基于动态风险预测的船舶引航任务分级管理  

Hierarchical management of ship pilotage tasks based on dynamic risk prediction

作　　者：郭云龙[1] 胡甚平[2] 轩少永[2] 李伟[1] 曹久华 

机构地区：[1]南通航运职业技术学院航海系,江苏南通226010 [2]上海海事大学商船学院

出　　处：《中国安全科学学报》2014年第12期63-69,共7页China Safety Science Journal

基　　金：国家自然科学基金资助(51149001);上海海事大学校基金资助(20120057);上海海事大学研究生创新基金资助(GK2013032);广东省交通运输厅科技项目(201202004)

摘　　要：为加强引航安全,实现引航任务的科学管理,需要对其进行有效的分级。基于引航风险的物理属性,从被引船舶和环境2方面入手,辨识船舶引航过程中的动态风险,考虑引航任务风险的动态特性和风险因素间的网络结构关系,引入动态贝叶斯网络(DBN)构建风险预测模型。根据模型预测得到的引航任务风险值,对其进行分级管理。以青岛港2002—2011年10 a间的引航事故数据为基础,应用模型对港口的引航任务进行预测,据此确定风险大小与引航员等级的匹配标准。结果表明:利用所建模型预测得到的引航任务动态风险与实际情况较吻合,说明根据风险大小对引航任务进行分级管理的方法具有可行性。In order to strengthen the pilotage safety and achieve the scientific management of pilotage tasks, pilotage tasks require effective classification. Based on physical properties of pilotage risk from aspects of piloted ship and environment, dynamic risks in the piloting process were identified. A risk prediction model was built through introducing DBN after considering the network structural relationship among risk factors and the dynamic characteristics of pilotage task＇s risks. The pilotage tasks will be managed based on sizes of predicted dynamic risks using the model. The model was used to predict risk of pilotage task of Qingdao port based on the pilotage accident data from 2002 -2011, the matching standard of risk size and level of pilot was determined. The results show that predicted dynamic risks in pilotage tasks using this model are in good agreement with actual situation, and managing the pilot tasks hierarchically through predicting the risks in them is feasible.

关 键 词：引航任务 分级管理 动态风险 预测 动态贝叶斯网络(DBN) 

分 类 号：X951[环境科学与工程—安全科学]