Emergency Core Cooling of PWR

  • Thread starter Muti
  • Start date
  • #1
55
7

Main Question or Discussion Point

For dealing with different types of loss of coolant accident (LOCA), pressurized water reactors (PWR) have emergency core cooling system (ECCS). ECCS are simply redundant pumps to inject water into core to cool fuel, thus preventing fuel melting in case of an accident (hypothetical) that involve break of primary coolant loop. Now these have further been divided into high head safety injection (HHSI), low head safety injection (LHSI) and accumulator (based on safety analysis study to fill core before LHSI start up). Some plant call HHSI as medium heads safety injection or simply safety injection (SI) because charging pumps flow is sufficient for up to some break size. Now my question after so may years in nuclear field is how all nuclear plant standardized on this arrangement of HHSI and LHSI. I have always wonder why redundant positive displacement pumps (like reciprocating pumps) have not been used to cater to all types of break (small, medium and large break). This way plant capital cost could have been saved. Am I right can some one prove me wrong. Can some one give me an example of other arrangement for ECCS.
 

Answers and Replies

  • #2
Astronuc
Staff Emeritus
Science Advisor
18,821
2,009
With regard to standardization, please remember that 40+ years ago, there were three PWR suppliers in the US: Westinghouse, Combustion Engineering and Babock & Wilcox. In addition, each had several different designs of different capacities and power densities. The ECCS designs were determined on the basis of postulated accidents. In the intervening years, plants have been uprated and cycle lengths have increased such that the demands on systems have increased.

Overseas, various foreign corporations licensed technology for W, CE and/or B&W, and there are yet alternate PWR designs but with many similarities.
 
  • #3
55
7
"I have always wonder why redundant positive displacement pumps (like reciprocating pumps) have not been used to cater to all types of break (small, medium and large break)." By this sentence I mean why a single pair of redundant positive displacement pumps and not two groups of HHSI and LHSI centrifugal pumps (minimum 04 pumps combined) have been considered. Is there some law for having HHSI and LHSI? I have not studied, but not thoroughly, the gen 4 (I think gen 5 reactors are way different) designs and they too mostly have this arrangement.
 
  • #4
55
7
i will further add that i stick to definition of safety class as given by ANS, but peoples define safety related equipment as the one that are used during normal operation (like charging pump and residual heat removal pump), but are also aligned for emergency service. I this case clearly only SI pumps (medium head SI) are in a true sense, meant for only safety service. Based on this assumption I think is it not possible to have a positive displacement pump (with sufficient head and flow rate) to cater to all three types of LOCA (Small, medium and large breaks)? I know this will have advantages as well as disadvantages. Do some one think that this been ever studied.
 
  • #5
Astronuc
Staff Emeritus
Science Advisor
18,821
2,009
"I have always wonder why redundant positive displacement pumps (like reciprocating pumps) have not been used to cater to all types of break (small, medium and large break)." By this sentence I mean why a single pair of redundant positive displacement pumps and not two groups of HHSI and LHSI centrifugal pumps (minimum 04 pumps combined) have been considered. Is there some law for having HHSI and LHSI? I have not studied, but not thoroughly, the gen 4 (I think gen 5 reactors are way different) designs and they too mostly have this arrangement.
Please provide an example of such a pump.

Pumps have to be seismically qualified, and have the capacity. One might wish to review the requirements for the HPSI/LPSI systems. In a PWR, the ECCS must inject coolant into the primary system. If there is a LOCA (particularly LBLOCA), at some point, sump pumps may be involved.

I hope QuantumPion will interject, or anyone else with PWR experience.

http://nrcoe.inel.gov/resultsdb/SysStudy/HPSI.aspx
https://nrcoe.inel.gov/resultsdb/publicdocs/CCF/CompBoundaries.pdf
http://dspace.mit.edu/bitstream/handle/1721.1/31333/MIT-EL-78-031-04842804.pdf

http://www.nucleartourist.com/systems/eccs.htm
 
  • #6
55
7
Astronuc I will check these document first and further refine my question.
 
