Chairmen: H.G. Riella /PEN. Sao Paul0
H. Roepenack Siemens AG, Hanau
AND FUEL CHARACTERISTICS
Journal of Nuclear Materials 178 ( 1991) 187-l 94 Noah-Ho1iand
Foundations for the definition of MOX fuel quality requirements H. Bairiot’, P. Deramaix’, N, Mostin’, E. Trauwaert’ and Y. Vanderborck’ ’ Be~gon~cle~ire,Rue du ~h~rn~s de Mars 25, B-IO50 Basils,
‘Re~gon~c~e~~re, Dessel, 3elgiu~
The quality of uranium-plutonium mixed oxide (MOX) fuel, as of any nuclear fuel, depends on the design optimization and on the fabrication process stability. The design optimization is essentially based on feed-back from irradiation experience through engineering assessment of the results; the stability of the process is necessary to justify minimal uncertainty margins in the fuel design. Since MOX fuel is quite similar to U02 fuel, the lessons learned from UOZfuels can complement the MOX experimental data base. MOX is however different from U02 fuel in some respects, among others: - the industrial fabrication scale is a factor 10 lower than for UOZfuel, - the fuel enrichment process takes place in the manufactu~ng plant, - the radioactivity of Pu imposes handling constraints, - Pu ages quite rapidly, altering its isotopic composition during storage, - the incorporation of Pu alters the material physics and neutronic characteristics of the fuel. In this perspective, the paper outlines some quality attributes for which MOX fuel may or even must depart from UOZfuel. 1. Introduction: LWR fuel quality achievements Nuclear fuel is now a reliable product, with an almost zero failure rate. Most failures originate from causes not attributable to the intrinsic quality of the fuel: debris carried by the coolant, baffle jetting, handling damage during reshuffling, waterside corrosion, etc. It has been demonstrated that even these external causes of failure can be minimized by remedies incorporated into the fuel assembly design: debris-filtering bottom end pieces, local additional spacing grids (or clamps), redesigned external straps of the spacer grids, improved cladding, etc. The fuel specific failure rates are now in the order of a few rods in 1O6irradiated fuel rods. Under these circumstances, the quality of fuel must be assessed in reference to a large population with adequate reliability record: fabricated fuel should not deviate from what was fabricated before and proven to be good. As a result, the specifications are increasingly based on a requkment that the product characteristics fall within the variability range observed in the past fabrications (and hence considered achievable by the manufacturer), rather than on a control of attributes in direct relation with good behaviour requirements. In summary, the product is of good quality, if QC data provide evidence that it does not deviate from Elsevier Science Publishers B.V. (North-Holland)
the product successfully utilized previously. This trend turns out to be cost effective, since some product quality controls quantifying attributes directly related to fuel behaviour can be replaced by process controls to assure stability and reproducibility. It has also been instrumental in improving the quality awareness of everybody involved: the acceptance or rejection of non confo~ing material is no longer based only on the results of the quality controls but requires proper explanation by the manufacturer of the reason(s) why this material is deviating and, based thereon, evaluation by the designer of potential side-effects unrevealed by the specified quality control (QC) plan but potentially hazardous for fuel behaviour. This approach is only possible and valid since the necessary quantity of fuel rods (i.e. half a million) on which confidence is based for the present failure level is fabricated in less than one year in the largest fuel manufacturing plants (over 1000 tU/year ) or in a few years in the smaller fuel manufacturing plants.
H. Buiriat et al. / F~und5tions~oF the de~nition
2. Specific situation of MOX fuel
~f~O~~e1 quality requj~e~e~t~
- MOX fuel must be an engineered quality product. 2.2. Similarities to .?Jfuel
2.1. The statistics
MOX fuel is always part of a reload, 70% (or more) of which consists presently of U fuel. In order to benefit to a maximum extent from the statistically demonstrated quality level achieved by the U fuel which constitutes the major part of the reload batch, as many as possible features of the U fuel assemblies are adopted unchanged for the MOX fuel assemblies. This is the case for the assembly structure (hardware, mechanical and thermal-hydraulic design of the assembly) and the fuel rod hardware (a.o. cladding tube). In these respects, MOX fuel is identical to the industrial U fuel whose quality is well established for the intended use. The same principle could not be applied for the MOX pellet. Trying to match the U pellet characteristics of each different fuel vendor would be impracticable and detrimental to a reproducible quality. As a result, the decision was taken to develop our own product (now the MIMAS fuel; MIMAS: MIcronited MASterblend) and to care for its quali~cation and its acceptability to customers unfamiliar with the product. After exploratory research in various directions in the 196Os, the decision was taken 20 years ago to utilize as raw material AUC Pu02 powder, in the first part because of its intrinsic characteristics that make it well fitted to MOX fuel manufacturing conditions, but in the second part also in the belief that MOX
MOX (uranium-plutonium mixed oxide) fuel manufacturing plants are subcritical in respect to statistically significant experience accumulation. The past operation experience of the MOX fuel produced by Belgonucleaire (fig. 1) and by Siemens [ 6,7 ] is impressive, but falis short in providing confidence at a level of 10m6 on the fuel rods basis. An industrial MOX plant of the present generation (i.e. with a 35 tM/year (M: heavy metal; i.e. U + PU + Au ) capacity [ 11) would need to operate at full capacity during 25 years to produce a total of half a million fuel rods. The largest MOX manufacturing plants now under construction (100 tM/year) would each need to operate at full capacity during 10 years to produce this statistically significant quantity of MOX fuel rods. The quality approach for U fuel (U: UOZ fuel made of enriched U ) can therefore not be applied as such to MOX fuel. Indeed, the MOX fabrication industry, today at a production level typical of U fuel fabrication 10 to 30 years ago, is confronted with the challenge to produce fuel of today’s quality and of assuring this quality to the same confidence level as U fuel. This challenge is met by adopting two guidelines: - the as-fabricated MOX should be as similar as possible to an industrial U fuel with good quality records;
assemblies i- Fuel~~~--
GRA 3&GRA 4
39 1 Fuel assemblies 41 141 Fuel rods
, . .._-_
Fig. 1,Belgonucleaire MOX fuel operation in LWRs.
H. Bairiot et al. /Foundations for the definition of MOXfuel quality requirements
fuel so produced would resemble in many respects the AUC type U fuel which is one of the wide-spread and proven industrial fuels. 2.3. Differences to U fuel One basic difference is that MOX fuel manufacturing combines in one MOX fabrication plant what is done in two plants for U fuel: the enrichment plant and the fabrication plant. A consequence is the necessity to control, at an adequate confidence level, the Pu content of the fuel and the uniform distribution of this Pu. Moreover, unlike enriched UF6, the as-received Pu is largely non-uniform raw material. Its isotopic composition varies considerably due to its origin and history [ 11. It depends on the reactor type from which the spent fuel assemblies are unloaded, the actual nuclear characteristics of this reactor, the bum-up reached by each reprocessed fuel assembly and the period since the end of the irradiation and since reprocessing (table 1) [ 41. Additionally, the specific characteristics of plutonium (toxicity, radioactivity, criticality, pyrophoricity) impose restrictions which influence the plant lay-out, the sequence of fabrication steps and the execution of the different control operations. All the fabrication steps and the execution of the different destructive control operations are performed in interconnected leaktight glove boxes with plexiglass windows and equipped with neoprene gloves, and, where ever required, additional gamma and neutron shielding. Indeed plutonium is essentially an alpha emitter but also a neutron, X- and gamma-ray emitter. The reprocessing produces “clean” plutonium oxide in
which the quantity of hard gamma emitting elements (daughter products of 236Pu), the amount of 24’Am (daughter of 24’Pu) and the neutron activity begin to increase from that moment. To keep the radiation levels to the personnel below a certain limit, it is important to fabricate the MOX fuel as soon as possible after the reprocessing of the fuel. The manufacturing and control technique must be adapted to achieve both a low dose rate for the personnel and a high reliability of the fabricated MOX pellets, fuel rods and fuel assemblies. 2.4. Engineered
Besides the differences pointed out under the previous heading, experience has shown that it is impossible to fabricate a MOX pellet completely identical to the U pellet taken as reference. The presence of Pu and its variable stochiometry with temperature, which is more pronounced than for U and not coherent with it, induce specific features distinguishing MOX fuel from U fuel. Moreover for identical irradiation conditions, MOX fuel would anyway respond differently than identical U fuel, for two reasons: - the different nuclear characteristics of Pu and 235U induce important differences in the time dependent power density distribution across the pellet; - the internal fuel rod chemistry and its evolution with burn-up are quite different for MOX and for U fuel. MOX fuel can therefore not be merely a copy of U fuel. It must be engineered on its own on the ground of an adequate experimental data base and, arising therefrom, a fundamental understanding of the fuel as fabricated. This
Averageisotopic composition of plutonium produced in uranium-fuelled thermal reactors [ 41 Reactor
Percentage of Pu isotopes at discharge
33000 43000 53000
1.3 2.0 2.7
56.6 52.5 50.4
23.2 24.1 24.1
13.9 14.7 15.2
4.7 6.2 7.1
Information not available.
