Wednesday, March 19, 2008

Pebble fuel grains pass performance testing

INL Advanced Test Reactor test site

An earlier post described the start of testing of the multiple-layer-coated fuel grains that form the billiard-ball-sized fuel pebbles in the pebble bed reactor. Idaho National Laboratory used its Advanced Test Reactor to expose these test fuel grains to radiation levels much higher than in an operational PBR, thus simulating years of exposure in a few months. The multiple, coated layers of silicon carbide and ceramic graphite contain the radioactive products of fission. These tested fuel grains have not failed, at the level of 9% burn-up of the uranium within. Tests will continue to see if a 12-14% burnup can be achieved by year-end.


aangel said...

Hi, Robert.

How do you see nuclear power being created as we are moving through Energy Descent?

For reference, please see


Robert Hargraves said...


I know that historically GDP has been positively correlated with energy use. I believe that OECD nations can gradually use less energy without impacting economic standards of living. But this matters little because the much larger rest of the world seeks to achieve a better standard of living akin to that of OECD nations, and they will consume energy to do this. Conservation will not work for them. Nuclear power is one answer to providing the energy needed in the developing world. With proper controls, nuclear power plants exports could help the developing world.

aangel said...

Hi, Robert. Thanks for your response. However, I may not have given you enough to go on to answer my question.

As background, I'm presuming that one or more critical factors for our civilization to continue are soon going to be limiting factors, per Limits to Growth: The 30-Year Update.

The most immediate one appears to be the decline in oil and the loss of energy from it (although close on its heels appears to be water depletion and soil depletion). With oil about to decline, or more precisely, with new oil production no longer able to keep up with the already-occurring depletion of approx. 4.5% per year, how do you see the rollout of pebble-bed reactors impacted?

Perhaps another way of getting at the answer to that question is to answer 'what is the EROEI for a pebble-bad reactor?' Dennis Meadows puts forward the following ratios:

Some Net Energy Yields
• US Oil 1930 - 100; 1970 - 30; 2005 - 15
• Imported Oil - 30
• Coal - 10 - 80
• Nuclear - 10
• Firewood - 25
• Photovoltaics - 15-45
• Oil Sands - 2-3

Are pebble-beds higher or lower than conventional nuclear?

That makes a difference for the following calculation:

Years for Nuclear to Give Back Energy
• Assume 4 year construction time.
• Assume 40 year operating life.
• Assume energy payback is 10.
• One plant starts to give net energy in the 9th year.
• A system building one plant/year gives positive net energy in the 13th year.
• A system building 10% more plants each year gives positive net energy in the 15th year.
• A system building 20% more plants each year never breaks even.

I'm trying to get a sense of whether we will be in a position to build these reactors or whether they will be too expensive in $'s and energy in the new context we will soon be operating in. At what rate of buildout do we never reach break-even?


Robert Hargraves said...


I suppose pebble bed reactors have returns on invested energy very approximately the same as for other nuclear power sources. In this example of the Vattenfall nuclear power plant the lifetime energy consumed is 1.4% of the lifetime output.

Credibility of this number is reasonable good, given the audit by EPD. Check
for more information about ISO 14025 standards for climate declarations.

Enrichment is the biggest energy consumer. This will be reduced an order of magnitude by switching from diffusion to centrifuge technologies.

The total energy debt investment in building a power plant is a small fraction of the first year's energy production, if you look at the numbers in this example. I'm not sure where the 13 year number you mention comes from.

aangel said...

Hi, Robert.

Yes, I'm not sure how he derived the 13 number either. I'd have to play around a bit. Here is a link to the presentation from which I got those numbers:

For the system to be net energy positive my presumption is that he is estimating a build start of one per year for a sequence of years. On year 5 the first reactor starts producing power and some portion of that is used to build the other ones not yet complete. Build them too fast and more energy is being drawn from the grid and other sources (like oil) to build the reactors and there is no net energy gain available for other uses. Build them slowly enough and there is net energy available to the grid AND to build more reactors.

I'd like to see his math. Perhaps he would respond to an email.

The energy ratio from the LCA on the Vattenfall is 1.4% as you say but they don't list the EROEI. The sample calculation above shows a thermal ratio of 17, which is close enough to Meadow's 10 that perhaps they are the same or similar number.

