Why are Power Grids so Large? 5

[Mbrooke]

Why are power systems so large and interconnected? For example, what technical obstacles prevents the US eastern interconnection from being 8 isolated islands? Why not separate them by ISO/RTO? Why does every power grid in the world strive to be as large as geography allows?

 

[crshears]

I'm no ex-nuke, but if I understand it correctly it has something to do with reactor physics and xenon poisoning; we have participants here vastly more qualified than myself to explain this, so I'll say no more.

As my shift winds to a close and I'm not back until 2 Jan 2020, I wish you all the best in the New Year.

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[Mbrooke]

Please post more when you come back! Highly educational and just the right guy for the topic.

 

[waross]

Why are Power Grids so Large?

The power grid must be larger than the greatest load.
Interconnecting adjoining grids into one larger grid often gives a better diversity factor.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[crshears]

Please post more when you come back! Highly educational and just the right guy for the topic.

Happy New Year, everyone!

So: does this post qualify as more? If not, what did you want to know?

I happen to be on shift with two ex-nukes today, so if nobody else offers a rejoinder to Mbrooke's query regarding the aversion of nukes to maneuvering their power outputs, I'll ask them for their thoughts once the morning flurry dies down.

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[DM61850]

I think that a better question would be "why are power grids so small?"

With small grids, you get the benefit of segregating problems. Cascades for the most part can't carry over due to the dc ties between regions usually being very small. In the U.S., we have three grids, the Eastern and Western Interconnects, and ERCOT (most of Texas). Any of these would really benefit from connecting with their neighbors and all of them have AC ties that they can close in to restart their grid. By Houston, there is a line that goes off into the eastern interconnect that is left open between the two intectonnects but if either side goes down, they can use the other side to help bring the system back up. On the eastern side of the Western Interconnect you have issues related to the fact that a portion of the system is relatively week since it is sandwiched between the Rocky Mountains and the Eastern Interconnect, which only has very weak ties.

The main reasons that I believe they are not larger is due to:

1. Politics
2. The feeling the need for control
3. Different control philosophies
4. It would take a lot of work to connect two interconnects physically and policywise. I could see that being a 40 year project.
5. Distrust that your neighbor is not going to bring you down. If the grid is operated correctly and problem areas are segmented quickly, this really shouldn't be an issue. FirstEnergy had all the time in the world to stop the 2003 blackout.
6. Making it easier for power to feed more loads can raise the price of electricity. People living by a coal plant don't like it when their power is basically being sold to people in another state.


Pros for having larger systems

1. Easy transfer of power between regions without conversion equipment
2. Sharing of generation reserves for dispatch or contingency.
3. A more economical dispatch of generation.
4. More consistent regulations
5. Better systemwide planning. There is a lot of wind generation produced in the midwest that goes to the coasts. A single grid would allow power on the western side of the Eastern Interconnect to flow west and vice versa. As it is now, there are pockets of captive generation due to transmission congestion.

 

[crshears]

Why can't nuclear plants follow load well? What would have to be done to them to do so?

Modifying what Keith posted earlier in this thread, nuclear plants can realistically be described as consisting of a nuclear reactor paired with a control scheme, serving as a heat-releasing source used to generate steam, which is then fed to a steam turbine driven generator. As such, they consist of a heat [reactivity] regulator "that happens to spew electrical power that can be used as long as it's connected to a large system that can regulate around its essentially non-load based output."

I know that not all nuclear power plants are created equal, but I'm being told that in my neck of the woods, for reactor safety, the trip settings for each of the various zones within the reactor are adjusted manually, by a painstaking and laborious process that satisfies the unit operating licence provided by the Canadian Nuclear Safety Commission [CNSC]. During unit start-ups, a Protection and Control Specialist is constantly adjusting the reactor trip points in such a way as to just stay ahead of the prevailing reactor output, and this work continues iteratively until such time as the reactor reaches full and stable output, at which point the P&CS can be released to other duties. Any reduction [or subsequent increase] to unit loading requires the P&CS to return to the unit and iteratively adjust the trip settings while the unit ramps up or down, and remain in attendance for a time until stable steady output is again achieved.

I'm also told that during start-ups, shutdowns and loading ramps, be they up or down, units suffer from thermal cycling of the reactor's fuel channel tubes, shortening the reactor's service life and hastening the time at which a reactor shutdown for fuel channel metal sampling [ which literally involves taking scrapings of numerous tubes at discrete locations within the calandria ] and, eventually, actual fuel channel re-tubing, will be required. Not only that, but at loadings other than full, differences in reactivity within the various zones of the calandria can cause control system instability and, if not caught in time, a reactor trip.

