University home > Advanced Composites Collaboration for Innovation and Science > Research > Design, analysis and failure
Fundamental experimental studies yield understanding of physical and mechanical behaviour that provides the foundation for analysis methods to predict performance, and tools to design and manufacture optimal structures. Select from the options below for an overview of work undertaken in ACCIS under this theme.
With the use of increasingly large composite components on aircraft comes the need for reliable design guidelines to be able to extrapolate data from laboratory tests with confidence to full size components, thus reducing the need for full-scale testing. Notches or holes are common in aircraft components for joints, access, or to allow wires to run through solid pieces and cause significant stress concentrations. Experimental and numerical studies allow us to understand the mechanisms responsible for failure from notches and to develop models to predict the resulting scaling effects.
Textile composites encompass 2D and 3D woven materials as well as Non-crimp fabrics and braids. 3D weaving provides a method for adding through thickness reinforcement to composite components. It additionally creates the ability to produce net-shape pre-forms, thus reducing manufacturing and assembly costs. In order to use textile composites in engineering designs it is necessary to understand the effect of the reinforcement on mechanical performance and to characterise the mechanisms by which they fail. It is possible to produce a great variety of weave architectures which can cause considerable variation in properties. On-going work aims to develop a better understanding of the failure mechanisms, methods of testing and numerical modelling.
Optimisation of laminates is made difficult due to the large number of potential lay-ups available and the multi-modal nature of the design space. A two level optimisation strategy is followed whereby continuous gradient based methods are used at the first level and stochastic methods, such as genetic algorithms, in conjunction with lamination parameters are used at the second level. This approach offers great increases in computational efficiency in the optimisation of multipart composites structures.
Anisotropic composites provide a means for weight efficiency by elastic tailoring. By exploiting anisotropy we can create benign stress fields that delay the onset of buckling phenomena. On-going work involves analytical, finite element and experimental investigation of buckling and post buckling. Work on analytical and finite element modelling of prismatic structures with flexural/twist coupling has lead to aeroelastic tailoring. Many laminates exhibit orthotropic properties but are not symmetrically laminated, and are often precluded from design because of their lack of symmetry. Non-standard asymmetric laminates can however sometimes have superior properties.
One major disadvantage of composite materials is that they are very susceptible to impact damage both at low and high velocity. The damage which occurs is complex and is made up of a combination of surface indentation, fibre breakage, matrix cracking, fibre pull-out and delamination. Work at the University of Bristol aims to develop a better understanding of the mechanisms of failure under the most severe loading scenarios. Our laboratory equipment includes an instrumented drop weight impactor, tensile Hopkinson bar and high speed digital video cameras. Advanced numerical models can help enhance our understanding of the way materials fail under high rate loading.
The ability to predict failure and mechanical properties is of great importance for structural design of components. In many cases this requires the development of new numerical methods, for example, new material models for finite element analysis, closed form solutions or techniques that account for the multi-scale nature of composite materials. These all require a sound understanding of the physical behaviour of materials and are carefully validated against high quality, data rich, experimental results. Examples of previous work include the development of cohesive elements for delamination, fatigue formulations of cohesive elements and multi-scale modelling of textile composites failure.
Ideally, composite components should be joined together without having to drill a multiplicity of holes for fasteners – to achieve this we can use adhesive bonded joints. Such joints may also be used to repair composite structures that have been damaged in service, for example by impact events. Bonded structures can be made to carry very heavy loads (our work has designed and manufactured a bonded beam capable of carrying more than 40 tonnes under test). To generate the confidence to use such approaches a significant effort has been applied to understanding failure and crack propagation in bonded joints.
Composites are typically designed using properties obtained from well controlled laboratory scale specimens. However when more complex components are manufactured in a production environment it is possible the defects can occur in the material, resulting in a knock-down of mechanical performance. For the safety, reliability and integrity of composite structures it is important to have an understanding of the effect of defects. Work in ACCIS has manufactured and tested a range of artificial defects such as in and out of plane wrinkling, automated fibre placement (AFP) gaps and overlaps, and voids. These tests further the fundamental understanding of the nature of defects and also drive forward the development of advanced numerical models for failure prediction.
