A. James Clark School of Engineering
Permanent URI for this communityhttp://hdl.handle.net/1903/1654
The collections in this community comprise faculty research works, as well as graduate theses and dissertations.
Browse
33 results
Search Results
Item OPTIMAL PROBING OF BATTERY CYCLES FOR MACHINE LEARNING-BASED MODEL DEVELOPMENT(2024) Nozarijouybari, Zahra; Fathy, Hosam HF; Mechanical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)This dissertation examines the problems of optimizing the selection of the datasets and experiments used for parameterizing machine learning-based electrochemical battery models. The key idea is that data selection, or “probing” can empower such models to achieve greater fidelity levels. The dissertation is motivated by the potential of battery models to enable theprediction and optimization of battery performance and control strategies. The literature presents multiple battery modeling approaches, including equivalent circuit, physics-based, and machine learning models. Machine learning is particularly attractive in the battery systems domain, thanks to its flexibility and ability to model battery performance and aging dynamics. Moreover, there is a growing interest in the literature in hybrid models that combine the benefits of machine learning with either the simplicity of equivalent circuit models or the predictiveness of physics-based models or both. The focus of this dissertation is on both hybrid and purely data-driven battery models. Moreover, the overarching question guiding the dissertation is: how does the selection of the datasets and experiments used for parameterizing these models affect their fidelity and parameter identifiability? Parameter identifiability is a fundamental concept from information theory that refers to the degree to which one can accurately estimate a given model’s parameters from input-output data. There is substantial existing research in the literature on battery parameter identifiability. An important lesson from this literature is that the design of a battery experiment can affect parameter identifiability significantly. Hence, test trajectory optimization has the potential to substantially improve model parameter identifiability. The literature explores this lesson for equivalent circuit and physics-based battery models. However, there is a noticeable gap in the literature regarding identifiability analysis and optimization for both machine learning-based and hybrid battery models. To address the above gap, the dissertation makes four novel contributions to the literature. The first contribution is an extensive survey of the machine learning-based battery modeling literature, highlighting the critical need for information-rich and clean datasets for parameterizing data-driven battery models. The second contribution is a K-means clustering-based algorithm for detecting outlier patterns in experimental battery cycling data. This algorithm is used for pre-cleaning the experimental cycling datasets for laboratory-fabricated lithium-sulfur (Li-S) batteries, thereby enabling the higher-fidelity fitting of a neural network model to these datasets. The third contribution is a novel algorithm for optimizing the cycling of a lithium iron phosphate (LFP) to maximize the parameter identifiability of a hybrid model of this battery. This algorithm succeeds in improving the resulting model’s Fisher identifiability significantly in simulation. The final contribution focuses on the application of such test trajectory optimization to the experimental cycling of commercial LFP cells. This work shows that test trajectory optimization is s effective not just at improving parameter identifiability, but also at probing and uncovering higher-order battery dynamics not incorporated in the initial baseline model. Collectively, all four of these contributions show the degree to which the selection of battery cycling datasets and experiments for richness and cleanness can enable higher-fidelity data-driven and hybrid modeling, for multiple battery chemistries.Item ASSESSING THE IMPACT OF ELECTROCHEMICAL-MECHANICAL COUPLING ON CURRENT DISTRIBUTION AND DENDRITE PREVENTION IN SOLID-STATE ALKALI METAL BATTERIES(2023) Carmona, Eric Alvaro; Albertus, Paul; Chemical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)The relationship between mechanical stress states and interfacial electrochemical thermodynamics of Li metal/Li6.5La3Zr1.5Ta0.5O12 and Na metal/Na-β”-Al2O3 systems are examined in two experimental configurations with an applied uniaxial load; the solid electrolytes were pellets and the metal electrodes high-aspect-ratio electrodes. Our experimental results demonstrate that (1) the change in equilibrium potential at the metal/electrolyte interface, when stress is applied to the metal electrode, is linearly proportional to the molar volume of the metal electrode, and (2) the mechanical stress in the electrolyte has negligible effect on the equilibrium potential for an experimental setup in which the electrolyte is stressed and the electrode is left unstressed. Solid mechanics modeling of a metal electrode on a solid electrolyte pellet indicates that pressure and normal stress are within ~0.5 MPa of each other for the high aspect ratio (~1:100 thickness:diameter in our study) Li metal electrodes under loads that exceed yield conditions. To assess the effect of electrochemical-mechanical coupling on current distributions at Li/single-ion conducting solid ceramic electrolyte interfaces containing a parameterized interfacial geometric asperity, we develop a coupled electrochemical-mechanical model and carefully distinguish between the thermodynamic and kinetic effects of interfacial mechanics on the current distribution. We find that with an elastic-perfectly plastic model for Li metal, and experimentally relevant mechanical initial and boundary conditions, the stress variations along the interface for experimentally relevant stack pressures and interfacial geometries are small (e.g., <1 MPa), resulting in a small or negligible influence of the interfacial mechanical state on the interfacial current distribution for both plating