Naslov
Multiscale modelling of additively manufactured composite material behaviour

Autori
Gljušćić, Matej

Vrsta, podvrsta i kategorija rada
Ocjenski radovi, doktorska disertacija

Fakultet
Tehnički fakultet

Mjesto
Rijeka

Datum
15.12

Godina
2022

Stranica
148

Mentor
Franulović, Marina ; Lanc, Domagoj

Ključne riječi
additive manufacturing, fiber-reinforced composites, material behaviour model- ling, material damage modelling, experimental assessment

Sažetak
Additive manufacturing is a process of joining material in successive layers in order to make objects from three-dimensional model data. Due to its versatility, the technology has initially been used only for rapid prototyping in research and development. However, recent improvements by introducing various types of reinforcing constituents in the fused deposition process expanded its application to geometrically complex, lightweight, and durable composites in engineering practice. However, the main drawbacks of this approach are weakened intralaminar and interlaminar contact zones due to successive material deposition followed by reinforcement distribution irregularities and voids. Consequently, these deficiencies lead to the overall reduction of load- bearing capabilities in additively manufactured composites in comparison to their traditionally manufactured counterparts. Therefore, to model this behaviour, a systematic analysis of the state-of-the-art in additive manufacturing and composite material mechanics was conducted in this thesis, based on which a multiscale modelling approach was proposed. It starts with micro- structural analysis based on homogenized representative volume elements designed according to microscopic inspections, followed by unidirectional and shear lamina properties identification through standardized destructive testing, while concluding with damage model calibration and validation on multidirectionally reinforced laminates. The microstructural investigation has been conducted on three distinctive cases of additively manufactured composites reinforced with continuous carbon, glass, and aramid fibers, respectively. The specimens have been produced using a Markforged-X7 3D printer utilizing a fused filament fabrication approach, while the microstructures have been inspected using SEM in cross-sections longitudinal and perpendicular to the fiber direction. The SEM images have been examined using machine-learning algorithms while the acquired results have been statistically analysed and compared with the relevant literature. The acquired data has been adopted to generate a representative volume element in Abaqus CAE environment for each of the test cases. The homogenization has been conducted using python scripting by adopting each of the constituent’s constitutive models from the literature, while the cohesive interactions between the constituents have been calibrated according to the experimentally acquired data, acquiring good agreement between the experimental and numerical homogenization results for longitudinal, transversal, and in-plane shear behaviour. Consequently, further analysis of multidirectinally reinforced laminates on the macroscale has been proposed. Therefore, three distinctive cases of additively manufactured carbon fiber reinforced laminates have been designed as standardized open-hole specimens, with the selection of lamina stacking sequences that enforce in-plane multiaxial stress state during uniaxial tensile loads. Subsequently, digital twins have been designed within the Abaqus CAE environment while the laminate constitutive behaviour has been described with a continuum damage model utilizing Puck failure theory implemented in the Abaqus CAE framework through the UMAT Fortran subroutine. The model has been modified for a better description of shear- influenced damage in AM composites and calibrated on a specific test case using response surface algorithms of multiparametric central composite design of experiments. The results have been validated experimentally and compared with other damage models, confirming the significant influence of shear stress on the failure of AM composites which has been ac- counted for by the proposed modifications. The performed research provides a significant scientific contribution to the field of additive manufacturing and material science, proposing a multiscale protocol for the identification of material properties in heterogenic and anisotropic composites produced by additive manufacturing, and the modification of failure criteria for a more accurate damage prediction and safer application of additive manufacturing technologies in engineering practice.

Izvorni jezik
Engleski

Znanstvena područja
Strojarstvo, Temeljne tehničke znanosti

POVEZANOST RADA