THERMAL DEFORMATIONS OF CARBON/EPOXY LAMINATES FOR TEMPERATURE VARIATION

Transcription

1 HERMAL DEFORMAIONS OF CARBON/EPOXY LAMINAES FOR EMPERAURE VARIAION K. Joon Yoon and Ju-Sik Kim Department of Aerospace Engineering, Konkuk Universit Kwangjin-gu, Seoul 43-7, Korea SUMMARY: Effects of anisotrop of thermal epansion properties on the thermal distortion of Carbon/Epo composite stiffner structures were investigated and analzed. Coefficients of thermal epansion and elastic properties in the material principal directions were measured and characterized for temperature variation. B appling the characterized properties to the classical lamination theor, the spring-forward distortion of a L-section laminate was predicted. he spring-forward angles of L-section laminates of various angle plies were measured and compared with those predicted. Eperimental results show that the process induced spring-forward of L-section laminate can be predicted quite well b using the computational method. KEYWORDS: hermal Deformation, Coefficient of hermal Epansion, Carbon/Epo, Formulation of CEs for emperature Variation INRODUCION Carbon Fiber reinforced composites are widel developed and used as structural materials of the modern aircraft and spacecraft because of their high specific strengths, moduli, and design fleibilities. Furthermore, the low coefficient of thermal epansion(ce) of carbon fiber composites can afford precision alignment and dimensional stabilit to aerospace structures. However, due to the anisotrop of materials, fiber reinforced composites have several problems during the manufacturing process. he process induced distortion brings about assembling problems of composite structures that need a precise dimension shape. his problem should be corrected to increase the performances of composite parts and to save cost and time consuming. o solve this problem, tool dimensions must be sized to correct the dimensional distortions occurring during the manufacturing process. Anisotrop of thermal epansion properties, residual stresses, chemical shrinkage and thickness change of plies affect the final shape of fiber reinforced laminate structures cured at elevated temperature [,,3,4,5]. Barnes et al.[] showed that the difference of coefficients of thermal epansion and residual stress were the main sources of process induced distortion of curved laminates. Nelson[3] proposed a relativel simple set of analtical equations for the prediction of spring back in thermoset composite laminates, which included the effects of thermal dilatation strain, chemical shrinkage of the matri, and residual stresses. Yoon et al.[5] proposed computational model to predict a spring-forward angle of L-section AS4/PEEK thermoplastic composite laminates b incorporating the difference of thermal

2 epansion coefficients and frictional residual stresses effects. However, it has not been reported et that the thermal distortion of thermoset composite structures were predicted b considering the changes of basic CEs and moduli for temperature variation. In this paper, the effects of anisotrop of thermal epansion properties on the process induced distortion of L-section Carbon/Epo thermoset composite structure were investigated. A computational method to predict the distortion of curved laminate due to the difference of CEs was proposed. o verif the proposed computational method, the spring-forward angle changes of Carbon/Epo L-section were measured and compared with those predicted. ANALYSIS FOR PROCESS INDUCED DISORION Anisotrop of thermal epansion properties of a material induces thermal distortion when the temperature of the structure changes. Antismmetric cross-pl laminates distort in clindrical or saddle shape and laminates of unbalanced stacking sequence twist after cured in a flat mold at an elevated temperature. hese distortions are shaped b the residual stresses occurred from the mismatch of CE between plies. Even in the case of a smmetric stacking sequence, if a laminate is cured in a curved mold, the curvature of a laminate differs from the curvature of a mold after cure. Fig. : hermal Distortion of Curved Section of Anisotropic Material Structure

