Derivation of ν_d data from β_eff measurements: interest and limits

The ν_d data for the three important isotopes, ²³⁵U, ²³⁸U, and ²³⁹Pu, included in the JEF2 library have been calculated, as a function of energy, by using the so-called LENDEL's model. In the framework of a general benchmarking of the JEF2 file, these evaluated data have been used to calculate 22 effective delayed neutron fraction (β_eff) measurements in the thermal and fast energy ranges. In order to make the calculated β_eff value dependent on the ν_d values alone, the ERALIB1 set of cross section data adjusted for the main nuclear parameters but the ν_d has been used in the neutronics calculations for these 22 benchmarks. A reanalysis of the experimental conditions has been performed in order to produce a realistic estimate of the covariance matrix on the one hand and consistent values for the calculated part Pc of the 'experimental' value of β_eff on the other hand. It appears that all β_eff data but one (experiment SNEAK 9C2) are reproduced by the calculation using the JEF2 ν_d data with a discrepancy lower than 3%, which is the performance required by Reactor Physicists. A statistically consistent adjustment procedure has been used to quantify the small corrections to apply to the evaluated ν_d data. Recommended ν_d values have been derived for ²³⁵U,²³⁸U and ²³⁹Pu. They are affected by an uncertainty lower than 3% except for ²³⁸U (∼4%). This result raises the question of the capability of current calculational systems to calculate accurately the ²³⁸U fission rates. Further measurements and basic investigations are highly encouraged to improve the situation for the higher Pu isotopes and the minor Actinides. (author)