Insight Vol 48 No 3 March 2006 139

1. Introduction

The repair and rectification of corrosion damage continues to be a major factor in the life cycle costs of aircraft. The present approach

to corrosion control is based on ‘Find-and-fix’. Once an area of corrosion is found, rectification is carried out as soon as possible. This often involves the blending away of corrosion damage and the re-protection of the blended area. Where the depth of corrosion is beyond the permissible level (in the case of wing and fuselage skins this is typically 10% of the skin thickness) it is necessary to replace components. The ‘Find-and-fix’ approach can result in aircraft being taken out of service for long periods of time between regular maintenance schedules.

Attempts are being made to move away from the ‘Find-and-fix’ approach and to manage the corrosion so that it can be rectified during scheduled maintenance periods. Cole, Clark and Sharp(1) put forward the concept of an ‘Identify-and-manage’ approach to corrosion on military aircraft. Similarly, Peeler and Kinzie(2) have outlined the USAF shift towards an ‘Anticipate-and-manage’ approach to airframe corrosion control. The two approaches are essentially the same and are aimed at significantly reducing aircraft non-availability and at decreasing maintenance costs. The move from a ‘Find-and-fix’ approach requires a careful assessment of the effects of the corrosion on structural integrity of the aircraft. To achieve this, reliable models are needed which can be used to predict the long-term effects of corrosion on properties such as static strength and fatigue life. Significant progress has been made towards this goal through research being conducted in the US, Canada and Australia. Models allowing the effects of pitting and exfoliation corrosion on fatigue life and the role of corrosion on the degradation of lap joints are currently under development. Eventually these will allow the impact of different levels of corrosion on structural integrity to be assessed and will permit the time at which rectification must be carried out to be determined.

The ‘Identify-and-manage’ or ‘Anticipate-and-manage’ approaches will only be possible if effective non-destructive evaluation (NDE) methods are available to detect, characterise and monitor the corrosion that is present within the airframe. The aim of the present project is to assess the capabilities of a range of NDE techniques for detecting and quantifying corrosion present in aircraft structures. As part of this programme, the initial step has been to establish the corrosion metrics that need to be determined so that the ‘Identify-and-manage’ approach can be followed. The metrics will depend on the particular models that are being employed to estimate the residual life of the structure.

This paper presents the results of a review of the some of the models and tools that are currently under development and the basic corrosion data or corrosion metrics that are employed. The results from an initial feasibility study of pitting corrosion and crevice corrosion coupons are also presented.

2. Identification of corrosion metrics

A review of the literature has identified four areas where research is in progress to develop models for use in assessing the effects of corrosion on airframe structural integrity. These are:q Influence of corrosion on static strength.qInfluence of corrosion on fatigue life.

NDE IN THE AEROSPACE INDUSTRY

NDE corrosion metrics for life prediction of aircraft structures

E A Birt, L D Jones, L J Nelson and R A Smith

Recent research in NDE has led to the development of a number of techniques for detecting and measuring hidden corrosion, which could be used to underpin life extensions for aircraft structures. This work has reviewed the important corrosion metrics that need to be measured so that the remaining life can be calculated from corrosion and fatigue crack growth models. It has been found that the first corrosion metric which needs to be established is the type of corrosion taking place. In addition, experimental data in the review indicated that relatively small levels of corrosion can have a very detrimental effect on the fatigue life. A set of corrosion specimens has been fabricated to demonstrate the ability of various NDE techniques to measure the corrosion metrics. Two types of corrosion were investigated: pitting corrosion and crevice corrosion. The corrosion was produced by placing the samples in a salt fog chamber for different lengths of time. Ultrasonic, eddy current and X-ray inspection techniques have been undertaken on the corrosion samples. Both conventional (in-service) and advanced (state-of-the-art) NDE methods have been investigated. This paper will describe the initial results of the study.

On graduation, Anne Birt took up a position in the Non-Destructive Evaluation Sciences Branch at NASA Langley Research Centre. She was involved in the operation of a unique X-ray CT facility. In 1994 she returned to England and took up a position at QinetiQ. She is currently a Principal Scientist in the NDE group. She is a member of the British Institute of NDT and the Institute of Physics and is also currently a member of Council and the Technical Committee of the British Institute of NDT.

