Improvement of the Reliability of Fatigue Crack Detection on Holes of Typical Aircraft Structures by using Multi-Frequency EC-TechniqueDieter Schiller, Theodor Meier
Airbus Industrie, Germany, Bremen
Airbus Industrie, France, Toulouse
Corresponding Author Contact:
|TABLE OF CONTENTS|
Testing holes with rotating eddy current probes (single frequency technique) is a method widely used for the inspection of aircraft structures to detect small fatigue cracks. As small defects are to be found, the method must be very sensitive on the one hand, but on the other hand a large number of disturbing parameters occurring at in-service conditions can be easily interpreted as defects (false calls).
This lecture presents the results of rivet hole inspections on typical aircraft structures during a tear down inspection with the multifrequency eddy current technique. The focus is on the classification of readings and the evaluation of the crack length in the area up to 2 mm. The test equipment used allows a dynamic eddy current inspection with four simultaneously operating frequency channels and signal storage with high signal dynamics. Thus, analysis and evaluation of the signals is possible both on-line and after the inspection using the stored signals.
The low frequencies provide additional information for the determination of the crack length and are of advantage for the classification of signals and the detection of hidden defects. The results of the combined evaluation of signal parameters in the multifrequency inspection are compared to the results of the single frequency inspection. It is shown that the false call rate is considerably reduced due to the fact that a better distinction can be made between crack signals and disturbing signals.
An often used in-service inspection method on aircraft is the bore hole inspection with rotating eddy current probes. Here, mainly holes in aluminum alloys are inspected for cracks. The structures to be inspected may be single- or multi-layer. With this method it is also possible to inspect multi-layer aluminum structural areas with stiffening layers made from titanium and steel.
The advantage of this method is that the inspection can be carried out very quickly and that even small cracks can be detected.
The disadvantage of this method is that prior to inspection, fasteners (rivets) have to be removed.
As it is known for almost all NDT methods, in the case of bore hole inspection also, various parameters decisively influence the test result, especially during in-service inspections.
In-service experiences and various examples from practice show that it is often quite difficult to classify and evaluate a crack indication correctly. Additional evaluation steps such as evaluation of form of amplitude signal, evaluation of phase position of signal indication and other inspection steps with borescope or mechanical rework are often necessary.
Compared to the single-frequency method, the multi-frequency method described here makes an improved crack detection possible due to the use of an additional inspection parameter with several inspection frequencies. The false alarm is considerably reduced. The human factor is reduced and thus quality and operational efficiency of the inspection method is considerably increased. This has been achieved by developing an improved equipment electronics and by using corresponding equipment software. The increased analyzing algorithm is effected by the equipment and is not visible to the inspector so that no additional effort is generated for the human being.
A considerable improvement of results, i.e. covering of current findings rate and drastic reduction of false alarm, has been proved by an investigation of tear down results.
For a better general understanding of this report, the 'single-frequency' and 'multi-frequency' methods are to be called to mind and to be presented in the following. Furthermore, the real implementation of these two methods and a comparison of the results are to be dealt with.
Besides a short representation of the dynamic bore hole inspection with rotating eddy current probes operated in single-frequency technique, which inspection, today, is frequently used world-wide, this chapter is also to deal with the requirements for a reliable bore hole inspection, with the real existing inspection requirements on site and with the evaluation of signal indications.
Principle of dynamic bore hole inspection in single-frequency technique
Various types of inspection equipment from different equipment manufacturers are available world-wide for carrying out dynamic crack inspections in holes of typical aircraft structures.
|Fig 1: Bore hole inspection - state of the art|
The applied inspection method is based on the dynamic eddy current inspection and can be presumed as being generally known. Figure 1 shows in a simplified form the principle of operation. It is to be noted that this method is in principle a surface crack inspection with a high constant inspection frequency where the bore hole side is manually, carefully scanned with a rotating eddy current probe.
Requirements for a reliable dynamic bore hole inspection
Each NDT test method has its natural and real performance limits and should only be used and applied within these limits (performance spectrum). The same applies for the dynamic bore hole inspection and this is the reason why the conditions of use for the implementation and evaluation of a reliable bore hole inspection are to be dealt with in the following.
Specification of structure to be inspected
The spectrum of typical aircraft structures which are mainly to be inspected ranges from single-layer to multi-layer structural assemblies made from aluminum alloys. A multi-layer aluminum structure can also be fitted with stiffening layers made from materials with higher strength values than the aluminum structure to be inspected.
|Fig 2: Example of structure to be inspected|
Figure 2 shows a selection of structures to be inspected with possible, frequently occurring design for the dynamic bore hole inspection with rotating eddy current probes. The bore holes to be inspected must be cylindrical and may not be oval. The bore hole diameters may be between 3.5 and 25.0 mm. The conductivity of the aluminum structure to be inspected must be between 16.0 MS/m (27.6% IACS) and 28.0 MS/m (48.3 % IACS). The minimum layer thickness (of single layer) to be inspected must not be less than 1 mm.
Manual inspection with hand held rotor Single frequency Visual evaluation Fulfillment of aircraft specific standard requirements structure calibration standard probe crack rotor instrument signal representation
Defects to be detected
The bore hole inspection method with rotating eddy current probes has been developed to detect especially fatigue cracks that propagate in vertical direction to the structure layer surface and originate at the bore hole side. Figure 3 shows the defects to be detected with its orientation, position and dimension.
Fig 3: Detectable damage
Fig 4: Inspection conditions
Qualification programs with defined inspection procedures have proved that this inspection method will for sure not overlook defects with a depth and a length of more than 1 mm. This means that the reliability required world-wide in aeronautics for failure detection at a findings rate of more than 90% and with a confidence level of more than 95% has been reached. Verified test procedures in the Nondestructive Testing Manual (NTM), applied by sufficiently certified inspectors, comply with these requirements. The crack with the flat shape in the third layer can not be detected with the required reliability, because the crack depth of >= 1mm is not reached.
Fasteners of the bore hole to be inspected are to be removed prior to inspection. However, this big disadvantage is accepted in many cases because the detection of small defects, the often resulting advantage of long inspection intervals or a single inspection only is often much more important.
As already mentioned in 2.2.1, the bore hole to be inspected must of course be cylindrical and may not be oval. The inspection probe to be used must be chosen in such a way that the probe diameter is 0.1 mm ± 0.05 mm smaller that the bore hole diameter to be inspected.
Furthermore it is very important that the crack borders of the defect to be detected are open towards the bore hole side and that they are not closed or smeared over by the base material. The bore hole side must be even and clean for the inspection to be carried out.
Real inspection requirements on site
Bore hole inspections on structures of aluminum layers that have not yet been exposed to real in-service conditions, can be carried out with almost no problems and displayed indications can be evaluated mainly unambiguously.
This is not at all the case for inspections on cyclicly stressed in-service structures (especially with stiffening layers made from titanium or steel) during overhaul work on aircraft, structural modifications or the performance of repairs on site. In many cases, the bore hole conditions described in chapter 2.2.3 cannot be fulfilled under in-service conditions.
If the bore hole inspection with rotating eddy current probes is to be reliably used during in-service nevertheless, it is very important to consider possible inspection conditions even though they are not always determinable. Only if possible influencing parameters are considered, a reliable test result of high quality can be achieved. If possible influencing parameters are not considered, there is a risk of a reduced findings rate and especially a tendency towards a strong increase in the degree of false alarms.
|Fig 5: Disturbing parameters|
Besides the risk that, when selecting a too small probe diameter, cracks are left undetected due to the resulting lift off effect, the findings rate can also be reduced when cracks are closed or smeared over.
The removal of fasteners by drilling and pulling out generates, like possible rough bore hole sides or shim layers, an increased signal noise and therefore a more difficult signal interpretation.
Oval bore hole forms, pilot holes not completely removed as well as deposits of drilling chips or possible foreign particles (e.g. steel chips) within the inspection area can cause a misevaluation of the signal indication. Furthermore, bigger burr formations at the bore hole edge of structure layers as well as shim layers and panel butt joints within the bore hole area can cause misinterpretation.
