贝叶斯网络在海洋平台应急撤离成功率分析中的应用  

Estimating probability of success in EER from a offshore platform in emergency based on Bayesian network

作　　者：孔得朋[1] 王克[1] 陈国明[1] 刘鹏翔[1] 

机构地区：[1]中国石油大学(华东)海洋油气装备与安全技术研究中心

出　　处：《中国安全科学学报》2016年第9期169-174,共6页China Safety Science Journal

基　　金：国家自然科学基金资助(51504282);博士后科学基金会面上项目(2014M 560592)

摘　　要：为评估海洋平台应急撤离(EER)成功率,首先运用鱼骨图法定性分析影响应急撤离逃生、疏散、救援3个阶段的因素;进而建立3阶段的贝叶斯网络(BN)模型,并计算各阶段的成功率;最后对BN进行敏感性分析(SA)。结果表明:在影响3阶段因素状态未知的情况下,逃生、疏散、救援成功的概率分别为0.84,0.7和0.37,且因素状态对应急撤离成功率有较大影响;SA结果表明,疏散路线选取因素对逃生阶段的成功率的影响最大,而天气因素对疏散和救援阶段的成功率的影响最大。To estimate probability of success in EER from a offshore platform under an emergency condition, factors affecting the probability were identified by using the fishbone diagram method. Three BN models were built for the three stages of EER. Probabilities of success were calculated for the three stages. The results show that the probabilities of success in escape, evacuation, and rescue are 0.84, 0.7, 0.37 respectively, if the state of the influence factors are unknown. A sensitivity analysis was made for the probabilities.The results show that the state are great impact on the probability of success in EER, that the factor of ＂evacuation route selection＂ has the most important effect on the success in escape stage, and the ＂weather＂ factor is the most influential on the success in the evacuation and rescue stages.

关 键 词：应急撤离(EER) 影响因素 鱼骨图法 贝叶斯网络(BN) 敏感性分析(SA) 

分 类 号：X937[环境科学与工程—安全科学]