  • #7
1,751
818
...I mean why a single pair of redundant positive displacement pumps and not two groups of HHSI and LHSI centrifugal pumps (minimum 04 pumps combined) have been considered.
Well, you must have the two RHR pumps anyway for normal shutdown operation, so if you add two ECCS pumps then you will again have four pumps in the design.

Or, do you mean replace the RHR pumps with large positive displacement pumps? I don't know for sure, but PD pumps with that flow capacity may not be available. Remember that many (most?) of the mechanical components (pumps, valves, heat exchangers, etc) used in the nuclear designs were originally "off the shelf" items already developed for use in other applications. The high-head injection pumps are basically boiler feed pumps; the positive displacement charging pumps were oil-field equipment.

Plus, it seems positive displacement pumps are mechanical nightmares in comparison to centrifugal pumps - noisy vibrating monsters with pulsating suction and discharge pressures, and lots of maintenance required. The PD charging pumps are bad enough at 44 gpm capacity; I can't imagine the problems in scaling those up to the several thousand gpm capacity needed for RHR service.

Also, the ECCS analysis and regulations grew / evolved over time. The original thinking was that a system designed for a maximum size break would be able to accomodate a smaller break. When the details of the phenomena associated with smaller breaks were worked out, the need for high-head injection became apparent. But, the flow requirements are much lower, so additional pumps were "tacked onto" the system rather than replacing the low head pumps.

In the end, it is the overall cost that dictates the design -- cost of the pumps, the piping, the analysis. And there is the cost of delay in licensing. The nuclear field has a large inertia; proposing a system different than the already-licensed systems invites a lengthy regulatory review time and uncertainty in the outcome.
 
Last edited by a moderator:
  • #8
QuantumPion
Science Advisor
Gold Member
902
42
For SBLOCA you need a pump which can inject water into the 2250 PSI RCS. For LBLOCA you need a pump with very high flow rate at low pressure (< 600 PSI) to refill the RCS. These two types of pumps have mutually exclusive designs, there's no way to make one pump that can handle both regimes.
 
  • #9
2,170
572
To the OP: Why are you recommending positive displacement pumps? Such pumps have pulsating flow, both on the suction and discharge side, and this makes for difficult design, particularly at large flow rates.
 
  • #10
jim hardy
Science Advisor
Gold Member
2019 Award
Dearly Missed
9,839
4,875
gmax and Q'pion nailed it.
power required to pump is basically flowrate X pressure.

The flowrate that you need to pump into the system is the flowrate that's running out through the leak.
Small break by definition is a small leak, not much flow meaning pressure will stay high, so you need a small high pressure pump. Ours were a few hundred horsepower.
Large break is by definition a lot of flow, which will lower pressure, so you need a big low pressure pump that'll move a lot of water. Likewise, ours were a very few hundreds of horsepower. Basically they were irrigation pumps.



A high flow high pressure pump can certainly be built but it'd take a lot of power. Remember it's the product of flow and pressure.
You'll soon run into problems powering such a huge machine from a rapid start diesel generator as required for emergency service.
Our diesels were nominal three thousand horsepower. They had to start from cold and be ready to accept load in ten seconds. Basically they were locomotive engines.
Our boiler feed pumps were seven thousand horsepower. I don't think you could find a ten second diesel big enough to run them.

I hope that helps explain the reasoning behind different pumps for the different jobs.

Now an exercise for you - look into speed/torque characteristics of pumps, both centrifugal and piston, and of electric motors.

old jim
 
  • #11
55
7
Very well gmax. Very well old Jim. Very good explanations. I will look into speed/ torque characteristics. Beware! I think I might study all types of LOCA analysis again.... Hahaha .... My interest would be that after small/ medium break LOCA, generally after how much time operators have reduced primary pressure for LHSI operation (and can this be reduced or some modification to design may be required, and after all this would there be any benefit!).
 
  • #12
jim hardy
Science Advisor
Gold Member
2019 Award
Dearly Missed
9,839
4,875
after small/ medium break LOCA, generally after how much time operators have reduced primary pressure for LHSI operation
If the leak is small enough that HHSI and Charging Pumps can keep up with it, operators may prefer to cool down and depressurize by normal means... The less you flood containment the easier the cleanup.

