- Email: [email protected]
Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch
Engineering Fracture Mechanics xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Engineering Fracture Mechanics journal homepage: www.elsevier.com/locate/engfracmech
Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch A. Albedah a, Sohail M.A. Khan a, F. Benyahia a, B. Bachir Bouiadjra b,⇑ a b
Mechanical Engineering Department, College of Engineering, King Saud University, Riyadh, Saudi Arabia LMPM, Department of Mechanical Engineering, University of Sidi Bel Abbes, BP 89 Cité Ben M’hidi, Sidi Bel Abbes 22000, Algeria
a r t i c l e
i n f o
Article history: Received 28 October 2014 Received in revised form 7 April 2015 Accepted 8 April 2015 Available online xxxx Keywords: Bonded composite repair Fatigue life Constant amplitude loading Variable amplitude loading Adhesive disbond Stress ratio
a b s t r a c t In this study, we analyzed the fatigue behavior of a V-notch crack in an aluminum alloy 7075-T6 plate repaired with bonded composite patch under constant and stepped variable amplitude loading. The effect of adhesive disbond on the repair performance was analyzed. The obtained results confirm the need of repairing cracks at their early creation for a maximum repair performance and showed that the increases in the disbond width and the stress ratio have a large negative effect on the repair efficiency. The latter is low for increasing loading blocks but significantly high for decreasing blocks. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Bonded composite repair of aircraft structures was developed by Dr. Alan Baker and his team in 1972. The technology involves adhesively bonding patches of advanced fiber composite materials to repair damaged aircraft structures and to stop stress corrosion cracking. The repairs are structurally very efficient, can be applied rapidly and are cost effective. The technology has many advantages over traditional mechanical repair methods, such as bolting or riveting. Composite patches are lighter, offer more uniform load transfer, seal interfaces reducing corrosion and leakage, create minimal damage to the parent structure and facilitate non-destructive inspections. Since aircraft structures are often subjected to fatigue loading condition, analysis of the fatigue behavior for composite patches repair is thus essential to evaluate the repair performances. Several studies were carried out to analyze the improvement of the fatigue life of repaired structures compared to unrepaired structures [1–8]. Bonded patch repair reduces stresses near the crack by transferring stresses between the cracked structure and the composite patch throughout the adhesive layer, and therefore retards the crack growth. Jones and Chiu [9] showed that externally bonded composite repairs can be successfully used to extend the fatigue life of thick structural components. Hosseini-Toudeshky [10] analyzed the effect of the number of plies of the composite on the repair performance and concluded that the life of a cracked panel of 2.29 mm thickness may increase by about 65% and 236%, by implementing a 4 and 16 layers patch, respectively. However, for 6.35 mm thickness, the life of repaired structures may be improved by only 21–35% according to Hosseini-Toudeshky [10]. Aakkula and Saarela [11] analyzed the effect of the stiffness ratio between the aluminum structure and carbon/epoxy and boron epoxy patches on the repair performance. They showed that this ratio has a significant effect on the fatigue life of repaired structures. Indeed, and according to the results of Aakkula and Saarela [11], ⇑ Corresponding author. Tel./fax: +213 48544100. E-mail addresses: [email protected], [email protected] (B. Bachir Bouiadjra). http://dx.doi.org/10.1016/j.engfracmech.2015.04.008 0013-7944/Ó 2015 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008
2
A. Albedah et al. / Engineering Fracture Mechanics xxx (2015) xxx–xxx
Nomenclature a F DF N R w W L
crack length load difference between maximum and minimum amplitude of fatigue loading number of fatigue cycles load ratio specimen width disbond width disbond length
the fatigue life improvement, for center crack repaired with single-sided patch, was better with the wet-laminated carbon/ epoxy having a stiffness ratio of 1.04. For edge-cracked specimens the best repair performance was achieved with the double-sided boron/epoxy repairs having a stiffness ratio of 0.65. Wang et al. [12] conducted experimental fatigue tests on notched AA7075 and AA6061 substrates in the gigacycle regime, with and without boron–epoxy patches of 1-ply, 2-pliesand 4-plies. They observed that the patch repair improves the fatigue life of the substrate considerably. In addition, Wang et al. [12] concluded that the improvement in the fatigue strength with a 1-ply patch is not significant while it is improved nearly 100 times with 4-plies patching. Mall and Conley [13] investigated the fatigue behavior of repaired crack in thick and thin panels. They observed that, due to the asymmetric repair there was a significant bending moment causing anon-uniform crack growth between the un-patched and patched faces of the repaired thick panels. This moment is negligible for thin panels [13]. Chung and Yang [14] analyzed the behavior of inclined crack repaired with bonded composite patch by conducting fatigue tests. Their results showed that the fatigue life of patched plate with inclined crack increased approximately 2.4–5.0 times compared to the un-patched plate. According to Chun and Yang [14] the maximum effect of the patch was obtained for a 0° crack inclination (pure mode I) and the effect was relatively weak for a 45° inclined crack. The adhesive disbond during the repair process has recently attracted some attention in the literature. Several papers describing the effects of the adhesive disbond on the repair efficiency were published [14–20]. Bachir Bouiadjra et al. [21] analyzed the effect of adhesive disbond on the performance of bonded composite repair in aircraft structures. They showed that adhesive disbond significantly reduces the repair efficiency by increasing the stress intensity factor at the crack tip. Bachir Bouiadjra et al. [21] conducted a numerical investigation on the effect of the disbond width (normal to the rack) and length (parallel to the crack) on the repair efficiency. It was shown that the increase of the disbond width significantly increases the stress intensity factor at the tip of repaired cracks which drastically reduces the repair efficiency, however, the effect of the disbond length is not significant. Ouinas et al. [22] investigated numerically the behavior of a cracked aluminum plate repaired with an adhesively bonded composite patch under a full length disbond. They confirmed that the reduction of stress around the crack tip increases with the patch thickness for a disbond width higher than the crack size. Many of the structures that utilize adhesive bonding are subjected to variable amplitude loading. It has been well documented that fatigue life of metals is affected by the load history. In order to obtain accurate fatigue life predictions, we need to take into account the effect of the load amplitude variation. In this work, we hypothesize that the loading variation affects the crack growth of bonded structures. So far, the available research on variable amplitude loading of adhesively bonded joints is quite limited when compared to the studies conducted on constant amplitude loading. In this study, we evaluated the repair efficiency by analyzing experimentally the fatigue life of repaired V-notched aluminum plate under constant and variable amplitude loading. The effects of the initial crack length of patching and the adhesive disbond on the repair efficiency were highlighted. 2. Experimental setup 2.1. Specimen details In this research, we conducted fatigue tests on unrepaired and repaired single edged notched tension (SENT) specimens (see Fig. 1). The specimens, of dimensions 150 50 2 mm, were cut from Al 7075-T6 plates. An initial v-notch of 6 mm depth and 60° angle was created in the center of each specimen by milling, as shown in Fig. 1, according to ASTM E647 standards [23]. The specimens were then pre-cracked to different initial crack lengths under fatigue loading. The presence of the V notch facilitates the mode I propagation of the crack. Once the desired crack length is reached, the sample was unloaded and taken for surface preparation followed by bonding the patch. 2.2. Patch preparation Composite patches are made with 8 plies of unidirectional carbon/epoxy pre-pregs. The dimension of each ply was 250 250 mm. After laying up, the laminate was cured under a hot press at 120 °C for 90 min. The pre-pregs were Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008
A. Albedah et al. / Engineering Fracture Mechanics xxx (2015) xxx–xxx
3
Fig. 1. Specimen details.
