Multiterminal HVDC schemes

Discussieer mee over de wereld van het grootschalig elektriciteitsnet en alles wat daarmee te maken heeft
Plaats reactie
Toliman
150 kV
Loads: 202
Aangesloten op: 28 apr 2020 01:47

Multiterminal HVDC schemes

Load door Toliman » 04 nov 2020 15:41

Most HVDC schemes are point-to-point connections, as operating multiterminal HVDC schemes is not so easy. In fact there are just two multiterminal schemes, which are in operation sincle longer time, SACOI - the interconnection of Italy, Corse and Sardegna and the Quebec - New England interconnection in North America.
How well do these schemes really work? Is multiterminal HVDC really a good alternative, where AC installations are realizeable?

Mastbeeld van deze aansluiting
Hans
Site Admin + actief in het mainsite team
Loads: 3437
Aangesloten op: 03 dec 2011 20:49
Dichtstbijzijnde trafo: Veenoord 110/10 kV

Re: Multiterminal HVDC schemes

Load door Hans » 05 nov 2020 11:12

At this moment the clear answer is no. It is only an alternative if a comparable AC-system capable of transmitting a comparable amount of power is for some reasons impractical or impossible.

Multi-terminality is more complex, vulnerable and expensive than an AC-setup with more than two stations. However, this is mainly due to the fact that even a multi-terminal DC-system is still included in an even larger AC-grid. Therefore it requires power electronics to convert DC in AC back and forth. This may sound like a bummer in considering an optimal solution, but when just checking your bank account as grid operator, it is pretty significant.

For SACOI, the multi-terminal setup is actually something that 'just happened to develop that way'. :phew:

The reason for being DC alltogether is the usual suspect: water. When a large waterbody has to be crossed, AC cannot be constructed overhead. Since water or soil have a significant larger reactive effect than air, within a rather short distance of already 50 to 100 km, AC suffers so much losses that heating of the cable (since the energy has to go somewhere) and the sheer loss of power to be sold at the market becomes a critical issue, let alone grid stability. DC is not governed by reactive behaviour and it only has to deal with Ohm resistance and some capacitive issues. This enables DC to be used within any medium, simply avoiding any reactive behaviour at all after being juiced up to the operating voltage.

The second reason in SACOI is what we just can call the coarse of history. SACOI is already an old scheme, multiple times upgraded. The wet distance between Sardinia and the mainland of Italy is shortest when crossing Corsica (France) straight over, and also the water depth is smallest at that place. Water depth is a critical factor in sinking down power cables due to their weight and tensile strength, but also in repairing when a failure occurs. So it appears that France got some money from the Italians for allowing an overhead line (cheap, easy to maintain) over Corsica, just as happened with Konti-Skan between Denmark and Sweden crossing Læsø island overhead (see here).
In the first decades after commissioning, SACOI was just a normal point-to-point link. But in the late eighties, the French must have thought 'parbleu, that is actually a lot of capacity running there and now we got short at Corsica, let's join in' and they tapped in a third station at the existing link. And since it is DC, therefore turning it into a multi-terminal de facto.

There are more reasons thinkable for cases elsewhere in the world. The point-to-point link of ALEGRO is a good example. AC is governed by loadflow and impedance: raw physics. If there is loadflow in a geographical direction (loop flows for example), only the line resistance can influence the ampacity throughout that link. By using phase shifting transformers some tweaking is possible, but this is limited. For DC, every single percent of the loadflow can be controlled in every direction at every moment. When considering Belgium and Germany, a direct AC-link between both countries will result in uncontrollable high loads if the situation goes rogue during exceptional situations.

In proposed multi-terminal projects in China, the same issues are present. When designing the DC-system as multi-terminal, load from super heavy DC links from far away in the west can be split up and dumped at the AC grid in the east in geographical different locations, spreading the power to ampacities which are within the safety and operational margins of the existing AC-grid.
Trots geboren in het 110 kV-gebied

Toliman
150 kV
Loads: 202
Aangesloten op: 28 apr 2020 01:47

Re: Multiterminal HVDC schemes

Load door Toliman » 06 nov 2020 22:16

What are the main problems at operating multiterminal HVDC schemes and how were they solved at realized schemes?

Mastbeeld van deze aansluiting
Hans
Site Admin + actief in het mainsite team
Loads: 3437
Aangesloten op: 03 dec 2011 20:49
Dichtstbijzijnde trafo: Veenoord 110/10 kV

Re: Multiterminal HVDC schemes

Load door Hans » 07 nov 2020 15:34

You can look those things up for yourself in technical publications or grid descriptions.

