Review of Root Cause Diagnosis and Propagation Path Identification Techniques for Faults in Industrial Processes

Research of these two techniques not only enriches the fault diagnosis theory, but also has important significance for promotion and application of fault diagnosis technology in actual projects.

In this paper, the basic ideas, application conditions, advantages, as well asdisadvantages of knowledge based, data based and joint knowledge and data based methods are illustrated.Moreover, research status of related methods is classified and summarized. Finally, some urgent problems and future directions in this field are discussed, including: 1) Root cause diagnosis and propagation path identification under three-dimensional perspective; 2) Manufacturing big data analysis and causality mining based accurate qualitytracing; 3) Distributed diagnosis for propagative, coupled, and concurrent multiple faults; 4) Multi-source eterogeneous dynamic information fusion based timespace traceability visualization for abnormal conditions.