Looking for specs on Failed Fuel Detectors on NPPs in US

[mesa]

Hello, trying to get some specs on the FFDs typically found on a US reactor. I was told they run NaI scintillation crystals and measure just Bq's (which seems a little odd) with a follow-up with the Hot Lab once a week (at least for the reactors close to my home).

Seems a bit simple to me and the NaI detector doesn't seem the best fit, but many of these plants do run some pretty old tech and have their reasons for doing things the way they do, so it wouldn't be too surprising.

Any new information would be appreciated!

 

[Astronuc]

Each nuclear reactor is somewhat unique, so there really isn't a typical failed fuel detection system, and most utilities and suppliers don't put out details in public. Systems will vary, but the methodology is basically the same.

High Purity Germanium (HPGe) detector systems are common these days, and have been for some time. One example, Fuel INtegrity Evaluation and Surveillance System (FINESS) from about 1999/2000 for BWRs. Note that the detection system is set up near the steam jet air ejector system.
https://www.ortec-online.com/-/medi...tegrityevaluationsurveillancesystemfiness.pdf

PWR operators have to pull pressurized coolant sample from the primary system through a long sampling line, but the detection system is more or less the same.

 

[artis]

I wonder how do they tell which fuel assemblies have cracked whenever they spot increased radioactivity or the presence of fission elements within the cooling water, I would suspect one can only tell whether there is a leak in the core but not a specific place, so do they then have to physically open up the core and search for the faulty fuel assembly/rod manually?

 

[anorlunda]

I wonder how do they tell which fuel assemblies have cracked whenever they spot increased radioactivity or the presence of fission elements within the cooling water,

That may be true. But even if you could tell which fuel pins were leaking, what could you do with that information while the reactor is still operating?

The remedy is to remove the leaking fuel from the reactor. That requires a shutdown.

 

[Astronuc]

I wonder how do they tell which fuel assemblies have cracked whenever they spot increased radioactivity or the presence of fission elements within the cooling water, I would suspect one can only tell whether there is a leak in the core but not a specific place, so do they then have to physically open up the core and search for the faulty fuel assembly/rod manually?

With CANDUs, one can monitor coolant activity on the individual out-feeder lines, since groups of fuel assemblies are located in individual pressure tubes. CANDUs also use online refueling, fuel can be removed while the reactor operates, and the individual assemblies are then sipped.
https://www.nrc.gov/docs/ML0334/ML033450220.pdf

BWR operators use 'flux-tilting' in which the reactor power is reduced to around 60-65% and groups of control rods (or one-by-one) are inserted, and changes in coolant activity, notably the offgas (sum of 6 or 7: ^85mKr, ⁸⁷Kr, ⁸⁸Kr, ¹³³Xe, ^135,135mXe and ¹³⁸Xe are monitored. When a control rod is inserted next to the fuel assemblies in the cell (4 assemblies) the power decreases and activity release decrease. Furthermore, one looks at ratios of pairs of radionuclides, usually the short-lived to long-lived, and the ratio can indicate a tight or open leaker. The ¹³⁴Cs/¹³⁷Cs ratio is also checked as it correlates with burnup. The reactor must be shutdown and fuel retrieved in order to confirm failure.

With PWRs, it's a matter of looking at coolant activity and Cs-ratio to estimate burnup. Ulimately though, the reactor must be shutdown to retrieve the failure.

When LWRs shutdown and the vessel head (and hardware) are removed, fuel handling (refueling) machines are use to remove the assemblies. In the early days, 1990s and earlier, individual assemblies were places in cans/canisters, to which were attached gas sampling lines. Inert gas was pumped into the canister, and a gas sample was transferred to a gamma detector. Elevated activity would indicate a failure. Full core sipping might be necessary which can take days in a large reactor.