Laminated composites are inherently weak in the through-thickness direction. This has led to a number of strategies being proposed to enhance the through-thickness performance, to resist the initiation and propagation of inter-laminar cracks. Through-thickness reinforcements being investigated in ACCIS include z-pinning, tufting and 3D weaving. For these, investigations are centred on understanding their role in toughening of laminates, development of advanced numerical models, and creating novel material and geometric configurations. Additionally the potential for sensing and multi-functional applications of through-thickness reinforcement is being investigated.
Full field and non-contacting techniques for measuring surface displacements and strains are developing very rapidly. They offer scope to detect damage by the effect it has on the strain distribution, for example cracking around a hole in a composite loaded in tension. They can also be used to measure material properties from tests on specimens with non-uniform stress distributions by processing the results using the virtual fields method. Video extensometry is another powerful technique enabling deformations and strains to be measured remotely in real time by tracking the relative displacement of points on the structure or by post-processing high speed video from impact tests.
ACCIS has significant research activity linked to the development of composite wind and tidal turbine blades. In 2009, we launched a formal partnership with Vestas Wind Systems and we now have a range of projects to linked to structural design, analysis, manufacturing and in-service health monitoring. We are also collaborating with a number of industrial partners on the development of improved materials and design methods for tidal turbine blades, focusing on delamination modelling under static and fatigue loads. These activities directly complement our work with the Aerospace industry and there is strong potential for further knowledge transfer in the future.
Further information: Composite developments in the renewable energy sector (PDF, 2.22MB)
The performance of composites can degrade under repeated application of loads well below those required to cause static failure. Our research is developing new experimental methods for testing composites under high cycle fatigue as well as under fully reversed loading. Fatigue testing is used to gain understanding of the damage propagation from a range of features such as open holes, through-thickness reinforcement, ply drops and defects. Novel numerical models have been developed for the integration of fatigue algorithms into cohesive interface elements and these are now being deployed to predict fatigue damage growth from these more complex cases.
Structural engineering research is primarily concerned with existing capacity of structures and strengthening methods. Bridge strengthening entails adhesive bonding of plates of Advanced Composite Materials such as Carbon Fibre Reinforced Plastic, to the cracked zones of the existing structure. The bond characteristics of the adhesive-to-concrete connection are being investigated and new design procedures developed. Fatigue testing on novel lightweight fibre reinforced road bridge decks and composite reinforcement of masonry panels to reduce earthquake vulnerability are other areas of research.
Internal architectures of modern composites exhibit enormous diversity. The development of new manufacturing methods is primarily driven by two needs: (1) to minimise efforts of draping and laying-up, (2) to optimise composite performance by minimising defects and reinforcing weak directions. There are two major technological trends, which comply with both the needs: (1) creating 3D preforms (e.g. 3D weaving, 3D braiding, structural stitching, tufting), which reinforce out-of plane direction and suppress delaminations, (2) robotic placement of elementary fibre bundles along complex trajectories (e.g. embroidery, automatic tow placement), which optimise in-plane load flow.
Both the technology routes tend to produce peculiar fibrous geometries. In contrast with the composites obtained by laying-up of unidirectional or 2D textile plies, the elementary building blocks (fabric repeat, unit cell) of these composites is comparable to component dimensions or does not exist at all. The latter is particularly explicit for the cases of non-flat component geometries, locations of shape transitions, sharp angles, conjunctions, and the profiles of complex shapes. It leads to an interesting situation when there can be no clear representative volume element (RVE) on the yarn level and hence, the composite effective properties depend on the component shape and loading. These composites are not materials in the traditional sense of homogeneous or homogenizable medium. Hence, the analysis of these materials demands novel modelling approaches and scale separation techniques.
A key limitation of high performance composites is their inherent lack of ductility and the usually sudden and catastrophic failure with poor residual load bearing capacity if any. The unfavourable failure character of composites renders them unsuitable for applications, where sudden failure cannot be tolerated mainly for safety reasons. The key challenge of this research theme is to develop materials which exhibit pseudo-ductility, and fail in a safe, controlled and progressive manner with detectable warning before final loss of integrity. Extensive work is in progress within the framework of HiPerDuCT programme in designing and testing different architectures involving novel materials to demonstrate ductility in composites while maintaining high performance.
University of Bristol,
Bristol, BS8 1TH, UK
Tel: +44 (0)117 928 9000