12023 research outputs found
Sort by
Design, Fabrication, and Mechanical Analysis of Intertwined and Frictional Micro-Architected Materials
Get PDFNatural biomaterials, e.g., shells, bone, and wood, are typically comprised of hard and soft constituent materials that are hierarchically ordered to achieve mechanical resilience, light weight, and multifunctionality. Advanced fabrication techniques have enabled the creation of precisely architected materials with exceptional mechanical properties unattainable by their constituent materials, yet they are often designed with fully interconnected structural members whose junctions are detrimental to their performance because they serve as stress concentrations for damage accumulation and lower mechanical resilience. Most studies have also focused on understanding the stretching, bending, and buckling of the structural members, while explorations toward contact interactions within structural members remain limited. We address these challenges by (i) introducing a new three-dimensional (3D) hierarchical architecture in which fibers are interwoven to construct effective beams, (ii) introducing the concept of knots into the hierarchical architecture framework, and (iii) developing a model to study the effects of structural element length scale on the energy dissipation capability of a frictional architected material.
We first explore the effective lattice response of hierarchical woven microlattices, and we demonstrate the superior ability of woven architectures to achieve high tensile and compressive strains via smooth reconfiguration of woven microfibers in the effective beams and junctions without failure events. We study how fiber topology and constituent materials influence the mechanical behaviors of hierarchical intertwined structures, and we compare our results with theory. Our study reveals that knot topology allows a new regime of deformation capable of shape-retention, leading to increased absorbed energy and failure strain compared to structures with woven topology. Agreements between experimental results and a model for long overhand knots suggest that the model can aid the optimization of the mechanical performance of microwoven materials. We then adapt classical contact mechanics and adhesion models to explore the influence of the size of structural elements in a frictional architected material on its energy dissipation capability. Our model shows that the energy dissipation capability of our frictional architected material can be significantly increased when it is scaled down from the mm-scale to the sub-micron length scale.
Our woven hierarchical design offers a pathway to make traditionally stiff and brittle materials more deformable and introduces a new building block for 3D architected materials with complex nonlinear mechanics. The unique tightening mechanism introduced by knotted topology unlocks new ways to create shape-reconfigurable, highly extensible, and extremely energy-absorbing bulk, 3D architected materials with mechanical properties that can be tuned not only by their geometries and bulk properties, but also by the surface interactions experienced by the structural elements. Lastly, our modeling work shows the potential of creating highly dissipative architected materials with shape-retention capability via carefully architected structural elements.</p
Learning-Augmented Control and Decision-Making: Theory and Applications in Smart Grids
Get PDFAchieving carbon neutrality by 2050 does not only lead to the increasing penetration of renewable energy, but also an explosive growth of smart meter data. Recently, augmenting classical methods in real-world cyber-physical systems such as smart grids with black-box AI tools, forecasts, and ML algorithms has attracted a lot of growing interest. Integrating AI techniques into smart grids, on the one hand, provides a new approach to handle the uncertainties caused by renewable resources and human behaviors, but on the other hand, creates practical issues such as reliability, stability, privacy, and scalability, etc. to the AI-integrated algorithms.
This dissertation focuses on solving problems raised in designing learning-augmented control and decision-making algorithms.
The results presented in this dissertation are three-fold. We first study a problem in linear quadratic control, where imperfect/untrusted AI predictions of system perturbations are available. We show that it is possible to design a learning-augmented algorithm with performance guarantees that is aggressive if the predictions are accurate and conservative if they are imperfect. Machine-learned black-box policies are ubiquitous for nonlinear control problems. Meanwhile, crude model information is often available for these problems from, e.g., linear approximations of nonlinear dynamics. We next study the problem of equipping a black-box control policy with model-based advice for nonlinear control on a single trajectory. We first show a general negative result that a naive convex combination of a black-box policy and a linear model-based policy can lead to instability, even if the two policies are both stabilizing. We then propose an adaptive λ-confident policy, with a coefficient λ indicating the confidence in a black-box policy, and prove its stability. With bounded nonlinearity, in addition, we show that the adaptive λ-confident policy achieves a bounded competitive ratio when a black-box policy is near-optimal. Finally, we propose an online learning approach to implement the adaptive λ-confident policy and verify its efficacy in case studies about the Cart-Pole problem and a real-world electric vehicle (EV) charging problem with data bias due to COVID-19.
