The Diagnostic Effect of Localized-CECP Algorithm on Different Rotary Components

The Complexity-Entropy Causality Plane (CECP) is capable of converting given time series into two-dimensional vectors and simplifying their characteristics in the form of points on a two-dimensional plane. The CECP algorithm has a good performance in distinguishing normal and fault data of different rotary components. However, for different objects, its fault or normal characteristics are not the same. As a result, when applying CECP to a complex mechanical system, the extracted normal and fault features may be confused due to multiple faults. Therefore, this paper intends to improve the CECP algorithm by localizing it, so as to retain its fault diagnosis effect for different moving parts, and unify the fault and normal characteristics to diagnose system faults.