For the first time, the basis behind each of the cross section sets in SCALE is documented in one place. This is valuable for a user to know how the cross section set was collapsed, what type of assumptions were made, where inherent weaknesses are likely to be, etc.
A clear, concise description of what BONAMI (unresolved resonance region) and NITAWL are used for (resolved resonance region) is provided. The limitations of each method, and the fact that the data required for these modules is not always present for a given isotope in a given cross section set.
A good report, but the arbitrary assignment of excellent to very poor for selected cross section sets in the conclusions is not good. This assignment appears to suggest that overall the 44 group is better than the 238 group. However, I noticed that the error bars for the 238 group results are significantly larger than that for the 44 group. Could this be the difference between an assignment of excellent and very good? I would definitely recommend the use of the 44 group or for that matter the 238 group simply on the merits that this data represents the current state-of-the-art in cross section development.
To be (good) or not to be (good), that is the question? What makes something good? What is the difference between accuracy and precision? Both are good - one is just biased. In order to get good results one must have a good experiment. But the level of good varies in good experiments such that there is always variability in good results. Keep up the good work!!!
Despite a nice first step, I don't think the NRC got much of what they wanted or knew much of what they thought they were getting. Some nice work was done in this report but it tries to push conclusions that aren't easily drawn or can correctly be used. This cross-section set is good and this other is bad? Did the NRC think they were getting a report that would tell them what cross-sections sets individuals and facilities should be using? And was this going to be part of their basis for determining that an adequate validation exists?
Despite, the best efforts of left-brained individuals within this community to simplify validation into some simple set of numbers or a black and white picture of which cross-section sets to use, the validity of most analyses are not an easy thing to define, assess, or determine in a cookbook fashion. Validation is an art still requiring thinking and the use of the right-side of the brain (along with the left). The validity of a code calculation can not be assured by the rating of a cross-section set or a "magical" keff numerical value alone (or even a complex one that applies a least-squares fit). It is unfortunate that our regulators and the industry continue to look blindly for an easy answer to a complex subject.
| Cross-Section Library | |||||
| Problem Type | SCALE 16-group Hansen-Roach | 27-groupa ENDF/B-IV | 218-groupa ENDF/B-IV | 44-group ENDF/B-V | 238-group ENDF/B-V |
| Fast HEU | Very good 0.995 +/- 1% | Very good 1.00 +/- 1% | Very good 1.00 +/- 1% | Very good 1.00 +/- 1% | Very good 0.995 +/- 1% |
| Thermal HEU | Very poor 2% high to 4% low | Poor 3% high to 2% low | Poor 1.00 +/- 2% | Poor 1.00 +/- 3% | Poor 1.00 +/- 2% |
| LWR Lattices | Fair +/-1.5% | Good 0.995 +/- 1% small bias vs energy | Good 0.99 +/- 1% small bias vs energy | Very good 1.00 +/- 1% | Very good 0.995 +/- 1% |
| Homogeneous LEU | Poor +/-2% bias vs energy | Fair 1.00 +/- 1.5% | Fair 0.995 +/- 1.5% | Fair 1.01 +/- 1.5% | Fair 1.005 +/- 1.5% |
| 239Pu | Very poor up to 5% high bias vs energy | Poor up to 5% high bias vs energy | Poor up to 5% high bias vs energy | Fair 1.02 +/- 2% | Fair 1.015 +/- 2% |
| MOX Lattices | Very poor 3% high to 1% low | Fair 1.00 +/- 1.5% bias vs energy | Fair 1% high to 2% low bias vs energy | Excellent 1.003 +/- 0.6% | Very good 1.00 +/- 1% small bias vs energy |
| 233U | Poor thermal - 0.99 +/- 2% fast - 1.0 +/- 1% | Poor thermal - 1.02 +/- 1.5% fast - 0.97 +/- 1% | Poor thermal - 1.015 +/- 1.5% fast - 0.97 +/- 1% | Good thermal - 1.00 +/- 1.5% fast - 0.997 +/- 0.5% | Good thermal - 0.995 +/- 1.5% fast - 0.999 +/- 0.4% |
aUse of hafnium and gadolinium data not recommended. | |||||
| Source: Experience with the SCALE Criticality Safety Cross-Section Libraries, NUREG/CR-6686, ORNL/TM-1999/322, October 2000 | |||||