Numerical and Experimental Dynamic Characteristics of Thin-Film Membranes

Leyland Young, Perngjin Pai
2004 45th AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics & Materials Conference   unpublished
Presented is a total-Lagrangian displacement-based non-linear finite-element model of thin-film membranes for static and dynamic large-displacement analyses. The membrane theory fully accounts for geometric non-linearities. Fully non-linear static analysis followed by linear modal analysis is performed for an inflated circular cylindrical Kapton membrane tube under different pressures, and for a rectangular membrane under different tension loads at four comers. Finiteelement results show that
more » ... results show that shell modes dominate the dynamics of the inflated tube when the inflation pressure is low, and that vibration modes localized along four edges dominate the dynamics of the rectan-gular membrane. Numerical dynamic characteristics of the two membrane structures were experimentally verified using a Polytec PI PSV-200 scanning laser vibrometer and an EAGLE500 8camera motion analysis system. L G. Young el aL I Internationul Journai of Sol& und Structures xxx (2004) xxx-xxx 28 Antenna Experiment (ME) in 1996 (Dornhiem and Anselmo, 1996) . The membrane antenna having an in-29 flated diameter of 5 O H w k h three 93fl long struts was transported by the space shuttle Endeavour in a 30 7@ x 3 H x lSf$ container. Also infiatable membrane structures have been used in parabolic antennas, 3 1 radiators, solar concentrators, sun shields, habitats, radio-frequency structures, optical communication sys-32 tems, radars, lightweight radio-meters, telescopes, etc. Moreover, large balloons are also 33 tures that have been used for many scientific missions. Advantages of membrane structu 34 stowed volume, lightweight, low cost, and good thermal and damping properties (Palisoc 35 there are difficulties in the design of large scientific membranes (Damle et al., 1997) . 36 Over the last few decades, studying the dynamic behaviors of inflatable membrane structures has proven 37 to be a challenging job. Many researchers have studied the dynamic characterization of membranes using 38 numerical methods and, when possible, experimental approaches. Numerical methods such as finite differ-39 ence and boundary elements were used by some researchers to compute vibration modes and frequencies of 40 inflatable dams (Hsieh and Plaut, 1990) . The membrane material used in the nwnerical analysis was as-41 sumed inextensible and its weight was neglected in the determination of the equilibrium shape. They found 42 that the membrane's mass density is of little influence on the computed natural frequencies. Other research-43 ers used finite elements and boundary elements to model and compute natural frequencies and mode shapes 44 of a single-anchor inflatable dam (Mysore and Liapis, 1998). They found that the rigid foundation that an-45 chors the dam increases the frequencies whereas the presence unded water tends to reduce the fre-46 quencies. They noted that the natural frequencies are depen the internal pressure as well as the 47 hydrodynamic pressure of the impounding water. The pressure in an inflatable structure can also play a 48 critical role in the suppression of vibration (Choura, 1997) . This study found that the vibration suppression 49 of inflatable structures can be accomplished by varying the internal pressure and thus there is no need of 50 other external actuators for vibration suppression. 51 Some researchers tested extremely lightweight inflatable structures in a vacuum chamber and in the 52 ambient atmospheric condition (Slade et al., 2001). They found a lack of correlation between the two cases, 53 and they explained it to be caused by air damping. Because the coupling of a lightweight membrane and air 54 is a highly non-linear and localized fluid-structure interaction prcblem, it is difficult to perform accurate 55 numerical modeling and simulation of such problcms. Hence, testing inflatable stntctuies in vacuum con-56 ditions becomes necessary in order to verify numerical predictions. Moreover, because of the size limitation 57 on actual vacuum chambers, tests in vacuum conditions for large membrane structures are only possible by 58 using scaled models (Pappa et al., 20011. Johnson and Lienard (2001) obtained the natural frequencies and 59 mode shapes of a one-tenth d e Next Generation Space Telescope (NGST) using a finite-element model 60 developed using the cable n od. The difference between predicted and measured natural fre-61 quencies ranges from 2% to was noted that predicted mode shapes correlated well for strut-62 dominated modes, while membranedominated modes showed less correlation. The study of pre-stressed 63 membranes by Hall et al. (2002) showed that the natural frequencies in air are lower than the ones in vac-64 uum because air acts as a non-structural mass. But, the numerical natural frequencies obtained by Kuka-65 thasan and Pellegrino (2002) were lower than experimental vacuum ones and the error was attributed to an 66 inaccurate tension force or Young's modulus. However, they stated that the error reduced as the tension 67 force was increased. Experiments also showed that it is difficult to excite global vibration modes of a mem-68 brane applying excitations at inflatable components because these components have high local 69 flexibi nant frequencies may vary w i t h the excitation location (Pappa et al., 2001; Gaspar et al., 70 2002). Moreover, because the light weight of membranes, contact sensors cannot be used in testing and 71 non-contact sensors (e.g., scanning laser vibrometers) need to be used (Gaspar et al., 2002) . 72 In recent years many researchers used commercial finite-element packages to model and analyze non-lin-73 ear elastic problems of thin-thickness membrane structures (Wong and Pellegrino,
doi:10.2514/6.2004-1618 fatcat:qn3ciuwgurclzixttns3yvajeq