What question did this study set out to answer?

The study develops a method to automatically convert B-Rep models into CSG geometry for fusion reactors, enhancing efficiency and stability.

March 26, 2026Nuclear Techniques0 citationsOpen Access

Graph Neural Network Method for Reusing CSG Models in Monte Carlo Calculations

A graph neural network-based method for reusing Monte Carlo CSG models and its application

View Full Paper

Ask AI

Bookmark

Authors

An intelligent method improves CSG modeling efficiency in fusion engineering while ensuring physical accuracy.

The study develops a method to automatically convert B-Rep models into CSG geometry for fusion reactors, enhancing efficiency and stability.
Proposed a graph-based geometry representation to convert B-Rep into an attributed adjacency graph (AAG).
Trained a Graph Attention Network (GAT) for efficient classification and retrieval of CSG cell expressions.
Introduced a preprocessing pipeline for consistency in similar models.
Automated integration of retrieved CSG expressions into Monte Carlo input files.
Achieved an average retrieval time of 0.405 ms for CSG construction.
Ensured structural consistency with original CAD models.
Observed a maximum deviation of less than 0.5% in Tritium Breeding Ratio (TBR) compared to traditional methods.

Shen et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark

March 26, 2026Nuclear Techniques0 citationsOpen Access

A graph neural network-based method for reusing Monte Carlo CSG models and its application

View Full Paper

Ask AI

Bookmark

Authors

An intelligent method improves CSG modeling efficiency in fusion engineering while ensuring physical accuracy.

The study develops a method to automatically convert B-Rep models into CSG geometry for fusion reactors, enhancing efficiency and stability.
Proposed a graph-based geometry representation to convert B-Rep into an attributed adjacency graph (AAG).
Trained a Graph Attention Network (GAT) for efficient classification and retrieval of CSG cell expressions.
Introduced a preprocessing pipeline for consistency in similar models.
Automated integration of retrieved CSG expressions into Monte Carlo input files.
Achieved an average retrieval time of 0.405 ms for CSG construction.
Ensured structural consistency with original CAD models.
Observed a maximum deviation of less than 0.5% in Tritium Breeding Ratio (TBR) compared to traditional methods.

Shen et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark