Multi-spectrum fusion for single-probe blade tip timing

Abstract

Blade tip timing (BTT) is currently the most promising method for online turbine blade health monitoring. Single-probe BTT minimizes the number of probes required and has the advantages of low cost and high safety. However, the number of probes in single-probe BTT makes its analytical capability significantly weaker than that of multi-probe BTT. In this study, we propose a multi-spectrum fusion (MSF) single-probe BTT signal analysis method that considerably improves the frequency analysis capability of the single-probe BTT and provides a theoretical upper bound for the frequency analysis. The feasibility and robustness of the MSF are verified through numerical simulations and experiments. In addition, we present methods to determine the number of frequency components of the BTT signal, which further reduces the dependence of MSF on a priori information. Finally, we demonstrate that the single-probe BTT is almost identical to the multiprobe BTT in terms of analyzing the asynchronous vibration of a blade.