Pebble bed reactors are attracting worldwide interest because of their high gas outlet temperatures, allowing applications beyond electricity generation. Germany has the most experience with the technology, and a detailed examination of that country’s PBR experience is therefore appropriate, especially because of a lack of other long-term experience. Germany has constructed and operated two PBRs. The Arbeitsgemeinschaft Versuchsreaktor (AVR), or Working Group Test Reactor (46MWt, 15MWe), was operated over the period 1967-1988 at coolant (He) outlet temperatures up to 990°C, which are in principle suitable for process heat applications such as hydrogen generation by splitting water. Its availability (time basis) was 66%. After a construction time of 14 years, the Thorium-HTR THTR300 (750MWt, 300 MWe), was operated from 1985 to 1989 at coolant exit temperatures of 750°C, but for a total of only 1.2 full power years (fpy).

Over the period 1988-89 both German PBRs were taken out of operation. Mainly, it was safety concerns that led to the permanent AVR shutdown: lack of sufficient protection against external impact leading to an air ingress with a core fire, and a potential positive void coefficient of reactivity in water ingress events.

The permanent shutdown of the THTR300 was caused by a couple of technical problems, which were partly pebble bed-specific and made its operation very complex and costly. No agreement had been reached between the German government and industry on compensation for the economic risks of further THTR300 operation.

Because the interest in PBR development generally declined in Germany with the shutdown of the PBRs, associated R&D greatly reduced from 1989. Thus the operational experience of both German PBRs was not sufficiently evaluated. Nevertheless, several optimistic statements about the AVR operational experience have been communicated. Because of those statements, and in light of the increasing interest in PBRs, a study on the operational safety of the PBR prototype, the AVR reactor, was performed at FZJ and published in June 2008 [1]. An updated version of this study is in press [2]. A synopsis of these studies is presented here.

AVR operational experience

The main aim of the AVR was to demonstrate the general applicability of a pebble bed core and to test different types of fuel elements. At the beginning of AVR operation the fuel contained bistructural-isotropic (BISO) coated UC2; at the end it used UO2 with tristructural-isotropic (TRISO) and with BISO coatings.

However, the primary circuit of the AVR is heavily contaminated with dust-bound metallic fission products (strontium-90, caesium-137), which create major problems in the current dismantling effort. The end-of-life contamination reached several percent of a single core inventory, although the AVR was operated only for about four years at coolant temperatures equal to or greater than 900°C. This contamination, when adjusted to power rating, is more than five orders of magnitude greater than for current German PWRs. Despite the small size of the AVR, its beta-contamination (strontium-90) was found to be by far the largest of all nuclear installations worldwide, except for those which suffered from severe accidents. Furthermore, the mobile dust nature of the beta-contamination was found to be most problematic from the point of view of dismantling and safety. Because the AVR vessel, which contains the whole contaminated primary circuit, could not be dismantled, it was filled with light concrete in order to stabilize it and immobilise the dust. The whole vessel, 2000 metric tonnes, will be airlifted in about 2012 to an intermediate storage site and remain there for at least 30 to 60 years, until a decision about its final treatment can be made. The transport to an external intermediate storage site is required because in 1978 a small fraction of the contamination (mainly strontium-90) was accidentally released into the ground around the reactor in the course of a water ingress accident into the primary circuit. Thus, major decontamination of the soil has to be carried out, and that work requires the complete removal of the reactor.

The FZJ re-evaluation

A re-evaluation [1,2] concludes that the high AVR contamination was mainly caused by unacceptably high core temperatures, and not only by inadequate fuel quality, as presumed in the past. We can rule out the possibility that only the poor quality of the fuel was responsible for the fuel elements’ failure to retain fission products, because the same fuel type that in the AVR (BISO coating, UC2-kernel) has released a major fraction of strontium-90 was also used in core 2 of the US Peach Bottom block-type HTR without any significant strontium-90 release. Second, there was almost no correlation between the release of metallic fission products and noble gas release that indicates inadequate fuel quality. Third, AVR contamination increased by several orders of magnitude shortly after raising the coolant temperatures to 950°C.

The unacceptably high AVR core temperatures were not detected until 1987, one year before final shutdown, because a PBR core cannot be equipped with standard instruments. From 1986, temperature measurement in the AVR was performed by 190 monitor pebbles containing a set of melt wires. This method only determines the maximum core temperature a pebble has seen during its pass through the pebble bed; it gives almost no results on spatial- or time-dependent temperature distribution.

