Irrespective of the core or cycle designs and operating modes chosen by the utilities, the electric power market is demanding unprecedented levels of fuel reliability. Despite the decline of fuel failure rates to a low level, reliability remains an issue. Failures or problems with the fuel or core components can be extremely costly and very difficult to cope with. The pressure to simultaneously increase both fuel reliability and fuel cycle economy is industry-wide, affecting both PWRs and BWRs. Not only are the fuel materials and designs of concern; so are plant operations.
In the last decade both fuel suppliers and plant operators have expended great efforts to improve fuel performance. But despite this, fuel failures and related problems have persisted. Some of these problems have appeared as ‘surprises’ – new and unanticipated problems associated with the new designs and their interaction with plant operations. These new designs offer important economic benefits through extended burnup and the flexibility of longer fuel cycles. However, unexpected problems can interfere with the realisation of these benefits. Recent problems have included axial offset anomalies in PWRs, first cycle fuel failures in both a PWR and a BWR, incomplete control rod insertion in PWRs and failed fuel degradation in BWRs.
In addition to the need to ensure good fuel performance and reliability, as well as improved fuel cycle economics, there are persistent regulatory issues that need to be resolved. For example, the US Nuclear Regulatory Commission (NRC) has recently interpreted the 17% loss of coolant accident (LOCA) oxidation limit to include the sum of pre-transient and post-transient oxide thickness. The NRC is also conducting ‘confirmatory research’ to see if reactivity initiated accident and LOCA limits should be reduced for high burnup fuel. It is essential for the utilities to formulate sound technical positions to deal with the NRC positions on transient fuel behaviour in high burnup conditions.
In response to these needs, US utilities established the Robust Fuel Programme (RFP) in January 1998. The programme includes participation by the fuel suppliers and it is managed by the Electric Power Research Institute (EPRI). Its goal is to resolve presently known problems and avoid future surprises. It has a three-pronged approach: responding to known problems; developing key technologies to assure problem-free advanced fuel products at extended burnup levels (>62GWD/T); and assuring awareness of potential regulatory issues, and acquisition of the data needed to resolve them.
Addressing failures
Potential fuel-related problems are listed in the table at right.
To assess and monitor fuel performance EPRI has developed a fuel reliability database. An example of its output is shown in the diagram above right, which plots annual fuel assembly failure rates for both BWRs and PWRs since 1980.
Fuel failures have persisted throughout the 1990s, albeit at a lower rate than in the 1980s. Grid fretting was a major cause of failures in PWRs throughout 1998. Although vendors have applied engineering fixes to this problem, the effectiveness of the fix has yet to be convincingly proven. The problem involves the interaction between assemblies of different designs and core flow. To test the ‘fix’, flow test facilities that can accommodate more than one assembly are needed. One of the fuel suppliers has moved to create such a facility.
It is essential to examine failed fuel rods to determine the root cause of their failure in any plan to prevent, or if necessary to mitigate, similar failures. The rods should undergo non destructive examination at the plant sites and should be shipped to hot cell laboratory facilities for more detailed examination. This requires a high degree of co-operation between the affected utility, the fuel supplier, EPRI, and the hot cell laboratory. The RFP helps this co-operation by providing independent technical insight and a programme structure.
Changing times
New fuel designs, which have already been accepted by operators, involve new materials and manufacturing processes. While the designs have been changing, there have also been changes in water chemistry.
Chemical additives to the water are designed to protect the non-fuel core components and plant systems, as well as reduce radiation dose to operating personnel. Thus BWR plants are operating with increasing levels of hydrogen, with zinc and sometimes noble metal additives. In PWRs the higher enrichments in the newer fuel designs require higher concentrations of boron in the water, which must be balanced by higher concentrations of lithium to control pH. Increases in power density have resulted in an increasing number of PWRs operating with significant boiling. In PWRs, concern about cladding corrosion inhibits use of optimum pH. Although each of these changes was introduced gradually and with what was considered adequate testing before application, some very significant surprises have occurred in which water chemistry is an important factor.
