UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Spring-in of angled thermoset composite laminates Albert, Carolyne I. 2001

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2001-0305.pdf [ 7.07MB ]
Metadata
JSON: 831-1.0078710.json
JSON-LD: 831-1.0078710-ld.json
RDF/XML (Pretty): 831-1.0078710-rdf.xml
RDF/JSON: 831-1.0078710-rdf.json
Turtle: 831-1.0078710-turtle.txt
N-Triples: 831-1.0078710-rdf-ntriples.txt
Original Record: 831-1.0078710-source.json
Full Text
831-1.0078710-fulltext.txt
Citation
831-1.0078710.ris

Full Text

S P R I N G - I N O F A N G L E D T H E R M O S E T C O M P O S I T E L A M I N A T E S by C A R O L Y N E I. A L B E R T B . A . S c . M e c h a n i c a l Engineering, Universite de Sherbrooke j , 1997 A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F A P P L I E D S C I E N C E i n T H E F A C U L T Y O F G R A D U A T E S T U D I E S Department o f Metals and Materials Engineering W e accept this thesis as conforming the required standard T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A September 2001 © Carolyne I. Alber t , 2001 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Afrfak J?jWa/Arfafe The University of British Columbia Vancouver, Canada DE-6 (2/88) A B S T R A C T A d v a n c e d f i b r e - r e i n f o r c e d thermoset p o l y m e r s are p r i m a r i l y used in the aerospace industry for their excellent m e c h a n i c a l properties and l o w weight . T o jus t i fy the use o f these materials , l o w p r o c e s s i n g and m a n u f a c t u r i n g cost must counterbalance their h i g h base cost. T h e necessity to p r o d u c e parts wi thin tight d i m e n s i o n a l tolerances in the aerospace industry constitutes a chal lenge to cost -effect ive m a n u f a c t u r i n g as p r o c e s s - i n d u c e d distortions are encountered i n p r o d u c t i o n . T h e current w o r k focuses on s p r i n g - i n , a change in angle o c c u r r i n g d u r i n g p r o c e s s i n g o f c u r v e d thermoset c o m p o s i t e parts. S p r i n g - i n is encountered for vir tual ly al l c u r v e d thermoset c o m p o s i t e parts and there is currently p o o r unders tanding o f the variables and m e c h a n i s m s that affect this p h e n o m e n o n . T h i s thesis examines the effect o f a n u m b e r o f design and process parameters o n s p r i n g - i n . T h e parameters s tudied i n c l u d e part geometry, l ay -up , tool material , tool surface c o n d i t i o n , cure c y c l e , and the use o f a rubber caul sheet. T h e results show that process parameters have a s ignif icant impact on spr ing- in , in some cases more than d o u b l i n g the value. C l o s e e x a m i n a t i o n o f the parts m a d e in this study reveals that s p r i n g - i n c a n be d e c o m p o s e d into two c o m p o n e n t s : a true corner c o m p o n e n t and a warpage c o m p o n e n t . T h e corner c o m p o n e n t is s h o w n to be m a i n l y the result o f c h e m i c a l and thermal strain anisotropy d u r i n g p r o c e s s i n g . T h i s c o m p o n e n t is repeatable wi th little var iabi l i ty for a g i v e n part des ign . Parameters af fec t ing the corner c o m p o n e n t are the init ial part angle and the part l a y - u p . T h e warpage c o m p o n e n t , on the other h a n d , is f o u n d to be d r i v e n b y m e c h a n i c a l interaction between the tool and the part. T h i s c o m p o n e n t is affected b y the laminate thickness , the f lange length, the cure c y c l e , the tool material , and the tool surface c o n d i t i o n . B a s e d o n the results o f this study, a s i m p l e set o f guidel ines is p r o p o s e d to reduce the variabi l i ty o f s p r i n g - i n in a p r o d u c t i o n e n v i r o n m e n t . - i i i -T A B L E OF CONTENTS Abstract » List of Tables vii List of Figures ix Nomenclature xiii Chapter 1: Introduction 1 1.1 Autoclave Processing Overview 1 1.1.1 Lay-up 2 1.1.2 Vacuum bagging 2 1.1.3 Cure '. •. 3 1.2 Material Property E v o l u t i o n D u r i n g C u r e 3 1.3 D i m e n s i o n a l Control 4 1.4 Research Objective and Outline 5 1.5 Tables 7 1.6 Figures 8 Chapter 2: Literature Review 11 2.1 M e c h a n i s m s C a u s i n g Manufacturing Distortions 11 2.1.1 Anisotropy 11 2.7.2 Tool-part Interaction 13 2.1.3 Laminate Asymmetry 14 2.1.3.1 Lay-up Asymmetry 14 2.1.3.2 Fibre Volume Fraction Gradients 14 2.2 Effect of Intrinsic and Extrinsic Parameters on Spring-in 14 2.2.7 Effect of Intrinsic Parameters 15 2.2.2 Effect of Extrinsic Parameters 16 2.3 Warpage 17 2.3.1 Parameters of Influence 17 2.3.2 Compounding Effect of Warpage and Spring-in 20 2.4 Tables 21 - iv -2.5 Figures 22 Chapter 3: Experimental - Total Spring-in 25 3.1 Description of Experiments 25 3.1.1 Tool and part geometries 25 3.1.2 Lay-up and processing 26 3.1.3 Measurement Techniques 27 3.1.4 Experiments 27 3.2 Anisotropy Calculations 28 3.3 Results 30 3.4 Observations 31 3.4.1 Effect of Intrinsic Parameters 32 3.4.1.1 Effect of Part Shape 32 3.4.1.2 Effect of Part Lay-up 32 3.4.1.3 Effect of Part Thickness 33 3.4.1.4 Effect of Flange Length 34 3.4.1.5 Effect of Part Angle 33 3.4.2 Extrinsic parameters 35 3.4.2.1 Effect of Cure Cycle 35 3.4.2.2 Effect of T o o l Surface 35 3.4.2.3 Effect of T o o l Material 36 3.5 Discuss ion 37 3.6 S u m m a r y 39 3.7 Tables 40 3.8 Figures 48 Chapter 4: Experimental - Spring-in Components 60 4.1 Separation of Spring- in into Warpage and Corner Components 60 4.2 Results 62 4.3 Observations 62 4.3.1 Effect of part shape on warpage 63 4.3.2 Effect of lay-up on warpage 63 - v -4.3.3 Effect of laminate thickness on warpage 64 4.3.4 Effect of flange length on warpage 64 4.3.5 Effect of initial part angle on warpage '• 65 4.3.6 Effect of cure cycle on warpage 65 4.3.7 Effect of tool surface on warpage 66 4.3.8 Effect of tool material on warpage 67 4.4 S u m m a r y 67 4.5 Tables 69 4.6 Figures 73 Chapter 5: Conclusions and Future Work 82 5.1 Future work 83 5.2 Tables 85 References 86 Appendix A: Variability 89 A . l Spring-in Variabi l i ty 89 A. 1.1 Effect of Moisture 89 A.l.2 Measurement Variability 90 A. 1.3 Variability Between Autoclave Runs 90 A . 2 Warpage Measurement Variabi l i ty 90 . A . 3 Tables 92 A . 4 Figures 94 Appendix B: Investigation of Viscoelastic Effects 95 B . l Objective 95 B .2 Experimental 95 B.3 Results and Discuss ion 95 B .4 Tables 97 B.5 Figures 98 - v i -LIST OF TABLES Table 1.1. T y p i c a l constituent materials of advanced thermoset composites [Agarwal and Broutman, 1990] 7 Table 2.1. T y p i c a l coefficients of thermal expansion for fibre materials [Agarwal and Broutman, 1990] 21 Table 2.2. T y p i c a l coefficients of thermal expansion for thermoset matrix materials [Agarwal and Broutman, 1990] 21 T a b l e 3.1. Descript ion of target cure cycles used 40 Table 3.2. Experiment 1 - parameters varied 40 Table 3.3. Experiment 1 - test matrix 41 T a b l e 3.4. Exper iment 2 - test matrix 41 Table 3.5. Experiment 3 - test matrix 42 Table 3.6. Experiment 4 - test matrix 42 T a b l e 3.7. Experiment 5 - test matrix 42 Table 3.8. Experiment 6 - test matrix 43 Table 3.9. Anisotropy component calculations - parameters and results 43 T a b l e 3.10. Experiment 1 - results 43 Table 3.11. Experiment 1 - first and second order coefficients 44 Table 3.12. Experiment 1 - first order effects 44 T a b l e 3.13. Experiment 2 - results 45 Table 3.14. Experiment 3 - results 45 Table 3.15. Experiment 4 - results 46 T a b l e 3.16. Experiment 5 - results 46 Table 3.17. Experiment 6 - results 46 Table 3.18. Lengths of influence, Ltot, when measuring spring-in for C and L-shaped parts. 46 Table 3.19. Effect of length of influence, Ltot, on spring-in. Parts are described in Tables 3.2 to 3.8 47 Table 3.20. Effect of cure cycle on spring-in. Parts are described in Tables 3.2 to 3.8 47 Table 4.1. Experiment 1 - spring-in component results 69 Table 4.2. Experiment 2 - spring-in component results 70 - vi i -Table 4.3. Experiment 3 - spring-in component results 71 Table 4.4. Experiment 4 - spring-in component results 71 Table 4.5. Experiment 5 - spring-in component results 72 Table 4.6. Experiment 6 - spring-in component results 72 Table 5.1. S u m m a r y of the contribution of the studied parameters on spring-in 85 Table A . 1. Experiment 1 - spring-in measurements before and after drying. T h e ranges shown represent 1/2 ( m a x i m u m measured value - m i n i m u m measured value). Parts are described in Tables 3.1 and 3.2 92 Table A . 2 . Variabi l i ty between autoclave batches. Duplicate parts were prepared in different autoclave runs. T h e ranges shown represent 1/2 ( m a x i m u m measured value -m i n i m u m measured value). Parts are described in Tables 3.1 to 3.4 92 Table A . 3 . Warpage component measurement variability for part W - 3 93 Table B . 1. Part description 97 Table B .2 . Heat treatment descriptions 97 Table B .3 . Spr ing- in results 97 - v i i i -LIST OF FIGURES Figure 1.1. Spr ing- in of a curved composite laminate 8 Figure 1.2. L a y - u p -example of prepreg stacking sequence, [0,45,-45,90]s 8 Figure 1.3. V a c u u m bag assembly 9 Figure 1.4. Schematic of the mechanical/physical behaviour of a curing matrix [adapted f rom Johnston, 1997] 9 Figure 1.5. Schematic of a typical spar/wing-skin assembly 10 Figure 2.1. Part angle definitions. Note that corner arc sector angle is geometrically equivalent to the part external angle 22 Figure 2.2. Schematic of the effect of anisotropy on spring-in. Radial (through thickness) strains are greater (negative) than circumferential (longitudinal) strains, leading to an increase in the corner arc sector angle, 9, or a reduction o f the part enclosed angle, 0' (see Figure 2.1 for angle definitions) 22 Figure 2.3. Tool-part interaction model proposed by N e l s o n [Nelson and Cairns , 1989] 23 Figure 2.4. Tool-part interaction model proposed by Ridgard [Ridgard, 1993]. (a) T o o l expands thermally on heating to prepreg cure temperature. Frict ion at the interface induces tensile fibre stresses in laminate, (b) Layer slippage relieves fibre stresses preferentially in layers furthest away from tool face. N o n - u n i f o r m stresses locked in when prepreg cures, (c) Laminate warps away f r o m tool face. 23 Figure 2.5. Schematic of the effect of asymmetry on warpage for a [0n,90n]t laminate 24 Figure 2.6. C o m p o u n d i n g effect of warpage and spring-in as proposed by Radford [Radford, 1995]. T h e author attributed the observed warpage to fibre volume fraction gradients resulting f r o m the bleeding of the top layer of the laminate 24 Figure 3.1. T o o l s , (a) 9 0 ° tool, (b) 4 5 ° tool 48 Figure 3.2. Part geometries, (a) C-shaped part, (b) L-shaped part 49 Figure 3.3. Schematic of cure cycles used. A more detailed description of the cure cycles is given in T a b l e 3.1 49 Figure 3.4. Digi ta l image analysis - cross section profile of part W - 3 . (a) Image before brightness and contrast adjustment, (b) Image after brightness and contrast adjustment 50 Figure 3.5. Schematic of flange angle measurements, (a) C-shaped part, (b) L-shaped part. Straight lines were fitted to two points on each segment 50 - ix -Figure 3.6. Digi ta l image analysis - example of good and bad part profile measurement locations. Section taken f rom the tool side edge of the cross section profile of part W - 3 shown in Figure 3.4.. 51 Figure 3.7. Composi te parts, (a) C-shaped parts, (b) L-shaped parts 52 Figure 3.8. Measurement of apparent spring-in of initially flat parts 52 Figure 3.9. Experiment 1 - measured spring-in. S o l i d and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. T h e ranges shown represent 1/2 ( m a x i m u m measured value - m i n i m u m measured value). Parts are described in T a b l e 3.10 53 Figure 3.10. Experiment 2 - measured spring-in. S o l i d and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. T h e ranges shown represent 1/2 ( m a x i m u m measured value - m i n i m u m measured value). Parts are described in T a b l e 3.13 53 Figure 3.11. Experiment 3 - measured spring-in. S o l i d and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. T h e ranges shown represent 1/2 ( m a x i m u m measured value - m i n i m u m measured value). Parts are described in T a b l e 3.14 54 Figure 3.12. Experiment 4 - measured spring-in. S o l i d and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. T h e ranges shown represent 1/2 ( m a x i m u m measured value - m i n i m u m measured value). Parts are described in Table 3.15 54 Figure 3.13. Experiment 5 - measured spring-in. S o l i d and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. T h e ranges shown represent 1/2 ( m a x i m u m measured value - m i n i m u m measured value). Parts are described in Table 3.16 55 Figure 3.14. Experiment 6 - linear fit of measured spring-in vs. part angle for a luminum and steel tooling. Anisotropy predictions were obtained using Equation 2.1, see Table 3.9. T h e ranges shown represent 1/2 (maximum measured value - m i n i m u m measured value). Parts are described in Table 3.17 55 Figure 3.15. Effect of part shape on spring-in. T h e ranges shown represent the m a x i m u m and m i n i m u m values for each data set. Parts are described in Tables 3.10 and 3.12 56 Figure 3.16. Effect of part lay-up on spring-in. T h e ranges shown represent the m a x i m u m and m i n i m u m values for each data set. Parts are described in Tables 3.10 to 3.16. 56 Figure 3.17. Effect of part thickness on spring-in. T h e ranges shown represent the m a x i m u m and m i n i m u m values for each data set. Parts are described in Tables 3.10, 3.14, and 3.15 57 - x -Figure 3.18. Effect of length of influence, Ltot, on spring-in. T h e ranges shown represent the m a x i m u m and m i n i m u m values for each data set. Parts are described in Tables 3.13 and 3.16 57 Figure 3.19. Experiments 4 and 5 - effect of part angle on spring-in. Spring- in of parts F-2, F-3 , W - 2 , and W - 3 . T h e ranges shown represents 1/2 ( m a x i m u m measured value - m i n i m u m measured value). Parts are described in Table 3.15 and T a b l e 3.16. Anisotropy predictions were obtained using Equation 2.1, see Table 3.9 58 Figure 3.20. Effect of cure cycle on spring-in. T h e ranges shown represent the m a x i m u m and m i n i m u m values for each data set. Parts are described in Tables 3.10, 3.13, 3.14, and 3.17 58 Figure 3.21. Effect of tool surface on spring-in. T h e ranges shown represent the m a x i m u m and m i n i m u m values for each data set. Parts are described in Tables 3.10 and 3.14 to 3.16 59 Figure 3.22. Effect of tool material on spring-in. T h e ranges shown represent the m a x i m u m and m i n i m u m values for each data set. Parts are described in Tables 3.10, 3.13 to 3.15, and 3.17 59 Figure 4.1. Flange warpage of part T S 2 - 4 73 Figure 4.2. Spring- in components (not to scale) 73 Figure 4.3. Experiment 1 - spring-in components. S o l i d and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. T h e ranges shown represents 1/2 ( m a x i m u m measured total spring-in value- m i n i m u m measured total spring-in value). Parts are described in Table 4.1 74 Figure 4.4. Experiment 2 - spring-in components. S o l i d and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. T h e ranges shown represents 1/2 ( m a x i m u m measured total spring-in value- m i n i m u m measured total spring-in value). Parts are described in Table 4.2 74 Figure 4.5. Experiment 3 - spring-in components. S o l i d and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. T h e ranges shown represents 1/2 ( m a x i m u m measured total spring-in value- m i n i m u m measured total spring-in value). Parts are described in Table 4.3 75 Figure 4.6. Experiment 4 - spring-in components. S o l i d and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. T h e ranges shown represents 1/2 ( m a x i m u m measured total spring-in value- m i n i m u m measured total spring-in value). Parts are described in Table 4.4 75 Figure 4.7. Experiment 5 - spring-in components. S o l i d and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. T h e ranges shown represents 1/2 ( m a x i m u m measured total spring-in value- m i n i m u m measured total spring-in value). Parts are described in T a b l e 4.5 76 Figure 4.8. Experiment 6 - spring-in components. S o l i d and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. T h e ranges shown represents 1/2 - xi -( m a x i m u m measured total spring-in value- m i n i m u m measured total spring-in value). Parts are described in Table 4.6 76 Figure 4.9. Effect of part shape on warpage. T h e ranges shown represent the m a x i m u m and m i n i m u m values for each data set. Parts are described in Tables 4.1 and 4.2 77 Figure 4.10. Effect of lay-up on warpage. T h e ranges shown represent the m a x i m u m and m i n i m u m values for each data set. Parts are described in Tables 4.1, 4.4, and 4.5. 77 Figure 4.11. Effect of laminate thickness on warpage. T h e ranges shown represent the m a x i m u m and m i n i m u m values for each data set. Parts are described in Tables 4 .1 ,4 .3 , and 4.4 78 Figure 4.12. Warpage component, Dqwarpage, vs. length of influence, Ltot, for all the parts presented in Tables 4.1 to 4.6 78 Figure 4.13. Effect of length of influence, Ltot, on warpage. T h e ranges shown represent the m a x i m u m and m i n i m u m measured values for each data set. Parts are described in Tables 4.3 and 4.4 79 Figure 4.14. Effect of initial part angle on warpage. T h e ranges shown represent the m a x i m u m and m i n i m u m measured values for each data set. Parts are described in Tables 4.1 and 4.4 to 4.6 79 Figure 4.15. Effect of cure cycle on warpage. T h e ranges shown represent the m a x i m u m and m i n i m u m measured values for each data set. Parts are described in Tables 4.1, 4.2, and 4.5 80 Figure 4.16. Effect of tool surface on warpage. T h e ranges shown represent the m a x i m u m and m i n i m u m measured values for each data set. Parts are described in Tables 4.1 and 4.3 to 4.5 80 Figure 4.17. Effect of tool material on warpage. T h e ranges shown represent the m a x i m u m and m i n i m u m measured values for each data set. Parts are described in Tables 4.1 to 4.3, and 4.6 81 Figure A . 1. Experiment 1 - measured spring-in before and after drying. Parts are described in Tables 3.1 and 3.2. T h e ranges shown represent 1/2 ( m a x i m u m measured value - m i n i m u m measured value) 94 Figure A . 2 . Spring- in variability between autoclave batches. Duplicate parts were prepared in different autoclave runs. Parts are described in Tables 3.1 to 3.4. T h e ranges shown represent 1/2 (maximum measured value - m i n i m u m measured value). ...94 Figure B . l . Experimental set-up 98 Figure B .2 . F lange A spring-in results 98 Figure B . 3 . Flange B spring-in results 99 - xi i -N O M E N C L A T U R E C, importance of parameter i on the experimental spring-in Cij importance of interactive effect between parameters / and j on the experimental spring-in CTE coefficient of thermal expansion L m total part length Lweb web length Lflange flange length V s h T total bulk resin volumetric cure shrinkage Xj coded level for parameter i (equal to +1 or -1) AT change in temperature Ad spring-in A6cl,mer corner spring-in component AOCTE thermal strains anisotropy spring-in component Adas c u r e shrinkage strains anisotropy spring-in component Ad/iangei warpage warpage spring-in component related to flange i A0mean mean experimental spring-in A6warpage warpage spring-in component Ad,otai total measured spring-in ai longitudinal coefficient of thermal expansion a, through thickness coefficient of thermal expansion 5 deflection (pi longitudinal chemical shrinkage - x i i i -<f>, through-thickness chemical shrinkage 9 part initial (external) angle or corner arc sector angle & part initial enclosed angle - x i v -C H A P T E R 1: I N T R O D U C T I O N F i b r e r e i n f o r c e d thermoset composi tes offer so me signif icant advantages over traditional materials, i n c l u d i n g h i g h spec i f i c strength, h i g h speci f ic stiffness, and g o o d durabi l i ty . T h e s e advantages make thermoset composi tes a g o o d c h o i c e in weight cr i t ical applicat ions t y pi ca l o f the aerospace industry. T h e i r h i g h base cost, h o w e v e r , remains an obstacle to a more extensive use o f these materials i n aerospace as w e l l as higher v o l u m e industries such as the automotive industry. T o jus t i fy the use of advanced thermoset composi tes , part manufacturers must o v e r c o m e the h i g h base cost b y a c h i e v i n g l o w m a n u f a c t u r i n g costs. A n obstacle to l o w cost m a n u f a c t u r i n g is the necessity to produce parts w i t h i n tight d i m e n s i o n a l tolerances. M a n u f a c t u r i n g distortions are u n a v o i d a b l e and must be compensated for in the m o u l d design . T h e p r o b l e m is that these distortions are not constant for al l parts processed on a g i v e n tool . D i s t o r t i o n variabi l i ty is p r o b l e m a t i c because it must be addressed o n a part-to-part basis , increas ing assembly t ime and cost. T h e w o r k presented in this thesis focuses o n spring-in, a reduct ion o f e n c l o s e d angles (F igure 1.1) encountered i n c u r v e d autoclave processed thermoset c o m p o s i t e parts. T h e variables and m e c h a n i s m s affec t ing s p r i n g - i n are e x a m i n e d and expla ined , and the present w o r k serves as a first step to a c h i e v i n g c o n t r o l l e d s p r i n g - i n in a p r o d u c t i o n environment . 