ProTopormer: Toward Understandable Fault Diagnosis Combining Process Topology for Chemical Processes

In modern chemical processes, fault detection and diagnosis(FDD)is a key part of abnormal situation management (ASM). As researcherscontinue to improve the fault diagnosis performance of different models,the emphasis on model interpretability and explainability studieshas increased in recent years. In this paper, a novel model, ProTopormer,was proposed for fault diagnosis of chemical processes. Self-attentionmechanism and process topology knowledge were fully combined to achievehigh model performance and good interpretability. Experiments on TennesseeEastman process showed that the model achieved a high diagnosis rateand a low false alarm rate. Attention weights were visualized andquantitatively analyzed to identify key variables for fault diagnosis,which showed strong explanations for the root causes of differentfaults.