Eddy Current Technique In Non Destructive Testing

Published Date: 02 Nov 2017

Critical Review of 3 Academic Journals

Solomon Oluwatobi Obadimu

Abstract

This report carries out a critical review of three different journals. An introduction of Non-Destructive Testing and Eddy Current method are provided in the first section. Literature reviews of three academic journals with the same NDT technique are carried out in the second section. Then in the third section, the journals are critically analysed by referring to other sources as well. Finally, the practical applications of the eddy current technique in aviation industry are justified.

Table of Contents

List of Figures

Figure 2.1 Rivet lap Joint ………………………………………………………………….5

Figure 2.2 Filter to suppress edge effect…………………………………………………..6

Figure 2.3 Influence of edge effect ……………………………………….........................6

Figure 2.4 Probe Design …………………………………………………………………..7

Figure 2.5 F/A-18 inner wing Spar ………………………………………………………..8

Figure 2.6 Fastener layout …………………………………………....................................8

Notations

CS Controlled Signal

NDT Non-Destructive Testing

PCA Principal Component Analysis

PEC Pulsed Eddy Current

Introduction

Eddy Current Technique

NDT has been in existence for a couple of decades now; it was used to detect cracks in wheels and axles of railcars, NDT is used in measuring and inspecting materials without causing any damage (NDT resource centre, 2012).There are several types of NDT methods, but Eddy Current Testing technique, principles and its applications will be discussed in this report. According to ( ,2013), Eddy currents "are induced electrical currents that flow in a circular path. An alternating current is passed through a coil and this results in a magnetic field being generated in the component. If another electrical conductor is brought into the close proximity to this changing the magnetic field, a current will be induced in this second conductor. Any flaw will disrupt the flow of the eddy currents in the second conductor and be sensed by an impedance change".

Eddy Current is used in the aviation industry to inspect conducting components of an aircraft such as aluminium (Horan et al, 2013). To inspect bolt-hole in the conventional eddy current technique, the fasteners must be removed, due to the limited depth of penetration (Lemire et al., 2010). However, removal of fasteners could disassemble and cause damage to the structure hence Pulsed Eddy Current method was developed. Pulsed Eddy current testing is new method mainly used to detect defects in multi-layered structures (Moulder et al., 1996). A frequency field excitation is used in the conventional eddy current method while pulsed excitation characterised mainly by the frequency contents are used in PEC method, (Tian et al., 2005). In PEC, zero-crossing time and the peak amplitude are used in characterising and detecting defects (Bieber et al., 1997) because the output signal contain defects (Sophian et al., 2003). Increase in defect volume lead to the increase of the peak amplitude of output signals and short zero-crossing time is caused by (deep) defect depth; with regards to this principle, the response signals of PEC in time domain are set to detect flaws in riveted structures (He et al., 2009).

Principal Component Analysis: in this statistical technique, data with large numbers are re-oriented into smaller dimensions in such a way that its original information is not lost and thus remains relevant (Lattin et al., 2003). PCA method was applied in the research of Sophian et al. (2003) to detect cracks in multi-layered structure and when it was compared to time-domain analysis, they discovered that the PCA method provided more reliable flaw detection. They reduced data with large numbers into single points in order to compare defects.

The main purpose of this report is to carry out a critical review of three different journals. Firstly, a literature review of three journals with the same NDT technique will be carried out in the second section. Secondly, the journals will be critically analysed by referring to other sources as well. Thirdly, the practical application of the eddy current technique in aviation industry will be justified. Finally, conclusions will be drawn out from the review.

Literature Review

Controlled Signal-Pulsed EC to inspect cracks on Multi-layered Aluminium Structure of an Aircraft

Bischoff et al. (1998) investigated the application of Eddy Current Non-Destructive Method to detect cracks on the joints in the aluminium alloy of the aircraft structure. Controlled Signal-Pulsed eddy current was used for the inspection of the rivets joints of the aluminium structure of the aircraft. The experiment also demonstrated the performance of the conventional sensor and the rotating field sensor in eddy current testing (technique).