The key point is that we were getting 100 barrels of oil energy back from an investment of 1 barrel when we started our industrialization. Now it appears that the EROEI (energy return on energy invested) is somewhere like 10 or 17 for nuclear — this is a dramatic drop for us. (Conventional oil is still 30 but tar sands are 2-3.)

So the number to be concerned with isn't 1.4% which leads one to believe that nuclear (or any source) is better than it actually is. The EROEI, which is 17 going by the WNA or 10 using Meadow's number, demonstrates that we are moving down the EROEI curve — and that means complex society is soon going to be poorer in terms of energy.

Robert Hargraves said...


Co-incidentally, I was a person who helped recruit Meadows to Dartmouth. He used a simulation language, Dynamo, at MIT. I was the Associate Director of the computer center at the time, and I asked a hotshot student, Phil Koch, to develop a version of Dynamo for the Dartmouth Time Sharing System, which helped attract Meadows.

Today I presented the first session of Energy Policy and Environmental Choices: Rethinking Nuclear Power. Slides and audio are available at

aangel said...

Robert, re: Meadows, thank you for doing that!

I downloaded and viewed your first presentation...very informative. The table of various energy expenditures was particularly interesting. Thank you for making that available.

I also went through your previous posts, but only quickly. How much is a pebble bed reactor projected to cost? For what size? Sorry if you covered that already in a previous post.


Roland said...

I am very impressed with all the good work you have done.
You might be just able to make a difference in the Climate Debate that is predominately directed by paranoia.

I hope that every politician on this planet takes a look at your summary and get informed rather then being told by the some of the environmentalists what to think.


Robert Hargraves said...

Thank you. The next segment, "Fear", will be posted after the 3/31/08 class at

John said...

I have just discovered this blog and am very impressed. Your ideas about nuclear power and specifically pebble bed reactors are aligned well with my own thoughts about the world's energy future. As a Ph.D in Engineering (U of Mich, 1968) I plan to be an active participant in this discussion in the future. I believe energy (how much we need, how we use is, where we get it) is the over-arching issue for the world today.

Robert Hargraves said...

John et al,

This blog is not very active. If you like, you can visit my course website at

There I have posted a list of active blogs that I have found useful to me.

Axil said...

These tested fuel grains have not failed, at the level of 9% burn-up of the uranium within. Tests will continue to see if a 12-14% burnup can be achieved by year-end.

To minimize the waste problem, I would think that 100% burn-up is an ideal goal. What is achievable? What design parameters are important? How much PU(239) can be consumed?

Robert Hargraves said...


Burnup is limited by two things. One is that as the amount of U-235 is reduced the unit drops below criticality. Second, the products of fission build up, and some are neutron absorbing, slowing the chain reaction. To achieve higher burnup requires reprocessing, such as the Integral Fast Reactor provides.

jimcrea said...

I took a closer look at the referenced article and also thought about what you said in your reply to my last post. I think that the testers are increasing the U(235) enrichment in stages to see how much burn-up they can get before the silicon carbide packaging of the pebbles fail. The testing is tedious but the testers are excited that the pebbles are holding up so well at this stage in the testing.

Nuclear fuel packaging in its many forms seems to be the key to waste minimization and associated energy efficiency; the higher the enrichment that can be supported by the packaging, the higher the efficiency that can be achieved.

The US is hugely afraid of reprocessing because of the perceived proliferation tread. It is also expensive, so in a one-through fuel cycle the fuel packaging is most important.

Thank you for your attention to my last post and I greatly appreciate any feedback to this post that you feel is warranted.

jimcrea said...

I have done some further research to try for myself to answer the original question I posted to you.

The Pebble Bed Modular Reactor (PBMR) is specified for a burn-up of 90,000 MWd/t of uranium. I think that is at 9% burn-up as tested as per the INL article.

The key burn-up parameter is the hardness of the neutron spectrum (fast average speed of the neutron flux) that the fuel/reactor can operate at.

The U(235) enrichment is constant at 9.6%

Nuclides with an even number of neutrons have very small fission cross section for neutron energies below 1 MeV. Hence, in the (epi) thermal spectrum of a light water reactor (LWR), these nuclides will tend to accumulate.

