damage tolerance and durability of selectively stitched stiffened composite structures

Available online at www.sciencedirect.com International

www.elsevier.com/locate/ijfatigue

International Journal of Fatigue 30 (2008) 483–492

JournalofFatigue

Damage tolerance and durability of selectively stitchedstiffened composite structures

Yi Tan a, Guocai Wu a,b,*, Sung S. Suh a, Jenn Ming Yang a, H. Thomas Hahn c

a Department of Materials Science and Engineering, University of California, Los Angeles, CA 90095, United Statesb Cummins Inc. Columbus, Indiana, 47201, United States

c Department of Mechanical and Aerospace Engineering, University of California, Los Angeles, CA 90095, United States

Received 8 September 2006; received in revised form 19 April 2007; accepted 22 April 2007Available online 10 May 2007

Abstract

Through-the-thickness stitching is one of the promising approaches to improve the impact damage tolerance of stiffened compositestructures. In this research, the effect of selective stitching on impact damage tolerance and fatigue durability of stiffened composite pan-els was extensively investigated and compared with unstitched stiffened composite panels. Two-blade stiffened composite panels werefabricated by resin film infiltration technique. Impact damage was inflicted on the stiffened panels using a drop-weight with the impactenergy of 30 J from the skin-side over the stiffener and the flange. The experimental results indicate that selective stitching is an effectiveway to improve the compressive failure strength and fatigue strength of stiffened panels with both clearly visible flange damage (CVFD)and clearly visible stiffener damage (CVSD). The bulking and post-buckling finite element analyses were performed to predict the staticcompressive strength of the selectively stitched stiffened panels. A good agreement was obtained between analytical and experimentalresults. The failure behaviors of the undamaged, CVFD and CVSD panels were also investigated.� 2007 Elsevier Ltd. All rights reserved.

Keywords: Selective stitching; Stiffened composite panels; Impact damage tolerance; Finite element analysis

1. Introduction

Composite materials have been used in the aerospaceindustry over past thirty years. However, their utilization inthe primary load-bearing applications has been impaired bytheir low damage tolerance, i.e., high sensitivity to the out ofplane failures resulting from low interlaminar strength. Stitch-ing through the thickness has been demonstrated as an effec-tive approach to improve the delamination resistance [1–9].To develop the knowledge base required for certification ofstitched composite structures used in air transportation sys-tems, the compression behavior and constant-amplitude fati-gue behavior of both unstitched panel and fully stitched panelwith two stiffeners was extensively investigated in our previ-ous studies [10–14]. The results showed that the most ben-efit of stitching is manifested in the strength of barely

0142-1123/$ - see front matter � 2007 Elsevier Ltd. All rights reserved.

doi:10.1016/j.ijfatigue.2007.04.008

* Corresponding author.E-mail address: [email protected] (G. Wu).

visible flange damage panels as stitching prevents stiffenerseparation in impact and during compression loading.The stitching induces stiffener damage instead of stiffenerseparation when impact is inflicted on a stiffener. Althoughthe fully stitched panels have a better delamination resis-tance, higher compression strength and more tolerance offatigue than unstitched panels, the fully stitched panelsare more costly due to excessive amount of time and laborinvolved in stitching. The aim of the present study is to sys-tematically investigate the effect of selective stitching onimpact damage tolerance and fatigue durability of stiffenedcomposite panels.

2. Specimen preparation and experimental procedures

2.1. Materials

Hexcel type 282 AS4 (3k)/3501-6 carbon/epoxy wasused for this study. Hexcel type 282 is a plain weave fabric

mailto:[email protected]

484 Y. Tan et al. / International Journal of Fatigue 30 (2008) 483–492

that has the same number of warp and fill yarns, at 4.92yarns/cm. The yarn consists of AS4-3k fiber tow. TheAS4/3501-6 is a well-characterized composite and amena-ble to the resin film infiltration process.

