Strength and Damage Tolerance of Adhesively Bonded Joints: Stress-Based and Fracture-Based Approaches

Transcription

1 Strength and Damage Tolerance of Adhesively Bonded Joints: Stress-Based and Fracture-Based Approaches Hyonny Kim Assistant Professor School of Aeronautics and Astronautics, Purdue University presented at the FAA Workshop on Key Characteristics for Advanced Material Control September 16-18, 2003, Chicago, IL Funding Acknowledgements: FAA, NASA AGATE, WSU/NIAR, NASA Langley Research Center, ASEE, Purdue Research Foundation FAA Workshop Chicago

2 Outline Introduction Bonded Joint Stress Analyses in-plane shear loading combined shear + tension loading plastic strain-based ultimate load variable bondline thickness tension loading in general unbalanced single lap joint Damage Tolerance of Disbonded Joints Buckling/Disbond Growth experiments models New Directions FAA Workshop Chicago

3 Introduction General Joint Loading Shear Due to Fuselage Torsion Tension Due to Internal Pressure Segment of Joint Loaded by Biaxial Tension/Compression and Shear N y Bonded Doubler Compression Due to Fuselage Bending N x N xy N xy N x N y in-plane shear + tension produces adhesive shear stress FAA Workshop Chicago

4 Introduction Damage Tolerance: Buckling/Disbond Growth compression causes buckling subsequent disbond growth possible FAA Workshop Chicago

5 Introduction Damage Tolerance: Buckling/Disbond Growth Stringer-Stiffened Panel Wing Assembly P FAA Workshop Chicago

6 Overarching Objectives bridge knowledge gap between coupon data and full scale structural performance provide engineering and scientific community with advancements in: the mechanics of bonded joints damage tolerance design and analysis philosophy assist FAA and GA companies by providing research support FAA Workshop Chicago

7 In-Plane Shear Loaded Lap Joint Outer Adherend 5 z y 4 Adhesive x-z Shear for Aluminum Joint FEA SLBJ Theory N xy x N xy, MPa a τxz 3 2 Inner Adherend 1 in-plane shear load transfer across joint produces τ xz shear stress in adhesive y/c at x = L/2 Kim, H. and Kedward, K.T., Stress Analysis of Adhesive Bonded Joints Under In-Plane Shear Loading, Journal of Adhesion, Vol. 76, No. 1, 2001, pp FAA Workshop Chicago

8 In-Plane Shear Loaded Doubler inherently 2D problem 2 τ o xy 2 o λ τxy + C o = 0 τ a xz = t o τ o xy y τ yz shear peaks along x = 0, a τ a yz = t o τ o xy x z b y Bonded Doubler - Outer Adherend N xy 0 0 N xy Base Structure - Inner Adherend a x τ xz shear peaks along y = 0, b (not shown) Kim, H. and Kedward, K.T., Stress Analysis of Adhesive Bonded Joints Under In-Plane Shear Loading, Journal of Adhesion, Vol. 76, No. 1, 2001, pp FAA Workshop Chicago

9 Combined Loading: Shear + Tension 800 Von-Mises Based Failure Envelope Outer Adherend N (lb/in) y t a = N y N xy z y N xy N (lb/in) xy x N y Inner Adherend Kim, H. and Kedward, K.T., The Design of In-Plane Shear and Tension Loaded Bonded Composite Lap Joints, Journal of Composites Technology and Research, Vol. 24, No. 2, 2002, pp Invited paper to special issue in honor of Don Oplinger. FAA Workshop Chicago

10 Plastic Strain-Based Ultimate Load in-plane shear loading ductile adhesive joint carries greater load than elastic-limit design more conservative than Hart-Smith elastic-perfectly plastic model Shear Stress (MPa) adhesive test data from J. Tomblin, WSU 5 Data Fitting Curve Shear Strain Kim, H. and Lee, J., Adhesive Nonlinearity and the Prediction of Failure in Bonded Composite Lap Joints, Joining and Repair of Composite Structures, ASTM Special Technical Publication STP1455, submitted January FAA Workshop Chicago

11 Failure Prediction vs. Test Data box beam torsion lap shear coupon experiments conducted by Wichita State University (John Tomblin) Max. Shear Flow (lbf/in) * Experimental Data Purdue Analysis Maximum Shear Flow Constitutive behavior for 0.20 was not avail. Used 0.12 data, with reduced failure strain Bondline Thickness (in) Tomblin, J., Seneviratne, W., Kim. H., and Lee, J., Characterization of In-Plane Shear Loaded Adhesive Lap Joints: Experiments and Analysis, FAA Final Report, DOT/FAA/AR-03/21, May FAA Workshop Chicago