Crud related fuel failures in PWRs and BWRs are summarised in the table on p17. The problem has persisted, at least with BWRs, into 1999. These fuel rod failures are due to the heavy, dense crud deposition that raised the temperature at the oxide-metal interface, leading to accelerated waterside corrosion of the fuel cladding.
The failure statistics indicate that PWRs have not experienced severe crud-induced accelerated corrosion since incidents at Three Mile Island 1 in 1995 and Seabrook in 1997. However, PWRs have been affected by crud-related operational problems such as axial offset anomalies, which are thought to be caused by boron trapped in the crud layer deposited on the high temperature sections of fuel rods. This is an example of problems that can affect high-efficiency core designs, which push the limits of operating margins.
Water chemistry remains difficult to understand and is plant-specific. Diligent monitoring and scrupulous control of water chemistry is essential for the success of plant operations and for fuel protection. Questions have been raised as to whether the current version of the EPRI Water Chemistry Guidelines is sufficient to provide full protection to the fuel in all modern, high efficiency cores. The RFP is working closely with the EPRI Water Chemistry Guidelines Committee to re-evaluate and revise them as needed. Proposed and presently used water additive agents are also being considered, tested and evaluated.
Dealing with surprises
In BWRs the last decade has been marked by some fuel rod failures that have resulted in remarkably high offgas. Several plants were forced into unscheduled outages as a result. The copious release of fission product and transuranic species from failed BWR fuel rods results from the formation of long splits in the cladding.
One of the earliest published reports of this phenomenon reported on the degradation of a sponge Zr-liner fuel rod in Oskarshamn 3 in Sweden, which resulted in a 133Xe offgas level of around 200MBq/s; around 100 times a ‘normal’ leak rate for a failed rod. Examination of that rod revealed splits extending almost the entire rod length – although they were not continuous. A study of an extensive General Electric database on the performance of failed fuel rods in BWRs confirmed the correlation of long axial splits with the intolerably high releases of radioactivity into the coolant.
It was known that the degradation was related to the secondary hydrating of the cladding, and the Zr-liner design was suspect because of its rapid rate of corrosion and contribution to hydrogen formation. However, the phenomenological understanding of the cause of these long axial splits and high activity releases to the coolant have come to light only in the past two years. There is still disagreement as to whether the rate of crack growth is limited by the diffusion of hydrogen in the cladding. Even in the absence of full mechanistic understanding, it is clear that the rate of hydrogen generation by the corrosion of the Zr-liner is an important parameter. All BWR fuel suppliers have instituted engineering solutions, by adjusting the composition of the liner material to reduce the corrosion rate in steam. There is good reason to believe that the newest fuel products will not be susceptible to such post-failure degradation, but the industry will have to monitor the situation to be certain.
The degradation mechanism has been studied with an in-reactor test at the Halden Project, under the auspices of the Nuclear Fuel Industry Research group. The NFIR/Halden test and procedure could be used to evaluate new fuel designs for their susceptibility to post-failure degradation. This is another example of the potential of industry wide co-operation.
The introduction of Zr-liner – or barrier – fuel for BWRs solved the problem of fuel rod failure in response to sudden power ramps. The failure mechanism is known as pellet-cladding interaction (PCI). Fuel with Zr lining has been in service since the early 1980s, with a very good record of successfully resisting failure by PCI. However there have recently been reports of this fuel failing by PCI. It has always been known that the barrier fuel product is resistant to PCI and not immune to it, so it was not too surprising when cladding failure by PCI was confirmed and then associated with missing pellet surface. Irrespective of pellet irregularities, as the compositions of the Zr-liner materials are altered to reduce the tendency for post-defect degradation, the possibility that the fuel cladding will have less than full PCI resistance must be considered. The RFP has confirmed that barrier fuel is not immune to PCI failure and is evaluating the PCI effectiveness of additive fuel.
