High-energy neutron shielding analysis using curve fitting and machine learning: a Monte Carlo-based study in concrete and Ferro-Boron configurations

Creative Commons License

Gökcan A. O., Sarıyer D.

ANNALS OF NUCLEAR ENERGY, cilt.232, sa.2026, ss.1-15, 2026 (SCI-Expanded, Scopus)

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 232 Sayı: 2026
  • Basım Tarihi: 2026
  • Doi Numarası: 10.1016/j.anucene.2026.112244
  • Dergi Adı: ANNALS OF NUCLEAR ENERGY
  • Derginin Tarandığı İndeksler: Scopus, Science Citation Index Expanded (SCI-EXPANDED), Chemical Abstracts Core, Compendex, Environment Index, INSPEC
  • Sayfa Sayıları: ss.1-15
  • Açık Arşiv Koleksiyonu: AVESİS Açık Erişim Koleksiyonu
  • Manisa Celal Bayar Üniversitesi Adresli: Evet

Özet

This study aims to develop an effective shielding approach to control secondary neutron radiation generated

under abnormal conditions in high-energy proton accelerators. Neutrons, being neutral, deeply penetrate matter

and pose risks to radiation protection and biological safety. Neutron dose distributions were obtained using

FLUKA Monte Carlo simulations for four shield–environment setups (standard concrete–air, standard concrete–

shield, ferro-boron–air, ferro-boron–shield) and four proton energies (50, 100, 250, 1000 MeV). The data

were analyzed using a classical exponential attenuation model via curve fitting and then used to train regression

models based on Linear Regression, Decision Tree, and Random Forest algorithms. Model performance was

evaluated with R2, RMSE, MAE, and 5-fold cross-validation. Random Forest achieved the highest accuracy,

especially in shielded setups (R2 ≈ 0.92–0.97). These results show that combining machine learning with

physics-based modeling reduces computation time while maintaining high accuracy, providing a practical

framework for radiation shielding design