Corrosion Detection in Aircraft Structures using Guided Lamb Waves

NDTnet - November 1996, Vol.1 No.11

Corrosion Detection in Aircraft Structures using Guided Lamb Waves

Authors: A. Chahbaz, V. Mustafa and D. Hay, Tektrend International (Canada) (Homepage)

Abstract

Non-destructive testing methods for rapid and reliable corrosion detection in complex metallic assemblies is an on-going challenge due to practicalities of inspection and geometric complexity. This work demonstrates the benefits Lamb waves for detecting and locating corrosion in aluminum metallic structures (such as aircraft structures). We experimentally elaborate a guided wave testing method to determine quickly, cost- effectively and reliably where these multilayered structurally significant parts are in need of repair. This method is a global and non-conventional ultrasonic technique to locate defects such as corrosion and disbonds. To demonstrate detectability and sensitivity of guided wave techniques, experiments were performed on 1.0 mm thick aluminum plate s. Artificially induced thinning in designated areas of laboratory specimens was performed chemically. Size and shape of the corroded areas were also varied. The experimental and analytical findings indicate that the most sensitive guided wave modes to variation of the specimen thickness and surface uniformity were the lowest symmetric and antisymmetric modes (S_o and A₁).

Contents

Objectives
Conventional Ultrasonic Inspection
Guided Wave Inspection
Experiment Results
Conclusion

Objectives
To develop a global NDT procedure
To outline speed and efficient of inspection
To introduce imaging interpretation

Non-destructive testing methods for simple, rapid and reliable corrosion detection in complex metallic assemblies is an on-going challenge, this is due to the size and geometric complexity of these assemblies.
Nondestructive ultrasonic testing technique based on velocity change, attenuation and backscattering is been successfully applied by using ultrasonic bulk waves. However, plate waves whose velocity changes with frequency and thickness product can equally be used to detect defects and corrosion in multlayered metallic structures.
This work demonstrates the benefits of Lamb waves for detecting corrosion in aluminum multilayered structures.
The main objective therefore, was to develop NDT method, through a theoretical and experimental work, to detect corrosion in multilayerd aluminum structures.
We experimentally elaborated Guided wave testing method to determine quickly and reliably where these multilayered structurally significant parts are in need of repair.
To demonstrate detectability and sensitivity of guided wave techniques, experiments were performed on 1.0 -2.0 mm thick aluminum plates and facilitate interpretation results are presented through imaging.
Fig.1: Conventional Ultrasonic
With Conventional ultrasonic method like C-scan the area under interrogation at any instant is limited to region covered by the transducer. Therefore, it is a localized point by point inspection technique.
This method is an efficient conventional inspection technique but it is very time consuming for large structural areas.
C-scans also have difficulty to inspect non uniform and buried structures, since with a C-scan, a transducer needs access to each point of the inspected area.
Fig.2: Guided Waves
Lamb waves also known as plate waves are based on plate wave natural resonant modes. Lamb waves are two dimensional stress waves are guided by the geometry of the plate-like structures whose surfaces are free of stresses. They can propagate in plate-like structures that are only a few wavelengths thick (d<=3l) where l represents the incident wavelength.
Particle displacements and stresses in the Lamb waves occur throughout the thickness of the plate. Their propagation properties depend on the density, the elastic properties and geometrical structure of the inspected object and are also influenced by the thickness of the material and the wave cyclic frequency.