Fig. 2. Configuration of repaired specimen assembly.
sandwiched in between Teflon sheets and woven matrix of thermo plastic polymers to absorb the extra resin which comes out under curing. The composite plate is cut into 50 50 mm square patches. 2.3. Adhesive The surface of SENT specimens was prepared according to Bell Process Specification method [24]. The pre-cracked specimens were repaired with the composite patches bonded to the specimens using the bi-components Permabond ET515 adhesive such that the lay-up principal direction is perpendicular to the loading direction, as shown in Fig. 2. The SENT specimen were cleaned using acetone only as there were no indents or scratches on the specimen and it was received in mirror polished condition. After cleaning the specimen, Permabond ET515A, the epoxy resin, is mixed with Permabond ET515B, the polyamine hardener, in 1:1 ratio. The mixture is applied to pre-cracked area and the composite patch was bonded to it. Permabond ET515 gets cured without any external heat and thus, prevents the galvanization between the composite and the metal. The assembly of aluminum specimen and patch was allowed to cure under pressure for 72 h, without any heat to obtain the full handling strength. 2.4. Fatigue tests procedure Fatigue experiments were carried out on a 100 kN capacity Instron 8801 servo hydraulic machine. All the fatigue tests on un-repaired and repaired specimen were performed at room temperature using a sinusoidal waveform at a loading frequency of 20 Hz under constant and variable amplitude loading. The crack length vs number of cycles were recorded in each Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008
4
A. Albedah et al. / Engineering Fracture Mechanics xxx (2015) xxx–xxx
Fig. 3. Photo taken during the fatigue test.
case. A digital camera was used to monitor the crack propagation (see Fig. 3). No damage was detected in the Carbon fibers during the fatigue tests. 2.4.1. Constant amplitude loading Fatigue tests were conducted at a frequency of 20 Hz under constant amplitude loading DF equal to 6.3 kN and a load ratio R = Fmin/Fmax = 0.1. This ratio is a decisive factor in the fatigue life of structures. Non-negative stress ratio was chosen to avoid the closure of the crack under the minimum load Fmin, and thereby ensure its full opening during the test. Recall that the amplitude of solicitation DF does not vary whatever the state of the specimen is: repaired or unrepaired. The choice of the initial crack length of patching is important in achieving the repair objectives. To analyze the effect of the initial crack length of patching on the repair efficiency, we have conducted fatigue tests with different initial crack lengths: 3, 5, 10 and 15 mm. Knowing that the plate width is 50 mm, the ratios between these crack lengths and the plate width (a/W) are: 0.06, 0.1, 0.2 and 0.3. To analyze the effect of the adhesive disbond we have conducted several fatigue experiments on pre-cracked aluminum specimens, repaired with imperfectly bonded composite patches. We have introduced the adhesion defectiveness such that we can determine parametrically the patch repair performance for different possible imperfection scenarios. We used 0.1 mm thickness Teflon material on the areas on the cracked specimen in which we planned to prevent the adhesive from taking place (Fig. 4). Disbond was located from the edge of the specimen with different lengths and widths. The intensity of maximum applied stress is 70 MPa and load ratio R is equal to 0.1. 2.4.2. Variable amplitude loading Two cyclic amplitude-loading blocks (increasing and decreasing), as shown in Fig. 5, were applied to each specimen in two sequences. The tests were performed using the first load amplitude, low or high (L or H, 7 kN or 12 kN) until the crack reached 3 mm length, then the second load amplitude, high or low, (H or L, 12 kN or 7 kN) is applied and maintained until fracture. For the repaired specimens, the same procedure is used except that the composite patch is bonded after the crack length reaches a0 = 3 mm. The data obtained from these tests were analyzed and compared. 3. Results and discussion 3.1. Constant amplitude loading 3.1.1. Comparison between patched and un-patched cracks Fig. 6 presents the fatigue life (crack length vs. number of cycles) for unrepaired and repaired plates. The initial crack length of patching is 6 mm. From this figure, we note that the fatigue life of repaired plate is increased eleven times Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008
A. Albedah et al. / Engineering Fracture Mechanics xxx (2015) xxx–xxx
5
Fig. 4. Adhesive disbond details.
compared to the unrepaired one, hence, the composite repair can highly improve the global fatigue life of aircraft structures. In the literature, several ratios for the life extension were reported: four times, ten times, twenty times and even one hundred times. These differences are probably the result of adhesive disbond during the crack propagation. The adhesive disbond over the crack region can significantly reduce the stress transfer between the repaired plate and the composite patch causing a reduction of the repair efficiency. A well-designed repair can only be effective if the patch is strongly bonded to the parent adherent and therefore the issues of adhesive bond strength and bond durability are absolutely crucial for a successful repair. 3.1.2. Effect of the initial crack length on the repair performance Fig. 7 presents the fatigue life of repaired structures for different initial crack lengths of patching. The repair efficiency is found to depend strongly on the length of the initial crack. The fatigue life of the plate increases for small initial crack lengths. For initial length of 3 mm (a/W = 0.06), the number of cycles to failure is about: 200,000 cycles, which is thirteen times greater than the number of cycles for un-repaired crack. For a = 5 mm (a/W = 0.1), the fatigue life is improved ten times. This level of the fatigue life improvement reduces to 2.3 times when a = 15 mm. It can also be noted that the number of cycles to failure decreases linearly with the increase of the initial crack length of patching. This behavior may make it easier for repair designers to predict the fatigue life of repaired structures. 3.1.3. Effect of the adhesive disbond on the repair performance Fig. 8 presents the fatigue life of specimens with full lengths of the disbond and different disbond widths: 3, 6 and 9 mm. From this figure, it can be noted that the increase of the disbond width reduces the fatigue life. For example the fatigue life of repaired specimen with 9 mm of disbond width is practically the same compared to that of unrepaired specimens. This means that the disbond growth may completely revoke the repair efficiency. These results are in concordance with those of Ouinas et al. [22], who showed that the propagation of the adhesive disbond perpendicular to the crack has a negative effect on the repair efficiency. For comparison, the fatigue life of repaired specimens falls from 113,500 cycles to 91,000 cycles when the disbond width increases from 3 to 6 mm. The relative difference is about 20%. This difference is about the same (20%) when the adhesive disbond increases from 6 to 9 mm. We can thus deduce that the reduction of the fatigue life vary linearly with the growth of the adhesive disbond width. In order to illustrate the effect of the adhesive length on the repair efficiency, Fig. 9 presents the fatigue life graphs of repaired specimens with different disbond lengths (15, 30 and 45 mm) for a fixed disbond width of 6 mm. The results of Fig. 8 show clearly that the disbond length is of lower significance compared to that of the disbond width. This confirms Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008
6
A. Albedah et al. / Engineering Fracture Mechanics xxx (2015) xxx–xxx
Fig. 5. Variable amplitude loading sequence of two blocks. (a) Decreasing amplitude blocks (H–L); (b) increasing amplitude blocks (L–H).
