Detecting and Estimating Local Corrosion Damages in Long-Service Aircraft Structures by the Eddy Current Method with Double-Differential Probes

Corrosion damages (CDs) due to corrosion-induced metal loss represent one of the most important factors influencing the integrity, safety, and projected service life of aircraft structures (ASs) that have been in service for significant durations. The outer frame of many ASs consists of several aluminum alloy fastened sheets [1–3]. During an aircraft’s service life, the infiltration of moisture initiates CDs between these sheets. Most susceptible to CD are aluminum alloy components like double-layer joints and the skins of the fuselage and wings. For example, various types of CD, such as pitting, delamination, intergranular, crevice, contact, and corrosion cracking, have been detected in Antonov airplanes after long-term operation [2]. In aluminum alloy AS, pitting (local point damages, the diameter of which is smaller than their depth) and delamination type CD (swelling and delamination of the metal) are often found. Therefore, the well-timed detection and estimation of hidden CD is very important for ensuring accident-free aircraft operation and projecting the residual service life.

Therefore, monitoring AS corrosion by means of nondestructive testing (NDT) methods is an important part of a maintenance program for aging airplanes, and substantial efforts have been made by NDT researchers and engineers to find reliable ways to detect and quantitatively estimate such hidden corrosion [3–6]. Different types of CDs originating on the external parts of ASs can be detected visually [3]. To detect the external skin thinning concerned with uniform CD presence, the external skin thickness can be measured, for example, by ultrasonic NDT methods [3]. The most complicated situation for corrosion monitoring, however, occurs when internal hidden CD is situated on the internal surfaces of multilayer ASs (the bottom surface of the first/top layer or the top surface of the second layer). The eddy current (EC) method is one of the most promising techniques for detecting and estimating subsurface CD in ASs, as it makes non-contact and productive NDT inspection possible without the need to disassemble multilayer aircraft units [3–7]. This makes further improvement of the EC method based on the development of EC probes particularly sensitive to local CDs as well as advances in quantitative EC signal processing highly desirable.

This study first analyzes the state of the art in detecting CDs in ASs, then investigates the effectiveness of low-frequency double-differential EC probes in detecting hidden local CD detection for inspection carried out from the undamaged surface. It also considers correlations between the EC signal phase and depth of location of the hidden CD, as a potential method for the residual thickness estimation of the skin due to the local CD presence.

The influence of CDs on EC probe signals allows for corrosion damage in ASs to be categorized as either uniform or local. This categorization depends on the ratio of the damaged area’s size to the EC probe’s area of sensitivity, which determines the locality and the spatial resolution of the EC method [8]. When the corroded area (the area of corrosion-induced metal loss) is larger than the EC probe’s sensitivity area, CD can be evaluated as the total corrosion using an EC probe of the absolute type [9]. In this case, CD can be interpreted as the thinning of a metal sheet and evaluated by EC sheet thickness measurements. In study [10], for example, aircraft joints of the stringer-skin type were inspected by low-frequency EC flaw detectors supplied with an absolute-type EC probe. This EC technique was efficient only for the detection of CD more than 1 mm deep, when the damaged area was larger than the diameter of the EC probe. Therefore, the sensitivity of this technique to local CDs was insufficient. Furthermore, absolute-type EC probes have the disadvantage of high sensitivity to interfering factors such as changes in electrical conductivity, thickness of the dielectric coating, and changes in coating or clearance between the tested multilayer sheets, etc. For simulating such CDs, researchers typically use metal sheets of different thicknesses, fabricated from the same structural material as the AS to be investigated (aluminum or titanium alloy).

Local CDs are usually simulated in the form of dead-ended flat-bottom drillings, and usually detected by EC probes of the differential type. These probes are insensitive to changes in thickness, reacting only to local defects like cracks, inclusions, pores, local CD, etc. [7]. The approach of applying hybrid EC probes has also been proposed, with absolute and differential EC probes being combined in a common design. This promising technique has been used to obtain information about uniform and local CD originating in copper ship pipes [11].

The problem is most complicated when CD is hidden, occurring on the inner sides (surfaces) of a multilayer AS and moreover under a layer of dielectric paint coating. Typical examples of double-layer ASs with uniform CDs situated on the bottom surface of the first layer (a) or the top surface of the second layer (b) and local CDs, like pitting, situated on the bottom surface of the first (top) layer (c) are presented in Figure 1. As is evident, even for a relatively simple two-layer AS, the conditions for CD detection are deteriorated due to the weakening of the useful EC probe signal because of the skin effect. The EC probe signal is also distorted by a change in the dielectric coating thickness between the metal layers. Research efforts have focused on reducing the influence of these factors. A general analysis of the influence of the thickness and the clearance between the layers on the signal of absolute-type EC probes in the complex plane was carried out in the study [12].