H. Bairiot et al. / Found~~i~ns~orihe d~~niiion @“.340Xfuel quality r~quire~~nis
can only be achieved after a long period since typically 10 years (table 2) are required between the perception of a need-to-know-more and the feed-back into industrial fuel. This is the reason why the existing MOX manufacturing facilities are backed by a strong fuel engineering team, which provides the required assistance on the basis of timely acquired experience and integrates the lessons learned into progressive fuel improvements or extended fuel utilization.
Table 3 On-going MOX research programmes Fuel thermomechanical behaviour BWR fuel
DOMO Dodewaard MOX irradiation to 60 GWd/t Ramp testing
PRIM0 Phase 1 BR3 MOX irradiation to 55 GWd/t Ramp and transient testing PRIM0 Phase 2 BR2 water Ioop irradiation to 60 GWd/t Ramp testing GEMINI NOK MOX irradiation to 60 GWd/t Ramp testing
2.5. The e,~per~menta~data base Fabricators of MOX fuel cannot afford the luxury of keeping static or following a slower evolution than U fuel vendors. For a MOX fuel manufacturer, the competitive challenge is not another MOX fuel manufacturer, but the U fuel itself: the MOX fuel cycle cost must be competitive to the U cycle cost. If U fuel is improved or is demonstrated to be capable of extended performances, MOX fuel most equally and with a minimum delay be similarly improved or proven to be capable of equivalent performances. Meeting such challenge requires a strong R & D and quali~~tion program, inco~rating tests to match all the aspects of fuel behaviour, at the limit and even beyond design basis requirements. Such a factfinding effort has been pursued without interruption for almost 30 years [ 21 and today is able to back MIMAS fuel with an adequate data base [ 61, Several organizations world-wide have perceived the need for a coordinated development and data acquisition effort to broaden the MOX fuel utilization field and reduce the uncertainty margins to be taken into account in design or safety evaluation. As a consequence, the growth of the MOX data base relies to a significant extent on international pro~ammes pursued as collaborative ventures between the partners (table 3) and sponsored by the most representative organizations from Belgium, Table 2 Timescale required for experimental data acquisition on high bum-up fuel Year(s) Programme definition + fuel fabrication & characterization Irradiation Post-irradiation examination Evaluation of the results+ implementation into the technology package
1-3 5-7 2-3
Neutronic behaviour VIP-BWR VENUS critical facility experiments BWR MOX and UO1 configurations
VIP-PWR VENUS critical facility experiments PWR MOX and UOz configurations
3. Specific quality attributes The balance of this paper will illustrate some quality attributes linked to fabrication or to control aspects which are specific to MOX fuel. 3. I. UO, and [email protected]
powdery To assure the quality of the MIMAS pellets, specific requirements for U02 and PuOp have been developed in agreement with the suppliers.
3.1.1. Powder supplies The powder suppliers have been qualified and a quality audit is performed at least every 2 years. In this way, all the inspection tests are made by the powder suppliers according to inspection, testing procedures and frequency mutually agreed upon. Accordingly the incoming inspection performed at the MOX manufacturing plant is limited to nondestructive testing (NDT) to verify compliance of the delivered products with the certificates, However a check analysis is performed at least every year on one sample. The NDT control of Pu quantities benefits furthermore from the in-
H. Bairiatet al. / Foundationsfor the de~~iti5nof~OXfue~ qualityretirements
dependent checks made by the joint team EURATOM/ IAEA for safeguards purposes. 3.1.2. Powder characteristics The U02 powder quality is one of the major factors which determines the quality of the final MOX fuel pellet microstructure and the consistency with the data base. Unlike PuO*, U02 is fortunately delivered in quite large homogeneized lots. The impurity content of the raw material is limited both as a total contamination and in many case as individual values. Particular attention is paid to: - boron, cadmium, gadolinium, rare earth contents (absorption aspects); - chlorine, fluorine (corrosion aspects); - silicon, aluminium (risk of eutectic formation). For the other impurities, MOX fuel is less restrictive than U fuel (neutronic aspects). The standard isotopic composition corresponds to depleted uranium (0.25 O /o 235U)_This is a sensitive attribute, as it cannot be verified in the final product: the presence of Pu masks the response of the rod scanner to the 23sU content and its potential variations from pellet to pellet. The destructive tests performed on MOX pellet samples are inadequate for providing confidence in the uniformity of the *“U content of the product. As a result, the only quality assurance approach, is to ascertain that only one single U feed was utilized during the manufacturing campaign. This restricts the flexibility of the plant in switching back and forth from depleted to natural uranium and prevents the use of reprocessed uranium. Anyway in most of today’s MOX specifications, the tolerance limits of 235Uare unduly restrictive and should be broadened by proper engineering backup either on a generic basis or through assessment of any non conforming pellet lot. The Pu02 powder is delivered in small lots (fig. 2) [ 11. The responsibility for reliability of the chemical analysis (in particular for the rare earths) and isotopic composition is on the reprocessor side because a repetition of such analysis as an incoming inspection would be prohibitive in cost and delays and incompatible with an industrial production. Particular attention is paid on halide content (corrosion aspects). For the other impurities, the’subsequent large dilution of PuOZ powder by U02 allows less severe limits. To avoid an unacceptable dose rate for the personnel, which is a result of hard gamma emitters (increasing with age), soft gamma emitters (almost proportional to age) and neutron emitters (increasing with age), the 24iAm is
Nbr. OF LOTS
Fig. 2. Typical weight variation of received Pu lots.
taken as a eonvenient expression of the plutonium age and is presently limited to 10000 ppm maximum, 12 months after being made available for fabrication. 3.2. Powder treatment and pellets The MIMAS process (fig.3) adopted since the end of 1984 is characterized by the ball milling of a master blending ( UOZ, PuOZ powders and scraps) at 25-300/o Pu content followed by a secondary blending where the master blend is diluted with UOZ to give a mixed powder for pelletizing at the required Pu content ( l-10% Pu). Concerning the specific cross blending approach outlined in fig. 4, a unique computerized system has been developed and demonstrated, to achieve a high level of homogeneization in Pu isotopic composition for a large amount of powders. Indeed, the reproducibility of the manufactured product [ 1] relies a.o. on the constancy of the feed material and must be reached with feed materials delivered in batches of very different sizes and isotopic compositions. A particular characteristic of MOX fuel is the plutonium distribution at the microscopic level. Whereas the early fabrications showed important heterogeneities of the plutonium distribution, the present MIMAS fabrication is almost free of heterogeneities and is from this point of view also much closer to U fuel than the MOX fuel prepared by the previous processes. The average effective diameter of the plutonium rich
H. Bairiot et al. /Foundations for the dejinition of MOXfuel quality requirements
ROD FILLING END PLUG WELDING
Fig. 3. MOX fuel fabrication process (MIMAS). particles is smaller than 50 pm, 95% of these particles have a diameter smaller than 100 pm and no particle reaches the 400 pm size threshold for safety aspects [ 5 1. It must be emphasised that the maximum plutonium content of these particles is limited to 25-30% as given by the master blend composition. During irradiation, the as fabricated heterogeneities disappear progressively by the burn-up effect (production of plutonium in the U matrix and burnup of plutonium in the Pu rich particles) and by U-Pu interdiffusion process. During the sintering step, the control of the furnace parameters, such as temperature profiles and gas chemistry is one of the key factors to guarantee an appropriate and stable microstructure (for the specified density) [ 3, 81 and a low hydrogen content of the pellets. By progressive improvement of the microstructure, the open porosity must be minimized to reduce the hydrogen pick-up, the fission gas release (which tends normally to be larger for MOX fuel than for U fuel but is now within the scatter band of U fuel), and the dust retention during the dry centerless grinding (which could induce a weld contamination). The homogeneity of MOX fuel microstructure from pellet to pellet has a paramount importance for the fuel behaviour.
