Micro-scale models of unidirectional fibre-reinforced composites accounting for strain gradients in the epoxy resin
Topic(s) : Material and Structural Behavior - Simulation & Testing
Co-authors :
Igor André RODRIGUES LOPES (PORTUGAL), Nathan KLAVZER , Carolina ERDEM DINLER , Francisco Manuel ANDRADE PIRES (PORTUGAL), Pedro CAMANHO (PORTUGAL), Thomas PARDOEN (BELGIUM)Abstract :
Multi-scale modelling of unidirectional fibre-reinforced composites relies on accurate simulation of the microscopic deformation mechanisms. Representative Volume Elements (RVEs) of the microstructure represent the geometry of the fibres, the matrix and their interfaces, and can be employed in finite element simulations to extract ply-level properties, to build databases of homogenised responses for surrogate models, and to provide a deeper insight of the micro-scale mechanisms, as well as of their interactions. To that end, appropriate high-fidelity constitutive models are required for each constituent. For instance, non-linear elastic models accounting for transverse isotropy can be employed for the fibres, elasto-plastic-damage models are usually applied for the matrix, and cohesive zone models are included in the fibre-matrix interfaces.
Concerning the matrix models, recent research [1,2] revealed that classical elasto-plastic constitutive models are not able to capture the strain fields accurately, especially in regions of the epoxy matrix confined between carbon fibres. In this case, the strain localisation is overpredicted by such classical models, which cannot cope with the large strain gradients observed. Alternatively, strain gradient plasticity models have the potential to better predict the deformation fields in these critical microscopic regions. Consequently, the macroscopic strength that is often underestimated with classical models [1] for the matrix can be captured more accurately, resulting in more reliable multi-scale models.
Therefore, the pressure-sensitive elasto-plastic model proposed Melro et al. [3] is endorsed with a strain-gradient plasticity approach, where a non-local plastic variable is included. An implicit-gradient-plasticity formulation is considered, and its implementation in the commercial finite element software Abaqus will be detailed in the present contribution. The impact of this modification in the matrix model will be demonstrated by means of parametric studies, where the new material parameters related to the length-scale are varied. Another important aspect that will be addressed lies on the calibration of such constitutive parameters. Moreover, the real deformation of a composite microstructure, obtained experimentally through nano-scale digital image correlation [2], will be the reference for a careful validation and verification of the novel model.
This project has received funding from the European Union’s Horizon Europe research and innovation programme under grant agreement No. 101056682.
Concerning the matrix models, recent research [1,2] revealed that classical elasto-plastic constitutive models are not able to capture the strain fields accurately, especially in regions of the epoxy matrix confined between carbon fibres. In this case, the strain localisation is overpredicted by such classical models, which cannot cope with the large strain gradients observed. Alternatively, strain gradient plasticity models have the potential to better predict the deformation fields in these critical microscopic regions. Consequently, the macroscopic strength that is often underestimated with classical models [1] for the matrix can be captured more accurately, resulting in more reliable multi-scale models.
Therefore, the pressure-sensitive elasto-plastic model proposed Melro et al. [3] is endorsed with a strain-gradient plasticity approach, where a non-local plastic variable is included. An implicit-gradient-plasticity formulation is considered, and its implementation in the commercial finite element software Abaqus will be detailed in the present contribution. The impact of this modification in the matrix model will be demonstrated by means of parametric studies, where the new material parameters related to the length-scale are varied. Another important aspect that will be addressed lies on the calibration of such constitutive parameters. Moreover, the real deformation of a composite microstructure, obtained experimentally through nano-scale digital image correlation [2], will be the reference for a careful validation and verification of the novel model.
This project has received funding from the European Union’s Horizon Europe research and innovation programme under grant agreement No. 101056682.