As i'm just an instrument guy, I'll defer further answers to someone who's closer to current off-normal operating procedures.

old jim

PS DrD's comment regarding pulsating flow from piston pumps is spot on. The piping is excited at the pulsation rate and any mechanical resonance in the piping system shows itself by tearing pipe hangers off the wall. Pulsation dampers are a must, even on our little 50gpm pumps..
 
Last edited:
  • #13
QuantumPion
Science Advisor
Gold Member
902
42
If you want to reduce the complexity of safety equipment than you should go with an ESBWR-type design which relies exclusively on passive, gravity fed accumulators and natural circulation.
 
  • #14
1,751
818
If the leak is small enough that HHSI and Charging Pumps can keep up with it, operators may prefer to cool down and depressurize by normal means...
There is a small break size where the break flow is not sufficient to remove the decay heat. For breaks smaller than this size, the operators must maintain heat removal via the steam generators and ultimately the RHR/shutdown cooling system. For larger breaks, the system depressurizes to equilibrium with the containment pressure, and the ECCS pumps make up (replace) the inventory lost out the break. Fortunately, there is a lot of overlap in the break size that can be accomodated in either fashion, so the operators dont really have to know exactly how big the break is. The system response tells them which way to go.
 
  • Like
Likes jim hardy
  • #15
55
7
Ok I add, I am a PSA person with some maintenance experience. Nowadays my assignment is related to deterministic analysis. So, it can be said that positive displacement pumps were not used because of their limited flow rate, at high pressure. Astronuc asked about example of positive displacement pump. Well not in a true sense but atleast is example of positive displacement charging pump mostly used for hydrostatic testing of primary system. Some plants around world have modified these for "non safety class' (powered mostly by non emergency bus) emergency need. But i sure their flow rate may be less then charging pump.
 
  • #16
jim hardy
Science Advisor
Gold Member
2019 Award
Dearly Missed
9,839
4,875
I'm retired now but nostalgia is irresistible..
My former plant has positive displacement charging pumps. They've run since 1972 for charging service but take a lot of maintenance. Piston seals wear, and when they ingest air they must be manually vented. There was a hydraulic torque converter between them and their motors which both provides variable speed for flow control and allows their motors to start against reasonable torque.
We also had multistage centrifugal HHSI pumps that were run only for periodic surveillance testing. With a small recirc line they'll purge themselves of air. Recall also that a centrifugal pump requires remarkably little starting torque making things easier for the diesel generator that has to start them in an emergency.

There's a LOT of thought behind these systems. You might see if there's an old Westinghouse PWR Technology Manual around, it explains much of the basic reasoning.
USNRC's "General Design Criteria" are an elucidating introduction to the straight thinking of the industry pioneers (and Rickover).

In your maintenance experience surely you saw how interconnected all these systems are. Change anything at your own peril.....

old jim
 
  • #17
403
182
Coming from a bwr background, at what point would emergency depressurization be required and how is it accomplished?

For a BWR plant, we always say "all roads lead to blowdown", at the end of every EOP leg is a blowdown statement, and the blowdown allows you to break cool down rates and use low pressure injection systems. Generally you are blowing down to either maintain adequate core cooling, blowing down to ensure containment integrity is not challenged, or blowing down to reduce or stop an unisolable primary system leak. we blowdown by activating the automatic depressurization system, which opens several Safety relief valves or depressurization valves and venting steam to the suppression pool (basically a containment sump that already has water in it).

How do Pwrs handle this? How do they handle it with a loss of aux feed and high pressure injection? (No line break).

I'm just curious. And If anyone is interested in Bwrs let me know.
 
  • #18
I'm just curious. And If anyone is interested in Bwrs let me know.
I am interested. It's been 3.5 years since Fukushima. Are you guys know for sure you can blow down your BWRs in a prolonged SBO? Did you test it? What sort of batteries, portable generators and/or compressed air tanks will you need to do that?
 
