DEPENDENCE OF GEOMETRICAL CONFIGURATION ON MECHANICAL BEHAVIOUR OF GFRP COMPOSITE SANDWICHES

Transcription

1 DEPENDENCE OF GEOMETRICAL CONFIGURATION ON MECHANICAL BEHAVIOUR OF GFRP COMPOSITE SANDWICHES Antonino Valenza 1, Vincenzo Fiore 1 and Luigi Calabrese 2 1 Dept. o Ingegneria Chimica dei Processi e dei Materiali, University o Palermo, 9128 Palermo, Italy 2 Dept. o Chimica Industriale e Ingegneria dei Materiali, University o Messina, Messina, Italy ABSTRACT Aim o this research work is to ind the dependence o geometrical coniguration on ailure mechanism o composite sandwiches subjected to three point bending test. All tested structures have been realized at room temperature, with a single lamination, by vacuum bagging technique. These sandwiches are made up o two skin o E-glass/epoxy composite and closedcell PVC oam cores with a density o 6 Kg/m 3. Experimental tests have been carried out at varying skin thickness (t) and span length between support (l) with the purpose to show the dependence o mechanical behaviour on these geometric parameters. From experimental results a ailure map in the plane t - l has been created identiying three regions o typical ailure modes o the sandwiches. 1. INTRODUCTION When sandwich composites are subjected to three point bending, several ailure mechanisms may occur: a ailure mechanism will be shown when its breaking load results smaller in comparison o those ones o competitive mechanisms. In literature ive main ailure mechanisms have been identiied: skin compressive/tensile ailure, core shear ailure, skin wrinkling, indentation and delamination, i.e skin-core debonding [1]. Consider a sandwich beam (Figure 1) with skins o equal thickness subjected to a bending moment M and shear orce V. Figure 1: Experimental Schematic conigurations o the oam core composite sandwich beam: (a) loading coniguration and (b) cross section For relatively thin skins and relatively low core stiness, the bending moment is mainly taken up by the acings, while the shear stress is mainly taken up by the core. The compressive or tensile skin stress σ, is [2]: M M σ = b t ( t + c) b t c Since maximum bending moment is given by: P l M = (2) 4 the compressive or tensile skin breaking load is: (1)

2 ( ) 4 b t t + c σ 4 b t c σ P = (3) l l The shear stress τ c in the core is: V V τ c = b t + c b c ( ) Since shear is given by: P V = (5) 2 the core shear skin breaking load is [3,4]: ( ) (4) P = 2 b t + c τ 2 b c τ (6) CS C C Skin wrinkling is deined as a localized short-wave length buckling o compression skin. It must be viewed as buckling o the compression acing supported on one side by the core. The value o the racture load is given rom the ollowing expression [5]: 2 4 b t c E E C PW = (7) l 12 ( 3 ν C ) ( 1 + ν C ) Indentation o the core occurs at concentrated load, such as ittings, corners, or joints. Practically they can be avoided by applying the load over a suiciently large area [3]. This area can be estimated rom [6]: P A = (8) σ C Where σc is the compressive strength o the core material. Triantaillou and Gibson [6] assumed that the indentation load is dictated by plastic yield o the core but the contribution rom the skins is ignored. This approximation is valid or the case o thin skins but ails or practical sandwich beams. Ashby et al. [7] have considered the indentation strength due to the combined plastic collapse o the ace sheets and core. They treat both the skins and the core as rigid, ideally plastic solids, with the core undergoing compressive yield and the skins orming plastic hinges. This model was conirmed by a series o experiments and detailed numerical simulations on metallic oam cores and aluminium alloy skins [8,9]. In another work [1] it is shown experimentally and by inite element analysis that the indentation load is set by elastic deormation o the aces and compressive yield o the core. These ailure mechanisms are called ductile indentation and elastic indentation respectively [11]. In the case o ailures like delamination and skin-core debonding it is very diicult to perorm a theoretical strength prediction analysis rom the knowledge o the mechanical characteristics o the skin and the core. 2. EXPERIMENTS Four type o symmetrical sandwiches at room temperature by vacuum bagging technique are realised. Structures prepared consist o a common PVC oam core and GFRP skins with one (named M1), two (M2), three (M3) and our layer (M4) respectively. The core is constituted by a closed cell PVC oam (Herex C 7.55 rom Alcan composites) having density o 6 Kg/m3 and thickness c equal to 5 mm.

