1、- II I! 111111 111 ! NASA Technical Paper 1785 Tests of Graphite/Polyimide Panels in Uniaxial Edgewise Charles J. Camarda DECEMBER 1980 NASA TP 1785 c. 1 - - Sandwich Compression Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-TECH LIBRARY KAFB, NM 0
2、067732 NASA Technical Paper 1785 Tests of GraphitdPolyirnide Sandwich Panels in Uniaxial Edgewise Compression Charles J. Camarda Latzgley Research Center Halnptm, Virginia NASA National Aeronautics and Space Administration Scientific and Technical Information Branch 1980 Provided by IHSNot for Resal
3、eNo reproduction or networking permitted without license from IHS-,-,-SUMMARY An experimental and analytical investigation has been made of the local and general buckling behavior of graphite/polyimide (Gr/PI) sandwich panels simply supported along all four edges and loaded in uniaxial edgewise com-
4、 pression. Material properties of sandwich panel constituents (adhesive and facings) were determined from flatwise-tension and sandwich-beam-flexure tests. Results from the flatwise-tension tests established a suitable cure cycle for FM-34l polyimide film adhesive which was the adhesive used to fabr
5、i- cate the flatwise-tension, sandwich-beam, and buckling specimens. A cell-edge bonding technique using a liquid version of FM-34 polyimide adhesive was investigated and results indicated that a considerable mass savings may be pos- sible using a cell-edge adhesive. Tensile and compressive material
6、 properties of the facings (quasi-isotropic, symmetric, laminates ( 0,+45,90,-45,) of Celion2/PMR-1 5) were determined at 11 6 K, room temperature, and 589 K (-250F, room temperature, and 600F) using the sandwich-beam-flexure test method. Buck- ling specimens were 30.5 by 33 cm (1 2 by 13 in.), had
7、quasi-isotropic, symmetric facings ( 0,+45,90,), and a glass/plyimide honeycomb core (HRH-3273-3/8-4). Core thicknesses were varied (0.635, 1.27, 1.91, and 2.54 cm (0.25, 0.50, 0.75, and 1.00 in.) and three panels of each thickness were tested at room temper- ature to investigate failure modes and c
8、orresponding buckling loads. Specimens 0.635 cm (0.25 in.) thick failed by overall buckling at loads close to the ana- lytically predicted buckling load; all other panels failed by face wrinkling. Results of wrinkling tests indicated that several buckling formulas were uncon- servative and therefore
9、 not suitable for design purposes; a recommended wrin- kling equation is presented. INTRODUCTION Preliminary structural studies of advanced space transportation systems using advanced composite structural materials of high-strength fibers and polyimide resin matrices indicate that a reduction of up
10、to 25 percent in vehi- cle structural mass is obtained by the direct replacement of aluminum panels with graphite/plyimide (Gr/PI) panels (refs. 1 and 2). Furthermore, prelimi- nary studies of the aft body flap of the Space Shuttle Orbiter (ref. 3) indicate that compression loads are the primary des
11、ign condition for this structural com- ponent and because a biaxial state of stress exists in the cover panels a sand- wich panel was chosen. The present study focuses on Gr/PI structural sandwich panels which may have application as cover skins on lightly loaded components such as the aft body flap
12、 of the Space Shuttle Orbiter. Based on the low magni- FM-34 film adhesive: manufactured by American Cyanamid Company, Bloomingdale Division. 2Celion: registered trademark of Celanese Corporation. 3HRH 327: registered trademark of Hexcel Products, Inc. Provided by IHSNot for ResaleNo reproduction or
13、 networking permitted without license from IHS-,-,-IllllIllIlllllll Ill1 I I lIlllIlllIIll I1 11111l1l1ll11l1l1l1 tude and biaxial nature of these loads, a minimum-gage, quasi-isotropic, SF- metric Gr/PI laminate ( Of +45 , 901 s) was chosen for the facings of these sand- wich panels in the present
14、study. The purposes of the present study are to analytically and experimentally investigate the local and general buckling behavior of minimum-gage Gr/PI sand- wich panels capable of use at temperatures ranging from 116 to 589 K (-250 to 600F), to verify the fabrication method used in manufacture of
15、 the panels, and to determine the material properties of the 0,+45,901s Gr/PI sandwich panel facings. Buckling specimens 30.5 by 33.0 cm (12 by 13 in.) were designed and fabri- cated with various core thicknesses to study local and general instability failure modes. The buckling specimens were teste
16、d in uniaxial edgewise compres- sion at room temperature (R.T.) and were simply supported along all four edges. Several analysis methods (refs. 4 to 8) were used to determine upper and lower bounds on critical stresses relating to intracellular buckling (dimpling), wrin- kling, shear crimping, and g
