Wednesday, October 18, 2006

New Nukes and Grid Recovery

The recent NERC report on grid reliability should focus our attention on one of the advantages of the next generation of nuclear build – “net load rejection.”

While only a very few nuclear units today can withstand a grid blackout without scramming, the newly designed plants will all be able to. Net load rejection allows a nuclear unit to automatically disconnect from a dead transmission grid, decrease reactor power to only internal demand, and keep running. Those “hotel” loads are typically 5% of maximum power and include pumps, fans, control rod drive mechanisms, etc – everything to keep the plant running. The generator keeps spinning at rated voltage but only energizing plant loads.

Once the grid is stabilized, the system dispatchers then will have the option of using the large nuclear units to help bootstrap the grid (and the tripped generating units) back up to normalcy. Restoring a dead grid from a system’s large nuclear units is much easier than doing the same from the typically very small units equipped with “blackstart” capability or outside transmission links. (Hydro plants can be an exception in being very large.) During the great blackout of 2003, all the nuclear units in the affected region tripped. It took a minimum of 24 hours to restart them after transmission power was locally available and then bring them back on line. At least one required special NRC dispensation.

The as yet unbuilt new designs (EPR, AP1000, and ESBWR) all have this feature as part of their licensing basis. New European and Asian nukes will often have it too at grid operator insistence. Some Gen III+ units offer it as an option too.

One issue is the institutional arrangements for incentizing reactor owners to invest in this feature and to use it. A nuclear reactor owned and operated as a merchant plant could save a few million in construction costs and minimize risk of plant damage by just tripping off an unstable grid. They would only lose a day or three of lost revenue otherwise. Merchant plants in general have been criticized for not being good "grid citizens." Maybe the new nukes can avoid that PR trap.

This is an issue that NERC, FERC, and NRC need to address at the regulator level. Nukes with net load rejection should be justly compensated for that service.

Technorati tags: , , , , , , , , ,

7 comments:

Anonymous said...

I'm glad to see this change.

IIRC, the Canadian units use a separate generator for hotel loads so they can remain critical even with the main generator tripped. This helped with the 2003 blackout recovery since they were immediatly available.

M Beasley

Anonymous said...

Good thing, too, 'cause if they have to shutdown they'll be down for days. They don't have enough excess reactivity to startup at peak xenon.

Farkas said...

Interesting ... "Accept a 100 percent load rejectionfrom full power to house loads without reactor tripor operation of the pressurizer or steam generator safety valves. The design provides for a turbine capable of continued stable operation at house loads. " See AP1000

Kirk Sorensen said...

Another nice thing about a fluid-fueled reactor--xenon comes out of solution all the time, so there's no xenon transient to override after shutdown. Also lets you go to much higher fluxes in the core. See page 680 of "Fluid Fuel Reactors" for discussion of this effect in an actual reactor.

Anonymous said...

I'm curious if there is more documentation available backing this claim. I have looked at the AP1000 and ESBWR FSAR available on the NRC website. The detailed information on the EPR isn't on the site as far as I can tell. For the AP1000 and ESBWR It appears that from the nuclear side the units may be able to remain critical through a 100% load rejection.

I wonder about the electrical side. The condensate, feed, circ water and many other smaller pumps need to remain powered to provide the condenser to dump to and provide SG/Reactor feedwater allowing continued operation when disconnected from the grid. When I looked at both designs' electrical systems, the normal and alternate sources of onsite power are supplied by the high voltage system in the switchyard. This is different from current units where normal power source is supplied by stepping down the generator terminal voltage. The significance of this is in that the switchyard would have to remain energized during a grid collapse if the unit were to remain online after a grid failure. This isn't normally the case. It's quite common for the generator output breaker to trip on out of step as the system fails. I'm not saying it can't be done, but the protective relaying on the HV system would have to be rethought to keep the generator connected to the switchyard. Far easier would be to have the current design with generator terminals supplying the power. Even more reliable would be a separate generator for house loads. I have heard one source that indicates that the CANDU units have such a setup.

I'm skeptical that this design aspect has been run by utility protective relaying and system stability engineers. It may have been and some smart guys in the relaying field may have come up with good solutions. That's why I'm wondering if there is more info.

As a side comment, what's with the complexity of the EPR? It seems that one of the lessons learned with the current fleet is that excessive complexity hurts. Ranch Seco comes to mind as an example. The EPR has FOUR active safety trains. SIX trains of EDG. If you want to stay with active safety systems, at least go the direction of System 80+ where the intent was simpler designs. The nuclear industry needs the fourth generation plants to be a success from the start. It doesn't seem the EPR is the way to get that result.

M Beasley

Joseph Somsel said...

The notion as I've seen it is for the house loads to normally come off a unit aux transformer that connects between the main generator and the main generator breaker (and stepup xmfrs.)

A grid disturbance would trip the main output breaker. This causes a 95% load decrease - hence "net load rejection." The house loads remain connected to the generator and energized. The reactor has to dump energy until neutron flux declines or else the pressure spike will cause a scram. This can be done with steam bypass valves to the condenser or by lifting the steam generator safeties. Sometimes, condensate and feedwater heaters have to be bypassed too.

All the grid sees is another unit offline.

The whole process is done and over in seconds.

Brian Mays said...

Anonymous asked:

"As a side comment, what's with the complexity of the EPR?"

Complexity? The design of the EPR is simpler than the current generation of plants. There are significantly fewer valves, fewer moving parts, etc. Sure there are four active safety trains, but what does that buy you? Reliability. The plant can run with plenty of margin, even if one of the safety systems has to be down for maintenance.

How does this work? Economy of scale. The EPR works economically because ... well because it's BIG. It's essentially the familiar PWR technology, which we know from over 30 years of experience works well, built to a larger scale with significant improvements (and simplifications) to the design that have resulted from the knowledge gained from these years of experience.

The AP1000 and ESBWR, while very clever designs, are essentially untried technology. (Although the GE can claim some experience from the advanced BWR's that they built in Japan.) As with anything new, I'm sure they will have a "break in" period, when the reactor designers and the operators will learn exactly how the new plants will perform. The EPR, on the other hand, builds on proven technology to provide a reliable solution, and by the time the first new plant is built in the US (whatever design it might be) there will already be an EPR that has been running in Finland for several years.

If you want to bet on a success from the start, my money is on the EPR.