Power plant design

PBMR (Pty)’s perspective

1 April 2009



The developer of South Africa’s pebble bed modular reactor, Pebble Bed Modular Reactor Pty Ltd, responds to queries about pebble bed reactor safety in a series of questions and answers submitted in February. It remains upbeat about the project, reactor design, technology and fuel.



PBMR cutaway
Cutaway of PBMR reactor with containment building

Q. Does the PBMR have a containment structure that will prevent the release of radiation to the environment?

A. The safety characteristics of the fuel used in high temperature reactors such as the PBMR ensure that no large-scale fuel damage can occur. In common with similar projects, total containment of radioactivity was thus deemed unnecessary. It is, however, important to protect the building from overpressure. Whilst containment is an appropriate concept for reactors which use water as a coolant, PBMR (Pty) has determined that it is more important to ensure the building integrity by initially releasing the helium though filters and then revert to a low-pressure, closed containment. This ensures that – for all circumstances – practically all the radioactivity outside the fuel is still retained inside the building.

Q. Is it true the US Nuclear Regulatory Commission will not license a PBMR without a containment structure?

A. PBMR is not aware of any such decision by the NRC. Recent NRC publications indicate, however, that the NRC is intending to set rules relating to building performance that would allow a non-low leakage building if the requirements for public dose are met.


Q. What happened with fuel tests at the Petten, Netherlands research reactors in 2008? Why did they fail?

A. As part of the European Union development programme, enriched fuel previously produced for the German programme is being irradiated in Petten to establish if fuel burn-ups higher than those used in the past can be achieved without additional fuel failures. Tests were performed with five fuel spheres relevant to the PBMR fuel. The first four tests were completed successfully with exceptional fuel performance observed. In one test, fuel was tested outside the PBMR operational conditions (burn-up and temperature), and as could be predicted, a single particle failed near the end of the test. All test results were evaluated in depth with PBMR models and published internationally. All observations could be explained, which confirmed the boundaries of the fuel. The fifth test, containing both German and Chinese fuel spheres, is still ongoing with excellent fuel performance observed so far.

Q. What is PBMR doing to ensure that the fuel to be manufactured at Pelindaba (near Pretoria) is the required quality?

A. Coated particles from the fuel plant will be irradiated in various reactors over the world as part of an international collaboration effort. To this end, the first batch of coated particles containing 9.6 percent enriched uranium was sent to the US in early 2009 for irradiation testing at the Idaho National Laboratory. Once production line fuel is produced, a number of spheres will be irradiated in a suitable material test reactor under PBMR operating conditions and for the maximum indicated burn-up. The failure rate must be within the stated limits before the fuel can be loaded in the reactor.

PBMR has instituted a full fuel qualification programme that includes coated particle, fuel material and complete fuel sphere testing. Fuel will be independently tested under PBMR operating conditions in material test reactors in both the Netherlands and Russia. PBMR (Pty) will only be allowed to load fuel into the reactor once the fuel qualification programme has been completed successfully and the PBMR models have been validated.


Fuel element design for PBMR
Fuel element design for PBMR


Q. The Thorium High Temperature Reactor (THTR) in Germany, which was intended to be the front-runner to become the world’s first commercial pebble bed machine, operated between 1985 and 1988. Why was it closed down after only three years of operation?

A. The utility’s contract with the German government stipulated that the government would pay for all the costs of operation above the costs of the equivalent coal-fired plant next door. The calculated cost for decommissioning as well as other items increased every time this was recalculated. The government consequently felt it was being coerced and refused to deposit the funds. The utility threatened to stop operation. When neither party budged, the reactor was stopped. The government minister then responsible for the project, Dr Klaus Töpfer, was quoted as saying that he thought he had made a mistake in halting the Germany’s HTR programme at the World Economic Forum in Davos, Switzerland in 2003.

Q. It is reported that the THTR had an accident that released a large quantity of activity to the environment. Is this true?

A. There was an event in May 1986 when a manual action to insert absorber spheres into the core was done in the wrong order. A quantity of the primary helium coolant was released to the building and from the stack to outside. The released activity was below the limit where the authorities needed to be informed. The event was, however, reported in the quarterly review and nuclear opponents subsequently accused the company and the government of hiding the facts.

