How Calvert Cliffs is dealing with the PV embrittlement issue

27 August 1998



Baltimore Gas & Electric’s Calvert Cliffs station is the first nuclear power plant in the United States to make a licence renewal application. Before BGE could submit its application, the plant had to undertake a major engineering and programmatic evaluation. One of the most critical and complicated issues is neutron embrittlement of the reactor vessels. To resolve this issue, BGE developed a distinctive, rigorous approach – which included going back to the past – to demonstrate that the vessels will meet NRC’s tough safety criteria for a further licensing period of 20 years.


The US Nuclear Regulatory Commission (NRC) licences nuclear plants for an initial period of 40 years. In order to extend operation a licence renewal application for a further 20 year period has to be made. Although the two Calvert Cliffs units had only been in operation since 1974 and 1976, BGE anticipates a rather lengthy procedure before a final decision is taken, as its application will be the first to be dealt with by the NRC.

The ageing issue of most concern is embrittlement of the pressure vessels. (A description of vessel fabrication is given in the panel.)

During operation, the beltline of the reactor vessel is subjected to a relatively high flux (>1010 n/cm2s) of fast neutrons (E>1 MeV) which leads to a potential for neutron embrittlement. This level of exposure to fast neutrons can cause significant increases in the nil-ductility transition temperature as well as substantial reduction in the upper shelf energy. Industry and NRC research has shown neutron embrittlement sensitivity can be related to the presence of copper and nickel.

VESSEL MATERIAL EVALUATION

Evaluation of material and vessel susceptibility to neutron embrittlement is performed according to NRC regulations (Chapter 10 of the Code of Federal Regulations, Part 50.61, Appendices G and H, and Regulatory Guide 1.99). Reg Guide 1.99 provides the essential framework for preparing an analysis. Basically, an initial, unirradiated value is determined for a reference nil-ductility temperature (RTNDT). Neutron irradiation causes the RTNDT to shift to a higher (less tough) value. A shift that is either calculated or measured is added to the initial reference toughness value. Finally, a margin term is added so that the resultant value for RTNDT is very conservative. Calculation of shift is accomplished using values for the copper and nickel content and the neutron fluence in a correlation formula.

Alternatively, the shift may be measured if the correct material is included in the vessel surveillance programme, that is, irradiated in surveillance capsules. Since the capsules are closer to the core than the vessel wall, they accumulate fluence at a greater rate than the wall. The shift measured in the material within the capsules will therefore be greater than that measured at the vessel wall.

THE YANKEE ROWE AFFAIR

In 1991, when Yankee Rowe nuclear plant pursued licence renewal, uncertainty about vessel material composition and properties meant that the embrittlement of the vessel could not be established. Uncertainty in material properties and composition eventually became severe enough that the problem grew from a relicensing issue into an intractable operating one. Ultimately the utility elected to cease operations as no cost-effective way forward to resolve the issue could be found. Against that backdrop, the NRC decided to review and revise its reactor vessel integrity regulations, and BGE began to pursue licence renewal.

NRC regulations (10 CFR 50.61) contain screening criteria that identify the maximum value of RTNDT (shifted for irradiation and with the margin term added). The shifted reference toughness is termed “RTPTS” and may not normally exceed 132°C for axial welds. In the late 1980s the NRC published Reg Guide 1.99 revision 2 which modified the correlation used to predict embrittlement shift. BGE evaluated the impact of the revised correlation and recognised that it would cause unit 1 axial weld seam 2-203 (see figure) to exceed the 132°C axial weld screening criterion before 2000.

At that time, a number of activities were identified that would support extended operation. First, Calvert Cliffs unit 1 had been one of the demonstration vessels for original determination of the screening criteria, so results of probabilistic analysis were known, and these supported safe operation above the screening criteria. The NRC had Regulatory Guide 1.154 that described how to perform a probabilistic safety assessment, and prescribed this analysis as one acceptable course of action for materials that would exceed the screening criteria. Since BGE had already performed the analysis, and knew the plant specific results would yield a higher screening criteria than the generic analysis performed under 10 CFR 50.61, our initial plan was to perform the Reg Guide 1.154 analysis.

However, the experience at Yankee Rowe forced us to rethink this. When Yankee Rowe strove to extend their operating licence, they attempted to use the Reg Guide 1.154 analysis to justify operation with a hypothetical RTPTS above the screening criteria. They failed. The reason was that analysis results can vary significantly with input assumptions. Yankee Rowe and the NRC were ultimately unable to agree on a set of realistic input assumptions, and so could never achieve similar results. For all practical purposes, the possibility of successfully performing a Reg Guide 1.154 analysis was eliminated by the Yankee Rowe experience. BGE needed to develop a different approach.