45
Unrepaired Repaired
40
crack length (mm)
35 30 25 20 15 10 5 0 0
20000 40000 60000 80000 100000 120000 140000 160000 180000
Number of cycles Fig. 6. Fatigue life for repaired and un-repaired plate under constant amplitude loading.
the results of Bachir Bouiadjra et al. [21] whom showed that a parallel propagation of the adhesive disbond, has no significant effect on the repair performance. From Fig. 9, we can note that the fatigue life is reduced by 8.5% when the disbond length varies from 15 to 30 mm and by about 9% when the disbond length increases from 30 to 45 mm. In conclusion, we recommend the adhesive disbond must be taken into account in the repair design and must be monitored during the patch inspection. The authors have analyzed in precedent work the evolution of the stress intensity factor at the tip of repaired crack with presence of the adhesive delamination [21]. They showed that the stress intensity factor increases with the increase of the patch width, which means that the fiber bridging effect is attenuated by the presence of the adhesive disbond. Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008
A. Albedah et al. / Engineering Fracture Mechanics xxx (2015) xxx–xxx
7
patched for initial crack of 15 mm patched for initial crack of 10 mm patched for initial crack of 5 mm patched for initial crack of 3 mm unpatched crack
45 40 35
a (mm)
30 25 20 15 10 5 0 0
50000
100000
150000
200000
250000
Number of cycles Fig. 7. Fatigue life for different initial crack lengths of patching.
50 Unrepaired
Crack length (a, mm)
45
L=45mm,W=3mm
40
L=45mm,W=6mm
35
L=45mm,W=9mm
30 25 20 15 10 5 0 0
20000
40000
60000
80000
100000
120000
No. of Cycles (N) Fig. 8. Effect of the adhesive disbond width on the fatigue life of repaired specimens.
50 Unrepaired L=15mm,W=6mm L=30mm,W=6mm L=45mm,W=6mm
Crack length (a, mm)
45 40 35 30 25 20 15 10 5 0 0
20000
40000
60000
80000
100000
120000
No. of Cycles (N) Fig. 9. Fatigue life curves of repaired specimens for different disbond lengths.
3.2. Variable amplitude loading Fig. 10 presents the effect of the stress ratio on the fatigue life of specimens subjected to low–high stepped variable loading. The results shown in Fig. 10 indicate that the improvement in fatigue life of repaired cracked specimens is not significant for increasing fatigue blocks. The difference in the number of cycles to failure between unrepaired and repaired specimens is about 5000 cycles for R = 0 and 10,160 cycles for R = 0.1. Increasing the applied load after patching to 12 kN counterbalances the extra stiffness offered by the composite patch which reduces the repair efficiency. The load ratio R was found to have significant effect on the repair efficiency; the fatigue life of repaired structures was significantly reduced when the load ratio Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008
8
A. Albedah et al. / Engineering Fracture Mechanics xxx (2015) xxx–xxx 50 7075-R0.1-Repaired 7075-R0.1-Unrepaired 7075-R0-Repaired 7075-R0-Unrepaired
45
Crack length (a, mm)
40 35 30 25 20 15 10 5 0 0
5000
10000
15000
20000
25000
30000
35000
No. of cycles (N) Fig. 10. Fatigue life of repaired and un-repaired specimen under increasing (L–H) blocks.
7075-R0.1-Repaired
50
7075-R0.1-Unrepaired
Crack length (a, mm)
45
7075-R0-Repaired
40
7075-R0-Unrepaired
35 30 25 20 15 10 5 0 0
25000
50000
75000
100000
125000
150000
No. of cycles (N) Fig. 11. Fatigue life of repaired and un-repaired specimen under decreasing (H–L) blocks.
Table 1 Fatigue life summary. Loading sequence
Stress ratio (R)
Un-repaired (number of cycles to failure)
Repaired (number of cycles to failure)
Repair efficiency (%)
Increasing block
0 0.1
20,627 22,470
25,384 32,630
23.06 45.2
Decreasing block
0 0.1
45,194 77,300
123,514 135,060
173 74.7
decreases. The relative difference of the fatigue life for R = 0.1 and R = 0 is about 8.9%. This is because for R = 0 the difference between the maximum and the minimum applied stresses (Dr = rmax–rmin) is higher, which leads to a faster fatigue crack growth. Baker and Jones [25] showed that the peel stress in the adhesive layer decreases rapidly with the increase in the applied load ratio, whereas the maximum stresses increase slightly with the increase of load ratio. Fig. 11 represents the performance of bonded composite repair and the effect of the stress ratio on the fatigue life of Al 7075-T6 specimens subjected to decreasing blocks of loading. Similar to the increasing loading case, two stress ratios were applied R = 0 and R = 0.1. It is evident that the fatigue life of repaired specimens increases considerably in the case of decreasing blocks of loading. This is because, in addition to the increase of strength due to the patch, the plastic zone, formed around the crack tip due to the overload, retards the crack growth. The size of the plastic zone generated by the high loads would generate a retardation of the crack growth even for the unpatched specimen. The retardation is thought to be even higher for reinforced specimen due to fiber bridging. According to Fig. 10, the fatigue life is improved from 25,000 cycles to 35,000 cycles by the composite repair for R = 0.1, the relative increase of the fatigue life is about 40%. This improvement is due exclusively to the fiber bridging effect because the crack was patched without any plastic zone around its tip. For the decreasing blocks (Fig. 11), the crack retardation effect due to the overload significantly increases the fatigue life. The combination of the retardation and the fiber bridging effects can further improve the fatigue life of repaired structures. However, we estimate that the presence of the composite patch reduces the plastic zone around the crack tip by the bridging effect and consequently the retardation effect will be attenuated. This behavior can reduce the repair efficiency for materials of higher ductility. In the literature, Nolting et al. [26] analyzed the behavior of bonded aluminum joint under variable amplitude loading. They showed that the application of overload cycles increased the damage caused by subsequent small cycles for both substrate and adhesive failures. In future work, we plan to study the retardation effect for repaired structure in aluminum alloy Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008
9
A. Albedah et al. / Engineering Fracture Mechanics xxx (2015) xxx–xxx L-H, UR, R0 L-H, UR, R0.1 L-H, Repaired, R0 L-H, Repaired, R0.1
da/dN (m/cycles)
1.00E-04
1.00E-05
1.00E-06
1.00E-07 0
0.002
0.004
0.006
0.008
0.01
0.012
0.014
0.016
0.018
Crack length (a, m) Fig. 12. da/dN vs a for increasing blocks (L–H), repaired and unrepaired.
da/dN (m/cycles)
5.00E-05
H-L, UR, R0 H-L, UR, R0.1 H-L, Repaired, R0 H-L, Repaired, R0.1
5.00E-06
5.00E-07
5.00E-08
5.00E-09 0
0.005
0.01
0.015
0.02
0.025
0.03
Crack length (a, m) Fig. 13. da/dN vs a for decreasing blocks (L–H), repaired and unrepaired.
Fig. 14. Microscopic observation of failed unrepaired specimens loaded with decreasing and increasing blocks. (a) Increasing 1; (b) increasing 4; (c) decreasing 1; (d) decreasing 4.