Q. Some publications and web sites state that the THTR had serious operating problems and that this is why it was discontinued.

A. The THTR was a demonstration power plant in several respects. As could be expected, some teething problems were experienced, all of which could be rectified. In the last full year of operation, the THTR had an availability of 70 percent, which compared favourably with other type reactors at the time. None of the technical problems can be described as ‘serious’, nor were they the reason for the discontinuation of the THTR.

Q. It is stated in some reports that the THTR suffered from compaction of the fuel, which could lead to high fuel temperatures. Your comment?

A. The initial loading of the fuel in THTR was done manually, with people walking on the pebble bed. This compacted the fuel, but it was corrected when fuel circulation was started, and the pebble bed attained the expected density of about 61%. No high fuel temperatures were experienced.

Q. Some scientists claim the AVR research reactor in Jülich, on which the South African pebble bed concept is based, cannot be used as an example for high temperature reactor design because it was so contaminated by bad fuel performance. Your response?

A. The AVR contamination was mainly due to the use of experimental fuel during its operating life. The unexpected high fuel temperatures were only discovered shortly before shutdown in 1987. Because of the shutdown, the reason for the high temperatures was never explained. PBMR is consequently in the process of re-analysing the reactor’s design and operating history.

Hot spots in reactors are defined as small areas of relatively high power compared to the average. In light water reactors, hot spots need to be prevented because they can cause local fuel cladding failure. For HTRs with ceramic fuel, hot spots are of lesser importance since the graphite material as well as the fuel can withstand large temperature differences. In pebble bed reactors, the random movement of spheres can cause small power deviations, but analysis has shown these to be of minor importance.

For over 30 years, the German fuel manufacturer has continuously improved the quality of the coated particle fuel. This was done by reducing the amount of free uranium in the graphite, by switching from BISO to TRISO fuel and by improving the quality of the SiC layer in the TRISO fuel. The latest fuel was tested for HTRs and not a single particle out of 88,000 failed in the irradiation test. PBMR fuel is to be manufactured to the same standard as the latest German production fuel.

Q. Some scientists believe pebble bed reactors produce a lot of graphite dust that traps radioactivity and can be released in an accident. Is this true?

A. The modular pebble bed reactors are designed with fuel that will not melt even at very high temperatures. Like the fuel in all nuclear reactors, however, there is a phenomenon known as diffusion where fission products can migrate from the fuel to the outside. In LWRs this is noticed when fuel rods start to leak. In pebble bed reactors the silicon carbide layer that protects the uranium kernels provides a strong barrier to most fission products, but at high temperatures (greater than 1000°C) some fission products start to diffuse through the layer into the matrix graphite. These fission products may end up in the coolant and deposit on cool parts of the system.

Due to the movement of the pebbles, some graphite is rubbed off from the surface of the spheres. This ‘dust’, which contains some of the released fission products like caesium-137, settles in stagnant or low-flow areas of the main coolant loop. Should there be a sudden break of a medium or large pipe, some of the dust may become suspended in the escaping gas and be carried out. For this reason the building has a filter in the stack to catch the escaping dust, despite the fact that only a small proportion would escape in such an event. The actual activity of the dust is low and only becomes a factor when a 50-year dose is calculated.

Diffusion can be defined as a mechanism whereby a concentration of an element at one spot in a material slowly spreads across the whole world until the concentration is the same everywhere. In theory, this would make a homogenous composition of the earth’s crust and waters. The diffusion speed of most materials is, however, both very low and very temperature-dependent and not applicable to crystals, for instance. For the coated particle, the effect is that some fission products diffuse through the SiC layer and out into the gas stream. While most of the fission products are firmly held within the uranium kernel or stopped by the coatings, a few materials like caesium, strontium and silver, which all have radioactive nuclides formed in the fission process, are more mobile than most. This is one reason why the maximum fuel temperature for normal operation is fixed at less than 1150°C, thereby preventing undue contamination of the coolant circuit. Measuring the diffusion constant is very difficult and not necessarily a good predictor of what will happen in the real application. By determining the fuel temperature and the gas activity, the predictions for PBMR can be verified and operating parameters may be adjusted upwards later, as the initial conditions for operation are chosen conservatively. There are also known ways to improve the retention capability of the fuel, but these need to be proven in long-term research efforts.