Try for example Google Scholar (accessible without the need for an account, other than Scopus or Web of Science) and take a look in concepts like rotating mass kinetic energy or AC grid inertia.

Let's do one short example.
Rotating mass (generator cores) are causing inertia in an AC-grid. Consider a large complex AC-grid - and I mean really large, like the full ENTSO-E Interconnected Supergrid, in which is typically 600 GW of load. Now we trip a nuclear unit, causing to drop instantly one gigawatt of power. A lot, but in the whole grid this is only 1/600'th of power in the full system. When demand exceeds production, energy is 'missing' in the system. That has to come from somewhere, so it will be stolen from the physical rotational energy of all the active generators. This will cause them to steadily slow down and the grid starts to drop in frequency. But due to the enormous scale of this grid, even shedding a full gigawatt of production is just a glimpse, causing a frequency to drop 1/10 Hz in a stretch up to about fifteen to twenty seconds.

Result: precious time is bought. Time in which we can act to stop the drop before it grows out of hand. In reality, FCC-power has to be kicked in within fifteen seconds. That is quick, but on the other hand long enough for computer systems to check, communicate, switch and kick in before the drop grows so much out of hand that load shedding would be necessary or consumers start to notice.

Such inertia is not present in DC, simply because there is no rotating mass in which energy is preserved. This makes it harder to act if there is imbalance. There is no 'vault' from which for a brief moment energy can be stolen to buy time.
For a multi-terminal DC embedded in AC, this might not be a critical drawback. But in the future, where it can be expected that more energy will be DC-generated or spiced into the grid by DC connectors, fighting the decline of grid inertia is one of the issues they already consider today.
Trots geboren in het 110 kV-gebied

Mastbeeld van deze aansluiting
arie
150 kV
Loads: 191
Aangesloten op: 30 jan 2016 19:18
Dichtstbijzijnde trafo: Heemskerk

Re: Multiterminal HVDC schemes

Load door arie » 07 nov 2020 18:16

one other thing is de switch station's
hvdc can not done with a normal switch
fore one switch you need a a lot of expensive components
link
https://library.e.abb.com/public/4571d0 ... _72dpi.pdf

the reason you can not have a normal switch
link
https://circuitglobe.com/hvdc-circuit-breaker.html

Toliman
150 kV
Loads: 202
Aangesloten op: 28 apr 2020 01:47

Re: Multiterminal HVDC schemes

Load door Toliman » 08 nov 2020 01:31

Can a capacitator bank or a superconducting coil storage in a multiterminal HVDC scheme work as power storage as the rotating masses of the generators in an AC grid?

Mastbeeld van deze aansluiting
Hans
Site Admin + actief in het mainsite team
Loads: 3437
Aangesloten op: 03 dec 2011 20:49
Dichtstbijzijnde trafo: Veenoord 110/10 kV

Re: Multiterminal HVDC schemes

Load door Hans » 08 nov 2020 20:17

Superconducting coils are even in the best case something of the future.
There might be some experimental setup active somewhere, but just the proof of concept is a far cry from a commercially attractive, mature and proven method. It might well be, but not today. So we can leap over it in considering today's HVDC multi-terminal schemes.

Again, try to look the technical things up in publications. There are literally thousands of sources and papers out there. [KCD]

Just because we can over here: let's do a back-of-the-envelope or bierviltjesberekening. :logisch:

Assume we have an imaginary island with a complex multi-terminal DC-grid without significant inertia and not interlinked to foreign inertia by a converter link to the AC supergrid. The DC-grid is there all at its own. It is loaded with (say) 300 MW of power. It is not relevant how we produce power in this grid as long as the method is static. We just assume it is balanced and it runs without any buffer. (Don't you even dare to touch anything bigger than a light switch!) Now we trip off a large solar park, separating 50 MW of production. Since there is no significant inertia and no storage, the grid will immediately collapse.

Avoiding this collapse can be done by a pack of big batteries. Static components, no startup-time. Using batteries is a common method for uninterruptable power supplies (UPS) like we see in hospitals and other critical applications. If the grid collapses and the subgrid has no own production (or just lacking sufficient mass and scale for FCC), a UPS is capable of powering the grid for a short brief in which auxiliary power units like diesel generators can be started to take over for the long run. For hospitals, the UPS installations usually have the size of half a sea container and they are capable of running the load for five to ten minutes before they are exhausted.
For our DC-island, the trip of the solar farm requires big batteries to buy significant time - and with big, I mean big. The biggest battery arrays in the world at this moment, such as Jardelund, are capable of about 60 MWh. If our island has such a 60 MWh battery (which in practice would be an absolute must just for balancing the time lags in demand and production anyway) and if it was fully loaded, it will buy us just an hour of time to desperate get 50 MW from somewhere and hook it up to the grid. That's shivering, but technically feasible.

Now assume we want to do the same trick with high voltage grid capacitors instead of batteries.
Big capacitor banks as we see them today in AC-grid are there for power quality and stabilizing purposes of the sine wave form, not in any way for storing bulk power such as batteries do. But let's assume we pick one of those stacks and use them for storage purposes anyway. We hook them up to a DC source and just 'fill them'. Now we place them at our island in the place where our battery just was and we trip the solar park again.

For this case, I did not look up the exact characteristics of such a stack of capacitors, but also with a proper guesstimation we will get the order of magnitude right. (Man, I love Fermi estimations. 8-) )
The capacitors are designed for AC, applied to compensate deformation of the sine characteristic of the AC waveform. This means, they will never have to store more energy than a single period of the AC can carry. (And probably just a fraction of it, but since we don't have numbers, I assume the worst case scenario which is the energy wrapped in 1/50'th of a second of a loaded grid.)
Now our grid is loaded with 300 MW (ignore AC or DC here and ignore the proper way by Joule as well, and assume this results in 300 MWh.) We just want to know how much energy is in the grid in 1/50'th of a second. Devide 300 MWh through 3600 and then 50. Result is 1.666 Wh.
The result is disappointing: the total stored energy in watt-hour is barely sufficient to run a single coffee machine for an hour. Even if you give or take a factor ten, or even hundred, you see it buys us barely a few seconds.

To conclude: superconducting is, just like the abbrevative nano, not the solution for all of our problems today. For instant static power backup in a big grid without inertia, batteries are at this moment the only way to go.
Trots geboren in het 110 kV-gebied

BaDu
Globale moderator
Loads: 1996
Aangesloten op: 18 mar 2014 13:55
Dichtstbijzijnde trafo: Kattenberg

Re: Multiterminal HVDC schemes

Load door BaDu » 09 nov 2020 09:41

Hans voegde de load toe:
08 nov 2020 20:17
To conclude: superconducting is, just like the abbrevative nano, not the solution for all of our problems today. For instant static power backup in a big grid without inertia, batteries are at this moment the only way to go.
Batteries will be the short term solution for storage of electrical energy, but for frequency balancing (keeping the 50Hz) there are some extra possibilities to compensatie for the loss of inertia that are being looked at the moment:

For solar generators you can electronically put more energy (by temporarily delivering more power) into the grid when the frequency drops below 50Hz en deliver less when the frequency rises again. This implies that that a solarinverter doesnt deliver maximum power but instead a few percent less, so there is wriggle room to deliver more power when needed. This might not sit well with the owner as he loses some energy delivery to the grid and thus some money.
However there are some TSO's in western europe will pay you handsomly for helping with balancing. The main problem here is that that the TSO has to be able to rely on this power being availabe when he needs it, as you know this can be a problem for renewable energy sources.

For wind gerenators there is of course a fair amount of rotating mass available when the wind is blowing. The problem is that due to the electrical construction of a windturbine this can only be made available electronically (as above), so this again requires the inverter (which is also in a modern windturbine) to behave as described above. So the same problems arise.

Great discussion!
KML-generator generator

Toliman
150 kV
Loads: 202
Aangesloten op: 28 apr 2020 01:47

Re: Multiterminal HVDC schemes

Load door Toliman » 18 nov 2020 14:32

Concerning SACOI, there is the interesting fact, that the inverters of this scheme used from 1986 to 1992 different valve technologies: the stations at San Dalmazio (Italian mainland, replaced in 1992 by Suvereto) and Codrongianos (Sardinia) used at those days still mercury arc valves, while Lucciana on Corse used thyristors.

Mastbeeld van deze aansluiting
kevelek
110 kV
Loads: 125
Aangesloten op: 05 jun 2019 20:00
Dichtstbijzijnde trafo: Hoeilaart 36/11kV

Re: Multiterminal HVDC schemes

Load door kevelek » 19 dec 2020 16:31

accidentally came across on Youtube today: https://www.youtube.com/watch?v=DY2v0fQ ... sTech.Info
Meten is Weten, Gissen is Missen. Raak niets aan wat niet geaard is en aard nooit zonder te meten.

Plaats reactie