为有效管理海洋环境污染,需要识别船舶溢油风险(SOSR)因素,并研究其对事故后果影响的动态特性。在辨识SOSR的基础上,构建危险源网络图,并考虑其动态性。建立动态贝叶斯网络(DBN)预报模型。运用该模型,量化SOSR因素与溢油事故的可能性和后果之间的逻辑组合关系,推理风险变量的后验概率,并获取风险因素时间状态转移效果和溢油事故与事故后果的时序变化规律。以某港口1999-2012年间船舶溢油事故数据为基础,运用模型分析表明:船舶溢油事故风险波动性很强,重大溢油事故的风险现阶段不稳定性突出,相对风险值为-41%~53%,溢油事故有扩散性和自繁衍特征。
It is held by authors that in order to manage marine pollution effectively, the factors affecting SOSR should be analyzed, the dynamic characteristics of factors and consequences of accidents should be studied. A hazards network diagram of SOSR was drawn after identification of factors for ship oil spill. A DBN prediction model of SOSR(SOSR-DBN) was built after considering the dynamic character. Quantitative logic relations between possibilities and consequences of risk factors and accidents of ship spill can be measured using the model. Then, posterior probability of risk variables can be reasoned, time state transition effect of risk factors can be obtained, and temporal variation of risk accidents and consequence of accident can be obtained as well. The model was used to analyze the ship oil spill data of a certain port area from 1999 -2012. Analysis shows that risk of ship oil spill is highly volatile, and instability of risk of ma- jor accidents is salient at present stage, and that relative risk values vary from -41% to 53%, indicating that ship oil spill has characteristics of diffusion and self-reproduction.