Mechanical Engineering Commons^™

Deep Reinforcement Learning For The Design Of Structural Topologies, Nathan Brown

All Dissertations

Advances in machine learning algorithms and increased computational efficiencies have given engineers new capabilities and tools for engineering design. The presented work investigates using deep reinforcement learning (DRL), a subset of deep machine learning that teaches an agent to complete a task through accumulating experiences in an interactive environment, to design 2D structural topologies. Three unique structural topology design problems are investigated to validate DRL as a practical design automation tool to produce high-performing designs in structural topology domains.

The first design problem attempts to find a gradient-free alternative to solving the compliance minimization topology optimization problem. In the proposed …

Physics-Based Machine Learning Methods For Control And Sensing In Fish-Like Robots, Colin Rodwell

All Dissertations

Underwater robots are important for the construction and maintenance of underwater infrastructure, underwater resource extraction, and defense. However, they currently fall far behind biological swimmers such as fish in agility, efficiency, and sensing capabilities. As a result, mimicking the capabilities of biological swimmers has become an area of significant research interest. In this work, we focus specifically on improving the control and sensing capabilities of fish-like robots.

Our control work focuses on using the Chaplygin sleigh, a two-dimensional nonholonomic system which has been used to model fish-like swimming, as part of a curriculum to train a reinforcement learning agent to …

Improving Sizing Resolution Of Particle Impactors In The Nanoparticle Range, Shivuday Kala

All Dissertations

The application of particle size measurement extends across many fields: air quality measurement, pharmaceutical studies, paint and coating production, and nanoparticle formulation to name a few. Therefore, accurate measurement of nanoparticles is critical to aerosol science. While devices currently exist that can size and count nanoparticles such as electrical mobility spectrometers, dynamic light scattering devices, and small angle X-ray scattering devices, their high costs, complex operation, and lack of outdoor usability, restrict their use in practical applications. Among the devices that can size aerosols down to the nanoscale, cascade impactors stand out because of their robustness, relatively simple design, low …

A Digital Triplet For Utilizing Offline Environments To Train Condition Monitoring Systems For Rolling Element Bearings, Ethan Wescoat

All Dissertations

Manufacturing competitiveness is related to making a quality product while incurring the lowest costs. Unexpected downtime caused by equipment failure negatively impacts manufacturing competitiveness due to the ensuing defects and delays caused by the downtime. Manufacturers have adopted condition monitoring (CM) techniques to reduce unexpected downtime to augment maintenance strategies. The CM adoption has transitioned maintenance from Breakdown Maintenance (BM) to Condition-Based Maintenance (CbM) to anticipate impending failures and provide maintenance actions before equipment failure. CbM is the umbrella term for maintenance strategies that use condition monitoring techniques such as Preventive Maintenance (PM) and Predictive Maintenance (PdM). Preventive Maintenance involves …

Impacts Of Connected And Automated Vehicles On Energy And Traffic Flow: Optimal Control Design And Verification Through Field Testing, Tyler Ard

All Dissertations

This dissertation assesses eco-driving effectiveness in several key traffic scenarios that include passenger vehicle transportation in highway driving and urban driving that also includes interactions with traffic signals, as well as heavy-duty line-haul truck transportation in highway driving with significant road grade. These studies are accomplished through both traffic microsimulation that propagates individual vehicle interactions to synthesize large-scale traffic patterns that emerge from the eco-driving strategies, and through experimentation in which real prototyped connected and automated vehicles (CAVs) are utilized to directly measure energy benefits from the designed eco-driving control strategies. In particular, vehicle-in-the-loop is leveraged for the CAVs driven …

Cfrp Delamination Density Propagation Analysis By Magnetostriction Theory, Brandon Eugene Williams

All Dissertations

While Carbon Fiber Reinforced Polymers (CFRPs) have exceptional mechanical properties concerning their overall weight, their failure profile in demanding high-stress environments raises reliability concerns in structural applications. Two crucial limiting factors in CFRP reliability are low-strain material degradation and low fracture toughness. Due to CFRP’s low strain degradation characteristics, a wide variety of interlaminar damage can be sustained without any appreciable change to the physical structure itself. This damage suffered by the energy transfer from high- stress levels appears in the form of microporosity, crazes, microcracks, and delamination in the matrix material before any severe laminate damage is observed. This …

Improving Hexapod Platform Pose Accuracy - A Photogrammetry-Based Approach, Sourabh Karmakar

