Publications

Direct Numerical Simulations (DNS) data of Moderate or Intense Low-oxygen Dilution (MILD) combustion are analysed to identify the contributions of the autoignition and flame modes. This is performed using an extended Chemical Explosive Mode Analysis (CEMA) which accounts for diffusion effects allowing it to discriminate between deflagration and autoignition. This analysis indicates that in premixed MILD combustion conditions, the main combustion mode is ignition for all dilution and turbulence levels and for the two reactant temperature conditions considered. In non-premixed conditions, the preponderance of the ignition mode was observed to depend on the axial location and mixture fraction stratification. With a large mixture fraction lengthscale, ignition is more preponderant in the early part of the domain while the deflagrative mode increases further downstream. On the other hand, when the mixture fraction lengthscale is small, sequential autoignition is observed. Finally, the various combustion modes are observed to correlate strongly with mixture fraction where lean mixtures are more likely to autoignite while stoichiometric and rich mixtures are more likely to react as deflagrative structures.

The interaction of multiple injections in a diesel engine facilitates a complex interplay between freshly introduced fuel, previous combustion products, and overall combustion. To improve understanding of the relevant processes, high-speed Planar Laser-Induced Fluorescence (PLIF) with 355-nm excitation of formaldehyde and Polycyclic Aromatic Hydrocarbon (PAH) soot precursors is applied to multiple injections of n-dodecane from Engine Combustion Network Spray D, characterized by a converging 189-µm nozzle. High-speed schlieren imaging is applied simultaneously with 50-kHz PLIF excitation to visualize the spray structures, jet penetration, and ignition processes. For the first injection, formaldehyde (as an indicator of low-temperature chemistry) is first found in the jet periphery, after which it quickly propagates through the center of the jet, towards the jet head prior to high-temperature ignition. At second-stage ignition, downstream formaldehyde is consumed rapidly and upstream formaldehyde develops into a quasi-steady structure for as long as the momentum flux from the injector continues. Since the first injection in this work is relatively short, differences to a single long injection are readily observed, ultimately resulting in high-temperature combustion and PAH structures appearing farther upstream after the end of injection. For the second injection in this work, the first formaldehyde signal is significantly advanced because of the entrained high-temperature combustion products, and an obvious premixed burn event does not occur. The propensity for combustion recession after the end of the first injection changes significantly with ambient temperature, thereby affecting the level of interaction between the first- and second injection.

Distribution systems with high levels of solar PV may experience notable changes due to external conditions, such as temperature or solar irradiation. Fault detection methods must be developed in order to support these changes of conditions. This paper develops a method for fast detection, location, and classification of faults in a system with a high level of solar PV. The method uses the Continuous Wavelet Transform (CWT) technique to detect the traveling waves produced by fault events. The CWT coefficients of the current waveform at the traveling wave arrival time provide a fingerprint that is characteristic of each fault type and location. Two Convolutional Neural Networks are trained to classify any new fault event. The method relays of several protection devices and doesn't require communication between them. The results show that for multiple fault scenarios and solar PV conditions, high accuracy for both location and type classification can be obtained.

A computational simulation of various mixing laws for gaseous equations of state using planar traveling shocks for multiple mixtures in three dimensions is analyzed against nominal experimental data. Numerical simulations use the Sandia National Laboratories shock hydrodynamic code CTH and other codes including the thermochemical equilibrium code TIGER and the uncertainty qualification and sensitivity analysis code DAKOTA. The mixtures are 1:1 and a 1:3 molar mixtures of helium and sulfur hexafluoride. The mixing laws to be analyzed are the ideal-gas law, Amagat’s law, Dalton’s law, the Becker–Kistiakowsky–Wilson equation of state (EOS), the exponential 6 EOS, and the Jacobs-Cowperthwaite-Zwisler EOS. Examination of the experimental data with TIGER revealed that the shock strength should not be strong enough to turn the mixture nonideal because the compressibility factor z was essentially unity (z ≈ 1.02). Experimental results show that none of the equations of state are able to accurately predict the properties of the shocked mixture; similar discrepancies have been observed in previous works. Kinetic molecular theory appears to introduce a parameter that offers an explanation regarding the discrepancies. Implementation of the kinetic molecular theory parameter into the EOS is left for future work.

