This work is focused on the prediction of elastic behavior of short-fiber reinforced composites by mean-field homogenization methods, which account for experimentally determined and artificially constructed microstructure data in discrete and averaged form. The predictions are compared with experimental measurements and a full-field voxel-based approach. It is investigated, whether the second-order orientation tensor delivers a sufficient microstructure description for such predictions.

Performance-based design strategy allows the use of composites in fire risk situations. In this approach, the fire behaviour of the composite materials and components can be comprehensively qualified and accurately envisaged. The number of tests required to qualify a new composite material can be minimised by developing a model for prediction of the thermal response of composite materials, thereafter, validated against standard fire-test methods. Additionally, once the fire behaviour under the heat transfer conditions of the specific case of a standard furnace fire test is well comprehended, then the same model may be generalised to foretell thermal response under more realistic fire situations. A comprehensive set of experiments to measure the fire performance of composites can be cost-intensive and time-consuming. In particular, large-scale fire tests, such as a room corner test, requires special settings with advanced facilities to implement the fire scenario in a building and obtain reliable data. The experimental results cannot always capture the complex scenario that occurs in real life. Hence, the development of models to predict thermo-structural response of a burning composite is of necessity to provide an insight into the reaction to fire properties and to the fire-induced damage processes of the composites. State-of-the-art fire models based on computational fluid dynamics (CFD) typically include pyrolysis and combustion sub-models. The models predict the fire growth and the spread quite accurately. In addition, the fire models also have mass, momentum and energy balance sub-models. Moreover, the phenomena of pyrolysis and combustion of materials are severely complex in nature. The scenario gets even worse when fire and natural fibre composites are compounded to that complexity. In this work, comprehensive experimental and numerical analyses have been carried out to compare the fire reaction properties of wool and flax based polypropylene composites under horizontal and vertical orientations. Two natural fibres, namely flax (cellulose based) and wool (protein based), polypropylene (PP) and epoxy matrix were used to prepare short and long fibre reinforced composites to evaluate the fire performance under different orientations. Moreover, extensive numerical analysis results on the fire reaction properties, burn time and surface temperature distribution have been reported. Fire Dynamics Simulator (FDS®) was employed to develop a numerical predictive model for the cone calorimeter test. Moreover, a systematic methodology has been established to evaluate the validity of applying test data from “horizontal” material orientation to predict the fire performance of composite materials. The study consists both experimental and numerical investigations to assess the fire performance of natural fibre reinforced composites (NFRCs) with respect to the sample orientation, fibre type and fibre volume fraction under different heating irradiances. In addition, the effects of orientation on the ignitability, heat and smoke production of flax fibre reinforced composites were investigated in the cone calorimetry tests. Finally, a coupled fire-structure model combining the finite volume and finite element methods, which captures the essential physics of the problem has been developed. The model is based on a multi-physics framework where the essential physics pertaining to the combustion process of the fire and resultant thermo-mechanical response of the structure of natural fibre reinforced composites, has been incorporated. Additionally, a relatively new concept of adiabatic surface temperature has been introduced as a practical means to transfer data between fire and thermal/structural models at the gas-solid interface. The model has been able to predict the temperature, stress distribution and deformation behaviour of composite beams under combined thermal and mechanical loads.

A discontinuous fiber-reinforced thermoset material produced by the Sheet Molding Compound process is investigated. Due to the process-related fiber orientation distribution, a composite with an anisotropic microstructure is created which crucially affects the mechanical properties. The central objectives are the modeling of the thermoelastic behavior of the composite accounting for the underlying microstructure, and the experimental characterization of the pure resin and the composite material.