Neutron spectrum unfolding based on neutron activation reactions is an inherently ill-posed inverse problem that is difficult to validate experimentally. This study extends the performance evaluation of the SpecKit software previously published in SoftwareX, with a focus on quantifying the effect of energy-group structure on unfolding accuracy. The ground-truth neutron spectrum was generated using MCNP simulations. The prior spectrum was modeled under the same geometric configuration; however, compared with the ground-truth spectrum, the cement layer in the 60-, 113- and 240-group cases was uniformly reduced by 3 cm to simulate scattering interference and geometric simplification commonly encountered in real-world applications. Comparative unfolding with SpecKit using 60-, 113- and 240-group structures shows that the mean activity error (ERRE) of calculated activities is significantly reduced from 0.165 to 0.050, while the integrated flux also approaches the ground-truth value more closely (ground-truth: 1.606 × 10 6 n/cm 2 /s; 60-group: 1.065 × 10 6 ; 240-group: 1.578 × 10 6 ). These empirical results demonstrate that, within the SpecKit framework, adopting a finer energy-group structure effectively mitigates the ill-posed nature of the problem, enhances numerical stability and reconstruction fidelity, and provides a reliable, reproducible software platform for nuclear engineering education and practical neutron field analysis. The open-source SpecKit toolkit is freely available at GitHub.
No takes yet. Share an insight, caveat, or question.
Chen et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: