G-SHARE: A Guideline-Based Structured Reasoning Framework for Human-Factor Event Diagnosis

G-SHARE：一种用于人为因素事件诊断的基于指南的结构化推理框架

Abstract: Human-factor event diagnosis is essential for learning from operational events in nuclear power plants, yet its quality depends strongly on expert interpretation of narrative reports and guideline-based analysis. Current data-driven or one-shot large language model approaches often lack structured reasoning, have limited alignment with formal diagnostic guidelines, and may generate logically inconsistent conclusions.

摘要： 人为因素事件诊断对于从核电站运行事件中汲取经验至关重要，但其质量在很大程度上取决于专家对叙述性报告的解读以及基于指南的分析。当前的数据驱动或“一次性（one-shot）”大语言模型方法往往缺乏结构化推理，与正式诊断指南的对齐程度有限，且可能产生逻辑上不一致的结论。

To address this issue, this study proposes G-SHARE, a guideline-based structured reasoning framework that operationalizes the CNNP nine-step human-factor event diagnosis guideline into a multi-stage diagnostic process. The framework consists of evidence extraction, stepwise diagnostic reasoning, and post-hoc consistency repair, enabling explicit use of report evidence, intermediate rationale generation, and logical validation of diagnostic outputs.

为了解决这一问题，本研究提出了 G-SHARE，这是一个基于指南的结构化推理框架，它将中核集团（CNNP）的九步人为因素事件诊断指南转化为多阶段诊断流程。该框架由证据提取、逐步诊断推理和事后一致性修复组成，能够明确利用报告证据、生成中间推理依据，并对诊断输出进行逻辑验证。

A dataset of real human-factor event reports was constructed from Chinese nuclear industry sources, and a gold-standard subset annotated by domain experts was used for evaluation. Results show that G-SHARE substantially outperforms one-shot prompting and traditional machine learning baselines, with the strongest version achieving the best overall accuracy and macro-F1.

研究人员利用中国核工业来源的真实人为因素事件报告构建了一个数据集，并使用由领域专家标注的黄金标准子集进行了评估。结果表明，G-SHARE 在性能上显著优于“一次性提示（one-shot prompting）”和传统的机器学习基准方法，其最强版本在整体准确率和宏观 F1 分数上均达到了最优水平。

Ablation results further indicate that structured reasoning and consistency enforcement are critical to robust diagnosis, especially under weak prompting conditions. The findings demonstrate the value of transforming expert diagnostic guidelines into auditable reasoning workflows, providing a practical pathway for intelligent human-factor analysis in safety-critical industries.

消融实验结果进一步表明，结构化推理和一致性强制执行对于稳健的诊断至关重要，尤其是在提示条件较弱的情况下。这些发现证明了将专家诊断指南转化为可审计推理工作流的价值，为安全关键型行业中的智能人为因素分析提供了一条切实可行的路径。