News update | Safety
Cracks found at Doel 31 November 2012
Thousands of defects have been found in the reactor pressure vessel of the Belgian PWR Doel 3. While it is likely that the flaws appeared during fabrication, further investigation is needed to assess their significance. There may also be implications for other units with vessels from the same manufacturer. By Caroline Peachey
In August, it was revealed that ultrasonic inspection of the reactor pressure vessel at Doel 3 had identified the presence of potential cracks. “Numerous flaw indications in the basic steel material of the reactor vessel were detected in late June, in particular in the bottommost ring,” Belgium’s Federal Agency for Nuclear Control (FANC), said in a 16 August press release, referring to the way the vessel was made of several steel cylinders welded together. “We are considering these flaw indications very seriously,” it said.
The findings prompted the plant’s operator, Electrabel, and the regulator to launch additional investigations, which are due to be completed by end of September. Meanwhile, initial indications suggest that flaws are associated with manufacturing.
FANC has pledged to work alongside regulators of the countries concerned to provide them with information about Doel 3. On 16 August, a technical meeting of international experts was held in Brussels, aimed at taking stock of the situation and enabling the sharing of knowledge on reactor vessel integrity and inspections. FANC said it wanted to be able to “take action in a completely unbiased manner about the future of Doel 3.”
The reactor, along with a second Belgian unit, Tihange 2, remains offline for further study. Although it is too early to draw any conclusions about the ultimate fate of the units, the government has said they will not be restarted until there is proof that the defects pose no risk to the public, workers and the environment. This will not be until at least October, when experts meet again to discuss the findings of the supplementary inspections, and could be even longer. The Belgian prime minister’s office said in a statement on its website that “Belgium will not face a blackout this winter or next, even if Doel 3 and Tihange 2 remain shut throughout.”
It is not just Doel 3 that could be affected. The reactor vessel of the Doel 3 nuclear power station was built in the early 1970s by Dutch firm Rotterdam Drydock Company (RDM). That same firm (now bankrupt) also manufactured reactor pressure vessels for up to 21 other reactors around the world, according to the OECD Nuclear Energy Agency. They are: Brunsbuettel and Philippsburg 1 in Germany (both permanently shut down); Borssele and the shuttered Dodewaard in the Netherlands; Santa María de Garoña and Cofrentes in Spain, Ringhals 2 in Sweden; Leibstadt and Muehleberg in Switzerland as well as ten reactors in the United States: Catawba 1, McGuire 2, North Anna 1&2, Quad Cities 2 in part, Sequoyah 1&2, Surry 1&2, and Watts Bar 1. (The US NRC’s own list of plants with RPVs mostly built by RDM?is similar, but excludes Quad Cities 2 and includes Watts Bar 2 (which is under construction)).
Despite sharing the same manufacturer, the vessels often used different materials or fabrication techniques, so are not necessarily likely to be affected by similar phenomenon, regulators have said. The unit at most risk appears to be Belgium’s Tihange 2, whose vessel was also forged by RDM at the same time and under the same contract. That unit went offline 16 August, and will undergo the same inspections as Doel 3, according to FANC. Similar inspections are planned at Tihange 1 (in 2013) and ‘eventually’ at Doel 4 and Tihange 3, according to the regulator.
The US Nuclear Regulatory Commission told NEI on 6 September that it sees no reason for US reactors to take action at this stage; other national regulators have made similar statements, although some (such as Spain’s CSN) are carrying out further investigations, and others (such as the Swedish Radiation?Safety Authority)?are requiring utilities to do so.
There could also be wider implications for other reactors. The NEA said in a statement, “Inspections of the reactor pressure vessel (RPV) are usually carried out about every ten years, but it would be prudent to use this type of operating experience to ensure that future inspections check the reactors for these kinds of micro-fissures.”
Doel: the facts
On 2 June, the Doel 3 pressurized water reactor was shut down for its ten-yearly overhaul. The outage entailed more than 10,000 maintenance tasks and checks, including examination of the 13-metre-high, 4.4-meter diameter, 330-tonne reactor pressure vessel.
Since 1982, when the reactor begun operation, several in-service inspections have been performed on the vessel. Different areas are inspected on each occasion, which means that after each period of ten years all the sensitive zones have been inspected. During the recent inspection, examination focused on zones prone to cracking, including weld zones (between the rings of the vessel) and duct tubes in the reactor circuit.
French company Intercontrole (a member of the AREVA group) conducted this and all prior vessel inspections at Doel 3. The company is also in charge of the same inspections at French reactors. During the recent outage, Intercontrole carried out an innovative ultrasonic inspection for Doel 3, aimed at detecting defects under the cladding; an approximately 7 mm thin stainless steel liner that protects the vessel from corrosion. (The vessel walls are up to 205 mm thick). A FANC spokeswoman said that the inspection technique was an immersion technology with pre-focus mono-crystal probes operating in pulse-echo mode in four directions with three focal depth zones. She also said that similar UT inspections of the RPV head and upper rings in the 1990s found only a few indications.
