Any thoughts/information on what happened? Was it a lack of spinning reserve?
Was it " The Portuguese operator, REN, said the outage was caused by a “rare atmospheric phenomenon”, with extreme temperature variations in Spain causing “anomalous oscillations” in very high-voltage lines."
as is written in the Guardian?
The Italien blackout from a few years ago had a definate cause in the tripping of interconnetors from Switzerland during a storm.
The next report from ENTSO-E on the Iberian blackout has now been published:
This suggests that the cause was a large quantity of various renewables going off line.
'Several important generation trips occurred from 12:32:00 onwards. Between 12:32:00.000 and 12:32:57.000, there was a loss of 208MW identified distributed wind and solar generators in northern and southern Spain, as well as an increase in net load in the distribution grids of approximately 317MW, which might be due to the disconnection of small embedded generators <1MW (mainly rooftop PV) or to an actual increase in load or to a combination of both. The reasons for these events are not known. From 12:32:57.000 until 12:33:18.020, major disconnection events occurred in the regions of Granada, Badajoz, Sevilla, Segovia, Huelva, and Cáceres, which resulted in an additional loss of generation of at least 2GW (the effects of frequency deviation suggest a loss of even 2.2GW).'
The reasons for for the trips are not stated.
As you say there is a lot of detail in the report. There were clearly some stability problems before the trip.
Thermal 4–Centre /South-West was meant to be online for voltage control but was unavailable due to internal problems. Thermal 5–Centre/South-West was left connected during the previous night for voltage control but was shut down on the morning of the trip.
The need for more conventional generation with power system stabilisation was recognised and a CCGT plant was ordered to start up, but the trip occurred before it was up to speed.
This is stated here:
2.6.8 Technical Constraints in Spain
“Even if there was not a remunerated voltage control service in place, power plants scheduled by the TSO under PO3.2 to solve situations of lack of dynamic voltage control receive the technical constraints remuneration (pay as bid) for their active power redispatch. In security studies conducted on 27 April, for 28 April, the combined cycle “Thermal 4–Centre /South-West” was scheduled for the entire day to regulate voltage in Western Andalusia. At 19:52 on 27 April, the unit was declared unavailable due to an internal problem, initially until 22:00 on 27 April and later extended to 00:00 on 30 April. The connection of “Thermal 5–Centre/South-West” was extended during the night to secure voltages. During the morning of 28 April, RE considered that the “Thermal 5–Centre/South-West” plant was not needed. There is no operational procedure approved in Spain where a minimum number of generation units coupled is required, and there is also no maximum limit. The criterion to decide the coupling of an additional generation unit is the fulfilment of Operational Procedure 1.1 with foreseen scenarios (generation, demand, and network). At 12:20 on 28 April, RE ordered the connection of an additional thermal power plant equipped with PSS, following the detection of system oscillations. The selected group was a combined-cycle gas plant in centre/south-west, which indicated that it could be connected in 90 minutes. At 12:26, the confirmation was issued to the power plant to connect at 14:00. Due to the blackout occurring before 14:00, this connection never occurred. In general, RE is aware of the start-up times of the combined-cycle gas plants in its control area.”
The initiating event for this cascade failure was the disconnection of various renewables as is shown in the table starting on page 103. The reasons for the disconnections are not given.
It is interesting to note that the restart was carried out using hydro and thermal generation. The renewables were not allowed back online until everything was running stably again.
The initiating event for this cascade failure was the disconnection of various renewables as is shown in the table starting on page 103. The reasons for the disconnections are not given
As engineers we should also ask WHY things happened thus a reason must be obtained so that future occurrences can be minimised or eliminated altogether.
It may well be that the problems relate to expecting inverter driven generation, usually renewable, to behave in the same way as traditional sources. That is however a problem of control algorithms and setting trip limits, rather than any intrinsic "renewable problem" with solar power, batteries, or whatever.
It maybe worth noting that private solar power is really advancing in leaps and bounds in some sunnier countries where traditionally folk have been less worried about 'first world' niceties such as frequency, voltage, pssc and so on, and the local infrastructure is generally more "flaky" anyway.
Mike
There is a fundamental difference between traditional thermal or hydro generation systems and inverter connected renewable sources, the overload ‘ride through’ capabilities. As you say here:
https://engx.theiet.org/f/discussions/31419/iberien-peninsular-blackout/150004
“there are a number of things about our grid management that are assuming that all the sources behave much like conventional generators, i.e. the frequency tries to slow down when overloaded leading to phase shifts between current and voltage that can be detected and acted upon in slow time, for a very short duration overloads can be massive > 100% without anything bad happening, and critically as here, a spinning thing can switch from motor to generator and back, to absorb energy either for several cycles or during one part of the cycle, while trying to speed up a bit and push the power station round, and then give the excess energy back slightly later, much like a mechanical flywheel.”