Lyn Jones joined the NDE group at the Defence Research Agency (now QinetiQ Ltd) in Farnborough in 1987. He is currently a Senior Scientist specialising in ultrasonic, eddy current and low-frequency vibration techniques. Lyn was heavily involved with the design and development of the ANDSCAN® portable scanning arm. Prior to this he spent three years at Marconi Underwater Systems concentrating on ultrasonic transducer development. Lyn is a member of the British Institute of NDT.

Dr Luke Nelson graduated in 2000 with a BEng in Materials Science from the University of Bath and then undertook a PhD researching piezoelectric materials. After completing a post-doctorate position working on transient thermography for NDT applications, he joined QinetiQ where he works in the non-destructive evaluation group.

Robert Smith is Capability Leader of the NDE Group of QinetiQ Ltd. He studied Physics at the University of Cambridge, and Applied Acoustics at Kings College, London, moving to the NDE Group at Farnborough in 1989. Author of over 85 publications, he was awarded the John Grimwade Medal in 1995, 2001 and 2004 and the Roy Sharpe Prize for 1996 by the British Institute of NDT where he is a Fellow. He is also a Fellow of the Institute of Physics, a Chartered Physicist, Scientist and Engineer.

Contact details: NDE Group, QinetiQ Ltd, A7 Bldg, Cody Technology Park, Farnborough, Hants, GU14 OLX, UK. Tel: +44(0) 1252 396206; Fax: +44(0) 1252 393053; E-mail: eabirt@QinetiQ.com

140 Insight Vol 48 No 3 March 2006

qCorrosion growth modelling.qLap joint corrosion.

The influence of corrosion on fatigue life has been modelled using two different approaches. The first of these is described by Koch(3). Using this approach it is proposed that corrosion sites in aluminium 2024 alloy sheet in the T3 condition (2024-T3) can be considered equivalent mechanical damage sites and that fracture mechanics can be applied using an Equivalent Initial Flaw Size, (EIFS) to compute the residual stress concentration and loss in residual strength.

The second approach was developed by researchers at the Defence Science and Technology Organisation (DSTO)(4,5). The

concept of the ‘process zone’ was introduced to account for the reduction in fatigue due to exfoliation corrosion. Laboratory-based testing has shown that the major effect of exfoliation on fatigue life was to cause a dramatic reduction in life with small corrosion depths. They proposed that the reduction in fatigue life was due to the presence of small pits rather than the more general stress concentration associated with the bulk of exfoliation. They indicated that the typical pit depth leading to fatigue failure is ~100 μm deep and at the base of the pits intergranular cracking extends to a depth of 10-30 μm. The combined pit/intergranular cracking regions was defined as the ‘process zones’, see Figure 1.

Table 1 details the main elements of some of the models described in the literature. It is clear from this table that the emphasis has been on sheet and plate 2024-T3 and 7075-T6 aluminium alloys. Further, the main type of corrosion considered has been pitting attack, although there appears to be growing interest in intergranular attack and exfoliation corrosion. Examples of the metrics employed are included in Table 1.

From Table 1 the main corrosion metrics may be summarised as follows:qInfluence of corrosion on static strength – depth of corrosion

attack, number of pits.qInfluence of corrosion on fatigue life – depth, width and

distribution of pits, surface roughness, average thickness loss and extent of damage.

qCorrosion growth modelling – initial pit depths.qLap joint corrosion – surface topography, average thickness

loss and extent of damage, volume of corrosion product present, pit depth and location.

Most of the reviewed work has focused on two areas: the interaction between corrosion and fatigue, in particular the effects of pitting corrosion on fatigue crack initiation and growth, and on the problem of corrosion within lap joints. What is clear from the review is that the first corrosion metric that needs to be established is the type of corrosion taking place. No single model is available at present that encompasses all types of corrosion; rather there are several models at various stages of development taking as their starting point a particular type of corrosion. For each model there are specific corrosion metrics, which may include depth of attack, number of pits, thickness loss etc.

