ANS-19.10 AFast Neutron Fluence in PWR Reactor Cavities@

Marriott Wardman Park Hotel

Room 8209

Washington, DC

November 15, 1998

The fifteenth meeting of ANS-19.10 was convened at 9:00 a.m. on Sunday, November 15, 1998. Nine committee members and Shawn Coyne-Nalbach (ANS) were in attendance. The names, affiliations, e-mail addresses, and fax and phone numbers of the attendees are given in Attachment-1.

Chairman L. Lois welcomed the Committee members and expressed appreciation for the continued interest and contributions to the Standard. L. Lois summarized the progress on the Standard. He stated that we have contributions on the Transport, Dosimetry and Monte Carlo Sections, but still require a clear discussion for the Synthesis Section.

L. Lois stated that M. Weber has retired and that Shawn Coyne-Nalbach is attending the meeting as a representative of the ANS. He stated further that Dick Rubin has developed a web site for ANS-19.10 at http://home.att.net/~rmrubin/, and that when you arrive at this address you will see AClick here for the ANS-19.10 home page.@ The latest version of the ANS-19.10 standard is available at this site. He asked the committee members to use the site to access the Standard and to send any comments to Dick Rubin or the Chairman. He thanked Dick for his efforts in developing the ANS-19.10 site.

S. Anderson proposed a motion to approve the minutes of the previous meeting. R. Cacciapouti

seconded the motion, and the Committee approved the minutes unanimously.

L. Lois asked for comments on the Standard. A. Haghighat stated that he sent his contribution on the Monte Carlo section to A.Ford. L.Lois stated that R. Little will also be providing comments on the Monte Carlo section.

L. Lois asked R. Cacciapouti for comments on the dosimetry section. R. Cacciapouti stated that the latest version of the dosimetry section is on the web site and he considers this his final version. J. Adams indicated that he is recommending a few additional minor changes to the dosimetry section.

Q. King stated that J. Worsham has provided a new contribution on the process of unfolding measured fluences for the dosimetry section (Attachment-2).

S. Anderson stated that he believes additional improvements will be required before the synthesis section will be ready for release. Specifically he noted that the standard, as presently formulated, provides a methodology for performing calculations and measurements but does not provide a methodology for reducing the fluence uncertainty. He stated further that the measurements are being used to Acheck@ the calculations rather than to reduce the uncertainty in the calculations.

S. Anderson also noted that the standard allows the analyst a choice of methods (discrete ordinates v.s. Monte Carlo, BUGLE-93 v.s. BUGLE-96, measured v.s. calculated core power distributions, etc.), but does not provide a recommendation on which prediction to use when they are different. He stated further that the two predictions could easily disagree by 3%.

A. Haghighat stated that there are methods issues that result in differences of ~ 10% and that he finds the discrete ordinates predictions are in better agreement with measurement than the Monte Carlo calculations. He stated further that he believes the multigroup cross section libraries can be improved.

I. Remec stated that he expects the different methods to disagree by as much as 10 %.

L. Lois stated that he believes two benchmark codes should give very good agreement.

J. Carew stated that, in order to be applicable to a reasonably broad range of problems, the standard must not be overly prescriptive and must allow the analyst freedom to apply engineering judgement to his problem. As a result, the calculational method is not completely specified and the analyst is given a choice of approaches to solve the problem.

I. Remec stated that he believes it is acceptable for the standard to allow the use of engineering judgement and the selection of alternate methods.

A. Haghighat stated that the possibility of multiple calculational methods and solutions can affect the response cross sections. For example, if the dosimetry cross sections are adjusted to provide agreement between the calculations and measurement, the adjustment of the dosimetry cross section will depend on the specific type of calculation (e.g., discrete ordinates v.s. Monte Carlo). Consequently, a response cross section adjusted using a Monte Carlo calculation is only valid when used in a Monte Carlo analysis.

R. Little stated that for a given nuclear data set ( e.g., ENDF/B-VI), a continuous energy cross section library generally involves fewer approximations and is considered more accurate.

S. Anderson stated that if comparisons with measurement are going to be made acceptance criteria are required. For example; (1) an acceptance range within which measurements must fall (e.g., two standard deviations) to be considered acceptable and/or (2) an acceptance range for M/C comparisons.

J. Carew noted that the statement in the standard that the calculation and measurements should agree to within 20% ( 1-sigma) provides a criteria on the M/Cs.

J. Adams stated that there are two issues concerning uncertainty: (1) improving the precision of the calculation and (2) improving the accuracy of the calculation relative to Atruth.@ The measurements provide a truly independent estimate and provide the basis for improving the accuracy of the prediction.

A. Haghighat suggested that the calculation and measurement be considered equally valid estimates of Atruth.@

L. Lois stated that it is not clear what approach should be taken if the iron and nickel measurements disagree by 35%.

I. Remec stated that when the measurement and calculation disagree the most conservative value should be used.

J. Adams stated that, while you don=t want to be overly prescriptive, the M/C data should be used to improve the best-estimate fluence prediction.

Q. king made the following observations: (1) the interpretation of the measurements is difficult because the measurements are typically not at the location of interest - the vessel inner-wall (2) additional assurance can be obtained by the comparison of the benchmark problem calculation with the reference solution (3) it is unlikely that two calculations that follow the guidance of the standard will differ by 10% (4) if the measurements are reliable, they provide an indication of the Atrue@ fluence and an estimate of the calculational accuracy (5) having measurement data on both sides of the vessel can improve your benchmarking significantly.

L. Lois stated that: (1) if both discrete ordinates and Monte Carlo calculations are performed and the predictions disagree, the differences must be explained (2) when there is a large number of accurate and reliable measurements, the measurements can be used to adjust the calculations.

Q. King stated that we take a similar approach; we compare the M/Cs, determine the calculation and measurement uncertainty, and evaluate the acceptability of the calculation.

S. Anderson stated that a plant-specific benchmarking can be performed, if you have a statistically significant number of measurements for the plant.

L. Lois asked A. Haghighat to include these comments in the synthesis section.

Chairman Lois thanked all present for their continued efforts and adjourned the meeting at approximately 12:00 noon.

November 15, 1998