Although no defects were found under the cladding, numerous flaws of a different type were discovered, especially in the bottommost of the three forged rings that make up the vessel. The flaws appeared to be ‘almost circular in shape’ and ‘running more or less parallel’ to the inner/outer surface of the vessel, covering a region about 1.5 meters high, 120° large, and 13.5 cm deep, according to the Belgian regulator’s technical support agency, Bel V.
The event was initially classified as a level 1 (anomaly) occurrence on the International Nuclear Event Scale (INES). Regulator FANC said that the flaws are ‘theoretically not dangerous’ as they run parallel to the vessel wall and, as such ‘are normally not subject to stress.’
More detailed ultrasonic testing commenced 16 July, ultimately revealing over eight thousand ‘quasi-laminar’ flaws in the upper and lower rings, combined. Preliminary findings confirm that the bulk of the flaws are located in the base material, outside the weld regions, in a zone extending from about 30 mm from the inner surface to one half of the RPV thickness. The circular flaws have an average diameter of 10-14 mm, although some have diameters as large as 20-25 mm, Bel V said. There is a marked disparity in the flaw densities between the upper and the lower shell rings, according to Bel V.
However, it noted that the shape of the flaw distribution is ‘very similar’ in both the upper and lower shells. “The core lower shell is the most affected with a total of 7776 indications. The core upper shell contains 931 indications. The other parts of the reactor vessel contain some indications, but to a lower extent, and these are of a different nature in the transition ring.”
As of 3 September Electrabel was continuing to examine these additional inspection results.
To put the discovery in perspective, the technique to inspect for defects under the cladding has been carried out in France since 1991, according to the French regulator, ASN. To date, 37 defects have been detected on the forged rings (core shells) that make up the French reactors’ vessels, including 20 on Tricastin 1. “The undercladding defects identified in France are isolated and well-defined. They do not affect the quality of metal over a large area and the reasons for their occurrence are known. As such, they are not comparable to the defects detected on Doel 3's vessel, where several thousands of clustered defects have been observed,” ASN said in a statement on 20 August.
ASN also said that the manufacturing defects detected in France are “periodically checked and do not evolve while operating.”
Causes of the cracks
FANC said that the defects found in the Doel RPV “are usually associated with manufacturing and are not due to ageing.” It is ‘improbable’ that the flaws have evolved since their formation, according to Electrabel.
A number of different companies participated in the fabrication of the Doel 3 reactor pressure vessel. German firm Krupp supplied the raw material for the reactor shells; Dutch firm Rotterdamsche Droogdok Maatschappij (RDM) performed the forging; while cladding and assembling was carried out by Cockerill for the lower part (two core shells, transition ring and bottom plate) and by Framatome for the upper part (RPV head, nozzle shell), and the final assembly.
Documentation recovered by Electrabel and Tractebel Engineering gives information on most of the fabrication steps, including data on the chemical composition, and in particular hydrogen content of the blooms, semi-finished metal castings, supplied by Krupp. Certificates proving the conformity with the fabrication specifications have also been retrieved. But some information is missing, notably detailed documentation about the first thermal heat treatment performed by RDM, an intermediate UT inspection and some RDM specifications, according to Bel V. Metallurgical investigations support conclusions that the flaws appeared during fabrication, and principally due to high hydrogen content.
Bel V explains: “During the casting of the ingots, some segregation zones inevitably develop. During the cooling steps, hydrogen diffuses more into the segregation region, and accumulates there. If the initial hydrogen content is sufficient, the hydrogen accumulation may imply the formation of flaws...Experience gained over several decades shows that it is possible to avoid the formation of these flaws if the hydrogen content is kept below a certain level, if long annealing is applied, and if a dehydrogenation step is carried out,” Bel V said. “During the fabrication of the shells of the Doel 3 reactor vessel, no trace of a dehydrogenation treatment at RDM can be found.”
Inspections carried out during the fabrication stages of the vessels did not lead to detection of the flaws. This could be because the inspections, which involved UT controls and volumetric examinations as well as some other controls for welds, were performed manually and from the external surface of the vessel. The recent inspections were more sensitive and carried out from the inner surface of the vessel wall, according to FANC.
FANC said that the flaws observed during the 2012 inspections would not necessarily have lead to reject the shells according to the applicable design criteria of the ASME Code. The inspections at the fabrication stages were carried out following the rules of the ASME Code Section III, edition 1974.
Electrabel plans further non-destructive examinations to obtain more information on the flaws and on their behaviour.
This article was first published in the October 2012 issue of Nuclear Engineering International