  • #19
jim hardy
Science Advisor
Gold Member
2019 Award
Dearly Missed
9,839
4,875
I'm not current on today's ONOP's and EOP's. Will defer to somebody who is.
 
  • #20
jim hardy
Science Advisor
Gold Member
2019 Award
Dearly Missed
9,839
4,875
Coming from a bwr background, at what point would emergency depressurization be required and how is it accomplished?
How do Pwrs handle this? How do they handle it with a loss of aux feed and high pressure injection? (No line break).
When i went through ONOP course thirty+ years ago
emergency depressurization was via pressurizer by sprays or as last resort the power operated relief valves (PORV's).
of course with attention to maintain fifty degreesF of subcooling to keep natural circulation going.
so to that end i pulled the reactor pressure gauge and added to its face adjacent each major scale division the equivalent saturation temperature, making a 'poor man's subcooled margin monitor'.

If there's no line break or leak then the system isn't depressurizing itself so you don't need high pressure injection or emergency depressurization.
If there's a medium break that you cannot keep up with the approach was to make it into a larger break by opening PORVs and letting pressurizer relief tank rupture,
in other words turn it into a large break that engineered safeguards can handle via accumulators and injection. We could align the high head injection pumps to recirc as well.
If we lost aux feed (awful to contemplate) then feed with either main feed pumps or if pressure is low enough the condensate pumps.
Aux feed was so fail-safe and redundant that so long as one could make steam he could use it to run the pumps.
In the very early days we had a feedwater tie from adjacent fossil units, subsequently removed. The old-timers had put it there for last-ditch backup. It was handy at hot shutdown because the fossil feedwater was preheated which made boiler level easier to control, and especially during turbine rollup and synch.

Please understand my old memories are now painfully weak in detail and surely procedures have evolved.

getting old ,

jim
 
  • #21
403
182
I am interested. It's been 3.5 years since Fukushima. Are you guys know for sure you can blow down your BWRs in a prolonged SBO? Did you test it? What sort of batteries, portable generators and/or compressed air tanks will you need to do that?
The generic answer is that the FLEX requirements for Fukushima require that you have the capability to depressurize the vessel and use low pressure injection sources. How each BWR gets there is a little different, because it depends on what type of pressure relief system they have.

There are 3 main types of relief valves. There are Electromatic relief valves, where a solenoid is picked up and reactor steam is used to pilot open the valve, there are pure solenoid type valves where large solenoids pick up to open a reverse seated valve disk, and then there are the air piloted dikkers safety relief valves which require DC power and AIR. The first two types can function with just AC power, usually 120VAC. The third type requires pressurized air and DC power. Many/most plants have some form of pre-installed backup air system. My plant has 2 banks of bottles, either one can give us 100 lifts with the last lift having at least 7 days hold time, and the bottles automatically go into service when a LOCA signal is initiated. The dikkers SRVs also have built in air accumulators which have enough air for a few lifts. The BWR 5s and some 4s use their containment inerting air nitrogen bottle farms to also supply air to the ADS valves, this should supply hundreds of valve lifts.

This should give you enough air. I've also seen a few plants with outdoor air hookups for a portable compressor. DC power is a little harder. For my plant, we have portable DC power supplies that have battery packs installed, and we would have to open up the containment electrical penetration and hard wire them to the SRV contacts.

The FLEX procedures aren't implemented in my plant yet. We are testing them now in the simulator with live crews and finding/working out many of the questions and kinks that you tend to not see when you are just writing them at a desk. The focus in FLEX is to use your portable equipment to go from station blackout (4 hour coping time) to FLEX equipment (not on site for 24 hours). The strategies you use change depending on what equipment is available. RCIC is our preferred feed source, and the new EOPs have override orders to defeat all of RCIC's safety interlocks (run to failure) and even halt blowdown to maintain pressure reactor pressure ~150 PSIG to keep RCIC running. If we don't have RCIC, and no level indication, we would use one of those portable DC supplies to hook up to a level transmitter. If we can't determine level in a specific timeframe (there's a graph, based on last known level, time since scram), then we blow down with all available SRVs and line up to either use our on-site condensate transfer pumps (powered by portable generator) to feed through all the ECCS flush lines, or use the portable FLEX pump we bought to in. The FLEX connections allow us to use quick connects to put the FLEX pump directly into the LPCI injection headers, and we can also get an RHR heat exchanger in service for decay heat removal. None of this requires entry into high rad areas or containment, and uses all portable equipment.
 