1.1 AUTOCLAVE PROCESSING OVERVIEW A u t o c l a v e p r o c e s s i n g is often the m e t h o d o f c h o i c e for fabr ica t ing h i g h p e r f o r m a n c e thermoset c o m p o s i t e structures. T h e base material used in autoclave p r o c e s s i n g t y p i c a l l y c o m e s i n the f o r m o f prepreg, or sheets o f cont inuous h i g h performance fibres that have been impregnated wi th a contro l led a m o u n t o f part ial ly c u r e d thermosetting resin . Prepreg is avai lable i n rol ls o f u n i -Chapter 1 Introduction direct ional or fabric tape, c o m p o s e d o f various f ibre and matrix materials ( T a b l e 1.1). T h e material used in this study is a u n i d i r e c t i o n a l c a r b o n / e p o x y prepreg, T-800/3900-2 , m a n u f a c t u r e d b y the T o r a y C o m p a n y . T h r e e steps are i n v o l v e d in autoclave process ing : l ay -up , v a c u u m b a g g i n g , and cure. A br ief descr ipt ion o f each step is presented in the f o l l o w i n g sections. 1.1.1 L A Y - U P L a y - u p is the s tacking o f plies o f prepreg onto a tool (mould) o f the desired shape. T h e m o u l d , typica l ly m a d e o f a l u m i n u m , steel, invar, or a thermoset composi te , is in i t ia l ly coated with a release agent to ensure easy r e m o v a l o f the part f r o m the tool after process ing . A release p l y , e.g. f luor inated ethylene p r o p y l e n e ( F E P ) sheet, c a n be a d d e d o n top o f the release agent to further h e l p part r e m o v a l . Pl ies o f prepreg are cut into shape with special attention to the f ibre orientation. A major advantage o f the anisotropic nature of the prepreg material is that part properties can be tai lored to best fit the anticipated l o a d i n g c o n d i t i o n s . F o r this reason, the c h o i c e o f a f ibre orientation s tacking sequence is an integral part o f d e s i g n i n g c o m p o s i t e structures. In a lay -up , each p l y is c o m m o n l y referred to as lamina, w h i l e the stack o f pl ies , i .e. the part, is c a l l e d laminate. T h e prepreg plies are p l a c e d one at a t ime either b y h a n d or b y automated means onto the m o u l d f o l l o w i n g the d e s i g n e d s tacking sequence (Figure 1.2). E a c h p l y is pressed onto the m o u l d b y hand or with a r o l l to e l iminate entrapped air and wrinkles . It is also r e c o m m e n d e d to a p p l y v a c u u m every 1 to 4 plies for a more thorough c o n s o l i d a t i o n a n d e l i m i n a t i o n o f entrapped air. 1.1.2 V A C U U M BAGGING A f t e r l a y - u p , the laminate is c o v e r e d with a release p l y to prevent the part f r o m c h e m i c a l l y b o n d i n g to the v a c u u m a s s e m b l y d u r i n g cure . A layer o f breather fabric is p l a c e d over the tool , part, a n d release ply to e v e n l y distribute the v a c u u m a r o u n d the part as w e l l as to p r o v i d e a path f o r the r e m o v a l o f air Chapter 1 Introduction and volat i le gases d u r i n g cure . T h e l a y - u p and tool are put inside a v a c u u m b a g and v a c u u m is a p p l i e d . F i n a l l y , the w h o l e assembly is p l a c e d inside an autoclave. A n e x a m p l e o f a v a c u u m bag assembly is s h o w n in F i g u r e 1.3. 1.1.3 CURE C u r i n g is done inside the autoclave u s i n g a careful ly contro l led temperature and pressure c y c l e , or cure cycle, i n an inert gas such as ni t rogen. V a c u u m is mainta ined d u r i n g cure to c o n f o r m the part to the m o u l d as w e l l as to r e m o v e volati les d u r i n g cure. In some cases, v a c u u m is released once the pressure is suff ic ient to take over the funct ions o f the v a c u u m . A typical cure c y c l e m a y last more than 5 hours , d e p e n d i n g o n the size and the heating capaci ty o f the autoclave as w e l l as the thermal mass o f the tool and part. A baseline cure c y c l e is usual ly p r o v i d e d for a g i v e n prepreg b y the supplier , and m a y be adjusted b y the part manufacturer . A n ideal cure c y c l e w i l l m i n i m i z e the cure c y c l e t ime as w e l l as part defects (e.g. voids) , distortions (e.g. s p r i n g - i n , warpage) , and undesi rable residual stresses. 1.2 M A T E R I A L P R O P E R T Y E V O L U T I O N D U R I N G C U R E W h i l e f ibre properties essentially r e m a i n constant, resin properties change t remendously throughout the cure c y c l e . In the ini t ial u n c u r e d state, the resin consists o f m o n o m e r s w h i c h interact with each other through w e a k V a n der W a a l s b o n d s . D u r i n g cure, the resin p o l y m e r i z e s into a c r o s s l i n k e d network o f strong covalent b o n d s . T w o major events o c c u r d u r i n g resin p o l y m e r i z a t i o n : gelation and vi t r i f ica t ion ( F i g u r e 1.4). G e l a t i o n corresponds to the f o r m a t i o n o f an infini te network o f covalent b o n d s , w h i l e v i t r i f i ca t ion occurs w h e n the instantaneous glass transition temperature o f the resin reaches the part temperature [e.g. B e r g l u n d and K e n n y , 1991; Montser ra t , 1992]. B o t h of these events are a c c o m p a n i e d b y drastic changes in the p h y s i c a l and m e c h a n i c a l properties o f the resin. Chapter 1 Introduction The resin can be viewed as going through three states during cure: viscous, visco-elastic, and elastic. In reality there is overlap between these different states. At gelation (a = agei, where age] depends on the resin system), the polymer is transformed from a liquid to a gel and its mechanical behaviour changes from viscous (Region I in Figure 1.4) to viscoelastic, with short relaxation times (Region II). After gelation, the resin modulus begins to develop [e.g. Kim and Hahn, 1989]. Other changes that occur in the resin are a decrease in specific volume (cure shrinkage) and a significant increase in the instantaneous glass transition temperature (Tg). Upon vitrification, the mechanical behaviour changes from viscoelastic to effectively elastic (Region III). The neat resin is typically viscoelastic even in Region III, but relaxation times are much greater than the time scale of processing [Kim and White, 1996]. At vitrification, the cure reaction becomes diffusion driven and slows down considerably [Kenny etal., 1991]. 1.3 DIMENSIONAL CONTROL During autoclave processing of fibre-reinforced polymer composite structures there is an inevitable build-up of residual stresses, mainly due to differential thermal expansion between the fibres and the matrix, which often causes distortions of the fully cured structure. These distortions can be classified as warpage, or deviation from flatness, and spring-in, or a decrease of enclosed angles. Spring-in will in general occur in angled autoclave processed parts whereas warpage may or may not be an issue, depending on the design of the structure and the process conditions. Spring-in causes problems in assembly because of poor fit-up of mating sub-components of a structure (Figure 1.5), which results in increased assembly time and cost. A typical approach to dealing with spring-in is to compensate for it when designing the process tooling. This can be done using an experience based compensation factor or iterative modifications of the mould shape until the part matches the intended shape. The absolute magnitude of the spring-in, however, is difficult to Chapter I Introduction determine a priori and is furthermore variable in production. This variability reduces the benefits of a compensative tool design approach and variability must therefore be minimized. A solid understanding of the underlying mechanisms that drive spring-in is essential to minimize this variability. 1.4 RESEARCH OBJECTIVE AND OUTLINE The objective of the current research is to advance the understanding of the parameters and mechanisms that affect spring-in of angled thermoset composites during autoclave processing. The research presented in this thesis is organized as follows: Chapter 2 summarizes the literature found regarding the effect of design and process parameters on spring-in of angled thermoset composites as well as mechanisms that drive spring-in. A brief literature review of process-induced warpage is also presented. Chapter 3 presents a set of experiments that examine the individual and interactive effect of 5 design parameters (part shape, lay-up, flange length, thickness, and part angle) and 4 process parameters (tool material, tool surface, cure cycle and the use of a rubber caul sheet) on spring-in of curved thermoset composite parts. Chapter 4 shows that tool-induced warpage of nominally flat sections can contribute significantly to the apparent spring-in. In this chapter, the spring-in measurements of the parts presented in Chapter 3 are separated into corner and warpage components. It is also shown that these components are driven by two distinct mechanisms: material anisotropy and tool-part mechanical interaction, respectively. Chapter 5 provides a summary of the observations and conclusions obtained from this study. Based on these findings, a set of guidelines is proposed to minimize spring-in variability. Finally, recommendations for future work in this area are presented. Chapter I Introduction Appendix A addresses different types of variability encountered in the experiments. Appendix B investigates whether viscoelasticity is a concern for the material used in this study, in its fully cured state. The experimental results presented show that no significant viscoelastic relaxation occurs in fully cured parts within the temperature and time scales studied. Chapter 1 Introduction 1.5 T A B L E S Table 1.1. Typical constituent materials of advanced thermoset composites [Agarwal and Broutman, 1990]. Fibre materials Matrix materials Carbon Polyester Aramid (Kevlar™) Epoxies Chapter 1 Introduction 1.6 FIGURES Figure 1.1. Spring-in of a curved composite laminate. larnina stacking sequence Figure 1.2. Lay-up -example of prepreg stacking sequence, [0,45,-45,90]s. Chapter I Introduction vacuum bag breather sealant tape prepreg release ply tool/mould vacuum port Figure 1.3. Vacuum bag assembly. 220 200 180 160 140 120 100 80 60 4 0 -20" 0 I. Viscous ffl. Elastic instantaneous part temperature (TO autoclave temperature (Ta) \ degree of cure instantaneous T g --1 - 0.9 - 0.8 - 0.7 ire (a) - 0.6 U o -0.5 <u o in degi - 0.4 in degi - 0.3 Res - 0.2 - 0.1 r 0 50 100 150 200 250 Time (minutes) 300 350 400 Figure 1.4. Schematic of the mechanical/physical behaviour of a curing matrix [adapted from Johnston, 1997]. Chapter I Introduction fastener location Figure 1.5. Schematic of a typical spar/wing-skin assembly. - 10-C H A P T E R 2: L I T E R A T U R E R E V I E W 2.1 MECHANISMS CAUSING MANUFACTURING DISTORTIONS 2.1.1 ANISOTROPY M a t e r i a l anisotropy is u n d o u b t e d l y the best unders tood source o f s p r i n g - i n . U n l i k e isotropic materials, anisotropic materials do not retain their shape as they undergo v o l u m e t r i c changes such as c h e m i c a l shrinkage or thermal expansion/contrac t ion . F o r c u r v e d thermoset laminates, material anisotropy leads to a decrease in e n c l o s e d angles or an increase in corner arc sector angles (see F i g u r e 2.1 for angle def ini t ions) . D u r i n g cure, thermoset resins exhibi t c h e m i c a l shrinkage result ing f r o m the p o l y m e r i z a t i o n process . In b u l k resin the resul t ing c h e m i c a l strains are the same in al l direc t ions . In a f i b r e - r e i n f o r c e d thermoset c o m p o s i t e , however , the c h e m i c a l strains seen o n the laminate level are constrained to a certain degree b y the st iff f ibres , l e a d i n g to m u c h greater c h e m i c a l strains in the out -of -plane than in the in-plane direc t ions . S i m i l a r l y , because o f the s ignif icant ly l o w e r coeff ic ient o f thermal expans i on (CTE) o f the fibres ( T a b l e 2.1) c o m p a r e d to that o f the matrix ( T a b l e 2.2), thermoset laminates have greater out -of -plane than in-plane thermal contract ion d u r i n g the c o o l - d o w n stage o f a cure c y c l e . T h e thermal a n d cure shrinkage strains in c u r v e d laminates d u r i n g autoclave p r o c e s s i n g are thus s ignif icant ly greater in the radial ( through-thickness) di rec t ion than in the c i rcumferent ia l ( longitudinal ) d i r e c t i o n . F i g u r e 2.2 illustrates h o w material anisotropy leads to the spr ing- in o f c u r v e d laminates . A s i m p l e equat ion ( E q u a t i o n 2.1) has been p r o p o s e d for predic t ing the effect o f anisotropy on the spr ing- in o f a n g l e d laminates [ N e l s o n and C a i r n s , 1989; R a d f o r d and D i e f e n d o r f , 1993]. T h i s - 11 -Chapter 2 Literature Review equation accounts f o r the difference between process and ambient temperatures, the anisotropy o f thermal e x p a n s i o n and cure shrinkage, and the init ial part angle. A0 = A8CTE + A9CS =6 ((CTEl-CTEl)AT\0U,-#^ l + CTE, AT 1 + 0, (2.1) where 9 = ini t ia l part angle ( ° ) A6 = s p r i n g - i n ( ° ) A # C r £ = thermal expansion anisotropy c o m p o n e n t ( ° ) A&cs = cure shrinkage anisotropy c o m p o n e n t ( ° ) AT - change i n temperature ( ° C ) (pi = l o n g i t u d i n a l c h e m i c a l shrinkage (-) <p, = through-thickness c h e m i c a l shrinkage (-) CTEi = l o n g i t u d i n a l coeff ic ient o f thermal expansion ( m / m / ° C ) CTE, = through thickness coeff ic ient o f thermal expansion ( m / m / ° C ) T h e first term o f E q u a t i o n 2.1, A6CTE, is the thermal expansion anisotropy c o m p o n e n t , w h i c h is the result o f res idual stresses that d e v e l o p d u r i n g c o o l - d o w n , w h e n the laminate is f u l l y cured . T h e second term, A9Cs, o n the other hand, is the result o f cure shrinkage and is associated with stresses that b u i l d up earlier in the cure c y c l e , as the matrix p o l y m e r i z e s . W o r k presented in the literature indicates that in s o m e cases E q u a t i o n 2.1 provides a reasonable estimate for s p r i n g - i n [ R e n n i c k and R a d f o r d , 1996; H u a n g and Y a n g , 1997]. T h e cure shrinkage c o m p o n e n t , h o w e v e r , does not consider the m o d u l u s d e v e l o p m e n t o f the resin. R e s i n cure shrinkage o c c u r r i n g early in the cure c y c l e is assumed to have the same effect as shrinkage o c c u r r i n g w h e n the resin has hardened. Fur thermore , E q u a t i o n 2.1 does not account for extrinsic factors such as t o o l i n g effects. - 1 2 -Chapter 2 Literature Review 2.1.2 T O O L - P A R T INTERACTION T o o l - p a r t interaction has been p r o p o s e d as a m e c h a n i s m d r i v i n g both s p r i n g - i n and warpage. T h e concept o f tool-part interaction has been i n v o k e d to account for the warpage o f n o m i n a l l y flat s a n d w i c h panels [Pagliuso, 1982]. In that study, the author attributed the warpage to a t o o l - i n d u c e d stretching o f the tool side skin d u r i n g heat-up w h i l e the bag side skin was d a m p e n e d b y the c o m p l i a n t h o n e y c o m b core . A s imi lar m e c h a n i s m was p r o p o s e d as a contributor to s p r i n g - i n [ N e l s o n and C a i r n s , 1989], see F i g u r e 2.3. B a s e d o n the higher thermal exp an sio n o f the tool , the laminate is stretched b y the tool d u r i n g heat-up. T h e resul t ing residual stresses are re l ieved at the outer layers b y a c o m b i n a t i o n of resin f l o w a n d f ibre m o t i o n through the thickness at the corner, as w e l l as interply shear. T h e d e v e l o p e d stress gradients are b e l i e v e d to cause s p r i n g - i n w h e n they are r e l i e v e d after cure . Pressure has been int roduced as a necessary c o n d i t i o n for tool-part interaction to o c c u r [Ridgard , 1993]. T h i s m e c h a n i s m was suggested as a source o f warpage o f n o m i n a l l y flat parts as is illustrated in F i g u r e 2.4. It has been s h o w n that for s y m m e t r i c a l c ross -ply laminates, i .e. [0 n ,90 n ] s or [90 n ,0 n ] s , the stacking sequence has a strong effect o n warpage [ C a n n and A d a m s , 2001] . W h i l e parts processed with a 0 ° ply (fibres a l o n g the length o f the part) near the tool surface had signif icant warpage, parts processed with a 9 0 ° p l y (fibres a l o n g the w i d t h o f the part) near the tool surface had near zero warpage. T h i s observat ion indica ted that, in the latter lay -up , the tool-part shear strain was thus effec t ively dissipated in the b o t t o m 9 0 ° layer. T h i s result lead to the c o n c l u s i o n that for t o o l - i n d u c e d warpage o f flat parts to occur , the p l y located at the tool surface must be such that it is capable o f c a r r y i n g tensile stresses. - 13 -Chapter 2 Literature Review Tool-part interaction as a distortion mechanism has been reported by a number of other studies [e.g. Johnston, 1997; Yoon and Kim, 1997; Fernlund and Poursartip, 1999; Radford and Rennick, 2000; Twigg, 2001; Cann and Adams, 2001 ]. 2.1.3 L A M I N A T E A S Y M M E T R Y 2.1.3.1 Lay-up Asymmetry Asymmetry of the lay-up is a well-known cause of warpage. The effect of asymmetry on laminate warpage has been observed in a number of studies [Sarrazin et al., 1995; White and Hahn, 1993], and is the result of thermal and cure strain asymmetry through the thickness, which leads to warpage of the laminate (e.g. Figure 2.5). 2.1.3.2 Fibre Volume Fraction Gradients Fibre volume fraction (V f) gradients have been found to contribute to manufacturing distortions [Radford, 1990, 1993, and 1995]. In these studies, it was shown that the use of a bleeder fabric in processing leads to a resin poor region at the laminate top surface, near the bleeder, and a resin rich area adjacent to the tooling. The resulting V f gradient leads to an unsymmetrical laminate that warps when subjected to thermal and chemical strains. The effect of this mechanism has also been observed to affect spring-in in top bleed processes [Huang, 1997]. 