Reliability study of friction stir welded copper canisters containing Sweden's nuclear waste

Lars Cederqvist, Tomas Öberg

Forskningsoutput: TidskriftsbidragArtikel i vetenskaplig tidskriftPeer review

17 Citeringar (Scopus)


The plan for the final disposal of spent fuel front Swedish nuclear power plants is to place it in copper canisters that are sealed and stored in a deep repository. The canisters will be sealed by friction stir welding (FSW) and the reliability of this process has now been evaluated. The reliability study was performed in three steps: first an optimization experiment to identify optimal process settings and establish the process window; then a demonstration experiment with welding tinder production-like conditions; and finally a postdemonstration series to evaluate an adjustment in the welding equipment. A process window was defined around the optimal process setting, i.e., the limits within which the welding variables must lie in order for the process to produce the desired result. In the demonstration experiment, a series of 20 scaling welds was carried out under production-like conditions. The maximum discontinuity in each weld-detected by non-destructive testing-was fitted to a generalized extreme value distribution. The 95% confidence interval for the maximum discontinuity in a production series of 4500 canisters was estimated at 4.5-7.7mm. The best estimate from the postdemonstration series suggests that the maximum size of a discontinuity will not exceed 2.3 mm. A main conclusion is that the FSW process produces reliable results, fulfilling the predetermined requirements for minimum copper thickness by a very good margin.
Sidor (från-till)1491-1499
TidskriftReliability Engineering & System Safety
StatusPublished - 2008

Ämnesklassifikation (UKÄ)

  • Produktionsteknik, arbetsvetenskap och ergonomi


Utforska forskningsämnen för ”Reliability study of friction stir welded copper canisters containing Sweden's nuclear waste”. Tillsammans bildar de ett unikt fingeravtryck.

Citera det här