Teamwork With Industry - nuclear boiler repaired using multi-disciplinary approach - Part 2
TWI Bulletin, July/August 2000
John Wintle is a Consultant Engineer in Structural Integrity. He is responsible for developing reliability engineering at TWI and Inspection Qualification.
Richard Jones is Manager of the Arc and Surfacing Section of TWI where he has been involved in development and application of arc welding process technologies.
Part one of this two part feature on nuclear boiler repairs at Chapelcross power station examined the requirements of the job, the repair strategy and the validation of the repair procedure. In the final episode John Wintle and Richard Jones describe how their strategy was implemented and the way in which associated safety and inspection issues were addressed.
Guaranteeing the inspection
It was essential that the repair weld was demonstrated to be defect free by a programme of non destructive examinations in whose reliability the utmost confidence could be placed. Reliability was achieved by:
- Applying a range of diverse inspection techniques
- Building in redundancy through repeat inspections, sometimes using multiple operators
The diverse inspection techniques included magnetic particle, radiography, and ultrasonic manual pulse echo and time of flight diffraction (TOFD). Whilst the manual pulse echo ultrasonic had proven capability, the TOFD inspection was particularly powerful as an automated technique for achieving repeatability and recording data for later analysis and comparison. Redundancy was obtained by manual pulse echo inspections carried out separately by inspectors from MBEL and Eagle Star.
An Inspection Qualification Body (IQB) led by TWI generated the necessary confidence in the inspection by proving the capability of the inspection in accordance with an established European methodology. For the critical inspections, the key elements of this qualification were:
- A technical justification report providing a theoretical capability statement
- Open trials on test pieces with known defects
- Blind trials where defects in test pieces were not known to the operators
The trials were undertaken under controlled conditions in front of the IQB who later assessed the performance of the individual operators. All the operators assessed met the requirements. The IQB later reviewed the performance of the on-site inspections of the repair and satisfied itself that they had achieved their objectives. No defects of concern were detected ( Fig.1 and 2).
Fig.1. Chapelcross power station
Fig.2. Heat exchangers in gantries outside the reactor building
Development of the safety case
In order to complete the safety case, it was necessary to show that any defect that could threaten structural integrity of the repair could be reliably detected by the inspection. Consequently, it was necessary to determine a limiting size of defect that would be of no concern and to show that the inspection was capable of detecting this size of defect reliably. Smaller defects that might not be so reliably detected would be tolerable. An engineering critical assessment of the limiting defect size required the necessary material properties and stress data. Since there was no data on the fracture toughness of the dome end parent metal, weld metal and HAZ materials of the repair, a programme of testing was undertaken at TWI to determine the properties. There was particular interest in the properties of the sub critical HAZ since the dome end material was a silicon killed boiler quality steel, equivalent to BS 1501: 161C, and therefore susceptible to strain ageing embrittlement in the as-welded condition. It was also necessary to show that the controlled deposition repair procedure had produced a fine grain microstructure in the HAZ regions closer to the fusion boundary with good toughness.
As no spare plate from the dome end remained, a quantity of plate of similar vintage to that of the dome end was procured and used to manufacture a test weld. This was used to determine the properties of the weld metal and HAZ regions. Core samples of the dome plate, 5mm in diameter, were extracted from the shell and these were electron beam welded to manufacture small specimens for testing the plate properties.
The most important result from the fracture toughness testing programme was a shift in the fracture toughness of the sub critical HAZ relative to the parent plate increasing the transition temperature by approximately 70°C ( Fig.3). Welding had strain aged the material. This embrittlement observed in the as-welded condition had shed some light on the cause of the original arrested brittle fracture. Normally, this type of embrittlement is removed by post weld heat treatment, but this was not possible for the Chapelcross repair. Fortunately, the toughness was within the bounds required by the safety case. The fracture toughness of the other HAZ regions was good showing that the grain refining weld buttering procedure had been effective.
