panel production, unfortunately. I'm not a free-market purist, and am in favor of protectionist tariffs to build a nascent industry, but this is not a way to do to it and we are only harming our installer industry with these terrible forms of tariffs and uncertainty.
Edit: here's an article from a little while ago on this
https://www.reuters.com/world/us/us-solar-industry-warns-slo...
Re: Datadog dashboard for the Texas power grid
#28The surprising inference from me, as an energy layman, is that solar output coincides with periods of peak demand. Obviously this is seasonally adjusted, but in an era when the Saudis have solar contracts being supplied for $10USD/MWh [0], the potential profits for a solar provider at periods of peak demand seem more than sufficient to encourage investment, especially with interest rates still at a historic low for b…
> So, what gives? Where's the solar? Being installed as fast as the fabs can churn it out? Production capacity isn't infinitely elastic, factories take time to build. Solar is growing very rapidly, for reasons exactly like this, but it's not magic.
Also huge tariffs on Asian solar panels cuts demand.
Re: Datadog dashboard for the Texas power grid
#29I was looking at the source code linked from dashboard, and surprised a bit that it's a gist, rather than a repository. Is it just to facilitate discussions on the code? Otherwise it feels a bit ad hoc that way, wouldn't consider a gist with this many files more ergonomic compared to a repository. Also, trying to follow to the data source of the Ercot website https://www.ercot.com/ I get blocked with things like this…
> This seems ... pretty old school as well.
This doesn't surprise me. One of my projects as an intern was to generate reports from public data from ERCOT and a few other North American ISOs (organizations that control a region's power grid). Some of them are happy to reply to a basic cURL request. Others are more particular about things like your user agent or cookies set (even sites that didn't require authentication would set a cookie on first load and would reject requests to certain pages without that cookie). A few were very particular and ran some Javascript code to serve requests that made it so something like Selenium was required.
It was a bit surprising to me how many hoops we had to jump through considering this is public data that tax payers are entitled to.
If I recall correctly PJM was the best to work with because they had a well-designed REST API and provided a developer guide with sample code.
> I don't see anything that would _really_ be drain on those ERCOT resources...
I'm not sure if these controls are in place to safeguard against abuse. Most people don't even know these companies exist and I doubt they get too many requests. I'm assuming this is just a symptom of software sold to the lowest bidder.
Re: Datadog dashboard for the Texas power grid
#30Earlier quoted context omitted.
ERCOT didn't declare any capacity emergencies on that day, or any recent day. Is it possible you just suffered from a stochastic outage, and not a capacity-related intentional disconnection? https://www.ercot.com/services/comm/mkt_notices/opsmessages
Oooooor maybe they're not publishing notices when they should be. The grid being under-frequency for days at a time shows they didn't have enough generation capacity.
Pretty unlikely. Outages are common, especially in this above average summer, due to higher transformer load and temps.
Ever since the freeze, people who have even planned outages, refer to them as “rolling blackouts”. ERCOT is a pretty open data organization aside from QSE and generator data, and it’s unlikely that this data would have been fudged. We would have seen EEA3 in addition to $5k settlement prices if there was enough of a shortage to invoke rolling blackouts. These prices directly affect end users, e.g. businesses, datacenters, end users still on RTM rates, and solar buyback users. Generators definitely have offers open up to the $5k price point, and there is no way they’d be able to lock the settlement price without it being obvious to businesses/end users, or to generators/REPs/TDUs that there is market manipulation.
As far as sagging frequency, this is also wrong, because frequency sag calls up additional reserves, and we haven’t run below 1 gigawatt of reserves since the winter freeze. The number to watch for is 59.5 hertz, which is automatic EEA3 (we haven’t hit this since the winter storm, where we reached a whopping 59.3 hertz) This is also directly measurable by TDUs and (maybe) end users.