In applications spanning across aerospace, marine, automotive, energy, and space travel domains, composite materials have become ubiquitous because of their superior stiffness-to-weight ratios as well as corrosion and fatigue resistance. However, from a manufacturing perspective, these advanced materials have introduced new challenges that demand the development of new tools. Due to the complex anisotropic and nonlinear material properties, composite materials are more difficult to model than conventional materials such as metals and plastics. Furthermore, there exist ultra-high precision requirements in safety critical applications that are yet to be reliably met in production. Towards developing new tools addressing these challenges, this dissertation aims to (i) build high-fidelity numerical simulations of composite assembly processes, (ii) bridge these simulations to machine learning tools, and (iii) apply data-driven solutions to process control problems while identifying and overcoming their shortcomings. This is accomplished in case studies that model the fixturing, shape control, and fastening of composite fuselage components. Therein, simulation environments are created that interact with novel implementations of modified proximal policy optimization, based on a newly developed reinforcement learning algorithm. The resulting reinforcement learning agents are able to successfully address the underlying optimization problems that underpin the process and quality requirements. / Doctor of Philosophy / Within the manufacturing domain, there has been a concerted effort to transition towards Industry 4.0. To a large degree, this term refers Klaus Schwab's vision presented at the World Economic Forum in 2015, in which he outlined fundamental systemic changes that would incorporate ubiquitous computing, artificial intelligence (AI), big data, and the internet-of-things (IoT) into all aspects of productive activities within the economy. Schwab argues that rapid change will be driven by fusing these new technologies in existing and emerging applications. However, this process has only just begun and there still exist many challenges to realize the promise of Industry 4.0. One such challenge is to create computer models that are not only useful during early design stages of a product, but that are connected to its manufacturing processes, thereby guiding and informing decisions in real-time. This dissertation explores such scenarios in the context of composite structure assembly in aerospace manufacturing. It aims to link computer simulations that characterize the assembly of product components with their physical counterparts, and provides data-driven solutions to control problems that cannot typically be solved without tedious trial-and-error approaches or expert knowledge.

CubeSats, specially the 6U standard, is nowadays the tendency where many developers point towards. The upscaling size of the standard and payloads entail the increase of the satellite overall mass. Composite materials have demonstrated the ability to fulfill expectations like reducing structural masses, having been applied to different types of spacecraft, including small satellites. This Thesis is focused on designing, manufacturing, and dynamic testing of a 6U CubeSat made of carbon fiber, fiberglass, and aluminum. The main objective of this study was obtaining a mass reduction of a 6U CubeSat structure, maintaining the stiffness and strength. Considering the thermal effects of the used materials an outgassing test of the used materials was performed and the experimental results are presented. The CubeSat structure was entirely manufactured and tested at Cal Poly Aerospace Engineering Department facilities. A mechanical shock test and random vibration test were performed using a shock table and a shake table respectively. Results of both tests are presented. A correlation between the Experimental data and the Finite Element Model of the satellite was carried out. Finally, a comparison between 6U structure studied and aluminum 6U structures available in the market is presented.

A Study Of Shock Analysis Using The Finite Element Method Verified With Euler-Bernoulli Beam Theory; Mechanical Effects Due To Pulse Width Variation Of Shock Inputs; And Evaluation Of Shock Response Of A Mixed Flow Fan David Jonathan González Campos For many engineers that use finite element analysis or FEA, it is very important to know how to properly model and obtain accurate solutions for complicated loading conditions such as shock loading. Transient acceleration loads, such as shocks, are not as common as static loads. Analyzing these types of problems is less understood, which is the basis for this study. FEA solutions are verified using classical theory, as well as experimental results. The complex loading combination of shock and high speed rotation is also studied. Ansys and its graphic user interface, Workbench Version 14.5, are the programs used to solve these types of problems. Classical theory and Matlab codes, as well as experimental results, are used to verify finite element solutions for a simple structure, such as a cantilevered beam. The discrepancy of these FEA results is found to be 2.3%. The Full Method and the Mode Superposition Method in Ansys are found to be great solution tools for shock loading conditions, including complex acceleration and force conditions. The Full Method requires less pre-processing but solutions could take days, as opposed to hours, to complete in comparison with the Mode Superposition Method, depending on the 3D Model. The Mode Superposition Method requires more time and input by the user but solves relatively quickly. Furthermore, a new representation of critical pulse width of the shock inputs is presented. Experimental and finite element analyses of a complete mixed flow fan undergoing ballistic shock is also completed; deformation results due to shock loading, combined with rotation and aerodynamic loading, account for 32.3% of the total deformation seen from experimental testing. Solution methods incorporated in Ansys, and validation of FEA results using theory, have great potential implications as powerful tools for engineering students and practicing engineers.

