Assessment of the safety of MDF fuel in use

An Investigation into the Falsification of Pellet Diameter Data in the MOX Demonstration Facility at the BNFL Sellafield Site and the Effect of this on the Safety of MOX Fuel in Use

4.1 - Factors Affecting Fuel Safety in Use
4.2 - NII's Views

74. When NII was first informed of the falsification of fuel pellet diameter data, one of the major concerns was to gain an understanding of the potential impact upon the safety of fuel that was either operating in reactors or, as was the case in Japan, about to be loaded into the reactor. NII asked BNFL to provide a justification for why it believed fuel safety would not be prejudiced, and NII fuel specialists assessed the response. In this section of the report the key fuel rod characteristics which can affect the safe performance and reliability of the fuel assembly in the reactor are discussed, along with NII views on the measures taken in MDF to ensure that the key parameters which affect these characteristics are within the required fuel assembly specification.

75. BNFL commenced production of LWR MOX fuel via the Short Binderless Route (SBR) in 1990 and since then it has been engaged in a MOX fuel development programme to demonstrate the safety of SBR-MOX fuel under normal and abnormal reactor conditions. This programme has included:

characterisation and physical property measurements on unirradiated pellets;
in-pile testing in both test and commercial reactors;
post irradiation examination of commercially irradiated fuel; and
the development of fuel performance modelling capability through the use of computer codes.

Factors Affecting Fuel Safety in Use

76. Most manufacturing defects in fuel will not precipitate a nuclear accident. These require a significant power-coolant mismatch and occur as a result of a loss of cooling due to a variety of causes, or from a loss of control of neutron power causing the fuel rods to overheat. Manufacturing defects or irregularities, if outside specific tolerances can cause the fuel rod cladding to fail giving rise to the release of fission products into the reactor coolant circuit. This does not cause an immediate threat to the public or the power plant workers but it will increase the contamination levels in the reactor and make maintenance activities more difficult. Most nuclear reactor designs anticipate some level of fuel failure and have coolant cleanup systems to remove fission products. However, to avoid unnecessary circuit contamination, nuclear plant operators require their fuel to have high levels of reliability.

77. There are a number of fuel pellet and fuel rod characteristics which can influence reliability and hence safety and operability performance in the reactor. Tables 2 and 3 list the quality characteristics (for fuel pellet and fuel rod respectively) which the fuel manufacturer has to address. Fuel rod failure mechanisms and how they can be influenced by fuel pellet and rod quality characteristics are detailed in Appendix 3.

78. Fuel rods are designed to produce the required power output within conservative specifications which are aimed to produce very low manufacturing defects (which can result in clad failure as discussed above), and be tolerant both to normal operation and to a range of fault conditions. The designers, in specifying the fuel requirements, make allowances for the effects of irradiation during normal operation. Therefore, the most important characteristics which can have an influence on fuel rod integrity during operation in the reactor are:

Fuel pellet cracking Fuel densification Fuel pellet swelling Fission gas release from the fuel matrix Fission product migration Clad creepdown Fuel rod growth Clad corrosion and crud deposition Thermal conductivity (of the fuel/clad gap and the fuel itself)

79. In relation to the MDF fuel, the fuel rod cladding and other assembly components are supplied by the customer so for the purpose of looking into the likely implications of fuel pellet diameter overinspection falsification, NII concentrated on the role and significance of the fuel pellet characteristics.

80. The key characteristics are those associated with the physical properties of the fuel pellet, ie pellet density, fuel grain size, porosity, and homogeneity of the uranium and plutonium oxides, the surface condition including chips and cracks, the shape of the pellet and the pellet dimensions.

Fuel Pellet Physical Properties

81. A number of quality checks are made of the physical properties of the pellet. Geometric density must be within a specification range to a 95/95 confidence level; this is calculated for a random sample of 20 pellets per Lot by measuring diameter, length and weight. Resinter behaviour is tested to demonstrate that the material is thermally stable; 10 pellet samples taken at regular intervals must meet the specification on geometric density following high temperature extended sintering. PuO₂ particle size and grain size are measured using colour alpha autoradiography of ceramography samples; two pellet samples are taken at regular intervals. Sellafield Analytical Services measure a number of chemical characteristics, in particular the plutonium enrichment. At a later stage in the process, when the pellets have been inserted in fuel pins, the enrichment of each pellet is checked using a special detection system which records any unusual enrichments.

Fuel Pellet Surface Condition

82. Every pellet is visually inspected for chips, cracks, surface defects and major shape abnormalities. Surface roughness must be within the specified limit, and random samples of five pellets are taken at regular intervals. The surface condition of the pellets is important because pellet chips can cause clad stress intensification, and a fuel crack can cause clad wall-thinning during operation.

