Development of Fatigue Failure Criteria for Darlington Fuel Bundle End Plates

Abstract

In November of 1990, normal refuelling operations at the Darlington Nuclear Generating Station Unit 2 (DNGS2) were prevented by fuel bundle failure at the outlet end of channel N12. Subsequent inspections revealed that a number of high-powered channels had end plate cracks on the outlet fuel bundles predominantly initiating at the notch formed by the weld between the end plate and the fuel pencil end caps. To identify the cause of fuel bundle failure, a multi-disciplined investigation was launched into the problem. As examinations on the failed fuel bundles in the hot cells at Chalk River Laboratories later confirmed fatigue as the mechanism causing end plate failure, an experimental out-reactor program was conducted to investigate the response of fuel bundles under dynamic loading conditions. The objective of the program was to identify whether axial and/or lateral vibration of fuel elements could contribute to end plate (fatigue) failure and if so which of the two is the likely mode causing end plate failure observed at DNGS2. The room temperature axial bundle fatigue tests conducted in this experimental program produced through wall end plate failures. The end plate fractures were associated with the welds and the point of initiation appeared to be at or near the tip of the weld notch between the end cap spigot and end plate. Fracture surface examination revealed some similarities to the field failures, however distinct differences were noted in fracture location on the end plates and the presence of a "Black Eye" fracture pattern associated with most DNGS2 failures. Results from the lateral vibration tests were also similar to field failures; however significant differences existed in the amount of fretting and the higher degree of secondary cracking associated with the tests compared to field failures. The results from the tests indicated that both axial and transverse bundle vibrations can cause end plate (fatigue) failure. However, the tests were unable to duplicate all of the fracture features unique to the DNGS2 failures. It is conceivable that the inability to accurately reproduce particularly mode of hydraulic loading, differences in fractures. in-reactor conditions contributed to the marked differences in fractures.