3 ms·
This is one of those rare posts on hacker news that I'm qualified to reply to (I normally lurk - had to make an account). I work as an expert in power system op
by probablypower 7y ago
This is one of those rare posts on hacker news that I'm qualified to reply to (I normally lurk - had to make an account). I work as an expert in power system operations for a national grid.
So with that disclaimer out of the way, I want to say that this research is incredibly misleading, particularly when it is presented as 'ways to avoid blackouts'.
What they present is an 'adequacy' model. That means, essentially running an accounting simulation of energy in and energy out. It is what power system engineers refer to as 'copper plate' model, in that it replaces the entire transmission and distribution network with a hypothetical infinite conductor where everything connects to everything else. Such a copper plate would never have a 'blackout' as long as Energy In = Energy Out.
Therein lies the bullshit. Blackouts don't occur just because Energy In =/= Energy out. They are incredibly complicated and nuanced, and often result from multiple minor events cascading into something larger [0].
You need, at a minimum, a steady-state power flow simulation to be able to analyse power system security. Particularly when assessing low inertia power systems (such as the systems proposed in this paper). The fact that the word inertia is not mentioned in a paper advocating high-penetration of wind and solar is a massive red flag (I'm talking, Moroccan palace style large).
I laughed out loud when I read "LOADMATCH runs quickly (e.g., ~2.6 min for a 5-year, 5.3 milliontime-step, simulation on a single Nehalem 5580 3.2 GHz processor)" knowing that it takes around that time to simulate a load flow on the pan-european grid, or around that long to do a basic security assessment on a single country's grid.
I worry that policy makers will be 'informed' by this research and will unknowingly sweep a lot of complexity under the rug until blackouts become more common.
[0] - https://www.entsoe.eu/fileadmin/user_upload/_library/publications/ce/otherreports/Final-Report-20070130.pdf https://www.entsoe.eu/fileadmin/user_upload/_library/publica...