2024-T3 which is more ductile than the 7075-T6. The fatigue life of repaired specimen is summarized and presented in Table 1. We observe that the improvement of fatigue life for H–L blocks is 173% higher for R = 0, whereas for L–H blocks, the repair efficiency is just 23%. So the combined effect of patch, reduce in load and the plastic zone due to overload for H–L significantly blocks improve the repair efficiency. For better explaining the crack growth for repaired and unrepaired cracks under variable amplitude loading, Figs. 12 and 13 present the variation of the fatigue crack growth rate (da/dN) as a function of the crack length for increasing and Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008
10
A. Albedah et al. / Engineering Fracture Mechanics xxx (2015) xxx–xxx
Fig. 15. Microscopic observation of failed repaired specimens loaded with decreasing and increasing blocks. (a) Increasing 1; (b) increasing 4; (c) decreasing 1; (d) decreasing 4.
decreasing blocks respectively. From Fig. 12, we can see that for increasing blocks, the dispersions of the results are very significant and the variation of da/dN is not uniform when the crack length increases. In addition, the crack growth is faster for unrepaired cracks compared to the repaired ones for R = 0 and R = 0.1. However, for decreasing blocks (Fig. 13), the variation of the crack growth rate (da/dN) is more uniform and varies linearly as a function of the crack length (a). The difference in the crack growth rate between unrepaired and repaired cracks is not significant. This is mainly due to the retardation effect generated by the plasticity around the crack tip. Figs. 14 and 15 present microscopic observation of fracture surfaces for unrepaired and repaired specimens respectively and for the two cases of variable loading (increasing and decreasing blocks). Before patching, these figures shows a brittle fracture for increasing blocks and ductile fracture for decreasing block. Fig. 15 shows a more regular crack growth at the top face (patched face) than the bottom one (unpatched). The shear band at the unpatched face is clearly deeper and growing faster than that at the reinforced face. In Fig. 14, the two faces of the specimen are clearly behaving symmetrically and the shear band seems to be faster growing than that created in the patched specimen. The microscopic observations and comparison suggest that there is definitely a shear failure tunneling effect caused by the CFRP patch. 4. Conclusions This study was carried out to evaluate the performances of bonded composite repair of lateral V-notched7075 T6 aluminum alloy plate. Fatigue tests, with constant and stepped variable amplitude loadings, were performed to evaluate the composite repair performance. From the results of these tests we can conclude that: – The composite patches increase considerably the fatigue life of cracked aluminum structures but the level of improvement depends strongly on the adhesive layer between the aluminum structure and the composite patch. The Adhesive disbond do highly decrease the fatigue life of repaired structures. – The initial crack length of patching has a significant effect on the repair efficiency and a crack should be repaired as soon as it is detected. – As the load ratio (R) increases, the fatigue life increases for both repaired and unrepaired specimen because of the difference between the maximum and the minimum applied stresses. The higher the difference between the maximum and the minimum stresses, the lower the fatigue life. – The patch repair efficiency is not improved for increasing amplitude loading blocks, whereas a relatively small improvement is noticed for decreasing loading amplitude blocks. – The plastic zone formed by the early high loads (overload) of the decreasing loading blocks retards the crack growth for unrepaired specimens. Nonetheless, for repaired specimens, the patch carries the stress and diminishes the effect of the overload local plasticity, which decreases the patch repair efficiency.
Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008
A. Albedah et al. / Engineering Fracture Mechanics xxx (2015) xxx–xxx
11
Acknowledgment The Authors extend their appreciation to the Deanship of Scientific Research at King Saud University for funding the work through the research group No. RGP-VPP-035. References [1] Oudad W, Bachir Bouiadjra B, Belhouari M, Touzain S, Feaugas X. Analysis of the plastic zone size ahead of repaired cracks with bonded composite patch of metallic aircraft structures. Comput Mater Sci 2009;6:950–4. [2] Tsai GC, Shen SB. Fatigue analysis of cracked thick aluminum plate bonded with composite patches. Compos Struct 2004;64:79–90. [3] Sabelkin V, Mall S, Avram JB. Fatigue crack growth analysis of stiffened cracked panel repaired with bonded composite patch. Engng Fract Mech 2006;73:1553–67. [4] Okfar C, Singh N, Enemuoh UE, Rao SV. Design, analysis and performance of adhesively bonded composite patch repair of cracked aluminum aircraft panels. Compos Struct 2005;71:258–70. [5] Lee WY, Lee JJ. Successive 3D FE analysis technique for characterization of fatigue crack growth behavior in composite-repaired aluminum plate. Compos Struct 2004;66:513–20. [6] Liljedahl CDM, Fitzpatrick ME, Edwards L. Fatigue crack growth in integral structures with crack retarders. Mater Sci Engng, A 2009;523:152–9. [7] Tsouvalis Nicholas G, Mirisiotis Lazarus S, Dimou Dimitris N. Experimental and numerical study of the fatigue behaviour of composite patch reinforced cracked steel plates. Int J Fatigue 2009;31:1613–27. [8] Baker Alan, Rajic Nik, Davis Claire. Towards a practical structural health monitoring technology for patched cracks in aircraft structure. Compos A Appl Sci Manuf 2009;40:1340–52. [9] Jones R, Chiu WK. Composite repairs in metallic components. Compos Struct 1999;44:17–29. [10] Hosseini-Toudeshky H. Effects of composite patches on fatigue crack propagation of single-side repaired aluminum panels. Compos Struct 2006;76:243–51. [11] Aakkula Jarkko, Saarela Olli. An experimental study on the fatigue performance of CFRP and BFRP repaired aluminium plates. Compos Struct 2014;118:589–99. [12] Wang QY, Sriraman MR, Kawagoishi N, Chen Q. Fatigue crack growth of bonded composite repairs in gigacycle regime. Int J Fatigue 2006;28:1197–201. [13] Mall S, Conley DS. Modeling and validation of composite patch repair to cracked thick and thin metallic panels. Compos A Appl Sci Manuf 2009;40:1331–9. [14] Chung Ki Hyun, Yang Won H. Mixed mode fatigue crack growth in aluminum plates with composite patches. Int J Fatigue 2003;25:325–33. [15] Genest M, Martinez M, Mrad N, Renaud G, Fahr A. Pulsed thermography for non-destructive evaluation and damage growth monitoring of bonded repairs. Compos Struct 2009;88:112–20. [16] Qing XP, Beard SJ, Kumar A, Hannum R. A real-time active smart patch system for monitoring the integrity of bonded repair on an aircraft structure. Smart Mater Struct 2006;5:66–73. [17] Ouinas D, Bouiadjra BB, Serier B. The effects of disbonds on the stress intensity factor of aluminium panels repaired using composite materials. Compos Struct 2007;78:278–84. [18] Ouinas D, Bachir Bouiadjra B, Achour T, Benderdouche N. Influence of disbond on notch crack behaviour in single bonded lap joints. Mater Des 2010;31:4356–62. [19] Alderliesten RC. Damage tolerance of bonded aircraft structures. Int J Fatigue 2009;31:1024–30. [20] Jones R, Callinan RJ, Aggraval KC. Analysis of bonded repairs to damage fiber composite structures. Engng Fract Mech 1983:37–46. [21] Bachir Bouiadjra B, Oudad W, Albedah A, Benyahia F, Belhouari M. Effects of the adhesive disbond on the performances of bonded composite repairs in aircraft structures. Mater Des 2012;37:89–95. [22] Ouinas D, Bachir Bouiadjra B, Himouri S, Benderdouche N. Progressive edge cracked aluminium plate repaired with adhesively bonded composite patch under full width disbond. Compos B Engng 2012;43:805–11. [23] Standard A. E647. Standard test method for measurement of fatigue crack growth rates. Annual book of ASTM Standards, 3; 2000. [24] Specification BP. Surface preparation of materials for adhesive bonding. BPS FW, 4352; 1972. [25] Baker AA, Jones R. Bonded repair of aircraft structures, vol. 7. Springer; 1988. [26] Nolting AE, Underhill PR, DuQuesnay DL. Variable amplitude fatigue of bonded aluminum joints. Int J Fatigue 2008;30:178–87.
Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008
Contents lists available at ScienceDirect
Engineering Fracture Mechanics journal homepage: www.elsevier.com/locate/engfracmech
Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch A. Albedah a, Sohail M.A. Khan a, F. Benyahia a, B. Bachir Bouiadjra b,⇑ a b
Mechanical Engineering Department, College of Engineering, King Saud University, Riyadh, Saudi Arabia LMPM, Department of Mechanical Engineering, University of Sidi Bel Abbes, BP 89 Cité Ben M’hidi, Sidi Bel Abbes 22000, Algeria
a r t i c l e
i n f o
Article history: Received 28 October 2014 Received in revised form 7 April 2015 Accepted 8 April 2015 Available online xxxx Keywords: Bonded composite repair Fatigue life Constant amplitude loading Variable amplitude loading Adhesive disbond Stress ratio
a b s t r a c t In this study, we analyzed the fatigue behavior of a V-notch crack in an aluminum alloy 7075-T6 plate repaired with bonded composite patch under constant and stepped variable amplitude loading. The effect of adhesive disbond on the repair performance was analyzed. The obtained results confirm the need of repairing cracks at their early creation for a maximum repair performance and showed that the increases in the disbond width and the stress ratio have a large negative effect on the repair efficiency. The latter is low for increasing loading blocks but significantly high for decreasing blocks. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Bonded composite repair of aircraft structures was developed by Dr. Alan Baker and his team in 1972. The technology involves adhesively bonding patches of advanced fiber composite materials to repair damaged aircraft structures and to stop stress corrosion cracking. The repairs are structurally very efficient, can be applied rapidly and are cost effective. The technology has many advantages over traditional mechanical repair methods, such as bolting or riveting. Composite patches are lighter, offer more uniform load transfer, seal interfaces reducing corrosion and leakage, create minimal damage to the parent structure and facilitate non-destructive inspections. Since aircraft structures are often subjected to fatigue loading condition, analysis of the fatigue behavior for composite patches repair is thus essential to evaluate the repair performances. Several studies were carried out to analyze the improvement of the fatigue life of repaired structures compared to unrepaired structures [1–8]. Bonded patch repair reduces stresses near the crack by transferring stresses between the cracked structure and the composite patch throughout the adhesive layer, and therefore retards the crack growth. Jones and Chiu [9] showed that externally bonded composite repairs can be successfully used to extend the fatigue life of thick structural components. Hosseini-Toudeshky [10] analyzed the effect of the number of plies of the composite on the repair performance and concluded that the life of a cracked panel of 2.29 mm thickness may increase by about 65% and 236%, by implementing a 4 and 16 layers patch, respectively. However, for 6.35 mm thickness, the life of repaired structures may be improved by only 21–35% according to Hosseini-Toudeshky [10]. Aakkula and Saarela [11] analyzed the effect of the stiffness ratio between the aluminum structure and carbon/epoxy and boron epoxy patches on the repair performance. They showed that this ratio has a significant effect on the fatigue life of repaired structures. Indeed, and according to the results of Aakkula and Saarela [11], ⇑ Corresponding author. Tel./fax: +213 48544100. E-mail addresses: [email protected], [email protected] (B. Bachir Bouiadjra). http://dx.doi.org/10.1016/j.engfracmech.2015.04.008 0013-7944/Ó 2015 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008
2
A. Albedah et al. / Engineering Fracture Mechanics xxx (2015) xxx–xxx
Nomenclature a F DF N R w W L
crack length load difference between maximum and minimum amplitude of fatigue loading number of fatigue cycles load ratio specimen width disbond width disbond length
the fatigue life improvement, for center crack repaired with single-sided patch, was better with the wet-laminated carbon/ epoxy having a stiffness ratio of 1.04. For edge-cracked specimens the best repair performance was achieved with the double-sided boron/epoxy repairs having a stiffness ratio of 0.65. Wang et al. [12] conducted experimental fatigue tests on notched AA7075 and AA6061 substrates in the gigacycle regime, with and without boron–epoxy patches of 1-ply, 2-pliesand 4-plies. They observed that the patch repair improves the fatigue life of the substrate considerably. In addition, Wang et al. [12] concluded that the improvement in the fatigue strength with a 1-ply patch is not significant while it is improved nearly 100 times with 4-plies patching. Mall and Conley [13] investigated the fatigue behavior of repaired crack in thick and thin panels. They observed that, due to the asymmetric repair there was a significant bending moment causing anon-uniform crack growth between the un-patched and patched faces of the repaired thick panels. This moment is negligible for thin panels [13]. Chung and Yang [14] analyzed the behavior of inclined crack repaired with bonded composite patch by conducting fatigue tests. Their results showed that the fatigue life of patched plate with inclined crack increased approximately 2.4–5.0 times compared to the un-patched plate. According to Chun and Yang [14] the maximum effect of the patch was obtained for a 0° crack inclination (pure mode I) and the effect was relatively weak for a 45° inclined crack. The adhesive disbond during the repair process has recently attracted some attention in the literature. Several papers describing the effects of the adhesive disbond on the repair efficiency were published [14–20]. Bachir Bouiadjra et al. [21] analyzed the effect of adhesive disbond on the performance of bonded composite repair in aircraft structures. They showed that adhesive disbond significantly reduces the repair efficiency by increasing the stress intensity factor at the crack tip. Bachir Bouiadjra et al. [21] conducted a numerical investigation on the effect of the disbond width (normal to the rack) and length (parallel to the crack) on the repair efficiency. It was shown that the increase of the disbond width significantly increases the stress intensity factor at the tip of repaired cracks which drastically reduces the repair efficiency, however, the effect of the disbond length is not significant. Ouinas et al. [22] investigated numerically the behavior of a cracked aluminum plate repaired with an adhesively bonded composite patch under a full length disbond. They confirmed that the reduction of stress around the crack tip increases with the patch thickness for a disbond width higher than the crack size. Many of the structures that utilize adhesive bonding are subjected to variable amplitude loading. It has been well documented that fatigue life of metals is affected by the load history. In order to obtain accurate fatigue life predictions, we need to take into account the effect of the load amplitude variation. In this work, we hypothesize that the loading variation affects the crack growth of bonded structures. So far, the available research on variable amplitude loading of adhesively bonded joints is quite limited when compared to the studies conducted on constant amplitude loading. In this study, we evaluated the repair efficiency by analyzing experimentally the fatigue life of repaired V-notched aluminum plate under constant and variable amplitude loading. The effects of the initial crack length of patching and the adhesive disbond on the repair efficiency were highlighted. 2. Experimental setup 2.1. Specimen details In this research, we conducted fatigue tests on unrepaired and repaired single edged notched tension (SENT) specimens (see Fig. 1). The specimens, of dimensions 150 50 2 mm, were cut from Al 7075-T6 plates. An initial v-notch of 6 mm depth and 60° angle was created in the center of each specimen by milling, as shown in Fig. 1, according to ASTM E647 standards [23]. The specimens were then pre-cracked to different initial crack lengths under fatigue loading. The presence of the V notch facilitates the mode I propagation of the crack. Once the desired crack length is reached, the sample was unloaded and taken for surface preparation followed by bonding the patch. 2.2. Patch preparation Composite patches are made with 8 plies of unidirectional carbon/epoxy pre-pregs. The dimension of each ply was 250 250 mm. After laying up, the laminate was cured under a hot press at 120 °C for 90 min. The pre-pregs were Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008
A. Albedah et al. / Engineering Fracture Mechanics xxx (2015) xxx–xxx
3
Fig. 1. Specimen details.