Fig.3. Transition toughness temperature shift between parent plate and as-welded subcritical HAZ of Si-killed steel
Another technical advance was the development of a method for predicting residual stresses in and around the repair weld by finite element modelling. The temperature and stress fields were incrementally updated as each pass was simulated and the model allowed for plasticity and creep effects. Using the model it was possible to show how the residual stress distribution could be optimised by altering the sequence in which the weld beads were placed and lengthening the time allowed for cooling the weld after each layer had been deposited.
The model was validated by residual stress measurements made on a simple welded test plate and the weld manufactured on the mock-up. Measurements were made using the block layering and splitting technique that had been developed at TWI. The results show that residual stresses in the weld region were generally of yield magnitude, as expected, but that they rapidly decayed outside the repair area.
AEA Technology undertook a finite element stress analysis of the boiler as a whole, and these results were combined with the residual stresses. An assessment of the stresses and the repair to the requirements of current codes of practice showed a high degree of conformity. Putting these results together with material properties data in an engineering critical assessment using the R6 methodology, BNFL determined a limiting defect size of 12mm. The inspection qualification had shown that defects as small as 6mm could be detected with high reliability.
The repair is made
It was not until all elements of the repair had been designed and justified and the processes qualified that the actual repair work on the boiler shell could proceed. A key factor to ensure that the good intentions of the preparatory work were put into practice was proper monitoring and supervision of the repair work at all stages. For this reason, the work was controlled by a detailed quality assurance programme.
A temporary weather-proof well heated enclosure was constructed on scaffolding around the repair location. The enclosure was designed to ensure that the welders were working in a comfortable position as confirmed on the full scale mock up. The machining and preparation of the repair excavation were prepared in advance ( Fig.4).
Fig.4. Machining an opening in the shell to remove the cracked material
Three teams of two welders carried out the repair welding. This enabled welding to continue around the clock without interruption. The welders operated in pairs, with only one welding at any time whilst the other recorded the welding parameter details and the other teams rested.
The deposited bead position and run out length of each electrode were recorded by observation, and the current voltage and deposition time were recorded automatically. In addition, continuous recordings were taken from strain gauges attached to the boiler shell to confirm that the stresses generated during welding corresponded with the values predicted by analysis. The welding operation was completed in about 72 hours without any undue event.
Following the initial inspection, the backing bar was removed carefully by manual grinding. The remaining site inspections were carried out according to procedures that had been qualified by the IQB. The Eagle Star inspector together with representatives of the Technical Group and BNFL plant management witnessed the whole operation.
After completion of welding and inspection ( Fig.5), the Technical Group met to review the results in the light of the safety case that BNFL had proposed. It was agreed that the welding had been to a very high standard with no significant defects. After the relevant authorities had approved the safety case, the decision was made to reconnect the heat exchanger to the reactor and recommence nuclear steam generation. The decision was reviewed and confirmed when a further inspection after one cycle of operation detected no change in the repair.
Fig.5. Section of the repair weld showing the fine grain buttering
Achievements of the project
Technically, this project made a number of significant advances and showed:
- A significant opening in a safety critical boiler could be repaired by filling with weld metal in a way that can meet the pressure systems regulations and additional requirements appropriate to nuclear plant
- Vintage silicon killed boiler quality steel, equivalent to BS 1501: 161C, can be successfully repair welded using a controlled deposition weld buttering procedure to produce a fine grain HAZ close to the fusion boundary. However,strain ageing embrittlement of the sub critical heat affected zone in the as-welded condition is an issue to be addressed in practice and by further research.
- The normal requirement for a pressure test after such a repair could be avoided on a basis including a rigorous reliable inspection proved by formal qualification.
- The residual stress analysis found that there was benefit in allowing a little more time for the weld to cool between adjacent layers.
The repair was judged by BNFL to have been an outstanding success that had returned an old vessel to full working order. For the organisations involved, this was a most satisfying result. It showed that a multi-disciplinary approach combined with careful preparation and attention to detail, and, most importantly, co-operation, could win the day.
Acknowledgements
The authors would like to acknowledge the contributions made by all the organisations involved, and particularly by Peter Mills of Matsui Babcock and Tony Pennick of BNFL.