Temperature measurement with monitor pebbles is also a time-consuming process. The first alarming results were not obtained until 15 months after starting measurements. The whole measurement campaign was never finished. At final AVR shutdown, about 25% of the monitor

pebbles remained in the core and thus were not examined. In those monitor pebbles that were examined, a significant fraction had wires that were completely melted. This means that the core temperatures were more than 200K higher than previously calculated, although the maximum core temperatures are not known. Crude estimates indicate that maximum core temperatures were about 300K higher than predicted. This accelerated fission product release from fuel elements.

Furthermore, azimuthal temperature differences at the margin of the active core were measured in the side reflector, pointing to a power asymmetry. Also, unpredictable hot gas currents with temperatures greater than 1100°C were measured in the hot gas chamber below the steam generator. These currents may have damaged the steam generator, and caused the leak in 1978 that lead to a steam/water ingress accident. Despite some effort spent in the past decades, these temperature problems are not yet understood. External bypass flows of the coolant in the core have recently been assumed to be the main reason for the high AVR temperature. If external flows reduce the core cooling, it is reasonable to expect a homogeneous core temperature increase. This was not found, so external bypass flows cannot explain the high AVR temperatures. In fact, temperature measurements in the hot gas chamber indicate that the location of hot spots vary over time, in periods that match variations in pebble bed arrangement. Examination of the distribution of fission product release in spent AVR fuel elements also indicates the presence of hot spots. These observations hint at an inherent safety problem with pebble beds.

First, then, is the issue of pebble bed compaction. A stochastic pebble bed shows a void fraction of 0.4, but the densest possible packing has a void fraction of only 0.26. Pebble flow may lead to compaction, as experiments indicate. Such compaction results in higher temperatures not only because of the higher power density but also because of the significantly larger pressure drop in compacted regions. The pressure drop tends to reduce the coolant flow through compacted regions. Second, fuel elements with high and low burn-up need to be sufficiently mixed over the entire pebble bed. Accumulations of fuel with relatively little burn-up (and high power) raise temperatures significantly. Third, irregularities in pebble flow may change the local heat production and the corresponding core temperature profile. Retardation of fuel element flow, which is unavoidable near reflectors, may lead to unacceptably high burn-up, which enhances fission product release.

The pebble bed mechanics problems are related to an error made at the beginning of PBR design: it was assumed that the benign lubrication features of graphite are intrinsic properties. PBRs require low friction in the core for smooth pebble flow. However, it was discovered in 1948 (but not sufficiently noticed by AVR designers) that the lubrication behaviour of graphite occurs only in the presence of sufficient humidity, or, in a less efficient manner, in the presence of oxygen. Under the inert conditions required in PBRs, the friction coefficient is four times as large, and the wear rate is up to 10,000 times as large. This leads to the formation of large amounts of graphite dust. Unfortunately all out-of-pile experimental studies on graphite pebble beds were performed under low friction conditions. The major underestimation of graphite friction became obvious after some years of AVR operation because of the large amounts of graphite dust found, because of pebble movement problems in the fuel reshuffling facilities, and because of a pronounced discrepancy between calculated and observed residence time spectra of fuel elements in the core.

Similar phenomena were later observed in the THTR300 too. However, the THTR300’s average coolant temperatures were about 200K lower than in the AVR, so temperature-induced effects were not as pronounced as they were in the AVR. Activity release in THTR300 also remained limited due to the low fuel burn-up achieved in its short operation time. However, in THTR300, pebble flow was virtually restricted to the core axis, and hardly occurred at all in the outer core zones. Because of this effect, temperatures rose substantially higher than expected close to the reactor axis, and led to hot gas currents that probably caused the damage to the hot gas duct’s metallic components which were observed after less than 1 fpy. In THTR300, unexpected pebble bed compactions were found. An evaluation of the THTR300 operating experience is still incomplete and is urgently required.

The black box character of the PBR core makes it uncertain whether convincing explanations for the unacceptably high AVR temperatures can be identified without major experimental research and development work. A general disadvantage of PBRs in comparison with other reactors is that the position and local density of fissile materials in the core, which are important input parameters for neutronics and thermodynamics calculations, are not known a priori, but have to be estimated via insufficiently validated pebble bed mechanics codes.

After high AVR temperatures were detected, coolant temperatures had to be strongly reduced for safety reasons. Maximum fuel temperatures had significantly exceeded permissible limits. Therefore, the AVR should no longer be taken as a reference for a safe and reliable reactor operation at gas temperatures allowing process heat applications, as is often done. Ironically, the pebble bed HTR concept has probably survived until now mainly as a consequence of one of its weak points, its insufficient in-core instrumentation abilities. Had AVR core temperatures been known from the beginning of its operation, the AVR hot gas temperatures would have been limited to values far below 950°C. Its main advantage, its apparent capability for process heat generation, would not have been available.