Fig. 2. Configuration of repaired specimen assembly.
sandwiched in between Teflon sheets and woven matrix of thermo plastic polymers to absorb the extra resin which comes out under curing. The composite plate is cut into 50 50 mm square patches. 2.3. Adhesive The surface of SENT specimens was prepared according to Bell Process Specification method [24]. The pre-cracked specimens were repaired with the composite patches bonded to the specimens using the bi-components Permabond ET515 adhesive such that the lay-up principal direction is perpendicular to the loading direction, as shown in Fig. 2. The SENT specimen were cleaned using acetone only as there were no indents or scratches on the specimen and it was received in mirror polished condition. After cleaning the specimen, Permabond ET515A, the epoxy resin, is mixed with Permabond ET515B, the polyamine hardener, in 1:1 ratio. The mixture is applied to pre-cracked area and the composite patch was bonded to it. Permabond ET515 gets cured without any external heat and thus, prevents the galvanization between the composite and the metal. The assembly of aluminum specimen and patch was allowed to cure under pressure for 72 h, without any heat to obtain the full handling strength. 2.4. Fatigue tests procedure Fatigue experiments were carried out on a 100 kN capacity Instron 8801 servo hydraulic machine. All the fatigue tests on un-repaired and repaired specimen were performed at room temperature using a sinusoidal waveform at a loading frequency of 20 Hz under constant and variable amplitude loading. The crack length vs number of cycles were recorded in each Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008
4
A. Albedah et al. / Engineering Fracture Mechanics xxx (2015) xxx–xxx
Fig. 3. Photo taken during the fatigue test.
case. A digital camera was used to monitor the crack propagation (see Fig. 3). No damage was detected in the Carbon fibers during the fatigue tests. 2.4.1. Constant amplitude loading Fatigue tests were conducted at a frequency of 20 Hz under constant amplitude loading DF equal to 6.3 kN and a load ratio R = Fmin/Fmax = 0.1. This ratio is a decisive factor in the fatigue life of structures. Non-negative stress ratio was chosen to avoid the closure of the crack under the minimum load Fmin, and thereby ensure its full opening during the test. Recall that the amplitude of solicitation DF does not vary whatever the state of the specimen is: repaired or unrepaired. The choice of the initial crack length of patching is important in achieving the repair objectives. To analyze the effect of the initial crack length of patching on the repair efficiency, we have conducted fatigue tests with different initial crack lengths: 3, 5, 10 and 15 mm. Knowing that the plate width is 50 mm, the ratios between these crack lengths and the plate width (a/W) are: 0.06, 0.1, 0.2 and 0.3. To analyze the effect of the adhesive disbond we have conducted several fatigue experiments on pre-cracked aluminum specimens, repaired with imperfectly bonded composite patches. We have introduced the adhesion defectiveness such that we can determine parametrically the patch repair performance for different possible imperfection scenarios. We used 0.1 mm thickness Teflon material on the areas on the cracked specimen in which we planned to prevent the adhesive from taking place (Fig. 4). Disbond was located from the edge of the specimen with different lengths and widths. The intensity of maximum applied stress is 70 MPa and load ratio R is equal to 0.1. 2.4.2. Variable amplitude loading Two cyclic amplitude-loading blocks (increasing and decreasing), as shown in Fig. 5, were applied to each specimen in two sequences. The tests were performed using the first load amplitude, low or high (L or H, 7 kN or 12 kN) until the crack reached 3 mm length, then the second load amplitude, high or low, (H or L, 12 kN or 7 kN) is applied and maintained until fracture. For the repaired specimens, the same procedure is used except that the composite patch is bonded after the crack length reaches a0 = 3 mm. The data obtained from these tests were analyzed and compared. 3. Results and discussion 3.1. Constant amplitude loading 3.1.1. Comparison between patched and un-patched cracks Fig. 6 presents the fatigue life (crack length vs. number of cycles) for unrepaired and repaired plates. The initial crack length of patching is 6 mm. From this figure, we note that the fatigue life of repaired plate is increased eleven times Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008
A. Albedah et al. / Engineering Fracture Mechanics xxx (2015) xxx–xxx
5
Fig. 4. Adhesive disbond details.
compared to the unrepaired one, hence, the composite repair can highly improve the global fatigue life of aircraft structures. In the literature, several ratios for the life extension were reported: four times, ten times, twenty times and even one hundred times. These differences are probably the result of adhesive disbond during the crack propagation. The adhesive disbond over the crack region can significantly reduce the stress transfer between the repaired plate and the composite patch causing a reduction of the repair efficiency. A well-designed repair can only be effective if the patch is strongly bonded to the parent adherent and therefore the issues of adhesive bond strength and bond durability are absolutely crucial for a successful repair. 3.1.2. Effect of the initial crack length on the repair performance Fig. 7 presents the fatigue life of repaired structures for different initial crack lengths of patching. The repair efficiency is found to depend strongly on the length of the initial crack. The fatigue life of the plate increases for small initial crack lengths. For initial length of 3 mm (a/W = 0.06), the number of cycles to failure is about: 200,000 cycles, which is thirteen times greater than the number of cycles for un-repaired crack. For a = 5 mm (a/W = 0.1), the fatigue life is improved ten times. This level of the fatigue life improvement reduces to 2.3 times when a = 15 mm. It can also be noted that the number of cycles to failure decreases linearly with the increase of the initial crack length of patching. This behavior may make it easier for repair designers to predict the fatigue life of repaired structures. 3.1.3. Effect of the adhesive disbond on the repair performance Fig. 8 presents the fatigue life of specimens with full lengths of the disbond and different disbond widths: 3, 6 and 9 mm. From this figure, it can be noted that the increase of the disbond width reduces the fatigue life. For example the fatigue life of repaired specimen with 9 mm of disbond width is practically the same compared to that of unrepaired specimens. This means that the disbond growth may completely revoke the repair efficiency. These results are in concordance with those of Ouinas et al. [22], who showed that the propagation of the adhesive disbond perpendicular to the crack has a negative effect on the repair efficiency. For comparison, the fatigue life of repaired specimens falls from 113,500 cycles to 91,000 cycles when the disbond width increases from 3 to 6 mm. The relative difference is about 20%. This difference is about the same (20%) when the adhesive disbond increases from 6 to 9 mm. We can thus deduce that the reduction of the fatigue life vary linearly with the growth of the adhesive disbond width. In order to illustrate the effect of the adhesive length on the repair efficiency, Fig. 9 presents the fatigue life graphs of repaired specimens with different disbond lengths (15, 30 and 45 mm) for a fixed disbond width of 6 mm. The results of Fig. 8 show clearly that the disbond length is of lower significance compared to that of the disbond width. This confirms Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008
6
A. Albedah et al. / Engineering Fracture Mechanics xxx (2015) xxx–xxx
Fig. 5. Variable amplitude loading sequence of two blocks. (a) Decreasing amplitude blocks (H–L); (b) increasing amplitude blocks (L–H).
