Metal materials and structures are commonly inspected by the acoustic emission (AE) technique. Any event of crack
propagation or corrosion development results in the emission of elastic waves which are captured by sensors on the
surface of the material. Study of the AE signal incoming rate, as well as qualitative signal parameters, reveals crucial
information on the extent of damage and the cracking mode. Based on laboratory experiments, classification criteria are
established concerning the type of the active damage source. However, elastic wave propagation in metal plates is
dispersive, forcing different frequencies to propagate on different velocities. This leads to shape distortion of the signals,
altering their specific features, like duration, amplitude, number of cycles which are crucial for AE characterization. Due
to dispersion, an AE event from a single source will be acquired with very different shape in distinct sensor positions and
the differences will be accumulated as the propagation path increases. In the present paper numerical simulations of
wave propagation in metal plates are conducted. The aim is to investigate plate wave dispersion, not from the classical
ultrasonics, but from the AE point of view, quantifying the effect of distance on the shape of the propagating pulses and
the specific AE parameters. Indicative experiments on metal plates confirm the effect of distance on the wave parameters.
Consequently, a procedure to "correct" the AE parameters based on the propagating distance between the source crack
and the sensor is discussed. This way the original AE parameters of the signal as emitted by the source are calculated and
the effect of distortion which masks the original content is cleared out. It is shown that any classification scheme based
on AE parameters, should incorporate the information of the source location relatively to the sensor since long
propagation distance causes strong changes in the waveform, masking the information of the original source.
|