Electrical outages. cyber attacks ?

Yes, synchronous condensers may well be the way forward especially if fitted with large flywheels.

Perhaps also large batteries connected to the grid via rotating machinery rather than static converters.

There is an unfortunate effect here, where in the interests of energy efficiency, the large induction motors are now almost exclusively being fed from some sort of VSD or inverter drive, and the star-dela type starter of yore is very much on the wane. Now this tends to remove the inertial loads that do indeed switch from motoring to generating as the supply frequency dips. A VSD has no such effect  being a rectifier and some switching transistors. Actually if anything it may exacerbate an overload related dip, if the electronics tries to correct by  increasing the conduction angle, and therefore the average current, to maintain a constant horse power input, if the supply voltage droops - this is like a negative resistance, in that the current rises as the voltage falls, and is exactly the opposite of what is needed for stability.

So because of these loads that break the link between overload and falling frequency, we are far more vulnerable to the fact that much of our generation is also just inverters, this is the case on solar farms, and all wind turbines of significant size.

The upshot is that looking only at the frequency deviation may no longer be a safe proxy for state of national load and other parameters should be used instead or in addition.

I suspect the problem with large flywheels is the energy they use just to keep going, especially if they are going fast which they need to be to store significant amounts of energy.

I recollect working on a flywheel based UPS system which had multiples of 2MW generators each associated with a flywheel rotating at twice synchronous speed which was magnetically coupled to the generator / motor shaft. My recollection was that each unit consumed 50kW when not being used just overcoming air resistance. The flywheel carries the 2MW load for just 3 seconds but this was long enough to start a diesel generator.

Can't read the Times article as it's behind a paywall and I'm too cheap but it's interesting to compare the event last week with the ostensibly similar event of 2008, wherein 2 generators tripped within approx 2 minutes of each other, between them about 1.6 GW resulting in a frequency dip. Back then, a mere half a million consumers lost power from the pre-programmed load shedding, which continued for several hours... It should theoretically have been far more as less than half of the DNOs' relays operated!


If it turns out that it was indeed a similar story, then I expect the batteries currently operating under EFR, FFR and STOR contracts will have been rather welcome additional reserve capacity. I won't disagree that we've lost significant inertia but because batteries can respond very rapidly - far quicker than rotating reserve plant, and easily sufficient for the timescales of the 2008 event - they are strong contenders for the stability services National Grid has been tendering for lately for this very reason. These are competitive markets (albeit weighted against batteries due to lobbying from existing plant) so that surely demonstrates the viability of the technology.


Meanwhile, back in 2008 a large chunk of embedded generators tripped out on under-frequency as per then applicable G83/G59 settings. New settings were rolled out to HV connected plant over the last few years which will have enabled them to ride through last week's event but I don't think the roll-out to LV G59 customers has started yet; I'll be interested to see how much disruption, if any, was caused by secondary tripping revealing "hidden loads", and whether this causes Ofgem to accelerate the programme!


Lastly, in all this it's worth noting that after the 2008 event, the investigation report ended up concluding that yes, it would be theoretically be possible to contract additional (conventional) reserve capacity but no, it was not economically worthwhile to do so for what they considered a 4-5 year event (11 years ago). No supply is 100%, hence backup generators and UPSs for essential supplies! To me the question is why so much basic infrastructure (rail and road signals) didn't appear to account for this and took so long to recover, when the actual Grid appeared to work as expected and was back within the hour...

Of course a battery driven inverter, regardless of if it could support gigawatts or not, also does not have the mechanical generator characteristic of a frequency down-shift with overload. If we had a grid with no rotating generation, (and this is a thought experiment only, at least for now) we could have a constant frequency or even one that oscilatas or rises with increasing load.

Can anyone here describe how the distribution networks coordinate with the national grid in this scenario. Is there no real coordination and in the old days it didn't matter due to a lot more inertia and heaps more reserve capacity?

Try to keep it simple!

mapj1:

Of course a battery driven inverter, regardless of if it could support gigawatts or not, also does not have the mechanical generator characteristic of a frequency down-shift with overload. If we had a grid with no rotating generation, (and this is a thought experiment only, at least for now) we could have a constant frequency or even one that oscilatas or rises with increasing load.

We could, but wouldn't we need a central oscillator to keep a check on things if we have a significantly dominant proportion of asynchronous generation - ie the band leader needs to be able to set the time, fundamentally not such a problem with synchronous generation as it simply locks in to the grid frequency and all you need do is drive in more energy rather than worry about controlling speed.


The man who owns the clock, rules the world ?


Regards


OMS

whjohnson:

Can anyone here describe how the distribution networks coordinate with the national grid in this scenario. Is there no real coordination and in the old days it didn't matter due to a lot more inertia and heaps more reserve capacity?

Try to keep it simple!

As I understand it, the DNO's simply state what frequency their particular bit of the distribution effectively trips out at usually a degree lower than 49.5 Hz (along the lines of say 48Hz) - each of the DNO's report into the grid - from there is fairly automatic based on a nominal 1GW instant loss on the grid generation and consequent frequency droop -  and an action plan to shove more power onto the bars based on fast response generation (pumped Hydro, CCGT etc) if that's feasible depending on how bad the grid frequency droop is.


Regards


OMS

As has been reported, DNOs automatically disconnected about 5% of the load due to the falling frequency.

When this next happens, will the SAME 5% be disconnected or does some means exist to spread the pain and disconnect a DIFFERENT 5% next time.