2.2 E F F E C T O F INTRINSIC A N D EXTRINSIC P A R A M E T E R S ON SPRING-IN Many parameters are known or suspected to affect spring-in. They can be classified as intrinsic or extrinsic factors. In the scope of this work, intrinsic factors are defined as parameters related to part geometry and material properties, whereas extrinsic factors are defined as parameters related to tooling and processing. This chapter presents a summary of the literature found regarding the effect of intrinsic and extrinsic parameters and their effect on spring-in. - 1 4 -Chapter 2 Literature Review 2.2.1 E F F E C T O F INTRINSIC P A R A M E T E R S Initial part angle: As denoted in Equation 2.1, the initial part angle is known to have a significant effect on spring-in [Nelson and Cairns, 1989; Radford and Diefendorf, 1993]. Other published studies have also found that spring-in increases with increasing initial part (external) angle, 6 [e.g. Jain et al., 1997; Huang and Yang, 1997; Radford and Rennick, 2000]. Corner radius: There is disagreement over the effect of corner radius on spring-in. One study indicates that spring-in is smaller for smaller radii [Kim and Hahn, 1989]. Other studies have concluded that corner radius has little effect on the overall spring-in [Patterson et al., 1991; Wiersma et al., 1998]. Thickness: The effect of part thickness on spring-in is also unclear. While some studies have indicated that thinner parts have greater spring-in than thicker ones [Kim and Hahn, 1989; Ganley et al., 2000; Radford and Rennick, 2000; Zhu et al., 2001], other studies [Patterson et al., 1991; Jain et al., 1997; Wiersma et al., 1998] revealed little effect. Yet, in another study, doubling the part thickness increased the spring-in by over 20% [Stephan et al., 1996]. The differences between the findings of these studies suggest that the effect of thickness on spring-in may be sensitive to other interacting parameters. Lay-up: It has been reported that ply orientation has little effect on spring-in when comparing cross-ply, angle-ply and quasi-isotropic laminates [Patterson et al., 1991; Stephan et al., 1996]. It is generally agreed that spring-in is almost zero for 90° lay-ups, i.e. when all fibres are parallel to the elbow axis [Patterson et al., 1991; Stephan et al., 1996; Zhu et a l , 2001]. However, there is disagreement regarding the spring-in of 0° lay-ups (fibres perpendicular to the elbow axis). While one study [Stephan et al., 1996] indicated that 0° unidirectional laminates had the greatest spring-in, others [Patterson et al., 1991; Rennick and Radford, 1996; Johnston et al., 1996; Fernlund et al., - 1 5 -Chapter 2 Literature Review 2000; Zhu et al., 2001] revealed that the spring-in of unidirectional laminates was lower than that of multidirectional laminates. Stacking sequence: For symmetrical laminates, the stacking sequence has been reported to have little effect on spring-in [Patterson et al., 1991; Jain et al., 1997]. It was noted that when 0° and 90° plies were interchanged, spring-in remained virtually unaffected, provided that the lay-up remained symmetrical [Patterson et al., 1991]. Part shape: In an experimental and numerical study [Fernlund et al., 2000], C-shaped parts were found to have greater spring-in than L-shaped parts. This difference was explained in terms of a possible 'geometric locking' of the part on the male tool. 2.2.2 E F F E C T O F EXTRINSIC P A R A M E T E R S Cure cycle: The cure cycle has been found to affect spring-in. Published work has shown that if a 2-hold cure cycle is designed so that the resin gels during the first lower temperature hold, spring-in can be significantly increased compared to a single-hold cure cycle [Fernlund and Poursartip, 1999; Fernlund et al., 2000]. Tool surface: Numerical and experimental studies [Johnston et al., 1996; Fernlund et al., 2000; Fernlund and Poursartip, 1999] have shown that reducing the mechanical interaction between the part and tool at the tool-part interface reduces the spring-in angle significantly when compared to a situation where the part is bonded to the tool. Tool material: Numerical studies [Johnston et al., 1996; Wiersma et al., 1998; Zhu et al., 2001] have predicted that the tooling coefficient of thermal expansion (CTE) has a direct impact on spring-in. In these studies, parts prepared on male tooling had greater spring-in when processed on aluminum tooling (with a CTE of 25 um/m/°C) compared to lower CTE tooling. This trend is reversed for parts - 16-Chapter 2 Literature Review prepared o n f e m a l e t o o l i n g [ Z h u et a l . , 2001] . In another study, however , n o s ignif icant difference in spr ing- in was exper imenta l ly observed w h e n c o m p a r i n g a l u m i n u m , steel and c a r b o n - e p o x y composi te t o o l i n g [Jain et a l . , 1997]. Cure temperature: In an experimental study [Svanberg and H o l m b e r g , 2001] , the s p r i n g - i n o f parts processed u s i n g resin transfer m o u l d i n g ( R T M ) was f o u n d to be propor t ional to the cure temperature. T h i s observat ion is in agreement with E q u a t i o n 2.1, in w h i c h the predic ted thermal strains are propor t ional to the dif ference between cure and r o o m temperatures. Resul ts f r o m a n u m e r i c a l study, o n the other h a n d , indicated that cure temperature has little effect o n s p r i n g - i n o f autoclaved laminates [ Z h u et a l . , 2001] . Cooling rate: T h e same two studies [Svanberg and H o l m b e r g 2001 ; Z h u et a l . , 2001] reported that the c o o l i n g rate has little effect o n the spr ing- in o f resin transfer m o u l d e d and autoc laved parts. Pressure: In a n u m e r i c a l study, pressure was f o u n d to have little effect o n the s p r i n g - i n o f c ross -ply graphite /epoxy laminates [ Z h u et a l . , 2001] . 2.3 W A R P A G E 2.3.1 P A R A M E T E R S O F INFLUENCE T h e effects o f des ign and p r o c e s s i n g parameters on warpage have been addressed i n a n u m b e r o f p u b l i s h e d studies [Pagliuso, 1982; Sarrazin et a l . , 1995; F l a n a g a n , 1997; C a n n and A d a m s , 2001 ; T w i g g , 2001] . \ A s y m m e t r y o f the l a y - u p is k n o w n to contribute to warpage [ W h i t e a n d H a h n , 1993; Sar raz in et a l . , 1995]. W a r p a g e , h o w e v e r , is also observ ed in some sy mmet r i ca l laminates, i n d i c a t i n g that there are - 17-Chapter 2 Literature Review other important parameters affecting warpage [Flanagan, 1997; Cann and Adams, 2001; Twigg, 2001]. An experimental study showed that warpage of sandwich panels was greater when processed on aluminum compared to steel tooling [Pagliuso, 1982]. A similar observation was made in another study in which aluminum, steel and glass tooling caused substantial, intermediate and small warpage of flat symmetric laminates, respectively [Cann and Adams, 2001]. These experimental results indicate that warpage is greater for parts processed on a tool with greater CTE, supporting the proposed tool-part interaction mechanism presented in section 2.1.2. Results presented in a study done on asymmetrical laminates, however, did not show a difference in warpage between parts processed on different tooling materials [Sarrazin et al., 1995]. The authors believed that the effect of tooling material on warpage is related to a difference in tooling thermal conductivity, and that the tooling material did not affect warpage in their experiment because of a careful control of the part temperature. It is generally agreed that for symmetrical lay-ups, thinner parts exhibit more warpage than thick ones [Sarrazin et al., 1995; Flanagan, 1997; Cann and Adams, 2001; Twigg, 2001], and that warpage is greater for long parts than short ones [Flanagan, 1997; Twigg, 2001]. These observations support the proposed tool-part interaction mechanism. As discussed in section 2.1.3.1, the stacking sequence of symmetrical laminates was also found to influence warpage [Cann and Adams, 2001]. It was shown that for symmetrical cross-ply laminates, the position of the 0° plies (fibres along the length of the part) within the lay-up had a considerable impact on the resulting warpage. Parts with a 90° ply near the tool surface exhibited near zero negative (convex up) warpage, while those with a 0° ply near the tool surface had significant concave - 1 8 -Chapter 2 Literature Review up warpage. This observation shows that the presence of a layer capable of developing tensile stress must be near the tool surface for tool-induced warpage to occur. In a study by Flanagan [Flanagan, 1997], parts processed with an FEP sheet at the tool-part interface were observed to have less warpage than equivalent parts processed without FEP. Results presented in another study [Cann and Adams, 2001], however, indicated that no significant difference was found when comparing parts prepared on smooth versus abraded tools of various materials. Finally, no predictable tool surface effect was found in another study when comparing parts processed using release agent with our without an FEP sheet [Twigg, 2001]. It is generally agreed that warpage has a dependence on pressure. Parts processed under higher pressure have greater warpage than those processed under little or no pressure [Twigg, 2001; Flanagan, 1997]. From the proposed tool-part interaction mechanism, this can be explained by an increase of stresses transferred at the tool-part interface when processing under higher pressure. It has been reported that the cool-down pressure, however, has little effect on warpage [White and Hahn, 1993]. In that study, varying the cool-down pressure from 0.35 to 1.0 MPa had no noticeable effect on warpage. The effect of cure temperature on the development of residual stresses has also been addressed. Studies have shown that warpage can be significantly reduced by processing at a lower temperature for a longer time [White and Hahn, 1993; Sarrazin et al., 1995]. Processing at lower temperatures, however, requires longer curing times and a certain minimum temperature is needed for the crosslinking to reach completion. For a given cure temperature, it was shown that reducing the cure cycle time (thus the final degree of cure) can significantly reduce warpage [White and Hahn, 1993; Sarrazin et al., 1995]. Reducing the cure time, however, does not allow the curing reaction to complete which has a detrimental effect on the mechanical properties. A reduction in warpage was - 1 9 -Chapter 2 Literature Review also reported for asymmetric laminates when cooled down at a rate of 0 . 5 6 ° C / m i n compared to a rate of 5 .6°C/min [White and Hahn, 1993]. This was explained by viscoelastic stress relaxation being more significant for slower cool-down. 2.3.2 C O M P O U N D I N G E F F E C T O F W A R P A G E AND SPRING-IN T h e c o m p o u n d i n g effect of warpage and spring-in has been previously observed [ K i m and H a h n , 1993; Radford , 1995; H u a n g and Y a n g , 1997]. T h e first study [ K i m and H a h n , 1993] noted that spring-in measurements taken near the corner were smaller than those taken further along the flange. T h i s observation was explained by the presence of warpage, which was believed to be the result of fibre wrinkling. T h e two other studies [Radford, 1995; H u a n g and Y a n g , 1997] explained warpage as a result of volume fraction gradients resulting f rom an uneven bleeding of the resin. In one study [Radford, 1995], warpage was proposed as an explanation for the larger spring-in of parts prepared on convex tooling compared to that of parts prepared on concave tooling [Figure 2.6]. In the other study [Huang and Y a n g , 1997], volume fraction gradient induced warpage was proposed to explain the disparity between the calculated spring-in f r o m material anisotropy and experimental data. - 2 0 -Chapter 2 Literature Review 2.4 T A B L E S Table 2.1. Typical coefficients of thermal expansion for fibre materials [Agarwal and Broutman, 1990]. Fibre material Longitudinal C T E Transverse C T E (um/m/°C) (um/m/°C) Carbon and graphite -0.7 to -0.4 7 to 10 Aramid -2 60 Table 2.2. Typical coefficients of thermal expansion for thermoset matrix materials [Agarwal and Broutman, 1990]. Matrix material C T E (pm/m/°C) Polyester 55 to 100 Epoxies 45 to 65 - 21 -Chapter 2 Literature Review 2.5 FIGURES corner arc sector angle (6) part enclosed angle (9' = 180° - 6) Figure 2.1. Part angle definitions. Note that corner arc sector angle is geometrically equivalent to the part external angle. Cured Figure 2.2. Schematic of the effect of anisotropy on spring-in. Radial (through thickness) strains are greater (negative) than circumferential (longitudinal) strains, leading to an increase in the corner arc sector angle, 6, or a reduction of the part enclosed angle, 6' (see Figure 2.1 for angle definitions). - 2 2 -Chapter 2 Literature Review resin flow through the thickness reduces the arc length of the outside fibres, decreasing the stress Figure 2.3. Tool-part interaction model proposed by Nelson [Nelson and Cairns, 1989]. (C) Figure 2.4. Tool-part interaction model proposed by Ridgard [Ridgard, 1993]. (a) Tool expands thermally on heating to prepreg cure temperature. Friction at the interface induces tensile fibre stresses in laminate, (b) Layer slippage relieves fibre stresses preferentially in layers furthest away from tool face. Non-uniform stresses locked in when prepreg cures, (c) Laminate warps away from tool face. - 2 3 -Chapter 2 Literature Review 0°ply 90° ply low thermal and - v ^ chemical strains <C= o o o o o o o o o o o o o o o o o o l O O O O O O O O O O O O O O O O O o o o o o o o o o o o o o o o o o o l ^ > high thermal and < ^ chemical strains 0°ply „ "o—o o o o o o 6 " 5 rT ~ < T 0 0 ° o o o o o o o o ° o ° o 0 0 ^ o ° o o°jL-iL^ o o o o o , ^ p ° 0 o ° 90° ply \> Figure 2.5. Schematic of the effect of asymmetry on warpage for a [0n,90n]t laminate. V f gradient c o n c a v e t o o l i n g n d e c r e a s e d warpage V f gradient c o n v e x tool ing increased warpage Figure 2.6. Compounding effect of warpage and spring-in as proposed by Radford [Radford, 1995]. The author attributed the observed warpage to fibre volume fraction gradients resulting from the bleeding of the top layer of the laminate. - 2 4 -C H A P T E R 3: E X P E R I M E N T A L - T O T A L S P R I N G - I N A s discussed in Chapter 2, the effect of several parameters on spring-in have been examined in a number of published studies. It is, however, difficult to draw definite conclusions from these studies because of the large number of factors that vary between them. The research presented in this chapter aims to clarify ambiguities present in the literature regarding the effect of design and process parameters on spring-in and the major mechanisms that drive this type of distortion. 3.1 DESCRIPTION OF EXPERIMENTS The experimental work presented in this chapter examines the individual and interactive effects of five design parameters: part shape, lay-up, flange length, thickness, and part angle, and four process parameters: tool material, tool surface, cure cycle and the use of a rubber caul sheet. The study begins with an eight-factor fractional factorial designed experiment. This design allows the examination of eight parameters simultaneously to identify key parameters. Subsequently, a series of experiments were performed that focused on selected parameters in order to examine interactive effects in more detail. 3.1.1 T O O L AND PART GEOMETRIES This section describes the overall range of part and tool geometries used in this study. A more detailed description will be provided for each part in the experimental section (Tables 3.2 to 3.8). The tooling geometries used in this study are shown in Figure 3.1. The tools consist of solid blocks of aluminum or steel, two common tool materials in autoclave processing with dissimilar CT.Es and thermal masses. Each tool was machined to an angle of either 9 0 ° or 4 5 ° (one tool with 9 0 ° angles - 2 5 -Chapter 3 Experimental - Total Spring-in and two tools with 4 5 ° angles for both steel and a l u m i n u m ) , with a corner radius o f approximately 6 m m . T h e parts m a d e were either u n i d i r e c t i o n a l laminates [0]n, where all the fibres are p e r p e n d i c u l a r to the corner axis, or quasi - i sot ropic laminates [0,+45,-45,90] n s . T h e parts were C or L - s h a p e d wi th f lange lengths o f either 57 m m or 89 m m , and thicknesses of either 8 or 16 pl ies , i .e. 1.6 m m and 3.2 m m n o m i n a l thickness respect ively (F igure 3.2). T h e w e b length o f C - s h a p e d parts was 102 m m . S o m e parts were processed wi th a rubber caul (sheet) of 2.0 m m thickness w r a p p e d a r o u n d the part and tool . C a u l s are sometimes u s e d i n p r o c e s s i n g to even out the pressure dis tr ibution over the part. 3 .1.2 L A Y - U P AND PROCESSING Initially, the tool surfaces were c leaned wi th acetone to r e m o v e traces o f o i l and dirt. T w o or three coats o f release agent ( M u l t i s h i e l d ) were then a p p l i e d over the entire surface, a l l o w i n g 15 minutes for a i r - d r y i n g between each coat. F o r some parts (see T a b l e s 3.2 to 3.7), an F E P sheet was p l a c e d over the release agent before the prepreg was l a i d up o n the tool . A v a c u u m o f approximate ly -1 atmosphere (-30 i n H g ) was a p p l i e d for 15 minutes every 4 plies to consol idate the part and r e m o v e entrapped air. T h e r m o c o u p l e s were p l a c e d o n the parts, near the tool surface a n d near the bag surface. T h e parts were then c o v e r e d wi th an F E P sheet and breather fabr ic and p l a c e d inside a v a c u u m bag . O n c e the l a y - u p was complete , the assemblies (tools, parts, and v a c u u m bags) were p l a c e d inside an autoclave and c u r e d f o l l o w i n g a prescr ibed temperature and pressure c y c l e . T w o different cure cyc les were used : a o n e - h o l d and a t w o - h o l d c y c l e (F igure 3.3 and T a b l e 3.1). T h e o n e - h o l d cure c y c l e was des igned so that gelat ion of the resin occurs d u r i n g the temperature h o l d . T h e t w o - h o l d c y c l e was d e s i g n e d to achieve gelat ion pr ior to the s e c o n d heat-up (Figure 3.3). A f t e r p r o c e s s i n g , the parts were left to c o o l d o w n to r o o m temperature before they were debagged. - 2 6 -Chapter 3 Experimental - Total Spring-in 3.1.3 M E A S U R E M E N T TECHNIQUES T o o l angles were measured with a M i t u t o y o Pro 3600 digi tal protractor, wi th a measurement resolution o f 0 . 0 1 ° f o r the 9 0 ° tools, and a resolution o f 0 . 1 ° for the 4 5 ° tools . M e a s u r e m e n t s were taken 5 times at each locat ion o f interest, i .e. where the parts were processed . T h e s e measurements were repeatable w i t h i n ± 0 . 0 2 ° for the 9 0 ° tools, and w i t h i n + 0 . 0 ° for the 4 5 ° tools . A f t e r d e b a g g i n g , a p p r o x i m a t e l y 1 c m was t r i m m e d o f f the edge o f each part a l o n g the cross section o f interest u s i n g an I M E R C o m b i 250/500 d i a m o n d saw to el iminate the effect o f edge t h i n n i n g on the spr ing- in measurements . T h e surfaces were p o l i s h e d o n metal lographic p o l i s h i n g wheels wi th 600 grit paper. A H e w l e t t P a c k a r d ScanJet 4c scanner was used to create three images for each cross section at a resolut ion o f 600 d p i . T h e scanned images were digi ta l ly a n a l y z e d u s i n g the U l e a d PhotoImpact V e r s i o n 3.01 software. E a c h image was first o p e n e d with the digital analysis software (F igure 3.4a). T h e contrast and brightness settings were then adjusted to obtain a clear prof i le o f the part (F igure 3.4b). T w o point locations were recorded on each f lange and w e b b y p l a c i n g the cursor o n the points o f interest and n o t i n g the coordinates (F igure 3.5). T h e s e points were selected at the m a x i m u m distance f r o m each other, yet r e m a i n i n g w i t h i n the straight and f u l l thickness por t ion o f each f l a n g e / w e b to a v o i d the corner curvature as w e l l as the tapered f lange tip. C a r e was taken to c h o o s e points w i t h i n clear areas o f the part p r o f i l e (F igure 3.6). T h e part angles were determined f r o m these coordinates u s i n g the M i c r o s o f t E x c e l software. S p r i n g - i n was calculated as the difference between c o r r e s p o n d i n g part and tool angles. V a r i a b i l i t y i n part angle measurement is d iscussed in A p p e n d i x A . 3.1.4 EXPERIMENTS In experiment 1, eight parameters were investigated in a fract ional factorial exper iment ( T a b l e 3.2). T h e h i g h and l o w settings for each parameter were selected to reflect t y pi ca l process c o n d i t i o n s . - 2 7 -Chapter 3 Experimental - Total Spring-in A 2 8 " 4 resolut ion I V exper iment [ B o x et a l . , 1978] was used, a l l o w i n g the e x a m i n a t i o n o f the effect o f eight parameters s imul taneously in a s ignif icant ly reduced n u m b e r o f experiments . E a c h parameter was var ied between two values. A f u l l factorial two- level experiment w o u l d have required 2 8 (256) parts, whereas the fract ional factorial design used in this study reduced the required n u m b e r o f parts to 16. T h e detai led test matrix for this d e s i g n e d experiment is s h o w n in T a b l e 3.2 and T a b l e 3.3. Photographs o f four o f the parts f r o m experiment 1 are s h o w n in F i g u r e 3.7. In experiment 2, a total o f fourteen parts were prepared to e x a m i n e the effect o f part shape, tool material , cure c y c l e and a rubber caul sheet. T h e experiment is d e s c r i b e d in T a b l e 3.4. E x p e r i m e n t 3 consists o f fourteen parts e x a m i n i n g the effect o f part thickness , tool material and tool surface on s p r i n g - i n . T h e s e experiment is descr ibed in T a b l e 3.5. E x p e r i m e n t s 4 and 5 were d e s i g n e d to examine the effect o f f lange length f o r parts wi th ini t ial angles o f 9 0 ° and 4 5 ° , respect ively . T h e s e two experiments also p r o v i d e addi t ional i n f o r m a t i o n about the effect o f part thickness , part angle, l a y - u p , tool surface, and tool material . A d e s c r i p t i o n o f the parts prepared in experiments 4 and 5 is g i v e n in T a b l e 3.6 and T a b l e 3.7. S i x parts were prepared i n experiment 6 to examine the effect o f tool angle and tool material on spr ing- in . T h e exper iment is descr ibed in T a b l e 3.8. A n apparent s p r i n g - i n was measured for the ini t ia l ly flat parts, T A - 3 and T A - 6 , as indicated in F i g u r e 3.8. Parts T A - 3 and T A - 6 have the same total length as the other parts in this experiment . 