From the experimental data presented in the literature it was clear that relatively small levels of corrosion can have a very detrimental effect on fatigue life. From an NDE perspective, the challenge is therefore to determine suitable corrosion metrics that can be reliably measured and are able to characterise small levels of corrosion damage. This is fundamental to the successful application of the models being developed.

In conclusion, the review has identified two principal areas where significant progress has been made towards the modelling of the effects of corrosion on structural integrity. These are:qthe influence of pitting corrosion on fatigue life;qthe failure of lap joints due to crevice corrosion.

Therefore, this work has decided to focus on these two areas and to investigate NDE techniques capable of measuring the required corrosion metrics.

3. Assessment of NDE techniques to measure the corrosion metrics

3.1 Introduction

A set of corrosion coupons were fabricated to demonstrate the ability of various NDE techniques to measure the corrosion metrics. Two types of corrosion were investigated: pitting corrosion and crevice corrosion (the type of corrosion found in lap joints). Using the corrosion coupons, a range of both conventional and

Figure 1. Process zone as defined by Sharp, Clark and co-workers(4, 5)

Model Type of corrosion

Metrics Materials

Influence of corrosion on static strengthCole et al(1) pitting Depth and number of pits 2024-T3 sheetMills et al(6) exfoliation Depth of exfoliation Not specified

Influence of corrosion on fatigue lifeEquivalent flaw sizeMultiple Crack Model (MCM) model, Perez(7)

pitting Depth and distance between pits

7075-T7451 (6.35 mm plate)2124-T851 (6.35 mm plate)

Thru-The-Thick-ness Crack Model, Perez(7)

pitting Depth of largest pit

Scheuring and Grandt(8)

pitting General thickness loss + EIFS (EIFS a-parameter typically 125 μm), surface roughness

2.2 mm sheet 2024-T3

Tuegel and Mills(9)

pitting Weight-loss (~100 μg.cm-2 considered significant)

7075-T6 sheet

Process zoneSharp et al(4) and Clark et al(5)

exfoliation Depth of pit + intergranular crackingpit 20-100 μm IG cracking 10-20 μm pit 20-100 μm IG cracking 10-30 μm

6 mm plate 2024-T351 6 mm plate 7075-T651

Corrosion growth modellingCorrosion PredictionModel,Altynova(10, 11)

pitting ‘As-Is’ corrosion condition, initial pit depths

Lap joint corrosionCorrosion Pillowing Stress, Komorowski et al(12, 13)

crevice corrosion

volume of corrosion product/ thickness loss (5% thickness loss considered significant)

Bellinger et al(14)

pitting Average thickness loss Clad 2024-T3

Mills et al(6) pitting Pit depth, surface topographyWhaley(15) Thickness loss, area affected

by corrosion Clad 2024-T3 sheet

Table 1. Summary of corrosion metrics identified from the literature

Insight Vol 48 No 3 March 2006 141

advanced NDE techniques were investigated. The NDE techniques investigated were:qConventional ultrasonic C-scanning.qFull-waveform capture high-frequency ultrasonic scanning.qConventional eddy current inspection.qTransient eddy current inspection.qX-radiography.

3.2 Pitting corrosion coupons

Initially, a set of five pitting corrosion test coupons were manufactured to assess the ability to produce accelerated corrosion damage. The coupons were rectangular plates manufactured from unclad 1.6 mm-thick aluminium 2024 alloy sheet in the T3 condition. In order to produce accelerated corrosion damage in the coupons they were placed in a 5% neutral salt fog chamber for different lengths of time. The intervals chosen were 1 day, 3 days, 1 week, 2 weeks and 3 weeks. The 1 week specimen was processed first as an initial test and, as a result, the decision was made to protect the top-left corner of future coupons to give a smooth reference surface. The finished coupons displayed a good range of corrosion damage from fine pits (1 day) to gross pitting and significant corrosion product on the surface (3 weeks), see Figure 2.

