Nonstationary and nonlinear signals are often encountered in turbomachinery research and development. Sometimes the frequency of these signals changes with time. One such an example is the pulsating pressure and strain signals measured during engine ramp to find the maximum resonance strain or during engine transient in applications. As the pulsation signals can come from different disturbance sources, detecting the weak useful signals under a noise background can be difficult. For this type of signals, a novel method based on Empirical Mode Decomposition (EMD) and Teager Energy Operator (TEO) is proposed. First, the signals are processed by a self-adaptive Lifting Wavelet Transform (LWT) to remove noises and enhance the Signal to Noise Ratio (SNR). Then the EMD and Correlation Kurtosis (CK) are employed to select the sensitive Intrinsic Mode Functions (IMFs). In the end, TEO algorithm is applied to the selected sensitive IMF to identify the characteristic frequencies. A case of measured sound signal and strain signal from a turbocharger turbine blade was studied to demonstrate the capabilities of the proposed method. In this case the FFT failed to identify the blade vibratory signal at all, and the EEMD method was barely able to do so. The proposed method successfully captured the blade vibration from the both sound and strain signals.

This content is only available via PDF.
You do not currently have access to this content.