Resolution of Generic Safety Issues: Item B-22: LWR Fuel (Rev. 2) ( NUREG-0933, Main Report with Supplements 1–34 2)
Individual reactor fuel rods sometimes fail during normal operations and many fuel rods are expected to fail during severe accidents. To ensure that these fuel failures do not result in unacceptable releases to the public, regulations have been promulgated that require a coolable, rod-like geometry to be maintained. The behavior of fuel under normal and accident conditions must be predicted to ensure that the regulatory requirements are met. Because many factors impact the behavior of the fuel under both normal and accident conditions, e.g., fuel rod bowing, densification, cladding fragmentation, etc., this issue was identified in NUREG-04713 to determine whether such predictions were reliable.
Failure of fuel rods results in a radioactive release within a plant and is a potential source of release to the public. To remain within the regulatory requirements for acceptable release rates, a plant may be required to shut down or operate at a reduced power level. The behavior of reactor fuel must be predictable and these predictions must be reliable to ensure that: (1) operating limits are not exceeded; (2) activity releases are limited; and (3) the fuel system does not degrade by an unacceptable amount.
A possible solution to this issue involved developing methods to reduce fuel failures and to limit activity releases. These methods can work if fuel behavior can be predicted for normal operations and postulated accident conditions and if operating limits are set in accordance with these predictions.
In an attempt to control fuel failures, some design changes in fuel assemblies were made by GE. However, it was believed that most of the efforts to resolve this issue would be in trying to predict or alter power excursions and reduce releases by controlling such events as steam generator leakages. Although controlling power excursions may not be feasible under accident conditions, it is feasible during planned power maneuvers when the potential for fuel failures may be limited.
Accident frequency is principally a function of equipment reliability and the frequency of accident initiators. Neither of these factors is influenced by the presence of failed fuel. Thus, the influence of failed fuel on accident frequency is extremely small.
Fuel defects increase releases during normal operation and transients, but these increases would remain within 10 CFR 20 limits. The magnitude of a release from failed fuel during an accident is much larger than the release from a pre-existing fuel defect, and the release during a severe accident is not expected to be significantly affected by pre-existing fuel defects. Thus, the resultant consequence from failed fuel was determined to be very small.64
Industry Cost: The cost to implement the possible solution would be primarily for training. It was assumed that some minor fuel assembly design changes would be made and that power excursions would be altered during planning power maneuvers. Implementing these changes would only require training the appropriate operations staff.
Training personnel for minor fuel assembly design changes would require a minimal amount of staff labor. It was estimated that 2 man-weeks would be required to train all personnel handling fuel assemblies or supervising refueling techniques. Changes in operations reflected by alterations in power excursions would require that all reactor operators and supervisors receive appropriate training. Five shifts/plant were assumed to be available with one of these being the training shift. It was assumed that the following personnel would receive 10 man-hours of initial training, followed by additional training on their respective training shifts:
Shift Supervisor (1/shift)
Unit Supervisor (1/shift)
Assistant Shift Supervisor (1/shift)
Reactor Operators (2/shift)
Shift Technical Advisor (1/shift)
Reactor Engineer (1)
Day Shift Supervisor (1)
The initial training procedures for changes in operations would involve 35 personnel. Assuming a 10-hour initial training program, it was estimated that 350 man-hours or 8.75 man-weeks would be required. Thus, the total labor required was 10.75 man-weeks/plant and the industry cost to implement the possible solution was ($2,270/man-week)(10.75 man-weeks)/plant or $24,000/plant.
NRC Cost: The NRC cost was estimated for three categories of activities: (1) develop a fuel behavior model; (2) support implementation of the solution; and (3) review operations and maintenance activities.
Technical support would be needed to establish PCI failure predictions and their correlations to release rates. The cost of developing the appropriate models was estimated to be $200,000. NRC review of implementation procedures and training procedures at individual plants was expected to require approximately 2 man-weeks/plant or $4,540/plant. Once the solution is implemented, a review of the results from power-maneuvering changes was anticipated to be required for 5 years. In addition, the improvement in any fuel assembly design changes will need to be assessed. This review effort was estimated to require 1 man-week/RY over the 5 years subsequent to any changes at a cost of $11,350/plant. Thus, the total NRC cost to implement the possible solution was approximately $216,000/plant.
Total Cost: The total industry and NRC cost to implement the possible solution was estimated to be $240,000/plant.
No calculations were performed for this issue since the potential risk reduction from the possible solution was negligible.
Occupational Dose Factors: Localized dose rates are increased by fuel defects, as is the activity in liquid and gaseous release paths. If fuel failures exist, plants must maintain off-gas release rates within allowable limits. If a plant exceeds these limits or fails to clean up high activity in the coolant, a shutdown is required. In less severe cases, the plant will often stay within the allowable release limits by maintaining reduced levels of power until the next refueling outage. Because of the allowable release limits imposed on plant operations, the reduction in public risk resulting from the possible solution was expected to be negligible. In addition, the number of refueling outages was not expected to change significantly because of the resolution. Therefore, the changes in ORE were expected to be negligible.
Cost Avoidance: When a fuel failure occurs, all attempts are made to keep the plant in operation until a scheduled refueling outage. In most cases, this is possible; however, this procedure often means that a reduced level of power must be maintained in order to remain within allowable release limits. It is difficult to predict whether plants will shut down, remain operating at reduced levels for long periods of time, or remain operating at reduced levels for relatively short periods of time. In this issue, all plants were assumed to remain on line at the reduced level of 50% power for 30 days. This was assumed to be an average power level and an average period of time until a scheduled outage, considering the circumstances. Because of the reduced levels of output, the major potential expense was the cost of replacement power. The value assumed for the cost of replacement power during an outage was $300,000/day.64 Therefore, under conditions of 50% output, the cost of replacement power would be $150,000/day and the cost of replacement power for 30 days at a plant operating at a 50% power level would be $4.5M.
Although all plants would be affected to some degree by the possible solution, the influence of the solution on core damage frequency and public consequence was insignificant. Fuel manufacturers have taken an active role in resolving the issue ever since it was identified. As a result, fuel failures are now rare and the significance of this issue has been diminished. The cost savings from replacement power far outweighed the costs associated with implementing the reactor manufacturers' changes.
Several problems were identified in the effort to improve the predictability of fuel performance and these were addressed by the staff in its revision to SRP11 Section 4.2 in 1981. In addition to the economic incentive to the industry in reliably predicting fuel performance, existing requirements were adequate to ensure continued low fuel defect rates; additional requirements would not decrease the number of fuel defects significantly. Therefore, this issue was DROPPED from further pursuit. In an RES evaluation,1564 it was concluded that consideration of a 20-year license renewal period did not change the priority of the issue.