Why aren't jet pumps used as main coolant pumps in PWRs?

[rmattila]

Once again reading reports of PWR main coolant pump seal issues, my mind started to wander.. Is there some obvious reason, why jet pumps are not used as PWR main coolant pumps? This would eliminate the non-isolable leak issue and enable a more flexible pump arrangement.

 

[Hiddencamper]

First, a general talk about jet pumps. Jet pumps have no moving parts. The pumps are physically located in the reactor in the downcomers region. The pumps require a motive drive flow from another location for them to perform their active function. The pumps receive a drive flow, which then creates a jet flow through a nozzle to generate dP and ultimately core flow. The jet pumps are also sized and located such that under the worst pipe break conditions, the top of the jet pumps is about 2/3 of the top of active fuel, and prevents the core from getting fully uncovered during lbloca. (2/3rd TAF is shown to have enough submerged fuel combined with steam cooling to prevent peak clad temperature from exceeding 2200F, which is the requirement for ECCS in the Us).

That said, where does drive flow in a BWR come from? It comes from recirculation pumps. These pumps are multi staged seal pumps similar to a PWR, and require seal injection cooling to not fail while they are operating. The same concern with pump failure exists in BWRs, but I believe the consequences aren't as bad and that the seals on recirc pumps do not need cooling after they are shut down.

Now for PWRs, jet pumps add a new layer of complexity. PWRs don't have a downcomer region like BWRs do, and I don't believe the flow rates provided by jet pumps are of much use for a PWR, compared to a bwr where over 40% reactor power is tied up in jet pump flow. Jet pumps also are points of vulnerability. As they degrade and foul, they no longer make 100% core flow, they also have to be within 10% of each other to prevent vibrations and fatigue. A jet pump mechanical failure is significant and requires a shutdown. They are difficult to replace, and cleaning them doesn't offer much long term benefit.

I hope someone with more PWR knowledge can come in and give stronger reasons as to why jet pumps may not be the most compatible thing for PWR designs. I think at a minimum though is that jet pumps do not directly fix the problem of cooling pump seal failure.

 

[rmattila]

I was thinking of putting the jet pumps in the PWR cold legs, not in the RPV. That way the leakage point would always be above top of active fuel. The solution to the cooling pump seal failure issue would come from the ability to put each recirculation pump behind isolation valves, which is usually not possible with conventional MCPs due to the large-diameter pipes.

 

[Hiddencamper]

You don't get much benefit as you now increase the piping spectrum you need to deal with.

Additionally, and this is bwr specific knowledge, anyone who has been at a plant that isolated a recirc pump online can tell you it's a bad idea. The thermal transient is very likely to damage the valve, and it also precludes restarting the loop with the unit hot. You also lose any natural circulation you could get (but it prevent freewheeling the pump).

I don't believe we close our isolation valves on line anymore. In fact I think the last time we did it was the 90s.

 

[rmattila]

The coolant pump seal issue I have in mind concerns a station black-out, where the pumps have stopped. There are plant designs that can't tolerate a loss of seal injection water, because the resulting loss of inventory will prevent heat transfer to the secondary side, and all measures to get feedwater to the steam generators lose their usefulness.

Some old VVER-440:s had wet-motor MCP:s (as do the ASEA BWRs with internal recirc pumps), but I've understood they have a very low efficiency, and would be difficult to scale up (although it has been considered as one possible option in one NUREG which I'll try to locate). Some new designs have a specific stand-still seal, which will reduce the SBO risk to a degree, but ability to somehow get the pumps behind lines with isolation valves would feel tempting.

EDIT: link to the report I mentioned: http://www.osti.gov/bridge/purl.cover.jsp?purl=/10191677-Lzyqbh/native/10191677.pdf

 

[Hiddencamper]

The coolant pump seal issue I have in mind concerns a station black-out, where the pumps have stopped. There are plant designs that can't tolerate a loss of seal injection water, because the resulting loss of inventory will prevent heat transfer to the secondary side, and all measures to get feedwater to the steam generators lose their usefulness.

Some old VVER-440:s had wet-motor MCP:s (as do the ASEA BWRs with internal recirc pumps), but I've understood they have a very low efficiency, and would be difficult to scale up (although it has been considered as one possible option in one NUREG which I'll try to locate). Some new designs have a specific stand-still seal, which will reduce the SBO risk to a degree, but ability to somehow get the pumps behind lines with isolation valves would feel tempting.

