Validated modelling methods for carbon composite sandwich panels

Abstract

HeliMods provides modification services for helicopter customisation and are looking to acquire more rigorous design allowables for composite design. Carbon-fibre composite sandwich structures are commonly used in the aerospace industry and are a major construction output for HeliMods. These composites are a versatile material and can often exhibit a superior strength-toweight ratio compared with metals. Their mechanical properties are also capable of being tailored to a specific application. (Cunningham, 2014) The use of carbon-fibre composites has several benefits and this can lead to more freedom in design when harmonising the demands of safety, efficiency and performance. (Bowler, 2014) As the designs being implemented by HeliMods can be weight and size critical, it is of importance to know how the structure will behave in failure in order to optimise design.The failure strength of mechanical fasteners in a carbon-fibre composite sandwich panel has been investigated. While design handbooks give the facility for basic loading cases, when combination loads are the subject of design it is commonplace to use superpositions of the basic cases - potentially leading to conservative results. (European Cooperation for Space Standardisation, 2011) This ensures a safe design is achieved, however can be inefficient in terms of the cost and weight.A fully-potted aluminium fastener in a panel comprised of epoxy pre-impregnated SE84LV carbon with a Divinycell H80 foam core was the subject of the investigation. By testing this basic structure subject to failure loading at shear, pull-out and discrete combined loads; several results were intended. Firstly, to confirm existing models or substantiate a new model for combination loading of this structure. Additionally to attain design allowables at B-basis rigour and characterise the failure modes. This subsequently could allow a simple finite element model to be developed to predict the type and magnitude of failure expected for a given angle.To achieve these aims a bespoke testing jig has been designed to mate with an Instron tensile testing machine and provide the facility to load at discrete angles. The test method involved loading the specimens at one mm⋅min-1 until a critical failure occurred, while supporting the structure locally. Six specimens were tested at angles ranging from zero to ninety degrees at fifteen degree intervals. Another eighteen specimens for pull-out and shear loading were tested to failure to substantiate the B-basis demands and observe batch-to-batch variability. Testing data was in good agreement with existing models, however variability was encountered when extracting B-basis allowables of 1780N in pull-out and 2130N in shear due to foam residue leftover from the machining process. An outcome of the testing showed a transition between two types of failure occurs at approximately sixty degrees loading angle with a fifteen degrees margin of confidence. This is beneficial in design analysis as the stronger failure mechanism can be achieved within a thirty degree range from shear loading. The ANSYS finite element model provides a conservative estimate of the angular loading, seeing an agreement of angular testing data within 12.5%. It predicts the failure in shear to be 3960N and 2626N in pull-out – in comparison with experimental results of 4186N and 2764N. Additionally existing theoretical models approximate the failure load to be 4920N and 2990N for shear and pull-out based on their respective failure modes.This testing also showed the potential for more consistent failure properties under the loading types investigated by ensuring that the machining process removes all residual core material before the inserts are potted. Based off design guidelines, the ideal failure mode under a shear load is not observed for the current structure. (European Cooperation for Space Standardisation, 2011) Therefore - subject to tooling limitations - increasing the potting size might provide greater mechanical strength and this is suggested to be the subject of further investigation. A final recommendation is to alter the geometry of the insert so a small portion of the bottom face is removed. This will ensure that enough epoxy will reliably fill underneath the insert, but will not reduce the structural integrity of the insert itself.The aerospace sector is a continually advancing industry and with the need for lighter and stronger designs while still conforming to safety requirements, investigations into effective failure modelling is ever important.