12 Variable Bondline Thickness (VBT) FAA Workshop Chicago

13 VBT: Increased Stress Due to Bondline Thinning Min. f Max. f f = t 1 / t unif Kim, H. The Influence of Adhesive Bondline Thickness Imperfections on Stresses in Composite Joints, Journal of Adhesion, Vol. 79, No. 7, 2003, pp FAA Workshop Chicago

14 Tension Loading of Unbalanced Single Lap closed-form stress analysis of generalized unbalanced joint Joint Specs: - glass/epoxy, - overlap 2c = 25.4 mm, -t o = 2.49 mm, t i = 2t o, t a = 0.33 mm FAA Workshop Chicago

15 Damage Tolerance of Disbonded Joints damage tolerance is safety concern thin flanges typically 1 mm (0.04 in.) susceptible to buckling + disbond growth disbond not easily detectable adhesive properties affect growth critically processing of joint very important FAA Workshop Chicago

16 Buckling/Debonding Experiments Construction: glass/epoxy & carbon/epoxy face sheets 6.3 mm foam core two halves secondarily bonded with PTM&W ES6292 paste adhesive FAA Workshop Chicago

17 Experimental Measurements Kim, H., Kwon, H., and Keune, J., Buckling Initiation and Disbond Growth in Adhesively Bonded Composite Flanges, 44th AIAA / ASME / ASCE / AHS / ASC Structures, Structural Dynamics, and Materials (SDM) Conference, April 7-10, 2003, Norfolk, VA FAA Workshop Chicago

18 Disbond Growth Model one-edge free flanges fracture-mechanics based compute strain energy release rate assumed post-buckled mode shape w( x, y ) = 1 2 [ c ( y b ) + c ( y b ) + c ( y b ) 2 c ( y b ) c n 2 ( y b ) n 2mπx ] 1 cos a m = 1: m = 2: FAA Workshop Chicago

19 Comparison of Model and Experiments Glass/Epoxy [0] 4 ; b* = 25.4 mm 0.8 mm Flanges Applied Farfield Strain, ε o Buckling - SS Buckling - C Disbonding - SS Disbonding - C Experiment - Buckling Experiment - Disbond G c = 578 J/m Disbond Length, a (mm) FAA Workshop Chicago

20 Non-Uniform Disbond Growth Front FAA Workshop Chicago

21 Strain Energy Release Rate Profile detailed view of G along disbond front predicts corner disbond initiation comparison with VCCT/FEA as programmed by R. Kreuger Extrapolated G peak ε o = 2,450 με FEA Model G (J/m 2 ) y/b* FAA Workshop Chicago

22 New Directions: Intrinsic Material Properties vs. Joint Behavior* understand relationship between intrinsic material properties vs. properties inferred from structural (joint) behavior intrinsic material properties should be independent of joint configuration, e.g., bondline thickness, mode of loading resolve differences observed from different test methods: tensile test dogbone napkin ring Krieger Gage / ASTM 5656 use nonlinear analyses and correlation with tests to extract TRUE, consistent, description of intrinsic material behavior Shear Stress (MPa) t a =2.08 mm t a =1.07 mm Shear Strain t a =0.33 mm * Adhesive test data from J. Tomblin and W. Seneviratne, WSU * this topic was raised at the ASTM Symposium on Joining and Repair of Composite Structures, March 2003, Kansas City, MO FAA Workshop Chicago

23 New Directions: Bonded Joint Impact (a) (c) (b) Damage on 38.1 mm Overlap Joint Due to 40 J Impact; (a) Impact Side; (b) Back Side, (c) C-Scan Showing Disbond Area basic guidance studies needed: characterize damage modes identify damage formation mechanisms determine governing parameters advanced: model predictions FAA Workshop Chicago

24 New Directions: Shear Buckling and Disbonding A B Disbond in Joint Between Fuselage Halves investigate in-plane shear buckling and subsequent disbond growth o N xy leads into combined compression + shear loading FAA Workshop Chicago

25 New Directions: Health Monitoring of Bonded Joints real time (or time interval based) monitoring of joint health existing: ultrasonic scanning common method for discrete interval inspection real-time implementation difficult electrical resistance external or embedded sensors / crack gages / fiber optics acoustic waves useful for internal, non-accessible locations correlate measurements with damage shape and location input for damage tolerance prediction models FAA Workshop Chicago

26 New Directions: Full-Scale Behavior Prediction Disbond P Path to get from Coupons & Elements to Full Scale Component Level Understand Experimental Data??? Other Factors Disbond in Joint Between Fuselage Halves A Refined Models: Adhesive Plasticity, Peel Stress Failure Under Multiaxial Stress, Fracture Mechanics B Side Load at 25% Chord o N xy Sandwich Construction - Halves Bonded Along Top and Bottom C-L Vertical Load develop predictive models: maintain balance between simplicity and complexity FAA Workshop Chicago