All of the foregoing applies to CANDU units; I would however not be at all surprised to learn that nuclear reactors of any other design would likely be subject to many of the same strictures.

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[Mbrooke]

Yup, see that in practice everywhere. Nuc units used mostly for base load. Typically I've seen over 1000MW units feed into a 345kv or 500kv system far away where near metropolitan areas its stepped down to 230, 138 or 115kv. Power for peak load comes from local fossil plants connected to the local 230/138/115kv transmission system.


Though it would be nice if everything was nuclear.

 

[DM61850]

I think the comment of connected to a "large system that can regulate around is essentially non-load based output" would be clearer if the reactor was just referenced as a base unit. Base units typically are run as close to 24/7 and near max output. They are your cheapest form of generation. Even then, you cheapest units will be backed off from rated output for security reasons so that there is enough margin that can be brought online quickly if something happens like a generating unit trips out. You don't have much system flexibility if near all your units are running at 100%.

 

[crshears]

Hello DM, our nukes do in fact run full out, as there are numerous other units available to provide spinning and ready reserve, since we currently are not at all tight when it comes to generating resources; indeed, the net power flows from our province into the US quite routinely run well into four figures, demonstrating that we have lots of juice available to export...

I agree, base unit is the correct name for these nuclear generators; I'd sort of forgotten that term, as I haven't used it in some time, load / generation balance not having been one of my accountabilities for a number of years.

Note however that the Saunders and FDR generating stations mentioned previously will operate as either base load, shallow peaking or deep peaking plants, depending upon the prevailing river flow.

As to the OP that started all this discussion, the larger the interconnection is electrically, the more mass it has to provide the stability to ride through its Most Severe Single Contingency [ MSSC ].



CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[Mbrooke]

So I want to ask, what would it take to design a smaller system? New Zealand and other Islands are on the small side. How do those work?

How infeasible would a bunch of islanded nucs be?

 

[DM61850]

I think what you are asking is basically what the united states was like prior to complete electricification. Some of the rural co-ops operate still in a similar manner of taking care of themselves. Larger utilities with strong systems take care of themselves and all their small neighbors just hang on and often are parasitic. Some small municipalities take care of themselves or have generation on site for when they lose non-firm power.

 

[crshears]

How infeasible would a bunch of islanded nucs be?

Quite, I'd think; there'd be no convenient way of ramping the reactors' power outputs up or down, so unless one or all of the units had condenser steam discharge valves to dump excess steam direct to condenser, there'd be no quick way to reduce the output of a steam turbine without having the boiler safeties blow, which would lead to a major loss of condensate, and if the water treatment facilities couldn't keep up with the loss, there'd be no choice but to shut the entire place, or at least some units, down due to lack of demin water, which could mean customers either left black as a stack, or subject to rolling blackouts, for an indeterminate length of time.

New Zealand and other Islands are on the small side. How do those work?

A co-worker took early retirement and went to work in NZ as a power system senior supervisor. While he was getting his ducks in order and before he left, we had time on some night shifts to chat about what he would be doing there, and what their grid is like [ he had been flown there a time or three ]; what little I offer here is based on my recollection of those conversations.

NZ's South Island is generation heavy, much of it hydraulic, while the North Island is generation deficient. As a consequence there is one great honking undersea transfer cable that sends excess SI generation to the NI. This can pose some issues, as having a power system that somewhat resembles a barbell can be problematic due to instability sometimes developing between the two islands, causing nasty flow fluctuations in the cable, the mitigation details of which my co-worker began to learn about but was not at liberty to divulge...

If I recall it correctly, their power system is integrated, meaning not divided up into generation, transmission and distribution entities, which simplifies considerably addressing any issues encountered.

 

[Mbrooke]

What if you were to use steam dump valves? And how would they respond to one turbine trip putting more load on the rest of the units?

 

[crshears]

That would depend considerably on the way the control system was configured...

The "steam dump valves" are what I was referring to as Condenser Steam Discharge Valves [CDSVs]. Note that the duty on these valves is very severe, and designing one to have a long life is quite the challenge. An additional aspect of harnessing these is that a design decision must be made as to what percentage of the reactor steam output can be dumped; the ex-nuke I'm on shift with says the ones he knows of are capable of dumping 60% of the reactor's full steam production capability, and these have been used in the past when the associated turbines were taken off line for a few hours overnight in the spring and fall seasons due to Surplus Base Generation. Returning the turbines to service the following morning as the load comes back in is quite a careful exercise, as there can be issues of turbine/casing differential expansion to address.

