Load drops, the big spinny generators stop having so much resistance and accelerate.
Load increases and the big spinny generators have more load, synchronously slowing down.
So in simplest system just feeding your generator more when frequency is below nominal and less when it is above is enough
I'm glad there's at least two of us ;)
Almost all cheap / "simple" consumer mains appliances, including non-"smart" microwaves, ovens, alarm clocks, etc. still use mains frequency as their time reference.
Due to the growing complexity of power grids and in Europe, international power-grid politics and infighting, the grid frequency is becoming less stable and you see these devices fluctuate badly more often than they used to. https://hackaday.com/2018/03/09/europe-loses-six-minutes-due...
I find this incredibly hard to believe. Crystals are incredibly cheap and waaaaaay more accurate than anything that the grid does.
If you're not an actual motor, sensing the grid is both an engineering challenge and a non-trivial expense.
The grid has poor short-term performance, but the long-term frequency is averaged out to be exactly 50/60Hz.
Turns out it had a button combo for “switch between 50Hz and 60Hz mode” that was similar to its combo for (what I thought I was setting) “switch between 12- and 24-hour mode”. (I think the 12/24 switch was “hold the hour button down while plugging in” and the 50/60 switch was “hold both buttons down while plugging in”.)
Some systems still use grid frequency for timekeeping.
If you own a mechanical rotary timeswitch, it's got a synchronous motor spinning 50 or 60 times a second, then a series of gears, gearing it down to 1 rotation per day.
https://www.theguardian.com/world/2018/mar/08/european-clock...
Very, very well done
"What happens when I turn on an electric device in my house?"
"A turbine in a power plant spins more slowly for just a moment."
By far and away the easiest way to control production - i.e. brake/feather turbines or dump PV, is frequency shifting, as I’ve batteries and inverters in several locations and while networking them would be possible, it’s unnecessary. It’s typically intended for grid-tie operations, but here I use it to control our tiny isolated grid.
It’s a pretty small range (50.2-53hz) over which they shift, but it’s more than enough.
I also discovered that some cheaper LED drivers are extremely sensitive to frequency shifts, and bulbs just flat out flat out die within weeks.
Huh, that's just outright bizzare. The most adding few hz should do is change power by few %
Practical engineering has a great video about power black starts that give some insight into this complicated machine.
PS One of the answers in the SO thread mentions JET in the UK. I spent a few summers there as an electrical engineering student (it's home to the MAST and JET fusion reactors). When the JET tokamak ignites a plasma, it can't sustain it for very long (we are not yet at the point of extracting enough energy to sustain the reaction). As a result they need to ignite the plasma and keep it hot. They can't do it for more than 1-10 seconds. During that time, they draw massive amounts of power - they're permitted to draw up to 1% of the UK's capacity for a short period, whilst they simultaneously dump all the energy stored in two gigantic flywheel generators housed in a nearby building. I've never been there when the flywheels are running but I've climbed around beneath them. There's nothing quite like massive engineering :)
Other problem is that for renewables to be profitable you want to push all the energy out all the time, especially in peak. Even now solar installation users have problem with that when there is too many small solar installations installed on same street the voltage goes too high and the inverters just trip and stop pushing the power to the grid, losing owner money.
We just need to have more cheaper storage solutions. Technically utilities could just put a bunch of batteries near concentration of residential solar and just basically sell the service of "storing the kilowatts" to them (say "you can receive 80% of what you put into it in next 48 hours"), all while having the capacity to use that stored joules in case a peak needs to be handled
Generator spins at 50hz Motor spins at 50hz
Add load to the motor, it starts slowing down both itself and the generator. Generator governor increases input energy, frequency goes up. Both are in phase all the time.
Remove load, both start spinning faster. Reduce governor to regulate it to 50hz again.
It's a bit harder with inverters but the idea is similar, you follow the grid phase and if you want to send energy to the grid you will be slightly early to the grid phase and if you want to take energy from the grid your phase will slightly lag.
The difficulty is not "keeping them in phase", that just happens (aside from initial connection), it's the whole load prediction and handling, when to tell which plant to start producing more or less energy, with variety of plants having shorter or longer ramp-up/down periods
The most difficult bit for an operator is to make sure the generator is synchronized before they actually connect it. Only after it's synchronized can they start actually feeding power in.
You can't just turn these things on and off at will like a regular motor. To extend the gears analogy, gears need to be synchronised before they can engage - just like synchronous AC generators.
[0] https://arstechnica.com/tech-policy/2021/05/texas-power-outa...
Battery production is currently shifting from hundreds of gwh per year to twh per year worth of production. They last quite long too. If you cycle 1 twh of battery every day, it would run out in about a decade (assuming 3000 cycles or so). And if you do that, that amounts to absorbing and serving about 365 twh.
The worlds entire electricity production is measured in peta watt hour (pwh). About 25 phw. or 25000 twh. So, with battery production scaling int he tens of twh per year, there may soon be enough battery to absorb and produce that kind of volume. Doesn't mean it will be used exclusively for that. But it drives the point home that we are talking some very non trivial levels of standby power. That's car batteries, home batteries, grid batteries, etc. connected to the grid. Mostly neither charging or discharging most of the time.
Tesla and a few others are starting to experiment with virtual energy providers. That might go from a few GW to having in the order of TW of power available on demand very shortly. There are about half a million tesla home batteries out there apparently. Even a few thousand of those already add up to quite a bit of collective power. Imagine that scaling to millions of connected batteries. They'd be the largest power supplier, by far, in most markets.
That won't run the grid for hours/days obviously. But it does eliminate a lot stand by capacity that is expensive and poorly utilized.
That's also the future for energy intensive industries. They'll want to generate most of their power locally (because that's cheaper) and use batteries to buffer it and fall back to the grid only when that runs out and export to the grid whenever there's a surplus locally. So, that reduces load on the grid and evens out a lot of the load spikes. And a lot of the time they'll be feeding energy into the grid. Anything energy intensive is going to compete on cost. And that means investing in local production and batteries.