ATRIUM 10: ten years of operational experience3 September 2002
Nuclear plant operators and utilities expect three primary qualities from the fuel that they load into their reactor cores: a high degree of fuel reliability, excellent fuel utilisation, and good operational flexibility. By Norman L Garner and Peter Urban
By April 2002, more than 47,000 Framatome ANP BWR fuel assemblies containing a total of 3.3 million fuel rods have been irradiated in a total of 52 plants located in Europe, USA and Asia. This BWR fuel operating experience encompasses fuel assemblies with fuel rod arrays ranging from 6x6 to 10x10. Around 850,000 Framatome ANP BWR fuel rods are currently in service in 29 plants worldwide, almost exclusively in fuel assemblies of the ATRIUM family.
The frequent switch to more advanced fuel assembly designs with different rod arrays - which always has been a typical feature of BWR plants - has resulted in earlier fuel designs being successively replaced. This continuous development focused on enhancing fuel economics through increased burnup potentials, large operating margins, and a high degree of operational flexibility. Today's optimum was achieved in the spring of 2001, when four ATRIUM 10 lead fuel assemblies reached their target assembly burnups of 71MWd/kgU after eight operating cycles.
So far, ATRIUM 10 fuel assemblies have been supplied to a total of 20 plants in Europe, Asia, and the US giving Framatome ANP operating experience with nearly 4700 fuel assemblies of this type and, thanks to ongoing supply contracts, the experience will continue to be enhanced significantly. This applies particularly to European BWRs that have been supplied almost exclusively with ATRIUM 10 fuel assemblies since 1998. In the US and Taiwan, all BWR reload fuel currently being delivered is of the ATRIUM 10 design.
Optimised BWR fuel assemblies require advanced design methods in order to satisfactorily model modern features like internal water structures or part-length fuel rods, especially at high burnup. Such methods must be validated against data obtained through extensive testing and, where practicable, in-service examinations. In-service examinations also provide the basis for confirming performance of the cladding and structural materials under operating conditions. A comprehensive validation process explicitly supports ATRIUM 10, assuring neutronic, thermal hydraulic, and mechanical performance characteristics are accurately predicted with design methods.
Void fraction distribution
The unique geometry of ATRIUM 10 creates a complex distribution of water and fuel both over the assembly cross-section and between the top and bottom of the fuel assembly. A detailed understanding of the performance of a specific assembly design, especially void distribution characteristics under various operating conditions, is critical to assuring neutronic methods are properly applied. It is therefore essential to support the validity of the void fraction correlation used in the development and design of advanced fuel assemblies and, in addition, generate void distribution information that general correlations cannot provide. Therefore, Framatome ANP extended the capabilities of its KATHY thermal hydraulic test loop by implementing an advanced void fraction measurement device.
Framatome ANP has developed neutronic methods to model the effects of void distributions and flux gradients inherent in the BWR environment with great precision. The capability of these methods, with the integrated void distribution model, to accurately model ATRIUM 10 has been directly validated through in-service gamma scan measurements on irradiated fuel assemblies. These examinations take advantage of the ability to detect concentrations of indicator isotopes from unique gamma emissions. Use of indicator isotopes is a well-accepted means to determine whether neutronic methods accurately calculate concentrations of the wide range of isotopes present in an operating assembly. Framatome ANP's MICROBURN-B2 core simulator is the only methodology currently available that includes explicit measurement data for a 10x10 fuel design.
Measurements were taken in 1998 for a once-burned ATRIUM 10 assembly near its peak reactivity. A total of 600 measurements were taken from 49 rods across the assembly and at four axial levels to span the range of existing local water-to-fuel conditions and void fractions. Separately, Framatome ANP had performed core tracking calculations for cycles leading up to and including the cycle in which the examined assembly operated. Very good agreement was obtained in the comparison of Ba-140 distributions simulated with MICROBURN-B2 versus those measured with consistent accuracy in all positions. The top right Figure on this page presents comparisons of the measured and calculated concentrations of the Ba-140 indicator isotope for a part-length and a full-length rod.
The ability to accurately predict dryout behaviour of a fuel assembly is critical to maximizing economic efficiency and operating flexibility while remaining in strict compliance with licensing conditions. Improvement of dryout behaviour with the ATRIUM 10 was the result of considerable investment in the development process.
In order to account for different axial power shapes, dryout tests were conducted for cosine, bottom peak and top peak axial power distributions. Inclusion of several top-peaked power profiles is a critical factor due to the complexity of dryout behaviour since this represents power profiles during the typically limiting end-of-cycle condition. This data was then integrated into a dryout correlation that accurately reproduces the experimental results. Additional top peaked tests were then conducted to demonstrate that the correlation remains reliable for predicting results not included in the correlation database.
Several examination campaigns have been conducted to assure that ATRIUM 10 fuel operates in line with expectations. Most notably, the initial lead assemblies were examined after each of eight consecutive cycles of operation. Examinations have included general visual assessment, growth evaluations, oxide measurements, and the gamma scan measurements noted earlier. Overall performance has been excellent with no unusual conditions encountered with assembly burnups in excess of typical batch discharge levels. The Figure on the right shows a representative photo of an ATRIUM 10 assembly that has achieved a burnup of 54MWd/kgU, typical of fuel at discharge burnups.
ATRIUM 10 rod measurement data has been incorporated into the Framatome ANP database that is partly based on the insertion of lead fuel rods which have been irradiated to burnups far above the range of commercial operation. The maximum peak pellet burnup currently achieved by Framatome ANP is 105MWd/kgU. Hot cell examinations have been completed for fuel rod burnups of 90MWd/kgU with the peak pellet burnup correspondingly higher. Comparison of ATRIUM 10 rod data to this database shows performance well within the expected range. This expanded database is being used to extend the fuel rod modeling code to explicitly high burnup effects including burnup dependence of the fuel thermal conductivity, pellet rim formation and rim structure, and fission gas release at high burnup.