Global awareness and preservation have spurred increasing interest in utilizing environmentally friendly materials for high-performance structural applications. Biocomposites pose an appealing solution to this issue and are characterized by their sustainable lifecycles, biodegradable qualities, light weight, remarkable strength, and exceptional stiffness. Many of these structural qualities are found in applications that exhibit flexural loading conditions, and this study focuses on improving the bending performance of engineered biocomposite structures. The current application of biocomposites is increasing rapidly, so this expanding research explores other natural constituent materials for biocomposite structures under flexural loading. The renewable material investigated in this study was experimentally and numerically validated by optimizing the mechanical characteristics of bamboo fibers in biocomposite structures under flexural loading conditions through various thermal and organic chemical treatment methods. Therefore, bending performance of a biocomposite truss and I-beam are analyzed to demonstrate the benefits of utilizing optimally treated bamboos in their design. To accomplish this goal, the first task consisted of treating bamboos by thermal and chemical means to determine the resulting effects on the compressive and tensile mechanical properties through experimental testing. Results indicated a significant improvement in strength, stiffness, and weight reduction. An extensive analysis determined the optimal treatment method that was utilized for flexural loading conditions. The second task entailed studying the flexural behavior of the optimally treated bamboo in two geometric configurations, a hollow cylinder and veneer strip, to determine the resultant properties for the truss and I-beam structure. The effect of node location on flexural performance was also studied to establish design guidelines for the applied structures. Bending tests indicated that node location affects the strength and stiffness of the hollow cylindrical configuration but has minimal effects on the veneer strip. Observations discovered by this study were employed into the designs of the applied structures that yielded excellent mechanical performance through flexural testing. The final task required conducting a finite element analysis in Abaqus/CAE on the performance of each structural application to validate experimental results. A conclusive analysis revealed good agreement between the numerical method and experimental result.

A study was undertaken to create a finite element model of a cooled throat converging/diverging rocket nozzle to be used as a tool in designing a cooled aerospike nozzle. Using ABAQUS, a simplified 2D axisymmetric model was created featuring only the copper throat and stainless steel support ring, which were brazed together for the experimental test firings. This analysis was a sequentially coupled thermal/mechanical model. The steady state thermal data matched closely to experimental data. The subsequent mechanical model predicted a life of over 300 cycles using the Manson-Halford fatigue life criteria. A mesh convergence study was performed to establish solution mesh independence. This model was expanded by adding the remainder of the parts of the nozzle aft of the rocket motor so as to attempt to match the transient nature of the experimental data. This model included variable hot gas side coefficients in the nozzle calculated using the Bartz coefficients and mapped onto the surface of the model using a FORTRAN subroutine. Additionally, contact resistances were accounted for between the additional parts. The results from the preliminary run suggested the need for a parameter re-evaluation for cold side gas conditions. Parametric studies were performed on contact resistance and cold side film coefficient. This data led to the final thermal contact conductance of k=0.005 BTU/s•in.•°R for contact between metals, k=0.001 BTU/s•in.•°R for contact between graphite and metal, and h=0.03235 BTU/s2•in.•°R for the cold side film coefficient. The transient curves matched closely and the results were judged acceptable. Finally, a 3D sector model was created using identical parameters as the 2D model except that a variable cold side film condition was added. Instead of modeling a symmetric one or two inlet/one or two outlet cooling channel, this modeled a one inlet/one outlet nozzle in which the coolant traveled almost the full 360° around the cooling annulus. To simplify the initial simulation, the model was cut at the barrier between inlet and outlet to form one large sector, rather than account for thermal gradients across this barrier. This simplified nozzle produced expected data, and a 3D full nozzle model was created. The cold side film coefficients were calculated from previous experimental data using a simplified 2D finite difference approach. The full nozzle model was created in the same manner as the 2D full nozzle model. A mesh convergence study was performed to establish solution mesh independence. The 3D model results matched well to experimental data, and the model was considered a useful tool for the design of an oxidizer cooled aerospike nozzle.

Firearms are present in two-thirds of United States households. As of 2003, roughly 500,000 projectile wounds occur annually in the United States. This costs an estimated 2.3 billion dollars of medical spending. The best treatment of gunshot wounds relies heavily on experience, but even with experience the unpredictable nature of ballistics can make treatment difficult. Wound ballistics studies the injury pattern of a particular bullet. Ballistic gelatin tests are used to analyze this pattern. A block of 10 or 20% ballistic gelatin is set and a bullet is fired through the block. Key characteristics of the wound profile seen in this test include: depth penetration, permanent cavity, and temporary cavity. Even with ballistic gelatin tests, there is still confusion and many unknowns throughout wound ballistic literature. Finite element analysis (FEA) can be used to reproduce the wound profile of a ballistic gelatin test. A .38 lead round nose was chosen to model. The bullet was assigned as an elastic plastic material and the ballistic gelatin block was assigned as an elastic plastic and viscoelastic material. SolidWorks®, TrueGrid®, and LS-DYNA® were used to create the models. Two elastic plastic and two viscoelastic simulations were developed from these models. Elastic Plastic 2 and Viscoelastic 1 were able to reproduce a depth penetration, temporary cavity, and permanent cavity. Elastic Plastic 1 and Viscoelastic 2 were unable to reproduce the temporary cavity. These simulations provided hopeful results, but further investigation is needed for contribution to the advancement of bullet wound treatment.

