Machine-Learning-Assisted Evaluation of the Am(n,f) Cross Section in the 0.1–20 MeV Neutron Energy Range
Journal of Fusion Energy, cilt.45, sa.2, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 45 Sayı: 2
- Basım Tarihi: 2026
- Doi Numarası: 10.1007/s10894-026-00601-y
- Dergi Adı: Journal of Fusion Energy
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Aerospace Database, Chemical Abstracts Core, Compendex, INSPEC, Engineering Source (EBSCO), Materials Science & Engineering Collection (ProQuest), Technology Collection (ProQuest)
- Anahtar Kelimeler: Americium-241, Fast neutron energy range, Machine learning regression, Neutron-induced fission, Nuclear data evaluation, TALYS nuclear reaction code
- Sivas Cumhuriyet Üniversitesi Adresli: Evet
Özet
Reliable cross sections are essential for reactor design, transmutation performance, and nuclear data evaluation, yet significant discrepancies persist among experimental datasets, evaluated libraries, and theoretical models. This work provides a systematic assessment of the fission cross section in the fast-energy range by integrating EXFOR measurements, TALYS-2.0 calculations using five fission barrier formulations, evaluated libraries (ENDF/B-VIII.0 and TENDL-2019), and machine-learning-based predictions. A physically informed feature set including separation energies, deformation parameters, and fission barrier characteristics was used to train six regression models: CatBoost, Gradient Boosting Machine, Random Forest, Support Vector Regression, k-Nearest Neighbours, and XGBoost. GBM delivered the highest predictive accuracy and generalization, closely reproducing the experimental excitation function up to 20 MeV. TALYS results showed strong sensitivity to the barrier model. Comparative analysis demonstrates that advanced machine learning approaches can reduce high energy uncertainties and complement reaction model calculations. The findings highlight the value of data driven methodologies for improving nuclear data related to long-lived minor actinides and for supporting next generation fuel cycle and transmutation applications.