Digital twin-driven blade rub-impact diagnosis using blade tip timing

Abstract

Blade Tip Timing (BTT) has become a promising method for online monitoring of rotating blades due to its non-contact measurement characteristics. However, due to the undersampling and non-uniform characteristics of blade tip timing signals, there are few studies on using them to diagnose rub-impact fault. This paper focuses on this topic and proposes a rub-impact diagnosis process. The contents of this paper are as follows: (1) A review of the existing rub-impact diagnosis technology and signal processing methods, then the advantages of diagnosis rub-impact using BTT are given. (2) The digital twin model framework for rotating blades has been established, including physical entity, digital entity, and the connections between them. (3) Dynamics analysis of both single blades and entire blade disks under rub-impact states has been carried out for the digital twin model, and six signal characteristics of blade tip under rubbing condition are summarized which provide prior knowledge for subsequent diagnosis. (4) Aiming at that traditional Enhanced Sparse Decomposition (ESD) method is prone to falling into local optimal solutions for BTT undersampled data, an envelope spectrum kurtosis enhanced sparse decomposition (ESKESD) method is proposed. (5) A complete rub-impact fault diagnosis process using digital twin is proposed. Finally, the effectiveness of the proposed method and diagnosis process are verified through simulation and experimental data.