PDC IC WELD FAILURE EVALUATION AND RESOLUTION

During final preparations for start of the PDCF Inner Can (IC) qualification effort, welding was performed on an automated weld system known as the PICN. During the initial weld, using a pedigree canister and plug, a weld defect was observed. The defect resulted in a hole in the sidewall of the canister, and it was observed that the plug sidewall had not been consumed. This was a new type of failure not seen during development and production of legacy Bagless Transfer Cans (FB-Line/Hanford). Therefore, a team was assembled to determine the root cause and to determine if the process could be improved. After several brain storming sessions (MS and T, R and D Engineering, PDC Project), an evaluation matrix was established to direct this effort. The matrix identified numerous activities that could be taken and then prioritized those activities. This effort was limited by both time and resources (the number of canisters and plugs available for testing was limited). A discovery process was initiated to evaluate the Vendor's IC fabrication process relative to legacy processes. There were no significant findings, however, some information regarding forging/anneal processes could not be obtained. Evaluations were conducted to compare mechanical properties of the PDCmore » canisters relative to the legacy canisters. Some differences were identified, but mechanical properties were determined to be consistent with legacy materials. A number of process changes were also evaluated. A heat treatment procedure was established that could reduce the magnetic characteristics to levels similar to the legacy materials. An in-situ arc annealing process was developed that resulted in improved weld characteristics for test articles. Also several tack welds configurations were addressed, it was found that increasing the number of tack welds (and changing the sequence) resulted in decreased can to plug gaps and a more stable weld for test articles. Incorporating all of the process improvements for the actual can welding process, however, did not result in an improved weld geometry. Several possibilities for the lack of positive response exist, some of which are that (1) an insufficient number of test articles were welded under prototypic conditions, (2) the process was not optimized so that significant improvements were observable over the 'noise', and (3) the in-situ arc anneal closed the gap down too much so the can was unable to exhaust pressure ahead of the weld. Several operational and mechanical improvements were identified. The weld clamps were changed to a design consistent with those used in the legacy operations. A helium puff operation was eliminated; it is believed that this operation was the cause of the original weld defect. Also, timing of plug mast movement was found to correspond with weld irregularities. The timing of the movement was changed to occur during weld head travel between tacks. In the end a three sequential tack weld process followed by a pulse weld at the same current and travel speed as was used for the legacy processes was suggested for use during the IC qualification effort. Relative to legacy welds, the PDC IC weld demonstrates greater fluctuation in the region of the weld located between tack welds. However, canister weld response (canister to canister) is consistent and with the aid of the optical mapping system (for targeting the cut position) is considered adequate. DR measurements and METs show the PDC IC welds to have sufficient ligament length to ensure adequate canister pressure/impact capacity and to ensure adequate stub function. The PDC welding process has not been optimized as a result of this effort. Differences remain between the legacy BTC welds and the PDC IC weld, but these differences are not sufficient to prevent resumption of the current PDC IC qualification effort. During the PDC IC qualification effort, a total of 17 cans will be welded and a variety of tests/inspections will be performed. The extensive data collected during that qualification effort should be of a sufficient population to determine if additional weld process optimization is necessary prior to production release.« less