and stripping. However, we find that the current distribution is sensitive to interface geometry, with sharper (i.e., smaller tip radius of curvature) asperities experiencing greater current focusing. In addition, the effect on the current distribution of an identically sized lithium peak vs. valley geometry is not the same. These interfacial geometry effects may lead to void formation on both stripping and plating and at both Li peaks and valleys. This work advances the quantitative understanding of alkali metal dendrite formation within incipient cracks and their subsequent growth, and pore formation upon stripping, both situations where properly accounting for the impact of mechanical state on the equilibrium potential can be of critical importance for calculating the current distribution. The presence of high-curvature interface geometry asperities provides an additional perspective on the superior cycling performance of flat, film-based separators (e.g., sputtered LiPON) versus particle-based separators (e.g., polycrystalline LLZO) in some conditions.Item DETERMINING ELONGATION AT BREAK OF CABLE INSULATIONS USING CONDITION MONITORING PARAMETERS(2022) Gharazi, Salimeh; Al-Sheikhly, Mohamad; Chemical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Many United States nuclear power plants are seeking to renew life licenses to extend the operational life of the plant to an additional 20 or 40 years. Degradation of insulation and jacket of cables, which are originally designed for 40 years in the second round of operation, is a critical issue which can impair the safe and reliable function of cables and ultimately the plant. The main criterion for assessing the end of life of these insulations is defined when the elongation at break reaches 50% of its original value. However, measuring the elongation at break is done by tensile tests, which are destructive and need large samples; the feasibility of these tests is significantly limited on installed cables at nuclear power plants. A new model was developed to relate the changes in the activation energy corresponding to EAB in terms of the changes in the activation energies corresponding to non-destructive condition monitoring, NDE-CM, parameters. The coefficients of the model are obtained by normalizing the calculated activation energy of each CM parameter’s changes with the activation energy of EAB changes. Therefore, it is possible to estimate EAB values, in the present developed equations, from the substitution of activation energy corresponding to EAB changes with the correlated activation energy of the non-destructive condition monitoring parameters. Cable Polymer Aging database, C-PAD, which is provided by Electric Power Research Institute, and supported by the U.S. Department of Energy, along with experimental results done in the University of Maryland, UMD, laboratory was used as the database. While taking advantage of C-PAD database which contains condition monitoring parameters of insulation cables such as Elongation at break, Modulus and Density provided by many U.S. and international research institutes, extensive aging experimental results on two cables, each with two grades provided us with not only a database but also a better understanding of the aging mechanism. The published experimental results of cable insulations are used to validate the model. A good fit between the experimental and modeled results confirms the validity of the model.Item ENERGY CONSUMPTION REDUCTION OF COMMERCIAL BUILDINGS THROUGH THE IMPLEMENTATION OF VIRTUAL AND EXPERIMENTAL ENERGY AUDIT ANALYSIS(2022) Bae, Ji Han; Ohadi, Michael; Mechanical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)According to the U.S. Energy Information Administration (EIA), about 38 quads of the total U.S. energy consumption was consumed by residential and commercial buildings in 2017, which is about 39% of the total 2017 annual U.S. energy consumption (EIA, 2018). Additionally, the building sector is responsible for about 75% of the total U.S. electricity consumption as well as for about 70% of the projected growth in the U.S. electricity demand through 2040. It is clear that the potential for energy savings and greenhouse gas emissions reduction in existing buildings today remain largely untapped and that there is still much left to explore in respect to determining the best protocols for reducing building energy consumption on a national and even a global scale. The present work investigates the effectiveness of coupling an initial virtual energy audit screening with the conventional, hands-on, energy audit processes to more quickly and less costly obtain the potential energy savings for high energy consumption buildings. The virtual screening tool takes advantage of a customized cloud-based energy efficiency management software and the readily available building energy consumption data to identify the buildings that have the highest energy savings potential and should be given priority for performing onsite walkthroughs, detailed energy audits, and the subsequent implementation of the identified energy conservation measures (ECMs). By applying the proposed procedure to a group of buildings, the results of this study demonstrated that a combination of the software-based screening tools and a detailed experimental/onsite energy audit as necessary can effectively take advantage of the potential energy consumption and carbon footprint reduction in existing buildings today and that the low-cost/no-cost energy conservation measures alone can oftentimes result in significant savings as documented in this thesis. However, selection of the appropriate software was deemed critically important, as certain software limitations were observed to hinder the obtainment of some energy savings opportunities.Item Trajectory Optimization of a Tethered Underwater Kite(2021) Alvarez Tiburcio, Miguel; Fathy, Hosam; Mechanical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)This dissertation addresses the challenge of optimizing the motion trajectory of a tethered marine hydrokinetic energy harvesting kite in order to maximize its average electric power output. The