A Bayesian-Neural-Network Prediction for Fragment Production in Proton Induced Spallation Reaction

TL;DR

This paper introduces a Bayesian neural network approach to improve predictions of fragment cross sections in proton-induced spallation reactions, enhancing accuracy across various nuclear systems and energies.

Contribution

The paper develops a BNN model that learns residuals from empirical parameterizations to accurately predict fragment production in proton spallation reactions.

Findings

BNN predictions align well with experimental data

Method improves upon existing empirical models

Applicable across a range of nuclei and energies

Abstract

Fragments productions in spallation reactions are key infrastructure data for various applications. Based on the empirical parameterizations {\sc spacs}, a Bayesian-neural-network (BNN) approach is established to predict the fragment cross sections in the proton induced spallation reactions. A systematic investigation have been performed for the measured proton induced spallation reactions of systems ranging from the intermediate to the heavy nuclei and the incident energy ranging from 168 MeV/u to 1500 MeV/u. By learning the residuals between the experimental measurements and the {\sc spacs} predictions, the BNN predicted results are in good agreement with the measured results. The established method is suggested to benefit the related researches in the nuclear astrophysics, nuclear radioactive beam source, accelerator driven systems, and proton therapy, etc.