18 TH INTERNATIONAL CONFERENCE ON COMPOSITE MATERIALS CONCEPTION OF CRASH TESTS FOR COMPOSITE TUBULAR STRUCTURES H. Zabala 1 *, J. Aurrekoetxea 1 , M. Mateos 1 , G. Castillo 2 , L. Aretxabaleta 1 1 Mechanical Engineering and Industrial Manufacturing Department, Mondragon Unibertsitatea, Mondragon, Spain 2 Civil, Materials and Manufacturing Engineering, University of Malaga, Malaga, Spain * Corresponding author (hzabala@eps.mondragon.edu) Keywords : Crash, composites, crashworthiness Abstract to get the energy dissipation in a more controlled way. One of these test configurations consists in the use of different shape plug initiators like radius or cone shaped A lot of research work on geometrical, material and plugs. loading factors’ influence on composite tubular structures Most of these experimental works are carried out in quasi- behaviour under crash conditions has been undertaken. static conditions, but most of the real life applications of Most of these works have dealt with tubular structures energy dissipation structures occur at impact velocities subjected to compressive loads, which can induce local and/or energies. The dependence of energy dissipation buckling phenomena. Under these conditions only a small capabilities on strain rate has been reported in a few amount of material is degraded resulting in a small energy works, as for example in [16], where rectangular section dissipation level in a non-controlled way. The tests 3D braided E-glass/epoxy composites have been crushed proposed in this communication allow measuring the at different strain rates, showing a clear higher energy crashworthiness of tubular structures for the automotive absorption capability as strain rate increases. industry subjected to different contour conditions. Three The aim of this work is to develop an experimental test different initiator plugs have been proposed to carry out methodology that will allow testing tubular composite the compression tests: flat, conical and radial ones and structures at different strain rates (from quasi-static to low tests at different strain rates have been performed. For the velocity impact) and different boundary conditions (with selected initiators dimensions, no influence of the strain different shape plug initiators). The objective is to be able rate on the dissipative behaviour of compressed carbon- to get information about the material damage behaviour in epoxy 0/90 tubes has been observed. different conditions that will allow validating existing damage initiation and degradation models. In this 1. Introduction communication only the preliminary results of this work are shown. Composite structures have been widely used to produce energy dissipation during a crash [1-3]. Energy is 2. Material and experimental test dissipated through damage mechanisms that include intra- laminar failure (fibre and matrix rupture) and inter- 2.1. Material laminar failure (delamination). There are several numerical approaches to predict the complex damage behaviour of composites, like Hashin’s damage initiation Tubular structures of two different materials have been and degradation model [4]; nevertheless, the information used in this study. In the first step, with the aim of needed to feed these models in a reliable way is often optimizing the testing equipment and plug initiators, a difficult to get through non standard tests carried out at commercial non reinforced PVC straight pipe with an conditions different from the ones that happen in real life outer diameter of 40 mm and 3.2 mm thickness has been (i.e. quasi static tests for feeding crash models). used. Specimens were cut at lengths of 40, 60 and 80 mm. The most common way of dissipating high amounts of Testing tools with different geometries have been built for energy is to subject the structure to a compressive load analyzing their influence on the crushing behaviour of the high enough to produce the crushing of the structure [5- PVC pipes. 8]. Different configurations of tubular structures [9-13] or tests [14-15] have also been studied in order to obtain The second structure has been a commercial pultruded higher energy dissipation. In some cases, the goal can be unidirectional carbon fibre composite pipe, with an outer
0 0/90 fabric lay yer. The dime ensions of the tubes have b been Ne evertheless, th hese buckling failure modes s were achieve ed a an outer diame eter of 40mm m and a thickn ness of 2 mm m. In at different stag ges of strain, as shown in n Fig. 3: in th he th his case, all t the specimens s have the sa ame length of f 50 thr ree cases the maximum for rce value corr responds to th he m mm. ini tiation of the e buckling fai ilure, but for the flat plat es con nfiguration oc ccurs for a lo ow strain leve el compared to F Finally tubes of 0/90º wov ven carbon f fibre-epoxy w were bot th i) the flat p plate and coni ical plug initia ator, and ii) th he m made by hand lamination. A A tubular stru ucture of 8 lay yers fla t plate and rad dial plug initia ator. o of a 0/90 fabri ic layup was b built with an i inner diamete r of 4 40 mm and 2. .2 mm thickn ess. In this ca ase the length h of th he samples ha as also been 50 0 mm. 2 2.2. Quasi-st tatic compre ession tests T The tests met thods consist in compress sing the samp ples su upported by d different initia ator plugs in th heir bottom (F Fig. 1 ). The tests are carried o out in a univ ersal tensile test machine config m gured in comp pression mode e. Three differ rent (a) (b b) (c) te est configurat tions have bee en performed: : a) Compress sion Fi gure 2: Tested d samples with h local buckli ing failures: (a a) between flat p b plates, b) com mpression betw ween a flat p plate Fla at plates, (b) f flat plate and c conical tool an nd (c) flat plat te tool ( α =55º) a a and a conical t and c) compre ession betwee en a and radi ial tool. fl flat plate and a a radial tool ( R R =22 mm). (a) (b) ( (c) Figure 1: Rep presentation o f the develope ed test method ds: (a) Flat plates ( , (b) flat plate and conical p plug initiator a and Fig gure 3: Force- -displacement t curves for di fferent initiato or (c) f flat plate and r radial plug ini itiator. plugs on PVC p specimens: (P P) Flat plates, (C) flat plate and conical in nitiator plug an nd (R) flat pla ate and radial T Tests have be een performed d at a consta ant velocity o of 1 initiato or tool. mm/min for P m PVC specimen ns and at test t velocities o f 1, 1 20 and 480 m mm/min for the e carbon fibre e-epoxy ones. Fro om Fig. 3 it ca an be conclud ded that even i if the maximu um for rce achieved i s the same, in n the tests for the conical an nd rad dial initiator p plug the energy y absorption f for PVC occu urs 3. Results an 3 nd discussion n in a more contro olled way. It c can be seen th hat this value is hig gher for larg ge displaceme ents in coni cal and radi ial 3 3.1. Tests on PVC tubes ini tiator plugs due to the f fact that buck kling thresho old occ curs for a high her deformatio on of the struc cture. This wa ay The tests on PV T VC samples h have demonstr rated that tubu ular it i is possible to avoid the un npredictable be ehaviour of th he st tructures are subjected to a very diffe erent stress s state stru ucture once th he first local c crash is given. depending on d the initiator p plugs used. I t was found t that th here was no influence of f the sample length on f final Th hese results co nfirm that dif fferent initiato r plugs are ab ble fa failure behavio our, since it w was a conseq quence of a lo ocal to improve the c crashworthine ss of tubular P PVC structure es. b buckling phen nomenon in the three cases (Fig. 2). Th he energy d dissipation be ehaviour of carbon-epox xy
Recommend
More recommend