As shown in the figure 1, the inspected specimen was a rivet lap joint in an airplane. The distance between the rivets ranged from 3 mm-18 mm, the cracks on the specimen was up to 1.44 mm in depth. The testing system of CS-pulsed eddy current was divided into three main categories: measurement device, Computer software and sensors. The computer software monitored all the measuring procedures when the CS-pulse was generated. Once the measurement was taken by the measuring device, the signal (analogue) was transferred to the sensor hence, the reading was recorded

. C:\Users\sollysoo\Desktop\NDT eassy\NDT Essay\Capture6.PNG

Figure 2.1 Rivet lap Joint (Bischoff et al., 1998)

The first experiment conducted with a conventional sensor revealed that the CS-pulse had a frequency range of 1 KHz-2 KHz. Linear grey and threshold scaling were used during the experiment to detect defects on the specimen. In addition, higher amplitudes were recorded on the edges of the rivets compared with the areas within the rivets. The lengths of the cracks between the rivets were indicated with grey (semi-dark) colour as shown in figure 2. The experiment concluded that there was a crack in rivet 3 going in the direction of rivet 4. Rivet 6, 9, and 12 had cracks that were 6 mm, 9 mm and 12 mm in length on the specimen. Rivets 15 and 16 on the other hand had gaps very close to each other due to the length of the cracks. Continuous cracks were also observed between the 18th and 19th rivets.

The second experiment was conducted with a rotating field sensor. The frequency range of the CS-pulses was within the range 1 KHz-3 KHz just like the first experiment. In this experiment, the receiving and sending time of the pulses were taken into consideration by adjusting the frequency. The entire time was 26 metre per second (ms) for a 360° rotation. When a sensor is positioned directly above a crack, a vertical eight ‘8’ figure will appear indicating there is a crack. The ‘C’ sensor in figure 3 received a couple of pulses, however, because of the field sensor (rotating), the rivets showed a low amplitude. The sensors T, L, R and B on the other hand were placed 2 mm from the centre of the rivets and thus higher amplitudes were recorded. There was higher amplitude in sensor ‘B’ because it was positioned close to the edge of the rivet. The edge influence of the rivet in crack detection was solved by filtering. As shown in figure 2, the position of the sensors clearly affected the results before it was filtered and thus the presence of a crack was detected. However, after it was filtered, no crack was detected. According to the second part of figure 2, the sensor on the rivet detected a crack before and after it was filtered and thus exhibited a vertical ‘8’ figure.

Bischoff et al. (1998) concluded that the combination of the conventional sensor and the rotating field sensor in Controlled Signal-Pulsed eddy current technique is a good approach to detecting (bending) cracks on multifaceted aluminium. It was further concluded that the CS-eddy current technique to detect cracks on joints in the aluminium alloy of the aircraft structure had very promising results.

C:\Users\sollysoo\Desktop\NDT eassy\NDT Essay\Capture8.PNG

Figure 2.2 Filter to suppress edge effect (Bischoff et al., 1998)

C:\Users\sollysoo\Desktop\NDT eassy\NDT Essay\Capture9.PNG

Figure 2.3 Influence of edge effect (Bischoff et al., 1998)

Pulsed EC for the detection of cracks in F/A-18 inner wing spar

Horan et al. (2013) investigated the application of pulsed eddy current to detect cracks in F/A-18 inner wing without removing the wing skin. As shown in fig 1, a probe design with a round driver coil encircled by a 9.6 mm diameter ferrite and two pickup coils (diameter of 1mm) was used for measuring. In addition, the Principal Component Analysis (PCA) was applied to reduce the sum (residual) of squares and get Eigen function scores that were used to detect cracks on the fastener holes. The coils were turned 400 times to get a 44 gage wire. The aim of the design was to generate eddy currents by using the fastener (ferrous) to expedite the transfer of fluids (magnetic) to the conducting region (aluminium spar) via the (skin layer) non-conducting region (Desjardins et al., 2012).

