基于F-BN的微型顶管凿岩作业人员疲劳风险预警

F-BN-based fatigue risk early warning method for workers in micro pipe-jacking rock drilling operations

摘要

针对微型顶管凿岩作业人员疲劳风险的多维度、动态性和不确定性特征，提出一种基于模糊贝叶斯网络(Fuzzy-Bayesian Network, F-BN)的预警方法，实现疲劳风险的精准评估与实时预警。基于系统工程理论，从人因、设备、环境、组织管理四大维度系统识别出24个关键风险因素，构建具有四层拓扑结构的贝叶斯网络模型。采用改进的模糊多属性决策法(FMADM)处理专家评估数据的模糊性与不一致性，通过三角模糊数转化与解模糊化处理建立先验概率分布，结合贝叶斯网络逆向推理与敏感性分析揭示风险传导机制。研究表明，“工作组织与安排不合理”“温度、湿度或气压不适”等为后验概率高的关键致因因素，与疲劳风险紧密相关，占据核心地位；过度紧张→心理异常、负荷超限→事故发生等为高作用强度的主要传导链路，建立的F-BN模型为地下工程施工职业安全健康管理提供了兼具理论深度与工程适用性的智能决策工具。

Abstract

In view of the multi-dimensional, dynamic and uncertain characteristics of the fatigue risk of micro-tunneling rock-drilling workers, an early-warning method was proposed based on the Fuzzy-Bayesian Network(F-BN) to achieve accurate assessment and real-time early-warning of fatigue risk. By comprehensively considering factors from four aspects: human, physical, environmental, and organizational management, 24 main risk factors are determined. The fuzzy multi-attribute decision-making method is used to integrate expert evaluation data, and in-depth analysis is carried out in combination with Bayesian network reasoning. The research shows that factors such as “unreasonable work organization and arrangement” and “inappropriate temperature, humidity or air pressure” are key causal factors with high posterior probabilities. They are closely related to fatigue risk and occupy a core position. The main transmission links with high intensity include excessive nervousness→psychological abnormality, overload→external manifestation, etc. The proposed method provides a theoretical basis for the occupational safety and health management of pipe-jacking construction.

关键词

贝叶斯网络 / 顶管施工 / 疲劳风险 / 风险预警 / 模糊多属性决策法

Key words

Bayesian network / pipe jacking / fatigue risk / risk early warning / fuzzy multi-criteria decision-making

Author summay

毛三军(1970—), 男, 正高级工程师, 硕士, 主要研究方向为水利水电工程智能化与信息技术研究。E-mail: [e-mail]

基金资助

国家自然科学基金面上项目(52179136)

中国长江三峡集团有限公司企业科研项目(202103551)