Starting in the 1990s, fuel manufacturers (frequently the reactor designers) developed in-mast sipping systems, in which the telescoping arm of the refueling machines were fitting with gas sampling lines. As an assembly is retrieved vertically from the core, a change in pressure with elevation causes fission gases to be released from the failed fuel.

Once an assembly is identified as containing a fuel failure (or failures), the fuel is either deposited in the spent fuel pool or moved to an inspection stand where it is visually inspected. If the fuel is 'reconstitutable', i.e., the one or both end-fittings can be removed, the failed fuel rod may be removed, IF the failed rod is not severely degraded (hydrided). Otherwise, all of the good fuel rods can be removed and placed in a new skeleton (if one is available). If a failed fuel assembly is not reconstituted, then a used assembly with similar burnup may be retrieved from the spent fuel pool and placed in the core.

If a failed fuel rod is removed, then an inert stainless steel rod is placed in that location, and the assembly can be returned to the core. With respect to the reconstitution, economics drives the utility's decisions, i.e., how much downtime is acceptable.

 

[artis]

@Astronuc thanks for the info, it seems channel type reactors like CANDU and the RBMK have an advantage in this regard, although do they really have radiation monitors on each channel? Because even if you can "stop" an individual channel which you can before you do that you have to know which channel is leaking and to do that under reactor operating conditions whereby water is flowing through under pressure one would need a detector either within or downstream of the channel, are you saying they do have individual detectors for all the hundreds of channels?

I wonder whether there is a method whereby you can spray or clad the individual fuel rods within the assembly before insertion with a substance that for example glows , fluoresces or changes some visible detectable feature upon contact with one of the fission products that would escape if there was a crack/leak, a reaction with the escaping gas for example. In this way broken rods could be located easier.

 

[anorlunda]

I wonder whether there is a method whereby you can spray or clad the individual fuel rods within the assembly before insertion with a substance that for example glows , fluoresces or changes some visible detectable feature upon contact with one of the fission products that would escape if there was a crack/leak, a reaction with the escaping gas for example. In this way broken rods could be located easier.

A reactor owner must be extremely cautious and conservative with water chemistry. The reason is that even single atoms that come loose and are carried by the water might carry radioactivity to undesirable places.

50 years ago, there was an infamous case where GE used abrasion resistant stellite buttons on the ends of their control rods. Those increased the practical life of the rods. But stellite contains cobalt, and cobalt atoms became radioactive cobalt isotopes that got carried around and contaminated the whole plant. It took many years to track down that source of contamination in GE plants.

In other words, a leak detection remedy could itself create the same contamination problem that leaking fuel causes.

 

[Astronuc]

@Astronuc thanks for the info, it seems channel type reactors like CANDU and the RBMK have an advantage in this regard, although do they really have radiation monitors on each channel? Because even if you can "stop" an individual channel which you can before you do that you have to know which channel is leaking and to do that under reactor operating conditions whereby water is flowing through under pressure one would need a detector either within or downstream of the channel, are you saying they do have individual detectors for all the hundreds of channels?

I'm not familiar with RBMKs, but as far as I know, the CANDU operators move a portable system adjacent to a cooling channel. I don't know if it is mounted on the fuel handling machines, but I expect so, and I can contact someone familiar with the operation to confirm.

I wonder whether there is a method whereby you can spray or clad the individual fuel rods within the assembly before insertion with a substance that for example glows , fluoresces or changes some visible detectable feature upon contact with one of the fission products that would escape if there was a crack/leak, a reaction with the escaping gas for example. In this way broken rods could be located easier.

No, not really. It would be impossible to place a 'camera' or visual system in an operation core, especially in a cooling channel, nor would want to do so.

Firstly, there is an abundance of 'bright blue' Cerenkov radiation, not too mention an abundance of beta radiation, X-rays, gamma rays, and neutrons. Secondly, an probe would be blocked by the end-fittings and spacer grids. Thirdly, inserting a probe into the cooling channel would starve cooling flow from the adjacent fuel rods and potentially cause failures.