3 The sandwich beam skins were realized with E-glass mat (randomly oriented ibres, with density o 3 g/m 2 ) and a SR 81 epoxy resin combined with the hardener SR 8822, both supplied by Sicomin, used as matrix. Experimental tests have been carried out by using a Universal Testing Machine (UTM) mod by Instron, equipped with a load cell o 5 KN, with a speed o 1 mm/min. For every sandwich, three point lexural tests were carried out on rectangular samples with width b = 15 mm placing span distance l equal to 4, 5, 6, 8, 1, 12, 14, 16, 18 and 2 mm. For each condition, ive samples have been tested. Average values o skin thickness t are reported in Table 1. M1 M2 M3 M4 3,326 ±,23 1,128 ±,33 1,655 ±,64 2,517 ±,76 4,365 ±,11 1,14 ±,78 1,82 ±,3 2,284 ±,126 5,362 ±,18,95 ±,35 1,823 ±,64 2,419 ±,88 6,358 ±,41,93 ±,3 1,739 ±,24 2,316 ±,39 8,341 ±,24 1,17 ±,58 1,68 ±,14 2,28 ±,65 1,376 ±,26,975 ±,31 1,776 ±,46 2,399 ±,49 12,367 ±,33 1,17 ±,44 1,471 ±,59 2,472 ±,12 14,324 ±,41 1,37 ±,32 1,547 ±,13 2,444 ±,78 16,351 ±,38 1,18 ±,35 1,559 ±,35 2,268 ±,44 18,257 ±,16,849 ±,56 1,545 ±,37 2,395 ±,67 2,277 ±,31 1,45 ±,138 1,67 ±,29 2,41 ±,89 Table 1. Average values o skin thickness or each condition investigated 4. EXPERIMENTAL RESULTS The experimental mechanical tests show that three ailure mechanism have been take place: two dierent indentation ailures (called type I and type II) and tensile ailure o the lower skin. In Figure 3 are showed three typical and representative load/width ratio versus displacement curves, each one relative to a ailure mechanism experimental obtained P/b [N/mm] Indentation type I 8 Indentation type II 4 Tensile skin ailure Strain [%] Figure 2. Flexural test graphs

4 All curve show an initial linear trend with dierent slope. In particular samples ractured or tensile skin ailure and indentation type II have similar slope while the other one shows smaller slope. Indentation type II and tensile skin ailure occur, the irst at higher loads than the second one, with a signiicant loss o the load-carrying capability. In Figure 3 is shown the irst type o indentation. From this picture it is possible to notice that the core yields without skin ailure under localized load. This is a non catastrophic ailure mechanism due to progressive collapse o the cell structure o the oam. In the load application point the densiication o PVC oam takes place due to the initial buckling racture o the cell walls o PVC core; this brings to a strong local width reduction o the sandwich. At high delection the lower skin racture appears. Figure 3. Indentation type I In the case o indentation type II (Figure 4) the upper skin is heavy deormed due to the local applied load. This condition inluence the ailure mechanism that occur in this region due to the collapse o the upper skin. Ater that the crack propagates across the core reaching occasionally the lower skin. In all cases no delamination was evidenced at the skin-core interace. Figure 4. Indentation type II In the case o tensile skin ailure (Figure 5) the crack start in the lower skin. It propagate usually at the skin-core interace beore to propagate across the core. The crack inish at the upper skin-core interace.

5 Figure 5. Tensile skin ailure 4.1 Flexural tests at varying the parameter l The lexural tests results obtained at varying distance span l or a ixed value o skin thickness t are reported in Figure M 1 (a) 2 16 M 2 (b) II T.S. 2 4 I II I I II M 3 (c) T.S Figure 6. Maximum load at varying l or sample M1 (a), M2 (b), M3 (c) and M4 (d) From Figure 6 (a), it s possible to observe that samples M1 (characterized by one layer skin) show only indentation type II ractures except or the sample with l = 3 mm where the ailure occurs or indentation type I. With increasing distance span l a reduction o the maximum load P is obtained As shown in Table 6 samples M2 (characterized by two layer skin) show indentation I ailures or l inerior to 6 mm and indentation type II or the other samples except or the sample with l = 2 mm where tensile skin ailure occurs. The maximum load P decreases with increasing l or the samples that show indentation I instead it remains constant in the range o indentation II (Figure 6 (b)). Samples M3 (three layer skin) show indentation type II ailures or l between 8 mm and 16 mm, indentation type I or l less than 8 mm and tensile skin ailures or l greater than 16 mm. In the range o indentation I, the maximum load P decreases with increasing l. For higher values o l, P assumes constant values or each ailure I II M 4 T.S. (d)