In the Nuclear Security Enterprise (NSE), many high reliability components must be stored for long periods of time before being called on to function a single time. During dormant storage, changes in the performance of these components may occur due to environmental exposures. These exposures may enhance the natural degradation of materials or result in shifts in the performance of electronics. Ongoing assessment of these components is necessary to inform the need for upgrades or replacements to ensure high reliability requirements are being maintained. This paper presents several assessment methodologies that are used and have been proposed for this problem. We also present methods that we believe to be most appropriate for the assessment of nuclear weapons components subjected to dormant storage.

The Source Physics Experiment (SPE) Phase I conducted six underground chemical explosions at the same experimental pad with the goal of characterizing underground explosions to enhance the United States (U.S.) ability to detect and discriminate underground nuclear explosions (UNEs). A fully polarimetric synthetic aperture RADAR (PolSAR) collected imagery in VideoSAR mode during the fifth and sixth explosions in the series (SPE-5 and SPE-6). Previously, we reported the prompt PolSAR surface changes cause by SPE-5 and SPE-6 explosions within seconds or minutes of the underground chemical explosions, including a drop of spatial coherence and polarimetric scattering changes. Therein it was hypothesized that surface changes occurred when surface particles experienced upward acceleration greater than 1 g. Because the SPE site was instrumented with surface accelerometers, we explore that hypothesis and report our findings in this article. We equate explosion-caused prompt surface expressions measured by PolSAR to the prompt surface movement measured by accelerometers. We tie these findings to UNE detection by comparing the PolSAR and accelerometer results to empirical ground motion predictions derived from accelerometer recordings of UNEs collected prior to cessation of U.S. nuclear testing. We find the single threshold greater than 1 g hypothesis is not correct for it does not explain the PolSAR results. Our findings show PolSAR surface coherence spatial extent is highly correlated with surface velocity, both measured and predicted, and the resulting surface deformation extent is corroborated by accelerometer records and the predicted lateral spall extent. PolSAR scattering changes measured during SPE-6 are created by the prompt surface displacement being larger than the spall gap.

Two key mechanical processes exist in the formation of powder compacts. These include the complex kinematics of particle rearrangement as the powder is densified and particle deformation leading to mechanical failure and fragmentation. Experiments measuring the time varying forces across a densifying powder bed have been performed in powders of microcrystalline cellulose with mean particle sizes between 0.4 and 1.2 mm. In these experiments, diagnostics measured the applied and transmitted loads and the bulk powder density. Any insight into the particle behavior must be inferred from deviations in the smoothly increasing stress-density compaction relationship. By incorporating a window in the compaction die body, simultaneous images of particle rearrangement and fracture at the confining window are captured. The images are post-processed in MATLAB® to track individual particle motion during compression. Complimentary discrete element method (DEM) simulations are presented and compared to experiment. The comparison provides insight into applying DEM methods for simulating large or permanent particle deformation and suggests areas for future study.

Abstract–Climate field reconstructions (CFRs) attempt to estimate spatiotemporal fields of climate variables in the past using climate proxies such as tree rings, ice cores, and corals. Data assimilation (DA) methods are a recent and promising new means of deriving CFRs that optimally fuse climate proxies with climate model output. Despite the growing application of DA-based CFRs, little is understood about how much the assimilated proxies change the statistical properties of the climate model data. To address this question, we propose a robust and computationally efficient method, based on functional data depth, to evaluate differences in the distributions of two spatiotemporal processes. We apply our test to study global and regional proxy influence in DA-based CFRs by comparing the background and analysis states, which are treated as two samples of spatiotemporal fields. We find that the analysis states are significantly altered from the climate-model-based background states due to the assimilation of proxies. Moreover, the difference between the analysis and background states increases with the number of proxies, even in regions far beyond proxy collection sites. Our approach allows us to characterize the added value of proxies, indicating where and when the analysis states are distinct from the background states. Supplementary materials for this article are available online.