A current problem unique to PWRs is the phenomenon of incomplete control rod insertion. Obviously this can have safety and licensing implications and must be resolved completely. The basic cause is believed to be the irradiation growth and excessive compressive stress on the fuel assembly hardware, which can cause bending of the guide tube. Each of the PWR fuel suppliers has reevaluated their designs and ‘fixed’ the problem. The industry, in concert with the RFP, is monitoring the results.
Pressures for improvement
The recent trends to increased burnup and cycle lengths at PWRs and BWRs is being pushed by the high cost of outages, the need for high plant capacity factors, and the need to limit the amount of discharged spent fuel. Cycle lengths of 18 months are now common for BWRs and PWRs; 24 month cycles have been implemented by some units.
A reasonable current target is burnup levels of >62MWd/kgU lead rod average. One important limit is the 5%U-235 maximum enrichment in fabrication, shipment and storage; there has been no serious move to increase this.
Within the RFP there was a study of the optimal burnup for the Oconee and Catawba PWRs in the Duke Power grid. Optimum levels were 60-70MWd/kgU batch average. However, as one increased the costs for spent fuel disposal or for fabrication, the optimal burnup increased, limited only by the 5% enrichment barrier. Similar studies for different utility power mixes and economic circumstances are contemplated by the RFP. A key goal of the RFP is to provide assurance that economically attractive fuel also has adequate margins for safety and technical licence limits, as well as complete reliability.
Achieving consensus
An important aim of the RFP is to achieve an industry consensus on performance requirements for future fuel designs. EPRI has issued a report, Robust Fuel Program Technical Requirements for Nuclear Fuel, with inputs from utilities and fuel suppliers. The objectives of this study were to identify challenging, yet achievable performance requirements that will assure adequate margins, and also the means for determining whether desired performance requirements are being met. The consensus reached involved the requirements on performance, not design. These technical requirements involve expected margins set so that the next generation of fuel products are less likely to cause performance ‘surprises’.
The EPRI document also considered the process for implementing fuel design changes. For example, it stipulates that whenever a fuel design is to be pushed to burnup >5MWd/kgU more than was previously attained, lead fuel assemblies (LFAs) should be specified. These should be operated to the target burnup and evaluated before subjecting additional assemblies to the new burnup level. Further, if a specific fuel design is to be used at a more severe duty cycle than previously experienced, the assemblies subjected to the most severe duty should be designated as the LFAs. In both cases, increased burnup or increased duty severity, a surveillance programme should be instituted to provide early information on potential problems related to that product and in those service situations. The requirements for the surveillance programme are also stipulated.
The target for peak fuel rod discharge burnup with an enrichment limit of 5% is 75MWd/kgU rod average for PWRs and 70MWd/kgU for BWRs. The burnup rates that could be associated with fuel to reach these levels are also stipulated. To achieve these target burnup rates will require extension of the allowable burnup limits by the NRC. The EPRI document should provide an industry discipline which will help in obtaining NRC approval.
As the industry adopts new designs and pushes fuel to high burnup under severe service conditions, LFA programmes are essential to gain the experience base and confidence needed for success. This is formally recognised in the report, but we must cautiously acknowledge that LFA programmes, as currently contemplated, might not be sufficient to catch an unanticipated, low-probability problem. LFAs intrinsically involve relatively small numbers of fuel rods – usually between four and 16 assemblies. Any event with <1% probability could easily be missed and yet be the source of significant difficulty in full commercial operation, with more than 50,000 rods in a single core. The answer to this dilemma is not only very thorough evaluation of the LFAs, but also extensive nondestructive examinations of new products at the reactor sites after the first full reload of a partial core has been placed in service. A cautious approach with co-operative, thorough evaluations before the adoption of new products for full-core service is needed to avoid surprises.
Obtaining the data
To realise high performance targets, data is necessary. The RFP will attempt to obtain the data needed and detailed examinations of poolside and in hot cells are planned.
The presently available BWR fuel is from Limerick LFAs, with a rod burnup average of 57MWd/kgU. Hot cell examination of this fuel is scheduled during 2000, with completion in mid-2001. This will be an interim examination. Similar fuel from Limerick with burnup levels at or near the target burnup will be discharged from the core in 2001. Their hot cell examination cannot begin until 2002.
For PWR fuel we are planning to obtain samples with advanced cladding types at high burnups, but since these designs and materials are relatively new, the burnup levels have not yet been achieved. Data on Zirlo clad fuel will not be available at burnups in excess of 70MWd/kgU until the latter part of 2001. The examination of fuel from other suppliers, such as the Framatome Cogema Fuels M5 clad fuel, the ABB-CE Alloy ‘A’ and Siemens’ fuel with Duplex cladding is at the planning stage.
Under the auspices of the Nuclear Fuel Industry Research Group, basic data is being obtained at a variety of laboratories. These data are used to supplement those from integral fuel rod tests and examinations. Cladding issues being examined by the NFIR include: the effects of hydrides on mechanical properties and on crack propagation; transient response of high burnup cladding in rapid burst tests; in-reactor cladding creep and growth; and development of a qualification test for fracture behaviour. The issues for fuel pellets include: post irradiation examinations on samples irradiated to 50, 75 and 100 MWd/kgU; various pellet compositions such as UO2, UO2 with gadolinia, soft pellets (additive fuel); the effect of grain size on rim formation; and the impact of ZrB2.
Nondestructive examination
Because it is difficult to ship spent fuel to hot cells, and the examinations are prohibitively expensive, as much information as possible has to be obtained by nondestructive examinations at poolside. The RFP is evaluating and enhancing poolside examination capabilities, focussing on performance limiting parameters. For example, there is a need for constant oxide thickness data processing and reporting so that measurements reported by one group are directly comparable to others. There is also the need to advance eddy current techniques, for correction for crud and the development of calibration standards for oxide thickness measurements. Another need is for rapid nondestructive measurement of released fission gas within many rods and for the measurement of fuel assembly deformation. This effort to enhance the industry’s capability to acquire poolside data should be completed in 2001.
UO2 pellets with 0.25% aluminosilicates have been shown to be virtually immune to failure by PCI in severe power ramp tests at modest burnup levels. However, since the aluminosilicate additive displaces four times its volume in urania, optimum additive concentrations are needed to make this fuel type economically attractive. A follow-on power ramp testing programme has been initiated to further quantify the PCI effectiveness of this fuel at different additive concentrations. These will include power ramp tests at burnup to 21MWd/kgU. Of course, further work on this fuel type must await the results of these early tests. If successful, this fuel type offers the potential for an economically attractive new fuel optimised for PCI resistance with increased resistance to post defect degradation.
Complementary work
The NRC is conducting confirmatory research to determine the effect of burnup on safety criteria such as those for LOCA and RIA. Some RIA simulation experiments have resulted in relatively low failure thresholds. The NRC has also begun an experimental programme at the Argonne National Laboratory (ANL) to assess the effect of burnup on the criteria for LOCA and the mechanical properties relevant to RIA. EPRI, through the RFP, is providing fuel for the NRC’s test programme. BWR fuel from Limerick with 57MWd/kgU will be shipped to ANL for tests. There are also plans to provide high burnup PWR fuel from HB Robinson at 72MWd/kgU. A goal of the RFP is to assure availability of the data needed to support licensing of future fuel designs to targeted burnup levels.
The industry has mobilised resources to create very reliable fuel that is economically attractive and free from surprises. Under the RFP reactor operators and fuel suppliers are co-operating. Agreement has been reached on the technical requirements for future fuel designs, and an ambitious test plan has been started for examination of fuel at appropriate burnup and of materials and designs that are relevant to the current and immediately foreseeable future.
The industry is also co-operating with the NRC and ANL to obtain data to aid in the resolution of safety issues for high burnup fuel. The RFP has established criteria and procedures to assure that new fuel products will have adequate performance margins.