All Dissertations

The aim of this research is to make a newly constructed Stewart-Gough Platform-based test frame Tiger 66.1 operational by developing control software and estimating the error in its pose accuracy. The accuracy of the platform is affected by one source or multiple sources. The typical error sources are kinematic and structural, some of them originate from manufacturing imperfections, assembly deviations, elastic deformations, thermal deformations, and joint clearances which change the expected kinematic behavior of the manipulator. Also, some non-mechanical errors like transmission error, sensor accuracy, algorithm error, and truncation error in calculation contribute significantly in some cases. Using pose deviations …

Controlled Manipulation And Transport By Microswimmers In Stokes Flows, Jake Buzhardt

All Dissertations

Remotely actuated microscale swimming robots have the potential to revolutionize many aspects of biomedicine. However, for the longterm goals of this field of research to be achievable, it is necessary to develop modelling, simulation, and control strategies which effectively and efficiently account for not only the motion of individual swimmers, but also the complex interactions of such swimmers with their environment including other nearby swimmers, boundaries, other cargo and passive particles, and the fluid medium itself. The aim of this thesis is to study these problems in simulation from the perspective of controls and dynamical systems, with a particular focus …

Damage Detection With An Integrated Smart Composite Using A Magnetostriction-Based Nondestructive Evaluation Method: Integrating Machine Learning For Prediction, Christopher Nelon

All Dissertations

The development of composite materials for structural components necessitates methods for evaluating and characterizing their damage states after encountering loading conditions. Laminates fabricated from carbon fiber reinforced polymers (CFRPs) are lightweight alternatives to metallic plates; thus, their usage has increased in performance industries such as aerospace and automotive. Additive manufacturing (AM) has experienced a similar growth as composite material inclusion because of its advantages over traditional manufacturing methods. Fabrication with composite laminates and additive manufacturing, specifically fused filament fabrication (fused deposition modeling), requires material to be placed layer-by-layer. If adjacent plies/layers lose adhesion during fabrication or operational usage, the strength …

Extensional Flows Of Polymer Solutions In Planar Microchannels, Mahmud Kamal Raihan

All Dissertations

Non-Newtonian fluids such as polymer solutions often flow under microscale extensional conditions in many natural and engineering flow fields such as in microfluidic chips, porous rocks, biological membranes and filters, printheads in additive manufacturing, etc. The changing cross sectional areas of the internal flow passages therein exert additional extension on the flow along with the shearing. Numerous studies have been dedicated to understanding the extensional flows of polymer solutions over the years. However, most of these studies only focused on flexible polymers exhibiting elasticity in their macroscopic rheology, whereas rigid polymers that portray shear-thinning but often elude elasticity in the …

Investigation Of Fatigue Response With Analytical And Machine Learning Models And Hygroscopic Analysis Of Asymmetric Bistable Cfrp Composites, Shoab Ahmed Chowdhury

All Dissertations

Asymmetric bistable carbon fibre reinforced plastic (CFRP) composites enable a broad range of applications as they can sustain multiple stable configurations and have small snap-through load requirements. These unique features, coupled with their light strength-to-weight and stiffness-to-weight ratios, have made them preferred options for multifunctional systems. This study investigates the fatigue and hygroscopic response of 2-ply, [0/90] bistable CFRP laminates and proposes predictive modeling approaches for improved performance.

While previous studies widely researched and documented the fatigue of general composites in axial loading, fatigue analysis of asymmetric bistable composites in the out-of-plane snap-through direction is inadequate. This study performs fatigue …

Applications Of Large Eddy Simulations To Novel Internal Combustion Concepts, Patrick O'Donnell

All Dissertations

Computational fluid dynamics (CFD) simulations of internal combustion engines (ICEs) are becoming an increasingly popular tool in the automotive industry to either explain experimentally observed trends or perform lower cost design iterations. The convenience of commercially available CFD software and advancements made in computing hardware have been the impetus behind this growing popularity. However, obtaining accurate results using these software packages is not a trivial process and requires an in-depth understanding of the underlying numerical methodology and sub models for various physical phenomena. Specific to the ICEs, CFD simulation often entails the use of models for detailed chemistry and combustion, …

Vibration-Based Fault Diagnostics In Wind Turbine Gearboxes Using Machine Learning, Abdelrahman Amin

All Dissertations

A significantly increased production of wind energy offers a path to achieve the goals of green energy policies in the United States and other countries. However, failures in wind turbines and specifically their gearboxes are higher due to their operation in unpredictable wind conditions that result in downtime and losses. Early detection of faults in wind turbines will greatly increase their reliability and commercial feasibility. Recently, data-driven fault diagnosis techniques based on deep learning have gained significant attention due to their powerful feature learning capabilities. Nonetheless, diagnosing faults in wind turbines operating under varying conditions poses a major challenge. Signal …

Multiscale Modeling And Gaussian Process Regression For Applications In Composite Materials, Joshua Arp

All Dissertations

An ongoing challenge in advanced materials design is the development of accurate multiscale models that consider uncertainty while establishing a link between knowledge or information about constituent materials to overall composite properties. Successful models can accurately predict composite properties, reducing the high financial and labor costs associated with experimental determination and accelerating material innovation. Whereas early pioneers in micromechanics developed simplistic theoretical models to map these relationships, modern advances in computer technology have enabled detailed simulators capable of accurately predicting complex and multiscale phenomena.

