基于多态模糊BN的石化动设备故障概率定量风险评估

石化行业的压缩机、泵等动设备故障模式多样,故障概率的不确定性较大,常用的风险评估方法无法实现对石化行业动设备失效概率的定量评估,限制了设备安全管理的准确性和有效性.基于贝叶斯网络(BN)可定量计算复杂系统失效概率的特点,利用故障模式与影响分析(FMEA)方法获取动设备风险事件的因果关系,将其映射为BN,并利用BN中节点...

Quantitative Risk Assessment of Mechanical Equipments Failure Probability in Petrochemistry Industry Based on Multiple-state Fuzzy BN

Compressors,pumps,and other mechanical equipments in the petrochemistry industry have diverse failure modes and uncertain failure probability.Therefore the commonly methods cannot achieve quantitative assessment of failure probability,which limits the accuracy and effectiveness of risk control and maintenance.Based on the characteristics of Bayesian network(BN)that can quantitatively calculate the failure probability of complex systems,this paper uses Failure Mode and Effects Analysis(FMEA)to get the relationship between risk events of mechanical equipments and map it to BN.Furthermore,the paper adopts the polymorphism of BN nodes to characterize the diverse failure modes of subsystems and components.Finally,the paper uses the fuzzy probability subset to replace the exact value of the edge probability to express the uncertainty,and proposes a feasible method of conditional probability judgment.Thus,a quantitative failure probability assessment method of the mechanical equipments in oil and gas stations is formed.Finally,the paper conducts a case study by analyzing a reciprocating compressor in a natural gas station with the method,showing that the failure probability of gas valve and cylinder are 0.0040 and 0.0045 respectively;the probability of sheet and spring damage and the bush loose are 1.15×10^-5 and 9.92×10^-6 when the compressor fails,which confirms to the reality verifying the validity and feasibility of the method.