 

[mesa]

Each nuclear reactor is somewhat unique, so there really isn't a typical failed fuel detection system, and most utilities and suppliers don't put out details in public. Systems will vary, but the methodology is basically the same.

Indeed, and most of the information I found so far cames from retired techs, especially the INC guys.

The FINESS system is pretty nice, and exceeds what would be required for pulling a gamma spec on the primary loop of a typical PWR when looking for fission products.

We are running some tests on an old NaI detector to mimic what is installed on a PWR to see what the resolution looks like going in through the side (overall quite good), through a thick plate of steel and through a thick plate of steel under a few inches of water.

So far it looks like we can pull decent enough gamma specs for early detection of FPs with just an electronics and software upgrade to the current systems. The water and thick steel tests are coming online shortly, fun stuff!

Thanks for the info.

 

[artis]

No, not really. It would be impossible to place a 'camera' or visual system in an operation core, especially in a cooling channel, nor would want to do so.

I was thinking for it to ease finding the faulty rod once reactor has been shut down and vessel head opened up, not during operation, but then again if you done the pain of removing the head you might as well go through with a detector on each assembly I guess.

 

[Astronuc]

I was thinking for it to ease finding the faulty rod once reactor has been shut down and vessel head opened up, not during operation, but then again if you done the pain of removing the head you might as well go through with a detector on each assembly I guess.

There has been a lot of effort to develop techniques to look for failed fuel. Radiation hardened underwater cameras have been employed, but sometimes it is difficult to peer along rows of fuel rods, and some tight leakers can be hard to detect. There are various ultrasonic based systems that pulse UT signals through the cladding with the expectation that the failed fuel rod will contain water, which will attenuate the UT signal. One problem is that high burnup fuel may have pellets that bond with the cladding giving a false positive signal, or a water may not have water insight, if the defect is 'tight', in which case one obtains a false negative. One can also use eddy-current technology (ECT) with encircling coils to detect discontinuities (defects) in cladding. That's not perfect either.

One challenge for failed LWR fuel is the potential for secondary degradation due to hydriding of the Zr-alloy cladding. If severely hydrided, the cladding may fracture (if not already) during retrieval and pellets fall out. I have seen many nasty pictures.

I spent the better part of two decades studying failed fuel, especially degraded fuel, and fuel inspection technologies. Most of my work involved fuel behavior, fuel degradation and ways to address it during operation. Other work involved mitigation fuel failures, i.e., preventing fuel failure in the first place.

 

[mesa]

I was thinking for it to ease finding the faulty rod once reactor has been shut down and vessel head opened up, not during operation, but then again if you done the pain of removing the head you might as well go through with a detector on each assembly I guess.

One possibility for in situ measurements would be to try to utilize the Rhodium detectors. The CE design had 61 detector sets in 5 layers, or 305 total throughout the reactor core on their big plant. The issue is the reactivity difference would be subtle, so I am not sure if it can get picked up with fluxs being as high as they are, but it would be interesting to try.

I would love to put some modern electronics on this system and there are oscilliscope rated outputs (post amplification unfortunately, but still usable) on the back of the NSSS cabinets just before the PPS (if I remember correctly).

Unfortunately I have been told the maitenance has been less than desirable on this detection system because the implementation offered little value compared to the cost of upkeep on the detector side.

Regardless, good 3-D rendering would be pretty neat and might offer some evidence of where a failure may be located, especially if it produced some level of measureable Xe/reactivity oscillations.

 

[Astronuc]

One possibility for in situ measurements would be to try to utilize the Rhodium detectors. The CE design had 61 detector sets in 5 layers, or 305 total throughout the reactor core on their big plant.

In-core detectors (neutron and/or gamma) would not work. A fuel failure has a negligible effect on the neutron flux. One fuel rod out of ~200 or more, would be a signal of about 0.5% in one assembly, and much in a core of ~50K fuel rods. A failed fuel rod does not introduce a measurable Xe/reactivity oscillation effect.