If temperature limits for specific metallic fission products are exceeded, they diffuse through fuel kernels, coatings and graphite over the long term. This weak point of HTRs and their fuel, in contrast to other reactors, is unresolved. It is particularly problematic in PBRs, which have large amounts of mobile, and fission product-absorbing, dust. At high temperatures, the diffusion of metallic fission products through intact coatings of TRISO-coated fuel becomes significantly larger than fission product release from particles with defective coatings. This is consistent with calculations using standard diffusion codes and data, and is particularly relevant for caesium. Reducing the fraction of defective particles, which was found to correlate during AVR operation to reduction of noble gas release, hardly stops the release of metallic fission at high PBR temperatures. Moreover, this holds true even if temperature anomalies such as in AVR are not considered. Current TRISO fuel is thus not suitable for high temperature applications such as process heat generation, in contrast to earlier assumptions. A recent experiment on high-quality German TRISO fuel elements underlined this [3]. When the fuel was irradiated at temperatures near the upper limit of the fuel design range, researchers recorded the release of unexpectedly high amounts of fission products. The upper applicability temperature of TRISO fuel should therefore be reduced by 100-150K from present assumptions.

An important lesson to be learned from the AVR water ingress incident of 1978 relates to reactivity effects. In order to dry the primary circuit, the reactor was operated for several days at low power and temperatures (500°C and 10MWt). During that time, 30,000kg of liquid water ran out of the steam generator leak toward the core. A fraction of about 3% of liquid water in the void volume of a pebble bed may lead to a positive void coefficient of reactivity, which might induce a reactivity excursion. These conditions were not reached in the AVR water ingress accident, partly because the steam generator leak did not enlarge. But in order to prevent such potential reactivity problems, future designs must not place steam generators on top of the core.

The consequences

The limited retention of metallic fission products in HTR fuel, and the presence of released fission products in mobile dust inside PBRs, have to be taken into account in design and licensing. For a PBR of 950°C to 1000°C hot gas temperature using modern fuel over 32 fpy of operation, the caesium contamination is expected to exceed the contamination in the AVR, even if temperature anomalies are not considered. This creates major problems with design-basis accidents, for maintenance, and for dismantling. Design basis accidents involving the release of radioactive dust into the environment in the course of depressurization are a major concern. Even for low coolant exit temperatures (700°C) and low fuel burn-up, the primary circuit caesium-137 contamination of HTRs remains about three orders of magnitude larger than in PWRs when compared on a basis of equal power rating. Moreover, as long as the inherent causes for the high temperatures in pebble beds cannot be prevented, they have to be conservatively accounted for in operation and design-basis accidents. Temperature anomalies, as in the AVR, significantly increase the release of caesium and initiate strontium release. Additionally, for steam-generating HTRs, the formation of burnable gases in the course of a water ingress accident, the probability of which is significantly enlarged by high reactor temperatures, has to be carefully examined. Temperatures during core heat-up accidents may become higher than expected, in case of hot spots. The application of German dose criteria on advanced PBRs leads to the conclusion that PBRs need a gastight containment, even without considering the AVR’s unacceptably high temperatures. Gastight containments are included in almost all other modern reactor systems. The PBR containment should be explosion-proof or inertized in order to prevent potential dust releases or burnable gas explosions in the course of an accident.

In addition to the design of a gastight containment, additional measures are needed. A reduction of hot gas temperatures in fuel burn-up is one option. For example, a recent study [4] opts for an HTR based on the well-proven British AGR design (CO2-cooled, 640°C hot gas temperature), but using coated particle fuel and standard steel vessel. An additional heating step using conventional means to reach temperatures desired for process heat applications is proposed. Another option is an elaborate research and development programme to solve unresolved problems related to operation and design-basis accidents, and to update safety analyses before construction. The following tasks remain. (This list takes also results of the NRC and the NEA Committee on the Safety of Nuclear Installations (CSNI) evaluations of PBRs into account.)

• Full evaluation of the operational experience and problems of AVR and THTR300.

• Development of a new fuel element that sufficiently retains metallic fission products over long-term operation.

• Development of reliable quality control for fuel elements.

• Experiments on iodine release from fuel elements in core heat-up accidents.

• Examinations of unexpected particle failures as observed in experiments with realistic core heat-up transients.

• Full understanding and reliable modelling of core temperature behaviour, and of pebble bed mechanics, including pebble rupture.

• Experimental and theoretical examinations of dust formation under real PBR conditions.

• Development of a fast and reliable local measurement method (whether direct or indirect) of safety parameters in the pebble bed core, such as temperature.

• Full understanding of fission product transport in the coolant circuit, including dust influence.

• Development of measures to avoid activity accumulation in the circuit.