EDIT: link to the report I mentioned: http://www.osti.gov/bridge/purl.cover.jsp?purl=/10191677-Lzyqbh/native/10191677.pdf

I know westinghouse has a product that they can install to basically make a pressure boundary around the RCP to prevent leakage, and several US plants have installed or have plans to install it. I'm not completely sure how a jet pump design would help here, because I don't know where you could even put jet pumps to have that kind of effect in a PWR design. The problem really doesn't have to do with jet pumps at all, it has to do with RCP seal leakage. I'm kind of surprised those pumps don't seat on their bottom seal to make a pressure boundary (the pumps at the BWR I work at will kind of 'back seat' on the outer seal while pressurized, then when depressurized will seat on the front seal to maintain a boundary). Jet pumps used in a BWR as a safety feature are only for LOCA type events, and in that case your 'natural circulation' is water boiling from the fuel. In a BWR under station blackout they serve no purpose. In a PWR under station blackout it wouldn't help, as you need to have a suction line in the RCS directly lined up with the drive flow pump, so you still have a RCP in the loop, (just not as big of an RCP...maybe that would help with seal leakage?)

I'm kind of thinking this out as I'm typing, very interesting. I still hope someone with PWR experience can come in and provide more info.

edit: Apparently the AP1000 solved this issue using a new RCP design. See the AP1000 DCD section 5.4.1.2.1 http://pbadupws.nrc.gov/docs/ML1117/ML11171A454.pdf

 

[rmattila]

Attached is a sketch of what I was thinking about. The idea is not to replicate the exact BWR application, but to use the jet pump principle to reduce the pipe size (enabling isolation valves) and to make it possible to use two smaller pumps per loop instead of one huge.

 

[rmattila]

edit: Apparently the AP1000 solved this issue using a new RCP design. See the AP1000 DCD section 5.4.1.2.1 http://pbadupws.nrc.gov/docs/ML1117/ML11171A454.pdf

So it seems. I've been told that wet motors (which have been used in BWR internal pumps since 70's) would have so low efficiency that they would not be suitable in PWR applications where main coolant pumps are larger. But apparently that problem can be overcome.

 

[Hiddencamper]

So it seems. I've been told that wet motors (which have been used in BWR internal pumps since 70's) would have so low efficiency that they would not be suitable in PWR applications where main coolant pumps are larger. But apparently that problem can be overcome.

BWR recirc motors use multi-staged seals with about a 5 gpm leakage through them. the water protects the seals by cooling them, but the water itself is not a seal and does not form the seal.

Apparently PWRs use some sort of wet labrynth seal?

 

[gmax137]

PWR RCPs use multistage seals with controlled bleed off. Some have seal injection and some designs do not. Also some have a "thermal barrier" cooler to cool the controlled bleed off before it enters the seal area.

I don't know anything about jet pumps but rmattilla's sketch seems like a lot of additional complication compared to the existing RCPs. More parts, more pipes to fail, etc.?

 

[gmax137]

By the way, what's the vessel flow rate in a BWR? In a 3400 MW PWR core it's around 400,000 gpm, at 570F and 2300 psia. That's with four RCPs.

 

[Hiddencamper]

By the way, what's the vessel flow rate in a BWR? In a 3400 MW PWR core it's around 400,000 gpm, at 570F and 2300 psia. That's with four RCPs.

holy **** that's a lot of flow.

well a BWR5, which has the highest flow rate of the BWR series, runs their recirc pumps (2 pumps) around 84000 gpm. This is just for the jet pump drive flow though. There are 20 jet pumps which then generate an actual core flow through the reactor (2 phase flow) at about 105 million lb/hr. About 15 million lb/hr of steam actually escapes and goes to the turbine. The rest is separated and mixes in with the feedwater entering the vessel, providing some reheat to that water and reducing subcooling.

edit: that's for a 3486 MWth core. 525F at 1025 psia

 

[rmattila]

BWR recirc motors use multi-staged seals with about a 5 gpm leakage through them. the water protects the seals by cooling them, but the water itself is not a seal and does not form the seal.

See http://www.tvo.fi/uploads/File/nuclear-power-plant-units.pdf page 20 for a picture of a wet(="canned") internal BWR recirculation pump.

 

[Hiddencamper]

See http://www.tvo.fi/uploads/File/nuclear-power-plant-units.pdf page 20 for a picture of a wet(="canned") internal BWR recirculation pump.

omg I've been looking for a good resource on those BWRs. Thank you so much for posting that! I've been curious about some of the differences between GE BWRs and other BWRs.

Ironically they are still using GE-14 fuel lol.