Iberien Peninsular Blackout

Many sources in the last 24 hours quoting a Spanish government report, e.g. this from Reuters, but no link to the actual government report.  Has anyone found this?

Paul Adkins

Thanks, this makes interesting reading.

I hope this is not a stupid question: Why would disconnecting generating plant from the grid cause the voltage to increase?  

Paul Adkins

presumably because that generator is actually motoring- or acting as an inertial drag on the system.If you have light loads and embedded generation the traditional slope of the voltage drop from generator to load can be reversed  and in a sense the tail tries to wag the dog.

Its not a condition you want to persist for extended periods however.

Mike

The issue seems to be about managing reactive power flows. If the service providers who were absorbing VArs trip off, then the remaining excess VArs will raise the voltage across the system

- yes no significant real power is generated, rather absorbed at one point in the mains cycle and pushed out in another, but in effect providing a giant power factor correction at grid level.

I'm also not sure if this includes waveform quality correction here - I suspect the inverter derived wave-forms are a lot dirtier in terms of square edges  than sine-waves from rotating magnets.

Certainly both the line- and self- commutating thyristor based DC link stations generate some pretty unpleasantly rectangular current bursts and rely on other grid elements to spread what are in effect  current impulses more uniformly across the cycle. In the hopelessly simplified circuit below the upper wave-forms near the in and out terminals illustrate this. (The pale grey sine is voltage and the darker bumpy thing represents the general form of the current into or out of the grid) 

Instructions always say that a low impedance grid is needed to keep the voltage waveform from becoming more like that of the current. (sometimes expressed as harmonic content vs supply impedance)
Removing e.g. the French AC connection and leaving a large DC station on would perhaps  have been problematic in this way
M

This is the E&T magazine's summary of the report.    E&T magazine Iberian blackout

David

I note that mentions reactive power as well - but the authors have put it in quotes, as if they are not sure if 'imaginary' currents are quite as 'real' as real ones ;-)

Hehe.

I've always thought that calling things 'imaginary' because they are phase shifted to be orthogonal relative to 'reality'  a very misleading choice of naming convention. Thanks maths department.  However, I'm not sure that the engineering folk calling it 'reactive' power is much better. Perhaps the metalanguage we use needs a bit of a rethink. 'Early and late' currents or 'tidal circulation' might be clearer. We need to imply the sort of rolling pay day loan of electrons with a lot of effort but nothing useful being done.
Mike

Edited for typos. I probably should not post things quite so late and / or after a beer...

Fascinating stuff, thanks for this.

I suppose the next question is, could it happen in the UK?

De we have different/ better standards for generators to be connected to the Grid?

Paul Adkins

Probably not identical, but something rather like it certainly could happen here - 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.

An electronic set-up can be designed to emulate this to some extent, perhaps putting more current out near the voltage zero-crossing than at very the tops of the waves, or drawing current from one phase to pump up  the others at different points in the cycles, but it is not the easiest thing to get right, and does not occur as a natural feature of the mechanics, because there aren't any.  So to fake that,  it needs to be measured and acted on more or less on a cycle-by-cycle basis.

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 real situation is we have something over 100 years of running a conventional grid, and it is well understood and ridiculously reliable. The new stuff is by comparison less well known, and the rarer occurrence corners of the problem space are yet to be fully explored. 

To an extent, the rules on when things should disconnect, and how they should behave when over-under loaded need to be reviewed, and at the same time probably some odd looking extra buildings with measurement and telemetry and something heavy inside that emulates the mechanical inertia will need to be designed and connected.

 Mike.

Yes, the problem can be that wind turbines and solar all are designed to produce DC and then synchronise with the grid. But if hydro and atomic generation supply line suddenly disconnects then instability occurs and cascade tripping result in black outs everywhere.

Redesign future wind turbines to run AC at 3000/N revs would much improve stability in UK. 

In sunnier climates note that solar panels are only 25% efficient and costly in dwindling material. But using mirrors and central solar boiler 75% efficient can be obtained to heat water and run steam turbines that are mechanically connected to rotating inertia generators good for frequency control.

Mechanical generation is protected by Woodward systems which instantly check the balance of forces on the rotor bearings;      quite how you can digitally simulate that by just looking at volts, amps, and Cosine angle I am not sure nor what action could be taken to restabilise the frequency. 

Redesign future wind turbines to run AC at 3000/N revs would much improve stability in UK. 

Entirely sensible for hydropower where we control the flow. Harder for wind, as I am not sure what you intend to do to the climate to ensure a constant wind speed over time !! 

Truly synchronous wind turbines would have a very narrow range of wind-speeds with high efficiency.

Long term I think the solutions will come and will involve power electronics and more telemetry.

Mike

Redesign future wind turbines to run AC at 3000/N revs would much improve stability in UK. 

Entirely sensible for hydropower where we control the flow. Harder for wind, as I am not sure what you intend to do to the climate to ensure a constant wind speed over time !! 

Truly synchronous wind turbines would have a very narrow range of wind-speeds with high efficiency.

Long term I think the solutions will come and will involve power electronics and more telemetry.

Mike

for reassurance, look here www.neso.energy/.../reactive-power-market

Thanks for that link. Looking there & finding 

Publicly Available 1 Voltage Technical Requirements and Specification

and 

Publicly Available Stability Technical Requirements and Specification

Is actually only partly re-assuring - the  circulating current problems of waveform quality (things that are really not simple sine-waves) and phase shift are not separated, indeed reactive power is purely considered in terms of the fundamental frequency, with no mention of harmonic content
e.g.
" Reactive Power The product of voltage and current and the sine of the phase angle between them measured in units of voltamperes reactive"

For now, its good that it is being discussed but the technical language is still that of a traditional rotating grid where changes in frequency are synonymous with a load response...

An inverter based system need not present any RoCoF with load, or indeed you could design it to give a negative one, if you like.

Mike.

Electrical energy per cycle, is a volume of rotating Magnoflux inertia with voltage at right angles to current and both at quadrature to the flux or a Cosine quantum reduction will occur as VAR's.

Not sure how above Equation 1  helps resolve the problem of cubic volume synchronisation ?

Hello Clive:

What has "quantum" effects have to do with this situation?

Peter Brooks

Palm Bay FL   

I think Clive said Quadrature rather than Quantum...

   - Andy,

the guided photons have a very low individual energy and long autocorrelation function, and I don't see a virtual particle model as a very helpful way to describe wht e we have here either. I think voltages between things and currents flowing through things is the more applicable model. We just need to be very clear what we are correcting for.

Mike.

the guided photons have a very low individual energy and long autocorrelation function, and I don't see a virtual particle model as a very helpful way to describe what  we have here either. I think voltages between things and currents flowing through things is the more easily applicable model.
We just need to be very clear what we ought to be sensing and correcting for. Voltage and current Sines/ cosines shifted in the manner of KVAr are one form of correction required, flow reversals mid cycle to maintain a sine-wave shape from non sine-wave generators are quite another.

Certainly where I am now has lots of IT kit and lots of LED lights, and the voltage is quite trapezoidal. Not as bad as the experimental waveform shown here but getting on that way.

Mike.