Table 1 shows in a summarized form possible effects of given examples for disturbing influencing parameters.
|CAUSE||Possible effects on/due to|
|Findings rate||False alarm||Noise|
|Crack smeared over||x|
|Diameter of inspection||x||x probe too small|
|Rough bore hole side||x|
|Foreign particles, e.g. steel||x|
|Panel butt joints||x|
Fig 6: Single-frequency eddy current signals from cracks and disturbances
Fig 7: Evaluation steps
Evaluation criteria for the bore hole inspection with a constant frequency (single-frequency method) are the displayed signal level, the signal form and the phase position of the signal represented in the line and spot display mode. Figure 6 shows selected examples of typical evaluation displays for a real crack, for an oval form and the display of an interference indication caused by the formation of burrs. The supporting of signal evaluation by means of a visual reinspection with a magnifying glass and, if necessary, with a borescope and a possible mechanical treatment of the bore hole (e.g. burr removal) is imperative, as shown in figure 7 with an example.
Like in the previous chapter where the single-frequency method has been described, this chapter also gives a short representation of the bore hole inspection with the multi-frequency technique.
Since, for this basic inspection method itself and in case of manual performance with conventional commercial inspection probes which are still the same, the inspection requirements and conditions to be considered are also identical with those of the single-frequency technique.
Fig 8: Portable multi-frequency eddy current equipment
Fig 9: Manual rotor with integrated position encoder
Fig 10: Recording of the maximum form of the signals
Fig 11: Eddy Current Screen display
Fig 12: Windows based eddy current analysis software
Due to the fact that the four frequency channels work in parallel, the inspection is carried out within the same time as with the single-frequency technique. An analysis and evaluation of signal indications is carried out by the equipment by using the simultaneously recorded inspection data such as amplitude level, amplitude form and phase position, all dependent on the four inspection frequencies, and this analysis and evaluation is displayed to the inspector on-line in the current way. Figure 8 shows the inspection equipment performing a bore hole inspection on site.
The conventional hand rotator is equipped with an integrated position encoder which makes it possible to indicate also the corresponding penetration depth of the rotating probe into the bore hole, besides the indication of the circumferential position of the bore hole. This allows a two-dimensional representation of the bore hole side to be inspected and thus the exact determination of the location of possible defects, which can be documented. Figure 9 shows the hand rotator with the integrated position encoder.
A special storage option makes it possible to record the maximum signal form which can be obtained by optimally guiding the inspection probe. Figure 10 shows this optimal representation and documentation with an example of a double-sided crack with different crack lengths.
Besides the illustration of the signal in the line display mode, the inspected bore hole area is also displayed in a two-dimensional form. Areas not inspected and zones with defects can immediately be determined on-line.
Figure 11 shows an example of the representation of results on the equipment monitor.
Windows software specially adapted to the eddy current scan inspection makes it possible to optimally process the inspection signals without changing the basic data. It includes functions for a quick manipulation of the range of colors and special signal filters. By means of a pointer and a switchable impedance window, the position and form of the eddy current signals can be displayed. In addition, record sheets with test results can be generated in order to initialize special subsequent measures as soon as possible.. Figure 12 shows the equipment monitor with the representation of processed eddy current signals.
Within the scope of Airbus A330/340 tear down program, various structures have been inspected in detail with different NDT methods and subsequently with destructive fractography. Various bore hole inspections at door areas were also to be carried out on these structure parts to be inspected.
Especially in the corner areas of the doors, a great number of crack indications were documented with the bore hole inspection with the single-frequency technique. As all cracks were to be examined, we carried out a second bore hole inspection with the multi-frequency technique prior to these destructive inspections.
In the corner areas of the doors, a multi-layer aluminum structure with a reinforced steel layer, about 225 bore holes were inspected and in doing so, the inspection with the single-frequency technique resulted in evaluated crack indications at 46 bore holes and the inspection with the multi-frequency technique resulted in evaluated crack indications at only 15 bore holes.
Comparison of test results of single- and multi-frequency technique
After the bore hole inspections with both methods the structure was disassembled and a fractographic analysis was performed. The analysis of the signals of all frequency channels in the impedance plane gives much more information about the kind of the indication. Thereby it is possible to determine if the indication is caused by a crack or by a disturbance. Signals from cracks are nearly vertical with a small phase shift to the left at lower frequencies ( see figure 13). Ovality of the bore hole and aluminium burr result in signals with phase shift to the left on all channels and decreasing amplitude at lower frequencies. Another possible disturbance is the smearing effect, so that the crack is not open to the surface. The result from an earlier investigation [5, 6] is a good crack indication at the lower frequencies with decreasing amplitude and phase shift to the right for higher frequencies. A crack in a steel doubler (not critical) looks the same as ovality at the 400kHz channel but the phase shifts to the right at lower frequencies.
