whaeck
commented
6 years ago Multiple solutions are possible for this issue. By default (since NJOY2012), reconr will add the summation cross section when only individual levels are present. As such we can assume that the summation cross section will always be present on the PENDF tape.
The logic present in GASPR can thus be significantly simplified by explicitly skipping the individual levels by themselves. There will no longer be any need to set any flags indicating which section to use.
kahlerac
paulromano
brown170
An issue was uncovered in gaspr concerning the charged particle production cros ssections when the original ENDF tape does not contain the summation cross section when individual levels are present.
Cfr the following email from @kahlerac on the issue: All, Late yesterday Valentin came to me to say he'd spoken incorrectly about the 10B(n,a) issue. Rather in 10B it was tritium, not alpha that was two times to large. My initial look at the output I created seemed to (once again) contradict this assertion as the tritium production in mt205 and the tritium production in my ACE file were in agreement. However, when I looked at the (admittedly tiny) NJOY/GASPR output where a table of original mt's that contribute to the respective mt203 through mt207 is given I was surprised to see it indicated both mt105 and mt700 were processed to obtain GASPR's mt205. This should not happen since mt105 is redundant to the sum of mt700 through mt749. So why did the code logic apparently fail in this instance? It turns out that in the original 10B endf file we have data in (i) mt103 and mt600 to mt605, (ii) mt107 and mt800 & mt801, and (iii) mt700 but no mt105. GASPR properly recognizes the mt103 and mt107 redundancy but with no mt105 section there is no redundancy to discover. You might then ask "how can there be a problem?". Well, in reconr processing we know that mt103 through mt107 are redundant if mt 600 and above are present and, beginning with NJOY2012, to guarantee consistency the redundant mt103 through mt107 are recalculated before being written to the pendf tape. Therefore in the "pendf world" we end up with an mt105 section because of something being found in the mt700 to mt749 range. Hence when we get to GASPR we have an original endf tape with mt700 and without mt105 and a pendf tape with both mt105 and mt700. It is the original endf tape's dictionary that is used to determine what mt's GASPR should omit but it is the pendf tape that is read and processed to create mt203 through mt207. Since mt105 doesn't exist on the original endf tape there is no "omit" flag set for the mt700 to mt749 interval and in the pendf world we process both mt105 and mt700 to obtain mt205 ... which is now twice what it should be. For 235U it is an (n,a) issue as there we have a suite of mt800's and no mt107 on the original endf tape. I seem to recall from endf/b-vi.0 days that there were only a handful of technically unimportant evaluations that omitted one or more of mt103 to mt107 while providing mt600 and above; hopefully that remains true in endf/b-viii.0 because the sad reality is that for everyone of those evaluations where this is true we have incorrect gas production data, :(. I wonder whether TENDL consistently provides (or omits) mt103 through mt107 when providing mt600 and above? I expect that the fix will be to simply review the pendf dictionary rather than the endf dictionary but I'm sure Jeremy and Wim will want to study (and test) this more completely. Skip