3 o describe the curvature change of curved section of an anisotropic material structure, O'Neill [] showed that the arc angle change of a curved section can be epressed as Eqn from a geometric analsis of a curved geometr as shown in Fig., { α () α () } d Δθthermal θ z () f where θ is the angle of the arc sector of the curved laminate, α z and α are the CEs in the through thickness direction and in the geometrical principal direction of the laminate plane respectivel. o predict the distortion due to the difference of CEs precisel, it is important to characterize the change of the CE for temperature variation accuratel. It was reported that the CEs of a polmer composite is dependent on temperature[5,6]. Barnes et al.[] measured the CEs changes of the [±45//9] S AS4/PEEK laminates for temperature variation and predicted the induced distortion angle of a L-section laminate b using the measured CEs. In their method, CEs should be measured eperimentall for ever laminate of different stacking sequence to predict its distortion angle of a curved section. Yoon et al.[5] proposed analtical formulations to calculate the change of CEs of various angle-pl laminates and used the calculated CEs to predict the spring-forward angle of AS4/PEEK curved laminate structures. In this stud, an analtical procedure to calculate the CE of a general stacking sequence laminate proposed b Yoon et al.[5] was used. As the temperature of composite materials changes, the thermal epansion strains of unconstrained laminate can be epressed as Eqn using the classical lamination theor[7]. ε ε γ N [A' ] N N + M [B' ] M M () where [A'] and [B'] are matrices that are the parts of inverse of combined matri of etension, coupling and bending stiffness, {N } and {M } are the thermal resultant forces and moments respectivel. For temperature variation of Ä the thermal strains of a laminate can be epressed as ε ε γ α α (3) where [α ] are the coefficients of thermal epansion of a laminate. From Eqn and 3, the coefficients of thermal epansion of a laminate can be obtained as ( ) α ( ) α ( ) ( ) ( ) [ A' ( )] [ Q( )] k α ( ) dz + [ B' ( )] [ Q( )] k α ( ) z dz α ( ) k α ( ) k (4)

4 where α α k α [ ] k ε ( ( ) ) (5) α and α are the CEs in the fiber direction and fiber transverse direction respectivel. It should be noted that the values of [A'], [B'], [Q ] and {α} are dependent on temperature. o obtain the CEs of a laminate for temperature variation in Eqn 4, the basic mechanical and thermal epansion properties in the material principal directions need to be characterized for temperature variation. CHARACERIZAION OF MAERIAL PROPERIES Elastic Moduli and Poisson's Ratio for emperature Variation Elastic moduli and Poisson's ratios obtained from the tensile coupon tests at four different temperatures were fitted as linear functions of temperature for the convenience of simple formulations, which are shown in Fig. and 3. E E G υ (GPa) (GPa) (GPa) ( 5 C 4 C) ( 5 C 4 C) ( 5 C 4 ( 5 C 4 C) where E, E, G and υ are the longitudinal, the transverse, the shear moduli, and Poisson's ratio, respectivel. (6) 6 4 Logitudinal Modulus(E ).8 E (GPa) Poisson's Ratio 4 Poisson's Ratio. 5 5 emperature( o C) Fig. : Logitudinal Modulus and Poisson s Ratio for emperature Variation

5 4 Elastic Modulus of Fiber ransverse Direction(E ) Shear Modulus(G ) GPa emperature( o C) Fig. 3 : ransverse Modulus and Shear Modulus for emperature Variation Coefficients of hermal Epansion for emperature Variation he composites reinforced with carbon fibers have a significant difference in CEs between the fiber direction and the fiber transverse direction. his characteristic has been shown well b Barnes et al.[3] who measured the coefficients of thermal epansion of several PEEK composites and characterized the changes of coefficients of thermal epansion as functions of temperature in the range of 83 o K to 393 o K. In this stud, to measure the change of CEs of a laminate in the in-plane material principal directions for temperature variation, specimens were cut b 5mm from [] 8 laminate. wo strain gages(wk-6-6ap-35, Micro-Measurement) for high temperature testing were bonded in the longitudinal and transverse directions along the centerline of a specimen. o measure the CE in the out-of-plane direction Specimens cut b 5mm X mm from [] 64 laminate (thickness : 4.5mm) were used and strain gages were bonded in the through thickness direction. hermal strains induced b temperature change were measured using strain gage measurement techniques used b Kim and Crasto[8]. he measured data in Fig. 4 show that CE in the transverse direction doesn't change much below 8 C, but changes significantl after 8 C. In this stud the measured CEs in the longitudinal and fiber transverse directions were characterized as quadratic functions of temperature as shown in Fig. 4 for the convenience of formulation. α (.3 α (.4 α 3 ( ) ) ) ( m / m / C) (3 C 3 C) ( m / m / C) (3 C 3 C) ( m / m / C) (3 C 3 C ) (7)

