Reconstruction of neutron spectrum emitted by portable 14 MeV neutron generator by a combination of different methods

The presented paper is a continuation of the Tikhonov Regularisation, Minimum Fisher Information, Maximum Entropy, and Maximum Likelihood methods application to unfolding the neutron spectra emitted by the fusion devices from activation measurement. The aim of the analysis is the evaluation of the possible uncertainty improvement. The developed algorithms and methodology were validated with 14 MeV portable neutron generators before application to the analysis of the neutron spectrum emitted from tokamak. The analysis concerns plasma-target generator did not provide satisfactory uncertainty of the results. The modification of the irradiation foils' set and the improvement in the activity measurement precision progress the results. The sealed tube portable neutron generator irradiated the Al, Fe, Ni, Zr, Au, Mg, and Nb foils. The HPGe spectrometer measured the induced activity with an uncertainty of about 10 %. The synthetic data analysis was made to establish the ratio between the error of the spectrum reconstruction and the discrepancy between the input data and activities reconstructed for this spectrum by the FISPACT-II inventory code. Two bin structures were considered. One was compatible with the FISPACT-II code, but the energy resolution was not regular. The second was characterised by constant bin spacing connected with the modification requirement before simulation and additional uncertainty of about 5 %. The best results were obtained by calculation of the mean value from all methods. The corresponding reconstruction uncertainty is equal to 11 %. The reconstructed spectrum has a dominant deuterium tritium peak slightly shifted to the higher energies due to the beam-target reaction kinematics. The evaluation of the impact of the deuterium-deuterium reaction due to the long-term operation was not reachable due to the measured high intensity from scattered neutrons.