THE CALVERT CLIFFS APPROACH

While the NRC was revising Reg Guide 1.99, Rev 2, Calvert Cliffs took additional action to address predicted embrittlement. Initially, the composition of weld seam 2-203 was not well known, so we assumed conservative, generic values for copper and nickel content. The assumed values were much higher than typically measured for this material so we embarked on a rigorous search for material property information. BGE located all fabrication records associated with the entire NSSS system, including the reactor vessels. This provided insight into how and when the welds in question were made. Combustion Engineering fabricated a number of vessels simultaneously, side by side on the production line. Duke Power’s McGuire unit 1 vessel was fabricated using the same weld wire, flux, and weld processes as the controlling weld (2-203) for Calvert Cliffs unit 1. Not only had the weld been used in fabricating the McGuire vessel, but it had been included in the surveillance programme. This meant that the composition and mechanical properties of the weldment had been characterised. Furthermore, surveillance capsule analysis provided measurements of shift that facilitated more accurate determination of RTPTS.

Using the mechanical properties and chemical analysis information available from the McGuire surveillance weld in the prediction correlation from Reg Guide 1.99, BGE was able to demonstrate a significant reduction in RTPTS. McGuire had pulled and analysed two surveillance capsules that both provided embrittlement shifts that were significantly lower than the correlation would predict. If BGE used the McGuire surveillance data directly in evaluation of the Calvert Cliffs vessel, the RTPTS for weld 2-203 could be shown to be below the screening criteria indefinitely. BGE acquired some of the surveillance weld material, performed additional chemical analyses on it, and fabricated Charpy specimens for inclusion in a replacement surveillance capsule. The McGuire surveillance material was inserted into the Calvert Cliffs unit 1 vessel in 1987.

Precedence existed for applying surveillance data from one vessel to analysis of another vessel. For Babcock & Wilcox plants, flow induced vibration caused some surveillance capsule holders to fail. B&W set up an integrated surveillance programme where many B&W vessel welds were irradiated in redesigned surveillance capsules installed in a few “host” vessels. The resulting data were shared by all B&W plants. BGE developed an embrittlement management option to prepare a submittal to take similar advantage of surveillance material irradiated at McGuire.

As a final action to establish a basis for future embrittlement management options, BGE redesigned unit 1 core loading patterns in the late 1980s to reduce the neutron flux and fluence to the unit 1 beltline.

MATERIAL EMBRITTLEMENT SCHEMES

In 1991 the NRC amended 10 CFR 50.61 and changed the correlation used to calculate the shift in material embrittlement (making Reg Guide 1.99 Rev 2 into law) and required utilities to perform embrittlement analyses using the newer correlation. BGE submitted a response that indicated, pending further investigation, that the unit 1 reactor vessel would exceed the screening criterion for axial weld 2-203 before expiration of the current operating licence in 2014. In fact, the correlation predicted the unit 1 axial weld would exceed the screening criteria by the spring of 1999, without severe changes in core loading. BGE established an interdisciplinary team to develop an approach to embrittlement management that would allow both units to operate through a 20 year extended licence term. The team was made up of representatives of the utility’s design engineering, systems engineering, materials engineering, operations and nuclear engineering, as well as the newly created life cycle management organisation.

The team initially reviewed the status of all embrittlement management activities and programmes. Alternatives that had received no prior consideration were also explored. It then developed and evaluated cost benefit analyses for a variety of embrittlement management scenarios. Options included: annealing the vessel; radical flux reduction through exotic fuel management changes; permanently shutting down; replacing the reactor vessel; replacing the most brittle welds; performing a Reg Guide 1.154 analysis; and using the McGuire surveillance data. Cost benefit analysis showed use of McGuire’s surveillance data to have the best combination of lowest cost and highest likelihood of technical and regulatory success.

Over time, as the vessel material becomes less tough, the maximum pressure the vessel is allowed to experience at low temperatures must be reduced. Administrative controls (Heat Up/Cool Down (HU/CD) curves) specify the maximum permissable pressure for low temperature operation, while relief valves (Low Temperature Overpressurisation Protection, LTOP) prevent inadvertant overpressurisation. The high predicted level of embrittlement led to premature expiration of curves, and the need to perform plant modifications to install variable overpressurisation protection. This modification effectively eliminated neutron embrittlement as an issue for LTOP and HU/CD for the remaining plant licence term, and for the 20 year extended licence period.

