汽轮机故障链诊断与评估方法研究

Research on Fault Chain Diagnosis and Evaluation of Turbine

对汽轮机进行诊断与维修决策时,大多数方法将故障看作独立的事件或模式。而实际中,故障事件彼此并不孤立,常有故障彼此诱发的情况。因此,故障是一个连锁的变化过程。为更准确地诊断和评估设备故障严重程度,并通过检修有效降低机组故障风险,提出基于故障链诊断与评估的维修决策方法。首先,通过故障机理分析,以故障因果网描述设备可能的故障过程。然后,通过贝叶斯网络诊断模型,推理各故障事件发生的可能性;再结合故障静态风险与故障劣化度评估,计算故障独立严重程度和故障链严重程度。最后,基于故障链严重程度,为故障检修顺序提供建议。与预防性维修决策和传统的状态维修方法相比,该方法可更有效地降低设备运维风险。汽轮机故障案例应用表明,该方法可有效应用于复杂的汽轮机诊断与维修中,为运维工作提供更有价值的检修建议。

The vast majority of methods of eliminating faults are erroneously regarded as independent events or patterns,while diagnosis of faults for turbine or strategic decision for maintenance is made.But in fact,the fault events are not independently occur from each other,are chain-reaction process in disorder often induced by fault events.In order to diagnose and evaluate severity of equipment fault more precisely,and reduce the fault risk effectively through maintenance,a decision-making method for maintenance based on the fault chain diagnosis and evaluation is proposed.First of all,the possible equipment fault process is described through fault mechanism analysis by fault cause and effect network,moreover,the Bayesian Network is used to inference possibility of each fault event.Then,by combining the static risk and degradation degree,the severity of independent fault and chain are calculated.Finally,the maintenance sequence is recommended on the basis of severity of fault chain.Comparing with the preventive maintenance decision and traditional condition maintenance method,the new method is able to effectively reduce operational risk.The method is applied to a real turbine fault case,which shows that the method can be effectively applied to complicate turbine diagnosis and maintenance,and more valuable maintenance suggestions for operational work are provided.