融合贝叶斯网络与变权AHP的气量调控系统安全风险动态分析方法

Dynamic analysis method of safety risk of capacity control system based on Bayesian network and variable weight AHP

针对往复压缩机组无级气量调控系统故障率较高、对机组安全运行影响大,而现有系统运行风险评价方法难以完成实时、定量分析的问题,提出一种融合贝叶斯网络与变权AHP的气量调控系统安全风险动态分析方法。建立了包含故障类型、故障模式与监测信号的贝叶斯网络模型,得到故障实时发生概率;基于变权层次分析法建立了故障危害性半定量分析模型,计算故障模式影响量化指标;进一步,修正了FMEA方法故障模式风险计算公式,引入故障发生概率,提出了历史故障与实时运行风险计算方法。以实际故障案例数据进行测试验证,结果表明,提出的新方法可量化计算机组实时与历史运行风险,设置归一化的实时运行风险阈值为0.5,以判定检修必要性。研究成果可为往复压缩机及气量调控系统检维修计划的制定提供量化指标。

In view of the problem that the stepless capacity control system of reciprocating compressor has a high failure rate and great impact on the operation of the unit,and the operation risk assessment methods of the existing system are difficult to complete the real-time and quantitative risk assessment,a risk dynamic analysis method for the safety of stepless capacity control system was proposed,which integrates Bayes network and variable weight AHP.The Bayesian network model,which includes fault type,fault mode and monitoring signal,was established to obtain the real-time probability of fault occurrence.Based on variable weight analytic hierarchy process(AHP),a semi-quantitative analysis model of fault hazard was established to calculate the influence index of fault mode.Further,the fault risk calculation formula of FMEA method was modified,the probability of fault occurrence was introduced,and the calculation method of historical fault and real-time operation risk was proposed.Test verification was carried out using the actual fault case data.The results show that the new method can quantify the real-time and historical operating risks of computer units,and the normalized real-time operating risk threshold is set at 0.5 to determine the need for maintenance.The research results can provide quantitative indexes for the development of inspection and maintenance plans for reciprocating compressors and gas volume regulation systems.