基于因果分解的SOFC系统故障根源诊断方法

Root Cause Diagnosis of SOFC System Based on Causality Decomposition

固体氧化物燃料电池(solid oxide fuel cell,SOFC)具有高效率、低排放、长寿命运行等特点。SOFC系统发电过程中存在时变、非线性和非平稳性等现象,会引起性能演化衰变,进而导致不明故障而停机。鉴于此,文中提出了一种基于因果分解的故障根源诊断方法。该方法运用集成经验模式分解(EEMD)方法将过程数据分解为有限个固有模式函数(IMF)以表征原始信号在不同时间尺度上的局部特征,同时引入瞬时相位相关性概念以定量评估IMF分量间因果作用,从而达到因果分解的目的。此外,该方法在考虑时间优先原则的同时,强调因果交互作用的瞬时关系,因而增强双向因果关系检测与分析能力。实验结果表明,该方法通过可靠的故障传播路径分析可实现故障源准确定位,为故障演化机理分析提供支撑,具有较强实用性。

Solid oxide fuel cell(SOFC)has the characteristics of high efficiency,low emission and long lifetime.There are time-varying,nonlinear and non-stationary phenomena in the generation process of SOFC systems,which may cause the degradation of performance evolution and in turn leads to the shutdown of unknown faults.In view of this,we propose a root cause diagnosis method based on causality decomposition.In this method,the ensemble empirical mode decomposition(EEMD)is used to decompose the process data into a limited number of intrinsic mode function(IMF)to characterize the local characteristics of the original signal at different time scales.At the same time,the concept of instantaneous phase correlation is introduced to quantitatively evaluate the causal effect between the IMF components.The experiment shows that the proposed method can accurately locate the fault source through reliable fault propagation path analysis.It provides support for fault evolution mechanism analysis and has strong practicability.