The main measurable temperatures for PBMR are the outlet gas and the reactor pressure vessel temperatures. For the first reactor, there will be additional thermocouples placed in the reflector. From these measurements, a temperature profile across the reactor can be constructed. It will be impossible to determine the exact temperatures of individual kernels as these are location-, burn-up- and coolant flow-dependent. With modern computational tools, however, the prediction will be accurate enough to ensure that no fuel exceeds the specified normal operating temperature, including the uncertainties assumed in the safety analysis.

It should be noted that for LWR fuel the same conditions apply. The coolant temperature is around 290°C, but the centre fuel temperature can be as high as 2,000°C. This is a very steep slope and in-core flux measurements are regularly taken to ensure there are no unexpected deviations from the predicted temperature.

Q. Is it not better to have a containment building and not vent to the outside?

A. LWRs have a containment that can withstand high pressure for a short time. This pressure is due to hig temperature steam escaping from a break in the pressure boundary. In case of a serious accident, the loss of coolant may lead to part or total fuel damage (for example at Three Mile Island). This will release very high quantities of radioactive material, which must not escape into the environment. Studies over the years for the PBMR and other HTR projects have shown that keeping the gas at pressure for a long time creates a bigger potential danger to the public, as even small leaks or part containment failure will lead to a higher public dose than would be the case if the first gas volume is vented and filtered in a controlled way. Therefore all designs so far have selected a vented, but closable containment.

Q. Block-fuel type High Temperature Reactors do not produce graphite dust. Are they not better than pebble bed designs?

A. While it is true that block-fuel reactors produce little, if any, graphite dust, this is only one of the considerations when assessing the two fuel formats. While the spherical fuel sphere design allows for online refuelling, block-fuel reactors must regularly replace used fuel with new blocks in a complicated change-out. After any change, some of the fuel will, for the same gas temperatures, experience temperatures well above those seen in pebble bed reactors. This will also lead to enhanced caesium release into the matrix graphite and eventually deposits in the system.

From an economic viewpoint, the availability of online-fuelled reactors like the PBMR is advantageous for applications where continuous operation is needed, such as the petrochemical industry.

Q. Why does PBMR aim for such high gas temperatures if this may produce added contamination?

A. For the direct cycle selected by PBMR for electricity production, the efficiency is directly dependent on the outlet gas temperature. For process heat applications, however, such high temperatures are not needed, except for the chemical production of hydrogen. PBMR aims to demonstrate that high temperatures can be achieved without serious contamination. Hydrogen production can therefore be a goal for high temperature reactors with some development of high temperature materials that are needed in the heat transport cycle. Note that the very high temperatures inadvertently achieved in the AVR over the years did not noticeably damage the fuel.

Q. Operators of the AVR research reactor were not aware of the high fuel temperatures that it experienced. Why would the PBMR not have the same problem?

A. In the AVR, the highest fuel temperatures were experienced where the fresh fuel entered the core. The gas flow was from bottom to top, thereby causing the highest power levels where the gas was hottest. The gas temperature, however, was not measured at the core exit, but at a position where returning bypasses had cooled the gas. Inadequate analytical tools and lack of incentive caused the problem to be unnoticed for a long time. For the PBMR, there will be more instrumentation, as well as advanced analytical tools to predict actual fuel temperatures much more accurately than was possible for the AVR. Furthermore, there is a large difference between the gas temperature and fresh fuel kernel temperatures. Fresh fuel has a high power per kernel, so kernel temperatures are much higher than gas temperatures. In planning the AVR, gas bypasses were ignored. Analysis being performed by PBMR (Pty) shows that the bypasses and the flow direction were mainly responsible for the high fuel temperatures.

In the PBMR, however, the high power of fresh fuel is at low gas temperatures. In the PBMR, where the gas flow is from top to bottom, hot bypass gas is diverted back to the core at the bottom. As there is no fresh fuel at the bottom of the core, the region where gas temperatures are high is well below the power peak region. The result is that there is only a small difference between the gas temperature (which can be measured) and the fuel kernel temperature.