Initially, the 1 week exposure coupon was investigated with a range of ultrasonic transducers. In order to scan the samples in an immersion tank, each sample was taped around the edges to provide a watertight seal to a larger aluminium sheet, with the corroded side in contact with the larger aluminium sheet. This arrangement simulated the inspection of hidden corrosion damage, as the corroded surface was now at the back face and also prevented further corrosion of the corroded side due to contact with the water. Conventional ultrasonic inspection of this sample was undertaken using a 10 MHz focused transducer (38 mm focal length in water) using a 100 μm step size. The resulting amplitude C-scan is shown in Figure 3.

The coupon was then scanned using both a 22 MHz focused transducer (focal length 25 mm in water) and a 50 MHz focused transducer (focal length 13 mm in water) using a 100 μm step size for both scans. Full-waveform data were collected for these scans so that analysis of the waveforms could be undertaken at a later time. The resulting amplitude C-scans are shown in Figure 4. As can be seen, the higher frequency C-scans were considerably better at imaging the individual pits compared to the C-scan at 10 MHz. The 50 MHz transducer was then used to C-scan the other pitting corrosion coupons. Again, full-waveform data were collected. Amplitude and depth C-scans of the 3 week exposure coupon taken with the 50 MHz focused transducer are shown in Figure 5. Two relationships between the parameters collected from the ultrasonic scans and the exposure times of the coupons have been identified. These are shown in Figure 6.

As can be seen, ultrasonic techniques have the ability to image and characterise the corrosion damage. The important limitation of the ultrasonic inspection method is that in multi-layer structure, ultrasonic inspection will be limited to the inspection of the top layer due to the significant impedance mismatch between air and aluminium. If a structure is well-bonded with adhesive or sealant it may be possible to inspect a second layer but further work to determine the limitations of the technique in this situation would be needed.

Eddy current and X-radiographic techniques have also been investigated which are more suitable to the inspection of multi-layer structure. Conventional eddy current data was found to be insensitive to the pitting and material loss in these test coupons. However, the transient eddy current data from TRECSCAN was found to show a trend of increasing material loss with increasing exposure time. Further work on a larger set of samples will be needed to confirm this relationship.

Low-kV X-radiography was also carried out on the pitting coupons. The plates were radiographed at 20 kV with Agfa D4 film

(8 mA beam current, 15 minute exposure). Low-kV X-radiography was found to produce good imaging of the individual pits, see Figure 7. A second 1.6 mm-thick aluminium sheet was placed on top of the corroded coupon to simulate a double-layer structure. This double layer was radiographed at 25 kV with Agfa D4 film (8 mA beam current, 19 minute exposure), see Figure 8. As can be seen, there is a noticeable loss of contrast for imaging the pits through an extra layer and at the higher kV. The same 1 cm2 area of each radiograph was investigated in detail. For the single layer at 20 kV, 35 individual pits were visible in this 1 cm2 area, however, from the double-thickness radiograph taken at 25 kV only 20 pits were visible. Therefore, the ability to characterise corrosion damage located in a thicker structure will be reduced. For X-radiography, the amount of material loss that can be detected should be dependent only on the thickness of the structure to be penetrated and not on the number of layers present. Generally, when parameters (film type, X-ray energy and exposure time) are optimised, X-radiography can detect a 2% thickness loss in a structure. Therefore, for a 1.6 mm-thick specimen, a 32 μm thickness loss should be detectable.

3.3 Crevice corrosion coupons

The crevice corrosion coupons aimed to simulate the type of corrosion that can develop in aircraft lap-joints. Two test coupons were prepared, consisting of pairs of panels cut from 1.6 mm-thick unclad 2024-T3 aluminium alloy sheet. The panels were protected on one side using a corrosion-resistant tape and sandwiched together so that the unprotected faces were in contact, either directly or separated by filter paper soaked in 5% sodium chloride solution. Each pair of panels was held together by corrosion-resistant tape. The pairs of panels were initially exposed to 5% neutral salt fog for 7 days. Examination of the unprotected panel faces revealed evidence of crevice corrosion, with the degree of attack being highest for the sample made by placing the panels in direct contact, see Figure 9. The samples were then replaced in the salt fog chamber for a further 14 days exposure.

Ultrasonic inspection of the crevice corrosion samples was undertaken after the 21 days exposure. Again, in order to scan the samples in an immersion tank, each sample was taped around the edges to provide a water tight seal to a larger aluminium sheet, with the corroded side in contact with the larger aluminium sheet. The samples were then C-scanned with the corrosion at the back surface of the sample. The samples were scanned using the same 50 MHz transducer as before, with a 100 μm step size. Ultrasonic amplitude and depth C-scans are shown in Figure 10. The corrosion damage is clearly of a different nature to the corrosion in the pitting samples (see Figure 5). In these crevice corrosion samples, areas of distributed material loss can be seen compared to the more localised pitting seen before.

4. Conclusions

Future corrosion control management on aircraft may in the future move from a ‘Find-and-fix’ approach to an ‘Identify-and-manage’ approach. Central to this change will be the development of suitable models that can be used to predict the effects of corrosion on structural integrity. From the current review of the literature, it is apparent that most of the research effort to understand the effects of corrosion on structural integrity has been concentrated on two areas: the impact of pitting corrosion on fatigue life and the failure of lap joints due to corrosion.

Corrosion metrics such as the type of corrosion, distribution and loss of thickness, distribution and material loss from pits, and surface roughness can be identified, which are required as inputs to the various models available. It is clear from experimental data and from the models that very low levels of corrosion can have a serious impact on the fatigue life of aluminium alloys. It is concluded that it will be necessary to develop and employ NDE methods capable

142 Insight Vol 48 No 3 March 2006

of detecting and characterising low levels of corrosion damage if the ‘Identify-and-manage’ approach is to be adopted.

A set of test coupons with pitting and crevice corrosion were fabricated to demonstrate the ability to produce accelerated corrosion damage. Initial results to characterise the corrosion in these coupons after inspection with both conventional and advanced NDE techniques have been presented. For corrosion in the top layer, ultrasonic techniques have been found to be a good method to characterise the corrosion damage. For corrosion deeper in the structure, both transient eddy current and X-radiographic inspection methods have been identified as having potential for corrosion characterisation. Further work is now underway on a larger set of pitting and crevice corrosion dog-bone samples, with the aim to relate corrosion metrics to fatigue life.

Acknowledgements

The authors acknowledge the contribution of Dr C J E Smith of the Corrosion group at QinetiQ for his help and advice with the literature review. This work has been carried out with the support of the Weapons, Platforms and Effectors Domain of the UK MOD Research Programme.

Figure 7. X-radiograph of 2 week pitting corrosion coupon taken at 20 kV (x ~4)

Figure 8. X-radiographs of 2 week pitting corrosion coupon at 20 kV (left) and the same coupon with an extra 1.6 mm Al sheet placed on top to simulate a double layer structure taken at 25 kV (right)

Figure 5. Ultrasonic amplitude (top) and depth (bottom) C-scans of 3 week pitting corrosion coupon using a 50 MHz focused transducer. The uncorroded corner of the plate can be seen on the left side of the images

Figure 6. Relationships obtained between measured ultrasonic data and exposure time for the five test coupons. (a) Comparison of the standard deviation in the reflection coefficient versus exposure time –from ultrasonic amplitude scans at 50 MHz, (b) Comparison of the maximum measured pit depth and exposure time – from ultrasonic depth scans at 50 MHz

Figure 2. Al 2024-T3 test coupons after exposure in a salt fog chamber

Figure 3. Ultrasonic amplitude C-scan of 1 week pitting corrosion coupon using a 10 MHz focused transducer

Figure 4. Ultrasonic amplitude C-scans of 1 week pitting corrosion coupon using a 22 MHz focused transducer (top) and a 50 MHz focused transducer (bottom)

Insight Vol 48 No 3 March 2006 143

References

1. G K Cole, G Clark and P K Sharp, ‘The implications of corrosion with respect to aircraft structural integrity’, DSTO-RR-0102, 1997.

2. D T Peeler and R C Kinzie, ‘Corrosion damage management and future outlook on corrosion prediction tools’, NASA/FAA/DoD 5th Aging Aircraft Conference, Kissimmee, Florida, September 2001.

3. G H Koch, ‘On the mechanisms of interaction between corrosion and fatigue cracking in aircraft aluminium alloys’, AD-Vol 47, Structural Integrity in Aging Aircraft, ASME, pp 159-169, 1995.