17、eneral panel instability and are evaluated in this study for their capability in predicting buckling loads and modes of Gr/PI sandwich panels. The panels were fabricated using a commercially available high-temperature film adhesive, FM-34, to bond the core to the facings. Flatwise-tensile tests were
18、 performed using the sandwich panel facing laminate orientation, core, and adhesive to determine a suitable fabrication cure cycle and the tensile adhesive bond strength in a core-to-facing bond situation. In addition, flatwise-tensile tests were used to evaluate BR-34, a liquid version of the FM-34
19、 film adhesive, as a cell-edge adhesive. Sandwich-beam-flexure tests were performed to determine modulus, strength, and Poissons ratio of the facings. The flatwise-tensile tests and sandwich- beam-flexure tests were conducted at temperatures of 116 K, R.T., and 589 K (-250F, R.T., and 600OF). Qualit
20、y control standards for fabrication of all specimens were high to minimize scatter in the data. Results of the tests are presented in tabular and graphical form. Results of the beam tests were ana- lyzed statistically and a best-fit third-order polynomial relating stress and strain was fit through t
21、he data. Certain commercial materials are identified in this paper in order to spec- ify adequately which materials were investigated in the research effort. In no case does such identification imply recommendation or endorsement of the product by NASA, nor does it imply that the materials are neces
22、sarily the only ones or the best ones available for the purpose. In many cases equivalent materials are available and would probably produce equivalent results. SYMBOLS Values are given in both SI and U.S. Customary Units. The measurements and calculations were made in U.S. Customary Units. 2 Provid
23、ed by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-DQx I DQy FC Fcu G Nx Ny stiffness matrices of sandwich panel width of plate coefficients of polynomials used in regression analysis flexural stiffness of composite facings transverse shear stiffness of sandw
24、ich plate in x- and y-directions, respectively flexural stiffness of orthotropic sandwich plate in x- and y-directions, respectively twisting stiffness of orthotropic sandwich plate elastic modulus modulus of core in z-direction facing modulus facing modulus in x- and y-directions, respectively tang
25、ent modulus average elastic moduli of laminate in x- and y-directions, respectively lower of flatwise core compressive or tensile strengths, or core-to-facing bond strength compressive ultimate strength shear modulus core shear modulus in xz-plane core shear modulus in yz-plane facing shear modulus
26、in xy-plane total number of points in regression analysis length of plate number of half sine waves in x- and y-directions, respectively resultant normal forces in x- and y-directions, respectively 3 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-P
27、SCJ/E S Tg t tC t f tfl tf2 th Vf VV XIYIZ 6 6 1-1 1-1xr1-1y Dxy pyx “ P 0 ocr im *d im *wr load standard error of estimate honeycomb cell size glass transition temperature thickness core thickness average facing thickness thickness of facing 1 thickness of facing 2 total sandwich panel thickness fi
28、ber volume fraction void volume fraction rectangular coordinates initial panel waviness strain Poissons ratio Poissons ratio of orthotropic plate associated with bending of plate in x- and y-directions, respectively average Poissons ratios of orthotropic plate associated with extension of plate in x
29、- and y-directions, respectively density stress critical stress associated with shear crimping critical stress associated with dimpling critical stress associated with wrinkling Subscripts : av average 4 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-
30、,-cr critical i index of summation max maximum ult ultimate XIY coordinate directions 1,2 directions parallel and perpendicular to fiber direction, respectively TEST SPECIMENS I APPARATUS, AND PROCEDURE Graphite/Polyimide Materials This program was conducted as part of the NASA program, Composites f
31、or Advanced Space Transportation (CASTS) (ref. 1). The CASTS effort focused on graphite/polyimide, and a significant part of the program included evaluating and characterizing various fiber and resin materials. As a result of these evaluations, the materials used in different phases of the present s
32、tudy varied as improved systems were identified. Specifically, for flatwise-tensile tests, the laminates were HTS4-l/PMR-15; for sandwich-beam-flexure tests, the laminates were Celion 6000/PMR-15; for buckling tests, the laminates were Celion 3000/PMR-15. The primary purpose of the flatwise-tensile
33、tests was to evaluate adhesive tensile strengths and therefore the difference in facing materials was not critical. The thinnest gage prepreg, Celion 3000/PMR-15, was chosen over the Celion 6000/PMR-15 to minimize mass of the sandwich panels, and the Celion fiber was chosen over the HTS fiber becaus