45
Unrepaired Repaired
40
crack length (mm)
35 30 25 20 15 10 5 0 0
20000 40000 60000 80000 100000 120000 140000 160000 180000
Number of cycles Fig. 6. Fatigue life for repaired and un-repaired plate under constant amplitude loading.
the results of Bachir Bouiadjra et al. [21] whom showed that a parallel propagation of the adhesive disbond, has no significant effect on the repair performance. From Fig. 9, we can note that the fatigue life is reduced by 8.5% when the disbond length varies from 15 to 30 mm and by about 9% when the disbond length increases from 30 to 45 mm. In conclusion, we recommend the adhesive disbond must be taken into account in the repair design and must be monitored during the patch inspection. The authors have analyzed in precedent work the evolution of the stress intensity factor at the tip of repaired crack with presence of the adhesive delamination [21]. They showed that the stress intensity factor increases with the increase of the patch width, which means that the fiber bridging effect is attenuated by the presence of the adhesive disbond. Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008
A. Albedah et al. / Engineering Fracture Mechanics xxx (2015) xxx–xxx
7
patched for initial crack of 15 mm patched for initial crack of 10 mm patched for initial crack of 5 mm patched for initial crack of 3 mm unpatched crack
45 40 35
a (mm)
30 25 20 15 10 5 0 0
50000
100000
150000
200000
250000
Number of cycles Fig. 7. Fatigue life for different initial crack lengths of patching.
50 Unrepaired
Crack length (a, mm)
45
L=45mm,W=3mm
40
L=45mm,W=6mm
35
L=45mm,W=9mm
30 25 20 15 10 5 0 0
20000
40000
60000
80000
100000
120000
No. of Cycles (N) Fig. 8. Effect of the adhesive disbond width on the fatigue life of repaired specimens.
50 Unrepaired L=15mm,W=6mm L=30mm,W=6mm L=45mm,W=6mm
Crack length (a, mm)
45 40 35 30 25 20 15 10 5 0 0
20000
40000
60000
80000
100000
120000
No. of Cycles (N) Fig. 9. Fatigue life curves of repaired specimens for different disbond lengths.
3.2. Variable amplitude loading Fig. 10 presents the effect of the stress ratio on the fatigue life of specimens subjected to low–high stepped variable loading. The results shown in Fig. 10 indicate that the improvement in fatigue life of repaired cracked specimens is not significant for increasing fatigue blocks. The difference in the number of cycles to failure between unrepaired and repaired specimens is about 5000 cycles for R = 0 and 10,160 cycles for R = 0.1. Increasing the applied load after patching to 12 kN counterbalances the extra stiffness offered by the composite patch which reduces the repair efficiency. The load ratio R was found to have significant effect on the repair efficiency; the fatigue life of repaired structures was significantly reduced when the load ratio Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008
8
A. Albedah et al. / Engineering Fracture Mechanics xxx (2015) xxx–xxx 50 7075-R0.1-Repaired 7075-R0.1-Unrepaired 7075-R0-Repaired 7075-R0-Unrepaired
45
Crack length (a, mm)
40 35 30 25 20 15 10 5 0 0
5000
10000
15000
20000
25000
30000
35000
No. of cycles (N) Fig. 10. Fatigue life of repaired and un-repaired specimen under increasing (L–H) blocks.
7075-R0.1-Repaired
50
7075-R0.1-Unrepaired
Crack length (a, mm)
45
7075-R0-Repaired
40
7075-R0-Unrepaired
35 30 25 20 15 10 5 0 0
25000
50000
75000
100000
125000
150000
No. of cycles (N) Fig. 11. Fatigue life of repaired and un-repaired specimen under decreasing (H–L) blocks.
Table 1 Fatigue life summary. Loading sequence
Stress ratio (R)
Un-repaired (number of cycles to failure)
Repaired (number of cycles to failure)
Repair efficiency (%)
Increasing block
0 0.1
20,627 22,470
25,384 32,630
23.06 45.2
Decreasing block
0 0.1
45,194 77,300
123,514 135,060
173 74.7
decreases. The relative difference of the fatigue life for R = 0.1 and R = 0 is about 8.9%. This is because for R = 0 the difference between the maximum and the minimum applied stresses (Dr = rmax–rmin) is higher, which leads to a faster fatigue crack growth. Baker and Jones [25] showed that the peel stress in the adhesive layer decreases rapidly with the increase in the applied load ratio, whereas the maximum stresses increase slightly with the increase of load ratio. Fig. 11 represents the performance of bonded composite repair and the effect of the stress ratio on the fatigue life of Al 7075-T6 specimens subjected to decreasing blocks of loading. Similar to the increasing loading case, two stress ratios were applied R = 0 and R = 0.1. It is evident that the fatigue life of repaired specimens increases considerably in the case of decreasing blocks of loading. This is because, in addition to the increase of strength due to the patch, the plastic zone, formed around the crack tip due to the overload, retards the crack growth. The size of the plastic zone generated by the high loads would generate a retardation of the crack growth even for the unpatched specimen. The retardation is thought to be even higher for reinforced specimen due to fiber bridging. According to Fig. 10, the fatigue life is improved from 25,000 cycles to 35,000 cycles by the composite repair for R = 0.1, the relative increase of the fatigue life is about 40%. This improvement is due exclusively to the fiber bridging effect because the crack was patched without any plastic zone around its tip. For the decreasing blocks (Fig. 11), the crack retardation effect due to the overload significantly increases the fatigue life. The combination of the retardation and the fiber bridging effects can further improve the fatigue life of repaired structures. However, we estimate that the presence of the composite patch reduces the plastic zone around the crack tip by the bridging effect and consequently the retardation effect will be attenuated. This behavior can reduce the repair efficiency for materials of higher ductility. In the literature, Nolting et al. [26] analyzed the behavior of bonded aluminum joint under variable amplitude loading. They showed that the application of overload cycles increased the damage caused by subsequent small cycles for both substrate and adhesive failures. In future work, we plan to study the retardation effect for repaired structure in aluminum alloy Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008
9
A. Albedah et al. / Engineering Fracture Mechanics xxx (2015) xxx–xxx L-H, UR, R0 L-H, UR, R0.1 L-H, Repaired, R0 L-H, Repaired, R0.1
da/dN (m/cycles)
1.00E-04
1.00E-05
1.00E-06
1.00E-07 0
0.002
0.004
0.006
0.008
0.01
0.012
0.014
0.016
0.018
Crack length (a, m) Fig. 12. da/dN vs a for increasing blocks (L–H), repaired and unrepaired.
da/dN (m/cycles)
5.00E-05
H-L, UR, R0 H-L, UR, R0.1 H-L, Repaired, R0 H-L, Repaired, R0.1
5.00E-06
5.00E-07
5.00E-08
5.00E-09 0
0.005
0.01
0.015
0.02
0.025
0.03
Crack length (a, m) Fig. 13. da/dN vs a for decreasing blocks (L–H), repaired and unrepaired.
Fig. 14. Microscopic observation of failed unrepaired specimens loaded with decreasing and increasing blocks. (a) Increasing 1; (b) increasing 4; (c) decreasing 1; (d) decreasing 4.