Deep neural networks (DNNs) have achieved state-of-the-art performance across a variety of traditional machine learning tasks, e.g., speech recognition, image classification, and segmentation. The ability of DNNs to efficiently approximate high-dimensional functions has also motivated their use in scientific applications, e.g., to solve partial differential equations and to generate surrogate models. In this paper, we consider the supervised training of DNNs, which arises in many of the above applications. We focus on the central problem of optimizing the weights of the given DNN such that it accurately approximates the relation between observed input and target data. Devising effective solvers for this optimization problem is notoriously challenging due to the large number of weights, nonconvexity, data sparsity, and nontrivial choice of hyperparameters. To solve the optimization problem more efficiently, we propose the use of variable projection (VarPro), a method originally designed for separable nonlinear least-squares problems. Our main contribution is the Gauss–Newton VarPro method (GNvpro) that extends the reach of the VarPro idea to nonquadratic objective functions, most notably cross-entropy loss functions arising in classification. These extensions make GNvpro applicable to all training problems that involve a DNN whose last layer is an affine mapping, which is common in many state-of-the-art architectures. In our four numerical experiments from surrogate modeling, segmentation, and classification, GNvpro solves the optimization problem more efficiently than commonly used stochastic gradient descent (SGD) schemes. Also, GNvpro finds solutions that generalize well, and in all but one example better than well-tuned SGD methods, to unseen data points.

Linear elastic fracture mechanics theory predicts a parabolic crack opening profile. However, direct observation of crack tip shape in situ for brittle materials is challenging due to the small size of the active crack tip region. By leveraging advances in optical microscopy techniques and using a soft brittle hydrogel material, we can measure crack geometry on the micron scale. For glasses and ceramics, expected crack opening displacements are on the order of nanometers. However, for hydrogels, we can achieve crack opening displacements on the order of hundreds of microns or larger while maintaining brittle fracture behavior. Knowing the elastic properties, we can use crack geometry to calculate the stress intensity factor, K, and energy release rate, G, during propagation. Assuming the gel is hyperelastic, we can also approximate the size of the nonlinear region ahead of the crack tip. Geometric measurement of fracture properties eliminates the need to measure complex boundary and loading conditions, allowing us to explore new methods of inducing crack propagation. Further, this allows us to define measures of fracture resistance in materials that do not fit the traditionally defined theories of fracture mechanics.

Several recent workshops conducted by the DOE Advanced Scientific Computing Research program have established the fact that the complexity of developing applications and executing them on high-performance computing (HPC) systems is rising at a rate which will make it nearly impossible to continue to achieve higher levels of performance and scalability. Absent an alternative approach to managing this ever-growing complexity, HPC systems will become increasingly difficult to use. A more holistic approach to designing and developing applications and managing system resources is required. This paper outlines a research strategy for managing the increasing the complexity by providing the programming environment, software stack, and hardware capabilities needed for autonomous resource management of HPC systems. Developing portable applications for a variety of HPC systems of varying scale requires a paradigm shift from the current approach, where applications are painstakingly mapped to individual machine resources, to an approach where machine resources are automatically mapped and optimized to applications as they execute. Achieving such automated resource management for HPC systems is a daunting challenge that requires significant sustained investment in exploring new approaches and novel capabilities in software and hardware that span the spectrum from programming systems to device-level mechanisms. This paper provides an overview of the functionality needed to enable autonomous resource management and optimization and describes the components currently being explored at Sandia National Laboratories to help support this capability.

Physical vapor deposition (PVD) of high explosives can produce energetic samples with unique microstructure and morphology compared to traditional powder processing techniques, but challenges may exist in fabricating explosive films without defects. Deposition conditions and substrate material may promote microcracking and other defects in the explosive films. In this study, we investigate effects of engineered microscale defects (gaps) on detonation propagation and failure for pentaerythritol tetranitrate (PETN) films using ultra-high-speed refractive imaging and hydrocode modelling. Observations of the air shock above the gap reveal significant instabilities during gap crossing and re-ignition.