Fig. 1.

The reliability of various EC techniques for CD detection based on the single-frequency, double-frequency, and impulse excitation modes was analyzed in study [13]. The reliability of determining the thickness of the upper skin by the single-frequency EC technique was found to be low, since a 10% skin thinning can be detected with a probability of less than 90%. The reliability of the double-frequency EC technique, in which the operational frequencies are chosen to obtain sufficient sensitivity to the thickness of the first and second layers and to minimize the influence of the gap, was not much higher than that of the single-frequency one. The prospect of the pulse EC technique was also examined – this approach requires the use of scanners, which slows down the inspection procedure.

Impulse excitation mode and a robotic scanner were used in an EC inspection system of the TRECSCAN type, in which an EC probe was designed with a Hall sensor instead of sensing coils as in conventional EC probes [14, 15]. To isolate a useful signal created by CD against the background of the interfering factors, the EC probe output signal was analyzed over time. In another approach, using an EC probe of the absolute type, the selectivity concerned with CD detection was ensured by analyzing the EC signal in the complex plane [16]. The double-frequency EC technique has been presented as effective tool for detecting CDs in the second layer even with a variable air gap between layers [17]. The pulsed EC technique with giant magneto-resistive sensors in conjunction with proposed signal processing was also presented as a prospective approach to subsurface CDs detection in study [18]. In study [19], in turn, the characteristic features in obtained EC data sensitive to CD and fatigue cracks in ASs and invariant to other noise factors were proposed and investigated. The presented results demonstrate a novel feature for CD characterization using first and second order derivatives of the impedance response.

The above publications all report on EC techniques for determining the level of uniform CDs based on the application of EC probes of the absolute type. The use of a low-frequency EC probe of the double-differential type to detect real local CD situated on the internal surfaces of stringer-skin type joints was investigated for the first time in our previous study [20]. The specimens were selected at a repair plant from among extensively exploited ASs produced by Tupolev Aircraft. A typical example of a selected specimen is shown in Figure 2a. The selected AS fragments with 1.5 mm thick skin and 3 mm thick stringer fabricated of D16AT aluminum alloy were evaluated at the operational frequency of 3.0 kHz. The depth and area of the detected CD were then determined by subsequent metallographic examination. Inspection was carried out both on the side of the skin and the side of the stringer without removing the coating. During the inspection, the EC probe was scanned along a dashed line (Fig. 2a) at a constant distance from the rivets. The depth of the minimal local CD detected on the bottom side of the stringer was 0.16 mm (5.2% of thickness) (Fig. 2b). The presented results demonstrated the high performance of low-frequency double-differential type EC probes to detect hidden CD originating on the bottom side of an inspected component.

Fig. 2.

The results so obtained were used to develop a step-by-step procedure for the underfloor part of fuselage inspection. The inspected AS was divided into 4 regions characterized by different probabilities of CD occurrence, estimated following selected statistical data. Initially, it was proposed to examine only the first region, which long-term in-service experience indicates is most often subject to CD. If no CD was detected in this region, EC inspection was not continued. If any CD was detected, the second region was inspected, and so on. This step-by-step procedure provides for improved inspection efficiency and productivity.

At the same time, our previous study did not investigate the detectability of low- frequency double differential EC probes concerned hidden local CD detection for the inspection carried out from an undamaged surface, nor did it propose a method for the residual thickness estimation of the skin due to the local CD presence.

EC techniques are widely being discussed as highly effective for non-contact detection and measurement of hidden CDs in AS without the need for disassembly. This study investigates and analyzes the advantages and limitations of double-differential EC probes in detecting CDs with different sizes and depths. The findings underscore the potential of the EC method for detecting hidden CDs in double-layer ASs.

Investigations into the relationship between EC signal characteristics (amplitude and phase) and depth of location (or residual thickness) of the subsurface local CDs revealed that the phase angle of the EC signal correlated with the depth of local CD, independently of CD diameter. In other words, the EC signal phase is a measure of local CD depth that is invariant of CD diameter changes. This means it is possible to estimate the residual thickness of the skin in the area of local CD (such as pits or pitting) by measuring the EC signal phase independently of the local CD diameter (size).

These findings pave the way for developing an innovative quantitative approach for the detection and sizing of subsurface local CDs arising in ASs during long-term service life. The demonstrated high detectability of the EC inspection technique offers useful and critical quantitative information for evaluating AS condition and forecasting its remaining service life. Moreover, quantitative information about CD depth can be used to more precisely define suitable inspection intervals.