3.2 ANISOTROPY C A L C U L A T I O N S A n i s o t r o p y is the best unders tood cause o f s p r i n g - i n . A s m e n t i o n e d in C h a p t e r 2, the effect of anisotropy o n s p r i n g - i n can theoretically be predic ted u s i n g a s i m p l e equat ion, E q u a t i o n 2.1. F r o m this equat ion, the expected a n i s o t r o p y - d r i v e n s p r i n g - i n was ca lculated f o r the parts prepared in the - 2 8 -Chapter 3 Experimental - Total Spring-in current study. T h e parameters used in the calculat ions and the results are s h o w n in T a b l e 3.9. T h e thermal e x p a n s i o n anisotropy s p r i n g - i n c o m p o n e n t L\6CTE, i e . the first term i n E q u a t i o n 2.1, develops d u r i n g c o o l - d o w n , w h e n the material is f u l l y cured . A 0 C r e represents a l o w e r b o u n d for the total anisotropy s p r i n g - i n c o m p o n e n t . T h e coeff ic ients o f thermal e x p a n s i o n u s e d to evaluate A6CTE are those of the f u l l y cured parts at 1 0 0 ° C , the average temperature d u r i n g c o o l - d o w n . T h e longi tudinal coeff ic ient o f thermal e x p a n s i o n , CTET, for the quasi - isotropic laminate was ca lculated f r o m the l a m i n a properties u s i n g laminate plate theory. T h e transverse coeff ic ient o f thermal e x p a n s i o n , CTE,, for the quasi - isotropic laminate was ca lculated f r o m the l a m i n a properties u s i n g the methods descr ibed b y P a g a n o [Pagano, 1974]. T h e cure shrinkage anisotropy c o m p o n e n t , A0CS, or the second term in E q u a t i o n 2.1, develops w h i l e the resin is c u r i n g . T h e evaluat ion o f A.6Cs is not as straightforward as A6CTE because it is dependent on the resin cure history. T y p i c a l total volumetr ic cure shrinkage ( V s h T ) values quoted in the literature for e p o x y resins range between 1% and 10% [Bogetti and G i l l e s p i e , 1992; K i m , 1996; Johnston, 1997]. P r i o r to s ignif icant m o d u l u s d e v e l o p m e n t the resin is u n a b l e to d e v e l o p residual stress, consequent ly o n l y a por t ion o f V s h T can contribute to s p r i n g - i n . V a l u e s f o r $ and 0, were estimated for a u n i d i r e c t i o n a l laminate b y s i m p l e rule o f mixtures , wi th the a s s u m p t i o n that a resin volumetr ic shrinkage ( V s h ) o f 2% contributes to s p r i n g - i n . T h e prepreg sys tem u s e d vitr if ies at a degree o f cure o f approximate ly 0.8 (80%) f o r the cure c y c l e s u s e d in this study. T h u s , o n l y about 20% of the total cure shrinkage occurs w h e n the resin has a s ignif icant stiffness. W i t h a total shrinkage o f about 10%, o n l y 2% occurs w h e n the resin has a h i g h m o d u l u s . T h e w i d e range o f values for cure shrinkage quoted in the literature is partly due to that some researchers measure the entire cure shrinkage, f r o m l i q u i d to s o l i d state, whereas others o n l y measure shrinkage w h e n signif icant m o d u l u s has d e v e l o p e d . T h e ranges o f <j>t and 0, for the quas i - i so t ropic parts were calculated u s i n g the same methods as were used to calculate CTE/ and CTE,. V a l u e s for A8CTE, &9CS, - 2 9 -Chapter 3 Experimental - Total Spring-in and AO were ca lcula ted f r o m these parameters for each l a y - u p u s i n g E q u a t i o n 2.1. T h e results are s h o w n in T a b l e 3.9. F r o m the results in T a b l e 3.9, it can be seen that anisot ropy-dr iven s p r i n g - i n , A9, is greater for the quasi - isotropic laminates . T h i s is because the difference between longi tudinal and transverse strains is greater for the quasi - i sot ropic laminates than for the u n i d i r e c t i o n a l ones. 3.3 RESULTS T h e results o f exper iment 1 are presented i n T a b l e 3.10 and F i g u r e 3.9. In F i g u r e 3.9, the predic ted spr ing- in , Ad, u s i n g E q u a t i o n 2.1 (see T a b l e 3.9) is s u p e r i m p o s e d o n the exper imenta l results. T h e f igure c lear ly shows that there are other sources o f s p r i n g - i n than those e n c o m p a s s e d in E q u a t i o n 2.1. F r o m the results i n T a b l e 3.10, the effect o f each parameter and the second-order effects (interactions between two parameters) were determined analyt ica l ly . A s s u m i n g that a l l third order and higher order effects are n e g l i g i b l e , the response equat ion, Y = f(Xj), for this f ract ional factorial experiment can be written as f o l l o w s : A6 = A9meim + 1 / 2 ( C , X , + C 2 X2 +.. . + C , 2 X , X2 +. . . + C 1 8 X , X 8 ) ( 2 ) where A 0 m e f l n = 1 . 0 8 ° Cy is the importance o f the c o r r e s p o n d i n g factor or c o m b i n a t i o n o f factors on spr ing- in ( T a b l e 3.11) X[ is the c o d e d level for the i t h parameter and is equal to +1 or -1 ( T a b l e 3.3) In other words , v a r y i n g each parameter f r o m its -1 to its +1 value w i l l change the s p r i n g - i n f r o m its m e a n value by the amount C , . - 3 0 -Chapter 3 Experimental - Total Spring-in In a fractional factorial design, each individual coefficient and each interaction has aliases. The calculated effect for each factor and interaction (C, and Q ) is identical to that of its aliases and the effects of aliased variables cannot be separated. Instead, the calculated effect of any parameter expresses the sum of the effects of the aliased variables. The alias patterns associated with the current design of experiment and the calculated coefficients C, ; are shown in Table 3.11. A summary of the first order effects is given in Table 3.12, assuming that third order and higher order effects are negligible. Note that the effects, C„ are average effects for the entire designed experiment. The interpretation of Table 3.12 is as follows, e.g., "the spring-in of quasi-isotropic parts is on average 0 . 2 6 ° greater than for unidirectional parts". The results of experiments 2 to 6 are presented in Tables 3.13 to 3.17, and Figure 3.10 to 3.14. Again, the spring-in predictions of Equation 2.1 are superimposed on the experimental results of experiments 2 to 6. 3.4 OBSERVATIONS From these experimental results it is evident that spring-in varies significantly as a function of design and process parameters and that there are mechanisms in addition to anisotropy of cure shrinkage and thermal expansion that contribute to spring-in. A possible mechanism is tool-part interaction. Based on the literature presented in section 2.1.2, mechanical tool-part interaction can be defined as the process in which the tool causes residual stresses to develop in the composite part through mechanical interaction at the tool-part interface. The driving force for this interaction is differential thermal expansion/contraction between the tool and the part in the plane of the interface. The magnitude of the mechanical interaction, and thereby the stresses built-up in the part, is dependent on the ability of the interface to transfer shear stress, and the in plane stiffness of the tool and the part. - 3 1 -Chapter 3 Experimental - Total Spring-in 3.4.1 E F F E C T O F INTRINSIC P A R A M E T E R S 3.4.1.1 Effect of Part Shape T h e results o f experiments 1 and 2 s h o w that part shape has little effect o n s p r i n g - i n , (see F i g u r e 3.15). In experiment 1 the overal l effect o f part shape IC/I is 0 . 0 2 ° ( T a b l e 3.12), w h i c h is less than the measurement accuracy ( A p p e n d i x A ) . E x p e r i m e n t 2 (Figure 3.10) shows s i m i l a r results with an average di f ference o f about 0 . 0 1 ° between C-parts and L-parts processed wi th a 2 - h o l d cure c y c l e . T h i s observat ion, h o w e v e r , disagrees with previous w o r k [ F e r n l u n d et a l . , 2001] i n w h i c h C - s h a p e d parts had greater s p r i n g - i n than L - s h a p e d parts b y about 30%. T h i s d i f ference in f i n d i n g s c o u l d be related to the different prepreg materials or other d i f f e r i n g parameters such as l a y - u p , cure cyc le , thickness, a n d f lange length between the two studies. 3.4.1.2 Effect of Part Lay-up F r o m the results o f experiments 1, 4 and 5 it is clear that quasi - isotropic parts have greater spr ing- in than their u n i d i r e c t i o n a l 0 ° counterparts (see F i g u r e 3.16). In experiment 1, quasi - i sot ropic parts have greater s p r i n g - i n than u n i d i r e c t i o n a l parts by an average o f 0 . 2 6 ° ( T a b l e 3.12). In exper iment 4, part W - 4 also has greater s p r i n g - i n than part W - 3 ( T a b l e 3.15) b y 0 . 4 3 ° , note, h o w e v e r , that these parts were processed wi th different t o o l i n g materials . S i m i l a r l y , part W - 4 has 0 . 4 4 ° greater s p r i n g - i n than part W - 2 , these parts d i f f e r o n l y in their shape and l a y - u p . B e c a u s e part shape was f o u n d to have little effect o n s p r i n g - i n , their dif ference in spr ing- in c a n be m a i n l y attributed to their dif ference in l a y - u p . F i n a l l y , i n exper iment 5, part F -4 , a quasi - isotropic l a y - u p , has greater s p r i n g - i n b y 0 . 4 1 ° than part F - 3 , an identical u n i d i r e c t i o n a l part ( T a b l e 3.16). O v e r a l l , the o bserv ed trend agrees with the results o f the anisotropy calculat ions ( T a b l e 3.9), w h i c h s h o w that anisotropy d r i v e n spr ing- in is greater for quasi - i sot ropic than for u n i d i r e c t i o n a l parts. H o w e v e r it is not c lear whether l a y - u p affects s p r i n g - i n through a m e c h a n i s m other than anisotropy. E x p e r i m e n t s 1 and 4 indicate that the effect o f l a y - u p is m a i n l y due to anisotropy because the observed effect is o f the same magni tude as the - 3 2 -Chapter 3 Experimental - Total Spring-in difference between the calculated anisotropy s p r i n g - i n for each l a y - u p ( T a b l e 3.9). O n the other hand, the s p r i n g - i n o f parts F -3 and F - 4 o f experiment 5 di f fer b y 0 . 4 1 ° , w h i c h is greater than the difference predic ted b y the anisotropy ca lcula t ion , 0 . 1 9 ° . L a y - u p m a y therefore affect spr ing- in through a m e c h a n i s m other than anisotropy. 3.4.1.3 Effect of Part Thickness E x p e r i m e n t 1 shows that 8-ply parts have o n average 0 . 0 9 ° greater s p r i n g - i n than 16-ply parts ( T a b l e 3.12). It is also c lear f r o m the results o f experiments 3 and 4 that 8-ply parts s p r i n g - i n cons iderably more than 16-ply parts (see F i g u r e 3.17). F r o m experiment 3, it is apparent that the effect o f thickness is s ignif icant ly greater w h e n n o F E P sheet is used. In this experiment , 8 -ply parts processed without F E P s p r i n g - i n o n average 0 . 4 5 ° more than 16-ply equivalent parts, c o m p a r e d to a difference o f 0 . 0 5 ° f o r the parts processed wi th F E P . Fur thermore , part W - 2 , a C - s h a p e d 8-ply part, has 0 . 7 5 ° more s p r i n g - i n than part W - l , a s imi lar 16-ply part. T h e s e two parts also d i f f e r in shape. S i n c e part shape was f o u n d to have little overa l l effect on spr ing- in in this study, the large di f ference in s p r i n g - i n can be m a i n l y attributed to the dif ference in thickness . N o t e that the di f ference in s p r i n g - i n between parts W - l and W - 2 is larger than the thickness effect seen in experiment 3. It is also worth n o t i n g that parts W - l and W - 2 have longer flanges than the parts f r o m experiment 3, i n d i c a t i n g that the effect o f part thickness increases wi th increasing f lange length. T h e s e observations are consistent with the p r o p o s e d m e c h a n i c a l tool-part interaction m o d e l . Increasing the f lange length increases the interface surface between the tool and the part, w h i c h w i l l increase the stresses b e i n g transferred to the part and the stress gradients f o r m i n g through the part thickness. T h e magni tude o f these stresses is also related to the tool surface c o n d i t i o n . 3.4.1.4 Effect of Flange Length W h i l e exper iment 1 indicates that f lange length has n o effect on s p r i n g - i n , the other experiments lead to a different c o n c l u s i o n . T h e s p r i n g - i n o f parts W - l , W - 2 , and W - 3 are n o w c o m p a r e d to parts R C - 6 , - 3 3 -Chapter 3 Experimental - Total Spring-in RC-8, TS-4, TS2-1, TS2-4, and TS2-5 with shorter flange length (Tables 3.13 to 3.15). To compare the effect of length for these parts, the total length of the "flanges" corresponding to a given corner (L,ol) was determined. For an L-shaped part, L,„, is equal to twice the flange length (Lflange), while for a C-shaped part, L,„, is equal to the sum of the web length (Lweh) and the flange length (Figure 3.2). Table 3.18 shows the values of L,„, for the 9 different C and L parts that are compared to examine the effect of flange length. Comparison of parts W - l , W-2, and W-3 with their shorter counterparts shows that spring-in increases with increasing L,„, (Table 3.19 and Figure 3.18). Furthermore, the effect of flange length is greater for 8-ply parts than for 16-ply parts. 3.4.1.5 E f f e c t o f P a r t A n g l e Experiments 1 and 6 show that spring-in is approximately proportional to the part angle. In experiment 1, 90° parts have greater spring-in than 45° parts by an average of 0.28° (Table 3.12). The results of experiment 6 show a direct correlation between part angle and spring-in. Figure 3.14, a plot of measured spring-in angle vs. part angle, shows that for aluminum and steel tooling, the relationship between part angle and spring-in is similar to that predicted from anisotropy calculations. However, while the anisotropy calculations done with Equation 2.1 give a zero spring-in for flat parts, the two data sets in Figure 3.14 (for aluminum and steel tooling) do not have zero y-intercepts. This is due to the presence of warpage, which translates into an apparent spring-in for the flat parts (Figure 3.8). Four parts from experiments 4 and 5 confirm this observation (Figure 3.19). Parts F-2 and W-2, processed on steel tooling, are identical with the exception of their initial angle. The same thing is true for parts F-3 and W-3, processed on aluminum tooling. Figure 3.19 shows that extrapolation of a straight line fit to each pair of data does not give a zero y-intercept. Extrapolation gives a predicted 0.77° and 0.91° spring-in for flat parts processed on steel and aluminum tooling respectively. This observation clearly shows that Equation 2.1 is not sufficient to predict spring-in and that there is a significant source of spring-in other than anisotropy. The slopes of these experimental data curves - 3 4 -Chapter 3 Experimental - Total Spring-in are also m o r e than 5 0 % greater than the anisotropy predict ions ca lculated u s i n g E q u a t i o n 2.1. T h e m a i n di f ference between F i g u r e 3.14 and F i g u r e 3.19 is that the parts i n F i g u r e 3.19 were processed u s i n g a 2 - h o l d cure c y c l e , c o m p a r e d to a 1-hold c y c l e for the parts in F i g u r e 3.14. 3.4.2 EXTRINSIC PARAMETERS 3.4.2.1 Effect of Cure Cycle In experiment 1, parts processed wi th a 2 - h o l d cure c y c l e have, o n average, 0 . 0 7 ° higher spr ing- in than parts c u r e d u s i n g a 1-hold c y c l e . F o r the C - s h a p e d parts i n experiment 2, the dif ference is 0 . 4 9 ° . T h e same effect is seen w h e n c o m p a r i n g other parts f r o m experiments 1, 2, and 6 (Figure 3.20). T h e s e results s h o w that the cure c y c l e has a large effect o n s p r i n g - i n . U s i n g a cure c y c l e where the part gels before it reaches the f i n a l cure temperature, spr ing- in can be s ignif icant ly greater than with a cure c y c l e where the part gels at the f i n a l cure temperature. T h e c o o l - d o w n rate was not altered in the present study, but based o n a viscoelast ic i ty test done o n f u l l y cured parts at elevated temperatures ( A p p e n d i x B ) , the c o o l - d o w n rate is not b e l i e v e d to s ignif icant ly affect s p r i n g - i n . 3.4.2.2 Effect of Tool Surface E x p e r i m e n t s 1, 3, 4 and 5 s h o w that the effect o f tool surface is substantial ( F i g u r e 3.21). Parts processed without an F E P sheet have cons iderably greater spr ing- in than those processed wi th an F E P sheet. In exper iment 1, tool surface has the largest effect o f a l l the studied parameters. In this experiment , parts processed without an F E P sheet have on average 0 . 3 6 ° greater s p r i n g - i n than those processed w i t h an F E P sheet. In experiment 3, parts processed without an F E P sheet have o n average 0 . 5 6 ° greater s p r i n g - i n than those processed wi th an F E P sheet. Further e x a m i n a t i o n o f the results o f experiment 3 ( F i g u r e 3.11 and F i g u r e 3.21) reveals that the effect o f tool surface is m o r e important for 8-ply parts than for 16-ply parts. F o r 8-ply parts, parts processed without F E P have 0 . 7 5 ° m o r e - 3 5 -Chapter 3 Experimental - Total Spring-in s p r i n g - i n than those processed with an F E P sheet. F o r the 16-ply parts, those processed without an F E P sheet have higher s p r i n g - i n than those processed with an F E P sheet b y an average o f 0 . 3 5 ° . It is thus evident that the effect o f tool surface increases as the part thickness decreases. In experiments 4 and 5, the effect o f tool surface is even greater than o b s e r v e d in the other experiments . Part W - 4 , processed without an F E P sheet, has 1 . 3 5 ° m o r e s p r i n g - i n than part W - 5 , a replicate part processed with an F E P sheet. S i m i l a r l y , part F -2 , processed without an F E P sheet, has 0 . 8 4 ° m o r e s p r i n g - i n than part F - l , an identical part processed with an F E P sheet. It is worth n o t i n g that the parts i n experiments 4 and 5 have f lange lengths o f 89 m m , c o m p a r e d to 57 m m for the parts in experiment 3. In exper iment 1 the f lange lengths were either 57 m m or 89 m m w i t h thicknesses of 8 and 16 plies c o m p a r e d to 8 plies for parts W - 4 , W - 5 , F - l and F - 2 . T h e s e results indicate that the effect o f tool surface is not o n l y dependent on the part thickness, but also o n the f lange length. T h e observed effect o f tool surface and its interaction wi th f lange length support the m e c h a n i c a l tool-part interaction m e c h a n i s m because a larger interface surface w i l l increase the stresses transferred to the part. T h e relat ionship between tool surface, cure c y c l e , part thickness and f lange length is addressed in m o r e detai l in C h a p t e r 4. 3.4.2.3 Effect of Tool Material E x p e r i m e n t 1 indicates that the c h o i c e o f tool material has little overal l effect o n s p r i n g - i n ( T a b l e 3.12). E x p e r i m e n t s 2, 3 and 6 lead to different observations (Figure 3.22). In exper iment 2, the tool material has little effect on s p r i n g - i n f o r parts processed with a 1-hold cure c y c l e . C o n v e r s e l y , the effect o f tool material is m o r e s ignif icant for the parts processed with the 2 - h o l d cure c y c l e , with a l u m i n u m t o o l i n g g i v i n g an average o f 0 . 3 3 ° higher spr ing- in than steel t o o l i n g . In experiment 6, parts processed o n a l u m i n u m t o o l i n g have on average 0 . 1 3 ° greater s p r i n g - i n than parts processed on steel t o o l i n g ( F i g u r e 3.14). T h e parts processed o n an a l u m i n u m tool in exper iment 3 have on average 0 . 0 9 ° greater s p r i n g - i n than those processed on a steel tool . A c loser l o o k at the results of - 3 6 -Chapter 3 Experimental - Total Spring-in experiment 3 reveals that for the 8 p l y parts processed without an F E P sheet, parts m a d e o n an a l u m i n u m tool have greater s p r i n g - i n than those made on a steel tool b y an average o f 0 . 2 7 ° . T h e effect o f tool material , however , is not present w h e n an F E P sheet is used or w h e n the thickness of the part is 16 pl ies . T h e s e results indicate that the effect o f tool material is dependent not o n l y on the cure c y c l e , but also on the tool surface as w e l l as the part thickness , w h i c h is consistent with the p r o p o s e d m e c h a n i c a l tool-part interaction m e c h a n i s m . W h e n c o m p a r i n g parts W - 2 a n d F - 2 wi th parts W - 3 a n d F - 3 , respect ively , however , tool material appears to have little effect o n s p r i n g - i n (see F i g u r e 3.22). N o t e that these pairs o f parts also di f fer in shape, a l though part shape was f o u n d to have little effect o n s p r i n g - i n . T h i s observat ion is not f u l l y unders tood and is p o s s i b l y i n f l u e n c e d b y the presence o f corner w r i n k l i n g i n some o f the parts, process-related var iabi l i ty s u c h as an uneven di spers i on o f release agent o n the tool surface, and p o s s i b l y b y the dif ference i n part shapes, a l though the mechanism(s ) i n v o l v e d are not clear. Nonetheless , as e v i d e n c e d in experiments 2 and 3, a l u m i n u m t o o l i n g c a n lead to greater spr ing- in than steel tool ing , and this effect appears to be related to the part thickness as w e l l as the tool surface c o n d i t i o n . T h e use o f a rubber c a u l is f o u n d to have little effect o n s p r i n g - i n (F igure 3.10). 