Q. Why was the fuel burn-up not measured accurately at the AVR? Why should we believe PBMR has a better method?

A. In simple terms, a reactor ceases to produce power when there is insufficient fissile material in the core (usually U-235). For continuously-fuelled reactors like AVR and the PBMR this means that old fuel must be removed and replaced with fresh fuel to keep the chain reaction going. If the burn-up is not measured well enough, the core will shut down due to lack of fuel. A serious consistent underestimate of the fuel burn-up is therefore impossible. The AVR method, however, was not very accurate; a small percentage of the fuel could possibly have stayed in the core longer than planned. The method selected for the PBMR is much more accurate, but in the fuel qualification it will be confirmed that longer periods in the core will not lead to additional fuel failure.

Q. Would inaccurate fuel burn-up contribute to fission product release?

A. If some of the fuel remains in the core well beyond its planned life, additional particle failures may occur. This is unlikely to contribute much to the source term due to the large number of particles (greater than 6bn). Should any additional failures be more than a few tenths of a percent, the level of fission products in the coolant gas will be noticed by the monitors.

Q. It is claimed that the fuel sphere flow was only measured with small glass spheres in a liquid. Can this be a good prediction for actual pebble flow in the core?

A. This is erroneous information. Extensive tests were done in experimental facilities to test graphite pebble flows in a helium environment. With newer analytical software, these experimental results are well reproduced and there is confidence that actual flows will not deviate significantly from the predicted values. Should a deviation occur, it would not significantly contribute to changes in fuel behaviour.

Pebble flow is like water flowing in a pipe. At the walls, the resistance is highest, slowing down the flow. If pebbles circulate on average six times, because of flow dynamics, some might circulate only four times, others eight. For fast-flowing pebbles, the pebble power per cycle is lower and for slow-flowing pebbles it is higher, but all are within the allowable power range.

The factors influencing pebble flow and resting position are well known and can be calculated by software. This computer model gives the variation of packing density, from top to bottom and radially. It shows that small-scale deviations are not important. It was found in experiments that after some time, a stable type of pebble lattice would form against the wall, which would not move for a long time. As a result, small dimples have been included in the walls of the THTR and follow-up reactors to break up any lattice formation.

Q. Reports state that the packing density of the core may be much higher that predicted by PBMR, and that this caused high AVR temperatures and may do so in PBMR. Is that correct?

A. For the PBMR spheres, there is a theoretical maximum packing density of 0.74. This, however, applies only to hand-packed spheres in an infinite array. There is plenty of experimental evidence that the average value of 0.61 used in PBMR calculations is a very good value to use and that deviations from this are predictable (near the walls) and vary little. The same is true for the AVR. It is very unlikely to have played a role in the high fuel temperatures found experimentally at the end of life.

Q. Is it not true that in the AVR the concentration of caesium in the outer layer of so-called modern fuel elements was very high?

A. Yes, but the concentration decreases towards the centre of the sphere. This indicates that the contamination is from the outside by deposition of Caesium released from bad fuel still in the core.

Q. Does the PBMR design include emergency cooling capabilities in case of a loss of gas coolant?

A. Yes, there is a double-redundant system to remove decay heat following a cessation of active cooling for any scenario. This is to prevent fuel temperatures rising to near the licensed limit, which could lead to long delays in restarting the reactor. The system is seen as mainly for investment protection.


Related Articles
PBR safety revisited
PBMR: hot or not?
Pebble bed reactor – safety in perspective
Small is not new
US-South African cooperation agreement good news for PBMR
A new look inside cores of UK AGRs
MHI to collaborate on PBMR development

PBMR cutaway PBMR cutaway
Fuel element design for PBMR Fuel element design for PBMR
Koeberg nuclear power plant Koeberg nuclear power plant


Privacy Policy
We have updated our privacy policy. In the latest update it explains what cookies are and how we use them on our site. To learn more about cookies and their benefits, please view our privacy policy. Please be aware that parts of this site will not function correctly if you disable cookies. By continuing to use this site, you consent to our use of cookies in accordance with our privacy policy unless you have disabled them.