Accurate knowledge of post-detonation fireball temperatures is important for understanding device performance and for validation of numerical models. Such measurements are difficult to make even under controlled laboratory conditions. In this work temperature measurements were performed in the fireball of a commercial detonator (RP-80, Teledyne RISI). The explosion and fragments were contained in a plastic enclosure with glass windows for optical access. A hybrid femtosecond-picosecond (fs-ps) rotational coherent anti-Stokes Raman scattering (CARS) instrument was used to perform gas-phase thermometry along a one-dimensional measurement volume in a single laser shot. The 13-mm-thick windows on the explosive-containment housing introduced significant nonlinear chirp on the fs lasers pulses, which reduced the Raman excitation bandwidth and did not allow for efficient excitation of high-J Raman transitions populated at flame temperatures. To overcome this, distinct pump and Stokes pulses were used in conjunction with spectral focusing, achieved by varying the relative timing between the pump and Stokes pulses to preferentially excite Raman transitions relevant to flame thermometry. Light scattering from particulate matter and solid fragments was a significant challenge and was mitigated using a new polarization scheme to isolate the CARS signal. Fireball temperatures were measured 35-40 mm above the detonator, 12-25 mm radially outward from the detonator centerline, and at 18 and 28 μs after initiation. At these locations and times, significant mixing between the detonation products and ambient air had occurred thus increasing the nitrogen-based CARS thermometry signal. Initial measurements show a distribution of fireball temperatures in the range 300-2000 K with higher temperatures occurring 28 μs after detonation.

Oxiranes are a class of cyclic ethers formed in abundance during low-temperature combustion of hydrocarbons and biofuels, either via chain-propagating steps that occur from unimolecular decomposition of β-hydroperoxyalkyl radicals (β-̇QOOH) or from reactions of HOȮ with alkenes. Ethyloxirane is one of four alkyl-substituted cyclic ether isomers produced as an intermediate from n-butane oxidation. While rate coefficients for β-̇QOOH → ethyloxirane + ȮH are reported extensively, subsequent reaction mechanisms of the cyclic ether are not. As a result, chemical kinetics mechanisms commonly adopt simplified chemistry to describe ethyloxirane consumption by convoluting several elementary reactions into a single step, which may introduce mechanism truncation error—uncertainty derived from missing or incomplete chemistry. The present work provides fundamental insight on reaction mechanisms of ethyloxirane in support of ongoing efforts to minimize mechanism truncation error. Reaction mechanisms are inferred from the detection of products during chlorine atom-initiated oxidation experiments using multiplexed photoionization mass spectrometry conducted at 10 Torr and temperatures of 650 K and 800 K. To complement the experiments, calculations of stationary point energies were conducted using the ccCA-PS3 composite method on ̇R + O2 potential energy surfaces for the four ethyloxiranyl radical isomers, which produced barrier heights for 24 reaction pathways. In addition to products from ̇QOOH → cyclic ether + ȮH and ̇R + O2 → conjugate alkene + HOȮ, both of which were significant pathways and are prototypical to alkane oxidation, other species were identified from ring-opening of both ethyloxiranyl and ̇QOOH radicals. The latter occurs when the unpaired electron is localized on the ether group, causing the initial ̇QOOH structure to ring-open and form a resonance-stabilized ketohydroperoxide-type radical. The present work provides the first analysis of ethyloxirane oxidation chemistry, which reveals that consumption pathways are complex and may require an expansion of submechanisms to increase the fidelity of chemical kinetics mechanisms.

Neuromorphic computing with spintronic devices has been of interest due to the limitations of CMOS-driven von Neumann computing. Domain wall-magnetic tunnel junction (DW-MTJ) devices have been shown to be able to intrinsically capture biological neuron behavior. Edgy-relaxed behavior, where a frequently firing neuron experiences a lower action potential threshold, may provide additional artificial neuronal functionality when executing repeated tasks. In this letter, we demonstrate that this behavior can be implemented in DW-MTJ artificial neurons via three alternative mechanisms: shape anisotropy, magnetic field, and current-driven soft reset. Using micromagnetics and analytical device modeling to classify the Optdigits handwritten digit dataset, we show that edgy-relaxed behavior improves both classification accuracy and classification rate for ordered datasets while sacrificing little to no accuracy for a randomized dataset. This letter establishes methods by which artificial spintronic neurons can be flexibly adapted to datasets.

The intentional spread of disinformation is not a new challenge for the scientific world. We have seen it perpetuate the idea of a flat earth, convince communities that vaccines are more dangerous than helpful, and even suggest a connection between the “5G” communication infrastructure and COVID-19.1 Nor is disinformation a new phenomenon in the weapons inspection arena. Weapons inspectors themselves are often forced to sift through alternative narratives of events and inconsistent reporting, and they regularly see their credibility and conclusions questioned in the face of government politics or public biases. But certain recent disinformation campaigns have become so overwhelmingly comprehensive and effective that they constitute a new kind of threat. By preventing accountability for clear violations of international law, these campaigns have created a challenge to the survival of arms control treaties themselves. If weapons inspectors cannot regain the trust of the international community in the face of this challenge, it will be increasingly difficult to ensure compliance with arms control and disarmament treaties going forward. In this essay, I will briefly discuss one of the most comprehensive disinformation efforts of the past decade: the disinformation campaign used to prevent accountability for Syria's repeated use of chemical weapons. After this discussion, I will propose one possible approach to help protect the credibility of disarmament experts and weapons inspectors in the face of pervasive disinformation. This approach will require a concerted effort to connect and support compliance experts and to understand and explain their expertise across cultural, political, national, economic, and religious divides.