Structural engineering is heavily dependent on the use of computers. When creating a building model using structural analysis software, it is required that the designer have an understanding of the system behavior and the modeling program capabilities. Some engineers in the Southern California region are taking steps towards incorporating the Dywidag ductile connector (DDC) and super hybrid systems into building practice due to the advantages found in these systems’ construction methods and seismic performance. As the DDC and super hybrid systems reach industry, the design engineer will need to model these systems using structural analysis programs. This report describes two DDC specimens that were each modeled two ways: (1) using elastic members in conjunction with nonlinear rotational hinges (lumped plasticity model), and (2) using finite elements (fiber model). The experimental pushover curve for each test specimen was compared to the corresponding analytical backbone curves. The fiber modeling focuses on providing a means to study the joint behavior as the parameters of the system change. The lumped plasticity model provides the design engineer with a means for modeling a three-dimensional DDC building in order to get acceptable global demand values. This project offers modeling suggestions for both the fiber models and the lumped plasticity models used to predict the seismic behavior of the DDC precast concrete system.

The purpose of this study was to investigate the influence of tibial stem geometry on stress shielding of the tibia for patients with a total knee replacement. Finite element analysis was used to study different tibial stem geometry types, as well as a vast array of different geometric sizes. Both a peg and stem type geometry were analyzed and compared in order to determine what type geometry causes the least amount of stress shielding. A static loading condition with a dynamic loading factor of three was used for the system and the stress responses were analyzed at regions of interest at various depths. Regions of interest include the posterior and medial regions, at depths ranging from the resurfaced tibial surface to 100 mm below the surface. It was found that the smallest stem/peg sizes produced the least amount of stress shielding, indicating that the less amount of foreign material within the tibia, the more natural the bending and stress response of the tibia. It was also concluded that for the loading conditions used in this study, peg type geometry yields a decreased amount of stress shielding when compared to stem type geometry. This is due to the fact that the peg type geometry allowed for more natural bending and a distributed loading transfer between two pegs rather than one long central stem. Further studies should be completed on other geometry types in order to understand how to best replicate the natural bending of the tibia.

A Comparison Study of Composite Laminated Plates with Holes under Tension A study was conducted to quantify the accuracy of numerical approximations to deem sufficiency in validating structural composite design, thus minimizing, or even eliminating the need for experimental test. Error values for stress and strain were compared between Finite Element Analysis (FEA) and analytical (Classical Laminated Plate Theory), and FEA and experimental tensile test for two composite plate designs under tension: a cross-ply composite plate design of [(0/90)4]s, and a quasi-isotropic layup design of [02/+45/-45/902]s, each with a single, centered hole of 1/8” diameter, and 1/4" diameter (four sets total). The intent of adding variability to the ply sequences and hole configurations was to gauge the sensitivity and confidence of the FEA results and to study whether introducing enough variability would, indeed, produce greater discrepancies between numerical and experimental results, thus necessitating a physical test. A shell element numerical approximation method through ABAQUS was used for the FEA. Mitsubishi Rayon Carbon Fiber and Composites (formerly Newport Composites) unidirectional pre-preg NCT301-2G150/108 was utilized for manufacturing—which was conducted and tested to conform to ASTM D3039/D3039M standards. A global seed size of 0.020, or a node count on the order of magnitude of 30,000 nodes per substrate, was utilized for its sub-3% error with efficiency in run-time. The average error rate for FEA strain from analytical strain at a point load of 1000lbf was 2%, while the FEA-to-experimental strains averaged an error of 4%; FEA-to-analytical and FEA-to-tensile test stress values at 1000lbf point load both averaged an error value of 6%. Suffice to say, many of these strain values were accurate up to ten-thousandths and hundred-thousandths of an in/in, and the larger stress/strain errors between FEA and test may have been attributed to the natural variables introduced from conducting a tensile test: strain gauge application methods, tolerance stacks from load cells and strain gauge readings. Despite the variables, it was determined that numerical analysis could, indeed, replace experimental testing. It was observed through this thesis that a denser, more intricate mesh design could provide a greater level of accuracy for numerical solutions, which proves the notion that if lower error rates were necessitated, continued research with a more powerful processor should be able to provide the granularity and accuracy in output that would further minimize error rates between FEA and experimental. Additionally, design margins and factors of safety would generally cover the error rates expected from numerical analysis. Future work may involve utilizing different types of pre-preg and further varied hole dimensions to better understand how the FEA correlates with analytical and tensile test results. Other load types, such as bending, may also provide insight into how these materials behave under loading, thus furthering the conversation of whether numerical approximations may one day replace testing all together.

The objective of this thesis is to determine the feasibility of a novel axial flux variable reluctance machine design. The design aims to compete with prevalent rare-earth permanent magnet machines while also implementing an innovative torque ripple minimization strategy. Given the fundamental operating principles, a selection of dimensions, materials, and excitations are prepared for the machine. Special attention is given to the rotor profile which is crucial to operation. Finite element analysis software is used to evaluate a three-dimensional model in terms of inductance and torque. The ultimate potential of the machine is discussed and recommendations for improvement are proposed.