Modelling the effect of voids in Composite Components

One of the key strategies in reducing the environmental impact of the aerospace and renewable energy sectors is to replace metal alloys with advanced fibre-reinforced composites. Such materials offer not only immediate performance benefits through lighter weight and improved aero-elastic behaviour, but also provide considerable life-cycle benefits through greater durability (e.g. better corrosion and fatigue resistance) and lower overall energy requirements for manufacture. One of the downsides of using composites, however, is their particular susceptibility to manufacturing defects, most commonly porosity caused by the entrapment or precipitation of gasses during final consolidation. In order to achieve acceptable production costs or scrap rates, some level of material porosity must be tolerated, which is often done by the introduction of conservative design allowables and safety factors mostly based on empirical methods. Conservative design rules tend to negate the benefits of adopting composite materials in the first place, a problem which may become critical as traditional autoclave manufacture is replaced by more energy-efficient 'out-of-autoclave' manufacturing techniques. This project will focus on the development of novel multi-scale analysis methods to predict the asmanufactured macro-scale mechanical performance of composite structures taking into account detailed (meso- and micro-scale) porosity characteristics such as local void content (relative to part geometry), local distributions of void sizes and morphologies, and the location of voids relative to the meso- and micro-scale material architecture. This project builds on previous work by this research group on porosity characterisation, process modelling, and multi-scale analyses of damage and fracture in composites. State-of-the-art techniques will be combined and further developed, including (but not limited to): - Process modelling of composites to predict the as-manufactured material architecture at the micro-scale (individual fibres), lower meso-scale (individual yarns or fibre bundles), upper mesoscale (individual layers or plies), and macro-scale (component-level); - High Performance Computing techniques and software for the analysis of damage and fracture at the micro- and meso-scales, enabling the analysis of thousands of different material and loading configurations, accounting for material variability and other uncertainties, which in combination with statistical inference methods can be used in design-of-experiment analyses; - Multi-scale modelling strategies based on response surface methods and/or machine learning, to implement the learned/inferred material behaviour above in the form of macro-scale predictive models suitable for the detailed design and design-for-manufacture of composite parts. Such a modelling capability has the potential to replace much of the empiricism currently involved in accounting for manufacturing defects in composites, leading to faster design cycles, fewer physical tests, more accurate design safety factors, lower scrap rates, and more efficient composite structures in terms of their life-cycle carbon footprints. This project falls mostly within the EPSRC research area 'Materials Engineering - Composites'. The work is conducted at the University of Bristol in collaboration with BAE Systems plc under the CerTest Programme Grant (EPSRC no. 18939/05 2018-4088).