This paper offers new insights into a partial fuel stratification (PFS) combustion strategy that has proven to be effective at stabilizing overall lean combustion in direct injection spark ignition engines. To this aim, high spatial and temporal resolution optical diagnostics were applied in an optically accessible engine working in PFS mode for two fuels and two different durations of pilot injection at the time of spark: 210 μs and 330 μs for E30 (gasoline blended with ethanol by 30% volume fraction) and gasoline, respectively. In both conditions, early injections during the intake stroke were used to generate a well-mixed lean background. The results were compared to rich, stoichiometric and lean well-mixed combustion with different spark timings. In the PFS combustion process, it was possible to detect a non-spherical and highly wrinkled blue flame, coupled with yellow diffusive flames due to the combustion of rich zones near the spark plug. The initial flame spread for both PFS cases was faster compared to any of the well-mixed cases (lean, stoichiometric and rich), suggesting that the flame propagation for PFS is enhanced by both enrichment and enhanced local turbulence caused by the pilot injection. Different spray evolutions for the two pilot injection durations were found to strongly influence the flame kernel inception and propagation. PFS with pilot durations of 210 μs and 330 μs showed some differences in terms of shapes of the flame front and in terms of extension of diffusive flames. Yet, both cases were highly repeatable.

Solving dense systems of linear equations is essential in applications encountered in physics, mathematics, and engineering. This paper describes our current efforts toward the development of the ADELUS package for current and next generation distributed, accelerator-based, high-performance computing platforms. The package solves dense linear systems using partial pivoting LU factorization on distributed-memory systems with CPUs/GPUs. The matrix is block-mapped onto distributed memory on CPUs/GPUs and is solved as if it was torus-wrapped for an optimal balance of computation and communication. A permutation operation is performed to restore the results so the torus-wrap distribution is transparent to the user. This package targets performance portability by leveraging the abstractions provided in the Kokkos and Kokkos Kernels libraries. Comparison of the performance gains versus the state-of-the-art SLATE and DPLASMA GESV functionalities on the Summit supercomputer are provided. Preliminary performance results from large-scale electromagnetic simulations using ADELUS are also presented. The solver achieves 7.7 Petaflops on 7600 GPUs of the Sierra supercomputer translating to 16.9% efficiency.

We analyze the coupling into a slotted cylindrical cavity operating at fundamental cavity modal frequencies overlapping with the slot’s first resonance frequency through an unmatched formulation that accounts for the slot’s absorption and radiation processes. The model is validated through full-wave simulations and experimental data. We then couple the unmatched formulation to a perturbation theory model to investigate an absorber within the cavity to reduce the interior field strength, also validated with full-wave simulations and experiments. These models are pivotal to understanding the physical processes involved in the electromagnetic penetration through slots, and may constitute design tools to mitigate electromagnetic interference effects within cavities.

Abstract not provided.

The accurate construction of a surrogate model is an effective and efficient strategy for performing Uncertainty Quantification (UQ) analyses of expensive and complex engineering systems. Surrogate models are especially powerful whenever the UQ analysis requires the computation of statistics which are difficult and prohibitively expensive to obtain via a direct sampling of the model, e.g. high-order moments and probability density functions. In this paper, we discuss the construction of a polynomial chaos expansion (PCE) surrogate model for radiation transport problems for which quantities of interest are obtained via Monte Carlo simulations. In this context, it is imperative to account for the statistical variability of the simulator as well as the variability associated with the uncertain parameter inputs. More formally, in this paper we focus on understanding the impact of the Monte Carlo transport variability on the recovery of the PCE coefficients. We are able to identify the contribution of both the number of uncertain parameter samples and the number of particle histories simulated per sample in the PCE coefficient recovery. