Detection of cracks in turbine blades with nonlinear ultrasonic frequency modulation

Non-destructive testing and structural health monitoring (SHM) techniques are becoming increasingly important for gas turbine manufacturers. Incipient cracks in the turbine blades have to be detected before catastrophic events occur. Linear ultrasonic methods are widely used to inspect structural integrity by monitoring the amplitude and phase variations of the mono-frequency input signal due to linear scattering caused by damage. However, closed cracks or small cracks cannot be easily detected due to a low impedance mismatch with the surrounding materials. Nonlinear ultrasonic methods have demonstrated the ability to detect early-stage damages. These methods investigate the distortion of the elastic waveform due to damage/material nonlinearity. This generates new signal components such as sub- and high-harmonics of the fundamental frequency in the frequency spectrum. The aim of this study was the development of a frequency modulated technique for the detection of cracks in turbine blades. Experimental work was carried out on flat samples with artificial defects and turbine blades excited by two different frequencies. A new global nonlinear parameter was used to determine a correlation between crack length and measured nonlinear features. The results were compared with samples without any defects. The findings showed a clear trend of increasing nonlinear parameters as a function of the crack size.