Aggregators have emerged as crucial tools for the coordination of distributed, controllable loads. To be used effectively, an aggregator must be able to communicate the available flexibility of the loads they control, known as the aggregate flexibility to a system operator. However, most existing aggregate flexibility measures often are slow-timescale estimations and much less attention has been paid to real-time coordination between an aggregator and an operator. In the second part of this dissertation, we consider solving an online decision-making problem in a closed-loop system and present a design of real-time aggregate flexibility feedback, termed the maximum entropy feedback (MEF). In addition to deriving analytic properties of the MEF, combining learning and control, we show that it can be approximated using reinforcement learning and used as a penalty term in a novel control algorithm--the penalized predictive control (PPC) that enables efficient communication, fast computation, and lower costs. We illustrate the efficacy of the PPC using a dataset from an adaptive electric vehicle charging network and show that PPC outperforms classical MPC. We show that under certain regularity assumptions, the PPC is optimal. We illustrate the efficacy of the PPC using a dataset from an adaptive electric vehicle charging network and show that PPC outperforms classical model predictive control (MPC). In a theoretical perspective, a two-controller problem is formulated. A central controller chooses an action from a feasible set that is determined by time-varying and coupling constraints, which depend on all past actions and states. The central controller's goal is to minimize the cumulative cost; however, the controller has access to neither the feasible set nor the dynamics directly, which are determined by a remote local controller. Instead, the central controller receives only an aggregate summary of the feasibility information from the local controller, which does not know the system costs. We show that it is possible for an online algorithm using feasibility information to nearly match the dynamic regret of an online algorithm using perfect information whenever the feasible sets satisfy some criterion, which is satisfied by inventory and tracking constraints.
The third part of this dissertation consists of examples of learning, inference, and data analysis methods for power system identification and electric charging. We present a power system identification problem with noisy nodal measurements and efficient algorithms, based on fundamental trade-offs between the number of measurements, the complexity of the graph class, and the probability of error. Next, we specifically consider prediction and unsupervised learning tasks in EV charging. We provide basic data analysis results of a public dataset released by Caltech and develop a novel iterative clustering method for classifying time series of EV charging rates.</p
Methods for Control of Granular Material Attributes
Get PDFA granular material is a collection of discrete, solid particles. This substance is ubiquitous in nature and industry, with examples ranging from soils, jointed rocks, foodstuffs, ball bearings, powders, and even asteroids. As such, understanding granular materials is necessary for making sense of the physical world. Tremendous progress has been made in directly simulating granular materials in the previous decades, in particular via the discrete element method (DEM). Nevertheless, there remains ample opportunity for manipulating granular materials to achieve specific outcomes by leveraging the DEM. The research presented in this thesis utilizes DEM simulations to develop tools and strategies for manipulating granular material to achieve desired attributes. These attributes include the shape of individual grains, the structure of granular tunnels, and mesoscopic packing characteristics such as packing fraction and coordination number. Optimization of granular materials is considered at 3 different scales: at the single grain scale (100 grains), at the scale of granular structures such as arches (101 grains), and at the mesoscopic scale (103 grains). The first component of this thesis considers automated design of individual grain shapes that embody user-specified morphological properties via genetic algorithms. Next, excavation in granular materials is considered. It is studied how ants can so successfully manipulate granular materials to achieve stable systems by mapping the forces around real ant tunnels. Ant tunnels are simulated using a DEM which can handle arbitrary shaped grains: the Level-Set Discrete Element Method (LS-DEM). Finally, tools are developed for controlling mesoscopic attributes of granular materials as a function of grain shape. To do so, genetic algorithms and a deep generative model are combined with LS-DEM. The methodologies introduced in this thesis serve as a foundation for controlling granular material attributes. Such techniques can be leveraged to engineer granular materials, with applications ranging from swarm robotics, robotic grippers, mechanically tunable fabrics for armor, and robotic excavation
Understanding the Cosmological Evolution of Galaxies with Intensity Mapping
Get PDFThe intensity mapping (IM) technique has been proven to be a powerful addition to the toolkit for understanding the cosmology and astrophysics behind cosmic structure formation. From the nearby universe to the epochs of cosmic dawn and reionization, by mapping the large-scale structure traced by a certain intensity field, IM provides an economical and a holistic view of the formation and evolution of galaxies in the cosmological text, in a way that is highly complementary to traditional methods based on individual galaxy detections. In this thesis, I present a number of theoretical perspectives on how the IM technique, especially line intensity mapping (LIM), can help us better understand the cosmological evolution of galaxies --- all the way to the intriguing era of first galaxy formation.
