PhD
thesis: Defect detection in plates using guided waves
Experiments
The setup consists of modular components controlled from a central computer,
shown below. The aim of the measurements is twofold and therefore two different
types of specimens were employed. To compare the measurements with theoretical
calculations, a large aluminum plate allows the choice of the experimental
parameters in a wide range. For the study of the crack detection, a tensile
specimen with a fatigue crack generated at the hole was used.
The
experimental sequence goes from excitation over measurement to data analysis.
The excitation pulse is generated as a voltage signal in the function generator,
amplified and applied to the excitation transducer, where it is converted into a
flexural wave in the plate. Usually piezoelectric transducers were used and
their transfer function was studied. Additionally, electromagnetic acoustic
transducers (EMATs) for a point source excitation of the A0 mode were
investigated and prototypes built. Different types of pulses were studied as the
excitation signal, especially sine cycles in a Hanning window and sinusoidal
sweeps, offering good control over the frequency content. A narrow bandwidth of the excitation pulse avoids extensive signal
distortion due to the dispersive character of the A0 mode. The energy
of the pulse is concentrated around the center frequency and a
good signal to noise ratio is achieved. Most of the experiments were carried out
using a narrow bandwidth excitation pulse.
The wave propagates along the structure and is scattered at the obstacle.
The wave front can be assumed to be a straight line when it reaches the hole,
facilitating the theoretical simulation. The incident wave is scattered at the
stress-free boundaries of the hole, and a scattered wave is generated. The
scattered wave consists of two parts: a boundary layer close to the hole and a
part propagating radially outwards from the hole. In the vicinity of the hole,
incident and scattered wave overlap in time, so that only a single pulse is
visible in the measurements. This is due to the length of the excitation signal
and the low group velocity of the A0 mode. The specimen size is
selected large enough that the reflections from the plate boundaries reach the
measurement area some time later. This way a time separation between the
scattered field and the boundary reflections
is achieved.
The
scattered field on a measurement grid around the hole is recorded using a
commercially available heterodyne laser interferometer. The demodulator output is a voltage signal proportional to the
velocity of the out-of-plane component of the displacement of the plate surface.
The measurement spot is defined by the laser beam diameter, which is well below
0.1 mm. This allows a point-wise measurement of variations in the scattered
field, as no implicit average over a rather large surface of the measuring
transducer is made. The laser interferometer is moved parallel to the specimen on a
positioning system, allowing a measurement in the vicinity
of the defect without disturbance. The measurements are well repeatable, with
the largest cause of variation due to an inaccurate positioning of the laser
beam relative to the hole center, which could only be achieved with an accuracy
of about 0.1 mm. Two types of measurement grids were used. For measurements in the
vicinity of the hole a radial grid was used, moving the measurement spot on
concentric circles around the hole. For the experiments on
the propagation characteristics in the tensile
specimens, a Cartesian grid was used.
The voltage signal is bandpass
filtered around the center frequency and
averaged in a digital storage oscilloscope. The function
generator triggers the oscilloscope, so that excitation and measurement start at
the same time. The measured time series are then transferred to the computer for
evaluation. At each point of the grid a time series
with usually 10000 values is stored. A time gating is applied to cut off the
reflections caused by the plate boundaries, which contain no information about
the scattering at the hole. The arrival times of the different pulses are
calculated from the theoretical group velocity. Fast Fourier transform (FFT) is
applied and the amplitude and phase values at the center frequency
of the excitation signal are
extracted for each measurement point. These values are the equivalent of the theoretical results, where an
infinite sinusoid is assumed for the incident wave.
They can either be displayed on a circle
around the hole or as a pseudo-colored surface around the hole, shown below. At
the free boundary of the hole the incident wave is scattered and a high
amplitude directly at the boundary results. A flaw like a crack or a notch
introduces additional free boundaries, from which result a local peak in
amplitude and a change of the scattered field. Through these measurements a good
understanding of the geometry of the scattered wave and the influence of a notch
or a crack can be gained.
The measurements were made
automatically, using a program written in LabView. The grid, different
excitation pulses and frequencies, and further measurement parameters can be
selected by the user. From the measurement of the
whole scattered field a good notion of the geometry of the scattered wave was
obtained. To characterize the influence of a notch on the scattered field,
measurements were made before and after the notch was cut. From the difference
between the measured scattered fields, the geometry and propagation characteristics
of the wave scattered at the notch could be obtained.
However, such measurements are
too time-consuming in the industrial application for nondestructive testing
purposes. For a fast detection of changes in the scattered field due to a notch
or crack, a measurement at one point or on a line is sufficient. Such
measurements were made for the tensile specimens during the tensile testing to
monitor the crack propagation.
