超大型油船作业过程溢油风险网络分析方法

Approach for Dynamic Network Analysis of Oil Spill Risk in Process of VLCC Operation

对超大型油船(VLCC)装卸作业安全有效地预控,有必要评估及预测VLCC靠港装卸作业过程风险。通过分析VLCC装卸作业流程及各任务环节之间的联系,提出装卸作业的形式化描述、层次化风险辨识模式及过程化建模,确定VLCC装、卸油风险预测的贝叶斯网络结构。利用贝叶斯算法推理,得到相应的后验概率,定量识别装、卸油作业过程的风险环节及相互依赖关系。基于风险环节的马尔可夫性假设,将静态贝叶斯网络拓展成带有时间参数的动态贝叶斯网络。结合场景数据,动态推理整个装、卸作业过程的风险时序变化规律。算例结果表明,VLCC装卸作业过程中溢油风险是马尔可夫过程。通过对实际VLCC装卸作业过程设置证据信息,可实现对装卸作业过程风险的定量评估及预测。

In order to pre-control VLCC operations safely and effectively,it is necessary to assess and predict the risks of VLCC loading and unloading operations in ports.By analyzing the VLCC loading and unloading operation process and the connection between each task link,the formal description is proposed with hierarchical risk identification mode and process modeling of loading and unloading operations.The Bayesian network structure of VLCC loading and unloading risk prediction is determined,The corresponding posterior probability is obtained by Bayesian algorithm inference,and the risk links and inter-dependence is quantitatively identified in the process of oil loading and unloading operations.Based on the Markov hypothesis of the risk network,the static Bayesian network is extended to a dynamic Bayesian network with time parameters.Combining scenario data,the risk sequence change law of the entire loading and unloading operation process are dynamically predicted.The results of the calculation example show that the oil spill risk during VLCC loading and unloading operations is a Markov process.By setting evidence information in the actual VLCC loading and unloading operations,the quantitative assessment and prediction of the risks can be realized in the loading and unloading operations.