Multiscale Damage Modelling of Natural Fibre Composites

Abstract

Natural fibre composites (NFCs) display several beneficial traits, including low density, renewability of sources, and full biodegradability when combined with polymers such as poly(lactic acid ). Environmental concerns and sustainability requirements have led to increased application of NFCs in load-bearing primary structures. Structural application of NFCs requires accurate analysis of structural responses and prediction of failure modes for structural integrity assessment. Composite materials often are comprised of inherently, hierarchical multiscale, multi-material structures. This leads to complex failures mechanisms, whereby cracks, defects or damage can involve different physical processes and occur over several length scales, contributing to the overall material damage and ultimately failure. For NFCs, the multiscale nature is complex because the yarns or fibres are often comprised of two or more scales. Understanding the complex nature of failures can assist in an improved design of NFC-based structural materials. This can be achieved through an integrated multiscale model development coupled with experimental tests at various scales. Computational multiscale modelling techniques are specifically meant to deal with the multiple scale nature of material failures, and have been applied to the failure of composites. Hence, with respect to dealing with NFC failures, the primary aim of this thesis was to understand the failure mechanisms at relevant scales and predict their overall failure using multiscale computational techniques. Part of the thesis also deals with quantification of damage in NFCs within a continuum damage framework. As a foundation, only unidirectional composites have been explored in this work. Among natural fibres, flax is widely cultivated, has desirable mechanical properties, and is readily available in forms suitable for composite production. The length scales affecting NFC behaviour could extend several levels down to the molecular level. However, in the interest of the practicality, i.e. relevant physical processes and reasonable solution times, the analysis scope in this work extends down to the fibre and interface level. At the lowest length scale (microscale) in this work, the fibre and fibre-matrix interface properties were studied. The intermediate scale or mesoscale consisted of studies on impregnated yarns. Finally, full composite specimens and models were studied at the highest scale (macroscale). The multiscale framework implemented as part of this work bridges the macroscale and mesoscale systems. Four composite material systems have been studied involving two different polymer types and two different volume fractions, namely flax and polypropylene systems, with volume fractions of 0.22 and 0.41, and flax and epoxy systems, with volume fractions of 0.41 and 0.51. The flax used in this work was obtained in the form of a unidirectional fabric. At the microscale, the elastic stiffness and failure behaviour of flax fibres were determined through tensile and shear tests on fibres. Failure of the fibre-matrix interface was also investigated using microbond tests on interface specimens. Flax, like other natural fibres, demonstrates a spread in properties that could be attributed to several factors, including the location of extraction of the fibres from the stem, the age of plants and the environmental factors during growth. Analytical modelling was applied to model the fibre properties, considering different compositions of the fibre cell walls. Numerical models were then constructed to model the interface properties, and compared to the predictions using a shear-lag model for the microbond test. The mesoscale considered involved yarns impregnated with polymer. Flax yarns from the fabrics were extracted and impregnated with polymer. These were then loaded to failure in tension. The impregnated yarn failure was modelled numerically, incorporating both the yarn geometry and the properties at the fibre and interface level, as obtained from the microscale studies. Finally, at the macroscale (structural) level, the tensile, compressive and flexural properties of the flax fibre fabric composite systems were investigated using relevant standardised tests. In this way, numerical and analytical models were combined with the experiments at the micro- and mesoscales, so that reliability and confidence could be established in the ability to model the material properties at those scales. Material damage evolution laws for the mesoscale impregnated yarns were estimated from the mesoscale finite element (FE) models once the fibre and interface properties were established from the microscale analytical and FE models. Material damage was also quantified experimentally via image analysis of specimens damaged by tensile loading. The damage laws obtained from mesoscale FE models are termed numerically estimated damage rules (NED rules) in the text. Similarly, the damage laws from damage quantification experiments are termed as experimentally quantified damage rules (EQD rules). The NED and EQD rules were used to simulate tensile failure of the flax fabric composites. A two-scale coupled multiscale homogenization technique was implemented to simulate the failure behaviour of flax composite specimens subjected to bending. The implementation was FE-based, with the data exchange required for coupling performed using a custom FORTRAN code with the UMAT interface of the ABAQUS FE solver. The damage evolution laws obtained from the microscale FE models were utilised to assign material damage. The multiscale models predicted the bending behaviour reasonably accurately, with the predicted flexural strength being within 12% of the experimental results. Validation studies using a glass/epoxy composite system yielded a flexural strength value 12.6% lower than the test average. In summary, a multiscale modelling framework has been developed and utilised to predict the failure of natural fibre (flax-based) composites with reasonable accuracy. The implementation had taken into account several complex factors, such as the inherent scatter in the mechanical properties and geometry of natural fibres (such as flax). The framework developed can be extended to assist in the prediction of NFC properties and behaviour in structural applications. The results of this work and its potential extension can be used to develop NFC components for structural applications with improved durability and safety, thus contributing to sustainable engineering materials and products development.