Behaviour of carbon/epoxy composite sandwich panels with sustainable core materials subjected to intermediate velocity impacts

Resumen

Composite materials are used by multiple industries to build lightweight structures due to their high specific strength and stiffness and the possibility to customize their properties for different design requirements. The aerospace industry has largely benefited from their use in multiple applications, ranging from small general aviation aeroplanes to commercial airliners and spacecraft. Likewise, other high-performance industries such the wind-energy, automotive, railway and maritime industries use large amounts of composite materials in the development of wind turbine blades, sport-cars, high-speed trains and high-speed boats. A sandwich panel is a type of lightweight structure made by two strong and stiff face-sheets separated by a lightweight core material. They are used in lightweight structures for load-carrying applications as a way to increase the bending stiffness while maintaining a low weight. Most of the materials available in the form of a thin plate can be used as a face-sheet in a sandwich panel this includes metallic and non-metallic materials such as plastic reinforced composites. The core, on the other hand, is usually made from either corrugated materials, balsa wood, honeycombs of different materials or cellular foam cores. During their lifetime these structures are subjected to impact events such as the accidental drop of tools during assembly, bird strikes, hailstone impact, or even Foreign Object (FO) impact of stones, debris, etc… Damage produced by impacts can compromise the integrity of a structure reducing its residual strength and stiffness, causing premature failure of a component under service loads. The structural damage created by foreign object impact is usually defined as impact damage. The study of impact damage in sandwich composite structures is complex since it involves not just penetration and delamination (as in conventional laminates) but other modes such as core crushing and face-sheet debonding. The damage varies across the sandwich thickness and in some cases it may affect one face-sheet while the other is still intact In the past decades, there have been many studies in the area of sandwich composite structures subjected to impact loads, some using purely experimental, analytical or numerical approaches or a combination of them. However, at this date, most of them have focused on Low-Velocity Impacts (LVI) on sandwich panels with metallic or composite face-sheets and metallic honeycomb cores. The response of sandwich panels subjected to Intermediate Velocity Impacts (IVI) has been little explored, despite it encompasses a broad range of impact events. At the fundamental physical level, IVI can be categorized as a transition region between Low-Velocity Impacts (LVI) and High-Velocity Impacts (HVI) in which the physical mechanisms taking place are a combination of those encountered separately in LVI and HVI events. Depending on the impacted material IVIs are considered to fall near the 50 m/s occurring usually with a blunt impactor and is characterized by short impact duration and a response dominated by flexural and shear waves. Typical IVI occurs from events such as road debris impact, a lower end velocity of a projectile, bird strikes, hailstone impacts, and FO impact that could affect general aviation aeroplanes, wind turbine blades, and high-speed trains. The energy absorbed in the system can be divided into two components: energy absorbed in creating damage and energy absorbed by the system through vibration, heat, elastic response. There is still limited understanding of the physical process occurring during an IVI on composite sandwich panels and multiple questions are still in debate in the scientific community such as the identification of the dominant physical mechanisms, the damage morphology and the influence of different impact parameters. These factors are relevant not just from a theoretical perspective but their understanding could bring advances in practical engineering applications. This PhD thesis studies the mechanical response, impact process, and damage mechanisms taking place in an IVI over a sandwich composite panel made of woven carbon/epoxy face-sheets with either agglomerated cork core or PET foam core. The selected carbon/epoxy face-sheet material (Hexcel AGP193-PW) is studied due to its widespread use in industry and because its brittle behaviour makes it highly susceptible to impact damage, limiting its use in applications where impacts loads are expected. The selection of two different types of core materials is motivated to study the difference in the mechanical response of the sandwich panel using core materials with different mechanical behaviour. A semi-rigid and highly elastic naturally derived cellular material as agglomerated cork (CORECORK NL-20) and a rigid polymeric foam with an elastic-plastic behaviour as PET foam (AIREX T92.200). This selection of core materials is also motivated by the increased public interest in environmental awareness and the demand for scientific research on more sustainable materials that could substitute conventional core materials in particular applications. In this regard, agglomerated cork and PET foam are potential candidates, as both can be produced by the residuals of other industries. This thesis applies a numerical-experimental methodology based on the building block approach used for aircraft certification. The building block approach is a step by step series of physical and virtual tests of increasing complexity that serves to create reliable simulation models that support the design of composite structures. The validation phase is intended to determine how accurate is the model compared to the real world keeping the perspective of its intended use. In this context, the thesis is divided into three main parts. In the first part of this thesis, the face-sheet and core are treated independently to understand their unique dynamic response and select appropriate constitutive material models for the FEA model implementation. For the face-sheets the continuous damage approach is used to model both inter-laminar and intra-laminar fracture behaviour. In particular, the cohesive damage model is used for delamination while the continuous damage model developed by Jhonson, Ladeveze and Le-Dantec is used to model intra-laminar damage in woven composite laminates. The suitability of these models are validated by implementing independent FEA models in Abaqus for fracture tests CT, DBC, ENF for modes I and II as well as ballistic impact test over monolithic carbon/epoxy laminates. The compressive and tensile response of the core materials (agglomerated cork and PET foam) is also studied by performing static and dynamic characterization tests. Tensile and compressive quasi-static characterization is carried out using a universal testing machine equipped with a cross-head displacement sensor while dynamic compression tests are performed using an instrumented drop-weight testing machine, high-speed video recording and image tracking. In all characterization tests, 2D Digital Image Correlation (DIC) is used to study the lateral Poisson effect. The collected data is then used for the implementation and validation of the non-linear material models by developing an FEA model for dynamic compression. The second part of this thesis studies the IVI event of the whole sandwich panel. This is done by performing a set of experimental impact tests and implementing a detailed explicit/nonlinear FEA model, which is validated against experimental results. The experimental tests are conducted using a gas gun and the response of the sandwich panel is studied using non-contacting state-of-the-art techniques such as high-speed video recording, 3D-DIC analysis, and image tracking. Different damage and failure mechanisms are analysed by studying the post-impact damage of the specimens using visual inspection and X-ray computed tomography. As a result, diverse qualitative and quantitative information is acquired and diverse hypotheses about the physics involved in the impact event are suggested. The obtained experimental data is then used to develop a validated FEA model for the whole sandwich panel which is used to study the phases and mechanisms of damage evolution present during the impact process, something that is not possible to obtain experimentally and provides a valuable tool to understand the phenomenon. A nonlinear/explicit finite element model is modelled with continuous damage models for intra-laminar and inter-laminar damage in the face-sheets while the non-linear behaviour of the core materials is included applying the validated constitutive material models and multiaxial Tsai-Wu failure criterion. This model is also used to perform a comparative analysis of different parameters such as impact velocity, core thickness, impact angle, and axial preloading and how they influence the mechanical response of the sandwich panels under IVI. Experimental and numerical results suggest that at the most general level, the impact process is dominated by different interacting physical mechanisms such as elastic deformation of the panel, inter-laminar and intra-laminar fracture of the face-sheets, non-linear core deformation, multiaxial core failure and core-face-sheet debonding. Impact velocity seems to be a critical parameter that dominates the mechanical response, impact process and damage extend. At impact velocities well below the penetration threshold (impact velocity required to penetrate the top face-sheet) most non-conservative energy dissipation mechanisms are negligible and the impact process is dominated by the elastic response of the panel and the subsequent projectile rebound. At impact velocities closer to the penetration threshold but still lower than it, delamination and face-sheet fracture start dissipating more energy. Both numerical and experimental results suggest that they do not act independently but interact with each other during the impact process. This idea is supported by the correlation between growth in intra-laminar crack length and delamination growth. Multiaxial compressive and shear deformation in the core produces local failure around the impact region triggering the formation of small internal indentations below the top-facesheet which are not easily detected from the surface of the sandwich panel. The effect of these internal core indentations could affect the post-impact performance and fatigue life of the panel under service conditions. At impact velocities higher than the penetration threshold the top face-sheet fractures in a petal (or diamond) shaped crack pattern which is associated with large delamination extend. Multiaxial core failure occurs around the impact region and in the interface between the bottom face-sheet and the core. This interface failure is not visually detectable from the surface of the panel but can be multiple times larger than the projectile diameter and constitute a risk to the post-impact integrity of the panel. About the differences in the panel response when changing the core material it is observed that the strength and stiffness of the core seem to influence the out-of-plane displacement of the top face-sheet during impact. Panels with agglomerated cork show larger displacements than the more rigid and stiff PET foam. Indentation and core failure are visible in both core materials at velocities near or higher to the penetration threshold, however, the physical mechanisms are different. For example, agglomerated cork is dominated by a large elastic response and fracture of the agglomerate grains while PET foam is dominated by an elastoplastic response with dynamic crushing near the impact region. FEA results suggest that agglomerated cork provides a beneficial effect on the top face-sheet delamination area if compared to PET foam; however, there is not sufficient experimental evidence to back up this hypothesis. The influence of other impact parameters is also studied using the developed FEA model. Core thickness seems to play an important role in the rebound of the projected below the penetration threshold and influence the fracture pattern in the core. Besides this, the damage and plastic dissipation energies seem to be unaffected by the core thickness for most of the analysed relative thicknesses. The obliquity of the impact seems to influence the extent of the damage across the panel thickness, as well as the rebound velocity of the projectile and the amount of energy dissipated as friction. For small impact angles, the panel response is similar to the perpendicular impact while for large impact angles the face-sheet damage is more superficial and the severity is lower due to the higher tangential rebound of the projectile. Another studied impact parameter is pretension (either tensile or compressive) which is found detrimental to the impact resistance of sandwich panels since it encourages the growth of intra-laminar cracks and speeds up the extension of the delaminated area during impact. The third part of this thesis studies the hailstone impact on the sandwich panel. First a dynamic/explicit FEA model using Smooth Particle Hydrodynamic (SPH) method is implemented to validate the hailstone material model which is compared with experimental results from the literature. This model is later implemented together with the FEA model for the whole sandwich panel to study different impact scenarios with different impact velocities and hailstone sizes. Results suggest that there is a strong interaction between the dominant physical mechanisms in the sandwich panel (e.g. elastic response, face-sheet damage, core failure, etc…) and the fragmentation of the hailstone. Core fracture in the intersection between the core and the bottom face-sheet is a dominant failure mode that extends quickly and appears early in the impact process. Besides this, damage extend is highly dependent on the impact velocity and the size of the hailstone.