dissertation focuses specifically on the “pumping” kite configuration, where the kite is periodically reeled out from a floating base station at high tension, then reeled in at low tension. This work is motivated by the significant potential for sustainable electricity generation from marine currents such as the Gulf Stream. Tethered systems can increase their energy harvesting potential significantly through cross-current motion. Such motion increases apparent flow speed, which is valuable because the instantaneous maximum power that can be harvested is proportional to the cube of this apparent speed. This makes it possible for tethered systems to achieve potentially very attractive power densities and levelized costs of electricity compared to stationary turbines. However, this also necessitates the use of trajectory optimization and active control in order to eke out the maximum energy harvesting capabilities of these systems. The problem of optimizing the trajectories of these kites is highly non-linear and thus challenging to solve. In this dissertation we make key simplifications to both the modeling and the structure of the optimal solution which allows us to learn valuable insights in the nature of the power maximizing trajectory. We first do this analysis to maximize the average mechanical power of the kite, then we expand it to take into account system losses. Finally, we design and fabricate an experimental setup to both parametrize our model and validate our trajectories. In summary, the goal of this research is to furnish model-based algorithms for the online optimal flight control of a tethered marine hydrokinetic system. The intellectual merit of this work stems from the degree to which it will tackle the difficulty of solving this co-optimization problem taking into account overall system efficiency and the full range of possible system motion trajectories. From a broader societal perspective, this work represents a step towards experimentally validating the potential of pumped underwater kite systems to serve as renewable energy harvesters in promising environments such as the Gulf Stream.Item CHARACTERIZATION AND ANALYSIS OF FLUIDIC ARTIFICIAL MUSCLES(2021) Chambers, Jonathan Michael; Wereley, Norman M; Aerospace Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Fluidic artificial muscles (FAMs) are a form of soft actuator that have been applied to an expanding number of applications, due to their unique characteristics such as low weight, simple construction, inherent compliance, and high specific force and specific work capabilities. With energy sourced from a pressurized fluid, contractile FAMs provide a uniaxial contractile force, while their morphing geometry allows them to contract in length. In a design environment where actuators have tight spatial requirements and must provide precise force and position control, it is becoming more important than ever to have accurate mathematical representations of FAM actuation behavior and geometric characteristics to ensure their successful implementation. However, geometric models and force analyses for FAMs are still relatively crude. Geometric models of FAMs assume a cylindrical geometry, the accuracy of which is suspect because there are no documented methods for effectively measuring FAM shape. Actuation force analyses are also relatively inaccurate unless they are adjusted to fit to experimental response data. Research has continually pursued methods of improving the predictive performance of these analyses by investigating the complex working mechanisms of FAMs. This research improves these analyses by first, making improvements to the experimental characterization of a FAM's actuation response, and then using the more comprehensive data results to test long-held modeling assumptions. A quantitative method of measuring FAM geometry is developed that provides 0.004 in/pixel resolution measurements throughout a characterization test. These measurements are then used to test common assumptions that serve as sources of uncertainty: the cylindrical approximation of FAM geometry, and assumption that the FAM's braid is inelastic. Once these sources of modeling error are removed, the model's performance is then tested for potential improvements. Results from this research showed that the cylindrical approximation of the FAM's geometry resulted in overestimations of the FAM's average diameter by 4.7%, and underestimations of the FAM's force by as much as 37%. The inelastic braid assumption resulted in a maximum 4% underestimation of average diameter and a subsequent 5% overestimation in force, while the use of softer braid materials was found to have the potential for much larger effects (30% underestimation in diameter, 70% overestimation in force). With subsequent adjustments made to the force model, the model was able to achieve a fit with a mean error of only 2.8 lbf (0.3% of maximum force). This research demonstrates improvements to the characterization of a FAM's actuation response, and the use of this new data to improve the fidelity of existing FAM models. The demonstrated characterization methods can be used to clearly define a FAM's geometry to aid in the effective design and implementation of a FAM-actuated mechanism, or to serve as a foundation for further investigation into the working mechanisms and development of FAMs.Item EVALUATION OF IMPACT OF NOVEL BARRIER COATINGS ON FLAMMABILITY OF A STRUCTURAL AEROSPACE COMPOSITE THROUGH EXPERIMENTS AND MODELING(2021) Crofton, Lucas; Stoliarov, Stanislav; Fire Protection Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Composites have become a integral part of the structure of airplanes, and their use within aircraft continues to grow as composites continue to improve. While polymer composites are an improvement in many facets to traditional airspace materials, their flammability is something called into question. The work performed for this study was to create a pyrolysis model for a particular aerospace composite, IM7 graphite fiber with Cytec 5250-4 Bismaleimide matrix (BMI), and three innovative composite barrier coatings that could be applied to the