摘要
针对远程控制船舶内河航行风险定量评价问题,提出了一种不确定性建模与船模航行试验相结合的风险评价框架,分为模型构建、量化和验证3个部分。在模型构建方面,模型节点来自内河环境、通信、岸端、船端和风险5个方面辨识的船舶远程控制航行风险因素,根据专家经验和航行试验对初始模型结构进行迭代修正,形成了远程控制船舶内河航行风险评价的贝叶斯网络模型;在模型量化方面,为量化父子节点之间线性或非线性关系提出了条件概率计算方法,在内河连续桥区开展船模航行试验,获取了部分根节点的先验概率、航行事故或险情,以及试验人员对航行风险的评价等数据;在模型验证方面,从情景分析、敏感性分析、对比分析3个角度对模型进行了验证。结果表明:在统计的40起航行试验案例中,触碰/碰撞事故占比最高,试验发现了新的失效模式,即控制冲突、程序错误、信号丢失等;构建的贝叶斯网络模型能够有效评价不同情景下航行风险水平,并辨识了与远程控制船舶内河航行风险相关的重要节点。研究结果可以为实现内河船舶远程控制作业模式提供理论依据,也为内河远程控制船舶航行安全提供研究思路。
To quantify the inland navigational risk of a remote control ship,a framework is proposed integrating uncertainty modeling and navigational trial,including three parts,i.e.,model construction,quantification,and verification.In the part of model construction,nodes of the model are navigational risk influencing factors of remote control ship identified from five aspects,i.e.,the environment,ship side,shore side,communication,and risk.An initial model is built and iteratively updated to construct a Bayesian Belief Network model for risk assessment on a remote ship in an inland river according to experience from experts and navigational trials.In the part of model quantification,a method is proposed to calculate Conditional Probability Tables for Bayesian reasoning,which can describe the linear or nonlinear dependencies between nodes with multiple states.The prior probabilities of some root nodes are derived from the navigational trials that are conducted in an inland river with multiple bridges.During the trials,the recorded items also include the situation of accidents or incidents and risk of navigation evaluated by a risk analyst onboard.In the part of model verification,the model is verified from three analyses,i.e.,scenarios,sensitivity,and comparison analysis.The results of this study are as follows:In the counted 40 cases in the trials,accidents,i.e.,contact or collision,occupy the largest percentage.During the trials,there are some emerging failure modes,i.e.,control conflict,software error,and signal loss.The built Bayesian Network(BN)model can effectively judge the risk levels in different scenarios.Furthermore,the BN model identified nodes that are important to the inland navigational risk of remote control ships.The results can not only provide a theoretical basis for inland ships facilitating the remote control operational mode but also show a methodology for the navigational safety of inland remote control ships.
作者
范存龙
张笛
翁金贤
付姗姗
韩冰
FAN Cunlong;ZHANG Di;WENG Jinxian;FU Shanshan;HAN Bing(College of Transport&Communications,Shanghai Maritime University,Shanghai 201306,China;School of Transportation and Logistics Engineering,Wuhan University of Technology,Wuhan 430063,China;National Engineering Research Center for Water Transport Safety,Wuhan University of Technology,Wuhan 430063,China;Academician Workstation,China COSCO Shipping Corporation Limited,Shanghai 200135,China;Shanghai Ship and Shipping Research Institute,Shanghai 200135,China)
出处
《安全与环境学报》
CAS
CSCD
北大核心
2024年第1期72-83,共12页
Journal of Safety and Environment
基金
国家自然科学基金项目(52072237)
上海市青年科技启明星计划项目(22QC1400600)。
关键词
安全工程
水路运输
远程控制船舶
内河航行
航行试验
风险评价
贝叶斯网络
safety engineering
water transportation
remote control ship
inland navigation
navigational trial
risk assessment
Bayesian Network(BN)