Off again for four days...


CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[Mbrooke]

Wow

I'll ponder some more and ask questions in 4 days.

 

[Mbrooke]

@DM61850: pretty much- yes. I've always wondered how those small grids survived, where a trip and re-close could mean a sizable percentage of system load dropped and then picked up.

 

[rcw retired EE]

M Brooke- Some small plants can also island and pick up radial loads. About thirty years ago a severe snow & ice storm took out 60 kV lines in the Northern California mountains putting the area around Susanville in a blackout. 20 miles from Susanville, the 30MW Honeylake geothermal and wood-fired power plant went to island mode when the 60 kV tie breaker tripped. The "sky valve" dumped excess steam until the single generator was stabilized carrying just in-plant parasitic load.

PG&E couldn't re-energize Susanville for a week. The Honeylake Plant Manager didn't like the idea of his home freezing. So he contacted the Lassen Municipal Utility District and together they developed a plan to pick up load in blocks from the power plant using distribution switches and cutouts. They bypassed some synch check interlocks to pick up the 20 mile 60 kV line and energize the 60 kV to 12(?) kV distribution transformers in the Susanville sub with all feeders open.

Communicating via radio, plant operators increased fuel to the boiler as LMUD line crews added each load group. The turbine ran in isochronous mode with minor frequency excursions. Operators adjusted excitation/voltage as needed. After many hours, most loads in the town were restored. The plant supplied 15 MW in isochronous mode for several days with the operators manually trimming fuel input as loads changed. This style of islanded operation was not in the original design but it worked to maintain power until PG&E lines were repaired. As long as load fluctuations were minor the HP steam bypass valve could dump excess steam to the condenser without lifting the sky valve and wasting demin water.

I don't recall if they had to dump load to reconnect to PG&E. The design didn't have any synchronizing equipment at the Susanville sub other than synch check functions in the SEL line relays for the breakers to PG&E.

In normal operation, the 60 kV line must be energized before the plant can connect. When the 13.8 kV generator breaker and the 60 kV line breaker are both closed, the turbine governor switches from isochronous to droop mode to operate in parallel with PG&E's other generation.

This anecdote may not provide any technical information on large grid, small grid, or island operation, but I like telling the story. I had a small part in the control and protection design.

 

[Mbrooke]

What is isochronous mode?

 

[itsmoked]

That means the frequency is maintained at 60 or 50Hz by A generator regardless of the load. It's normal mode for running an islanded generator. Otherwise droop mode provides only a approximate load related frequency.

Watching a generator run in isosynchornous is fascinating. Block loads are merely 100ms anomalies in engine speed.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[MatthewDB]

The reason nuclear plants are run all out is economical, not technical. The issue is that fuel purchase and disposal costs amount to less than 10% of the total annual expense. Once construction and decommissioning is added in, the variable cost of producing power drops to ~5-7% of the total cost. Having the power plant ready to go and not producing power is very expensive. The way to look at it is that it is 95% as expensive as saying "I know the wind is blowing and your turbine is ready, but don't turn it on" or "I know it's a nice sunny day, but don't turn on your solar inverter."

The one technical hurdle is that none of the designs support a drop to under 20% after high power operation right away and never operate under 10% due to xenon instability. If takes maybe 6 hours to burn through the xenon to allows for the descending from 20% to 10%. All North American designs have a steam dump with capacities over 50% and enough to survive a load dump. (A few seconds to a few minutes of dumping steam to the air before the reactor output drops to the capacity of the condenser steam dumps). Realize that a nuke burns somewhere between 5% and 10% of power to operate the equipment so the amount of steam to dump during non-output isn't the full 10% minimum power operation.

Boiling water reactors (BWR) are the most agile. They use core flow rate to adjust power level above 20% output. More flow removes the voids (steam bubbles) which increases the moderator effect and raises power output. Effectively, the amount of steam in the core is a constant, and the faster water is pushed in, the faster the steam is made to keep that steam / water ratio constant. BWRs are second only to hydro in power ramp rate when the BWR is in the 20% to 95% power band.

Pressurized water reactors are a bit slower, but France has proven that they pretty good too. Since they're 80% nuclear, they have to meet the daily load fluctuation using nuclear. Their Framatomme design is capable of 5% per minute. That's on par with an industrial turbine and beats a supercritical coal plant.