3.5 D I S C U S S I O N T h e s e results s h o w that s p r i n g - i n is greatly affected by a n u m b e r o f intrinsic and extrinsic parameters. T h e s e parameters inf luence s p r i n g - i n through two distinct m e c h a n i s m s : laminate anisotropy and m e c h a n i c a l tool-part interaction. L a m i n a t e anisotropy is the m a i n m e c h a n i s m through w h i c h part l a y - u p and part ini t ial angle affect s p r i n g - i n . Q u a s i - i s o t r o p i c laminates have greater spr ing- in than u n i d i r e c t i o n a l 0 ° laminates because o f the greater d i f ference between the through-thickness and longi tudinal thermal e x p a n s i o n and cure - 3 7 -Chapter 3 Experimental - Total Spring-in shrinkage. S p r i n g - i n was also f o u n d to be approximately proport ional to the ini t ia l angle wi th a slope s imi lar to that predic ted b y E q u a t i o n 2.1. M e c h a n i c a l tool-part interaction, o n the other hand, is the m a i n m e c h a n i s m through w h i c h part thickness, f lange length, cure c y c l e , tool surface and tool material affect s p r i n g - i n . T h e s e parameters are either related to the d r i v i n g force for the m e c h a n i c a l tool-part interaction, the abil i ty of the interface to transfer stress, or the abili ty o f the part to d e v e l o p stresses through the thickness . F l a n g e length, tool surface and tool material affect the stresses transmitted f r o m the tool to the part at the tool-part interface, as the thermal expansi o n o f the tool causes stretching o f the part d u r i n g heat-up. T h e s e stresses w i l l be greater with an a l u m i n u m tool than a steel tool because o f a greater thermal e x p a n s i o n . T h e use o f an a l u m i n u m tool was f o u n d to increase s p r i n g - i n b y as m u c h as 0 . 3 3 ° c o m p a r e d to a steel tool . T h e tool surface affects the abil i ty o f the interface to transmit stresses. Parts processed without an F E P sheet had greater s p r i n g - i n b y up to 1 . 3 5 ° c o m p a r e d to parts processed wi th an F E P sheet. It was also seen that b y increas ing the total length o f the f langes b y 32 m m , spr ing- in was increased b y as m u c h as 0 . 7 7 ° . C u r e c y c l e and part thickness are b e l i e v e d to i n f l u e n c e the abili ty o f the part to d e v e l o p stresses through the thickness d u r i n g cure . T h e cure c y c l e affects the temperature at w h i c h the part gels. W i t h a 2 - h o l d cure c y c l e where the part gels p r i o r to the second heat-up, the part d e v e l o p s greater stresses than for a 1-hold c y c l e . Parts c u r e d w i t h a 2 - h o l d cure c y c l e had greater s p r i n g - i n , b y u p to 1 . 0 6 ° , w h e n c o m p a r e d to parts c u r e d wi th a 1-hold cure c y c l e . F i n a l l y , thin parts s p r i n g - i n m o r e than thick parts. E x p e r i m e n t s s h o w e d that 8 -ply parts had larger s p r i n g - i n than 16-ply parts b y as m u c h as 0 . 7 5 ° . T h e s e exper imental results support the p r o p o s e d m e c h a n i c a l tool-part interaction m e c h a n i s m and s h o w that there are s ignif icant interactive effects between part thickness , f lange length, cure cyc le , tool surface, and tool material . - 3 8 -Chapter 3 Experimental - Total Spring-in 3.6 S U M M A R Y The following observations were made from the experimental results presented in this Chapter: • The following parameters were found to affect spring-in: part lay-up, part initial angle, part thickness, flange length, cure cycle, tooling material, and tooling surface conditions. • The effects of part lay-up and initial part angle were in agreement with the predicted trends from anisotropy spring-in calculations (Equation 2.1). • Process conditions were found to significantly affect spring-in. Of the process parameters examined, the tool surface condition and the cure cycle had the greatest influence on spring-in. • There was a significant interaction between the effects of part thickness, flange length, cure cycle, tool material, and tool surface condition. The observed trends were in general agreement with the proposed mechanical tool-part interaction mechanism. - 3 9 -Chapter 3 Experimental - Total Spring-in 3.7 T A B L E S Table 3.1. Description of target cure cycles used. Segment Description 1-hold cure cycle 1 Apply a vacuum of -101 kPa 2 Pressurize to 586 kPa (release vacuum when autoclave pressure reaches 103 kPa) 3 When pressure reaches 448 kPa, heat to 180°C at a rate of 2.2°C/min 4 Hold for 140 min (after lagging thermocouple reaches 174°C) 5 Cool down to room temperature 6 Depressurize when the lagging temperature reaches 82°C 2-hold cure cycle 1 Apply a vacuum of -101 kPa 2 Pressurize to 586 kPa (release vacuum when autoclave pressure reaches 103 kPa) 3 When pressure reaches 448 kPa, heat to 135°C at a rate of 2.2°C/min 4 Hold for 160 min (after lagging thermocouple reaches 130°C) 5 Heat to 180°C af a rate of 2.2°C/min 6 Hold for 120min (after lagging temperature reaches 174°C) 7 Cool down to room temperature 8 Depressurize when the lagging temperature reaches 82°C Table 3.2. Experiment 1 - parameters varied. Parameter ID +1 -1 Intrinsic Part shape 1 C L Lay-up 2 [0,+45,-45,90]ns (n = 1 or 2) [0]„ (n = 8 or 16) Flange length 3 89 mm 57 mm Part thickness 4 16 plies 8 plies Part angle 5 45° 90° Extrinsic Tool material 6 Aluminum Steel Tool surface' 7 Release agent (2) Release agent (2) + FEP Cure cycle 8 2 holds 1 hold ' The number within parenthesis denotes the number of coats of release agent applied. - 4 0 -Chapter 3 Experimental - Total Spring-in Table 3.3. Experiment 1 - test matrix. Intrinsic Extrinsic Part Lay-up Flange Thickness Part T o o l T o o l Cure shape length angle material surface cycle Part ID x, X 2 X 3 X4 x 5 x 6 X 7 x 8 D O E - 1 -1 -1 -1 +1 +1 +1 -1 + 1 D O E - 2 +1 -1 -1 -1 -1 +1 +1 + 1 D O E - 3 -1 + 1 -1 -1 + 1 -1 + 1 ' + 1 D O E - 4 + 1 + 1 -1 + 1 -1 -1 -1 +1 D O E - 5 -1 -1 + 1 + 1 -1 -1 + 1 + 1 D O E - 6 + 1 -1 + 1 -1 +1 -1 -1 + 1 D O E - 7 -1 +1 +1 -1 -1 +1 -1 + 1 D O E - 8 + 1 +1 + 1 + 1 + 1 +1 +1 + 1 D O E - 9 + 1 + 1 + 1 -1 -1 -1 + 1 -1 D O E - 1 0 -1 + 1 + 1 + 1 + 1 -1 -1 -1 D O E - 1 1 + 1 -1 + 1 + 1 -1 +1 -1 -1 D O E - 1 2 -1 -1 + 1 -1 + 1 + 1 + 1 -1 D O E - 1 3 + 1 + 1 -1 -1 + 1 + 1 -1 -1 D O E - 1 4 -1 + 1 -1 + 1 -1 + 1 + 1 -1 D O E - 1 5 + 1 -1 -1 +1 +1 -1 + 1 -1 D O E - 1 6 -1 -1 -1 -1 -1 -1 -1 -1 Note: +1 and -1 represent the high and low value for each factor. Table 3.4. Experiment 2 - test matrix. Part ID Part Lay- Flange Thickness Part Tool Tool surface' Cure Rubber shape up length (plies) (mm) angle O material cycle (holds) caul RC-1 C [0]8 57 8 90 Aluminum Release agent (3) 1 Yes RC-2 C [Oh 57 8 90 Aluminum Release agent (3) 1 No RC-3 C [0]« 57 8 90 Steel Release agent (3) 1 Yes RC-4 C [0]8 57 8 90 Steel Release agent (3) 1 No RC-5 C [0]« 57 8 90 Aluminum Release agent (3) 2 Yes RC-6 C [0]« 57 8 90 Aluminum Release agent (3) 2 No TS2-1 C [0]« 57 8 90 Aluminum Release agent (3) 2 No RC-7 C [0]« 57 8 90 Steel Release agent (3) 2 Yes RC-8 C [0]« 57 8 90 Steel Release agent (3) 2 No TS2-4 C [0]8 57 8 90 Steel Release agent (3) 2 No RL-1 L [0]« 57 8 90 Aluminum Release agent (3) 2 Yes RL-2 L [0]8 57 8 90 Aluminum Release agent (3) 2 No RL-3 L [0]« 57 8 90 Steel Release agent (3) 2 Yes RL-4 L [0]«- 57 8 90 Steel Release agent (3) 2 No The number within parenthesis denotes the number of coats of release agent applied. - 4 1 -Chapter 3 Experimental - Total Spring-in Table 3 . 5 . Experiment 3 - test matrix. Part ID Part Lay- Flange Thickness Part Tool Tool surface' Cure shape up length (mm) (plies) angle O material cycle (holds) RC-6 C [0]« 57 8 90 Aluminum Release agent (3) 2 TS2-1 C [0]8 57 8 90 Aluminum Release agent (3) 2 TS-1 c [0],6 57 16 90 Aluminum Release agent (3) 2 TS2-2 c [ ],« 57 16 90 Aluminum Release agent (3) 2 TS-2 c [0]s 57 8 90 Aluminum Release agent (3) + FEP 2 TS-3 c [0],6 57 16 90 Aluminum Release agent (3) + FEP 2 TS2-3 c [0],6 57 16 90 Aluminum Release agent (3) + FEP 2 RC-8 c [0]« 57 8 90 Steel Release agent (3) 2 TS2-4 c [0]« 57 8 90 Steel Release agent (3) 2 TS-4 c [0],« 57 16 90 Steel Release agent (3) 2 TS2-5 c [0],6 57 16 90 Steel Release agent (3) 2 TS-5 c [0]8 57 8 90 Steel Release agent (3) + FEP 2 TS-6 c [0]l6 57 16 90 Steel Release agent (3) + FEP 2 TS2-6 —1 c [0],6 57 16 90 Steel Release agent (3) + FEP 2 The number within parenthesis denotes the number of coats of release agent applied. FEP denotes the use of a fluorinated ethylene propylene sheet. Table 3 . 6 . Experiment 4 - test matrix. Part ID Part Lay-up Flange Thickness Part Tool Tool surface' Cure shape length (mm) (plies) angle O material cycle (holds) W-l L [01,6 89 16 90 Steel Release agent (3) 2 W-2 C [0]s 89 8 90 Steel Release agent (3) 2 W-3 L [0]8 89 8 90 Aluminum Release agent (3) 2 W-4 L [0,45,-45,90]s 89 8 90 Steel Release agent (3) 2 W-5 L [0,45,-45,90] s 89 8 90 Aluminum Release agent (3) + FEP 2 The number within parenthesis denotes the number of coats of release agent applied. FEP denotes the use of a fluorinated ethylene propylene sheet. Table 3 . 7 . Experiment 5 - test matrix. Part ID Part Lay-up Flange Thickness Part Tool Tool surface' Cure shape length (plies) (mm) angle (°) material cycle (holds) F-l C [Oh 89 8 45 Steel Release agent (3) + FEP 2 F-2 C [0]« 89 8 45 Steel Release agent (3) 2 F-3 L [0]8 89 8 45 Aluminum Release agent (3) 2 F-4 L [0,45,-45,90]s 89 8 45 Aluminum Release agent (3) 2 The number within parenthesis denotes the number of coats of release agent applied. FEP denotes the use of a fluorinated ethylene propylene sheet. - 4 2 -Chapter 3 Experimental - Total Spring-in Table 3.8. Experiment 6 - test matrix. Part ID Part Lay-up Flange Thickness Part Tool Tool surface' Cure shape length (mm) (plies) angle (°) material cycle (holds) TA-1 L [0]8 57 8 90 Aluminum Release agent (2) 1 T A - 2 L [0]8 57 8 45 Aluminum Release agent (2) 1 TA-3 L [0]8 57 8 0 Aluminum Release agent (2) 1 TA-4 L [0]8 57 8 90 Steel Release agent (2) 1 TA-5 L [0]8 57 8 45 Steel Release agent (2) 1 TA-6 L [0]8 57 8 0 Steel Release agent (2) 1 The number within parenthesis denotes the number of coats of release agent applied. Table 3.9. Anisotropy component calculations - parameters and results. Parameters [0]„ [0/45/-45/90]n s CTE, ( m / m / ° C ) y -2.55 x 1 0 s 2.53 x 10"6 CTE, (m/mrC)' 4.05 x 10"5 5.95 x 10"5 <pi (m/m) 8.05 x 10"5 2.91 x 10"4 <j>, (m/m) 3.76 x 10"3 5.43 x 10"3 Cure temperature ( ° C ) 180 180 R o o m temperature ( ° C ) 20 20 Initial angle, 9 (°) 90 45 90 45 A 9 C X E (°) 0.58 0.29 0.82 0.41 A 9 C s O 0.32 0.16 0.46 0.23 A9 = AOCTE + A 9 C S Q 0-90 0.45 1.28 0.64 C T E values are given for the fully cured material at 1 0 0 ° C . Table 3.10. Experiment 1 - results. Factor 1 2 3 4 5 6 7 8 Part ID Part Lay-up Flange Thickness Part Tool Tool surface' Cure Spring- Range2 shape length (mm) (plies) angle (°) material Cycle (holds) in (°) (°) DOE-1 L [0],6 57 16 45 Aluminum Release agent (2) + F E P 2 0.65 ±0 .01 DOE-2 C [0]8 57 8 90 Aluminum Release agent (2) 2 1.44 ± 0 . 0 3 DOE-3 L [0,45,-45,90]., 57 8 45 Steel Release agent (2) 2 1.41 ± 0 . 0 7 DOE-4 C [0,45,-45,90]2s 57 16 90 Steel Release agent (2) + FEP 2 1.10 ± 0 . 0 7 DOE-5 L [0] 1 6 89 16 90 Steel Release agent (2) 2 1.37 ±0 .01 DOE-6 C [018 89 8 45 Steel Release agent (2) + FEP 2 0.61 ± 0 . 0 7 DOE-7 L [0,45,-45,90]s 89 8 90 Aluminum Release agent (2) + FEP 2 1.09 ±0 .01 DOE-8 C [0,45,-45,90]2s 89 16 45 Aluminum Release agent (2) 2 1.26 ± 0 . 2 0 DOE-9 C [0,45,-45,90]s 89 8 90 Steel Release agent (2) 1 1.51 ± 0 . 0 3 DOE-10 L [0,45,-45,90]2s 89 16 45 Steel Release agent (2) + F E P 1 0.89 ± 0 . 0 0 DOE-11 C [0],« 89 16 90 Aluminum Release agent (2) + FEP 1 0.94 ± 0 . 0 8 DOE-12 L [0]8 89 8 45 Aluminum Release agent (2) 1 1.00 ± 0 . 0 9 DOE-13 C [0,45,-45,90]s 57 8 45 Aluminum Release agent (2) + FEP 1 1.03 ± 0 . 0 6 DOE-14 L [0,45,-45,90]2s 57 16 90 Aluminum Release agent (2) 1 1.41 ± 0 . 0 4 DOE-15 C [0]1 6 57 16 45 Steel Release agent (2) 1 0.70 ± 0 . 0 3 DOE-16 L [0]8 57 8 90 Steel Release agent (2) + FEP 1 0.91 ± 0 . 0 3 average 1.08 The number within parenthesis denotes the number of coats of release agent applied. F E P denotes the use of a fluorinated ethylene propylene sheet. 2 The range shown represents 1/2 (maximum value - minimum value). - 4 3 -Chapter 3 Experimental - Total Spring-in Table 3.11. Experiment 1 - first and second order coefficients. Aliased Parameters Coefficient (°) C | + (higher order effects) -0.02 0,2 + (higher order effects) +0.26 C3 + (higher order effects) 0.00 C 4 + (higher order effects) -0.09 C 5 + (higher order effects) -0.28 Q + (higher order effects) +0.04 C 7 + (higher order effects) +0.36 Cs + (higher order effects) +0.07 " C 1 2 " i.e. C|2+C37+C48+C5 6 + (higher order effects) +0.04 " C 1 3 " i.e. C 1 3 + C 2 7 + C 4 6 + C 5 8 + (higher order effects) +0.01 " C H " i.e. C 1 4 + C 2 8 + C 3 6 + C 5 7 + (higher order effects) -0.06 " C 1 5 " i.e. C 1 5 + C 2 6 + C 3 8 + C 4 7 + (higher order effects) -0.07 "Ci6" , i.e. C ] 6 + C 2 5 + C 3 4 + C 7 8 + (higher order effects) +0.15 " C , 7 " i.e. C 1 7 + C 2 3 + C 6 8 + C 4 5 + (higher order effects) -0.05 " C , 8 " i.e. C 1 8 + C 2 4 + C 3 5 + C 6 7 + (higher order effects) -0.01 T a b l e 3.12. E x p e r i m e n t 1 - f i rst order effects . Parameter X j = -1 X i = +1 Coefficient Value O Part shape L C C , -0.02 Lay-up [0]„ [0,+45,-45,90]ns c 2 +0.26 Flange length 57 mm 89 mm c 3 0.00 Thickness 8 plies 16 plies c 4 -0.09 Part angle 90° 45° C5 -0.28 T o o l material Steel A l u m i n u m c 6 0.04 T o o l surface Release agent (2) + F E P Release agent (2) +0.36 Cure cycle 1 hold 2 holds c 8 +0.07 - 4 4 -Chapter 3 Experimental - Total Spring-in Table 3.13. Experiment 2 - results. Part Part Lay- Flange Thickness Part Tool Tool surface' Cure Rubber Spring- Range2 ID shape up length (plies) angle material cycle caul in (°) (mm) O (holds) (°) RC-1 C [0]« 57 8 90 Aluminum Release agent (3) 1 Yes 1.09 ±0 .01 RC-2 C [0]8 57 8 90 Aluminum Release agent (3) 1 No 1.12 ± 0 . 0 2 RC-3 C [0]8 57 8 90 Steel Release agent (3) 1 Yes 1.06 ±0 .01 RC-4 C [0]8 57 8 90 Steel Release agent (3) 1 No 1.04 ± 0 . 0 5 RC-5 C [0]« 57 8 90 Aluminum Release agent (3) 2 Yes 1.62 ± 0 . 0 2 RC-6 C [0]« 57 8 90 Aluminum Release agent (3) 2 No 1.89 ± 0 . 0 3 TS2-1 C [0], 57 8 90 Aluminum Release agent (3) 2 No 1.64 ± 0 . 0 2 RC-7 C [0]8 57 8 90 Steel Release agent (3) 2 Yes 1.32 ± 0 . 0 4 RC-8 C [0]8 57 8 90 Steel Release agent (3) 2 No 1.36 ± 0 . 0 2 TS2-4 C [0]8 57 8 90 Steel Release agent (3) 2 No 1.63 ± 0 . 0 5 RL-1 L [0]8 57 8 90 Aluminum Release agent (3) 2 Yes 1.78 ± 0 . 0 2 RL-2 L [0]8 57 8 90 Aluminum Release agent (3) 2 No 1.76 ± 0 . 0 3 RL-3 L [0]« 57 8 90 Steel Release agent (3) 2 Yes 1.34 ± 0 . 0 4 RL-4 L [0]8 57 8 90 Steel Release agent (3) 2 No 1.37 ± 0 . 0 5 The number within parenthesis denotes the number of coats of release agent applied. 2 The range shown represents 1/2 (maximum measured value - minimum measured value). Table 3.14. Experiment 3 - results. Part ID Part shape Lay-up Flange length (mm) Thickness (plies) Part angle (°) Tool material Tool surface' Cure cycle (holds) Spring-in (°) Range2 (°) RC-6 C [0]8 57 8 90 Aluminum Release agent (3) 2 1.89 +0.03 TS2-1 C [0]8 57 8 90 Aluminum Release agent (3) 2 1.64 ± 0 . 0 2 TS-1 C [0],6 57 16 90 Aluminum Release agent (3) 2 1.11 ± 0 . 0 1 TS2-2 C [0],« 57 16 90 Aluminum Release agent (3) 2 1.20 ± 0 . 0 1 TS-2 C [0]« 57 8 90 Aluminum Release agent (3) + FEP 2 0.92 ± 0 . 0 2 TS-3 c [0],« 57 16 90 Aluminum Release agent (3) + FEP 2 0.79 ± 0 . 0 2 TS2-3 c [0],6 57 16 90 Aluminum Release agent (3) + FEP 2 0.90 ± 0 . 0 2 RC-8 c [0]8 57 8 90 Steel Release agent (3) 2 1.36 ± 0 . 0 2 TS2-4 c [0]g 57 8 90 Steel Release agent (3) 2 1.63 ± 0 . 0 5 TS-4 c [0],6 57 16 90 Steel Release agent (3) 2 1.15 ± 0 . 0 2 TS2-5 c [0],6 57 16 90 Steel Release agent (3) 2 1.25 +0.02 TS-5 c [0]8 57 8 90 Steel Release agent (3) + FEP 2 0.84 ± 0 . 0 4 TS-6 c [0],6 57 16 90 Steel Release agent (3) + FEP 2 0.78 ± 0 . 0 5 TS2-6 c [0]|6 57 16 90 Steel Release agent (3) + FEP 2 0.85 ± 0 . 0 2 The number within parenthesis denotes the number of coats of release agent applied. F E P denotes the use of a fluorinated ethylene propylene sheet. 2 The range shown represents 1/2 (maximum measured value - minimum measured value). - 4 5 -Chapter 3 Experimental - Total Spring-in Table 3.15. Experiment 4 - results. Part ID Part Lay-up Flange Thickness Part Tool Tool surface' Cure Spring Range2 shape length (plies) angle material cycle -in (°) (mm) O (holds) O W - l L [0],6 89 16 90 Steel Release agent (3) 2 1.38 +0.01 W-2 C [0]8 89 8 90 Steel Release agent (3) 2 2.13 ± 0 . 0 3 W-3 L [0]8 89 8 90 Aluminum Release agent (3) 2 2.14 +0.03 W-4 L [0,45,-45,90]s 89 8 90 Steel Release agent (3) 2 2.57 ± 0 . 0 2 W-5 L [0,45,-45,90]s 89 8 90 Aluminum Release agent (3) + FEP 2 1.22 ± 0 . 0 3 The number within parenthesis denotes the number of coats of release agent applied. F E P denotes the use of a fluorinated ethylene propylene sheet. 2 The range shown represents 1/2 (maximum measured value - minimum measured value). Table 3.16. Experiment 5 - results. Part Part Lay-up Flange Thickness Part Tool Tool surface' Cure Spring- Range2 ID shape length (plies) angle material cycle in O (mm) O (holds) O F - l C [0]8 89 8 45 Steel Release agent (3) + FEP 2 0.61 ± 0 . 0 3 F-2 C [0]8 89 8 45 Steel Release agent (3) 2 1.45 ± 0 . 0 7 F-3 L [0]» 89 8 45 Aluminum Release agent (3) 2 1.52 ± 0 . 0 2 F-4 / rr,,. L [0,45,-45,90]., 89 8 45 Aluminum Release agent (3) 2 1.98 ± 0 . 0 2 The number within parenthesis denotes the number of coats of release agent applied. F E P denotes the use of a fluorinated ethylene propylene sheet. The range shown represents 1/2 (maximum measured value - minimum measured value). Table 3.17. Experiment 6 - results. Part Part Lay-up Flange Thickness Part Tool Tool surface' Cure Spring- Range2 ID shape length (plies) angle material cycle in O (mm) O (holds) (°) TA-1 L [0]» 57 8 90 Aluminum Release agent (2) 1 1.03 ±0.00 TA-2 L [0]g 57 8 45 Aluminum Release agent (2) 1 0.77 ±0.02 TA-3 L 10]8 57 8 0 Aluminum Release agent (2) 1 0.21J ±0.02 TA-4 L [0]8 57 8 90 Steel Release agent (2) 1 0.92 ±0.02 TA-5 L [0]8 57 8 45 Steel Release agent (2) 1 0.56 ±0.00 TA-6 L [0]8 57 8 0 Steel Release agent (2) 1 0.13"' ±0.01 The number within parenthesis denotes the number of coats of release agent applied. 2 The range shown represents 1/2 (maximum measured value - minimum measured value). 3 Apparent spring-in for flat parts due to warpage. Table 3.18. Lengths of influence, Llot, when measuring spring-in for C and L-shaped parts. C-shaped parts L-shaped parts P a r t l D LjUmge Lweb L,nt Part ID Lflange (mm) (mm) (mm) (mm) (mm) W - 2 89 102 191 W - l , W - 3 , F-4 89 178 R C - 6 , R C - 8 , T S - 4 , 57 102 159 D O E - 3 57 114 T S 2 - 1 . T S 2 - 4 , TS2-5 - 4 6 -Chapter 3 Experimental - Total Spring-in Table 3.19. Effect of length of influence, L,„„ on spring-in. Parts are described in Tables 3.2 to 3.8. Longer part Equivalent shorter part Thickness Difference Difference Part Lot Spring-in Part Lot Spring-in (plies) in Lm in spring-in ID (mm) C) ID (mm) C) (mm) C) F-4 178 1.98 D O E - 3 114 1.41 8 64 0.57 W - 2 191 2.13 R C - 8 159 1.36 8 32 0.77 TS2-4 159 1.63 8 32 0.50 W-3 178 2.14 R C - 6 159 1.89 8 19 0.25 TS2-1 159 1.64 8 19 0.50 W - l 178 1.38 T S - 4 159 1.15 16 19 0.23 TS2-5 159 1.25 16 19 0.13 Table 3.20. Effect of cure cycle on spring-in. Parts are described in Tables 3.2 to 3.8. Part prepared with Similar part prepared with Difference a 1-hold cure cycle a 2-hold cycle in spring-in Part ID Spring-in (") Part ID Spring-in (") C) T A - 1 1.03 R C - 6 1.89 0.86 TS2-1 1.64 0.61 T A - 4 0.92 R C - 8 1.36 0.44 TS2-1 1.63 0.71 D O E - 9 1.51 W-4 2.57 1.06 D O E - 1 2 1.00 F-3 1.52 0.52 - 4 7 -Chapter 3 Experimental - Total Spring-in 3.8 FIGURES 22 m m 256 m m (b) Figure 3.1. Tools, (a) 90° tool, (b) 45° tool. - 4 8 -Chapter 3 Experimental - Total Spring-in Figure 3.2. Part geometries, (a) C-shaped part, (b) L-shaped part. 1 100 -I Time (not to scale) Figure 3.3. Schematic of cure cycles used. A more detailed description of the cure cycles is given in Table 3.1. - 4 9 -Chapter 3 Experimental - Total Spring-in Figure 3.4. Digital image analysis - cross section profile of part W - 3 . (a) Image before brightness and contrast adjustment, (b) Image after brightness and contrast adjustment. f l a n g e a n g l e f 7 f l a n g e angle r (a) N V (b) Figure 3.5. Schematic of flange angle measurements, (a) C-shaped part, (b) L-shaped part. Straight lines were fitted to two points on each segment. -50-Chapter 3 Experimental - Total Spring-in Chapter 3 Experimental - Total Spring-in (b) Figure 3.7. Composite parts, (a) C-shaped parts, (b) L-shaped parts. warped part Figure 3.8. Measurement of apparent spring-in of initially flat parts. - 5 2 -Chapter 3 Experimental - Total Spring-in o .5 1 6 0 •so. 0. 0 0. A 0 C T E + A0 C S AG CTE o Q N O i W o Q a o Q o Q w o Q o Q 45° parts I w o Q w O Q [0]„ [0, 45, -45, 90]n —> 90° parts UQ O Q >/-» W o Q W O Q [0]„ r— ON w tu w o O o Q Q Q J v. o Q [0, 45,-45,90]n Figure 3.9. Experiment 1 - measured spring-in. Solid and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. The ranges shown represent 1/2 (maximum measured value - minimum measured value). Parts are described in Table 3.10. 2.2 2.0 H 1.8 1.6 C 1 - 4 • •S 1.2 -oo .£ 1.0 #0.8 0.6 0.4 0.2 H 0.0 A^ CTE + A0 C S A 9CTE 2 < C N m u u U U U u oS oS oS OS OS OS on t--• U OS U OS ( N on H OS ( N i J OS i—I oS OS Figure 3.10. Experiment 2 - measured spring-in. Solid and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. The ranges shown represent 1/2 (maximum measured value - minimum measured value). Parts are described in Table 3.13. - 5 3 -Chapter 3 Experimental - Total Spring-in 2.2 2.0 1.8 1.6 •S1.2 bo •S 1.0 S—( CL, ^0 .8 0.6 0.4 0.2 0.0 aluminum tooling steel tooling — A 0 C T E + A0, A 0 C T E 3cs 35 % a. IS o OS CN 00 H OO CN 00 H oo CN OO H U OS oo 00 CN H oo H oo H O0 CN H oo H Figure 3.11. Experiment 3 - measured spring-in. Solid and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. The ranges shown represent 1/2 (maximum measured value - minimum measured value). Parts are described in Table 3.14. 