 

[mesa]

In-core detectors (neutron and/or gamma) would not work. A fuel failure has a negligible effect on the neutron flux. One fuel rod out of ~200 or more, would be a signal of about 0.5% in one assembly, and much in a core of ~50K fuel rods. A failed fuel rod does not introduce a measurable Xe/reactivity oscillation effect.

Indeed, a difficult measurement even under the best conditions, and the effect would of course be a fraction of the .5%.

Regardless, even if ineffective, it would still be interesting to have a real time 3D render of the core power using a modern analog to digital PCB and updated software right off the Rhodium detectors.

Either way, thanks for the information as usual Astronuc.

 

[mesa]

Each nuclear reactor is somewhat unique, so there really isn't a typical failed fuel detection system, and most utilities and suppliers don't put out details in public. Systems will vary, but the methodology is basically the same.

High Purity Germanium (HPGe) detector systems are common these days, and have been for some time. One example, Fuel INtegrity Evaluation and Surveillance System (FINESS) from about 1999/2000 for BWRs. Note that the detection system is set up near the steam jet air ejector system.
https://www.ortec-online.com/-/medi...tegrityevaluationsurveillancesystemfiness.pdf

PWR operators have to pull pressurized coolant sample from the primary system through a long sampling line, but the detection system is more or less the same.

A quick followup question, are you aware of what the FFD running temperature, pressure, and flow rate are? It is directly off the primary loop so I expect these to be elevated.

 

[Astronuc]

A quick followup question, are you aware of what the FFD running temperature, pressure, and flow rate are? It is directly off the primary loop so I expect these to be elevated.

I'll have to confirm with some utility folks, but PWRs have a 'Hot Leg Sampling Line'. An EPRI report describes some lines from different plants. The report focuses on measurements of Fe, Ni and other cations, which are indicative for corrosion in the plants.

https://www.epri.com/research/products/1003136

A typical hotleg sample line arrangement in a Westinghouse PWR is shown in Figure 3-1. The hot leg sample line is made of small bore (typically 3/8 inch) stainless steel tubing. Coolant is passed through a cooler and then on to the volume control tank (VCT).

As far as I know, pressurized gas (Xe, Kr) and I, Cs samples are taken from the same line. That is what I have to confirm. There is some cooling involved, but I don't know the pressure. The sampling lines vary in length from 300 to ~900 feet. Some plants may have in-line sampling, but the sampling lines are done outside of primary containment as far as I know.

Figure 6 in this paper - https://www.tandfonline.com/doi/full/10.1080/00223131.2014.973460 - has a schematic of a primary coolant sampling line. It shows two coolers and a pressure reducer in-line between primary coolant loop and on-line detector or grab sample port. On-line sampling was developed relatively late, i.e., not part of original LWR plant design.

 

[mesa]

I'll have to confirm with some utility folks, but PWRs have a 'Hot Leg Sampling Line'. An EPRI report describes some lines from different plants. The report focuses on measurements of Fe, Ni and other cations, which are indicative for corrosion in the plants.

https://www.epri.com/research/products/1003136As far as I know, pressurized gas (Xe, Kr) and I, Cs samples are taken from the same line. That is what I have to confirm. There is some cooling involved, but I don't know the pressure. The sampling lines vary in length from 300 to ~900 feet. Some plants may have in-line sampling, but the sampling lines are done outside of primary containment as far as I know.

Figure 6 in this paper - https://www.tandfonline.com/doi/full/10.1080/00223131.2014.973460 - has a schematic of a primary coolant sampling line. It shows two coolers and a pressure reducer in-line between primary coolant loop and on-line detector or grab sample port. On-line sampling was developed relatively late, i.e., not part of original LWR plant design.

I did get some more information on the specific system we are interested in, and it is indeed outside of containment as you stated.

