Inside-out weld repair at Forsmark 128 February 1998
Last autumn a highly innovative BWR vessel nozzle repair was done in Sweden, with a full load of fuel remaining in the core.
During the 1997 outage at the Forsmark 1 BWR in Sweden a crack was found on the inner surface of a reactor pressure vessel nozzle, in a circumferential Inconel 182 weld. Unexpectedly, the crack was orientated transversally.
A boat specimen was taken using underwater Electric Discharge Machining (EDM) equipment attached to the plant’s inspection manipulator. Unfortunately the sample was dropped so it was not possible to analyse the cause of cracking. This meant that there was no alternative but to repair the crack and the task became a critical path item for the outage.
An attempt was made to remove the crack by further use of EDM. But after reaching a depth of 16.4 mm the crack still remained and the remaining wall thickness was getting down to the minimum allowable according to stress analysis (the nozzle wall is 22.5 mm thick). It was therefore decided to remove the crack and subsequently make a weld repair, and to do this from inside the water filled vessel without removing the fuel.
The method chosen was to do the work from a “dry shaft” – a technique that NRG had successfully used a few months earlier at Oskarshamn 3 for installing nozzles on the steam drier.
Within five weeks the 7m high, 1.5m diameter, shaft, weighing 6t, had been designed and fabricated for the Forsmark 1 job, the welding equipment designed and assembled, qualification and factory acceptance testing done and the work on the vessel itself completed. Equipment in the dry shaft was operated remotely and these operations went particularly smoothly, the most critical task, the welding, being done in 16 hours.
WORKING IN THE SHAFT
Once at site the shaft was lifted into the vessel (which remains fully filled with water) and its top attached to the vessel flange with two bolts. The lower end of the shaft is forced against the vessel wall by two support beams linked to a hydraulic unit pushing against the opposite side of the vessel. This gives the necessary pressure for leak tightness between shaft and nozzle. After draining of the shaft, the external water provides additional pressure. No leaks were experienced during work on the vessel.
A special carrier was deployed at the bottom of the shaft for positioning of equipment for cleaning, brushing, welding and X-raying, while at the top a work platform was erected. A tarpaulin covering was placed below the work platform to prevent things from falling into the vessel (which remained fully loaded with fuel).
Further attempts were made to remove the crack with EDM. But at a depth of 17 mm the crack was still visible with TV. With just 5 mm of wall thickness remaining, the risk of melting through the wall was obvious. It was therefore decided to weld over the existing crack.
The welding itself took 16 hours. Subsequent X-ray inspection showed that it had been very successful, although it had not managed to completely melt out the crack. The weld collar was removed using further EDM and the surface eddy current tested.
Unfortunately, during this eddy current testing some indications were detected outside the welding area. However, after analysis of a boat sample these were put down to an oxidised depression that had existed since commissioning of the unit and not to welding stresses. The nozzle was therefore declared ready for operation.
NUCLEAR REPAIR GROUP
The work was carried out by ABB Atom and Uddcomb Engineering, two members of the recently established Nuclear Repair Group (the other member being Metalock Industriservice). The project was led by Ake Persson, general manager of NRG, and done in collaboration with ABB TRC.