Fig.3: Equipment and instrumentation
The basic equipment in the instrumentation set up was a tone-burst pulser/receiver system that can be used to excite a high power narrow-band width guided wave mode. The schematic diagram of a generalized tone-burst system set up is given in the above Figure.
The output of system is a tone-burst waveform which is fed to the sender (transmitting transducer). This burst is initially formed of a continuous sine wave from the function generator which is gated and subsequently amplified. The received signal (received from the receiving transducer) is transferred to the broadband receiver through the attenuator and the digital oscilloscope.
we used Lamb wave because they offer an improved inspection potential due to their:
variable mode structure and distributions
multimode character
sensitivity to different type of flaws
propagation for long distances
guiding character which enables them to follow curvature and reach hidden and/or buried parts
Fig.4: Inspection setups
The flexibility of guided wave approach is based on mode selection, criteria which are dictated by launch angle and excitation frequency.
The most widely employed method of guided wave excitation is the wedge or prismatic coupling block method, which is based on conversion (Cook, Valkenburg and Minton 1954).
Fig.5: Dispersions Curves - Phase and Group Velocities
Dispersion curves describe the natural resonance of a specific structure. Commonly presented as a plot of Vp versus the fd product of the structure.
They depend upon material properties and the particular geometrical model selected.
We use the dispersion curves to determine the incidence angle of the transducer and to select any desired mode.
Another type of dispersion curve is presented as a plot of group velocity versus frequency*thickness product of the structure.
These graphs are essential for signal interpretation and identification.

Fig.6: Wave Structure Diagrams
Wave structure of guided wave mode represent the ensemble of particles displacement, stresses and the density flux energy distributions. The wave structure is important for the determination of interaction of Lamb waves with defect in a structure. From wave structure, we can determine the eventual interaction of mode with a specified defect.
Fig.7: Experimental Results - Open Corrosion - Hidden Corrosion
In the following only the most important results from three sets of experiments will be presented to demonstrate the ability of the proposed ultrasonic guided waves method to detect corrosion in different structures, Aluminum plate, lap splice joint and tear straps with preselected wave modes.
The first experiment demonstrates the feasibility of the guided wave excitation procedure, and determines the optimum conditions of excitability and propagation.
Fig.8: Open Corrosion

Detectability of corrosion was investigated in two aluminum specimens with two types of simulated corrosion. The first specimen was an aluminum plate with dimension 460x405x1 mm with controlled thinning in desig- nated areas . This first type of corrosion, is named open surface corrosion because corrosion is visible to the naked eye. To demonstrate the sensibility of the excited wave modes, corrosions were induced in three places with different level of thinning (10%, 15% and 25%). Measurements were made using the pitch-catch setup which consists of two variable angle broadband transducers with central frequencies at 3.5 MHz, one of the transducers acts as transmitter used to generate the guided wave mode and the other one used to re- ceive the generated mode and its interaction with the corroded structure. The transducers are driven by a tone-burst pulser/receiver system. The first set of tests demonstrates detectability of the open corrosion on the aluminum plate using the pitch- catch setup with piezo-composite transducers to generate the A₁ mode at 2.2 MHz with an incident angle of Figures 3b, 3c and 3d show the RF waveforms obtained with transducers positioned perpendicular to the corroded areas (three locations), while Figure 3a shows RF waveform obtained with transducers perpen- dicular to the noncorroded area.