The most important indications were caused by ferrous particles pressed into the wall of the bore hole of aluminium layers. They gave the same signals as cracks at the 400kHz channel. Only the lower frequency signals with phase shift to the right and higher amplitudes allow the discrimination from crack indications. Figure 14 shows the results of the examination of the particle and the EC-signals. Most of the disturbances described above give EC signals similar to crack indications at the typical high inspection frequency. Significant signals of all these indications are shown in figure 15.
|Fig 13: 4- frequency eddy current signals from cracks and disturbances||Fig 14: Single-frequence technique, sample bore holes with fatigue crack and ferrous particle||Fig 15: 4-frequence technique, sample, bore hole with fatigue crack and ferrous particle|
|Number of inspected bore hole||225|
|Number of bore holes with crack indications|
- single-frequency technique
|Number of bore holes with cracks confirmed by fractographic analysis||15|
|Fig 16: Fuzzy classification|
|C1:Large crack||C2: Small crack||C3: Burr, Ovality||C4: Ferrous particle||C5: Smeared crack||C6: Crack in steel doubler|
|41 of 41||75 of 92||337 of 433||18 of 27||1 of 4||12 of 13||Total 610|
|Other classes||10 * C3|
7 * C1
|93 * C2|
3 * C1
|7 * C6|
1 * n.c.
|2 * C1|
|1 * C3|
|Fig 17: Calculated crack length versus crack length from fractographic analysis|
Investigations for the determination of the crack length from the eddy current signals were carried out on sample sheets with fatigue cracks using the multifrequency EC inspection system. The signals of all samples were recorded as 4-frequency circumferential scans simultaneously employing the inspection frequencies 400kHz, 133kHz, 67kHz and 22kHz. The X- and Y-components, amplitudes and phase angles of the signals of all frequency channels were evaluated, and the results were correlated with those gained from fractographic analysis using a multidimensional linear regression procedure. Most reliable results were obtained by a 4-dimensional evaluation of the Y-components. The investigations were made with scope of 52 samples (thickness 1.6 and 3.0 mm, diameter 4.9 and 6.4 mm) with 90 cracks up to 2.00mm length. The recording of the signals was performed 4 times by 3 different inspectors. Three different rotors with differences in the transducer bandwidth were used. The correlation coefficient between the Y-component of the 400kHz signal and the crack length is 0.73. The correlation coefficient obtained by the 4-dimensional correlation is 0.84 which means, that 95 % of the calculated crack lengths are within a range of ± 0.3mm from the real crack length. By special optimization a correlation coefficient of 0.90 is possible. The overall results are shown in figure 17. The evaluation process may be integr ated in the eddy current instrument to give the operator together with the classification result a reliable information about the indication.
The bore hole inspection with eddy current probes is today state-of-the-art and almost nobody talks about possible difficulties and deficiencies with regard to in-service implementation. Nevertheless, it can be assumed that the experienced inspector is for sure aware of bigger or smaller problems with regard to the implementation on site and knows of corresponding corrective actions.
We, as aircraft manufacturers, believe that it is quite informative to talk about our experiences on this topic within the scope of this conference, to describe necessary measures for confirmation of the inspection and to present new tendencies and possibilities of improving inspection methods.
As the report documents show, not only the reliability of the bore hole inspection is improved by the multi-frequency technique but also the quality of the inspection method is strongly increased. This reduces the costs considerably as, for example, due to the drastically reduced faults alarm, unnecessary repairs can be avoided. Information on the crack length of a detected defect, which information is often required for determining subsequent actions (repairs), is processed by the multi-frequency technique as an additional, important test result. Up to a crack length of 2.0 mm, the length can be reliably indicated.
Finally, it can be said that the bore hole inspection with the multi-frequency technique offers a much better inspection method for the future and makes it possible to considerably improve the reliability and quality of the bore hole inspection without additional, unnecessary efforts for the inspector. This is achieved by technical details such as documentation of results and signal evaluation, which considerably reduces the human factor (human element). The positive effect for the user is an increased operational efficiency and thus a considerable reduction in costs.