C:\Users\sollysoo\Desktop\NDT eassy\NDT Essay\Capture1.PNG

Figure 2.4 Probe Design (Horan et al., 2013)

The specimen (wing span) used was made from a 7075-T6 aluminium with fasteners placed 25 mm to hold the connected spar as shown in fig 2. In addition, the wing spar had several notches (0.2 mm width) which was extended from the centre over the head of the fasteners. The Pulsed Eddy Current (PEC) probe was placed inside the alignment block in order to align the pickup coil with the fasteners then the reading/measurement was taken. In total, readings were recorded from ten fasteners; five fasteners 17-21 with cracks and fasteners 13, 14, 15, 16 and 22 without cracks. Because of the presence of probe, Twenty one (21) readings/measurements were recorded for each fastener in other to acquire a statistical representation of the reading. The alignment block and probe were removed to take each reading and then replaced back for a new reading. The computer was configured to take readings every 6 seconds during the experiment.

The result of the experiment was divided into some categories: Firstly, due to the sensitive nature of the probe, 21 readings/measurements were carried out on fasteners without notch at their bore holes and fasters with notch originating from their bore holes (e.g. fasteners #16 and #17). The probe was replaced and then aligned after each reading/measurement. The method resulted in major variation in fastener #16 because the frequencies between the signals (such as peak amplitude) were not easily identified. Responses of fastener #16 and #17 were compared and it was concluded that such features should not be used for cracks detection. Further experiments were conducted to reveal the main source of variation; it was revealed that misalignment along the link connecting the coil (pickup). Secondly, the PCA was used to separate variations caused by misalignment and the presence of notches on some of the fasteners. For each of the 10 fasteners, 21 readings/measurements were taken which contained 200 data points that were gathered into a single matrix to obtain the PCA scores. Once the scores were obtained, the signals were separated from each other by plotting the scores and observing gathering of the data. Finally, the results/data obtained were verified by setting an alarm to go off whenever the data was invalid or due to improper alignment. However, a second alarm was set to go off if there was a crack. Horan et al. (2013) concluded that with the above results, the inspection of several fasteners in an aircraft is possible especially with the availability of a method to check the validity of data/ results obtained.

C:\Users\sollysoo\Desktop\NDT eassy\NDT Essay\Capture4.PNG

Figure 2.5 F/A-18 inner wing Spar (Horan et al., 2013)

C:\Users\sollysoo\Desktop\NDT eassy\NDT Essay\Capture2.PNG

Figure 2.6 Fastener layout (Horan et al., 2013)

Pulsed EC method for detecting defects in aircraft riveted structures

In the work carried out by He et al. (2009), the application of Pulsed Eddy Current (PEC) method (technique) to detect defect in riveted structures of an airframe was investigated. Three types of probes were designed to detect defect. The first was the differential hall probe: According to Tian et al. (2005a and b, 2009), studies, Hall Effect sensor is applicable in measuring magnetic field signals. The differential hall probe used consisted of two sensors (hall-effect) UGN3503 and an existing coil. The two sensors were placed symmetrically beside the exciting coils and connected in a differential manner. Eddy current was generated in the experiment by exciting the coils with repetitive broadband pulse. A sensor was placed on a defect-free material to generate an inherent signal (voltage) while the other sensor was placed on an area with defect to output changed voltage signal as a result of the defect.

The second probe designed was the differential coil probe: This consisted of two pick-up coils and an exciting coil. The exciting coil in the differential hall probe had the same shape and parameters with that of the differential coil probe. The pick-up coils were placed symmetrically beside the exciting coils and connected in a differential manner. A pick-up coil was placed on a defect-free material to generate an inherent signal (voltage) while the other coil was placed on an area with defect to output changed voltage signal as a result of the defect. Thus there was change in the differential output signal of the coil probe.

The third probe designed was a two-staged differential coil probe: It consisted of three reference coils. Deformation of the defect signal caused by the lift-off effect can lead to the misinterpretation of the Pulsed Eddy Current signals (Li et al., 2007). A defect signal and a defect-free signal were traditionally used in working out the differential signal. The probe was optimised in order to suppress the lift-off effect. The first stage was aimed at reducing the lift-off effect while the reference signal was obtained by subtracting the corresponding reference signal in the second stage.

The experimental (PEC) system used consisted of the specimen, pulse generator, data acquisition module, probe, power amplifier, signal process and software. To simulate the riveted (multi-layer) structure of an airframe, two plates were fixed together by a 3 mm diameter screw and rivet. The results of the experiment were categorised according to the different types of probes used. In differential hall probe, it was discovered that the hall response signals were noise that resulted into the inaccuracy of the results i.e. in detecting defects. After data processing, the peak values of crests rose when there was an increase in depth of defects without a change in its length and width. Hence it was considered an effective method of detecting defects in structures that are riveted. In the differential coil probe, it was discovered that the depth of defect increased when the peak value of the amplitude increased thus the time (zero-crossing) of response signal and peak amplitude were extracted to get the depths of defects in the riveted structures. The two-staged differential probe was used to detect deeper detect in an aluminium structure then it was concluded that the two-staged differential probe is more effective in detecting deep defects in riveted aluminium structure.

Critical Analysis (Evaluation)

3.1 Bischoff et al. (1998) investigated the application of EC Method to detect cracks on the joints in the aluminium alloy of the aircraft structure. Although the title was clearly stated, no abstract was included at the beginning of the article thus readers cannot ascertain purpose of the research. An abstract help readers quickly discover the purpose of a paper/article (Blake and Bly, 1993). The purpose of the experiment was followed up with an experiment, however, the experimental setup/method was not well detailed rather it was summarized. Compared to the research conducted by He et al. (2009) and Kreutzbruck et al. (1998); a more detailed explanation of their experimental method was given. It was stated in the experiment that inverse filter was carried out to suppress edge effect however, how the filtering was carried out was not explained.

Two types of experiments were carried out in their research; this enabled them to compare the results of the two experiments and hence justifying their results. Furthermore, the researchers concluded that the combination of the two provided a more reliable result. He et al. (2009) also combined three experimental results and discovered that the two-staged coil probe was more reliable. Although a total of eight references were listed in their experiment, however only six were cited in the article. Apparently, the main goal of the experiment was achieved; which according to the researchers seemed promising. Future works however was not suggested. In the research conducted by Krause et al. (1998), further research works to improve their research were suggested and thus will enable new researchers build up on their work.

3.2 Horan et al. (2013) carried out a research to apply PEC in detecting cracks in FA-18 inner wing span. The wing spar thickness is not same in all case; this was supported by the author in the introduction. However, in the experiment there was no effort to find the parameter changes in eddy current in case there is change in the thickness of the skin. More clearly the thickness of the skin is said to vary for different aircrafts and the authors gave a range of 9.1- 21.3 mm. In the experiment a polymer of 13 mm thickness was used to simulate the skin. So it can be argued that results of this experiment cannot be used to validate the applicability of this method on an aircraft with wing spar thickness of more than 13mm. In addition, the reason for using a polymer of 13 mm was not stated.

Furthermore, Nylon 6 polymer was used in the experiment to simulate the non-conducting composite of the wing skin. Different materials have different permeability for magnetic flux to pass through that is, there will be low permeability if a component contains a lot of carbon contents and thus retains more magnetic flux (NDT resource centre, 2012). Having said that, one would certainly seek the explanation of using a different material for the experiment and this was not found in the paper.

The researchers stated that the notches had various orientations and lengths however; figure 2.6 clearly contradicts the claims of the author. The notches used in the experiment were parallel to the bolt arrangement; the reason for this was not explained and thus could confuse a reader. If it was of uniform orientation as shown in figure 2.6, then the repeatability of the method with cracks (notches) in a different orientation can be questioned. Contrarily, if the notches were in various orientations as claimed in the article, then this question has no value. The fact that this is not clear leaves the paper/article with some ambiguity.

3.3 He et al. (2009) conducted a research using PEC technique to detect defects in aircraft riveted structures. The purpose of their experiment was justified; step by step analysis of method used was explained. In the article, three different probes were designed in order to compare their results; then they came to a conclusion that the two staged probes was more effective in detecting cracks in multi-layered structure.

The apparatus used in experiment were well stated and explained however; their mode of connection was not explained. This gives the reader a bit of confusion as to how the apparatus were connected. An aluminium was designed during the experiment to verify the performance of the differential coil probes, however, the basis in which it was designed was not pointed out Optimisation of the probe was stated in the report, the authors backed it up by explaining how the optimisation was carried out in their experiment.

Application of Eddy Current in the Aerospace Industry

Eddy current technique was used in testing fuselage parts although there were some draw backs,