In both the LWR and the BPMR, nuclides with an odd number of neutrons are burnt easily.

As a hard spectrum reactor, BPMR will burn PU(238),PU(240), PU(242). CM(244), well. The probability of this burn for the light water reactor is near zero, but for BPMR, it is about .5.

What adjusts the neutron spectrum is the thickness and density of the Pyrolytic Carbon( IPyC) and the Silicon Carbite (SiC) Barrier Coatings. For the prototype fuel, the SiC layer is 35um thick. For the optimized production fuel, the SiC layer will be 25um thick. This will make the neutron spectrum even harder for the optimized production fuel. I think the test was run with the prototype fuel.

If the end of the year test is successful at INL, a PBMR burn-up of 14% means 140,000 MWd/t. But that still leaves room for improvement in burn-up using the production fuel. This will leave very little plutonium products left at burn-up. This is very good because the waste is far less dangerous in terms of storage and proliferation.

Breeding ratio (k) is temperature dependent but at operating temperature (K=523) it is (k = 1.12). This seems high in terms of what the NRC wants to see.

I think, the design of TRISO fuel is very conservative in relation to what can be done. It is designing for safety and political correctness.

I would like to know what the INL varies to increase the fuel burn-up. Maybe it is the neutron spectrum from the test reactor? Changing the TRISO fuel seems like a lot of work.

I am only an amateur who is just starting and who is trying to understand the technology. I would greatly appreciate correction of any errors in my thinking.

Also any reaction to this info is also greatly appreciated.

Major reference:

jimcrea said...

I have done more research.

There is absolutely no doubt; the Bush nuclear plan will use TRISO fuel. Whether it is Pebbles or Prismatic is not yet determined.

As an environmentalist, the following results are disappointing in that plutonium isotopes are not complexly consumed in the spent TRISO fuel. I hope to correct any mistakes that I posted above.

In order to meet IAEA criteria for discard of fuel, Pu-238 must be 80% of the PU content.

TRISO Spent Fuel Discharge Isotopes Table
Spent fuel summary at 10% burnup

………Fresh Load Milligrams (mg)


………………Burnup mg

Pu-238………( 2.6)
Pu-239………( 65.6)
Pu-240………( 33.2)
Pu-241………( 31.1)
Pu-242.……..( 21.9)
Total Pu.……(154.4)

The fissile contents of TRISO fuel spent fuel are comparable to those of fresh light water reactor fuel. For the case without control rods, the fissile isotopes content is about 4.38 wt-percent if one considers all plutonium isotopes (since they are all fissile) or about 3.27 wt-percent if only Pu-239 and U-235 are considered. Since these contents include Pu-239 in addition to the residual U-235, they correspond to nominally higher enrichment levels. Furthermore, the fissile isotopes content of the TRISO fuel spent fuel is comparable to the fissile content of fresh light water reactor fuel.

It follows that the TRISO fuel spent fuel could be used as the fresh fuel of another Light Water Reactor provided some intermediate processing is carried out. That processing would involve the removal of the fission products and could involve the removal of the minor actinides if a future use in a thermal spectrum is considered.


In all cases, the reprocessing of TRISO-based spent fuel is a difficult and expensive undertaking. Notwithstanding the expense, the fissile content of TRISO-based spent fuel cannot be viewed as irrevocably irretrievable from the point of view of safeguard via waste storage.

The TRISO fuel design goal is full burnup in 664days at 18%


Robert Hargraves said...

Jim, you may be a amateur but you are a fast learner!

jimcrea said...

I have been doing some research. I could be wrong; let me know of any errors.

General Atomics has already tested TROSO fuel to 80% burnup. The PU(239) was reduced by 98%. Light Water Reactor(LWR) waste is the best for power production. An astounding 750,000 MwW/t was quoted. In comparison, a LWR gets 40,000 MwW/t. That is 18.5 times greater power output.

It is part of this project as follows:
General atomic Modular Helium Reactor (MHR) Development Program Underway in Support of Weapons Plutonium Disposal in Russia Beyond 34 MT Excess

All three TRISO types use 35um ceramic. The burnup must all depend on something other then the thickness of silicone carbide. More research needed here.