3.0 2.5 2.0 A Ml.5 c •c Cu 00 1.0 0.5 0.0 A 0 C T E + A 0 C S A 0 C T E I l l s ! * i [0]„ [0/45/-45/90]n Figure 3.12. Experiment 4 - measured spring-in. Solid and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. The ranges shown represent 1/2 (maximum measured value - minimum measured value). Parts are described in Table 3.15. - 5 4 -Chapter 3 Experimental - Total Spring-in 2.5 2.0 A 1.5 H 60 C 0.5 0.0 A 0 C X E + AG C S — A6 f J C T E c s.'st,i ri» C'S/St/R L/K/AI/R LVX'AL'R U-[0]„ [0/45/-45/90]n Figure 3.13. Experiment 5 - measured spring-in. Solid and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. The ranges shown represent 1/2 (maximum measured value - minimum measured value). Parts are described in Table 3.16. 1.2 1 aluminum tooling: A6 = 0.0099 + 0.259° steel tooling: A8 = 0.0099 + 0.140° anisotropy calculations: A9 = 0.0109 20 40 60 Initial part angle f ) 80 100 Figure 3.14. Experiment 6 - linear fit of measured spring-in vs. part angle for aluminum and steel tooling. Anisotropy predictions were obtained using Equation 2.1, see Table 3.9. The ranges shown represent 1/2 (maximum measured value - minimum measured value). Parts are described in Table 3.17. - 5 5 -Chapter 3 Experimental - Total Spring-in 3.0 -i 2.5 -2.0 -o C '5)1.5 _g 'C a, 00 1.0 0.5 A 0.0 2 /jcWs no FEP 8 plies 57 mm E x p 1 (average) E x p 2 (2 holds) Figure 3.15. Effect of part shape on spring-in. The ranges shown represent the maximum and minimum values for each data set. Parts are described in Tables 3.10 and 3.12. 3.5 3.0 2.5 :i.o 1.5 1.0 0.5 0.0 C OO E[0,45,-45,90]ns H[0]n 2 holds no FEP 8 plies 89 mm L-shaped W-4 C-2 holds no FEP 8 plies 89 mm steel WsSm. W-4 ••IF 1 \\ • I 2 holds no FEP 8 plies 89 mm F-4 i l l E x p 1 (average) E x p 4 E x p 4 E x p 5 Figure 3.16. Effect of part lay-up on spring-in. The ranges shown represent the maximum and minimum values for each data set. Parts are described in Tables 3.10 to 3.16. - 5 6 -Chapter 3 Experimental - Total Spring-in 2 holds FEP 57 mm T wMmmmm IBIH 2 holds no FEP 57 mm 2 holds no FEP 89 mm C-shaped .shaped] W-2 | W-l E x p 1 (average) E x p 3 (FEP) E x p 3 (no F E P ) E x p 4 Figure 3.17. Effect of part thickness on spring-in. The ranges shown represent the maximum and minimum values for each data set. Parts are described in Tables 3.10, 3.14, and 3.15. 2 hold no FEP steel 16 plies 1 17Hmm 159 mm l'S-4 TS2-5 W-l 2 hold no FEP steel 8 plies 4, 159 mm 191 nun 2 hold no FEP aluminum 8 plies Li, /.iy mm RC-6 TS2-1 17<S mm W-5 Figure 3.18. Effect of length of influence, LM„ on spring-in. The ranges shown represent the maximum and minimum values for each data set. Parts are described in Tables 3.13 and 3.16. - 5 7 -Chapter 3 Experimental - Total Spring-in 2.5 100 Initial part angle (°) F i g u r e 3.19. E x p e r i m e n t s 4 and 5 - effect o f part angle on s p r i n g - i n . S p r i n g - i n o f parts F -2 , F - 3 , W -2, and W - 3 . T h e ranges s h o w n represents 1/2 ( m a x i m u m measured value - m i n i m u m measured value) . Parts are d e s c r i b e d i n T a b l e 3.15 and T a b l e 3.16. A n i s o t r o p y predic t ions were obtained u s i n g E q u a t i o n 2.1, see T a b l e 3.9. 2.5 2.0 1*1-0 0.5 H 0.0 0 1 h o l d H 2 h o l d s fe*'-E x p 1 ( a v e r a g e ) no FEP 8 plies 57 mm E x p 2 ( C - s h a p e d ) no FEP 8 plies aluminum 89 mm no FEP 8 plies aluminum 57 mm D O E -12 F-3 . ! T-\-l l i IIP K L - l no FEP 8 plies steel 57 mm TA-3 Kl.-' F i g u r e 3.20. E f f e c t o f cure c y c l e on s p r i n g - i n . T h e ranges s h o w n represent the m a x i m u m and m i n i m u m values for each data set. Parts are descr ibed i n T a b l e s 3.10, 3.13, 3.14, and 3.17. - 5 8 -Chapter 3 Experimental - Total Spring-in 3.5 3.0 2.5 tio c "g.1-5 on 1.0 0.5 0.0 • Release agent + F E P U Release agent 2 holds 57 mm' 2 holds 16 plies 57 mm :r_ 4 -2 holds 8 plies 57 mm 2 holds 8 plies 89 mm 2 holds 8 plies 89 mm r - i F-2 E x p 1 E x p 3 E x p 3 E x p 3 (average) (average) (16 plies) (8 plies) E x p 4 E x p 5 Figure 3.21. Effect of tool surface on spring-in. The ranges shown represent the maximum and minimum values for each data set. Parts are described in Tables 3.10 and 3.14 to 3.16. 3.0 2.5 2.0 -I o _C '5)1.5 a •c Cu 00 1.0 0.5 0.0 • Steel • Aluminum / hold no FEP 8 plies 57 mm Ml l •ill 2 holds no FEP 8 plies 57 mm 1 hold no FEP 8 plies 57 mm rs 2 holds no FEP 8 plies 57 mm 2 holds J 6 plies 57 mm 2 holds FEP 57 mm 2 holds no FEP 8 plies 89 mm W-2 2 holds no FEP 8 plies 89 mm • l l mmm WBm lllllll 1 F-2 Exp 1 Exp 2 Exp 2 Exp 6 Exp 3 Exp 3 Exp 3 (average) (1 hold) (2 holds) (average) (8 plies (16 plies) (FEP) no FEP) Exp 4 Exp 4 Figure 3.22. Effect of tool material on spring-in. The ranges shown represent the maximum and minimum values for each data set. Parts are described in Tables 3.10, 3.13 to 3.15, and 3.17. - 5 9 -C H A P T E R 4: E X P E R I M E N T A L - S P R I N G - I N C O M P O N E N T S T o quantify s p r i n g - i n , the concept o f a spr ing- in angle is c o m m o n l y u s e d (F igure 1.5). H o w e v e r , what typica l ly matters in assembly is the relative m o v e m e n t between fastener locat ions on mat ing parts. T h i s chapter w i l l s h o w that s p r i n g - i n angles measured at the w e b - f l a n g e c o n n e c t i o n m a y not be suffic ient to quant i fy this because the deformat ion m a y be c o m p o u n d e d wi th warpage o f the w e b and f lange. T h e w o r k presented i n this chapter d e c o m p o s e s the spr ing- in angle measurements presented in C h a p t e r 3 into two c o m p o n e n t s : a corner c o m p o n e n t and a warpage c o m p o n e n t . T h e experimental results demonstrate that the corner c o m p o n e n t is m a i n l y the result o f anisotropy and depends o n the lay -up , the ini t ia l angle , and the c o m p o s i t e material properties. T h e warpage c o m p o n e n t , o n the other hand, is the result o f tool-part interaction and depends o n the cure c y c l e , the tool surface, the tool material , the part thickness and the f lange length. 4.1 SEPARATION O F SPRING-IN INTO W A R P A G E A N D C O R N E R C O M P O N E N T S C l o s e observat ion o f the parts presented in C h a p t e r 3 s h o w e d that al l flat sections w a r p e d away f r o m the tool , i.e. c o n c a v e side towards the tool . T o quantify this warpage, the prof i les o f each f lange and w e b were measured u s i n g digi tal image analysis . T h e scanned cross-sect ion o f each part was e x a m i n e d wi th the U l e a d PhotoImpact V e r s i o n 3.01 software. A f t e r the image contrast and brightness were adjusted to obtain a c lear black and white p r o f i l e o f the part (e.g. F i g u r e 3.4), the locations o f thirteen points o n the tool -s ide edge o f the prof i le were recorded a l o n g the length o f each f lange /web. B e c a u s e the p r o f i l e edges are not clear at a l l locations w i t h i n the images , the recorded - 6 0 -Chapter 4 Experimental - Spring-in Components points were careful ly selected w i t h i n clear, u n i f o r m portions o f the p r o f i l e (e.g. F i g u r e 3.6). T h e prof i le data was subsequently analysed u s i n g the M i c r o s o f t E x c e l spreadsheet software. Plots of f lange def lec t ion vs. locat ion a l o n g the f lan g e /web length were obtained through s i m p l e mathematical translation and rotation o f the data points (see F i g u r e 4.1). T h e scale was conver ted to mil l imetres (at a resolution o f 600 d p i , 1 p i x e l is equal to 4.23 x 10"2 m m ) . T h e measured prof i les s h o w e d that w h i l e the warpage was not direct ly observable with the n a k e d eye, it was i n m a n y cases great e n o u g h to s ignif icant ly increase the apparent (measured) s p r i n g - i n . F o r a part without warpage, the s p r i n g - i n measured at the corner is the same as that measured anywhere a l o n g the f langes . T h e s p r i n g - i n o f such parts is independent o f the measurement locat ion . W h e n warpage is present, however , s p r i n g - i n results w i l l be affected b y the measurement locat ions . W h e n a n a l y z i n g s p r i n g - i n , it is thus important to cons ider h o w the measurements are taken. In the presence o f warpage, the total s p r i n g - i n (A9,„mi) can be separated into two c o m p o n e n t s : a corner c o m p o n e n t (A6corner) and a f lange warpage c o m p o n e n t (A0wurpase), see E q u a t i o n s 4.1 and 4.2, and F i g u r e 4.2. A0,„,„, =A(?6w„f,. + A 0 U W (4.1) ^ ^ M o l — ^corner ~*~ A warpage flui\ge R warpage ) (4.2) T h e effect o f warpage o n s p r i n g - i n was measured as the angle f o r m e d b y the tangent to the f lange or w e b p r o f i l e o n the side adjacent to the corner and the straight l ine d r a w n between the two extremities of the f lange or w e b ( A - A ) , see F i g u r e 4.2. A l t h o u g h the def lec t ion o f the f lange s h o w n i n F i g u r e 4.1 is o n l y 0.2 m m for a f lange that is 89 m m l o n g , warpage o f the f lange affects the measured spr ing- in by a p p r o x i m a t e l y 0 . 4 8 ° . T h i s shows h o w sensitive the measurement technique is to warpage . It is very d i f f i c u l t to measure spr ing- in right at the corner and all c o m m o n l y u s e d techniques rely o n measurements o f the f l a n g e / w e b away f r o m the corner. T h e total A6warpage for a g i v e n angle w i l l be - 6 1 -Chapter 4 Experimental - Spring-in Components the s u m o f the warpage c o m p o n e n t o f each f lange , i.e. A8flange A warpage and A8flange B warpage, ( E q u a t i o n 4.2 and F i g u r e 4.2). T h e corner spr ing- in is ca lculated b y subtracting the measured warpage c o m p o n e n t f r o m the total measured s p r i n g - i n . 4.2 RESULTS T h e s p r i n g - i n o f the parts presented in C h a p t e r 3 was s u b d i v i d e d into a warpage c o m p o n e n t , A 0 w a r p a g e , and a corner c o m p o n e n t , A8cornen u s i n g the techniques descr ibed in the p r e v i o u s sect ion. T h e results are s h o w n in T a b l e s 4.1 to 4.6 and F i g u r e s 4.3 to 4.8. In these f igures , the predic ted spr ing- in f r o m E q u a t i o n 2.1 (see T a b l e 3.1) is also presented. 4.3 OBSERVATIONS B a s e d on material anisotropy, all parts with the same l a y - u p and init ial angle s h o u l d have the same s p r i n g - i n ( E q u a t i o n 2.1). T h e results in F i g u r e s 4.3 to 4.8 c lear ly show that it is not the case. T h e measured s p r i n g - i n is i n general m u c h larger than what is predic ted b y E q u a t i o n 2.1. W h e n the s p r i n g - i n due to f lange and w e b warpage, A8warpage, is subtracted f r o m the total s p r i n g - i n , however , the large variat ion in the measured s p r i n g - i n between parts is s ignif icant ly r e d u c e d and the r e m a i n i n g A8cim„ lies c lose to the expected anisotropy value for most o f the parts (Figures 4.3 to 4.8). W h i l e f lange length, thickness , part shape, and extrinsic parameters have little effect o n A8cornen l a y - u p and initial part angle , o n the other h a n d , affect A8corner in a manner consistent wi th E q u a t i o n 2.1. T h e spr ing- in o f the corner, A9cornen is m a i n l y the result o f anisotropy, and c a n be predic ted with reasonable accuracy u s i n g E q u a t i o n 2.1. T h e results presented in T a b l e s 4.1 to 4.6 and F igures 4.3 to 4.8 s h o w that the va lue o f A0warpage varies s ignif icant ly f r o m part to part, a c c o u n t i n g for most o f the observed var iabi l i ty in s p r i n g - i n between different design and process condi t ions . T h e value o f A6warpage is o b serv ed to vary as a f u n c t i o n of - 6 2 -Chapter 4 Experimental - Spring-in Components two des ign parameters: part thickness and flange length, and three process parameters: tool material , tool surface, and cure c y c l e . T h e r e are also s ignif icant interactive effects between these parameters. 4 .3.1 E F F E C T O F PART SHAPE O N W A R P A G E O v e r a l l , part shape ( C or L ) has little effect o n A9warpage. In experiment 1, C - s h a p e d parts have a p p r o x i m a t e l y the same A9warpage as L - s h a p e d parts, wi th an average di f ference o f o n l y 0 . 0 5 ° ' . In experiment 2, equivalent C and L - s h a p e d parts have a p p r o x i m a t e l y the same A9warpage, see T a b l e 4.2 and F i g u r e 4.4. In this experiment , the average difference between the A9warpage o f C - s h a p e d parts cured with a 2 - h o l d c y c l e and that o f equivalent L - s h a p e d parts is 0 . 0 3 ° . T h e s e observations are illustrated in F i g u r e 4.9. 4 .3.2 E F F E C T O F L A Y - U P O N W A R P A G E A s s h o w n in F i g u r e 4.10, the effect o f l a y - u p on A9warpage is i n c o n c l u s i v e . In exper iment 1, parts with a [0,45,-45,90] n s l a y - u p have a greater A0warpage b y an average o f o n l y 0 . 0 4 ° w h e n c o m p a r e d to [0] n parts. A c o m p a r a b l e observat ion is m a d e w h e n e x a m i n i n g parts W - 2 and W - 4 . Part W - 4 , a [0,45,-45 ,90] s l a y - u p , has a p p r o x i m a t e l y the same A9warpage than part W - 2 , a s i m i l a r part wi th a [0] 8 l ay -up , with a di f ference o f 0 . 0 2 ° . N o t e that these parts, however , have different shapes. O n the other h a n d , in experiment 5 (F igure 4.7), part F -4 , wi th a [0,45,-45,90] s l a y - u p , has a greater A9warpage than part F -3, an equivalent [0 ] 8 part by 0 . 5 1 ° . 4 .3.3 E F F E C T O F L A M I N A T E THICKNESS O N W A R P A G E O v e r a l l , thin parts have greater A9warpage than thick parts. In experiment 1, 8-ply parts have greater A9warpage than 16-ply parts b y an average o f 0 . 1 2 ° . T h e effect o f laminate thickness is c lear ly seen in ' As discussed in Chapter 3, because of the nature of fractional factorial designed experiments, individual parts cannot be compared in experiment 1 because of the large number of parameters that vary from part to part. All effects noted from experiment 1 represent average effects over all the parts. - 6 3 -Chapter 4 Experimental - Spring-in Components experiment 3 ( F i g u r e 4.5), where 8 -ply parts have a greater A9warpage than their 16-ply counterparts for all process c o n d i t i o n s . T h e same f igure shows that the effect o f part thickness is most s ignif icant for parts processed without an F E P sheet, wi th an average difference o f 0 . 4 7 ° between 8-ply and 16-ply parts, c o m p a r e d to a di f ference o f 0 . 0 9 ° for parts processed w i t h an F E P sheet. Fur thermore , c o m p a r i s o n o f parts W - l and W - 2 (experiment 4) shows that part W - 2 , an 8-ply part, has 0 . 8 1 ° larger AOwarpage than part W - l , a s imi lar part with a thickness o f 16 pl ies . T h e s e parts d i f fer o n l y in their thickness and shape. K n o w i n g that part shape has little effect o n A9warpage, the large difference in AQwarpage can therefore be m a i n l y attributed to the di f ference in thickness . F r o m these observations, it is clear that thinner parts have greater A9warpage than thicker parts, and that the effect o f laminate thickness is dependent o n the tool surface c o n d i t i o n (F igure 4.11). 4.3.4 E F F E C T O F F L A N G E L E N G T H ON W A R P A G E T o e x a m i n e the effect o f part length on spr ing- in , a length variable , L,0„ was d e f i n e d i n C h a p t e r 3 as the total length o f the " f l a n g e s " c o r r e s p o n d i n g to a g i v e n corner (see C h a p t e r 3 f o r m o r e details) . T h e same variable , Lm, is n o w used to examine the effect o f length o n A9warpage. F i g u r e 4.12 shows a plot o f the A9warpage vs. L,nt for a l l the parts presented in T a b l e s 4.1 to 4.6. It is evident f r o m F i g u r e 4.12 that Adwarpage is not necessari ly greater for longer parts and that A9warpage is dependent o n other parameters. O n the other h a n d , w h e n c o m p a r i n g equivalent parts processed wi th a 2 - h o l d cure c y c l e without the use o f an F E P sheet, part length has a clear effect o n A9warpage. In F i g u r e 4.13, the A 9 w a r p a g e o f parts W - l , W - 2 , and W - 3 is c o m p a r e d to that o f equivalent shorter parts. F r o m these results, it is clear that A9warpage is greater for longer parts. Fur thermore , the results i n F i g u r e 4.13 s h o w that w h i l e part length has a s ignif icant effect on the A9warpage o f 8-ply parts, it does not s ignif icant ly affect the A0 w a r p a ge o f 16-ply parts. F r o m these results, it is evident that the effect o f part length on A9warpage is dependent on the cure c y c l e , the tool surface, as w e l l as the part thickness . - 6 4 -Chapter 4 Experimental - Spring-in Components 4.3.5 E F F E C T O F INITIAL PART A N G L E ON W A R P A G E In experiment 1, parts with an initial angle of 9 0 ° have a greater A9warpage than 4 5 ° parts by an average of only 0 .06° . In experiment 6, A6warpage is approximately the same between 9 0 ° part sand 4 5 ° parts, with a difference of merely 0 .01° . Parts prepared in experiments 4 and 5, on the other hand, show different results. Parts W-2 and W-3 (Table 4.4), with an in initial angle of 90° , have greater A6warpage than parts F-2 and F-3 (Table 4.5), equivalent parts with an initial angle of 4 5 ° , by an average of 0 .34° . It is worth noting that parts W-2, W-3, F-2, and F-3 are 8-ply-parts processed with a 2-hold cure cycle without an F E P sheet. These results, illustrated in Figure 4.14, indicate that part angle has an effect on A8warpage, and that this effect could be related to part thickness, cure cycle, and tool surface. This effect, however, is not clearly understood at this time. 4.3.6 E F F E C T O F CURE C Y C L E ON W A R P A G E In experiment 1, parts processed using a 2-hold cycle have greater A0warpage than those processed using a 1-hold cycle by an average of 0 . 2 0 ° . The effect of cure cycle is also seen in experiment 2 (Table 4.2 and Figure 4.4). In experiment 2, the C-shaped parts prepared with a 2-hold cure cycle (RC-5, RC-6 , TS2-1, R C - 7 , RC-8 and TS2-4) have an average A6warpase of 0 .79° while those prepared with a 1-hold cure cycle (RC-1, RC-2 , RC-3 and RC-4) have an average A8warpage of 0 . 2 0 ° . Further examination of the results of experiment 2 reveals that for this experiment the effect of cure cycle is slightly greater for parts processed on aluminum tooling compared to steel tooling. Parts RC-5 , R C -6, and TS2-1, processed on aluminum tooling using a 2-hold cycle, have greater A6warpage than parts RC-1 and RC-2 , equivalent parts cured with a 1-hold cycle by 0 . 6 4 ° . When comparing parts RC-7 , RC-8 , and TS2-4, processed on steel tooling using a 2-hold cycle, with parts RC-3 and RC-4 , processed with a 1-hold cycle, the difference is 0 .54° . The effect of cure cycle is also seen when comparing other parts from experiment 1, 2, 5 and 6. Parts F-3, RL-1 and R L - 3 are equivalent to parts D O E - 1 2 , T A - 1 , and T A - 4 respectively, with the exception of their cure cycle. For all these - 6 5 -Chapter 4 Experimental - Spring-in Components parts, those processed using a 2-hold cycle have larger A9wurpage, with differences of 0 . 5 4 ° , 0 .68° , and 0 . 7 1 ° , respectively, compared to those processed using a 1-hold cycle. These results, however, do not indicate a greater A9warpage for parts processed on aluminum tooling compared to steel. It is worth noting that the parts from experiment 2, and parts F-3, R L - 1 , RL-3 , D O E - 1 2 , T A - 1 , and T A - 4 have a thickness of 8 plies and were processed without F E P . The effect of cure cycle on A8warpage is greater for these parts compared to the average effect seen in experiment 1, indicating that the effect of cure cycle is related to the part thickness as well as the tool surface. The parts from experiment 1 consist of a combination of 8 and 16-ply parts that were processed with or without an F E P sheet. This observation is consistent with the tool-part interaction mechanism. The results are plotted in Figure 4.15. 4.3.7 E F F E C T O F T O O L SURFACE ON W A R P A G E O f the studied parameters, tool surface has the most significant effect on A9warpage. In experiment 1, parts processed without the use of an F E P sheet have on average 0 .16° greater A9warpage than those processed with an F E P sheet. The results of experiment 3 show that parts processed without F E P have the greatest A6warpage (Table 4.3 and Figure 4.5). In experiment 3, parts processed without F E P have an average A9warpage of 0 .55° , while parts processed with F E P have an average A9warpage of 0 .11° . This effect is even more notable for thin parts. In experiment 3, thin parts processed without F E P have a greater A9warpage than those processed with F E P by an average of 0 . 