| 2. |
- Ronneteg, Ulf, et al.
(författare)
-
Reliability in sealing of canisters for spent nuclear fuel : Final report, Swedish Energy Agency Project 31634-1
- 2006
-
Rapport (övrigt vetenskapligt/konstnärligt)abstract
- The reliability of the system for sealing and inspection that has been developed for the future Encapsulation plant was investigated. In the investigation the occurrence of discontinuities that can be formed in the welds was determined both qualitatively and quantitatively. The probability that these discontinuities can be detected by nondestructive testing (NDT) was also studied. The friction stir welding (FSW) process was verified in several steps. The variables in the welding process that determine weld quality were identified during the development work. In order to establish the limits within which they can be allowed to vary, a screening experiment was performed where the different process settings were tested according to a given design. In the next step the optimal process setting was determined by means of a response surface experiment, whereby the sensitivity of the process to different variable changes was studied. Based on the optimal process setting, the process window was defined, i.e. the limits within which the welding variables must lie in order for the process to produce the desired result. Finally, the process was evaluated during a demonstration series of 20 sealing welds which were carried out under production-like conditions. Conditions for the formation of discontinuities in welding were investigated. The investigations show that the occurrence of discontinuities is dependent on the welding variables. Discontinuities that can arise were classified and described with respect to characteristics, occurrence, cause and preventive measures. To ensure that testing of the welds has been done with sufficient reliability, the probability of detection (POD) of discontinuities by NDT and the accuracy of size determination by NDT were determined. In the evaluation of the demonstration series, which comprised 20 welds, a statistical method based on the generalized extreme value distribution was fitted to the size estimate of the indications obtained with NDT. The predicted maximum discontinuity size in connection with the welding of 4500 canisters at the present stage of development of the process was conservatively determined to be less than one centimetre. All factors considered, the predicted minimum copper coverage for a 5 cm thick canister is 4 cm. Acceptance criteria for permitted process settings in the welding process in a future sealing system are proposed, as is the use of statistical process control based on nondestructive testing as an independent inspection system. Furthermore, principles for handling of process nonconformances are presented.
|
|