It is ran through schedule 80 stainless steel pipe that is 'u-bended' around the scinillator that is simply clamped into place with just an old school analog gauge (think early style volt meter) for total Bq's per a given timeframe.

The person I spoke with believed the inlet temperature to the FFD was only 120F to prevent damage to the resin bed for ion exchange post measurement, but I bet this is probably C. Luckily we have a radiation monitoring tech coming out next week to feed us more specifics.

I am still waiting on specs for pipe diameter (have an approximation), flow rate, etc.

The links and confirmation on some of the design specs are very useful, thank you as always Astronuc.

 

[Astronuc]

I am still waiting on specs for pipe diameter (have an approximation), flow rate, etc.

I believe the EPRI document reports those data. Typical tube ID is 3/8 inch (9.5 mm), but could be 1/4 (6.4 mm) to 1/2 inch (12.7 mm), or up to 3/4 inch (19.1 mm). See Table 3-4. I believe the sink line dimension is what one is looking for. Flow rates are given.

 

[artis]

But in theory if one is having such a small feed line from the core if it runs far enough away from the core one could just put a geiger counter and determine whether there is a leak right? Because away from the core if the core radiation is blocked (like outside of containment) then the water itself should be only minimally radioactive and failed fuel would result in the water having the gas or particulate fragments being present increasing the count rate?

It is ran through schedule 80 stainless steel pipe that is 'u-bended' around the scinillator that is simply clamped into place with just an old school analog gauge (think early style volt meter) for total Bq's per a given timeframe.

A this point if what I said is true do you even need a scintillator to know the fuel is damaged because increased count rate from core water can only mean fuel leak as there I think are no other sources for radioactivity for core water besides fuel in the core.

 

[Astronuc]

But in theory if one is having such a small feed line from the core if it runs far enough away from the core one could just put a geiger counter and determine whether there is a leak right? Because away from the core if the core radiation is blocked (like outside of containment) then the water itself should be only minimally radioactive and failed fuel would result in the water having the gas or particulate fragments being present increasing the count rate?

A Geiger counter online detects radiation and related activity, but does not indicate what radionuclide is causing the radiation. Far from the core, the level of radioactivity is much less, but there is radioactivity from activated nuclides. Certainly, gaseous and volatile fission products would increase the activity, but then question is, what is causing the activity.

Some volume has to be removed in order to provide boric acid (H₃BO₃) and buffer lithium hydroxide (LiOH) in PWRs, so this volume is passed though filter demineralizers, which are designed to capture cations (corrosion products). In addition, some PWRs add Zn to the coolant, so this is added with the boric acid and LiOH. Corrosion products are monitored since they change with the pH and power levels, especially with trips and during shutdown. Activated cations include isotopes of Fe, Cr, Ni, Mn, Co, which originate from the stainless steel and Ni-alloy (Inconel) surfaces. Stainless components, e.g., thimble tubes and/or in-core instrumentation absorb neutrons; in some early fuel designs, guide tubes were stainless steel and spacer grids were Inconel (718 or 625), but these materials were gradually replaced with Zircaloy-4, then ZIRLO and M5 (in US, EU and Asian fuel designs). The core support structure, core baffle and core barrel are stainless steel, and much of the primary system piping is lined with stainless steel, if not stainless steel. The steam generator tubes are typically Ni-alloy, typically Inconel 600 (high Ni), but later Inconel 690 (replaced 600) or Incoloy 800. The stainless steel and Inconels/Incoloys will release some metal to the coolant, which may settle on the fuel as 'crud', where the cations absorb neutrons and become 'activate'.

With respect to fuel failures, we are interested in the radioisotopes of Xe, Kr, the five I-radisotopes, Cs (¹³⁴Cs, ¹³⁷Cs), and others, including ²³⁹Np. There are various ratios (short-lived/long-lived radionuclides), which give an indication of a 'tight' leaker, or an 'open' or degraded leaker. There are limits on coolant activity due to Xe, Kr, and I, such that a reactor will must shutdown if activity exceeds a limit. In practice, utilities will often shutdown to remove a leaker(s), depending on many factors, such as load demand, cost or replacement power, time in the cycle, severity of failure, . . . .

Similar practices are employed in other types of reactors, although there are differences. For example, in fast reactor fuel, it has been proposed to add gas tags, i.e., gas mixtures, e.g., Ne, Ar, Kr to He in fuel rods. This would enable one to identify which group of rods might be leaking. In a liquid metal cooled FR, He gas above the liquid metal (usually sodium) would be sampled for gaseous activity. Gas tagging was considered for LWR fuel, but it would be expensive, and one would have to determine it's effect on fuel temperature under normal and accident conditions, so it was never implemented, except maybe for some test rods. I'm not aware of any program in a commercial reactor.

In gas cooled reactors, the gas coolant can be monitored on-line.

 

[artis]

So @Astronuc what your saying is that due to the activation of various particulate "debris" during operation the primary coolant activity is already sufficient that sampling it with a geiger tube won't produce much difference in case the leak is small so only a spectroscopic approach would show the specific products that can only come from within fuel rods and nowhere else?

Makes sense.

 

[Astronuc]

So @Astronuc what your saying is that due to the activation of various particulate "debris" during operation the primary coolant activity is already sufficient that sampling it with a geiger tube won't produce much difference in case the leak is small so only a spectroscopic approach would show the specific products that can only come from within fuel rods and nowhere else?

The Geiger counter only counts activity, and it does not indicate what is causing the activity. Each radionuclide has a unique signature for beta decay and gamma emission. We are interested in the specific nuclides producing the radiation, particularly those originating in the fuel.

It's a bit more complicated, partly for two reasons: 1) burnup/age of fuel and 2) background radioactivity. In an LWR, at the beginning the first cycle of operation, all the fuel is fresh (first core), i.e., unirradiated, and so there is little inventory of fission products. There may be some because the plant has had to go through criticality, low power testing and startup testing. With operation, fission products accumulate. If there is a failure in fresh or low burnup fuel, it might be difficult to determine. There is also the potential for 'tramp' uranium, i.e., uranium contaminating the surface of the fuel, i.e., outside the cladding in contact with the coolant. Certain signatures for 'tramp' uranium are the same for gross failures where the fuel finds its way into the coolant where it undergoes fission.

At the end of an operating cycle, some portion of the fuel is removed and fresh fuel added to the core. The fresh fuel has no fission products, except for that from spontaneous fission or low level neutron sources. If a failure occurs in the least irradiated fuel, the activity is so low that it is lost in the background. Furthermore, if failures have occurred in a previous cycle, particular in the most recent prior cycle, then there is a high level of background (failures add to the 'tramp' uranium) that could mask a current failure, or failures. We then look at signatures of short and long-lived radionuclides, or fission products, that tell us something about the failure.

There are also commitments from utility operators of NPPs to monitor certain activity levels. One such parameter is the does equivalent iodine (DEI) activity, which is a weighted sum of activities of ¹³¹I, ¹³²I, ¹³³I, ¹³⁴I and ¹³⁵I. So, the operator must account for specific radionuclides of I. The DEI must always be less than 1.0 μCi/ml, otherwise the plant must shutdown.

An example - https://www.nrc.gov/docs/ML2024/ML20246P829.pdfEdit/update: Historical perspective on coolant activity in NPPs, including gas and liquid metal systems.
https://www.osti.gov/servlets/purl/4844195

 

[artis]

@Astronuc yes pardon I wasn't clear I know geiger tube can't produce a signature for specific elements it only counts interactions per second. I was thinking that a polluted coolant would produce an increased amount of activity but you already pointed out clearly that for fresh fuel or very small leaks this might not be visible.