Estimations for (n,α) reaction cross sections at around 14.5MeV using Levenberg-Marquardt algorithm-based artificial...

Area of Science:

• Nuclear Physics
• Computational Physics

Background:

• Accurate prediction of neutron-induced reaction cross-sections is vital for fusion reactor technology, impacting nuclear transmutation, heating, and radiation damage assessments.
• Existing theoretical models require validation and faster computational methods for complex nuclear data calculations.

Purpose of the Study:

• To introduce and evaluate a novel artificial neural network (ANN) approach for predicting (n,α) reaction cross-sections at approximately 14.5 MeV neutron energy.
• To assess the efficiency and accuracy of the Levenberg-Marquardt algorithm-based ANN for this predictive task.

Main Methods:

• Utilized an artificial neural network (ANN) model, specifically employing the Levenberg-Marquardt algorithm for training.
• Trained and validated the ANN model using experimental data, achieving high correlation coefficients (R-values up to 0.99283).
• Compared ANN predictions against results from the TALYS 1.95 nuclear code for validation.

Main Results:

• The Levenberg-Marquardt algorithm-based ANN demonstrated high predictive accuracy, with an overall R-value of 0.98515.
• The ANN model proved to be computationally efficient, significantly faster than some advanced algorithms.
• ANN predictions showed good agreement with theoretical calculations from TALYS 1.95.

Conclusions:

• The developed ANN model is well-suited for the systemic study and prediction of (n,α) reaction cross-sections.
• This approach offers a fast and accurate alternative for nuclear data calculations in fusion reactor applications.
• The ANN methodology provides a robust tool for understanding neutron-induced reaction mechanisms.