I can speak for the GB case. Low Frequency Demand Disconnection (LFDD) occurs automatically and in stages when the frequency drops until it stabilises. The substations or feeders that are tripped off are not currently determined by real-time metering - instead they are pre-allocated based on their typical demand. This means that the system operator does not really know how much demand will be disconnected at any given time. If it's sunny, you could easily trip off a lot of solar generation connected on the low voltage network, causing the frequency to drop further. It is far from optimal!

They don't really disconnect it like that in these circumstances (as in they choose what to disconnect). As far as I know generation plants will start disconnect when grid frequency is <49Hz or >51Hz (at least in the UK) automatically as it's all starting to go very wrong. This is what causes this huge cascade effect. Roughly speak less frequency means there is more demand than supply and the other way round for higher frequencies.

This has changed a lot though, as even home batteries afaik will start discharging if they start noticing the frequency dropping to provide some support on generation. But if it's dropping too fast and too quickly it won't help.

But yes they do have very granular info on all the HV sources and how much load is on them.

Load shedding is an active measure taken when the grid operator knows that it will not be able to balance supply and demand. This is what happens in South Africa, where the operator preventively disconnects parts of the network at predetermined times to ensure network balance. As this is programmed, it is possible to rotate the areas under blackout.

In this case, we are dealing with a widespread grid incident. The various grid protection mechanisms have been triggered to prevent interconnection overload. In addition, the generators are trying to correct the grid frequency to exactly 50Hz. At 49Hz, more power must be generated; at 51Hz, less power must be generated. However, if the frequency varies too much, there are also protection mechanisms to prevent the turbines from overspeeding or amplifying frequency variations.

The grid is complex, and normally this type of incident is limited to one cell of the electricity distribution grid. A blackout is a domino effect, when a minor event triggers a chain reaction that disconnects more and more elements from the grid.

Th grid operator will have to restart or reconnect the power plants one by one, restore power to stations and sub stations. All of this must be done in a specific order before power can be restored to consumers. All of this takes time, requires resources (you need men on the ground), and the slightest error can lead to further outages.

Some consumers are prioritised, such as hospitals, transport infrastructure, telecoms and water networks. Many critical pieces of equipment have UPS systems, but these are not always designed for such long outages or have not been tested for years. There are patients with home equipment who will struggle.

This is why rotating load shedding is preferable. The outages are not too long and vital infrastructure is not affected (or less so).

In my area, "buckets" of average load are prepared ahead of time in coordination between the transporter (who gives the order to shed), the distributer (who implements the shedding operation as close to consumers as possible), and representatives of the state (who define priority areas that can't be load-shed, or that should be load-shed last).

When yhe time comes, you just shed enough "buckets" to stabilize. Load shedding is not a precise task, when you're at that point, you'd rather load a few more megawatt and be safe, than play with the limits and be sorry.

> So how do grid operators know what amount of load will be cut if they disconnect point A123

Opening a line within the grid isn't used to shed load, as the grid is mainly redundant. It's used 1) to protect the line itself (either by letting it trip, or by opening it preventively), or 2) to force power to flow differently through the network, by modifying its impedance. Opening a line from the transport network is not a way to load shed.

In this totally random example [1], opening the line A-B increases impedance in the right part of the network, forcing power to re-route through the left part, and reduces the load on another line in the right path that was overloaded.

[1]: https://excalidraw.com/#json=5l8OS96Wdke6l9YEClQt8,NJ-r2PtiE...

> Or is there continual measurement? (e.g. there's XYZ MW of load behind A123 as of 2:36pm)

The network is, indeed, monitored continually, and we do have potential network equipments failure simulations every few minutes, with contingency plans also simulated to make sure that they work in that particular case. This way, when shit hits the fan, we at least have a recently tested plan to start from.