贫数据条件下海底电缆故障概率评估方法*

为解决贫数据引起海底电缆失效概率评估的不确定性影响，实施有效的海底电缆故障风险管理，提出1种耦合模糊集理论、层次贝叶斯分析（HBA）和贝叶斯网络的海底电缆失效概率评估方法，识别海底电缆失效致因因素，梳理各因素之间的关联关系，并采用贝叶斯网络(BN)构建海底电缆失效模型；根据数据源特点将电缆失效因素分为数据完全缺失和具有稀少的先兆数据，采用模糊集理论(FST)计算完全没有可用数据的失效致因发生概率，通过HBA估计有稀少数据失效致因的发生概率；以失效致因发生概率为输入，通过贝叶斯网络实现海底电缆失效概率的动态评估。研究结果表明:FST-HBA-BN方法可以解决基本风险因素的数据稀缺问题，量化评估海底电缆失效概率，研究结果可为贫数据条件下的电缆失效风险管理提供支撑。

Failure probability assessment method of submarine cable under scarce data

In order to solve the influence on the uncertainty of failure probability assessment on the submarine cable caused by the scarce data,and implement the effective failure risk management of submarine cable,a method for the failure probability assessment of submarine cable by coupling the fuzzy sets theory,hierarchical Bayesian analysis (HBA) and Bayesian network (BN) was proposed.The causation factors of submarine cable failure were identified,then their correlation were sorted out,and BN was used to establish a failure model of submarine cable.According to the characteristics of data source,the factors caused cable failure were divided into two types: completely missing data and scarce precursor data,then the fuzzy sets theory (FST) was used to calculate the probability of failure causation factors completely without available data,and the probability of failure causation factors with scarce data was estimated through HBA.Finally,the failure probability of submarine cable could be dynamically assessed through BN by taking the probability of failure causation factors as input.The results showed that the FST-HBA-BN method could solve the data scarcity problem of basic risk factors and quantitatively assess the failure probability of submarine cable,so as to provide support for the failure risk management of cable under scarce data conditions.