The panels used in this study were designed to be a two-blade stiffened panel as shown in Fig. 1. Aside from its over-all dimensions, the panel is characterized by skin and stiffenthicknesses, stiffener spacing and height, and flange width.The skin and one half of the stiffener constitute a flange. Boththe length and width of the panel are 254 mm. The skin con-sists of 13 layers of fabric and the stiffener 12 plies. The lay-upsequences are as follows:

Skin: [0/90, ±45, 0/90, ±45, 0/90, ±45, 0/90, ±45, 0/90,±45, 0/90, ±45, 0/90]TStiffener: [±45, 0/90, ±45, 0/90, ±45, 0/90]SFlange: [0/90, ±45, 0/90, ±45, 0/90, ±45, 0/90, ±45,0/90, ±45, 0/90, ±45, 0/90, 0/90, ±45, 0/90, ±45,0/90, ±45]T

2.2. Manufacturing of stitched stiffened composite panels

The stitched stiffened panels were fabricated in the fol-lowing steps: (1) lay-up and stitching of skin and stiffener,(2) preparation of resin film and (3) resin film infusion inan autoclave. Stitching was done on a JUKI 200 indus-trial sewing machine with a working platform capableof handling up to 500 mm · 700 mm · 25 mm preform.A 1600-denier Kevlar thread was used for stitching anda 400-denier Kevlar thread for the bobbin. The stitchingspeed was maintained at 18 cm/min to minimize damageto the fibers near the stitch. The flange and stiffenerregion was stitched with 2.48 stitches per cm2 and 1.24stitches per cm2, respectively. The same Hercules AS4 car-bon fiber tows were used fill the gap at the joint betweenthe stiffener and the skin.

Prior to infusion, a film was prepared using 500 g ofresin, which is slightly more than required for 50% fiber

Fig. 1. The geometry and dimension of

volume content to ensure complete saturation of the drypreform. The refrigerated solid resin was crushed into fineparticle smaller than 5 mm in diameter. Then the particleswere spread evenly onto a mold and heated to 60 �C insidean autoclave for 20 min to form a flat film approximately5 mm thick. The stitched preform was placed into the moldcontaining the resin film, pressed down using a roller, andthe inner mold pieces were put into place. The assemblywas placed in an autoclave and cured using a cure cyclewhich was a slight modification of the manufacture’s rec-ommended cycle with the first isothermal dwell at 135 �Crather then the recommended 121 �C. The higher isother-mal condition was necessary to ensure low resin viscosityfor maximum penetration.

Nominal dimension of the manufactured panel were546.1 mm · 304.8 mm · 42.3 mm. Each panel was cut intotwo specimens using a high-pressure abrasive water-jet sys-tem. Overall dimensions of each specimen were 254.0 mm· 254.0 mm with two 25.4 mm stiffeners.

2.3. Experimental procedures

Impact testing was performed according to the ‘‘SACMARecommended Test Method for Compression After ImpactProperties of Oriented Fiber-Resin Composites, SRM 2-88’’ [15–18]. Impact loading was applied using a dropweight on a Dynatup 8250 drop-weight impact-testingmachine with a 4.3-kg indenter having a 12.7-mm diametertup. Before impact testing the panel was held between topand bottom end plates with a groove. To this end, 350 g ofCerrobend� was melt in a ceramic cup at 100 �C for 15 mininside an oven. The top and bottom plates were placedinside the oven and the molten alloy was poured into thegrooves in the heated plates. The plates were then removedfrom oven, placed on a Plexiglas support for better align-ment, and the composite panel was slipped into thegrooves. The entire panel assembly was then allowed tocool in the ambient temperature. The panel assembly wasplaced on the Dynatup base frame with the end plate

two-blade stitched stiffened panel.

Y. Tan et al. / International Journal of Fatigue 30 (2008) 483–492 485

providing support for the assembly. In this study, twoimpact locations were chosen. The one location was onflange, which is 25 mm far from the stiffener to insideand the center of the vertical direction. Because the inden-ter went through the flange of specimen, it showed a cleardamage around the indentation. So this damage was iden-tified as clearly visible flange damage (CVFD). The otherimpact location was directly on one of the stiffener, so thisdamage was named clearly visible stiffener damage(CVSD). For both cases, the impact point was on the skin-side opposite to the stiffener side of the panel, and the impactenergy is 30 J. Figs. 2 and 3 show an example of the panels

Fig. 2. The clearly visible flange d

Fig. 3. The clearly visible stiffener

with CVFD and CVSD, respectively. The X-ray radiogra-phy was used to detect the damage after impact.