In the first part of this thesis, I use the Tomographic Ionized-carbon Mapping Experiment (TIME), a pathfinder for LIM observations of the 158-micron [CII] line emission from the epoch of reionization (EoR), as an example to demonstrate the aspects of high-redshift star-forming galaxies that can be practically studied with LIM. In Chapter 2, I elaborate the science cases of TIME for the investigations of the EoR using the redshifted [CII] line as a star formation rate tracer, and the cosmic molecular gas content near cosmic noon using CO rotational lines redshifted into the same observing bandpass. The results also guide the design of future phases of TIME. In Chapter 3, I introduce and demonstrate an effective masking strategy for the cleaning of line interlopers such as CO from the [CII] data TIME will measure. Using proxies of CO emitters built from stacking analysis of deep, near-infrared selected galaxies, it provides a practical solution to the notoriously challenging line confusion problem for LIM data analysis.
The second part of this thesis focuses on the concept of multi-tracer LIM, namely the synergies among LIM observations of multiple distinct tracers. Forward modeling and inference tools based on semi-analytic models and semi-numerical simulations are developed to explore and showcase the scientific potential of multi-tracer LIM. In Chapter 4, I describe a self-consistent, semi-analytic framework for modeling a variety of LIM signals from the multi-phase interstellar medium (ISM) of galaxies, and use it to illustrate the potential application of LIM to shed light on mean ISM properties of galaxies. In Chapters 5 and 6, I present a new semi-numerical simulation called LIMFAST that is developed for efficiently and self-consistently simulating a plethora of IM signals in the high-redshift universe. The LIMFAST code is particularly tailored for revealing the connection between the EoR and the first galaxy formation with multiple cosmological probes.
Finally, in the last part of thesis, I show two example case studies where the IM technique is applied to investigate the astrophysics of star formation in galaxies. In Chapter 7, I present an updated analysis of the contributions from star-forming galaxies at z≳5 to the observed cosmic near-infrared background. Imprints that reveal the formation histories of first stars, including the prospects for detecting them with the forthcoming space missions, are also studied. In Chapter 8, I describe a novel way to constrain the global star formation law of galaxies using LIM measurements of the baryonic acoustic oscillations.
As an emerging technique in observational cosmology, IM is no doubt still in its early days, promising exciting scientific returns while facing various practical challenges. Studies described in this thesis represent only a tiny fraction of the theoretical efforts from the community, but they pave the way for more follow-up investigations that will eventually turn IM into a truly rewarding endeavor.</p
Coupled Oscillators with Generalized Dissipation
Get PDFWe theoretically study a mechanical system of two coupled harmonic oscillators with arbitrary damping kernels. We consider cases where the damping is of a Markovian nature as well as the case of generalized non-Markovian damping. Previous studies had been performed for specific and equal nonMarkovian damping kernels, namely an exponential and a power law kernel. We generalize this study for arbitrary and unequal damping kernels, finding that certain properties, namely the existence of a phase transition remain unchanged. This remains true for all non-zero values of the coupling strength between the modes. The study opens up new avenues for the experimental study of systems with hitherto unexplored system-bath interactions
Towards Integrated Molecular Machines: Structural, Mechanical, and Computational Motifs
Get PDFThe programmability of DNA has made it well-suited for building molecular machines, performing nanoscale self-assembly, and computing via biochemical circuits. In the last few decades, great strides have been made in characterizing the interactions between DNA molecules such that they can be predicted and engineered.
The development of frameworks for those interactions has enabled the construction of more complex molecular systems that can execute specified programs. Such programs have included mechanical tasks, like walking and sorting cargo; assembly and reconfiguration of 2D and 3D shapes; and computation, like Boolean logic and pattern recognition.