Fig.9: Hidden Corrosion -Patch joint

Lamb modes are dispersive waves and their velocities are function of the frequency-thichness product. Therefore, any material changes such as wall thinning due to corrosion will affect the propagating mode ve- locity, amplitude and its time-of-flight. In the following test we take advantage of amplitude and time-of-flight changes to examine a second specimen (Figure 9) with simulated hidden corrosion (similar to a tear strap structure). It represents a 460x450x1 mm aluminum plate on which two corroded regions of different dimen- sion were induced. An aluminum patch of 150x150x1 mm was adhesively bonded over the corroded areas to simulate hidden corrosion.
A Lamb wave manual scan was carried out over the specimen illustrated in Figure 4 using a pitch-and-catch setup by moving a transducer pair along the specimen in X-direction. Waveforms were acquired using two variable angle probes at 30\260 for the S_o mode excited with 1.45 MHz frequency attached to a manual scanner which is controlled by a computer. Signals were acquired and stored for color imaging and analyses (with Tektrend's ARIUS software).
Figure 9 Box 1 and Box 2 show the time response in terms of a sinusoidal waveform (RF radio frequency signal) for the corroded and noncorroded areas. The corroded area between the second and the first alumi- num layer, creates a disbond and permits good transmission of the generated mode from the sender toward the receiver without any energy leakage in the additional bonded aluminum layer. In the noncorroded area, there is a good bond between the second and the first layer, therefore, the transmitted signal amplitude is at- tenuated due to leakage of the transmitted energy into the second layer. Boxes 3 and 4 show B-scan plot and its 3-D representation of the acquired signals from the Lamb wave scan. The B-scan plot is function of transducer displacement (X-direction), time-of-flight (Y-direction) and signal amplitudes (Z-direction). In these pictures, signals with higher amplitudes on the ends represent the starting point of our scan. Amplitudes of signals are high at these points since they are generated by the sender and received directly by the receiver without any interaction with the second plate. In the remaining region of the scan, green represents a poorly bonded area while red (or higher gray level) represents the corroded area. The sensitivity and efficiency of this inspection are demonstrated from the repeatability of Lamb wave C-scan in Box 1, where repeatedly three lines of scans were reproduced.
Fig.10: Hidden Corrosion - Tear strap

The inspection of bonded structures with pitch-catch setup is based on the following physical principles:
a guided Lamb wave mode once generated will travel from sender to receiver, producing relatively high amplitude RF signal when a disbond exists between the two bonded layers; otherwise it will leak in the tear strap if the bond is good, preventing the generated wave mode from being received by the receiver.
relative amplitude changes which occur in the transmitted wave mode through bonded structures are an indication for the existence of disbond, corrosion or even missed tear strap.
Inspection was carried over the specimen of Figure 7 in pitch-catch setup. The above Figure shows the time response in terms of a sinusoidal waveform (RF radio frequency signal). These waveforms are measured using the variable angle probe at 30° for the S0 mode at 1.48 MHz.
Fig.11: Hidden Corrosion - Lap splice joint
The above figure shows the second inspected specimen. It represent an aluminum lap splice joint with 580x330x1.5 mm dimensions attached to a stringer. This specimen was taken out from a Boeing 727 aircraft with history flight hours of 57,872 and flight cycles of 48,589. This specimen contain corrosion in its middle region which apparently was initiated by a large disbond between the overlapped parts.
Lamb wave scan was perform in the indicated direction. In the following figure, results from this inspection indicate the existence of corrosion between the layers. This was observed since amplitude of the transmitted mode was relatively low.
Again, these results were confirmed with those of eddy current C-scan . Results from this scan match perfectly Guided Lamb wave scan.
The propagation of Lamb guided waves in plates and multilayered structures has been presented.
The generation of pure modes and appropriate mode selection have been discussed.
We have also demonstrated the tools to properly control, predict and launch guided waves in aluminum.
Under different conditions, three sets of experimental tests utilizing So, S1 and A1 modes with different frequency-thickness product were performed.
We showed how the ultrasonic guided waves method can be used as a highly sensitive, fast and cost effective inspection technique for disbond and corrosion detection.
We have demonstrated that the ultrasonic guided wave method can be used as appropriate and promising inspection technique.
Lamb wave inspection can detect disbond in lap splice joints and tear straps in a single scan and the procedure is suitable for presentation of the results as an image.
Similar results from Lamb wave inspections on real Lap splice joints from Boeing 727 aircraft were obtained and will be reported in future paper. In addition, these test results will be used to better define the physical in- terpretation on how guided Lamb waves interact with second and multi-layered (shared) corrosion in the joint area.

This subject was also introduced during the ASNT fall Conference 96 in Seattle WA. See our Report.

Refer to another article:
Imaging of Disbond in Adhesive Joints with Lamb Waves V. MUSTAFA, A. CHAHBAZ, D. R. HAY, M. BRASSARD

| UTonline | |Top |