基于余数定理的叶端定时欠采样信号处理方法

Sub-sampled Blade Tip-Timing Signal Processing Based on the Remainder Theorem

提出了一种基于余数定理的叶端定时欠采样信号处理方法,该方法仅根据两组欠采样频率下的振动信号频谱信息,通过简单运算即可直接计算叶片的振动频率,优于传统的"动频+静频"方法,并利用"3+2"两组叶端定时传感器数据进行了仿真,结果验证了其有效性。

High-speed rotating blades are key mechanical components in gas turbine compressors and turbine. Under continuous vibration stresses, high-cycle fatigue faults are prone to often happen in the blades, in particular under resonant conditions. Thus on-line detection of blade vibrations is very important in practice. Now blade tip-timing(BTT) method is a common non-contact method of detecting blade vibrations. However, blade vibration frequencies are always much higher than that of the rotating shaft, so blade vibration signals collected by BTT methods are one kind of typical sub-sampled signals.Traditional Fourier transform cannot be used directly here. In this paper, a new sub-sampled BTT signal processing method is presented based on the Remainder Theorem. This method can be used to calculate blade vibration frequencies directly only using two groups of sub-sampled BTT signals. It is much better than traditional methods based on' static and dynamic' frequencies. In the end, '3+ 2' BTT signals are simulated to validate the proposed method and the results verify its feasibility.