However, the continuing development of more complex molecular programs relies upon expanding the modules available for molecular systems to use to execute them. Expanded functionality of mechanical, structural, and computation modules are required in order to build compound systems that can interact with the physical world, reconfigure, and analyze signals in a variety of interesting ways. In this dissertation, we will discuss our contributions to this effort, which include exploring a motif for molecular robotic behavior, characterizing tile-tile interactions, and developing new capabilities for bimolecular circuits.
Within the framework of a maze-solving molecular robot, we aim to implement walking behavior on DNA origami that introduces a surface modification via a four-way strand displacement reaction. Surprisingly, our experiments suggest that the walking behavior is at least two orders of magnitude slower than expected. To understand why, we quantitatively explore to what extent the speed and completion level of the robot can be modulated by design considerations such as toehold lengths, track redundancy, and strand purity. Another factor affecting the reaction rate is the number of tethering points, and we demonstrate an order of magnitude speed up in the four-way strand displacement reaction when we remove one tethering point. The characterization of a surface-modifying four-way strand displacement reaction is a useful tool for the continued development of molecular robots with more complex functionality.
Free-floating DNA origami tiles, called invaders here, can swap out DNA origami tiles within larger assemblies via a technique called tile displacement, which has previously been demonstrated using single tile and dimer invaders with 4- and 9-tile arrays. We introduce initial structures and invading assemblies with more complex shapes. We explore the robustness of this reaction by testing a variety of edge configurations and comparing their reaction rates. We demonstrate tunable growth of one of the invaders, which can grow into polymers of arbitrary length or close into 3D structures. By a tile displacement reaction, we reconfigure the 3D structures into 2D. The invaders with complex shapes are able to reconfigure the original tile assembly at rates comparable to simpler tile displacement reactions, and two reconfiguration events can take place sequentially or simultaneously.
Finally, we build two new modules for use with biochemical circuits. The first, a loser-take-all circuit, yields binary outputs indicating which analog signal is the smallest among all inputs. We implement a signal reversal function that converts the smallest input to the largest output, which can then be composed with a previously developed winner-take-all function to achieve loser-take-all. By making concentration adjustments, we can mitigate biases in the circuit that are a result of sequence-dependent different in reaction rates. We experimentally demonstrate a three-input loser-take-all circuit with nine input combinations. With further development, this circuit could be used to implement the activation function in neural networks that perform pattern classification according to which memory an input pattern is least similar to.
The second circuit processes information using temporary memory. We design and implement a circuit that outputs distinct logic decisions based on relative timing information of a pair inputs and their logic values. We show that we can mitigate crosstalk in the circuit by utilizing mismatches and adjusting toehold lengths. The circuit is able to display clear ON-OFF separation at time intervals as short as one minute between the two inputs arriving.</p
Uncertainty and Decentralization: Two Themes in an Energy Transformation
Get PDFOver the last two decades, the rapidly decreasing units costs of solar, wind, and energy storage technologies have launched a fundamental transformation in how electric power is produced, distributed, and consumed. Proliferation of these technologies has effected a shift towards a more decentralized, flexible, and sustainable energy system that can meet the growing demand for energy while reducing greenhouse gas emissions from fossil fuels. The work in this thesis studies two principal themes in this transformation: uncertainty and decentralization.
Uncertainty is a key challenge in the modern grid resulting from the weather dependence of variable renewables and volatile loads like electric vehicles distributed throughout the grid. Electricity markets, whose function is to regulate the precise balance of supply and demand across the system, face a pressing need for dispatch mechanisms that account for uncertainty while providing participation incentives for generators and loads. We introduce a framework for multi-stage market dispatch and pricing under a general description of forecast uncertainty that enables system operators to explicitly incorporate uncertainty into market-clearing prices. In related work, we study mechanisms that guarantee feasibility of multi-interval dispatch under robust uncertainty and provide participation incentives for shiftable demand response in forward multi-interval markets.