“And of course, overload - the jokes about transistors being the fastest fuses on 3 legs are based on unfortunate facts. True cycle duration energy storage, without moving parts ,rather than emulated, requires large inductors and/or capacitors, and at grid scale this becomes a serious and bulky engineering exercise.”
The problem may well be more than control algorithms and trip limits. We have developed AC grid systems which work very well with rotating generation and can accept an amount of inverter sourced power.
The Iberian Peninsular Blackout suggests that there is a practical limit to the amount of inverted based power that can be fed into a conventional grid. The recovery action, that was unfortunately applied too late, was to bring in more conventional generation. Renewables were the last generation sources to be bought back in once everything had been stabilized.
There may be technical solutions, but by the time you have restructured the grid and added batteries and synchronous capacitors etc. the energy and resource payback time may become infinite.
There has been a lot of capital, political and real, invested in this attempt to decarbonize electricity generation. It would be very embarrassing to many if it was not realistically achievable.
on the surface, this doesn't seem to be a mismatch between generation and demand leading to frequency collapse. The immediate issue seems to be a rise in voltage which led to things tripping off to save themselves, so it may be unwise to point the finger at any particular generating technology until we know a little more
I look forward to the next report, which promises to look at the causes and remedies
There is a fundamental difference between traditional thermal or hydro generation systems and inverter connected renewable sources, the overload ‘ride through’ capabilities.
The problem here is that that difference is by historic design that didn't like the new-kid on the block.
By defining the inverter style equipment as 'grid-failing' (sorry, grid-following) and the rapid development of larger and larger renewable based generation capabilities, it has led to the 'poor' (misunderstood) management practices for grid forming.
Those approaches used historic assumptions about what and how the systems respond.
"If I were you I wouldn't start from here", "why did you come this way anyway?"
It's going to take a while and a few bits of CPD / refreshers on the new fundamentals before this all gets sorted (if it ever does - CO2 has a long memory).
Is the steam engine flywheel the latest in technology we consider?
We have developed AC grid systems which work very well with rotating generation and can accept an amount of inverter sourced power.
Don't forget that there are plenty of examples of systems that are 100% inverter driven - e.g so called "off grid" system, where proportionally, the problem of matching supply with demand is if anything a much greater problem (e.g. one appliance suddenly switching on or off could represent a sudden change of tens of percent of the generating capacity - swings of magnitude that the national grid would never have to deal with (and would probably collapse if it was asked to). And these "mini grids" probably less diversity in terms of power factor of loads and so on as well. For sure scale can bring extra challenges (e.g. oscillations on lines that are hundreds of miles long), but I've yet to see any insurmountable challenges.
- Andy.
Is the steam engine flywheel the latest in technology we consider?
Given the recent Rail 200 celebrations it's interesting to remember Richard Trevithick's 1804 locomotive, where one reason for its failure was because the cast iron rails (i.e. the infrastructure) couldn't take the weight. However a solution came along, which was to re-engineer the infrastructure.
And of course before that James Watt had tried to prevent Richard Trevithick's development of high pressure steam on the grounds that it wouldn't work - missing the technical developments in metallurgy which meant that it could.
It's really easy to fall into a "not invented here" mindset, I'm embarrassed to say I've done it myself over the years ("digital audio will never replace analogue" c. 1990, "axle counters will never replace track circuits" c. 2000). Of course new technology brings new challenges. But as you say AJJewsbury they're often not insurmountable - in fact they keep us employed!
Not to say being cautious is a bad thing. I might be persuaded one day that autonomous cars are not considerably more dangerous overall than human driven cars - just not today. But, unless it actually breaks the laws of physics, we should be wary of considering any technology as inherently unfeasible.
In some ways it's frightening to look at what the Chinese are achieving in many areas where they don't have our ball and chain historical infrastructure.
A while back I was looking at their MMMC (Modular Multilevel Converter)
long distance transmission line that did AC-AC frequency conversion. On the conventional generation we forget just how fast we can make electronics respond and provide phase jumps as required.
In my early years I worked at NEI Parsons (Turbine generators) and a key feature I saw was the need to avoid transformer saturation during startup, called V/f monitoring, and then later saw the same effects with zero-crossing solid state switching which had a similar hysteresis based saturation issue. It is very easy to get locked into a particular technology stack, that falls over when another techno-stack comes along (Software engineers every 6-months!)
There seems to be a general agreement that to just keep adding inverted coupled equipment to a conventional grid is not going to work and will result in various stability problems.