• Full understanding of the cobalt-60 contamination of the primary circuits of AVR and THTR300.

• Development of a fast detection system for metallic fission product release from core.

• Material development for nuclear process heat components.

• Development of HTR-specific dismantling and disposal items.

Although priority has to be given to solutions of safety problems related to operation and design-basis accidents, there are still unresolved questions concerning beyond design-basis accidents. In particular, graphite burning caused by a huge air ingress may lead to massive fission product releases into the environment. In such air ingress accidents, there is a grace period between the end of depressurization and damage to fuel elements, ranging between a few hours and some days. It is assumed that no significant activity release occurs due to the depressurization preceding the air ingress, that is that the reactor surroundings remain accessible and thus intervention measures can easily be performed during the grace period. However, the widespread and mobile AVR dust contamination may make interventions impossible.

It should be noted that block-type fuel HTRs do not show most of the problems discussed above (no graphite dust, in-core instrumentation possible, no uncertainties as to pebble flow). They also have a smaller proliferation problem than PBRs due to their discontinuous fuelling. Concerning safety in general, a comparative probabilistic safety assessment considering Generation III LWRs, PBRs and block type HTRs is required.


Improved knowledge of the safety of PBRs leads to the conclusion that previous comprehensive safety assessments were too optimistic. Previously, PBR safety was claimed to be superior to other nuclear systems, with an allegedly ‘catastrophe free’ and ‘inherently safe’ design. According to the facts presented above, there are doubts about whether these claims depict reality.

The claim of superiority is mainly based on PBR advantages in areas where conventional reactors have problems, without sufficiently considering the disadvantages of PBRs. The first advantage of PBRs is a ceramic core of low power density, which will not melt down in case of lack of cooling. This advantage is however achieved by a large core containing a huge amount of graphite. In reality, the high costs of this core design have to be compensated by omission of some safety measures outside of the core, such as a gastight containment.

In addition, graphite can burn: the sort of core fire possible in HTRs does not exist in conventional reactors. Also, the ultimate disposal of the graphite, with a remarkably high carbon-14 content, is an unresolved problem. Ultimately, cooling of the core is hindered by the high core mass. In this context claims of the stability of fuel elements up to 1600°C have to be questioned, too. This stability means only two things. First, that short term heating (100 hours) of modern fuel elements with low and medium burn-up at 1600°C does not lead to additional failures of coated particles. Second, that the diffusion of fission products, except of silver-110m, remains small. This temperature limit of 1600°C must however not be generalized. It is too optimistic for fuel elements with high burn-up, for fast transients and particularly for prolonged heat-up periods with diffusion break-through of fission products.

The second advantage of PBRs is their strong negative temperature coefficient of reactivity. It was demonstrated in the AVR (and later in the Chinese HTR-10) that in case cooling stopped, but operators had not shut down the reactor, the negative temperature coefficient shut down the reactor. This benign effect should however not be generalized: a fast withdrawal of control rods in current PBRs leads to a heavy increase of reactivity which at the least destroys fuel elements before the reactor shuts itself down. Furthermore, the positive void coefficient of reactivity, which may occur in steam generating PBRs in the course of water ingress accidents, is a disadvantage of PBRs not found in conventional reactors.

In summary, PBRs contain certain inherent safety features compared to conventional reactors, they are however to some extent compensated by inherent safety problems. Some of these inherent safety problems can be solved by adequate safety measures or by R&D, but it remains uncertain whether this is possible in an economic manner. With PBRs, there is a tradeoff between economy and safety.

Author Info:

Rainer Moormann, Forschungszentrum Jülich GmbH, 52425 Jülich, Germany

Related Articles
PBMR (Pty)’s perspective
2,300MW on the line
Small is not new


[1] R. Moormann, A safety re-evaluation of the AVR pebble bed reactor operation and its consequences
for future HTR concept Jul-4275 (2008). Short version in: Proc. 4th International Topical Meeting on High Temperature Reactor Technology HTR2008, Washington DC, September 28-October 1, 2008, paper HTR2008-58336

[2] R.Moormann, AVR prototype pebble bed reactor: A safety re-evaluation of its operation and consequences for future reactors. Kerntechnik 74 (2009), No. 1 in press

[3] M.A.Futterer, G.Berg, A.Marmier, E.Toscano, D.Freis, K.Bakker, S. de Groot. ‘Results of AVR fuel pebble irradiation at increased temperature and burn-up in the HFR Petten,’ Nucl. Eng. Des. (2008),Volume 238, Issue 11, November 2008, Pages 2877-2885

[4] A.Marmier et al., HTR with downgraded specifications for high temperature process heat applications, Proceedings of ICAPP 2007, Nice, France, May 13-18, 2007, Paper 7058