An Alternative Approach to Creating ACE Data Files for Use in Monte Carlo Codes

There is a strong desire in Member States to have access to 'open source' data processing system that would be well maintained and would avoid the danger of 'common mode failure' due to all data processing done with the same set of processing tool. At present, the only tool for generating libraries in ACE format for Monte Carlo transport codes is the NJOY data processing system. Self-shielding in the unresolved resonance range (URR) is treated by the probability table method (PTM), generated from statistically distributed ladders of resonances. If other codes could generate equivalent parameters, this would open the door to an alternative route for generating the ACE libraries. The methodology in PREPRO codes can produce multi-band parameters (MB) from self-shielded cross sections. The multi-band parameters have been used in the TART Monte Carlo code and in deterministic codes, for many years, but they were not applied broadly in other codes. The multi-band parameters were shown to be equivalent to the probability tables like those generated by the PURR module of NJOY. Generally two bands are here shown to be sufficient, in contrast to the probability tables in PURR, which typically use about 20 bins to represent the probabilities. The purpose of the present work is to show that the replacement of the probability tables in ACE files with two band parameters (based on the same self-shielded cross sections) reproduces reference results for the criticality of reactor assemblies with sufficient precision. Positive results of the exercise justify the development of alternative modules for generating multi-band parameters in the unresolved resonance range, and possibly extending the multi-band approach to lower energies in order to reduce the volume of the ACE libraries without sacrificing the accuracy of the calculations. The overall conclusion is that two-band parameters can be substituted for the probability tables in the ACE libraries for Monte Carlo transport calculations. Out of the 32 benchmark criticality cases sensitive to epithermal capture and fission in 2^35U and 2^38U the maximum difference in the results was about 20 pcm, which is much smaller in comparison to the actual uncertainties in real systems.