Abstract
Over the last two decades, composites have become a key material in aircraft design and construction. Thermoplastic composites are steadily gaining ground over thermosetting composites due to their high impact load resistance, weldability and recyclability. In order to maximise the benefits of thermoplastic materials and reduce production costs, it is necessary to manufacture large and uniform structural sections and to replace the mechanical fastening methods with alternative techniques of joining such as adhesive bonding, co-consolidation and thermoplastic welding. However, the above-mentioned alternative joining methods have not yet been certified for use in primary structural elements. One means of compliance, suggested by EASA, is the incorporation into the structure of crack arrest features that will decelerate the progression of interfacial damage beyond a critical size between the pre-defined inspection intervals. In this effort, an understanding of the mechanical behavior of t ...
Over the last two decades, composites have become a key material in aircraft design and construction. Thermoplastic composites are steadily gaining ground over thermosetting composites due to their high impact load resistance, weldability and recyclability. In order to maximise the benefits of thermoplastic materials and reduce production costs, it is necessary to manufacture large and uniform structural sections and to replace the mechanical fastening methods with alternative techniques of joining such as adhesive bonding, co-consolidation and thermoplastic welding. However, the above-mentioned alternative joining methods have not yet been certified for use in primary structural elements. One means of compliance, suggested by EASA, is the incorporation into the structure of crack arrest features that will decelerate the progression of interfacial damage beyond a critical size between the pre-defined inspection intervals. In this effort, an understanding of the mechanical behavior of the thermoplastic composite layers and the fracture mechanics of the thermoplastic interface is of major importance, as is the availability of reliable numerical simulation models. In the present thesis, numerical models were developed for the simulation of interfacial failure of thermoplastic co-consolidated joints subjected to quasi-static and fatigue loading conditions, which, after being validated upon experimental data, were used for the design and evaluation of joints with two different crack arrest features: Refill Friction Stir Spot Welds (RFSSW), and Induction Low Shear Friction Stir Rivets. The selection of the specific features was made after an extensive literature study which generally concerned crack arrest features in composite joints. The numerical models were based on the finite element method and were developed using the commercial code LS-DYNA. The cohesive zone method was used to simulate failure at the thermoplastic co-consolidated interface. For the study, the thermoplastic material LM-PAEK with T700 carbon fiber reinforcement was used. A major part of the thesis was conducted in the framework of the European research project “TORNADO” (Innovative Disbond Arrest Features for Long Thermoplastic Welded Joints, EC Clean Sky 2, 2021-2023). An extensive experimental campaign was designed and implemented to extract input data and validate the numerical models. Specifically, mechanical tests were conducted on Double Cantilever Beam (DCB), End-Notch Flexure (ENF), Single Lap Shear (SLS) and Crack Lap Shear (CLS) specimens, which were subjected to quasi-static and fatigue loading. The experiments were conducted on reference specimens and specimens with crack arrest features. The length of the interfacial crack during the tests was measured either by visual means or by acoustic ultrasound monitoring carried out at pre-defined loading intervals during the mechanical tests. Since most failure criteria and degradation laws for composite materials’ properties have been developed for thermosetting composites, an applicability study of the available criteria and laws for thermoplastic materials was initially carried out. Four different damage models were evaluated in terms of their ability to simulate the mechanical response, the failure evolution and the ease of application/required data. Optimal performance was achieved through the combination of Hashin-type failure criteria and degradation through a progressive damage model. In addition, this material model has the ability to simulate interlaminar failure, which is not the case for the other models. The results from the mechanical testing of SLS specimens revealed the phenomenon of fiber bridging at the failure surfaces of the thermoplastic joints. This type of failure is very common in thermoplastic interfaces. Since the available traction-separation laws in the cohesive zone method are only suitable for simulating cohesive failure, a modified tri-linear traction-separation law, resulting from the superposition of the bi-linear behaviors of the thermoplastic matrix and fibers, was developed in this thesis to simulate the occurrence of fiber bridging. Initially, the data from the mode I (DCB) and mode II (ENF) tests were used to construct the crack propagation resistance curves (R-curves) of the interface. The curves were incorporated into the numerical models using a user-defined material model developed in the LS-Dyna finite element code. An algorithm was then developed to derive the fiber bridging law directly from the simulation results, thus eliminating the need for continuous crack monitoring during the analysis. The final model predicted with significantly improved accuracy the failure of the SLS specimens. A new fatigue crack growth model based on the cohesive zone method was developed to simulate the crack evolution at the interface of co-consolidated thermoplastic joints due to mixed-mode loading conditions. The model was specifically designed to account for any loading mode-mixity, requiring input data from corresponding pure mode I and mode II loadings. The interfacial crack growth rate is continuously updated using a function of the energy release rate and the mode-mixity ratio of the loads which is calculated via a linear law. The model can be applied both for force-controlled and displacement controlled loadings, and for structures from coupon-scale to larger structural elements, such as stiffened panels. A user-defined subroutine was developed to implement the model in LS-Dyna FE software. Numerical results revealed that the model accurately predicts the fatigue crack propagation for mode I, mode II, and mixed-mode loadings in DCB and ENF, SLS and CLS specimens, respectively. After the validation of the numerical models of the fatigue crack-growth and fiber bridging, they were then used to evaluate the effectiveness and parametrically design the crack arrest features in CLS specimens and in a mono-stiffener element with initial interfacial damage subjected to quasi-static loading and fatigue. The parameters studied were the features’ diameter, the number of features and their distance from the pre-crack. The effect of both crack arrest features on the CLS specimens for quasi-static loading was negligible. However, in the case of fatigue loading both mechanisms achieved significant retardation of interfacial crack evolution. Refill Friction Stir Spot Welding performs better than Induction Low Shear Friction Stir Spot Riveting since it has been more lab-studied, thus more technologically mature and appears to perform well as a joining method as well. In the mono-stiffener element, the RFSSW feature achieved deceleration of the interfacial failure for fatigue loading, while the mechanism of ILSFSR had no effect. All the geometric parameters investigated were found to have a significant effect on the efficiency of the crack arrest features, however the most significant effect comes from their diameter. The larger the diameter the more significant the deceleration in the crack growth evolution. It is noted that similar results regarding the efficiency of the crack arrest features are also expected for thermoplastic welded interfaces due to the similar fracture mechanics behavior of the co-consolidated interface and the weld seam. In conclusion, both mechanisms show high potential for use in thermoplastic material structures, especially the RFSSW feature. In summary, in this thesis the interfacial failure of co-consolidated thermoplastic composites was experimentally studied, and numerical methodologies were developed for the analysis and design of thermoplastic joints with crack arrest features. The following key innovations emerge from the methodologies and results of the thesis: a complete experimental characterization of the co-consolidated joints was carried out for the LM-PAEK/T700 material; a new tri-linear traction-separation law for simulating fiber bridging at thermoplastic interfaces was developed, numerically implemented and experimentally validated; a new model for the simulation of mixed-mode fatigue crack growth along thermoplastic interfaces was developed, numerically implemented and experimentally validated; the efficiency of two new crack arrest features in co-consolidated thermoplastic joints in coupon-scale and mono-stiffener element level was evaluated, and a parametric study on the influence on the features’ efficiency of some key geometric parameters was performed.
show more