Static compression tests were performed on a 500-kNInstron test frame at a displacement rate of 1.27 mm/min.The top plate of the panel assembly was secured onto aT-shaped platen with four socket cap screws. The positionof the screws ensured that the centroid of the panel coin-cide with the loading axis. The T platen was held in ahydraulic clamp. Securing the specimen to the top hydrau-lic clamp prevented the upper half of the assembly fromfalling upon failure. At the bottom the assembly was notsecured to the T platen.

amage (CVFD) after impact.

damage (CVSD) after impact.

Table 1Nominal maximum compressive stress levels for constant-amplitudefatigue tests of selectively, stiffened panels

Load level (MPa)

1 2 3 4 5

CVFD 135.3 130.1 119.7 111.9 103.6CVSD 113.3 108.9 100.2 93.7 78.8

Fig. 4. X-ray images of unstitched panels with CVSD. (a) Plan-view; (b)stiffener view of (A–A).


Omega 900 series data acquisition modules were used torecord strains, lateral displacement, top displacement, andapplied load. Strain gages were placed back to back at thecenter of the panel to capture a bifurcation point duringloading. The buckling load is taken as the load at whichthe load-strain relation deviates from linearity. Anotherstrain gage was placed on the stiffener to monitor the stiffenerstrain. In addition to measuring the local strains, a shadowMoire fringe technique was implemented to observe the glo-bal deformation and mode shapes during loading.

Constant-amplitude fatigue tests were performed at eachof the five highest stress levels of the TWIST spectrum at afrequency ranging from 0.5 to 1.2 Hz, depending on theload levels, to minimize a twist during the loading cycle.TWIST (Transport Wing Standard Test) consists of tenbasic flights, which are repeated to make a single blockof TWIST composed of 4000 flights. Each flight consistof ten stress levels (1–10) ranging from 0.222 to 1.6 of flightmean load such that single block of TWIST consist of398,665 cycles, and each block was repeated ten times, upto 40,000 flight hours. The nominal stress levels of con-stant-amplitude fatigue test are shown in Table 1. Becauseof the testing difficulty, any excursion into tension regimewas truncated. Hence, the resulting fatigue stress ratiowas infinity on all panels, but the stress amplitude wasmaintained according to each load level.

3. Results and discussion

3.1. Impact damage

A variety of damage modes were observed near theimpact region based on visual and X-ray radiographicinspection of the flange and stiffener impacted panels.The impact on the flange region for both unstitched andselectively stitched panels was readily noticeable since theimpactor had fully penetrated the flange and skin. How-ever, impact damage over the stiffener was visually unobserv-able in the skin region. Nevertheless, there were damages inthe stiffener and flange regions. For the unstitched CVSDpanels, delamination damage on the flange regions wasreadily observable with oval ring around the impact siteas shown in Fig. 4. Additional fiber breakage in the flangeregions on either side of a stiffener was also observed.Furthermore, the damage to the stiffener was shown withfiber breakage running through the stiffener heightdirection.

However, the damage conditions changed for the selec-tively stitched CVSD panels compared to the unstitchedpanels, especially in the flange region as shown in Fig. 5.The damage inflicted in the flange region is much reducedand in some cases, the fiber breakage/crack deflection wasobserved due to the influence of stitch element. The damageobserved in the stiffener was similar to that of unstitchedpanels. The impact damage size and types based on theX-ray radiographic and visual inspection are summarizedin Table 2. It can clearly be seen that selective stitching iseffective in reducing damages inflicted in both flangeimpacted and stiffener-impacted panels.

