摘要
为了预报船载易流态化固体散装货物的风险态势,利用隐马尔科夫模型(HMM),定量化评判船舶营运安全风险。首先确定易流态化货物船舶运输风险的影响因素,构建评判结构模型,采用HMM刻画变量的关联特征;然后通过问卷调查数据和Baum-Welch算法确定模型参数,根据特定的航次信息,结合前向算法和风险衡准描述风险瞬态和风险态势;最后通过建立多场景的模拟航次,检验风险评判模型的有效性。研究表明:航次前阶段货物流态性对船运风险的影响明显,影响度为1.49;后阶段环境恶化对船运风险影响大,影响度高达1.65。
In order to forecast risk prospect of ship carrying liquefiable cargoes,the quantification of risk estimation in the process of ship safety operation was studied by HMM.The factors influencing ship transportation risk of liquefiable cargoes were identified,and an estimation structure model was built.The relevance of the variables was described by HMM.The model parameters were determined according to the questionnaire survey data and Baum-Welch algorithm.Voyage information was combined with forward algorithm and risk criterion to describe the risk transient and prospect.The sensitivity test of voyage risk estimation in multi-scene was carried out to verify the effectiveness of the model.The results indicate that liquefiable cargo has a big influence on the navigation risk in the pre-stage of the voyage,whose influence degree is 1.49,and the environmental deterioration has a huge impact on the risk in the post-stage,whose influence degree is as high as 1.65.
作者
吴建军
刘英学
胡甚平
赵义豪
WU Jianjun;LIU Yingxue;HU Shenping;ZHAO Yihao(College of Merchant Marine, Shanghai Maritime University, Shanghai 201306, China)
出处
《中国安全科学学报》
CAS
CSCD
北大核心
2017年第11期73-78,共6页
China Safety Science Journal
基金
国家自然科学基金资助(11671416)
中国博士后科学基金资助(2016M591651)
上海国际港务(集团)股份有限公司科技创新项目(引航站_17KY-04B-31Z)
