现代应用物理2025，Vol.16Issue(6)：54-71,18.DOI:10.12061/j.issn.2095-6223.202506013
基于深度学习的共振计算方法研究及其在堆芯分析中的应用
Research on Deep Learning-Based Resonance Calculation Method and Its Application in Core Analysis
摘要
Abstract
To address the challenge of calculating effective self-shielding cross-sections(XSs)for fuels with temperature distribution,a deep learning-based approach is integrated into the resonance calculation module of the NECP-X code.Six deep learning models are developed,optimized,and trained to handle temperature-dependent effects in both rod-shaped and annular fuel configurations.The XSs calculation accuracy of the proposed method is validated through benchmark problems,involving fuel cells.Further applications to annular fuel assemblies and a full-core small research reactor demonstrate its capability.Results show that the deep learning-based resonance method effectively resolves temperature-dependent self-shielding effects,with good agreement in eigenvalues and power distributions compared to reference solutions.关键词
NECP-X/深度学习/共振计算/棒状燃料/环形燃料/全堆芯分析Key words
NECP-X/deep learning/resonance calculation/rod-shaped fuel/annular fuel/full-core analysis分类
能源科技引用本文复制引用
曹璐,李达,王立鹏,胡田亮,姜夺玉,李华琪,陈立新,刘宙宇,曹良志..基于深度学习的共振计算方法研究及其在堆芯分析中的应用[J].现代应用物理,2025,16(6):54-71,18.基金项目
国家自然科学基金资助项目(12275219,12205237) （12275219,12205237）