Fuel Pellet Dimensions

83. As shown in Tables 2 and 3, there are several dimensional checks which are important to fuel quality. Accurate measurement and confidence in the pellet's diameter is important because if the pellet is too large it will not fit into the cladding tube. If it is too small it may move around and possibly cause clad collapse: cracking of the clad gives rise to fission product release. BNFL recognises the importance of this and provides several quality checks on the process. The key questions are how much reliance can be placed on the 100% automatic measurement of pellet diameter and the failsafe system for rejecting pellets which are out of specification; what additional value does the 200 pellet manual AQL check give; what value can be gained from the diameter measurements taken for density measurements, and what support in relation to diameter measurement can be derived from the 'throat bush' test for oversized pellets and the 100% fuel rod radiographs for undersized pellets.

84. There is a robust case for saying that the 100% primary diameter check alone will provide adequate confidence that all pellets are within specification. In order to obtain some indication of the sensitivity of fuel performance to pellet diameter, NII asked BNFL to provide a safety assessment. BNFL used the ENIGMA code which had been jointly developed with British Energy. This was backed up using the Westinghouse PWR design and licensing code, PAD. The range of pellet diameter examined was nominal ± 0.2 mm (this compares with the MOX fuel pellet specification range of nominal ± 0.0125 mm. The key performance parameters examined were:

peak in-life fuel centre line temperature;
end of life fission gas release;
end of life rod internal pressure;
peak in-life clad hoop strain;
peak in-life clad concentrated hoop stress.

85. The results showed that only beyond the range of ± 0.1 mm of nominal, ie 8 x the specification range, could the effects become major. BNFL claims that in the event of the 100% diameter check not working as intended, the throat bush used as part of the rod filling process would prevent oversized pellets being loaded. The company equally claims that undersized pellets of 0.1 mm less than nominal would be detected by the fuel rod radiographs. The overall conclusion by BNFL was that the accuracy of the 100% automatic check, plus the relatively low sensitivity of the relationship between fuel pellet diameter and fuel rod failure, was such that the absence of the AQL check would not impact on the ability of the fuel to perform safely in reactor operation. 86. The total length of fuel in a fuel rod is important because it determines the length of the fission gas plenum and hence has an effect on internal rod pressure, which if incorrect could result in fuel rod failure as explained in Appendix 3.

This is measured in the following way: approximately 730mm sub-stacks of pellets are length checked and weighed prior to being manually loaded into the neck of the fuel rod. After loading the pellets, the free space in the rod is checked with a bar with two notches defining the acceptable limits. This check is for process control only and is not part of the certification. 87. For certification the correct length of the fuel stack is checked by measurement of the plenum length and the overall length of the fuel rod. The plenum length is measured from the radiograph of the fuel rod. The radiograph is also examined to check the absence of gaps. For the current order, the customer checks every weld radiograph and a random selection of the full-length radiographs. 88. The fuel manufactured in MDF for commercial purposes is covered by customer approved Quality Control Plans. These define for each characteristic the requirements of the specification, the method of analysis or measurement, the frequency of measurement and/or sampling, and how the information is recorded There are a number of characteristics/parameters whose quality is controlled and assured. For the pellets, these include physical and chemical properties, surface condition, dimensions and shape. X-radiography of the rods is carried out to check that there are no gaps or unintended materials in the fuel column.

NII Views

89. NII is satisfied that the fuel manufactured in MDF will be safe in use in spite of incomplete QA records caused by the falsification of some AQL data by process workers in the facility. The NII takes this view on the basis of the robustness of the fuel manufacturing process and the totality of the checks made on the key parameters.

90. Firstly in relation to the physical properties of the SBR-MOX fuel pellets, the MOX demonstration programme coupled with BNFL's extensive expertise in oxide fuel manufacture and NII's examination of the plant and processes carried out in the MDF fuel pellet production area are such that NII is confident that the MOX fuel pellets produced in MDF are of the required quality and will perform as designed in the reactor.

91. Second in relation to surface condition, NII is satisfied that the rod radiography check (which is examined by both BNFL and the Japanese customer) will detect crack, chips or other defects which are outside the customer's specification.

92. Finally NII is satisfied that the 100% automatic check on fuel pellet diameter is sufficiently robust to ensure that only fuel pellets which will not prejudice the safety of fuel pins in operation are used in fuel rod assembly. Comfort is also taken from the throat bush which limits the upper diameter and the radiograph checks to detect undersized fuel pellets. The combination of pellet density, stack length and stack weight provide further reassurance as to mean pellet diameter.

Added to the HSE website on 18th February 2000