Detecting ice on plate-like structures via flow-induced random vibration: a dispersion curve shift identification approach

This paper proposes an ice detection framework for thin plate structures using flow-induced random guided waves measured by two passive sensors. Its principle is based on the fact that the ice accretion shifts the dispersion curve of guided wave, which can be extracted from the cross-correlation of ambient wave fields at two receivers. More specifically, the group velocity under various frequencies is identified from the prominent peak of the cross-correlation with a band-pass filter, which signifies the travel time of a narrow-band guided wave, and this forms a reconstruction of the dispersion curve. The ice thickness is then estimated by minimizing the weighted error between the reconstructed dispersion curve and its theoretical model solved from the Rayleigh–Lamb equations for the two-layer (ice-plate) structure. The weights are assigned to various frequencies according to the global sensitivity of the group velocity versus the ice accretion. The proposed method is assessed by a laboratory experiment where the ice accretion on an aluminum plate is carried out in an air-cooling environment and the random vibration is excited by spraying air jet onto the plate surface. Experimental results show that the ice thickness can be accurately estimated if it is of the same order or even lower than the plate thickness. The proposed method has the potential to realize the long-range real-time detection of aircraft icing conditions where the passive sensors can be tiny, light, free of energy supply, and mounted on the internal surface of fuselage skin without any effect on aircraft aerodynamics.