34、e Celion exhibits less material prop- erty degradation than HTS at elevated temperatures. Flatwise-Tensile Specimens Forty-six 7.62 by 7.62-an (3 by 3-in.) specimens were fabricated using pre- cured 0,+45,901s laminates of HTS-l/PMR-15 Gr/PI facings, glass/polyimide honeycomb core (HRH-327-3/16-6 or
35、 8) and the desired adhesive. A schematic dia- gram of a typical specimen is shown in figure l. Details of fabrication proce- dures and cure cycles are given in reference 9 and table I, respectively. Steel load blocks were bonded to the facings of the specimens and each block had a tapped hole for a
36、ttaching a loading rod. Universal joints were attached between the testing machine and the loading rods to assure proper alignment of the fixture in the loading machine. The specimens were tested in a universal testing machine operating in a displacement control mode at a constant rate of 0.13 cm/mi
37、n (0.05 in/min). Test temperatures other than room temperature were obtained using an environmental chamber positioned within the crossheads and posts of the testing machine. Specimens were held at desired test temperatures for 15 minutes prior to testing to insure thermal equilibrium. Maximum load
38、was recorded for each test and converted to a normal tensile stress. 4HTS graphite fiber: product of Hercules Incorporated. 5 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-Sandwich-Beam-Flexure Tests Specimens.- Sandwich-beam-flexure specimens cons
39、isted of Gr/PI facings and glass/polyimide honeycomb core as shown in figure 2. The honeycomb core was HRH 327-3/16-8 glass/polyimide and was cut into strips 2.54 cm (1.00 in.) wide by 55.88 cm (22.0 in.) long by 3.175 cm (1.25 in.) high. The test facing of the beam was a O,+45,9O,-45Is laminate of
40、Celion 6000/PMR-15 and the oppo- site facing had a laminate orientation of 02,+45,90,-451,. The additional Oo layers of the non-test facing insured failure of the test facing. These lami- nate orientations were chosen to avoid microcracks in the laminate which are believed to occur when adjacent lay
41、ers are stacked at an angle greater than or equal to 90 with respect to one another. The honeycomb core was filled with BR-34 liquid adhesive and glass beads throughout the length of the beams, except for the 7.62-cm (3.00-in.) test section in the center of the beams, to prevent premature adhesive f
42、ailure. Details of the fabrication of the sand- wich beam specimens are presented in reference 9. Apparatus and instrumentation.- Each specimen was instrumented at the mid- span of the beam with a high-temperature strain rosette (WK-03-060-WR-350) oriented at Oo, 45O, and 90 with the load axis and b
43、onded to the test facing, and a single strain gage (WK-03-125-AD-350) oriented at Oo with the load axis and bonded to the non-test facing. These gages were manufactured by Micro- Measurements Division of Vishay Intertechnology, Inc. The strain gages were bonded to the outer surfaces of the beam usin
44、g a polyimide adhesive (either “Bond 610 or PLD-700 available from Micro-Measurements and BLH Electronics, respectively) . The sandwich beams were placed in a four-point bending test apparatus (fig. 3) which supported the beam on rollers 48.26 cm (19.00 in.) apart with flat sections 2.54 cm (1.00 in
45、.) wide machined in them. Load was applied by a 222-kN (50-kip) capacity hydraulic testing machine which acted at two points on the top flange of the beam spaced 10.16 cm (4.00 in.) apart and symmetric about the beams center. For testing at temperatures other than room temper- ature the specimen was
46、 instrumented with a thermocouple attached to the test facing and the test fixture and specimen were completely enclosed in an envi- ronmental chamber and either heated or cooled to the desired test temperature. Specimens were allowed to soak at the test temperature for 20 minutes to insure thermal
47、equilibrium. A data handling system consisting of 40-channel scanner, digital voltmeter, plotter, printer, clock, and calculator was used to record and reduce data. Procedure.- The load signals from the load cell on the hydraulic testing machine were input to one channel of the scanner. Strain signa
48、ls were input to selected scanner channels and initally set to zero using Wheatstone bridge bal- ance (for non-room-temperature tests, initial strain signals were set to zero after thermal equilibrium). Strains were corrected for transverse sensitivity of the gages and nonlinearity of the bridge cir
49、cuit. Thermocouples were con- nected to the scanner through a 273 K (32OF) cold-junction reference. Beams were tested to failure during the test, load was applied at a rate of 80 N/sec (20 lbf/sec), data were recorded every 3 seconds, and a stress-strain 6 Provided by IHSNot for ResaleNo reproduction or networking permitted without license from IHS-,
