Zobeiry, NavidSchoenholz, Caleb2024-09-092024-09-092024-09-092024Schoenholz_washington_0250E_26672.pdfhttps://hdl.handle.net/1773/52090Thesis (Ph.D.)--University of Washington, 2024Despite significant advancements in materials formulation and manufacturing technologies, high levels of uncertainty persist in the raw material and production of composite-intensive aircraft. One such uncertainty source is the impact of material and processing variabilities on residual stresses and deformations in composite parts, which negatively affects the assembly process of aerostructures. This research investigates these phenomena, focusing on Toray T800S/3900-2B, an aerospace-grade material used in the production of several aircraft such as the Boeing 787. The initial research phase involves a comprehensive characterization of various material properties, manufacturing phenomena, and processing variables that may significantly impact process-induced deformations (PIDs) but are surrounded by high uncertainty. Investigations include assessing the impact of release coating on tool surface properties, examining the influence of processing conditions on T800S/3900-2B, and evaluating the role of processing variabilities in tool-part interactions. Next, a novel machine learning (ML) method is developed for accelerated composites characterization, which demonstrates substantial time and cost savings compared to traditional methods. A parametric exploration into the effects of layup and cure cycle procedures on PIDs is conducted, and potential mitigation strategies are proposed. Next, an innovative methodology for efficiently predicting PIDs and analyzing composites using multi-fidelity simulation and theory-guided machine learning (TGML) is devised. Lastly, a novel process optimization approach for minimizing PIDs in composite parts without the use of any material characterization or process simulation is introduced. This research aims to provide a comprehensive framework for further exploration and potential mitigation of PIDs in aerospace composites manufacturing.application/pdfen-USCC BYAerospace Composites ManufacturingProbabilistic Machine LearningProcess-induced Deformations (PIDs)Residual StressesMaterials ScienceEngineeringMaterials science and engineeringThesis