Another question is as follows: can the pebble be used to do deep burn? Maybe not because of all the additional ceramic involved; that is bad for reprocessing.

There are three types of TRISO fuel: commercial, LWR based, and, weapons material based. They have a one pass burning process and a two pass burning process that reprocesses TRIOS one pass fuel.

MHR neutronics allow mixing of any and all types of TRISO fuel in any ratio to avoid burnout, so very high levels of “Deep Burn” can be achieved

LWR waste processing extracts actinides and the Uranium is recycled into new LWR fuel.

The general atomic reactor is a prismatic design.

PU(239) and U(235) evaporate in TRISO waste between 1000 and 10,000 years. After 10,000 years there is not many daughter fission produces left in the waste.

This beep-burn technology fits within the operational envelope of commercial MHR operation, including long refueling intervals and the highly efficient production of energy (approximately 50%). To the plant operator, a Deep-Burn Transmuter will be identical to its commercial reactor counterpart.

The DOE as bought in to the tune of $7.3 million.

The US DOE Awards has awarded $7.3 million for “Deep-Burn” Nuclear Technology Research & Development to internal government research agencies; $1 million led by Argonne National Laboratory; and Transuranic Management Capabilities of the Deep-Burn VHTR at a cost of $6.3 million led by Idaho National Laboratory.

I am very encouraged. An answer has been developed to remove nuclear waste and weapons material from this earth.


jimcrea said...

Here is some opinion that is in the early stages of development in my mind as follows:

Opinion 1

There may be an advantage in the prism system because the TRISO fuel can be precisely arranged via an optimized load pattern in the reactor bed to provide good neutron utilization and uniform extended burnup. Also various kinds of TRISO fuel: standard, LWR, weapons material, can be used in a load plan to implement a mixed-core (boosted) system.

The pebble bed approach has a random fuel distribution with no control possible. New hot fuel cannot be added with precision to extend the burnup to a very high level.

Opinion 2

The extra ceramic makes the reprocessing of the pebble bed TRISO fuel more difficult than the prism fuel system. The less ceramic that you have to remove, the easier it is to get at the target material. How much easier is the make or break question.

Opinion 3

Because of the General Atomic deep-burn experience and vision with reprocessing, together with the $7.3 million award for deep-burn nuclear technology research & development which shows a degree of excitement by the government for deep-burn, General Atomic has the edge in the ‘Next Generation Nuclear Plant Design’ competition.

Since my expertise is minimal, any insight is greatly appreciated.

Rod Adams said...

jimcrea - If you are not the same person as the one who signs his name as Axil on Atomic Insights, the two of you really need to talk. You are seemingly sharing a mind.

Robert asked me to come and answer your comment about the use of prismatic cores to optimize burn-up. That is one of the selling points that General Atomics makes when comparing its fuel designs to pebbles.

It is certainly true that it is possible to shuffle fuel around both horizontally and vertically with the prismatic fuel design. It could lead to some interesting operations during a refueling evolution, but if that is the goal for the system, it might be worth it.

From my own perspective, there are other measures of effectiveness and economy that make pebble bed cores more attractive. Not saying there is anything wrong with the GA prisms, just that they are not easy (or cheap) to make.

They were also the source of several of the growing pains issues with the old Ft St Vrain reactor - the only large high temperature helium reactor ever built and operated in the US. I have visited with GA several times over the years and believe that they learned a lot from Ft. St. Vrain - though that experience is now so far into the past that the lessons may be fading from memory.

I also wanted to briefly touch on a misconception that I found in one of your earlier posts - the neutron spectrum in either a prismatic or a pebble bed reactor is not affected very much by the coatings on the particles themselves. In both cases, most of the graphite moderator in the core is outside of the fuel particles.

For pebble beds, the fuel particles are pressed together with a large amount of graphite to form the ball - the ratio of fuel to graphite is about 9 grams to 210 grams in the PBMR and the Chinese HTR fuel designs. For prismatic cores, there is less graphite in the actual fuel pellets, but the prisms themselves are made of dense graphite with machined channels for the pellets and for the cooling channels. I think that the metal to graphite ratio is about the same as that used in the pebble beds.

In both cases, it is possible to alter the fuel to moderator ratio to harden the spectrum without changing the fuel particle design.