There are certainly technical solutions available, but are they viable? If you are starting from scratch or from a very small area system there are usable solutions. If you have ‘Ball and Chain Historical Infrastructure’ (I like that phrase) then it is rather more difficult. How much time , money and resources would you have to invest to upgrade a 50 or 60Hz grid to remain stable under rapidly varying renewable feeds? Would you ever get a payback? It is often quoted that the energy payback for wind turbines is around 1 year and solar PV 2 years. This is at the connection point. If you have to add batteries, pumped storage, hydrogen systems, new grid segments, DC links, synchronous capacitors etc. would there be any payback on energy, resources or financially?
Another Ball and Chain Historical Infrastructure problem is the British Railway loading gauge. Britain was the first and started out with the smallest. Now this places significant restrictions on what can be carried on British tracks. Could the loading gauge be enlarged? Technically yes, practically no.
There seems to be a general agreement that to just keep adding inverted coupled equipment to a conventional grid is not going to work and will result in various stability problems.
There are certainly technical solutions available, but are they viable? If you are starting from scratch or from a very small area system there are usable solutions. If you have ‘Ball and Chain Historical Infrastructure’ (I like that phrase) then it is rather more difficult. How much time , money and resources would you have to invest to upgrade a 50 or 60Hz grid to remain stable under rapidly varying renewable feeds? Would you ever get a payback? It is often quoted that the energy payback for wind turbines is around 1 year and solar PV 2 years. This is at the connection point. If you have to add batteries, pumped storage, hydrogen systems, new grid segments, DC links, synchronous capacitors etc. would there be any payback on energy, resources or financially?
Another Ball and Chain Historical Infrastructure problem is the British Railway loading gauge. Britain was the first and started out with the smallest. Now this places significant restrictions on what can be carried on British tracks. Could the loading gauge be enlarged? Technically yes, practically no.
UK railways are a bad example - there is not one historical loading gauge for a start but many historical ones.
But, starting with the most heavily used routes, there is a concerted programme of raising bridges and widening tunnels to allow modern container traffic, (increasing to the euro W12 gauge is usually preferred) and enough height to permit overhead electrification with pantographs. It will however take decades, and probably like electrification, which has been on-going now for over a century and is not even 40% of the network track length, will falter on the hard bits.
Mike.
There are certainly technical solutions available, but are they viable? If you are starting from scratch or from a very small area system there are usable solutions. If you have ‘Ball and Chain Historical Infrastructure’ (I like that phrase) then it is rather more difficult. How much time , money and resources would you have to invest to upgrade a 50 or 60Hz grid to remain stable under rapidly varying renewable feeds? Would you ever get a payback? It is often quoted that the energy payback for wind turbines is around 1 year and solar PV 2 years. This is at the connection point. If you have to add batteries, pumped storage, hydrogen systems, new grid segments, DC links, synchronous capacitors etc. would there be any payback on energy, resources or financially?
We will need batteries. The good news is that they keep coming down in price. Once you have them, they aren't just a backup in case there's no wind and no sun. They can be used every day for peak supply clipping by charging them up whenever electricity wholesale prices are lowest.
In my experience the loading gauge is the least of our problems in UK railways! In 30 years plus in the industry I've never known it come up as an issue - falls well into the "nice to have". Personally I suspect that getting ERTMS properly rolled out would be a much more practicable solution to the immediate capacity problems. Which has taken 25 years so far...
the good news is that we have over 5GW of batteries connected to the GB system, plus all the Tesla power walls hiding behind the meter. There's a lot a more in the pipeline. yes, they're used to clip the peaks. they're also used to stabilise the system by responding to the frequency: the classical relationship between power and frequency still holds true
The rail gauge analogy isn't the right one. The wires aren't being changed. It's the signalling that's being upgraded as the steam engines are being phased out, and newer faster trains, with digital controls are being introduced... Let's not take the analogy too far.
a general agreement that to just keep adding inverted coupled equipment to a conventional grid is not going to work and will result in various stability problems.
I feel this is the 'Luddite' problem of not appreciating where the changes come from. When, by mandate, inverter coupled equipment is neutered as simply negative load (grid following), rather than being generational power that will support the grid, then these major grid trips will continue to happen. Yesterday's status quo was a worthwhile reference, but is not a vision for the future.
There are plenty of ways of creating grid forming 'inverter' based designs that don't have to depend on e.g. hand drawn Nichols charts and computer simulations based old analog approaches and equations.
The missing element in the discussion is the 'short term' storage of energy, and the confusion about the purpose of "Inertia" and various "Compensators". Once the system is far more 'electronic' (and with batteries;-) in its response characteristics, it will be the switch gear trip settings that will need to be addressed (as it always was).
Older simulations/models almost always fail when new technologies arrive because they have no way of integrating the new capability into the old capability approach. There are apocryphal stories of a kangaroo counting simulation firing the default beach ball weapon back at the approaching counting helicopter because they'd simply reskinned a war game simulation. Or the war games that showed that InfraRed technology had no effect on night fighting because the simulation didn't cover/allow time of night effects. Lots of hidden implementation details catching folks out.