Increase in discharge burnup
While always maintaining reliability as the first priority, nuclear fuel development work has focused on finding ways to cut down on fuel cycle costs. Primarily these efforts have led to extending discharge burnup because this allows a reduction in the number of assemblies required for each reload. Consequently, the volume of spent fuel requiring storage and, ultimately, disposal is reduced. For example, typical reload batch sizes were reduced by about 10% when batch average discharge burnups were increased from 45MWd/kgU to 50MWd/kgU. In the case of a large BWR plant, this allows a batch size reduction of at least 12 assemblies in annual cycles and about 28 assemblies in two-year cycles. Efforts to extend the batch average discharge up to 60 MWd/kgU are now underway and will more than double these reductions in batch size.
Significant progress has been attained with incremental extensions of burnup capability as design evolutions from 8x8 arrays to today's 10x10 arrays have been successively introduced. Material development efforts concurrent with design development efforts have assured that appropriate cladding and structural materials have been proven and implemented in production well in advance of these extensions in fuel burnup levels. Of particular interest in this context is an "ultra high burnup programme" launched in a European BWR in 1999. As part of the programme, different structural and cladding materials as well as fuel with different grain sizes and additives are studied. This programme will provide the basis for eventually extending burnups beyond 70MWd/kgU.
The most successful increase in burnup is attained by increasing the average fuel assembly enrichment while also improving fuel utilisation. Design modifications introduced during the transition from 8x8 rod arrays to ATRIUM 10 served, on the one hand, to provide the design margins necessary for extended burnup and, on the other, to increase capability to efficiently use the fuel. The Figure above clearly demonstrates this relation. Key to improving fuel efficiency in BWR fuel is an increase of the volumetric water-to-fuel ratio with an optimal distribution. The ATRIUM water channel feature alone provides an improvement in fuel utilisation corresponding to an energy production increase of about 4MWd/kgU. Effective design developments have incorporated features that yield this improved ratio and also improve thermal margins. Cycle design studies have shown batch average burnups as high as 65MWd/kgU can be achieved with ATRIUM 10 through optimised use of enrichments within the allowable limit of 5wt% U-235.
The progress already made in the realisation of the burnup capability is especially reflected by the figures for batch average discharge burnup. In the last 20 years, average annual discharge burnups have risen from less than 30MWd/kgU to 45MWd/kgU (see Figure below). This is projected to soon reach and, in some cases, exceed 50MWd/kgU as full cores of ATRIUM 10 are operated to currently approved burnup limits (ranging from 54-60MWd/kgU). Close to 4500 assemblies have already achieved burnups higher than 40MWd/kgU. In this context the four ATRIUM 10 fuel assemblies first loaded in 1992 deserve special attention. These assemblies have demonstrated the high burnup capability of this design with a burnup of 71MWd/kgU after eight annual cycles of operation. In addition, 48 ATRIUM 10 assemblies have currently achieved a burnup higher than 50MWd/kgU.
Optimised and flexible operation
Plants must have the flexibility to respond to the needs of the power market for nuclear power to remain economical. As far as the reactor core is concerned, this means that operating modes such as load following, frequency control and stretch out operation must be possible. Plants must also be capable of altering their in-core fuel management strategies on short notice.
Cores with ATRIUM 10 fuel have large cold shutdown and hot excess reactivity margins. When designed for a certain cycle length, ATRIUM 10 reload fuel can thus be operated in cycles that are several months longer or shorter than originally scheduled, without restrictions on full power operation.
In many of the markets served by ATRIUM 10, operators have elected to extend their operating cycles to as much as 24 months to achieve optimum plant capacity factors and reduce net operating costs. Additionally, these plants are typically operated with spectral shift strategies that start the cycle with power concentrated in the bottom of the core and later finish the cycle with power mainly in the top of the core. The effect on fuel is to increase axial peaking relative to an evenly distributed power profile. However, this strategy allows a significant reduction in the amount of enriched uranium that must be loaded to meet energy requirements.
Average annual fuel rod failure rates have been under 2x10-5 since 1991, with the average for the period from 1993 to 2001 dropping as low as 0.7x10-5. (Note that these failure rates include failures from all causes, including debris.) Two major contributing factors in this area were the introduction of new cladding materials and improvements in fuel pellet quality. Thanks to these advances as well as the excellent capabilities of today's core monitoring systems, the probability of fuel rod failures being caused by pellet-clad interaction (PCI), a significant fuel failure cause in the past, was greatly reduced.
Altogether only eight fuel rod failures have been encountered in ATRIUM 10 fuel assemblies over a period of ten years. In four of these cases debris fretting was identified as the failure cause. For the other rods the fuel examination is not yet completed. The FUELGUARD lower tie plate, which is now introduced with increasing frequency in ATRIUM 10 reloads, will further reduce the probability that debris will enter a fuel assembly and cause a fuel rod failure.
Further performance improvements
BWR operators are pursuing changes in plant equipment and operational strategies that will yield further improvements in power production economy. As a consequence, Framatome ANP is developing advanced versions of the ATRIUM concept under the ATRIUM 10XP programme.
Stability-optimised lead fuel assemblies of the ATRIUM XP design were loaded into a European plant in 2002. These lead assemblies realise top ranking stability characteristics without compromising fuel weight. In fact, a considerable increase in fuel assembly weight was achieved compared to ATRIUM 10 to reduce fuel cycle costs. Elements of this design will be carried forward in further design developments to add improved thermal hydraulic performance characteristics among other advancements.