The trend towards a more decentralized energy system stems from the inherent modularity of distributed energy resources (DERs), such as solar and storage, as well as the persistent growth in end-use loads. This evolution presents significant challenges to system operators who typically lack the tools and processes for managing a complex, distributed power system. To fill this gap, we introduce and implement a Microgrid Operating System (OS), a software platform for monitoring, modeling, and optimizing microgrids and distribution systems. The Microgrid OS is a central layer that links DER hardware, such as batteries, solar, and flexible loads, to energy applications like cost minimization, emissions reduction, and wholesale market participation. The core functions it provides are data acquisition and processing, system modeling and learning, and optimization and control. We present key modules of the Microgrid OS in the context of several implementation projects in microgrids, commercial buildings, and distribution networks.</p
The Reproduction of Homosocial Domesticity Aboard the Pequod in Melville’s Moby Dick
No full text[Introduction] In the 19th century, American society expected men and women to occupy entirely different worlds. Gender norms dictated that women belonged at home, in the domestic realm, where they performed routine household tasks and provided comfort and emotional support to their family. On the other hand, these norms encouraged men to pursue adventure and passion, and ideals of masculinity revolved around power and individuality. On the surface, the setting of a whaling boat is the perfect realization of the masculine sphere, as a group of men leave the domestic shore to travel the world, while reaping economic benefit and asserting power over nature by killing the seemingly indomitable sperm whale. This is the picture that Herman Melville paints at first in his 1851 novel Moby Dick, in which the main character Ishmael embarks with the crew of the Pequod to escape domesticity and takes part in Captain Ahab’s quest to kill the white whale. However, as Ishmael discusses his daily life, it becomes apparent that even in the absence of women, homosocial interactions on the ship recreate a new type of domesticity. Ishmael embraces the unexpected appearance of domesticity on the Pequod, which
enables him to survive the disastrous encounter with Moby Dick. This contrasts the tragic fate of Captain Ahab, who rejects all opportunities to take part in domestic affairs in favor of pursuing his individual quest to kill Moby Dick. By recreating domesticity on the Pequod and contrasting Ishmael with Ahab, Melville arrives at an important conclusion: domesticity persists even in masculine spheres and repeated attempts to eradicate it only result in disaster
Neural Coding of Finger Movements in Human Posterior Parietal Cortex and Motor Cortex
Get PDFWe use our hands constantly in our everyday lives. This seemingly simple ability is disrupted in individuals with cervical spinal cord injuries. By circumventing injured signal pathways, brain-computer interfaces (BCIs) promise to enable such individuals to control artificial limbs for everyday use. However, existing BCI limb control remains coarse and inflexible, because we do not understand how the recorded neural activity relates to dexterous movement. As a result, BCI control in physical settings remains frustratingly difficult for paralyzed users. To improve dexterous BCI control, I studied the neural coding of individual finger movements in the posterior parietal cortex and motor cortex of tetraplegic participants. These regions are directly involved in dexterous hand movements and are candidates for BCI recording implants. Finger coding matched the correlation structure and dynamics of able-bodied usage, reflecting preserved motor circuits even after paralysis. Individual finger movements of each hand were coded in a factorized, correlated manner that still allowed decoding. Participants controlled artificial fingers with state-of-the-art accuracy. Finally, we studied the temporal dynamics of neural control to understand how existing models of neural activity extend to BCI control. These findings contribute to the understanding of human hand movements and advance the development of dexterous BCIs
Synthetic Studies Toward the Total Synthesis of Enterocin
No full textAs part of a broader program aimed at the synthesis of complex and highly oxygenated natural products, we initiated a chemical synthesis of the natural polyketide enterocin. This dissertation will disclose our efforts to bridge that gap through the development of synthetic strategies for the total synthesis of the enterocin. The studies herein will address three unique strategies to access the tactical difficulties in the rich oxygenation patterns and caged core structure of enterocin. The program was first inspired by a SeO2 multioxidation reaction, and the methodology has been successfully applied to install bridgehead oxygenation patterns in enterocin. A strategy featuring a radical-polar crossover reaction as an annulation step to quickly construct the [3.2.1] bicyclic core of enterocin is detailed. Initial studies have successfully achieved the radical-polar crossover annulation reaction to forge [3.2.1]bicycles with bridgehead hydroxyl groups, and will guide the future development toward the total synthesis of enterocin. An intermolecular aldol approach will be discussed to address the challenge on pyrone installation and core structure synthesis. In summary, the development of an efficient and general approach will allow the development of novel reactions and a comprehensive evaluation of the potential of caged polyketides to serve as medicinally interesting molecules