6 8 7 6 CE of fiber direction CE of fiber transverse direction CE of through thickness CE( -6 / o C) emperature( o C) Fig. 4 : CEs in the Material Principal Directions VERIFICAION OF CEs OF ANGLE-PLY LAMINAES o verif the analtical formulations for the in-plane CEs of general laminates, specimens, mm 5mm in size, with [±5] S, [±3] S, [±45] S, [±6] S, and [±75] S stacking sequences were prepared and tested for the measurement of in-plane CEs. In Fig. 5 the measured CEs of each [ϕ ] specimen are compared with those predicted b using the S analtical formulations. We can see those agreements between eperimental data and predictions are quite good. If CEs obtained from the proposed formulations are used for an analsis of thermal deformation of a carbon fiber reinforced structure, it is epected that dimensional changes due to thermal deformations can be predicted more accuratel than the cases of conventional linear analsis, which do not consider the changes of CEs of laminates. EXPERIMENAL RESULS AND DISCUSSION L-section specimens (Fig. 6-a)of [] 8, [±3] S, [±45] S, [±6] S laminates were selected to investigate the spring-forward behavior of Carbon/Epo composite laminate structures. Carbon/Epo laminates of L-section were manufactured b using an autoclave manufacturing process(fig. 6-b). he angle changes of the manufactured specimens were measured at room temperature (5 o C) b stamping the L-section on a reference paper of right angle mesh.

7 8 6 Measured (35 o C) Measured (85 o C) Measured (5 o C) Predicted 5 o C CE( -6 / o C) 4 85 o C 35 o C Pl angle (ϕ) Fig. 5 : Measured and Predicted CEs of [±ϕ] S Laminate he Spring-forward angles of the various angle plies laminates were predicted b using the Eqn to see the effects of thermal deformations. he measured data are compared with those predicted b the computational method in Fig. 7. he predicted values from the thermal analsis show that the the difference CEs is the major sources of the process induced deformation of curved section laminates, even though the predicted values are still lower than the values from eperiments. o predict the thermal distortion angle more precisel, other effects such as chemical shrinkage, residual stresses, thickness change of plies, visco-elastic and plastic deformation effect etc. must be included, as mentioned b Nelson and Cairns[3]. CONCLUSIONS he effects of thermal deformation were investigated for L-section Carbon/Epo laminate structures. o see the effects of the difference of CEs, the basic mechanical properties and thermal epansion properties of Carbon/Epo composite were characterized for temperature variation. B incorporating the characterized mechanical and thermal properties into the classical lamination theor, it was possible to obtain the change of the inplane coefficients of thermal epansion of a general stacking sequence laminate for temperature variation. he trend of spring forward angle of L-section laminate structures of Carbon/Epo was predictable b appling the calculated CEs and chemical strains to the distortion models.

8 ϕ 46mm R 3 mm. mm 3 mm (a) L-section Specimen Shape (b) General La-Up Structure Fig. 6 : Specimen Dimensions and Autoclave La-Up Structure

9 5 Predicted Data with onl Difference of CE 4 Eperimental Data Angle change(degree) Pl angle [+ϕ/ ϕ] S Fig. 7 : Comparison of Predicted Distortion Angles with Eperimental Data REFERENCES. Barnes, J. A., J. A. Berl, M. C. LeBouton and N. Zahlan. 99. "Dimensional Stabilit Effects in hermoplastic Composites-owards a Predictive Capabilit," Composite Manufacturing, : O'Neill, J. M.,. G. Rogers and A.J.M. Spencer. 988 "hermall Induced Distortions in the Moulding Laminated Channel Section," Mathematical Engineering in Industr, : Nelson, R.H "Prediction of Dimensional Changes in Composite Laminates During Cure," Proceedings of 34th International SAMPE Smposium, 34 : Manson, J. A. E. and J. C. Seferis 99 "Process Simulated Laminate(PSL) : A Methodolog to Internal Stress Characterization in Advanced Composite Materials," J. of Composite Materials, 6 : Yoon, K.J and P.J.Kim, 997 "Process Induced Distortion of L-section thermoplastic Laminate Structure, " J. of Materials Processing 4 manufacturing Science, 5 : Barnes, J. A., I. J. Simms, G. J. Farrow, D. Jackson, G. Wostenholm and B.Yates. 99. "hermal Epansion Characteristics of PEEK Composites," J. of Material Science, 6 : Jones, R. M. 975, Mechanics of Composite Materials, McGraw Hill Co., New York, NY 8. Kim Ran Y. and Allan S. Crasto, 995. "Dimensional Stabilit of Composites in Space: CE Variation and heir Prediction", Proceedings of ICCM-, Whistler, B.C.