  • Like
Likes mheslep
  • #22
The FLEX procedures aren't implemented in my plant yet.
After *3.5 years*?

Why even bother implementing any post-Fukushima fixes, maybe plant owners should just wait until license to run the plant runs out?

I see no reason to return to my former stance of supporting nuclear power. Clearly, industry as a whole is not capable of running nuclear power plants safely enough (regardless of how well individual people in it do their job).
 
  • #23
55
7
How do Pwrs handle this? How do they handle it with a loss of aux feed and high pressure injection? (No line break).
Over the years many peoples proposed passive depressurization systems most of which were uneconomical or inapplicable to gen 2 and 3 reactors. But AP600/AP1000 have depressurization system (depressurization valves dumping in IRWST) similar to BWR. May be more experienced person can shed more light on differences between the two.
Now one very important point, I think US-APWR has only HHSI (4 trains) in its ECCS. I will further probe this point.
 
  • #24
jim hardy
Science Advisor
Gold Member
2019 Award
Dearly Missed
9,839
4,875
After *3.5 years*?

Why even bother implementing any post-Fukushima fixes, maybe plant owners should just wait until license to run the plant runs out?
That's Parkinson's Law of Delay !...

I see no reason to return to my former stance of supporting nuclear power. Clearly, industry as a whole is not capable of running nuclear power plants safely enough (regardless of how well individual people in it do their job).
there's more to that statement than meets the eye.
Nuclear science is simple compared to management science.
Organizations get too big to move quickly.
and i believe the same applies to organizations, companies, industries, nations and ciivilizations.
I suppose when the Dolphins have inherited the earth they'll look back and say "Mankind was just too much of a good thing."

Western civilization is only a hundred years into the age of huge machinery.
Management science I'd estimate is fifty years behind the technology.
Jimmy Carter may have been right to put the brakes on nuclear while that science catches up.
Even Gorbachev said: "Parkinson's laws apply everywhere."
If HiddenC's plant is still working bugs out out of their Flex procedures on the simulator i understand. TMI after all was caused by dilettante tinkering with operating procedures.

But thanks for your recognition of the little people in the trenches who are doing their level best. We live downwind too.
 
Last edited by a moderator:
  • #25
403
182
I'm not trying to defend the industry or the NRC, but I think it's important to see the timeline here and understand the whys behind how long its taken to implement all the Fukushima stuff.

From a pure construction standpoint, it is taking my plant 1 year to install all the Fukushima plant modifications. We have 3 full time construction crews right now (and a full time engineering crew for just Fukushima), and have installed generator pads for a generator to be airlifted onto (with portable connections), portable pump skids in bunkered reinforced domes. New fixed piping along with equipment storage areas, all using standardized fittings so that connections can be made easily. So right now we are on track for full FLEX implementation to be complete 4 years after Fukushima. The physical construction work is taking us 1 year. So that leaves 3 years for planning, regulation, design, contracts/bidding, along with needing to share construction firms with other plants (there are limitations on how many nuclear qualified construction firms there are).

3 months after Fukushima, INPO put a "Level 1 INPO Event Report" to all plants, detailing that they had 90 days to put together their plan to deal with a Fukushima like scenario. It had some extreme limitations that were well beyond the existing b.5.b program about where you could get water and fuel from, what systems and supplies worked, what power worked, etc. Each plant put their plants together 3 months later, and the industry started formulating those plans. If there were no regulations, and this was a non-nuclear field, all of this could have been done in another 1.5 years (1 year for construction, 1/2 year for design and bidding type stuff). And if you look in England and France, the countries where the regulatory structures are more about qualitative safety factors rather than deterministic ones, those countries got their Fukushima safety improvements installed much faster.

