Active aliasing ESPRIT: A robust parameter estimation method for low-intervention Blade tip timing measurement

Abstract

Rotating blades are critical but fragile components in aeroengine. Damage to rotating blades will sharply reduce the working efficiency and even endanger operational safety. Thus, it is significant to monitor the blades. Blade tip timing (BTT) is an emerging vibration measurement technique for rotating blades and is considered a promising approach for blade condition monitoring owing to its non-contact property and long service life. The key to BTT application is extracting vibration parameters from the measured signals that reflect the blade health condition. However, BTT signals are inherently undersampled and hard to analyze by traditional methods. Most existing BTT analysis methods require multiple probes, typically 4$\sim$7 probes. Due to weight, safety, installation, and maintenance costs, it is desired to implement BTT measurement and extract vibration parameters with as few probes as possible. In this paper, we propose a low-intervention BTT measurement-based signal post-processing technique, termed AA-ESPRIT, which is a practical variant of classic ESPRIT. Remarkably, AA-ESPRIT overcomes the limitations of ESPRIT in BTT application and significantly improves the estimation accuracy by actively utilizing aliasing instead of hastily suppressing aliasing. Both numerical and experimental results show the effectiveness of AA-ESPRIT in the presence of measurement noise and speed fluctuation. In addition to satisfactory estimation performance, AA-ESPRIT can work with only two probes and lead to a low usage cost; thus, it is expected to have its place in the BTT field.