Finally, our approach to resolve the screening criteria issue became an exercise in information collection and management. To use the McGuire data, a procedural framework needed to be developed, and all the necessary technical evaluations needed to be completed. BGE identified fluence estimation as a priority activity and embarked on a programme to develop the most rigorous, detailed and defensible fluence evaluation possible. Since embrittlement depends in part on the irradiation temperature, the day by day temperature history for both Calvert Cliffs 1 and McGuire 1 were obtained. McGuire operated an average of 10°F (5.5°C) hotter than Calvert Cliffs. This would eventually require raising the embrittlement estimates for Calvert Cliffs by 10°F (5.5°C). Finally, given the fluence analyses for both units, cycle average neutron spectra were compared to demonstrate neutron environmental equivalence between the units. A submittal to the NRC demonstrating the applicability of the McGuire data to Calvert Cliffs unit 1 weld 2-203 was prepared and approved. Approval of the submittal included NRC recognition that none of the materials in either Calvert Cliffs reactor vessel beltline would approach the screening criteria before expiration of the current licence, or before the end of an additional twenty years of operation. Calvert Cliffs successfully pioneered the application of “sister” plant data to an operating vessel. In the ensuing years, the NRC would make consideration of “sister” plant data not just permissible, but required.

In addition to vigorously investigating the McGuire surveillance data, BGE embarked on a search for material fabrication records, archive material, and other applicable surveillance data. Coincidentally, the ABB-Combustion Engineering Owners Group had initiated a programme to retrieve and catalogue fabrication records. While BGE participated in that task, it was discovered that the barely used Shoreham nuclear power plant reactor vessel, which the Long Island Lighting Co was decommissioning, was fabricated with a number of the same weld wire heats used to fabricate both Calvert Cliffs vessels. BGE bought portions of the Shoreham vessel containing welds of interest. These welds were studied extensively, including determination of some previously unmeasured mechanical properties and exhaustive analyses of chemical composition. The compositional investigations included statistical evaluation of variability of chemistry along the length, through the thickness and across the width of nearly every weld contained in Calvert Cliffs vessels’ beltlines.

As the material search continued, it was discovered that Pilgrim nuclear plant had the Calvert Cliffs unit 1 controlling weld in its surveillance programme. BGE obtained some of that archive material and performed additional compositional analyses. For the unit 1 weld seam 2-203, BGE went from having no chemical analysis information in 1985, to having scores of measurements on three different weldments, including statistical evaluations of variability by 1995.

Although the embrittlement management programme at Calvert Cliffs is able to demonstrate sufficiently low levels of beltline material embrittlement to support safe operation though the end of the current licence and an additional 20 years, BGE continues to search for material and information that would more rigorously define the beltline materials’ embrittlement state. In recent years a new method of measuring embrittlement has been developed. ASTM Standard E-1921 facilitates determination of cleavage fracture toughness from testing of precracked Charpy specimens. BGE has performed this testing on all but one of the beltline welds in the two reactor vessels. While there is neither a current need, nor a current regulatory method, to use this data, the results demonstrate even greater levels of toughness than had ever been imagined.

PROJECT COMPLETION

As BGE approached licence renewal, it was recognised that embrittlement management of the vessel materials would be complicated and critical. Calvert Cliffs deliberately chose to use a combination of plant modification and material property measurement to manage embrittlement. Installation of a variable LTOP system resolved operational issues with embrittlement. Modification of fuel management practices created less embrittlement. Calvert Cliffs rigorously sought chemical and physical property information on the vessel materials, and even procured archive material and made measurements. Material property research is continuing. The results of the research have demonstratred, and will continue to demonstrate that the Calvert Cliffs vessel beltline materials are tough.

Calvert Cliffs reactor vessel fabrication

The two Calvert Cliffs plants have, nominally, 172 inch (4.37 m) diameter, 8 inch (21 cm) thick, low alloy steel reactor vessels. The vessels were fabricated from formed ASTM A-533 Grade B plates welded together using the automated submerged arc process. This process produces coalescence by heating with an arc or arcs drawn between a bare metal (filler) electrode (or electrodes), and the work. The arc and weld are shielded by a blanket of granular fusible flux (for Calvert Cliffs, Linde 1092, 0091 and 124 fluxes were used). The weld wire conformed generally to a MIL B-4 specification, but was modified with various nickel contents ranging from 0-1.0%. The wire was coated with copper plating for corrosion resistance, which added copper to the resulting weldments.




Privacy Policy
We have updated our privacy policy. In the latest update it explains what cookies are and how we use them on our site. To learn more about cookies and their benefits, please view our privacy policy. Please be aware that parts of this site will not function correctly if you disable cookies. By continuing to use this site, you consent to our use of cookies in accordance with our privacy policy unless you have disabled them.