6 2 ° , compared to a difference of only 0.03 for 16-ply parts. The effect of tool surface is even greater when comparing some of the parts from experiments 4 and 5. Part W-4, processed without F E P , has greater A9warpage than part W-5, an equivalent part processed with F E P by 1.03° . Similarly, part F-2, processed without an F E P sheet at the tool-part interface, has 0 .73° higher A9warpage than part F - l , an equivalent part processed with an F E P sheet. It is worth noting that parts F - l , F-2, W-4, and W-5 have 89 mm - 6 6 -Chapter 4 Experimental - Spring-in Components flanges c o m p a r e d to 57 m m for the parts f r o m experiment 3. T h e s e observations are illustrated in F i g u r e 4.16. F r o m these results, it is evident that the effect o f tool surface is more s ignif icant f o r thinner parts as w e l l as parts w i t h longer f langes . T h e effect o f part thickness and f lange length o n A8warpage can therefore be s igni f i cant ly reduced with the use o f an F E P sheet on the tool surface. 4.3.8 E F F E C T O F T O O L M A T E R I A L ON W A R P A G E T o o l material also has an effect on A8warpage, h o w e v e r the effect is not as s ignif icant as other parameters. In experiment 1, parts processed o n a l u m i n u m t o o l i n g have a larger A8warpage than those processed o n steel t o o l i n g b y an average o f 0 . 0 5 ° . In experiment 6, the average difference is 0 . 0 8 ° ( T a b l e 4.6 and F i g u r e 4.8). In experiment 2, the average di f ference between a l u m i n u m and steel t o o l i n g is 0 . 0 7 ° w h e n c o m p a r i n g the parts processed u s i n g a 1-hold cure c y c l e ( T a b l e 4.2 and F i g u r e 4.4). In the same experiment , the difference is 0 . 1 5 ° for parts processed u s i n g a 2 - h o l d cure c y c l e , indica t ing a relat ionship between the effect o f tool material and the cure c y c l e . In experiment 3, the average d i f f e r e n c e between steel a n d a l u m i n u m is 0 . 0 2 ° . W h e n c o m p a r i n g o n l y the 8-ply parts processed without F E P , o n the other hand, the dif ference is 0 . 1 4 ° , suggesting that the effect o f tool material is also dependent o n the t o o l i n g surface and the part thickness . T h e s e observations are s u m m a r i z e d in F i g u r e 4.17 and c a n be rat ional ized i n terms o f m e c h a n i c a l tool-part interaction. 4.4 S U M M A R Y T h e f o l l o w i n g c o n c l u s i o n s can be d r a w n f r o m the results presented i n this chapter. • S p r i n g - i n at the corner is typica l ly measured based on the def lec t ion o f the flanges away f r o m the corner . T h e results presented in this chapter s h o w that this m e a s u r i n g technique is not very precise because o f the c o m p o u n d i n g effect o f warpage and s p r i n g - i n . - 6 7 -Chapter 4 Experimental - Spring-in Components • Even when not directly observable with the naked eye, the presence of warpage was in many cases great enough to significantly increase the apparent measured spring-in • Spring-in can be decomposed in two distinct components: a true corner component, and a flange/web warpage component. • The corner component is mainly the result of material anisotropy and can be predicted with reasonable accuracy using geometric arguments for the longitudinal and transverse strain at the corner (Equation 2.1). This component is governed by the material properties, the lay-up of the laminate, and the angle of the corner. • The warpage component, on the other hand, is a result of mechanical interaction between the tool and the part during processing. Unlike the corner component, the warpage component is highly variable and depends on the part design as well as the process conditions. Parameters affecting the warpage component are: part thickness, flange length, cure cycle, tool material, and tool surface. - 6 8 -Chapter 4 Experimental - Spring-in components IT) 3 c/} c cu e o I o o 6J) c c CD e CD cu x a <D as < <-> 3 cj o U 5 .2 a. 3 >> cd —1 Cu -C NO CN ON WO ON C^  CN -* NO ON oo Tt NO CN Tt CN wo o NO CN o m o CO o d d d d d d d d d d d d d d d ON co wo oo NO uo ON uo ON uo -* o ON 00 uo CN uo NO ON oo cn oo CO wo NO O d d d d d d d — d d d d — d d o co O o o o o o o CN ro •O o o oo o ON o o O CO o ro O ¥l ?l ?l ?l ?i ?l ?i ¥l ?l ? l ?! ?, •a- o o — ON NO — O N ^ r o r o — o — NO-^-'d- — COVOOCNWOOOONOOTtt^ON d — — _ ; _ ; o — : — : — ' d o — ^ d d CNCNCNCNCNCNCNCN a. cu a. a. cu cu cu CN CN + + C C CN CJ CJ ' 60 00 u cs. + + C CN CN C C N CS CJ ^ CO 1) ^ 60 *J « 60 60 w w 60 S S3 60 n tu Cd ™ O u u CJ S3 o u SS _u a cc _u 13 "u os ei cj cj c c cj cj 60 b0 So ca ta d cj CJ CJ CJ S 15 u u cd cd _> _> CJ CJ OJ 0i c c CJ CJ CJ 60 60 K 60 cd cd cd cd C CN C C CN CJ v—' CJ CJ s—' 60 ~ 60 60 ~ cd c cd cd c U U Mil cd cd cd CJ CJ CJ ^ CJSS u SS o u cj 02 8 OS OJ CJ CJ OS 05 CJ CJ OJ OJ E s 3 3 _ „ „ „ C3 C cj cj CJ cj '£ '5 V V V V § 3 M 71 Cfl M E E 3 3 C C 3 3 < < E E E E 3 3 3 3 _ _ C C C C ~S ~S | | § J £ £ <<< < woowooowoowoowoowowoouoo ^tONTtONONTj-ONTtON^t-ONTl-TtONTtON V l V l i n ^ M O O O O O O M M Q O O O I C l i n i n i O ON_ ox S «|2 m" o o t t £ cf O o ° o ° o ON ox ON ox ON ON £ »in ^" m vi £ oo uo ^ >£ oc g o t t t t o o t t o o "2 <rT "2 " "2 wo l — ' 1 — 1 X. Tf rt O o" O o o o" J U - J U - J O J U U J U J U J U - J —'CNroTtcnNor^ooov O —• CN ro Tt wo NO g o o o o o o o o g g g g g g g Q Q Q O Q Q Q Q Q Q Q Q Q Q Q Q u cj J S C/5 CD C _u ">N a. o CD c J 3 o c u -a cu W ^ CD > •5 "° a. CD a. CD r-60 C * & % 3 £ - E 3 « C3 3 8 g ° ^ U. !_ CD 3 ^ • c/> S C« CD i s » 1 s -a o ca u xB O us C CD CD on S3 CD ca i— ex cx S 2 IS c •I I ID XI CD CD CD H H -69-Chapter 4 Experimental - Spring-in components 3 c/3 CD C CD C o t o u c C N c CD CD Cu X W C N ed H NO CN 00 00 r» oo CN Os ON O m m 00 oo r» oo o 00 m cn NO O o d d d d d d d d d d d d cn oo Tt o\ ON OO NO OO m ON m oo CN SO oo in ON NO ON ON in Tt o o d d d d d d d d d d d d o cs o o m o CN O m O CN O Tt o CN O o CN O cn O Tt o m o ? i ? l ?, ?, ? l ? l ? l ? l ? l ? l ¥ i ?, ? l ?, o CN NO o Tt o CN NO ON 00 Tt SO CN Cl NO cn Cn NO 00 NO Tt ro r~ m CJ o U - e at p oo _B c c-3 >s a J I CU . Q <u ° <u 5 S ° o S > - Z > - Z > . Z Z > Z Z > - « 5 c n c n c n c n c n c n c n c n c n c n c n c n c n c n c c c c c c c c c c c c c c b 0 M N > b O b O b O b e b O b O b O b O b O b 0 b o | rtcdcclcdcaajcacdcdcdccjcdcaca l^piOiOSCiSOiOiOiCiiDiQiQiQiDi E £ 3 3 < < E E E 3 3 3 <<< ~at at at at at at ai c/5 E E 3 3 o o o o o o o o o o o o o o O N O N O N O N O N O N O N O N O N O N O N O N O N O N 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 i n i n i n i n i n i n i r i i n i n i n i n i n i n i n oo oo oo cc oo oo oo oo oo oo oc O O O O O O O O O O O O O O O U O U U U U U U U J J J J — c N c n T t m N O - 7 r - ~ . o o T f — c N c n T t u u u u u o S u u S J J c i a i c i c i c i c i H ^ ^ H 0 2 1 ^ ^ 1 ^ 3 ed. val pli •o pli CD ex tH 3 cfl gent mea C3 at E 3 ca C _CD S—. o "3 c o i Crt ue o "ca o > of ed u i_ CD 3 XI c/: E c3 CD nu £ E Xj 3 CA E CJ '>< O den B sis CN sis CD X ! c c CD CD S3 CD u a. D, c CD u IS C wit OW CD sh XI CD um bO c 03 s u CD CD x: x: H H -70-Chapter 4 Experimental - Spring-in components CJ X . 3 in CD c CD o a. o o bO G •c Cu c CD CD a, X m H 0) O u cs ON ON 00 cs >n cs Tt o cs o m in oo oo oo cn r-cn ON 00 Tt o © d d d d d d d d O d d d d c~ o m 00 m oo m ON 00 cs oo oo 00 in p~ r~ 00 00 m NO o NO 00 d d d d d d d d d d d d d d cn O cs o o o <N O cs o cs o cs o in o cs o cs o Tt o »n o cs o ?l ?l ?l ?l ?i ?: ?l ?l ?l ¥i ?l ?l ON OO Tt NO = o cs CS ON ON o ON NO cn cn NO >n m cs Tt 00 oo >n oo d d d d d d CS CS cs cs cs CS CS cs cs <s cs CN cs -— m 60 13 C 60 ca c p E 2 Si —1 C Q. ca ca 0. J = cn cn cn cn *—' *— e c B c CJ CJ CJ CJ 60 60 60 60 ca ca ca ca CJ CJ o CJ CO CO to CO ca ca ca ca 2 2 2 2 "u "53 T3 T3 Oi Oi Oi Oi 0 - 0 - 0 -+ + + cn cn cn m 0 - 0 - 0 -+ + + cn cn cn G cn cn cn cj o u cj * j * j * j 60 60 60 60*-ccccacacacac cj cj cj Vi Vi Vi 2 2 2 "u "o "CJ oi oi oi to CO CO Vi *~ ca ca ca ca ca ca ca ju u _U JJ tj u cj CJ cj cj cj %\ %\ £ Oi Oi Oi Oi CJ CJ CJ Oi Oi Oi E E S E E E E 3 3 3 3 3 3 3 3 3 3 3 < < < < < < < CJ ~o "o "u "cj "u "CJ s B u u u u u c/5 55 55 55 55 55 o o o o o ON ON ON ON ON O O O O O O O ON ON ON ON ON ON ON O O ON ON r- r» t— r- c— r- r-m m m in t/-> in m co ~ ~ o p ^ o o g j O o S P j O o O o o U U U U U U U U U O U U U U — CN c n r ? o o ' * r t , 9 m N O a. a, C3 c 60 cd e o S3 a. c CJ J 3 CO > CJ CS CJ E E 3 £ £ i CJ _3 C3 > -a ca CJ E £ 3 X ca c CJ CA Cj t« Cu CJ e o X ! co CJ 60 C C3 CJ CJ JC vi CJ C 2 >N U o H 3 CD C CD G O Cu s o cj bD G C CD CD Cu X W T t T t _CD IS cd H < u 1 " M r o- S -a « , E C 60 g ca c P E 2 £ C D. ca ca 0- xj cs cn Tt CN O — — — Tt w~i cs o cs cs NO o o cs o ON ON — cn O O O —' — — — cn cn CN cn O O O O O ¥i fi ?, £ ¥i — cs cs cs —• C S CN CN C S C S CS CS CS CS c c c c cs CJ CJ CJ CJ ^ 60 60 60 bO 4-. ca ca ca ca c « „, „ 60| ca ca ca ca ca u 2 2 u o "u " U "CJ "CJ 55 Oi Oi Oi Oi g "3 Oi E E _ _ 3 _ 3 CJ CJ C cj 3 3 < O O O O O ON ON ON ON ON OO 00 OO 00 CO CO 0 0 ON ON in in Tt Tt m" m" Tt Tt 0 O -J U J -J —1 — cs cn Tt m * Se gt gt -71-Chapter 4 Experimental - Spring-in components c CD c o cu e o o C c CD <D X W X) H OS 2 u c 2 o. s .a a, E « £ CJ c« ca 0. J= TT t— — CN ^ CO I— CN d d d -oo — NO T t TT) 0 0 t— d o d o — m CN oo NO Tt in CN o —' — —; CN CN CN CN c c CJ CJ , 0 0 6 0 ca ca LT n UJ UJ Of* C/> C/3 C/3 ca ca ca ca cj ju _ej _u SS "u ~3 "«3 8 05 OS OS E E 3 3 5 < iri in IT , in ^ ^ Tt OO 0 0 0 0 0 0 O 0\ ON CT\ OO 0 0 0 0 oo o > oo V"T o t ' 1 m" Tt U U —1 -J — CN co Tt l i [ll [i. tL. cu a CL o .2 "a, a. C3 c u 6 0 ca cj c D ca o-e IS c CD c o Cu E o cj c c CD S G CD X W NO •>*' _CD a H CJ 6 0 -c ; ca 05 cj cj O 3 J = t5 ^ _| ca Mc?1 0- S v-§ £ l l 4-. U ca ca 0. J5 0-C N — — Tt O rO — — CN O — —' d d d d d d Tt NO o oo NO — OO NO O OO Tt o d d d d d d O CN CN CN O — o o o o o o d d d d d d cn — CN NO co O r- CN ON U~I — - d d d d d CN CN CN CN CN CN c c c c c c CJ CJ6 0 6 0 ca ca cj cj t/i Vi ca ca 05 05 CJ CJ CJ6 0 6 0 6 0 ca ca ca cj cj cj Vi Vi Vi "o3 cj *a> 05 05 05 ( B u CJ c/5 0 0 t55 E E 3 3 c c £ 'g 3 3 O W l n O » 1 ON Tt ° Tt oo oo oo oo 0 0 oo oo oo oo oo oo oc| o o o o o o J J J J J J — CN ro Tt m NO < < < < < < H H H H H c--3 -6 ? CD > •5 7 3 Q, CJ CL >-« s 1 s & s ca _ C/3 CJ S g C 5 CU crt a cu ca D . CL E IS c 1 I <B "> X) cu cu CJ JS -C H H cu I CU c/l -72-Chapter 4 Experimental - Spring-in Components 4.6 FIGURES 0.25 _ 0.2 e^O.15 d .2 o 0 1 Q 0.05 0 comer flange tip — t y / \ AA - 0 48° / \ ^ w^eb warpage- u.to A i i 0 20 40 60 x (mm) Figure 4.1. Flange warpage of part TS2-4. 80 100 AOflange A warpage tool shape 'flange B warpage part tool shape Figure 4.2. Spring-in components (not to scale). -73-Chapter 4 Experimental - Spring-in Components o .5 1 60 •go.: ex, Mo. 0. 0 0 .6 .4 .2 .0 H 8 6 4 2 0 • Warpage component • Corner component — A B C T E + A0, 45° parts Jcs AG, CTE liiw mm T 90° parts • PJ o Q PJ O Q PJ o Q PJ O Q cn pj o Q oo PJ o Q tu o Q PJ O Q PJ O Q PJ O Q PJ O Q PJ O Q E-7 E-9 PJ E-7 E-9 o O O Q Q Q PJ o Q [0]„ [0,45, -45, 90]n [0]„ V [0,45,-45,90]n Figure 4.3. Experiment 1 - spring-in components. Solid and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. The ranges shown represents 1/2 (maximum measured total spring-in value- minimum measured total spring-in value). Parts are described in Table 4.1. 2.2 T 2.0 1.8 1.6 •5 1.2 .51.0 #0.8 0.6 0.4 0.2 0.0 ED Warpage component • Corner component A 0 C T E + A 0 C S A 0 C T E rh lllif Si U erf u Crf u erf u erf as H u erf c5 CN ?5 I r*4 U erf C N H U erf u erf CN oo H -J erf —i erf erf j erf Figure 4.4. Experiment 2 - spring-in components. Solid and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. The ranges shown represents 1/2 (maximum measured total spring-in value- minimum measured total spring-in value). Parts are described in Table 4.2. - 7 4 -Chapter 4 Experimental - Spring-in Components 2.2 2.0 1.8 1.6 HI •Sl.2 §1-0 ^0.8 HI 0.6 0.4 0.2 0.0 Aluminum tooling Steel tooling ED Warpage component H Corner component A 9 C T E + A0 C S A 9 C T E u <N or> H oo r-ro ro c/3 CN f— oo OO T t U C N u 2 0 0 C N C— OO H 0 0 CO Figure 4.5. Experiment 3 - spring-in components. Solid and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. The ranges shown represents 1/2 (maximum measured total spring-in value- minimum measured total spring-in value). Parts are described in Table 4.3. 0 Warpage component • Corner component — A 9 C X E + AO, Jcs AG, C T E as [0]„ < mSm 5» [0/45/-45/90]n Figure 4.6. Experiment 4 - spring-in components. Solid and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. The ranges shown represents 1/2 (maximum measured total spring-in value- minimum measured total spring-in value). Parts are described in Table 4.4. - 7 5 -Chapter 4 Experimental - Spring-in Components 2.5 2.0 5-1.5 6 0 C 0.5 0.0 ED Warpage component • Corner component a 0 c x e + a 0 c s a 0 c t e C/WSi/I-l-l' Oil, Si K ti. [01„ U»i\\ K P-[0/45/-45/90]n Figure 4.7. Experiment 5 - spring-in components. Solid and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. The ranges shown represents 1/2 (maximum measured total spring-in value- minimum measured total spring-in value). Parts are described in Table 4.5. 1.4 -| 1.2 1.0 H 11 Warpage component H Corner component I < HUB wSm < — < •y. < o 3 < o < Figure 4.8. Experiment 6 - spring-in components. Solid and dashed lines represent spring-in predictions using Equation 2.1, see Table 3.9. The ranges shown represents 1/2 (maximum measured total spring-in value- minimum measured total spring-in value). Parts are described in Table 4.6. < - 7 6 -Chapter 4 Experimental - Spring-in Components 1.2 1.0 0.8 I 0.6 s < 0.4 0.2 0.0 Figure S C 2 holds no FEP 8 plies 57 mm E x p 1 (average) E x p 2 (2 holds) 4.9. Effect of part shape on warpage. The ranges shown represent the maximum and minimum values for each data set. Parts are described in Tables 4.1 and 4.2. 1.8 1.6 1.4 1.2 '~i i .o a & i 0.8 0.6 0.4 0.2 0.0 m [0,45,-45,90]ns U[0]n E x p 1 (average) 2 holds no FEP 8 plies 89 mm E x p 4 2 holds no FEP 8 plies 89 mm L- USB shaped shaped -W-4 W-2 F-4 E x p 5 Figure 4.10. Effect of lay-up on warpage. The ranges shown represent the maximum and minimum values for each data set. Parts are described in Tables 4.1, 4.4, and 4.5. - 7 7 -Chapter 4 Experimental - Spring-in Components CD < 1.6 1.4 1.2 •s 1.0 ©0 I 0.8 0.6 0.4 H 0.2 0.0 H 8 plies • 16 plies 2 holds no FEP 89 mm 2 holds no FEP 57 mm 2 holds FEP 57 mm W-2 : W-l Exp 1 (average) Exp 3 (FEP) Exp 3 (no FEP) Exp 4 Figure 4.11. Effect of laminate thickness on warpage. The ranges shown represent the maximum and minimum values for each data set. Parts are described in Tables 4.1, 4.3, and 4.4. 1.4 1.2 1.0-1 i 0 6 H 0.4 0.2 0.0 100 120 140 160 180 200 (mm) Figure 4.12. Warpage component, A6warpage, vs. length of influence, LM„ for all the parts presented in Tables 4.1 to 4.6. - 7 8 -Chapter 4 Experimental - Spring-in Components CD < 1.6 1.4 1.2 . 1.0 tu aa I 0.8 0.6 0.4 0.2 0.0 Ll Short • L o n g 2 holds no FEP J 6 plies 159 mm TS-4 T.S 2-5 W-l 2 holds no FEP 8 plies 159-mm RC-8 TS2-4 178 mm 1 W-2 2 holds no FEP 8 plies 159 mm RC-6 TS2-1 191 mm W-3 Figure 4.13. Effect of length of influence, Lto„ on warpage. The ranges shown represent the maximum and minimum measured values for each data set. Parts are described in Tables 4.3 and 4.4. CT> < 1.6 1.4 1.2 1.0 I 0.8 0.6 0.4 0.2 H 0.0 H 4 5 ° • 9 0 ° E x p 1 (average) 1 hold no FEP 8 plies 57 mm 2 holds no FEP 8 plies 89 mm F-2 W-2 1 E x p 6 2 holds no FEP 8 plies 89 mm F 3 •Hi \\-Figure 4.14. Effect of initial part angle on warpage. The ranges shown represent the maximum and minimum measured values for each data set. Parts are described in Tables 4.1 and 4.4 to 4.6. - 7 9 -Chapter 4 Experimental - Spring-in Components GO 1 hold • 2 holds no FEP aluminum 8 plies 57 mm^ no FEP steel 8 plies 57 mnL. no FEP aluminum 8 plies 89 mm DOE 12 F-3 no FEP aluminum 8 plies 57 mm I \ I •PI RL-1. no FEP steel 8 plies 57 mm TA-4 RL-3 E x p 1 E x p 2 E x p 2 (average) (aluminum (steel C-shaped) C-shaped) Figure 4.15. Effect of cure cycle on warpage. The ranges shown represent the maximum and minimum measured values for each data set. Parts are described in Tables 4.1, 4.2, and 4.5. H Release agent + F E P • Release agent E x p 1 (average) 2 holds 16 plies 57 mm E x p 3 (16 plies) 2 holds 8 plies 57 mm 2 holds 8 plies 89 mm W-5- 4 ' E x p 3 (8 plies) 2 holds 8 plies 89 mm F-2 Figure 4.16. Effect of tool surface on warpage. The ranges shown represent the maximum and minimum measured values for each data set. Parts are described in Tables 4.1 and 4.3 to 4.5. - 8 0 -Chapter 4 Experimental - Spring-in Components 0.8 A e- 0.6 CD < 0.4 A 0.2 0.0 E x p 1 (average) / hold no FEP 8 plies 57 mm 1 hold no FEP 8 plies 57 mm 2 holds no FEP 8 plies 57 mm 2 holds 57 mm E x p 6 E x p 2 (1 hold) E x p 2 (2 hold) E x p 3 2 holds no FEP 8 plies 57 mm mm. •Hp E x p 3 (8 plies no F E P ) Figure 4.17. Effect of tool material on warpage. The ranges shown represent the maximum and minimum measured values for each data set. Parts are described in Tables 4.1 to 4.3, and 4.6. - 8 1 -C H A P T E R 5: C O N C L U S I O N S A N D F U T U R E W O R K The objective of this study was to determine the parameters and mechanisms that affect spring-in of angled thermoset composite parts during processing. The main findings from the current work are as follows: • Spring-in varies greatly as a function of design and process parameters. • For practical reasons, spring-in is usually measured based on the deflection of the flanges away from the corner. The current work showed that spring-in measured this way is highly sensitive to flange/web warpage. Even when the warpage is too small to be perceived with the naked eye, it can significantly contribute to spring-in. • A measurement method was developed to separate spring-in into two distinct components: a corner component and a flange/web warpage component. • The corner component of spring-in was shown to be primarily the product of thermal and chemical strain anisotropy. This component was found to be repeatable for a given part design and cure temperature, and can be predicted using a simple equation (Equation 2.1). For a given material and cure temperature, parameters affecting the corner component are the initial part angle and the part lay-up. • The warpage component of spring-in was shown to be mainly the result of tool-part mechanical interaction during processing. This component was found to be strongly dependent on process conditions. A number of parameters were found to affect the warpage component, including the laminate thickness and flange length, as well as the cure cycle, the tool material, and the tool - 8 2 -Chapter 5 Conclusions and Future Work surface condition. While the laminate thickness and flange length are inherent elements of a part design and thus cannot be modified during processing, the process conditions, on the other hand, can be altered to minimize this component of spring-in. • Table 5.1 summarizes the general importance of the effect of each parameter on total spring-in, and indicates which component of spring-in (i.e. corner or warpage) is affected by each parameter. • Based on the results of this study, three simple rules to minimize the variability of spring-in in production are proposed: 1. Control the cure cycle to avoid early gelation. Ensure that all parts gel on the final hold temperature. 2. Select tooling material that as closely as possible matches the coefficient of thermal expansion of the part in the longitudinal direction, i.e. perpendicular to the corner axis. 3. Use a fluorinated ethylene propylene sheet, or other release ply, in addition to release agent at the tool-part interface. This is especially important when processing long and thin parts. 5.1 F U T U R E W O R K The following recommendations for future work are based on observations made during the course of the current study: • Further investigation of the effect of lay-up. The number of quasi-isotropic parts prepared in this study is insufficient to draw clear conclusions regarding the effect of lay-up on the warpage component. Since most composite components manufactured in industry have multi-directional lay-ups, it would be valuable to expand this study with additional quasi-isotropic parts as well as other multi-directional lay-ups. - 8 3 -Chapter 5 Conclusions and Future Work • Investigation of size effects. The current study focused on parts with a thickness of up to 16 plies, with a maximum flange length of 89 mm. Within this range of geometries, the warpage component of spring-in was found to be significantly affected by thickness and length, with thinner and longer parts exhibiting more spring-in at certain process conditions. Future experimental work should examine a wider range of industrially relevant geometries. • Study of batch-to-batch variability. The presence of batch-to-batch variability has been observed in this study. This issue, however, is still not well understood. Future work should investigate the main factors responsible for batch-to-batch variability. • Examination of residual stress development during cure. The current work examined the final shape of parts of various geometries, processed under various conditions. Based on these results, the main parameters and driving mechanisms affecting spring-in were identified and described qualitatively. The development of the stresses responsible for spring-in and their distribution within the parts were not examined and should be studied in future experimental work. - 8 4 -Chapter 5 Conclusions and Future Work 5.2 T A B L E S Table 5.1. Summary of the contribution of the studied parameters on spring-in. Parameter Importance7 Affects A0c„rner Affects A0warpaee Type Tool surface High X Extrinsic (FEP/Release agent) Thickness High X Intrinsic (8/16 plies) Length of influence High X Intrinsic (-114 mm to 191 mm) Cure cycle High X Extrinsic (1/2-hold) Lay-up Medium X f Intrinsic ([0]A0,45,-45,90]ns) Part angle Medium X f Intrinsic (90745°) Tool material Medium X Extrinsic (Aluminum/Steel) Part shape Low Intrinsic (C/L) Rubber caul Low Extrinsic (With/Without) ' Indication of the impact of the parameter on total spring-in, based on the largest observed effects shown in Figures 3.15 to 3.22. Factors with a maximum effect of less than 0.20° are here classified as of "Low" importance, those with a maximum effect between 0.20° and 0.50° are classified as of "Medium" importance, and those with a maximum effect greater than 0.50° are classified as of "High" importance. 