Understanding the Sequence of Damage in Complex Hybrid Composite- Metallic Structures Subject to Out-of-plane Loading Investigated Using Computed Tomography

The progression of damage in a hybrid composite-metallic structure subject to out-of-plane loading has been examined throughout an interrupted quasi-static-indentation test using micro-focus computed tomography for damage assessment. Two key damage events were identified from the loaddisplacement tests. The CT data shows that the first damage mechanisms to occur are yielding of the aluminium substrate and matrix cracking in the CFRP; the onset of these damage mechanisms occurs at a similar point in the loading process. With increasing load, delamination occurred between all plies in the CFRP layer. The bonded interface between the aluminium and CFRP was maintained at lower loads a crack initiated preferentially within the innermost CFRP ply rather than at the interface between the two materials. The CT data indicates that the second damage event on the loaddisplacement curve can be attributed to CFRP fibre fracture, consistent with contact forces. This work indicates that relatively low-level out-of-plane loads are likely to cause plastic deformation to the aluminium substrate, whilst the integrity of the fibre in the CFRP is maintained until higher loads. 1 INTRODUCTION Hybrid composite-metallic structures (HCMS) beneficially combine the characteristics of multiple material systems. HCMS are used in a variety of industrial applications where strength, stiffness and weight are important properties. In service, undesired out-of-plane loading events, such as low velocity impact (LVI), can occur during manual handling and transportation. In order to improve the design for damage tolerance of these structures it is necessary to understand the damage processes when subjected to out-of-plane loading. The progression of damage mechanisms in composite materials is generally well known [1, 2]. However, the majority of impact research has been focused on plate specimens in order to reduce the effect of the specimen geometry on the response. Alderson and Evans investigated filament wound EGlass/Epoxy composite tubes under QSI and LVI showing that relatively low loads can cause matrix damage leading to delamination initiation [3]; the study concludes that local crushing was the cause of the first drop on the load-displacement curve. Matemilola and Stronge investigated carbon fibre reinforced pressure vessels with an internal layer of high-density polyethylene [4], concluding that the first load drop was caused by delamination. Curtis et al. [5] showed that QSI and LVI are comparable during out-of-plane loading of glass fibre reinforced plastic tubes, however Evans and Alderson showed that this equivalence was dependent on specimen support conditions [6]. The differences highlighted between these studies could be related to differences in the test specimen’s material, geometry and layup. T. Allen, S. Ahmed, P.A.S. Reed, I. Sinclair, S.M. Spearing The presence of multiple materials, a non-symmetrical composite layup and a residual stress state makes the response of HCMS complex in comparison to traditional composite plate or tube structures when subject to out-of-plane loading. The nature and sequence of damage mechanisms is not fully understood in HCMS. Many damage mechanisms such as: fibre breaks, delamination, intra-laminar cracks, metallic yield and interface debonding between the different materials may occur. Understanding the order in which these damage mechanisms occur, and the presence of any interdependencies between them will provide valuable insight for damage tolerant design of HCMS. LVI behavior of similar HCMS structures [7-9] has largely focused on the residual mechanical properties after impact events, with little consideration given to the mechanics and progression of damage during loading. The present work aims to improve understanding of the damage process in HCMS in order to better inform the design process for these structures. 2 METHOD AND MATERIALS 2.1 SAMPLE DETAILS The samples tested in this study were specially manufactured for the purposes of this work. The HCMS investigated was made of a 6061-T6 aluminium alloy shell, carbon fibre reinforced plastic (CFRP) and glass fibre reinforced plastic (GFRP). The thicknesses of the aluminium, CFRP and GFRP were 2mm, 5mm and 1mm respectively. The composite layers were filament wound in a combination of circumferential and wider angle helical warps onto the aluminium shell and cured in an out-ofautoclave process. 2.2 OUT-OF-PLANE LOADING QSI tests were conducted on a servo-mechanical test machine at a crosshead speed of 2mm/min. The sample was supported with a shallow angle V-groove block to centrally locate the specimen. A 16mm hemispherical indentor was used. Displacement measurement was taken at the crosshead of the machine. In order to account for any flexibility in the test machine a compliance test was completed. The indentor was loaded onto a rigid steel block of similar height to the sample. This was used to calculate a compliance correction that was then applied to all QSI tests to account for the compliance of the load train, and to allow for a better estimate of the responses of the HCMS specimens. The load data in this paper is presented on a scale normalised from the peak load. 2.3 COMPUTED TOMOGRAPHY The CT scans were completed on a Nikon Metrology HMX μCT scanner at the μ-VIS centre, University of Southampton. A custom jig was used to offset the sample from the centre of rotation to complete a local region-of-interest scan of regions of the sidewall. The scanner has a 225kV X-ray source and Perkin-Elmer 1621 2048 x 2048 pixel flat panel detector. An electron accelerating voltage of 170kV was selected, with a tungsten reflection target and a beam current of 151μA. 3142 projections were taken during the 360o rotation of the sample, with sixteen frames per projection taken to reduce noise. 3D reconstruction with a voxel size of 50μm was created using the filtered-back projection method within the Nikon CT Pro 2.0 package. Image processing and analysis was completed using the software packages ImageJ and VGStudio max 2.1. 2.4 EXPERIMENTAL METHOD A typical load-displacement curve for the HCMS test specimen reveals two key features of interest (Figure 1). The first is indicative of a reduction in the structural stiffness occurring at ~50% off the top load, the second a distinct load drop at ~95%. To describe the sequence of damage a QSI test was performed, with seven CT scan points along the load path for detailed non-destructive damage assessment ( Table 1); the peak load was selected as just after the load drop feature. An interrupted test on a single specimen was used in order to account for potential specimen variability seen in preliminary tests. The specimen was tested ex-situ from the CT scanner; i.e. the scans were completed 20 International Conference on Composite Materials Copenhagen, 19-24 July 2015 in an unloaded condition. After the first scan the sample was loaded to first load step and then rescanned, this was repeated through the remaining load steps. Scan 1 Scan 2 Scan 3 Scan 4 Scan 5 Scan 6 Scan 7 Scan 8 0 (reference) 0.40 0.45 0.50 0.60 0.75 0.95 1.00 Table 1: CT scanning points Figure 1: A typical uninterrupted load displacement plot for the HCMS test specimen 3 RESULTS 3.1 QSI LOADING The interrupted test (Figure 2) followed a very similar profile to the uninterrupted tests (Figure 1) indicating that the use of an interrupted test methodology is sufficient to describe the process of damage in the specimen. The first two load steps at 0.40 and 0.45 demonstrated a largely elastic structural response. After 0.45 the response became non-linear. Between the 0.50 and 0.75 load steps the slope of the load-deflection curve remains approximately constant. Increasing permanent deformation is seen with each load step when the structure was unloaded. At a load of ~0.98 a distinct load drop occurred. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 Fo rc e (n or m al ise d) Displacement (mm) T. Allen, S. Ahmed, P.A.S. Reed, I. Sinclair, S.M. Spearing Figure 2: Interrupted quasi-static indentation test 3.2 REFERENCE SCAN The first scan point was of the sample in the unloaded virgin condition. A cross section in the longitudinal plane shows a typical as manufactured-condition for these HCMS. Due to the out-ofautoclave process a void content of ~3.5% is typical for these structures [10]. Figure 3: Unloaded reference sample 3.3 PROGRESSION OF DAMAGE 3.3.1 INITIAL ELASTIC RESPONSE The first load step indicates a largely elastic response (Figure 2), however some permanent indentation was present on the surface of the sample. This was attributed to local indentation in the external gel coat layer and GFRP. No evidence of damage was present in the main structural components of the aluminium and CFRP (Figure 4); this was also the case in the 0.45 load step. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 Lo ad (n or m al ise d) Displacement (mm) Load Step 1