This work advances domain knowledge via two means: firstly, through the development of high-fidelity, physics-based finite …

Machine Learning-Based Data And Model Driven Bayesian Uncertanity Quantification Of Inverse Problems For Suspended Non-Structural System, Zhiyuan Qin

All Dissertations

Inverse problems involve extracting the internal structure of a physical system from noisy measurement data. In many fields, the Bayesian inference is used to address the ill-conditioned nature of the inverse problem by incorporating prior information through an initial distribution. In the nonparametric Bayesian framework, surrogate models such as Gaussian Processes or Deep Neural Networks are used as flexible and effective probabilistic modeling tools to overcome the high-dimensional curse and reduce computational costs. In practical systems and computer models, uncertainties can be addressed through parameter calibration, sensitivity analysis, and uncertainty quantification, leading to improved reliability and robustness of decision and …

Classification Of Electrical Current Used In Electroplastic Forming, Tyler Grimm

All Dissertations

Electrically assisted manufacturing (EAM) is the direct application of an electric current to a workpiece during manufacturing. This advanced manufacturing process has been shown to produce anomalous effects which extend beyond the current state of modeling of thermal influences. These purported non-thermal effects have collectively been termed electroplastic effects (EPEs).

While there is a distinct difference in results between steady-state (ideal DC) testing and pulsed current testing, the very definition of these two EAM methods has not been well established. A "long" pulse may be considered DC current; a "short" pulse may produce electroplastic effects; and even "steady-state" current shapes …

Deep Reinforcement Learning And Game Theoretic Monte Carlo Decision Process For Safe And Efficient Lane Change Maneuver And Speed Management, Shahab Karimi

All Dissertations

Predicting the states of the surrounding traffic is one of the major problems in automated driving. Maneuvers such as lane change, merge, and exit management could pose challenges in the absence of intervehicular communication and can benefit from driver behavior prediction. Predicting the motion of surrounding vehicles and trajectory planning need to be computationally efficient for real-time implementation. This dissertation presents a decision process model for real-time automated lane change and speed management in highway and urban traffic. In lane change and merge maneuvers, it is important to know how neighboring vehicles will act in the imminent future. Human driver …

A Value-Based Sequential Optimization Framework For Efficient Materials Design Considering Uncertainty And Variability, Maher Alghalayini

All Dissertations

Many problems in engineering and science can be framed as decision problems in which we choose values for decision variables that lead to desired outcomes. Notable examples include maximizing lift in airplane wing design, improving the efficiency of a power plant, or identifying processing protocols resulting in structural materials with desired mechanical properties. These problems typically involve a significant degree of uncertainty about the often-complex underlying relationships between the decision variables and the outcomes. Solving such decision problems involves the use of computational models or physical experimentation to generate data to make predictions and test hypotheses. As a result, both …

Mesoscale Modeling And Machine Learning Studies Of Grain Boundary Segregation In Metallic Alloys, Malek Alkayyali

All Dissertations

Nearly all structural and functional materials are polycrystalline alloys; they are composed of differently oriented crystalline grains that are joined at internal interfaces termed grain boundaries (GBs). It is well accepted that GB dynamics play a critical role in many phenomena during materials processing or under operating environments. Of particular interest are GB migration and grain growth processes, as they influence many crystal-size dependent properties, such as mechanical strength and electrical conductivity.

In metallic alloys, GBs offer a plethora of preferential atomic sites for alloying elements to occupy. Indeed, recent experimental studies employing in-situ microscopy revealed strong GB solute segregation …

Molecular Dynamics Simulation On Molybdenum Disulfide: Thermal-Mechanical Properties And Phase Transitions Under External Loading, Mahabubur Rahman

All Dissertations

Due to their remarkable properties, transition metal dichalcogenides (TMDs) have received much scientific interest throughout the past decade. Two layers of chalcogen atoms (S, Se, Te) sandwich a layer of transition metal atoms (Mo, W, Ta) to form the three-atom thick unit cell in TMDs. The interaction between TMD "single layers" is mediated by neighboring chalcogen planes and bonded by Van der Waals forces. Due to this weak out-of-plane interaction, bulk samples can be thinned down to a single layer by exfoliation. Among the TMDs, Molybdenum Disulfide (MoS₂) shows promise in the field of electronics, optics, and sensing …