Broadly speaking, electrical grids are the largest machines in the world and are held together by systems to keep everything in harmony at a target frequency (enormous spinning mass that slows when load is added, and has to spin back up, failsafes that have to decouple parts of the grid when maintaining voltage and frequency becomes impossible with committed capacity, etc). It’s actually wild it works flawlessly most of the time and is a testament to robust systems keeping them operational. There’s always room for improvement between here and some nebulous point of diminishing returns. That cost/resiliency discovery process is constant.
In this case, there is room for improvement, at some cost to be determined from a post mortem.
Batteries are taking over this role. Tesla has one near Houston, TX that can have an isolated electrical path established from it to geographically close thermal generators to provide blackstart capabilities (like jumpstarting a car). “Gambit Energy” is the project. To your point, distribution and grid segmentation solves for this, but it takes time and money to deploy batteries throughout a service territory, configure orchestration and transmission grade disconnects to rapidly isolate system components, etc. I see a large order of battery storage from BYD in Chile’s future.
Complete blackstarts are scary stuff. We have all this infrastructure to assist with pulling one off, but it's not the kind of thing you can test. As they saying goes "your backups are only as good as the last time you tried to restore them."
For example, you can't just turn on all the power plants. Ignoring the phase issue, power plants require some power to operate (which is where the flywheels and batteries come into play). This means you need to isolate them, and bring them up one-at-a-time (and you probably want to isolate consumers until the whole process is complete). How do you coordinate this in the absence of power/internet? Radios, but now you have to worry about the logistics surrounding that. Do you have a stockpile of petroleum that can be tapped without power, because your people aren't getting on-site in a reasonable time without a car. How familiar are people with the process, having never practiced it?
Couldn't you test it by either isolating an existing region and try to blackstart that particular region, intentionally out of phase with the main grid, and then try to synchronize and connect.
Black-start-region and join-regions should be enough to get a whole grid going?
If you can handle the consumer load in the region without local plants, maybe? That would still only be a 90% test because you can't completely blackout a region for testing purposes, consumers would get angry.
At times I wonder if Tesla, as far as the business goes, would be better off pivoting to being a battery company that happens to make cars than a car company that happens to make batteries.
They've been pushing the envelope on battery tech for some time now, and finding very useful, if at times novel, uses for it.
Developers want a solution, not cells. It’s why Tesla’s Megapack solution is so popular, and they have a years long pipeline even at 80GWh/year of manufacturing capacity (split equally between California and China). Autobidder is also state of the art for orchestrating the assets for developers.
With that said, BYD has a strong offering and recently landed a large contract (12.5 GWh) in Saudi Arabia.
AFAIK Tesla is already an energy tech company, some of their energy solutions happen to have wheels and some autonomy. We still think of Tesla as a car company because that is how they make money and also how most people experience Tesla given that they see them on the road. This is similar to Alphabet and Google Search situationship.
Perhaps if they have a CEO that focuses on the business such that BYD's CEO does. Without effective leadership, it's just an unattended machine that will not grow as fast as competitors will, although it'll continue to spit out cars and battery packs on existing lines as long as workers doing that work are happy enough to stay and quality stays "Tesla good enough” that consumers will continue to procure their products.
Hot take: Bring back JB Straubel (Former Tesla CTO). That’s who is the equivalent of BYD’s CEO. A driven, yet low key engineering leader. Culture comes from the top.
Well, currently the CEO is not only not doing his job at Testla, but actually doing his best to drive away Tesla's customers. And it's working, at least in Europe: https://www.bbc.com/news/articles/cvgd9v3r69qo
Musk fanboys are always impressed by how much he does, but it's becoming clear he can't have time to be the CEO of Tesla, owner of X, CEO of the Boring Company, CEO of SpaceX, CEO of SpaceX, CEO of xAI, owner of Neuralink, Reichsleiter of the DOGE, and top-20 Diablo IV player.
Great time to recall that Musk was unaware of the battery storage program at Tesla until someone pointed out the test installs they'd done of units in the car park to him. He immediately called for it to be shut down.
The battery storage part of the company is growing quickly with a new factory in China just coming online. That part of Tesla is growing faster than cars and Elon says it could be a bigger business in the long term.
Huh! Today I learned that inverter technology is advanced enough that batteries can also provide inertia and powergrid frequency control that flywheels were used for!
All (grid tied) inverters technically provide that function. An inverter doesn’t just generate 50/60 Hz sine, it supplies power by attempting to slightly “push” the frequency higher, i.e lead, to provide power.
if there’s no more power to provide, it “pushes less”.
AC voltage must stay within spec, and the grid has voltage control in place, but that’s not what determines the direction of power flow. Like, what makes the inverter output power rather than draw it?
In fact that’s how individual sections of the grid set power flow limits, by ensuring they stop leading when they reach their max capacity. fascinating really
Electric grid get more reliable with scale, not less (that's why parts of North Africa is interconnected with Europe, which itself is interconnected all the way to Russia).
Small grids (like islands) are always a nightmare to manage for that reason.
North Africa is interconnected to Europe because Europe thought that it would be a good idea to build huge solar plants there.
The project ultimately failed because long distance energy transportation is difficult and expensive, among other issues like general instability of the region.
No, this is independent: first of all the Spain-Morocco interconnection predates the Desertec project by half a decade, and survived after its demise, but it has never been designed to be able to carry that much electricity (it's barely capable of carrying 1.4GW): it cannot be used for significant export and is there for grid stability alone.
Had you read the Wikipedia page you posted, you'd have realized that they planned to use dedicated HVDC lines under the Mediterranean to carry the power, and not the Spain-Moroco interconnection.
In this case, there is room for improvement, at some cost to be determined from a post mortem.