Effect of nuclear data errors on the accuracy of measuring neutron spectra in reactor shields

The comparison of the accuracy of various methods of differential neutron spectra unfolding on the base of the activation analysis data employed for processing the experimental results on measuring the fast neutron fields in nuclear reactor shields is made. The sensitivity of the obtained spectra to uncertainties in initial data on cross sections and activation energies when using the express method, the method of statistical regulation and the modified maximum-likelihood fit method is investigated. The results of the experiments on the BR-10 reactor beam and NISUS installation are employed. A satisfactory agreement of the results obtained by various methods and a sufficiently high accuracy of real neutron spectra reconstruction are pointed out. The conclusion is drawn that the greatest contribution into the errors is made by uncertainties in determination of the activation integrals. In total these uncertainties lead to increase in the error in 1.5-2.0 times, and that for the low- energy field is 3-4 times. The error connected with the cross section variations constitutes from 1-4 to 5-8 %