3.2. Compression behavior of selectively stitched stiffenedpanels

Fig. 6 shows that the average buckling strengths of undam-aged, CVFD, and CVSD panels are 86.3 MPa, 73.3 MPa, and66.7 MPa, respectively. A fairly consistent buckling strengthwas observed for both the undamaged and CVFD panels.A wider experimental scatter was observed for CVSD panels

Fig. 5. X-ray images of selectively stitched panels with CVSD. (a) Plan-view; (b) stiffener view of (B–B).

0102030405060708090

100110

66.773.3

86.3

CVSDCVFDUndamaged

Bulk

ing

Stre

ngth

, MPa

Test 1,4,7 Test 2,5,8 Test 3,6,9 Average

Fig. 6. Buckling strengths of selectively stitched panels.

020406080

100120140160180200220

132.0

158.7166.3

CVSDCVFDUndamaged

Failu

re S

treng

th, M

Pa

Test 1,4,7 Test 2,5,8 Test 3,6,9 Average

Fig. 7. Failure strengths of selectively stitched panels.


due to differences in stiffener damage incurred on each panel.The impact damage on one of the CVSD panels was moresevere and various than expected; this resulted in a much lowerbuckling strength. The average failure strengths of undam-

Table 2Drop-weight impact damage on unstitched and selectively stitched stiffened pa

Specimen # CVFD C

Damage diameter (mm) F

Flange Stiffener

Stitched 11.8 None NStitched 12.1 NoneStitched 12.4 NoneUnstitched 14.0 None 2Unstitched 14.2 None 2Unstitched 14.7 None 1

aged, CVFD, and CVSD panels are 166.3 MPa, 158.7 MPa,and 132.0 MPa, respectively, as shown in Fig. 7. It is evidentthat the CVSD panels have a lower compressive strength thanthe CVFD panels. For example, the compression strength ofCVFD panels is almost 5% lower than that of the undamagedpanels, however, in the case of CVSD panels, a 20% strengthreduction was observed. This again demonstrates that thestiffener is very important to support the specimen fromfailure.

The compressive failure behavior of the undamaged,CVFD and CVSD panels were also investigated. For selec-tively stitched stiffened panels without impact damage, the

nels

VSD

lange delamination (mm2) Fiber fracture (mm)

Flange Stiffener

one None 25.4350 5.0 20.0250 5.0 22.0890 5.0 19.0135 10.0 25.4884 10.0 25.4


compressive failure was occurred when both stiffeners inthe mid-region ruptured and the crack extended into theflange and skin regions as shown in Fig. 8. For CVFDpanels, the typical compressive failure occurred due to acrack extension from the impact site on the flange thatpropagated bilaterally, and failure was imminent whenthe crack had extended into a stiffener as shown inFig. 9. Fig. 10 shows the typical final failure behavior ofthe CVSD panels. The flange crack in skin-side extendedbilaterally from the stiffener impact site due to compres-

Fig. 8. The failure mode of undama

Fig. 9. The failure mode of selectiv

sive loading, and the damaged stiffener crippled once thecrack extended approximately 25.4 mm. Immediately, theload was transferred to the other stiffener and that stiffenerfractured. So it is clearly seen that stiffener separation fail-ure was effectively suppressed with selective stitching inflange region.

Compared to our previous study of the unstitched stiff-ened panels [10–12,14], it is noted that the benefit of selectivestitching in improving buckling strength is marginal. How-ever, the failure strength increased significantly as indicated

ged, selectively stitched panels.

ely stitched panels with CVFD.

Fig. 10. The failure mode of selectively stitched panels with CVSD.

Table 5Impact damaged finite elements properties

Type Loadingdirection

EX (Pa) EY (Pa) GXY (Pa) mmXY

Fiberbreakage

Tensile 1.0 1.0 1.0 0.0

Compressive 1.0 1.0 1.0 0.0Matrix crack Tensile 38.9E9 38.9E9 1.0 0.0

Compressive 38.9E9 38.9E9 1.0 0.0Delamination Tensile 32.5E9 32.5E9 1.0 0.0

Compressive 1.0 1.0 1.0 0.0

Stiffener Flange region, 19 plies (skin and

stiffener overlap)