6 mechanisms (an higher value or indentation II than tensile skin) as shown in Figure 6 (c). Samples M4 (our layer skin) show indentation II ailures or l between 8 mm and 14 mm, indentation I or l less than 8 mm and tensile skin or l greater than 14 mm. In the ranges o all ailure mechanisms obtained, the maximum load P presents similar trends like those ones shown or samples M3 as shown in Figure 6 (d). In the cases o samples M3 and M4 (compared with samples M1 and M2) the range o indentation type I becomes larger and ailure by tensile skin becomes prevalent mechanism in a larger area or high values o l with the consequence that range o indentation II become smaller. 4.2 Failure map From the experiments it has been ound how geometrical parameters inluence the racture mode o the sandwiches. These results can be better evidenced creating a ailure map, plotting the ailure mechanisms as a unction o thickness and length. In this map it is possible to notice the existence o three regions in which one o the ailure mechanisms experimental obtained is dominant. 3 2,5 t [mm] 2 Indentation I Indentation II Tensile skin ailure 1,5 1,5 Figure 7. Failure map Particularly or high t and l values tensile skin ailure occurs. At decreasing o l two transition are observed: - Between tensile skin and indentation type II or l values comprised between 14 mm and 18 mm - Between indentation II and indentation I or l values comprised between 3 mm and 6 mm At decreasing o t a transition is observed between tensile skin ailure and indentation II. This transition occurs or t values comprised between,3 mm and 1,7 mm. For intermediate l values ailure mechanism predominant is indentation type II (any value o t). Decreasing l a transition between indentation II and I is observed or l values increasing with t. A ailure mechanism will take place when its breaking load results smaller in comparison o those ones o other competitive mechanisms. As shown in Figure 7, indentation mechanisms are usually obtained whereas tensile skin ailure happens in a limited area o the map (with limited geometrical conditions). This phenomenon can be explained by theoretical consideration. Indentation breaking loads are a unction o core

7 mechanical properties instead tensile skin breaking loads are independent rom these properties as seen in equation (3). Due to low mechanical properties o PVC oam core used in present work, indentations breaking loads result lower than tensile ailure one in most o the tests. 5. CONCLUSIONS In this work, the static ailure modes o E-glass/epoxy composite and closed-cell PVC oam cores sandwich were investigated experimentally by three point lexural tests. The study was ocused on the dependence o geometrical parameters, such as skin thickness (t) and span length between support (l), in the ailure mechanism. Four type o symmetrical sandwiches with dierent number o layers (and consequently dierent thickness) have been prepared and characterized by three point lexural tests varying span length, l (rom 4 to 2 mm). From experimental results a ailure map in the plane t - l has been created identiying three regions o typical ailure modes o the sandwiches. A urther our purpose is to develop in the uture a theoretical model able to orecast the experimental ailure mechanism. In this way predictive capacity o this model could be quite helpul in the sandwich structure design. REFERENCES 1- Corigliano A.Z., Rizzi E. and Papa E., Experimental characterization and numerical simulations o a syntactic-oam/glass-ibre composite sandwich. Composite Science and Technology, 2;6: Konsta-Gdoutos, M.S., and Gdoutos, E.E., The Eect o load and geometry on the ailure modes o sandwich beams. Applied Composite Materials, 25;12: Zenkert D., An introduction to sandwich construction. Sheield, UK: Engineering Material Service; Plantema F., Sandwich Construction. New York: Wiley; Allen H.G., Analysis and design o structural sandwich panels. Pergamon, Triantaillou T., Gibson L.J., Failure mode maps or oam core sandwich beams. Materials Science and Engineering. 1987;95: Ashby M.F., Evans A.G., Fleck N.A., Gibson L.J., Hutchinson J.W., Wadley H.N.G., Metal oams: a design guide. London: Butterworth, Heinemann; 2 8- Chen C., Harte A.M., Fleck N., The plastic collapse o sandwich beams with a metallic oam core. International Journal o Mechanical Sciences, 21;43: McCormack T., Miller R., Kesler O., Gibson L., Failure o sandwich beams with metallic oam cores. International Journal o Solids and Structures, 21;38: Steeves C.A., Fleck N.A., Collapse mechanisms o sandwich beams with composite aces and a oam core, loaded in three-point bending. Part II: experimental investigation and numerical modelling. International Journal o Mechanical Sciences, 24;46: Steeves C.A., Fleck N.A., Material Selection in sandwich beam construction. Scripta Materialia, 24;5: Gibson L.J., Ashby M.F., Sandwich structures, Cellular solids: structure and properties. 2nd ed, Cambridge U.P., Cambridge; 1997.