Date of Award
Fall 2024
Document Type
Open Access Dissertation
Department
Mechanical Engineering
First Advisor
Paul Ziehl
Abstract
Meeting the demands of emerging aeronautical sectors like urban air mobility (UAM) and cargo air vehicles (CAVs) necessitates the use of cutting-edge structures and materials. These sectors call for materials that can offer high performance, durability for shorter mission durations, and increased cyclic loading. Moreover, there is a requirement to attain higher production rates for these structures to cater to the increasing market demand. To fulfill these demands, novel manufacturing and joining techniques utilizing suitable material systems must be investigated. Thermoplastic composites have the potential to fulfill the increasing demands of the aerospace industry by providing high strength-to-weight ratios and can be produced at a faster rate. Hence, thermoplastic composites are a promising solution. In this dissertation, numerical prediction models for advanced manufacturing and joining techniques that are most suitable for thermoplastic composites are developed.
The initial phase of this study focuses on creating a thermal-optical simulation for the automated fiber placement process (AFP), integrating Eulerian finite element models with ray tracing analysis. First, a 2D finite element heat model is developed and the predicted temperature histories are validated against previous research and experiments. To address the time-consuming nature of these computational models, a numerical stabilization scheme is implemented, enabling these models to be solved in significantly shorter timeframes compared to previous studies. This approach reduces the computational efforts for the 2D models to less than 8% of what was needed in past research, making a full-scale 3D simulation feasible. A comparison is then made between the processing temperature history for a 2D and 3D case. Both the models show identical behavior along the irradiation region but deviate by about 17% at the nip point where the incoming tow meets the substrate. Additionally, while 2D models assume constant heat across the transverse direction, the 3D models indicate a 12% decrease in temperature 15 mm from the center of the tape width, with a more significant drop of 36% near the edges of the tape. Therefore, a 3D model offers an effective monitoring approach for assessing the overall temperature of the system over time, particularly for thicker tape widths and the incorporation of curved tooling.
The second part of this work focuses on the induction welding methodology and its application to assembling thermoplastic composites. Multiphysics induction heating models are initially developed for flat thermoplastic composite laminates. These process models are validated against various parameters, including frequency, excitation current, coil offset, and laminate stacking sequences. Comparisons with thermographic IR captures and thermocouple measurements demonstrate good agreement between the models and experimental results. Additionally, a dynamic moving mesh model is created using Lagrangian FEA to illustrate a full-scale dynamic weld. The results indicated a temperature drop of 5.7% when transitioning from a flat laminate to a highly curved one (Rc = 350mm) over a fixed time period. Increasing the ply count to 16 further reduced heating rates as curvature increased, with extreme curvatures resulting in a maximum heating reduction of 30%. The simulation was then expanded to assess complex double-curved aerostructures. This analysis revealed temperature contours elongated along curvilinear fiber paths, with lower peak temperatures compared to previous cases. Finally, the high-rate heating capability of induction heating was examined through the optimization of coil geometry and laminated stacking sequences. For the coil geometry, a numerical loop integrating a frequency prediction model with the induction model was implemented. This framework evaluates two categories of coils. The selected designs are then fabricated and tested. Ultimately, the framework serves as a tool for end users to define parameters critical for designing coils tailored to specific application requirements.
A genetic algorithm (GA) based framework was developed for optimizing laminate stacking sequences and integrated with the electromagnetic solver. The optimization process is applied to various benchmark cases, each with an increasing number of design variables. The objective function aimed to maximize interface heating while incorporating penalty values to minimize surface heating. The GA successfully identified optimal configurations within 25-30 iterations. The optimal solutions showed layers oriented at both 0◦ and 90◦, with skin layers predominantly at 90◦ for the specified coil configuration. This configuration resulted in interface temperatures exceeding surface values, which is desirable. While this configuration optimizes heat characteristics, it’s crucial to consider additional constraints for structural performance. A balance between structural integrity and weldability is necessary. Given the study’s focus on thermal aspects, it can be concluded that the GA provides a reasonable approximation within this context.
Rights
© 2025, Harikrishnan Mohan
Recommended Citation
Mohan, H.(2024). Towards High-Rate Manufacture and Assembly of Thermoplastic Composites via Numerical Modeling. (Doctoral dissertation). Retrieved from https://scholarcommons.sc.edu/etd/8137