a


in Table 3. In terms of fracture strength improvement com-pared to the unstitched panels, the failure strengthincreased by 6.8%, 15.9%, and 23.1% for undamaged,CVFD and CVSD panels, respectively. Residual strengthretention of selectively stitched CVSD and CVFD panelswith respect to undamaged panels is summarized in Table4. For CVFD panels, selectively stitched panels retain95% while unstitched panels only retain 88% of residualstrength. Furthermore, selectively stitched panel alsoretained higher buckling strength than that of theunstitched stiffened panels. The strength retention for selec-tively stitched and unstitched CVSD panels was 80% and69%, respectively. Therefore, selective stitching providesan effective way to improve the impact damage toleranceof stiffened composite panels.

Table 3A comparison of average buckling strength and failure strengths ofunstitched and selectively stitched panels

Properties Unstitched Selectively stitched

Undamaged CVFD CVSD Undamaged CVFD CVSD

Bucklingstrength

83.7 69.2 69.3 86.3 73.3 66.7

Failurestrength

155.7 136.9 107.2 166.3 158.7 132.0

Table 4Strength retentions of CVFD and CVSD panels compared to undamagedpanels

Panels CVFD CVSD

Bucklingstrength

Failurestrength

Bucklingstrength

Failurestrength

Selectivelystitched

0.85 0.95 0.77 0.80

Unstitched 0.83 0.88 0.83 0.69

Stiffener0.203mm gap

Flange

area,

b

Fig. 11. Finite element mesh of stitched stiffened panels (a) clearly visibleflange damage, damage area = 127 mm2; depth penetrated; (b) clearlyvisible stiffener damage.

Fig. 12. Buckling mode shapes of a stitched stiffened panel with CVSD.


3.3. Finite element analysis of selectively stitched stiffened

panels

Finite element buckling and post-buckling analyses ofselectively stiffened panels were performed using ABAQUSto support the experimental investigations by obtainingfull-field stresses that develop in the panels under compres-sion loading. A finite element model of a stiffened panelwas meshed with 8-node 3-D shell elements (S8R). In orderto improve the accuracy of the analysis, coupon tests ofunstitched were utilized for verification. For selectivelystitched panels, the properties of the skin and stiffenerregions were considered to those of unstitched values butthe flange regions were considered to be those of stitched val-ues. The displacement-loading compression was assumed inthat top and bottom edges were constrained to displaceuniformly in the loading direction. Furthermore, top and

bottom edges are assigned simply supported conditions. Asfor incorporating damages, those regions affected by animpact are assigned to have significantly lower materialproperties compared to the unaffected regions. The damageregion is defined according to the X-ray photos of specificpanels. The damaged properties are specified using modifiedproperties around the impact-affected zone. The modifica-tion of elastic properties of impacted panels in the regionswith the fiber breakage, delamination, and matrix crackwas discussed in detailed in [13]. Damage properties areassigned in relevance to the damage conditions as summa-rized in Table 5. These effects of varying degree are incorpo-rated in the stitched and unstitched stiffened panel to obtainthe analytical buckling and failure strengths for the impactdamaged panels.

The flange impact damage was modeled geometri-cally with 0.203 mm wide gap along the height of one of

0.60.70.80.91.01.11.21.3

Xo=107.2 MPa

lized

Stre

ss, S

(X/X

o)

100 1000 10000 1000000.20.30.40.50.60.70.80.91.01.11.21.3

xo=136.9 MPa

Nor

mal

ized

Str

ess,

S(X

/Xo)

Number of Cycles to Failure

Unstitched-CVFDSelectively stitched-CVFDBest-fit(unstitched-CVFD)Best-fit(Selectively stitched-CVFD)

Fig. 13. The maximum stress versus number of cycles for CVFD panels.


stiffeners are shown in Fig. 11a. Fig. 11b shows the finiteelement mesh for CVSD panels. For initial imperfectionshape required for the post-buckling analysis, a weightedfive buckling mode shape was used from a buckling analy-sis. A weight value of 0.005 is given for the lowest modeshape while the others are assigned a weight value of0.001. Fig. 12 shows the lowest buckling mode shapes forCVSD panels. Detailed information about finite elementmodels of stitched stiffened panels was reported in [11,13].