,A, Fig. 4. Plutonium
As fabricated, MOX fuel has a grain and pore size distribution similar to U fuel prepared with the same UOZ powder type and sintered in similar conditions [ 21. Scrap addition induces local lenticular porosities which have a beneficial effect for the plasticity of the MOX pellet and which probably partially explains the better behaviour of MOX fuel in ramping conditions. Though it appears [ 3 ] very difficult to judge MOX fuel reprocessability in the unirradiated state, the reprocessor imposes for the MOX a specific solubility test on fresh MOX fuel. Different testing procedures have been discussed. In order to test reprocessability, a test involving the dissolution of entire pellets in 5.5M HN03 for 6 h is used as proposed by the reprocessor to be representative of the dissolution unit of the reprocessing plant. In the routine production, this test provides usually, but not always, residues smaller than 0.2% expressed as the amount of undissolved Pu with respect to the initial quantity. It is very sensitive to the open porosity level and has no relation whatsoever with the existence or not of an adequate (U, Pu)02 solid solution. Conformance with such solubility test would encourage the MOX fuel manufacturer to maximize the open
H. Bairiot et al. /Foundations
for the dejinition
porosity but with a detrimental consequence on the fuel behaviour. Therefore strict enforcement of this test shouid be discouraged unless it is simply considered as a preliminary fast screening test, the final reception criteria being based on a more elaborate testing procedure performed on the lots which failed the screening test. This secondary more elaborated test procedure should determine the plutonium which is not in solid solution with UO2 and the rejection based thereon. The control of surface defects of the pellets refers to visual standards precisely defined in sketches, At present, the same defect levels as for U fuel are tolerated, relaxation of the criteria should be considered for MOX due to its better plasticity behaviour compared with U fuel. Remote control by cameras could then be applied. This may be necessary in the future, since this time consuming control step contributes significantly to the dose rate for the operators. For design reasons both for PWRs as well as for BWRs, a fabrication campaign contains MOX fuel rods with different plutonium contents. Hence care must be taken (as for BWR U fuel) to prevent pellet mixing during fabrication: only one plutonium content is in fabrication on the line at one time, the pellets are marked in the dish area for Pu fissile content and the line is cleaned and inspected when switching from one plutonium content to the next one.
appropriate device has been designed and built, to reduce the contact between the MOX pellets and the tube edge during loading and a procedure of cleaning the edge to be welded has been developed. The rewelding of a new end plug in case of rejection is not possible because the tooling in glove boxes does not allow it and the risk to have a too high level of fixed contamination after this operation would be so high that a proper decontamination of the tube end is in this case almost impossible [ I 1. To perform directly a high quality weld in a glove box and to optimize its X-ray examination sensitivity, a proprietary design has been adapted for the inner part of the upper end plugs (fig. 5). It has been used with success for all PWR and BWR rods supplies. The nitrogen retention at the weld level is strondy limited by the quality of the pure helium stream around the electrode and by the monitoring of the helium atmosphere in the glove box around the welding cell (no contact with ambient air). This quality aspect is checked by autoclave tests on fabrication samples. As a final control, the rod scanning integrates different measurements in one single automatic equipment. Besides the length measurements (fuel column, plenum lengths) and the evaluation of gaps between pellets, it allows a direct measure of the plutonium content in the rod on a pellet per pellet basis and makes the individual weighing of pellet stacks as in the U fuel fabrication unnecessary. Moreover, on the same equipment, a special
3.3. Fuel rods OLD END PLUG
The use of MOX instead of U fuel has no influence on the speci~~tion of the cladding tube or on the quality of the bottom end plug weld. Indeed, even for high burn-up objectives, the U and the MOX fuels behave similarly and no specific effect from the fuel side must be expected for the use of MOX fuel. The base policy is therefore to adopt the cladding and the bottom end plug from the U fuel manufacturer currentiy qualified for the power plant. The cladding tube can be Zircaloy-4, Zircaloy-4 “low tin”, Zircaloy-2 with or without liner in different metallurgical conditions and with different surface treatments (pickled, sand blasted, autoclaved, etc). Tubings with bottom end plug already welded are delivered and fully documented under the customer’s res~nsibility. Only the upper end plug weld is performed in glove box conditions. As a result, more care is needed to guarantee the cleanliness of the tubing inside and the welding equipment and staff must have a very high qualification level to cope with a multitude of different cladding materials and fuel rod geometries. To limit the fixed contamination in the upper weld, an
NEW END PLUG
Fig. 5. Comparison between standard UO1 and Belgonucleaire end plug design.