2 Question mark indicates that the effect is not clear based on the data available. r - 8 5 -R E F E R E N C E S A g a r w a l B D , B r o u t m a n L J . A n a l y s i s and performance o f f iber composi tes . J o h n W i l e y & S o n s , Inc., 1990. B e r g l u n d L A , K e n n y J M . P r o c e s s i n g science for h i g h performance thermoset composi tes . S A M P E Journal , V o l . 27, N o . 2, 1991. Boget t i T A , G i l l e s p i e J W , Jr. P r o c e s s - i n d u c e d stress and d e f o r m a t i o n in thick-sec t ion thermoset c o m p o s i t e laminates . J o u r n a l o f C o m p o s i t e M a t e r i a l s , V o l . 2 6 , N o . 5 , pp .626-659 , 1992. B o x G E P , H u n t e r W G , H u n t e r J S . Statistics for experimenters . J o h n W i l e y & S o n s , Inc., 1978. C a n n M T , A d a m s D O . E f f e c t o f part-tool interaction o n cure distort ion o f flat c o m p o s i t e laminates. 46th International S A M P E S y m p o s i u m and E x h i b i t i o n , L o n g B e a c h , C A , 2001 . F e r n l u n d G , Poursart ip A . T h e effect o f t o o l i n g material , cure c y c l e , and tool surface f i n i s h on spr ing-in of autoclave processed c u r v e d c o m p o s i t e parts. 12th International C o n f e r e n c e o n C o m p o s i t e M a t e r i a l s , 1999. F e r n l u n d G , R a h m a n N , C o u r d j i R , et a l . , E x p e r i m e n t a l and n u m e r i c a l study o f the effect o f cure c y c l e , tool surface, geometry , and the l a y - u p o n the d i m e n s i o n a l stability o f autoclave-processed c o m p o s i t e parts. S u b m i t t e d to C o m p o s i t e s Part A . F l a n a g a n R . T h e d i m e n s i o n a l stability o f c o m p o s i t e laminates and structures. P h . D . T h e s i s , Queen's U n i v e r s i t y o f Belfas t , 1997. G a n l e y J M , M a j i A K , H u y b r e c h t s S. E x p l a i n i n g s p r i n g - i n in f i lament w o u n d c a r b o n f iber /epoxy composi tes . J o u r n a l o f C o m p o s i t e M a t e r i a l s , V o l . 3 4 , N o . 1 4 , pp.1216-1239 , 2000. H u a n g C K , Y a n g S Y . Short c o m m u n i c a t i o n - w a r p i n g in a d v a n c e d c o m p o s i t e tools with v a r y i n g angles and r a d i i . C o m p o s i t e s Part A , V o l . 2 8 A , pp.891-893 , 1997. H u l l D , C l y n e T W . A n introduct ion to c o m p o s i t e materials . C a m b r i d g e U n i v e r s i t y Press, 1996. Jain L K , L u t t o n B G , M a i Y W , Paton R . Stresses and deformat ions i n d u c e d d u r i n g manufactur ing , part II: a study o f the s p r i n g - i n p h e n o m e n o n . Journal o f C o m p o s i t e M a t e r i a l s , V o l . 3 1 , N o . 7 , pp.696-7 1 9 , 1 9 9 7 . Johnston A . A n integrated m o d e l o f the d e v e l o p m e n t o f process i n d u c e d d e f o r m a t i o n i n autoclave process ing o f c o m p o s i t e structures. P h . D . T h e s i s , U n i v e r s i t y o f B r i t i s h C o l u m b i a , 1997. Johnston A , H u b e r t P, F e r n l u n d G , V a z i r i R , Poursart ip A . Process m o d e l i n g o f c o m p o s i t e structures e m p l o y i n g a vir tual autoclave concept . S c i e n c e and E n g i n e e r i n g o f C o m p o s i t e M a t e r i a l s , V o l . 5 , N o . 3 - 4 , pp.235-252 , 1996. - 8 6 -References K e n n y J M , T r i v i s a n o A , B e r g l u n d L A . C h e m o r h e o l o g i c a l and dielectr ic b e h a v i o u r o f the e p o x y matrix i n a c a r b o n f ibre prepreg . S A M P E Journal , V o l . 2 7 , N o . 2 , pp.39-45 , 1991. K i m K S , H a h n H T . R e s i d u a l stress d e v e l o p m e n t d u r i n g p r o c e s s i n g o f graphi te /epoxy composi tes . C o m p o s i t e S c i e n c e a n d T e c h n o l o g y , V o l . 3 6 , pp.121-132, 1989. K i m Y K . P r o c e s s - i n d u c e d viscoelast ic residual stress analysis o f graphi te -epoxy c o m p o s i t e structures. P h . D . T h e s i s , U n i v e r s i t y o f I l l inois at U r b a n a - C h a m p a i g n , 1996. Montserra t S. V i t r i f i c a t i o n and further structural relaxation in the isothermal c u r i n g o f an e p o x y resin. Journal o f A p p l i e d P o l y m e r S c i e n c e , V o l . 4 4 , pp.545-554, 1992. N e l s o n R H , C a i r n s D S . P r e d i c t i o n o f d i m e n s i o n a l changes in c o m p o s i t e laminates d u r i n g cure. International S A M P E S y m p o s i u m and E x h i b i t i o n C o v i n a , C a , B o o k 2, pp.2397-2410 , 1989. Pagano N J . T h i c k n e s s e x p a n s i o n coeff ic ients o f c o m p o s i t e laminates . J o u r n a l o f C o m p o s i t e M a t e r i a l s , V o l . 8 , pp.310-312 , 1974. P a g l i u s o S. W a r p a g e , a nightmare for c o m p o s i t e parts producers . Progress i n S c i e n c e and E n g i n e e r i n g o f C o m p o s i t e s , p p . 1617-1623, 1982. Patterson J M , S p r i n g e r G S , K o l l a r L P . E x p e r i m e n t a l observations o f the s p r i n g - i n p h e n o m e n o n . C o m p o s i t e s , D e s i g n , M a n u f a c t u r e , and A p p l i c a t i o n , 8th Interantional C o n f e r e n c e o n C o m p o s i t e M a t e r i a l s , 1991. R a d f o r d D W . M a n u f a c t u r i n g warpage i n flat u n i - a x i a l c o m p o s i t e laminates . A m e r i c a n Soc ie ty for C o m p o s i t e s , 5th T e c h n i c a l C o n f e r e n c e , pp.686-695 , 1990. R a d f o r d D W . C u r e shrinkage i n d u c e d warpage in flat u n i - a x i a l composi tes . Journal o f C o m p o s i t e s T e c h n o l o g y & R e s e a r c h , V o l . 1 5 , N o . 4 , pp.290-296, 1993. R a d f o r d D W . V o l u m e fract ion gradient i n d u c e d warpage i n c u r v e d c o m p o s i t e plates. C o m p o s i t e s E n g i n e e r i n g , V o l . 5 , N o . 7 , pp.923-934 , 1995. R a d f o r d D W , D i e f e n d o r f R J . S h a p e instabilities in composi tes result ing f r o m laminate anisotropy. Journal o f R e i n f o r c e d Plastics and C o m p o s i t e s 1995, V o l . 1 2 , pp.58-75 , 1995. R a d f o r d D W , R e n n i c k T . Separat ing sources o f m a n u f a c t u r i n g distort ion in laminated composi tes . Journal o f R e i n f o r c e d Plast ics and C o m p o s i t e s , V o l . 1 9 , N o . 8 , PP .621-641 , 2000. R e n n i c k T , R a d f o r d D W . C o m p o n e n t s o f m a n u f a c t u r i n g distort ion i n c a r b o n f i b r e / e p o x y angle brackets. 28th International S A M P E T e c h n i c a l C o n f e r e n c e , p p . 189-197, 1996. R i d g a r d C . A c c u r a c y and distort ion o f c o m p o s i t e parts and tools : causes and solutions . T o o l i n g for C o m p o s i t e 9 3 , 1993. Sarrazin H , K i m B , A h n S H , S p r i n g e r G S . E f f e c t s o f p r o c e s s i n g temperature and l a y - u p on springback. J o u r n a l o f C o m p o s i t e M a t e r i a l s , V o l . 2 9 , N o . 1 0 , pp.1278-1294 , 1995. - 8 7 -References Stephan.A, Schwinge E, Muller J, Ory H. On the springback effect of CFRP stringers: an experimental, analytical and numerical analysis. 28th International S A M P E Technical Conference, pp.245-254, 1996. Svanberg JM, Holmberg JA. An experimental investigation on mechanisms for manufacturing induced distortions in homogeneous and balanced laminates. Composites Part A, Vol.32, pp.827-838, 2001. Twigg G A . Tool-part interaction in composites processing. M.A.Sc. Thesis, The University of British Columbia, 2001. White SR, Hahn HT. Cure cycle optimization for the reduction of processing-induced residual stresses in composite materials. Journal of Composite Materials, Vol.27, No.14, pp.1352-1378, 1993. Wiersma HW, Peeters JB, Akkerman R. Prediction of springforward in continuous-fibre/polymer L-shaped parts. Composites Part A, Vol.29 A, No. 11, pp. 1333-1342, 1998. Yoon KJ, Kim PJ. Processing induced distortion of L-section thermoplastic laminate structure. Journal of Materials Processing and Manufacturing Science, Vol.5, pp.317-327, 1997. Zhu Q, L i M , Geubelle PH, Tucker C L . Dimensional accuracy of thermoset composites: simulation of process-induces residual stress. To appear in Journal of Composite Materials. - 8 8 -A P P E N D I X A : V A R I A B I L I T Y A . l S P R I N G - I N V A R I A B I L I T Y A . 1.1 E F F E C T O F M O I S T U R E The effect of moisture on part shape was examined to ensure consistency of the angle measurements. After processing, the parts from experiment 1 had been exposed to ambient air for a minimum of 15 days (parts DOE-3, DOE-4, DOE-5 and DOE-6) up to a maximum of 83 days (parts DOE-11*, DOE-12*, DOE-13*, and DOE-14*)1 before being measured. To determine the effect of moisture absorption on spring-in, each part was measured before and after drying the specimens in an oven. The results are shown in Table A . l and Figure A . l . Drying was achieved by placing the parts inside an oven at 75°C ± 10°C, recording their weight intermittently until the part weight had stabilized. The drying process lasted for 83 days. The parts most affected by the drying process were parts DOE-2, DOE-7, and DOE-14*. These parts saw an increase in spring-in of 0.16°, 0.11°, and 0.14° after drying respectively. These are 90° parts that were exposed to ambient air for the longest periods of time prior to measurement. This increase is consistent with Equation 1 as moisture absorption causes through-thickness swelling and spring-out of the laminate. Furthermore, based on Equation 1, the increase of spring-in due to anisotropy should be greater for 90° parts than for 45° parts. The overall effect of moisture on the spring-in angle was quite small. On average, spring-in increased after drying by 0.04°, which is less than the measurement accuracy (Section A. 1.2). Nevertheless, to reduce experimental error, only the measurements done on dried parts were used for the analysis of 1 Parts DOE-11*. DOE-12*, DOE-13* and DOE-14* are replicated of parts DOE-11, DOE-12, DOE-13 and DOE-14, respectively, and were processed in a separate autoclave run. - 8 9 -Appendix A Variability experiment 1. T h e parts prepared in experiments 2 to 6 were measured w i t h i n 7 days after b e i n g processed to m i n i m i z e the effect o f moisture absorpt ion. A.1.2 M E A S U R E M E N T VARIABILITY E a c h part was measured a m i n i m u m o f three times u s i n g a m i n i m u m o f two scanned images , taken at different locat ions on the scanner. T h e measurement variabi l i ty is s h o w n f o r each part in terms o f a range between the m i n i m u m a n d the m a x i m u m measurement , i.e. ( m a x - m i n ) / 2 . T h e measured ranges, s h o w n in T a b l e s 3.10 to 3.17, T a b l e s 4.1 to 4.6, F igures 3.9 to 3.14, and F i g u r e s 4.3 to 4.8 are w e l l b e l o w ± 0 . 1 0 ° for almost al l parts. Part D O E - 8 , h o w e v e r had a measurement variabil i ty o f + 0 . 2 0 ° . C l o s e r e x a m i n a t i o n s h o w e d that the spr ing- in o f the two flanges o f part D O E - 8 d i f f e r e d by 0 . 3 7 ° . T h i s a s y m m e t r y c o u l d be d u e to an u n e v e n dispers ion o f release agent o n the tool surface, or a possible m i s a l i g n m e n t o f the fibres a l o n g one or both flanges. T h e other C-parts are more symmetr i ca l , wi th a di f ference o f 0 . 0 1 ° to 0 . 1 0 ° between the s p r i n g - i n o f each f lange . F o r the results presented i n this study, the s p r i n g - i n value indicated for C-parts represents the average value for the two f langes . A.1.3 VARIABILITY B E T W E E N A U T O C L A V E R U N S A total o f ten parts were dupl ica ted i n different autoclave runs to investigate batch-to-batch variabi l i ty . A di f ference o f up to 0 . 2 7 ° was observed between c o r r e s p o n d i n g parts c u r e d in different autoclave runs (F igure A . 2 and T a b l e A . 2 ) . T h e batch-to-batch variabi l i ty f o r all the parts averaged 0 . 1 1 ° . T h e causes for this observed batch-to-batch variabil i ty are not unders tood . A.2 W A R P A G E M E A S U R E M E N T V A R I A B I L I T Y W a r p a g e c o m p o n e n t measurements were repeated for part W - 3 to evaluate the variabi l i ty associated wi th the measurement m e t h o d . A total o f 5 measurements were taken u s i n g two different digi tal - 9 0 -Appendix A Variability images o f the same cross sect ion: three measurements on image 1 and two measurements o n image 2. T h e results are s h o w n in T a b l e A.3. T h e warpage c o m p o n e n t measurement var iabi l i ty for part W-3 is approximate ly ± 0 . 0 9 ° . - 9 1 -Appendix A Variability A.3 T A B L E S T a b l e A . 1. E x p e r i m e n t 1 - s p r i n g - i n measurements before and after d r y i n g . T h e ranges s h o w n represent 1/2 ( m a x i m u m measured value - m i n i m u m measured value) . Parts are d e s c r i b e d i n T a b l e s 3.1 and 3.2. Before dryi ng After drying Days Change in Part ID Spring-in Range exposed to Spring- in Range spring-in C) C) ambient air C) C) C) D O E - 1 0.57 ± 0 . 0 3 64 0.65 ±0.01 0.08 D O E - 2 1.28 ± 0 . 0 6 64 1.44 ± 0 . 0 3 0.16 D O E - 3 1.39 ± 0 . 0 0 15 1.41 ±0 .01 0.02 D O E - 4 1.06 ± 0 . 0 8 15 1.10 ± 0 . 0 7 0.04 D O E - 5 1.37 ±0.01 15 1.37 ±0 .01 0.00 D O E - 6 0.61 ± 0 . 0 7 15 0.61 ± 0 . 0 7 0.00 D O E - 7 0.98 ±0 .01 64 1.09 ±0.01 0.11 D O E - 8 1.23 ± 0 . 1 8 64 1.26 ± 0 . 2 0 0.03 D O E - 9 1.45 ± 0 . 0 5 42 1.51 ± 0 . 0 3 0.06 D O E - 1 0 0.88 ± 0 . 0 2 41 0.89 ± 0 . 0 0 0.01 D O E - 1 1 1.03 ±0 .01 43 0.94 ± 0 . 0 0 0.09 D O E - 1 2 1.04 ± 0 . 0 0 43 1.00 ± 0 . 0 9 -0.04 D O E - 1 3 0.98 ± 0 . 0 6 43 1.03 ± 0 . 0 6 0.05 D O E - 1 4 1.37 ± 0 . 0 2 43 1.41 ± 0 . 0 4 0.04 D O E - 1 1 * 0.84 ± 0 . 0 4 83 0.76 ±0.01 -0.08 D O E - 1 2 * 0.99 ± 0 . 0 2 83 1.04 ± 0 . 0 3 0.05 D O E - 1 3 * 0.97 ± 0 . 0 6 83 1.04 ± 0 . 1 0 0.07 D O E - 1 4 * 1.25 ± 0 . 0 3 83 1.39 ± 0 . 0 2 0.14 D O E - 1 5 0.69 ± 0 . 0 4 42 0.70 ± 0 . 0 3 0.01 D O E - 1 6 0.91 ±0 .01 41 0.91 ± 0 . 0 3 0.00 T a b l e A . 2 . V a r i a b i l i t y between autoclave batches. D u p l i c a t e parts were prepared i n different autoclave runs. T h e ranges s h o w n represent 1/2 ( m a x i m u m measured value - m i n i m u m measured value) . Parts are descr ibed in T a b l e s 3.1 to 3.4. First part Duplicate part Part ID Spring-in Range Part ID Spring-in Range I Difference | (°) (°) (°) (°) (°) D O E - 1 1 0.94 ± 0 . 0 0 D O E - 1 1 * 0.76 ± 0 . 0 1 "0.18 D O E - 1 2 1.00 ± 0 . 0 9 D O E - 1 2 * 1.04 ± 0 . 0 3 0.04 D O E - 1 3 1.03 ± 0 . 0 6 D O E - 1 3 * 1.04 ± 0 . 1 0 0.01 D O E - 1 4 1.41 ± 0 . 0 4 D O E - 1 4 * 1.39 ± 0 . 0 2 0.02 R C - 6 1.89 ± 0 . 0 3 TS2-1 1.64 ± 0 . 0 2 0.25 TS-1 1.11 ±0.01 TS2-2 1.20 ±0.01 0.09 T S - 3 0.79 ± 0 . 0 2 TS2-3 0.9 ± 0 . 0 2 0.11 R C - 8 1.36 ± 0 . 0 2 TS2-4 1.63 ± 0 . 0 5 0.27 T S - 4 1.15 ± 0 . 0 2 TS2-5 1.25 ± 0 . 0 2 0.10 T S - 6 0.78 ± 0 . 0 5 TS2-6 0.85 ± 0 . 0 2 0.07 - 9 2 -Appendix A Variability T a b l e A . 3 . W a r p a g e c o m p o n e n t measurement var iabi l i ty f o r part W-3 . Measurement Image Total A0wari,age ( ° ) 1 1 1.04 2 1 1.11 3 1 1.17 4 2 0.98 5 2 1.06 Range = + (max - min) / 2 ± 0 . 0 9 - 9 3 -Appendix A Variability AA F I G U R E S 1.8 1-6 H 1.4 ^ . 1 - 2 .¥i.o OJ) •g0.8 a, ^ 0 . 6 0.4 0.2 0.0 ft ii i t j I B e f o r e d r y i n g 0 A f t e r d r y i n g i — CN co UJ UJ UJ UJ UJ UJ O O O O O O UJ w o o O Q UJ UJ O O UJ O Q UJ UJ o o UJ UJ UJ UJ O O O O UJ UJ o o F i g u r e A . 1. E x p e r i m e n t 1 - measured s p r i n g - i n before and after d r y i n g . Parts are descr ibed i n T a b l e s 3.1 and 3.2. T h e ranges s h o w n represent 1/2 ( m a x i m u m measured value - m i n i m u m measured value) . 1.6 1.4 H : 1 .2 O W CN * — CN Q 0 W Q c o * <—> c n U T OB Q O Q 2 ^ w Q VO —' • i I I 00 CN H oo H « C j <N 00 CN H oo H i 0O T t t j CN T t >0 00 CN F—i oo H \0 \o I I 00 C N H oo H F i g u r e A . 2 . S p r i n g - i n var iabi l i ty between autoclave batches. D u p l i c a t e parts were prepared i n different autoclave runs. Parts are descr ibed i n T a b l e s 3.1 to 3.4. T h e ranges s h o w n represent 1/2 ( m a x i m u m measured value - m i n i m u m measured value) . - 9 4 -A P P E N D I X B : I N V E S T I G A T I O N O F V I S C O E L A S T I C E F F E C T S B. l O B J E C T I V E The following experiment was designed and performed to investigate the presence of viscoelastic relaxation in a fully cured part within the time and temperature range encountered during autoclave processing. B.2 E X P E R I M E N T A L The spring-in of part V E - 1 (see Table B . l for part description) was initially measured at room temperature using the methods described in Chapter 3. The part was subsequently stretched onto the aluminum tooling on which it was processed, forcing each flange away from the tool with a steel wedge as illustrated in Figure B . l , where H = 10.2 cm, L = 30.5 cm, W = 10.2 cm, 0 = 9 0 ° , h = 51 mm, t = 2.1mm, and <p = 2 .4° . The part was left in this position at room temperature ( 2 0 - 2 5 ° C ) for 19h50min, after which it was removed from the tool and re-measured. The part was subsequently forced back onto the tool using the described wedges and subjected to the heat treatments described in Table B.2. Spring-in measurements were taken at room temperature following each treatment. B.3 R E S U L T S A N D D I S C U S S I O N The results of this experiment are shown in Table B.2 and Figures B.2 and B.3, where the ranges shown represent 1/2 (maximum measured value - minimum measured value). From these results, the maximum change in spring-in observed for these treatments is 0 . 0 9 ° . The change in spring-in occurring during each treatment is between - 0 . 0 2 ° and 0 . 0 4 ° , which is of the same magnitude as the - 9 5 -Appendix B Investigation of Viscoelastic Effects measurement var iabi l i ty . T h e s e observations s h o w that no signif icant viscoelast ic relaxation occurs in f u l l y cured parts m a d e o f T o r a y T-800 /3900-2 prepreg for the t ime and temperature ranges exper ienced in autoclave p r o c e s s i n g . It is worth n o t i n g that the part used i n this experiment had been e x p o s e d to ambient air f o r 154 days pr ior to the " i n i t i a l " measurement g i v e n in T a b l e B . 2 . A s discussed in A p p e n d i x A , the presence of moisture increases s p r i n g - i n . T h e effect o f moisture o n these measurements was not e x a m i n e d , and m a y to a certain degree obscure the results o f this experiment . N o n e t h e l e s s , based on the observations m a d e in A p p e n d i x A f r o m an 83 day d r y i n g c y c l e , the effect o f d r y i n g o n the measured spr ing- in can be a s s u m e d n e g l i g i b l e . - 9 6 -Appendix B Investigation of Viscoelastic Effects B.4 T A B L E S Table B . l . Part description. Part description Part ID VE-1 Part shape C Flange length 57 mm Lay-up [0]8 Thickness 8 plies Cure cycle 1-hold Tooling Aluminium with rubber caul Tool surface Release agent (3) Table B.2. Heat treatment descriptions. Treatment Temperature CC) Duration T 0 (initial) 20-25 n/a T, 20-25 19h50min T 2 60-70 24h08min T 3 140-150 24h T 4 180-190 24h T 5 215-225 24hl0min T 6 215-225 171h(~lweek) Table B.3. Spring-in results. Flange A Flange B Treatment Temperature Duration Spring-in Range Drop in Spring-in Range Drop in CC) C) C) spring-in Flange B C) spring-in C) C) C) T 0 (initial) 20-25 0 1.13 0.02 n/a 1.22 0.01 n/a T, 20-25 19h50min 1.14 0.02 -0.01 1.20 0.02 0.02 T 2 60-70 24h08min 1.14 0.00 0.00 1.19 0.01 0.01 T 3 140-150 24h 1.05 0.01 0.09 1.20 0.01 -0.01 T 4 180-190 24h 1.07 0.01 -0.02 1.16 0.01 0.04 T 5 215-225 24hl0min 1.05 0.00 0.02 1.14 0.00 0.02 T 6 215-225 171h(~lweek) 1.03 0.00 0.02 1.11 0.01 0.03 Note: Each angle was measured twice. - 9 7 -Appendix B Investigation of Viscoelastic Effects 0 0 c 1.4 j 1.2 -1.0 0.8 H 0.6 0.4 0.2 0.0 22.5"C 19h50 65'C 24h08 145"C 24h00 185 C 24h00 220 C 24M0 (initial) ' 2 ' 3 Treatment T 4 T 5 Figure B.2. Flange A spring-in results. - 9 8 -Appendix B Investigation of Viscoelastic Effects 22.5 C 19160 65'C 24h08 145'C 24h00 185 C 24h00 220'C 24M0 T 0 (initial) ' 2 ' 3 Treatment T 4 Figure B.3. Flange B spring-in results. - 9 9 -

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0078710/manifest

Comment

Related Items