The analytical buckling and failure strength predictionsfor selectively stitched stiffened panels are presented inTable 6. The maximum point stress criterion was used topredict the post-buckling failure strength. In comparisonto experimental results shown in Table 3, the predictedbuckling and failure strength of the undamaged selectivelystitched stiffened panels are in a good correlation with theaverage experimental results. However, a negligible reduc-tion in buckling strength of selectively stitched CVFD pan-els was predicted from its undamaged counterparts becausethe mode shape deformations for circular penetration typedamage are not different from the undamaged case.The predicted failure strength of selectively stitched panelswith CVFD is considerably low due to stress concentrationaround the notch damage. The finite element analysisfor selectively stitched panels with CVSD is complicatedbecause there are two types of damage conditions inducedby impacting the stiffener, and various damage statesduring the failure. Five different damage states rangingfrom fiber breakage and crack through the stiffener to lesssevere stiffener damage with varying degrees of delami-nated regions were simulated [13]. The average predictedbuckling strength and failure strength for selectivelystitched CVSD panels are 67.1 MPa and 143.3 MPa,respectively, which are in a good agreement with the exper-imental results.

3.4. Fatigue behavior of selectively stitched stiffened panels

Compression-dominated constant-amplitude fatigue behav-ior for selectively stitched stiffened panels with CVFDand CVSD were studied to investigate the effectiveness ofstitching in preventing damage growth under fatigue load-ing. The test results showed that a large scattering in fatiguecycles for both CVFD and CVSD cases. This is mainly due tothe differences in the impact damage and infiltration quality,resulting in different failure modes or premature fractures.To assess the effect of stitching on the fatigue lives, the resultson selectively stitched panels were compared with unstitchedpanels in our previous studies [10,12,13]. Figs. 13 and 14show the maximum stress amplitude versus the number of

Table 6Analytical prediction results of selectively stitched stiffened panels

Buckling strength (MPa) Failure strength (MPa)

Undamaged CVFD CVSD Undamaged CVFD CVSD

85.7 85.0 67.1 160.0 126.6 143.3

cycles to failure and best-fit curves for both CVFD andCVSD panels. The stress was normalized using its averagecompressive strength (Xo) value of the correspondingunstitched CVFD and CVSD panels. As shown in Fig. 13,the fatigue strength improvement at 105 cycles is marginalfor selectively stitched panels compared to unstitched panelswith CVFD. At transition between high to low fatigue stresslevels, the best-fit curves indicate approximately 5% increasein fatigue strength for selectively stitched panels compared tothe unstitched panels, respectively. However, at higher stresslevels, a significant (>20%) increase in the fatigue strength isshown for the selectively stitched panel compared tounstitched panels. The selectively stitched panels performmuch better at high stress levels, but their performance atlower stress levels are comparable to the level of unstitchedpanels.

For CVSD case, the fatigue strength improvement at 105

cycles for selectively stitched panels is almost 15% bettercompared to unstitched panels. However, the selectivelystitched panels perform at least 20% better than unstitchedpanels at higher stress levels. The selectively stitched panelsare more effective in preventing excessive damage in theimpacted stiffener compared to the unstitched panels, asreported in static compression test results.

1 100 1000 10000 100000 10000000.20.30.40.5

Norm

a

Number of Cycles to Failure

Unstitched-CVSD Selectively stitched-CVSD Best-fit(Unstitched-CVSD) Best-fit(Selectively stitched-CVSD)

10

Fig. 14. The maximum stress versus number of cycles for CVSD panels.


4. Conclusions

The effect of selectively stitching on compression and con-stant amplitude fatigue behavior of stiffened panels withCVFD and CVSD were extensively investigated. The follow-ing conclusions were obtained:

1. Selectively stitched is an effective way to improve thecompression strength and fatigue strength for bothCVFD and CVSD cases.