We are seeing the same in power system design where simulations & models, essentially are not yet fit for the disruptive nature of CO2 reduction and the renewables replacement. There are still plenty of hidden issues in power systems, such as the use of 'pu' (per unit; and it's rationale), phase jumps, etc.
There's also the V/f transformer saturation limitations.
Which 'classical' frequency - power relationship were you thinking (i.e. wondering about hidden assumptions that bring frequency into the I.ac * V.ac = P.ac ?)
the classical relationship that a power shortfall leads to a drop in frequency, which can be corrected by injecting more power into the system. it's how both steam turbine governors and modern frequency-responsive battery systems work: there are smarter engineers than me out there, who can successfully integrate the old and the new
Thanks; That, for me, is the flywheel temporary energy storage of the rotating inertia.
It also shows how we can easily leap past an 'obvious' issue in discussions that can become 'untrue' when there's a technology update. We only get rotation rate drops because we were using rotation based temporary energy storage. If we aren't using rotation based energy storage we don't (automatically) get a drop in frequency.
We can also fall into the frequency-phase trap where it's unclear where we have set the reference, both locally and globally, such as : "Are we still maintaining electric synchronous clocks to the level that we used to, given most have been phased out". Locally (to a generator), we talk of both frequency changes, and phase changes, in the same discussion not realising the confusion.
We can know the destination frequency/phase before we generate, now that most lines have optic fibre incorporated (Synaptec..).
Thanks again for clarifying.
There seems to be a general agreement that to just keep adding inverted coupled equipment to a conventional grid is not going to work and will result in various stability problems.
There are certainly technical solutions available, but are they viable?
Well if you look at the GB system, NESO are procuring stability services, to provide rotating inertia and short-circuit current contribution, services which previously would have only been available from conventional generation while it was running to generate active power. These services are now being provided, typically, from synchronous condensers (i.e, conventional generators connected to rotating mass, such as flywheels) which operate without generating active power, so the stability service is now split from, and independent of, the provision of active power so you can have the stability service without running coal/gas generation, so avoiding the need to curtail renewables in order to run conventional generation for stability purposes.
There's already a number of assets providing these stability services and they have been in operation for some years now.
In parallel, there's also been a number of new frequency response type services developed to take advantage of the very fast response times of battery assets, to help manage grid frequency. I think the first one was the Enhanced Frequency Response (EFR) service some years ago, where response times were sub-1 second. There's a new suite of Dynamic services which are being introduced geared around batteries and fast responding technologies.
Going back a few years, NESO was expecting it would be able to operate the grid with 100% green energy for periods of time around 2025 onwards, presumably through a combination of these new stability and frequency services but possibly benefitting from some conventional generation (nuclear). I'm not sure about the current status of this target, but they've been putting in place the building blocks to get there.
Another Ball and Chain Historical Infrastructure problem is the British Railway loading gauge. Britain was the first and started out with the smallest. Now this places significant restrictions on what can be carried on British tracks. Could the loading gauge be enlarged? Technically yes, practically no.
Another example, closer to home, is Britain's electricity system, which itself developed in a non-standardised way with a mixture of DC systems, AC systems using a range of different frequencies, distributed at numerous different voltages. Britain was slow to standardise the electricity system, but it got there in the end, first with the standardisation of frequency at 50Hz AC as part of the development of the National Grid, which required the replacement of significant amounts of non-50Hz equipment, particularly the large 40Hz system in the North East. Then various, more localised measures, to standardise distribution to AC and to 230V over following decades.
If there's enough benefit, then eventually it happens. Or, people innovate to live with the constraints.
thanks, that's a useful expansion
being pedantic, synchronous condensers are there to control reactive power
that doesn't detract from your point about NESO procuring some big flywheels, as proposed in previous posts. I don't have the figures to hand but, at least in terms of GW connected, I think that NESO are procuring much more battery capacity than flywheels
it feels like NESO are on top of this:
a) make sure that there are enough big flywheels connected to smooth things out a bit and to preserve the classical frequency - power relationship
b) procure enough responsive generation to keep the frequency within 2% of target, to cater for credible loss of in-/out-feed
c) look to the distributors to shed load if the frequency falls below that 2% threshold, to save the wider system
that point about condensers and flywheels brings us neatly back to the original post. while the final report is yet to emerge, it looks like the issue on the Iberian Peninsula was cascade tripping on over-voltage rather than any kind of imbalance between production and consumption
We're about to take you to the IET registration website. Don't worry though, you'll be sent straight back to the community after completing the registration.
Continue to the IET registration site