H. Bairiot et al. /Foundations
for the deJnition
camera system verifies the marking of the bottom end plug versus the plutonium content of the rod and a counter measures the fixed contamination of the top end plug. By using such an equipment, the overall quality of each individual rod is checked and recorded accurately with practically no impact on the dose exposure of the personnel. 3.4. Fuel assemblies The fuel assembly hardware requirements are the same for U fuel as for MOX fuel. The procedures for assembling BWR or PWR MOX fuel assemblies are similar to the ones followed for U fuel. Here again, the routine hardware and procedures for U fuel are applied without modification, except for additional radiation protection of the personnel. A special inspection is made of the positioning of the fuel rods in the loading magazines versus the Pu content identification of the rods. The rods are pulled through the different spacer grids from the bottom end plug to prevent any risk of scratches on the top end plug weld, for contamination transfer reasons. A record of the zoning pattern is performed by photography (the end face of the bottom end plugs is specially marked to visualize the plutonium content of the fuel rod). The final control of the fuel assembly is performed as for U fuel assemblies: for PWR fuel assemblies, it is performed on a control tower with motorized carriage moving vertically to drive the electronic transducers. A large use of different callipers is made to limit the duration of the controls and the resulting dose rate to personnel which is checked by gamma and neutron counters.
tered in similar conditions. These characteristics give MIMAS pellets in-reactor properties approaching very closely those of U pellets. The experimental results accumulated so far on MIMAS fuel demonstrate that the behaviour of MOX fuel has always been as good and even superior to U fuel. However, the presence of plutonium in these pellets means there are significant differences in the fabrication or control aspects which leads to the delivery of an engineered product fitting the most stringent customer specifications. By coping with the specific Pu-related constraints in the fabrication and/or control stages, high-quality MOX fuel can be readily fabricated with a consistent reliability. Industrial experience shows that it works.
References [ I ] E. Trauwaert,
[ 61 4. Conclusion [ 71 Due to its micro-structural characteristics - grain size and pore size distribution - MIMAS fuel is similar to the U fuel prepared from the same UOz powder type and sin-
N. Mostin and R. Lefevre, MOX fabrication experience at Dessel, Presentation at a Specialists’ Meeting, Cadarache, November 13- 16, 1989. H. Bairiot and M. Lippens, Commercial MOX fuel: The experimental and demonstration background, Presentation at a Specialists’ Meeting, Cadarache, November 13- 16, 1989. J. van Vliet, D. Haas, F. Eeckhout, Y. Verschuere and E. Trauwaert, MOX fuel quality and reprocessing requirements, Presentation at a Specialists’ Meeting, Cadarache, November 13-16,1989. Plutonium fuel: an Assessment (OECD/NEA, Paris, ISBN 92-64-13265-1, 1989). A.F. Renard and N. Mostin, in: Proc. Conf. on the Characterization and Quality Control of Nuclear Fuels. Karlsruhe, June 1978, J. Nucl. Mater. 8 1 ( 1979) 3 1. J. van Dievoet, H. Bairiot, G. Lebastard, H. Pekarek and H. Roepenack, MOX fuel and its fabrication in Europe, PBNC, San Diego/CA, March 4-8, 1990. P. Schmiedel, Recycling of uranium and plutonium, Nucl. Europe Worldscan 3-4 ( 1990 ) . R. Gilldner and H. Schmidt, in these Proceedings, J. Nucl. Mater. 178 (1991) 152.