2. The selectively stitched CVSD panels had a lower com-pressive strength than those with CVFD. The compres-sive strength of CVFD panels reduced about 5% fromthat of undamaged panels, and it reduced about 20%for CVSD panels.

3. The buckling and post-buckling analyses of finite ele-ment models were used to predict the bulking strengthand failure strength of selectively stitched stiffened pan-els. A good agreement was obtained between analyticaland experimental results.

Acknowledgements

This work is based on research supported by FAA throughAirworthiness Assurance Center of Excellence under GrantDTFA03-98-F-IA020. The authors wish to extend theirappreciation to Mr. Peter Shyprykevich, Federal AviationAdministration William J. Hughes Technical Center projectmanager for his support, guidance, and helpful comments.

References

[1] Dow MB, Smith DL. Damage tolerant composite materials producedby stitching carbon fibers. In: Proceedings of 21st internationalSAMPE technical conference; 1989. p. 595–605.

[2] Dow MB, Dexter HB. Development of Stitched, Braided and WovenComposite Structures in the ACT Program and at Langley ResearchCenter (1985–1997); Summary and Bibliography, NASA/TP-97-206234; November 1997.

[3] Sharma SK, Sankar BV. Effect of stitching on impact and interlam-inar properties of graphite/epoxy laminates. J Thermoplast ComposMater 1997;10:241–53.

[4] Khan MZS, Mouritz AP. Fatigue behavior of stitched GRPlaminates. Compos Sci Technol 1996;56:695–701.

[5] Dransfield K, Baillie C, Mai Y-M. Improving the delaminationresistance of CRRP by stitching-A review. Compos Sci Technol1994;50:305–17.

[6] Mouritz AP, Cox BN. A mechanistic approach to the properties ofstitched laminates. Compos Part A 2000;31:1–27.

[7] Mouritz AP, Bannister MK, Falzon PJ, Leong KH. Review ofapplications for advanced three-dimensional fiber textile composites.Compos Part A 1999;30:1445–61.

[8] Mouritz AP, Leong KH, Herszberg I. A review of the effect ofstitching on the in-plane mechanical properties of fiber-reinforcedpolymer composites. Compos Part A 1997;28:979–91.

[9] Larsson F. Damage tolerance of a stitched carbon/epoxy laminate.Compos Part A 1997;28:923–34.

[10] Hahn HT, Yang JM, Suh SS, Han NL. Design, Manufacturing,and Performance of Stiffened Composite Panels With and WithoutImpact Damage, FAA Report, DOT/FAA/AR-02/111; November2002.

[11] Suh SS, Han NL, Yang JM, Hahn HT. Compression behavior ofstitched stiffened panel with a clearly visible stiffener impact damage.Compos Struct 2003;62(2):213–21.

[12] Suh SS, Han NL, Yang JM, Hahn HT. Damage tolerance anddurability of stitched composites. In: Proceedings of the 17thAmerican society for composites technical conference, West Lafay-ette, IN; October 2002.

[13] Suh SS. Effect of stitching on resin film infusion and damage toleranceof stiffened composite structure, Ph.D. Dissertation, University ofCalifornia, Los Angeles; 2003.

[14] Suh SS, Han NL, Yang JM, Hahn HT. The effect of stitching oncompression behavior of stiffened composite panels. In: Proceedingsof the ASME-2001 international mechanical engineering conferenceand exposition (IMECE), NY; November 2001.

[15] SACMA Recommended Test Method for Compression After ImpactProperties of Oriented Fiber-Resin Composites, SRM 2-88.

[16] Jackson WC, Portanova MA. Out of plane properties. In: Mechanicsof textile composites conference, NASA Conference Publication 3311Part 2; 1995. p. 315–48.

[17] NASA/Aircraft Industry Standard Specification for Graphite Fiber/Toughened Thermoset Resin Composite Material, compiled byACEE Composites Project Office, NASA RP-1142; 1985.

[18] Standard Tests for Toughened Resin Composites, compiled by ACEEComposites Project Office, NASA RP-1092; 1982.

damage tolerance and durability of selectively stitched stiffened composite structures

Documents