[Back]

E. Jericha, D. Neudecker, S. Gundacker, H. Leeb, Th. Srdinko, Volker Wildpaner:
"Novel Techniques for Nuclear Data Evaluation";
Poster: Joint annual meeting of the Swiss and Austrian Federal Physical Society, Lausanne, CH; 2011-06-15 - 2011-06-17.

The development of novel nuclear technologies, recent advances in medicine, safety and dosimetry require the knowledge of a consistent set of nuclear reaction data and their uncertainties in an extended energy range up to 200 MeV. This represents a major challenge because experimental data beyond 20 MeV are scarce and any evaluation relies strongly on nuclear models which cannot be fully accounted for by the established evaluation methods. Recently, a well-defined technique for nuclear data evaluation based on modeling has been developed by our group at the Atominstitut. It is based on Bayesian statistics and starts from a well-defined prior which accounts for parameter uncertainties and model defects. Parameter uncertainties are obtained within the concept of maximum entropy and transformation invariance. Advances have been achieved concerning model defects and experimental systematic errors. The applicability is demonstrated by neutron-induced cross section on ⁵⁵Mn.