摘要
为进行垂直卡爪式水下连接器机械结构故障诊断及失效预测,提出一种3层动态贝叶斯网络的垂直卡爪式水下连接器故障诊断方法。针对垂直卡爪式水下连接器常见部位(卡爪、密封圈和驱动环)及失效方式,利用动态贝叶斯网络模拟材料状态随时间的退化,采用专家打分法获得连接器部件的失效数据,从而确定故障层状态的先验概率与故障发生后症状层状态的条件概率,并利用GeNle软件求解动态贝叶斯网络。结果表明:根据观测到的症状层状态,能够得到3种关键部件的后验失效概率变化趋势,并确定故障部件。
In order to diagnose fault and predict failure for mechanical structure of vertical collet subsea connectors,a fault diagnosis method based on three-layer dynamic Bayesian network is proposed.Targeting at common parts(collet,sealing ring and driving ring)and failure modes of connectors,Dynamic Bayesian network was used to simulate degradation of these materials over time,and failure data of components were obtained through expert scoring,so as to determine prior probability of fault layer state and conditional probability of symptom layer state after occurrence of fault as well as solve dynamic Bayesian network by using GeNle software.The results show that based on observed symptom layer state,change trend of posterior failure probability of three key components can be obtained,and faulty component can be found.
作者
陈志煌
刘国恒
王莹莹
朱春丽
单荐
翟小东
CHEN Zhihuang;LIU Guoheng;WANG Yingying;ZHU Chunli;SHAN Jian;ZHAI Xiaodong(College of Safety and Ocean Engineering,China University of Petroleum(Beijing),Beijing 102200,China;CNOOC Research Institute,Beijing 100027,China)
出处
《中国安全科学学报》
CAS
CSCD
北大核心
2020年第5期81-87,共7页
China Safety Science Journal
基金
国家高技术船舶科研项目(2018GXB01-07)
国家重点研发计划项目(2016YFC0303701)。
