Performance of neutron spectrum unfolding using deuterated liquid scintillator

Detector Characterization for Accurate Monte Carlo Simulation and Neutron Energy-spectrum Unfolding

“…where s i is the measured light-output, R i j is the detector response matrix, x j is the estimated unfolded neutron energy spectrum for iteration k [8]. Figure 10 shows the unfolded neutron energy results from the DT measurement.…”

“…These detectors preferentially rely on non-isotropic n-d back-scattering; this asymmetry results in distinct peaks in the observed light-output spectrum for discrete neutron energies. The result of the "more peaked" spectrum improves the condition number of the response matrix, leading to improved unfolding performance for complicated spectra [7,13,8]. This "more peaked" light output may also improve the performance of spectrum stripping, an alternative method for performing neutron spectroscopy under certain input energy spectrum conditions [11].…”

Data Augmentation for Neutron Spectrum Unfolding with Neural Networks

“…There are several radiation detection methods used to determine the energy spectrum of a neutron radiation source, including Bonner sphere detectors [3,4] and scintillation detectors [5][6][7]. Almost all neutron detectors, however, involve the use of materials with characteristic responses to radiation which preserve information about the energy of the incoming radiation.…”

“…Many commonly used unfolding algorithms employ iterative methods [4,7,12] to converge on the unfolded spectrum, given some a priori guess of the solution spectrum. The constraints are usually baked into iterative unfolding algorithms-often through initializing the iteration from a realistic-looking neutron energy spectrum or by only considering neutron energy spectra which follow from theoretical models [12].…”