4. K Sharp, T Mills, S Russo, G Clark and Q Lui, ‘Effects of exfoliation corrosion on the fatigue life of two high-strength aluminium alloys’, DoD/FAA/NASA 4th Aging Aircraft Conference, St Louis, Missouri, May 2000.

5. G Clark, P K Sharp and T B Mills, ‘Modelling of fatigue crack growth from exfoliation and pitting corrosion,’ International Committee on Aeronautical Fatigue, ‘Design for durability in the digital age’, Toulouse, France, Vol 1, pp 485-498, 2001.

6. T B Mills, K Honeycutt, C Brooks, I Hammad and D T Peeler, ‘Managing damage in the wing: modelling the interaction of exfoliation with static and fatigue loads’, FAA/DoD/NASA 6th Aging Aircraft Conference, San Francisco, California, September 2002.

7. R Perez, ‘Corrosion/fatigue metrics’, ICAF 97, Fatigue in new and ageing aircraft, Proceedings of the 19th Symposium of the International Committee on Aeronautical Fatigue, 18-20 June 1997, Edinburgh, Scotland, Editors R Cook and P Poole, published by Engineering Materials Advisory Services Ltd, Cradley Heath, West Midlands UK, pp 215-229, 1997.

8. N Scheuring and A F Grandt, ‘Quantification of corrosion damage utilising a fracture mechanics based methodology’, DoD/FAA/NASA 3rd Aging Aircraft Conference, Albuquerque, New Mexico, September 1999.

9. E J Tuegel and T B Mills, ‘Correlation of holistic structural assessment method with corrosion-fatigue experiments’, FAA/DoD/NASA 6th Aging Aircraft Conference, San Francisco, California, September 2002.

10. M M Altynova, R G Kelly, J R Scully, X Zheng, G R Cooke and D T Peeler, ‘Lap joint corrosion prediction model’ DoD/FAA/NASA 4th Aging Aircraft Conference, St Louis, Missouri, May 2000.

11. M M Altynova, R G Kelly, J R Scully and D T Peeler, ‘Engineering corrosion prediction model for aircraft structures’, 6th Joint FAA/DoD/NASA Aging Aircraft Conference, San Francisco, California, 2002.

12. J P Komorowski, N C Bellinger and R W Gould, ‘Local stress effects of corrosion in lap splices’, NATO RTO Meeting Proceedings 18, ‘Fatigue in the presence of corrosion’, Corfu, Greece, 7–9 October 1998.

13. J P Komorowski, N C Bellinger and R W Gould, ‘The role of corrosion pillowing in NDI and in the structural integrity of fuselage lap joints’, ‘Fatigue in new and ageing aircraft’, Proceedings of the 19th symposium of the international committee on aeronautical fatigue, Edinburgh, UK, pp 251-266, 1997.

14. N C Bellinger, D S Forsyth and J P Komorowski, ‘Damage characterisation of corroded 2024-T3 fuselage lap joints’, NASA/FAA/DoD 5th Aging Aircraft Conference, Kissimmee, Florida, September 2001.

15. P W Waley, ‘Corrosion damage tolerance methodology for C/KC-135 fuselage structure’ NASA/FAA/DoD 5th Aging Aircraft Conference, Kissimmee, Florida, September 2001.

© QinetiQ Ltd 2005.

(a) (b)Figure 9. Photographs of crevice corrosion plates after 1 week in a salt spray cabinet, (a) no filter paper, (b) filter paper covering whole panel

Figure 10. Ultrasonic amplitude (top) and depth (bottom) C-scans of crevice corrosion plates after 3 weeks exposure. Left side: no filter paper between plates and right side: filter paper between plates during exposure in the salt fog chamber

AEROSPACE NDT Symposium26-27 April 2006

BAWA Conference Centre, Filton, Bristol

Tickets:Single day – £70/Both days – £110

For information contactBINDT Conference Department

1 Spencer Parade, Northampton NN1 5AA, UK. Tel: +44 (0)1604 630124; Fax: +44 (0)1604 231489; E-mail: conf@bindt.org