But the NRC doesn't operate that way. They have to make a technical basis for determining what actions they feel they need to take, which get challenged on multiple levels. Then they need to determine what things they think need to be done. Then they have to make a regulation or order for it. Then they have to determine how to determine whether or not a plant adequately met the requirements. Then that has to get issued. I mean, the NRC will literally put together the paperwork to determine how they are going to determine something. And all of this adds a lot of time and overhead. Just like with the extensive damage upgrades (post 9/11, known as b.5.b), the physical security regulations and upgrades, and cyber security regulations and upgrades, the Fukushima regulations by the NRC are going to take up to a decade to fully implement. The most important ones (FLEX) are going to be done in the next year or two. This is not the industry trying to hold it up, but rather its the strict rules based regulatory process slowly trudging along. In all of these cases, the industry wanted to spend 1-2 years actually doing the upgrades and be done with it, because it costs a LOT of money to have all of these things going on for up to a decade. The industry tried to get ahead of the NRC and get their FLEX plan approved and get it done early on, and the NRC was not moving at that pace. The NRC didn't even fully determine what the FLEX requirements were until late 2012/early 2013 I believe. Until early to mid 2013, the NRC was still trying to figure out if FLEX should be just severe site damage, or if it should also include extended loss of AC power (ELAP).

Think about that, it took the NRC several years just to decide if post-Fukushima requirements should have included ELAP.

Even if the US nuclear industry was ready to build all the FLEX upgrades in 2012, the NRC wouldn't have known how to even regulate it. Until the NRC issues their regulatory guides and standard review plans, the industry is very likely to be told that any upgrades they make are no good, causing substantial rework. The way nuclear plants in the US work today with the regulator, is they want to know exactly what the NRC is looking for before they issue a license amendment request, because they do not want to go back and re-do stuff. (Lets also remember major plant stuff like many Fukushima changes cannot be physically installed until approved, because they involve potentially significant changes to the plant's licensing basis). The industry tried to get FLEX moving early on, but it didn't, and then they basically slowed the pace down until the NRC figured out how they were going to be dealing with it.

Anyways................moving on to EOPs (emergency operating procedures).

You do not just change EOPs or make changes to your emergency response. When you are in EOPs, they supersede many of the conditions of your operating license, and authorize or even require you to defeat safety functions, perform high risk evolutions, use plant systems in ways they were not designed for, in order to prevent core damage, preserve the integrity of the containment, prevent unmonitored/unfiltered releases, and stop release rates to the public. Some of the EOP changes for FLEX are actually pretty extreme, the biggest one being that we are now required to halt an emergency blowdown to maintain RCIC injection capability, and maintain the reactor pressurized (albeit at a reduced pressure). But because of this directive, it puts us into a position where we are going to be making an active choice to potentially destroy the containment in order to preserve the reactor core. This required extensive engineering analysis, testing, risk analysis, all to demonstrate that this is in fact the best action to take with respect to the health and safety of the public. This is just one item of many many items in the FLEX EOP changes that is fundamentally changing the way we respond during an ELAP.

Needless to say, getting EOPs approved takes months to years. There's a reason my plant has only had 3 revisions to our EOPs since we started up, because the engineering and analysis behind them is rock solid (thanks to the work upfront), and because they are very hard to change. The last thing anyone wants is a flawed EOP that actually causes a more dangerous situation than you originally had. Right now we are working out the kinks in the new procedures before we get them approved. (For comparison, the procedure on starting up the reactor has been updated over 300 times in the plant's life).

That's my wall of text.
 
  • Like
Likes mheslep, QuantumPion, Astronuc and 2 others

Related Threads on Emergency Core Cooling of PWR

Replies
8
Views
2K
  • Last Post
Replies
2
Views
2K
Replies
0
Views
866
Replies
2
Views
1K
  • Last Post
Replies
2
Views
3K
Replies
3
Views
2K
  • Last Post
Replies
4
Views
4K
Replies
5
Views
5K
  • Last Post
Replies
4
Views
4K
Replies
3
Views
946
Top