BMI to potentially improve its performance in fire scenarios. The composites were all tested individually, in a series of milligram-scale tests, and the test results were inversely analyzed to determine stoichiometry, chemical kinetics, and thermodynamics of their thermal decomposition and combustion. Gram-scale experiments using the Controlled Atmosphere Pyrolysis Apparatus II (CAPA II) were performed on the BMI by itself and then again with one of each of the composite barrier coatings applied in a defined thickness. This data were inversely analyzed to define the thermal conductivity of the sample and resolve it’s emissivity. It was found after fully defining a pyrolysis model for each composite material that the composite barrier coatings did not provide any benefit to the base composite BMI, and only added more fuel load which in turn contributed to a increase in heat release rate when computational simulations were run to mimic a airplane fuel fire.Item Modeling Syndromic Surveillance and Outbreaks in Subpopulations(2020) Pettie, Christa; Herrmann, Jeffrey; Reliability Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)This research is motivated by the need to assist resource limited communities by enhancing the use of syndromic surveillance (SyS) systems and data. Public health agencies and academic researchers have developed and implemented SyS systems as a pattern recognition tool to detect a potential disease outbreak using pre-diagnostic data. SyS systems collect data from multiple types of sources: absenteeism records, over the counter medicine sales, chief complaints, web queries, and more. It could be expensive, however, to gather data from every available source; subsequently, gathering information about only some subpopulations may be a desirable option. This raises questions about the differences between subpopulation behavior and which subpopulations’ data would give the earliest, most accurate warning of a disease outbreak. To investigate the feasibility of using subpopulation data, this research will gather and organize SyS data by subpopulation (separated by population characteristics such as age or location) and identify how well the SyS data correlates to the real world disease progression. This research will study SyS how reports of Influenza-like-illness (ILI) in subpopulations represent the disease behavior. The first step of the research process is to understand how SyS is used in environments with varying levels of resources and what gaps are present in SyS modeling techniques. Various modeling techniques and applications are assessed, specifically the Susceptible Infected Recovered “SIR” model and associated modifications of that model. Through data analysis, well correlated subpopulations will be identified and compared to actual disease behavior and SyS data sets. A model referred to as ModSySIR will be presented that uses real world community data ideal for ease of use and implementation in a resource limited community. The highest level research objective is to provide a potential data analysis method and modeling approach to inform decision making for health departments using SyS systems that rely on fewer resources.Item A Control-Theoretic Model of Hemodynamic Responses to Blood Volume Perturbation(2018) Lo, Alex Kai-Yuan; Hahn, Jin-Oh; Mechanical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)This thesis presents a mathematical model to reproduce hemodynamic responses of different endpoints to the blood volume perturbation in circulation system. The proposed model includes three sub-models, which are a control-theoretic model relating blood volume response to blood volume perturbation, a physiologic-based model relating cardiac output response to blood volume perturbation, and a control-theoretic model relating mean arterial pressure response to cardiac output perturbation. Two unique characteristics of this hemodynamic model are that the model can reproduce responses accurately even with its simplicity, and can be easily understood by control engineers because of its physiological transparency. With these two advantages, closed-loop resuscitation controller evaluation can be performed in model-based approach instead of evaluating results from animal studies, which are relatively costly and time-consuming. In this thesis, the hemodynamic model was examined and evaluated by using experimental dataset collected from 11 animals. The results of system identification analysis, in-silico evaluation and parametric sensitivity analysis showed that the hemodynamic model may faithfully serve as a evaluation basis for the closed-loop resuscitation controllers.Item A comprehensive study of multiplicative attribute graph model(2016) Qu, Sikai; Makowski, Armand; Electrical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Graphs are powerful tools to describe social, technological and biological networks, with nodes representing agents (people, websites, gene, etc.) and edges (or links) representing relations (or interactions) between agents. Examples of real-world networks include social networks, the World Wide Web, collaboration networks, protein networks, etc. Researchers often model these networks as random graphs. In this dissertation, we study a recently introduced social network model, named the Multiplicative Attribute Graph model (MAG), which takes into account the randomness of nodal attributes in the process of link formation (i.e., the probability of a link existing between two nodes depends on their attributes). Kim and Lesckovec, who defined the model, have claimed that this model exhibit some of the properties a real world social network is expected to have. Focusing on a homogeneous version of this model, we investigate the existence of zero-one laws for graph properties, e.g., the absence of isolated nodes, graph connectivity and the emergence of triangles. We obtain conditions on the parameters of the model, so that these properties occur with high or vanishingly probability as the number of nodes becomes unboundedly large. In that regime, we also investigate the property of triadic closure and the nodal degree distribution.