2024-T3 which is more ductile than the 7075-T6. The fatigue life of repaired specimen is summarized and presented in Table 1. We observe that the improvement of fatigue life for H–L blocks is 173% higher for R = 0, whereas for L–H blocks, the repair efficiency is just 23%. So the combined effect of patch, reduce in load and the plastic zone due to overload for H–L significantly blocks improve the repair efficiency. For better explaining the crack growth for repaired and unrepaired cracks under variable amplitude loading, Figs. 12 and 13 present the variation of the fatigue crack growth rate (da/dN) as a function of the crack length for increasing and Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008
10
A. Albedah et al. / Engineering Fracture Mechanics xxx (2015) xxx–xxx
Fig. 15. Microscopic observation of failed repaired specimens loaded with decreasing and increasing blocks. (a) Increasing 1; (b) increasing 4; (c) decreasing 1; (d) decreasing 4.
decreasing blocks respectively. From Fig. 12, we can see that for increasing blocks, the dispersions of the results are very significant and the variation of da/dN is not uniform when the crack length increases. In addition, the crack growth is faster for unrepaired cracks compared to the repaired ones for R = 0 and R = 0.1. However, for decreasing blocks (Fig. 13), the variation of the crack growth rate (da/dN) is more uniform and varies linearly as a function of the crack length (a). The difference in the crack growth rate between unrepaired and repaired cracks is not significant. This is mainly due to the retardation effect generated by the plasticity around the crack tip. Figs. 14 and 15 present microscopic observation of fracture surfaces for unrepaired and repaired specimens respectively and for the two cases of variable loading (increasing and decreasing blocks). Before patching, these figures shows a brittle fracture for increasing blocks and ductile fracture for decreasing block. Fig. 15 shows a more regular crack growth at the top face (patched face) than the bottom one (unpatched). The shear band at the unpatched face is clearly deeper and growing faster than that at the reinforced face. In Fig. 14, the two faces of the specimen are clearly behaving symmetrically and the shear band seems to be faster growing than that created in the patched specimen. The microscopic observations and comparison suggest that there is definitely a shear failure tunneling effect caused by the CFRP patch. 4. Conclusions This study was carried out to evaluate the performances of bonded composite repair of lateral V-notched7075 T6 aluminum alloy plate. Fatigue tests, with constant and stepped variable amplitude loadings, were performed to evaluate the composite repair performance. From the results of these tests we can conclude that: – The composite patches increase considerably the fatigue life of cracked aluminum structures but the level of improvement depends strongly on the adhesive layer between the aluminum structure and the composite patch. The Adhesive disbond do highly decrease the fatigue life of repaired structures. – The initial crack length of patching has a significant effect on the repair efficiency and a crack should be repaired as soon as it is detected. – As the load ratio (R) increases, the fatigue life increases for both repaired and unrepaired specimen because of the difference between the maximum and the minimum applied stresses. The higher the difference between the maximum and the minimum stresses, the lower the fatigue life. – The patch repair efficiency is not improved for increasing amplitude loading blocks, whereas a relatively small improvement is noticed for decreasing loading amplitude blocks. – The plastic zone formed by the early high loads (overload) of the decreasing loading blocks retards the crack growth for unrepaired specimens. Nonetheless, for repaired specimens, the patch carries the stress and diminishes the effect of the overload local plasticity, which decreases the patch repair efficiency.
Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008
A. Albedah et al. / Engineering Fracture Mechanics xxx (2015) xxx–xxx
11
Acknowledgment The Authors extend their appreciation to the Deanship of Scientific Research at King Saud University for funding the work through the research group No. RGP-VPP-035. References [1] Oudad W, Bachir Bouiadjra B, Belhouari M, Touzain S, Feaugas X. Analysis of the plastic zone size ahead of repaired cracks with bonded composite patch of metallic aircraft structures. Comput Mater Sci 2009;6:950–4. [2] Tsai GC, Shen SB. Fatigue analysis of cracked thick aluminum plate bonded with composite patches. Compos Struct 2004;64:79–90. [3] Sabelkin V, Mall S, Avram JB. Fatigue crack growth analysis of stiffened cracked panel repaired with bonded composite patch. Engng Fract Mech 2006;73:1553–67. [4] Okfar C, Singh N, Enemuoh UE, Rao SV. Design, analysis and performance of adhesively bonded composite patch repair of cracked aluminum aircraft panels. Compos Struct 2005;71:258–70. [5] Lee WY, Lee JJ. Successive 3D FE analysis technique for characterization of fatigue crack growth behavior in composite-repaired aluminum plate. Compos Struct 2004;66:513–20. [6] Liljedahl CDM, Fitzpatrick ME, Edwards L. Fatigue crack growth in integral structures with crack retarders. Mater Sci Engng, A 2009;523:152–9. [7] Tsouvalis Nicholas G, Mirisiotis Lazarus S, Dimou Dimitris N. Experimental and numerical study of the fatigue behaviour of composite patch reinforced cracked steel plates. Int J Fatigue 2009;31:1613–27. [8] Baker Alan, Rajic Nik, Davis Claire. Towards a practical structural health monitoring technology for patched cracks in aircraft structure. Compos A Appl Sci Manuf 2009;40:1340–52. [9] Jones R, Chiu WK. Composite repairs in metallic components. Compos Struct 1999;44:17–29. [10] Hosseini-Toudeshky H. Effects of composite patches on fatigue crack propagation of single-side repaired aluminum panels. Compos Struct 2006;76:243–51. [11] Aakkula Jarkko, Saarela Olli. An experimental study on the fatigue performance of CFRP and BFRP repaired aluminium plates. Compos Struct 2014;118:589–99. [12] Wang QY, Sriraman MR, Kawagoishi N, Chen Q. Fatigue crack growth of bonded composite repairs in gigacycle regime. Int J Fatigue 2006;28:1197–201. [13] Mall S, Conley DS. Modeling and validation of composite patch repair to cracked thick and thin metallic panels. Compos A Appl Sci Manuf 2009;40:1331–9. [14] Chung Ki Hyun, Yang Won H. Mixed mode fatigue crack growth in aluminum plates with composite patches. Int J Fatigue 2003;25:325–33. [15] Genest M, Martinez M, Mrad N, Renaud G, Fahr A. Pulsed thermography for non-destructive evaluation and damage growth monitoring of bonded repairs. Compos Struct 2009;88:112–20. [16] Qing XP, Beard SJ, Kumar A, Hannum R. A real-time active smart patch system for monitoring the integrity of bonded repair on an aircraft structure. Smart Mater Struct 2006;5:66–73. [17] Ouinas D, Bouiadjra BB, Serier B. The effects of disbonds on the stress intensity factor of aluminium panels repaired using composite materials. Compos Struct 2007;78:278–84. [18] Ouinas D, Bachir Bouiadjra B, Achour T, Benderdouche N. Influence of disbond on notch crack behaviour in single bonded lap joints. Mater Des 2010;31:4356–62. [19] Alderliesten RC. Damage tolerance of bonded aircraft structures. Int J Fatigue 2009;31:1024–30. [20] Jones R, Callinan RJ, Aggraval KC. Analysis of bonded repairs to damage fiber composite structures. Engng Fract Mech 1983:37–46. [21] Bachir Bouiadjra B, Oudad W, Albedah A, Benyahia F, Belhouari M. Effects of the adhesive disbond on the performances of bonded composite repairs in aircraft structures. Mater Des 2012;37:89–95. [22] Ouinas D, Bachir Bouiadjra B, Himouri S, Benderdouche N. Progressive edge cracked aluminium plate repaired with adhesively bonded composite patch under full width disbond. Compos B Engng 2012;43:805–11. [23] Standard A. E647. Standard test method for measurement of fatigue crack growth rates. Annual book of ASTM Standards, 3; 2000. [24] Specification BP. Surface preparation of materials for adhesive bonding. BPS FW, 4352; 1972. [25] Baker AA, Jones R. Bonded repair of aircraft structures, vol. 7. Springer; 1988. [26] Nolting AE, Underhill PR, DuQuesnay DL. Variable amplitude fatigue of bonded aluminum joints. Int J Fatigue 2008;30:178–87.
Please cite this article in press as: Albedah A et al. Experimental analysis of the fatigue life of repaired cracked plate in aluminum alloy 7075 with bonded composite patch. Engng Fract Mech (2015), http://dx.doi.org/10.1016/j.engfracmech.2015.04.008