PZT array for passive guided wave tomography of extended defects using ambient elastic noise cross-correlations
Résumé
Structural health monitoring (SHM) consists in embedding sensors in a
structure like an aircraft or a naval ship in order to detect defects (for example
cracks or corrosion in metallic materials or delamination in composite materials)
before a serious fault occurs in the structure. Guided elastic waves emitted by a
sensor and propagating to another one are often used as the physical way of
detecting the defect. However, the implementation of SHM systems is restricted in
many situations by the necessity to store or to harvest the electric energy necessary
to emit the waves.
A promising way to tackle this constraint is to use techniques based on the
cross-correlations of the ambient acoustic noise in place in the structure. It has been
shown that, under certain conditions, transient response between two sensors can be
estimated from cross-correlation of ambient noise, with purely passive
measurements.
The idea is to take advantage of the elastic noise naturally present in the
structure (due to engine vibrations or aero-acoustic turbulences on the fuselage of an
aircraft for example) in order to avoid the emission of the elastic waves by the SHM
system. The complexity of the embedded SHM system is therefore reduced.
We present here studies of noise cross-correlation techniques that have been
conducted with the aim of doing passive tomography of extended defects (such as
corrosion or delamination) using an array of piezoelectric (PZT) transducers. Noise
is generated by spraying compressed air on the surface of an aluminum plate.
Passive measurements are compared to active signals to demonstrate the
convergence of the cross-correlation technique to the Green function of the system.
Experimental results which come from tomographic time-of-flight imaging
algorithms will also be described. Finally, an extension of this technique using
purely passive guided wave sensors such as Fiber Bragg Grating (FBG) will be
presented.