Time-dependent, matrix-dominated failure of continuous fiber-reinforced thermoplastic composites
Due to the COVID-19 crisis measurements the PhD defence of Ozan Erartsin will take place online without the presence of an audience.
The recording of this defence will be added to the video overview of recent defences.
Ozan Erartsin is a PhD student in the research group Production Technology. His supervisors are prof.dr.ir. L.E. Govaert and prof.dr.ir. R. Akkerman from the Faculty of Engineering Technology ET.
Composite materials have become indispensable for many engineering applications thanks to their high specific strength and tailorability of properties. Thermoplastic composites, which offer additional benefits such as recyclability and suitability for mass-manufacturing, are increasingly preferred especially in the automotive industry, which strives to decrease the weight of the electric cars for fuel-efficiency. Despite their advantages, thermoplastic composites exhibit strong time-dependent behavior in the matrix-dominated transverse and off-axis loading: strength depends on the applied strain rate and they are highly prone to creep and fatigue failure. Hence, it is of utmost importance to predict their time-dependent behavior to use them effectively in engineering applications.
Accurate prediction of the time-dependent behavior requires the identification of the failure mechanisms. Failure mechanisms under creep and fatigue loading are well known for neat thermoplastics, which play a significant role in the transverse and off-axis failure of composites. Neat thermoplastics exhibit plasticity-controlled failure at high stresses and low failure times and crack growth-controlled failure at low stresses and long times-to-failure. Although crack growth controlled nature of the failure is widely mentioned for the continuous fiber-reinforced composites, plasticity-controlled failure had not yet been identified. Hence, this thesis aims to identify the role of plasticity-controlled failure in the matrix-dominated, time-dependent failure of continuous fiber-reinforced composites, enabling accurate prediction of their time-dependent behavior.
Firstly, we aim to identify the role of plasticity-controlled failure mechanism in several unidirectional (UD) material systems loaded in the transverse direction at room temperature. Based on the experience on the neat and short fiber-reinforced thermoplastics, identification of the failure mechanisms requires the comparison of the time-to-failure under creep and fatigue loading, at the equal value of maximum stress. For plasticity-controlled failure, lifetime is longer under fatigue loading, while the opposite is valid for the crack growth-controlled failure. Besides, the identicality of the failure kinetics in the tensile tests at constant-strain-rates and creep tests is an indication of the plasticity-controlled failure, which is also investigated to identify the failure mechanisms. Following these methodologies, failure mechanisms of glass/iPP, carbon/PEEK, and carbon/PEKK are identified. Plasticity is found to play a crucial role for glass/iPP, being the only mechanism observed over the whole load and timespan investigated. Unlike glass/iPP, carbon/PEEK and carbon/PEKK displayed the effects of plasticity at high stresses, while crack growth dominated the failure kinetics at low stresses.
In further studies, we aim to investigate in more detail the time-dependent behavior of the material system that shows the most evident signs of the plasticity-controlled failure, which is glass/iPP. In this aspect, the effect of elevated temperatures and off-axis angle on the time-dependent failure are investigated. Moreover, the relation between the time-dependent behavior of the neat matrix and the transversely loaded unidirectional composite is studied both experimentally and numerically with a micromechanical finite element model.
High-temperature tests are employed for transversely loaded UD glass/iPP to accelerate the failure for evaluating the failure mechanisms in long timescales and to determine the deformation processes contributing to the total time-dependent response. Similar to room temperature behavior, glass/iPP is revealed to display plasticity-controlled failure also at elevated temperatures. Moreover, similar to neat iPP, glass/iPP is shown to exhibit multi-process deformation, which is captured well by an analytical lifetime prediction method based on the Ree-Eyring approach and the concept of critical strain, in the framework of plasticity-controlled failure. Studying the relationship between the time-dependent behavior of the neat matrix and the composite, the activation energy is found to be different for the matrix and the composite. Such a trend is linked to the possible role of temperature-dependent effects such as fiber-matrix debonding and residual stresses in the time-dependent behavior of the composite.
The effect of the off-axis angle is characterized since the composite laminates used in practice have layers oriented at different angles for tailored properties. The effects of orientation and strain rate on the tensile strength are found to be factorizable, which is used for accelerated characterization of the time-dependent behavior. Employing the factorizability, the analytical lifetime prediction method is extended to take into account the effect of orientation. Moreover, an anisotropic, lamina-level constitutive model, which was developed for the short fiber-reinforced composites, is applied to predict the stress-strain response using an approach based on a single deformation process. The model successfully described the stress-strain response for a large range of off-axis angles. Further research is needed for certain angles that involve a distinct failure behavior and for taking into account the effect of multiple deformation processes on the time-dependent failure.
Lastly, the findings of the thesis are put in a broader perspective: the implication of several parameters related to loading, composite structure and the state of the matrix on the plasticity-controlled time-dependent behavior is discussed. A methodology for the prediction of the time-dependent behavior is presented.
