Material chemistry

Most BWRs in the U.S. were fabricated with plate from Lukens Steel. Fabrication records sent to GE by Lukens generally did not include copper content, but Lukens has kept copper information in internal records. As part of the recent BWR Owners’ Group RG 1.99 Impact study [6.9], Lukens provided GE with copper content information on most BWR beltline plates.

Weld fabrication records, when available usually have copper and nickel data.

However, in some plants, especially older plants, weld chemistry is unknown. When chemistry data are unavailable, an upper bound is mandated by Regulatory Guide 1.99 Revision 2 [6.4]

which specifies an upper bound chemistry of 0.35 Cu (% wt) and 1.0 Ni, which may be much

higher than the unknown chemistry. In 1998 the BWR Vessel and Internals Project (BWRVIP) surveyed all US BWR reactor vessel plate and weld chemistry information.

The purpose of this effort was to resolve missing data and establish best estimate chemistry values for the limiting plate and welds materials for each US BWR. This chemistry data and other related material property data was then used to judge the adequacy of the materials in each plant surveillance programme. For each vessel limiting weld and limiting plate, a best representative surveillance material was assigned, based on heat number, similar chemistries, common fabricator, and the availability of unirradiated data. The results of this work are reported in BWRVIP-86-A [6.10].

The following provides the procedures adopted by the BWRVIP for the determination and use of best chemistry information.

6.1.2.1 Determination and use of best estimate chemistry

Analysis of the embrittlement behavior of surveillance materials requires knowledge of the material’s chemistry content, so that measured embrittlement shifts can be compared to the shifts that would be expected based on chemistry. Historically, there has been little guidance regarding the proper techniques for characterizing a material’s chemistry when various chemistry measurements are available for a heat of material. To ensure consistency of analysis, the BWRVIP ISP has adopted the protocol provided below for estimating a surveillance material’s best estimate chemistry.

6.1.2.2 Vessel weld vs. surveillance weld material chemistries

It is important to differentiate between the best estimate chemistry for a vessel weldment and the best estimate chemistry of surveillance weld material. For vessel welds, the U.S. NRC has established that the best estimate chemistry is obtained from the average of all industry-wide data for the specific heat [6.11]. Because of the wide variations in chemistry that are often present in weldments, the average of all industry-wide data is regarded as the best representative for a vessel weld.

For surveillance weld chemistry, the U.S. NRC has held that the best estimate should be based on the chemistry data for that specific weld rather than the heat best estimate chemistry [6.11].

Chemistry measurements

As a standard practice, chemistry measurements should be made on one or more surveillance specimens of each surveillance material (e.g. plate and weld) whenever a surveillance capsule is tested. The new chemistry data should be documented in the report, and the data should be averaged with any previous data to obtain a refined best estimate chemistry.

The new, revised best estimate chemistry should then be used in the analytical evaluation of the surveillance material embrittlement behavior.

Standard practice for calculating best estimate chemistry

In general, there are two types of chemistry data available: material certification data, to qualify a material for use in the reactor vessel; and specific chemistry measurements made on surveillance specimens fabricated, in many cases, years after the original material certifications are performed. For surveillance plates, both types of data are useful for estimating the

chemistry. For surveillance welds, only the specific measurements on the surveillance specimens are preferred as the basis for calculating best estimate chemistry.

In the case of surveillance plates, the surveillance materials are taken from the actual vessel material and the certification chemistry measurements generally reflect the chemistry of the surveillance plate specimens. Therefore, any certification chemistry data for the plate heat should be averaged with chemistry measurements that have been made on specific surveillance plate specimens in order to determine the best estimate chemistry for the surveillance plate material heat.

In the case of welds, most surveillance welds were made well after the vessel weld certification test was performed. Weld certification data preceded vessel fabrication, so in many cases there is a period of years between weld certification chemistry measurements and fabrication of the surveillance weld. As a result, the surveillance weld was typically made from a different coil of weld metal wire, even though it bears the same heat number as the qualification weld. Due to the variability of chemistry in different coils of weld wire, chemistry reported for the certification weld can differ significantly from the chemistry of the surveillance weld. Therefore, weld certification data should not be used in the calculation of the surveillance weld best estimate chemistries (unless no other data is available). In the past, weld certification chemistry data was sometimes reported as the estimate because no measurements existed specifically for the surveillance weld specimens. For all ISP surveillance welds, however, specific chemistry measurement data for surveillance specimens now exist.

The guidelines for handling the combining of chemistry data were provided by the Nuclear Regulatory Commission (NRC) at a workshop in 1998 [6.11]. When there are multiple specimens with individual chemistry measurements, each measurement is added into the average for that material heat. In the cases where multiple measurements exist for one specimen, an average is first determined for that specific specimen, and then that specimen’s average value is averaged with the other specimens’ measurements to yield the overall average for the material.

In the example below, the average chemistry for Specimen JJ1 (two data sets) is determined first because it is a single specimen having multiple data points. This data point is then averaged in with the other two chemistry sets to determine the overall average for the material heat.

Source Specimen

ID

Cu Ni P S Si

Baseline CMTR 0.5 0.8 0.001 0.002 0.2

Capsule X JJ1 0.4 0.6 0.001 0.2

JJ1 0.5 0.7 0.002 0.1

Avg. JJ1 0.45 0.65 0.001

5 0.1

5

J25 0.4 0.7 0.003 0.003 0.1

Best Estimate Average 0.5 0.7 0.002 0.003 0.2

It should be noted that there might be multiple sources for surveillance chemistry data.

Most information comes from capsule test reports or surveillance programme documentation.

In some cases, supplemental chemistry and Charpy testing has been performed by other organizations (e.g. national laboratories such as Oak Ridge), and that data should also be considered if its provenance is adequate. Care should be taken, however, when the inclusion of such data is considered; testing conducted at a laboratory for research purposes may not have been conducted with intent (and necessary quality assurance) for results to be used in a nuclear surveillance programme.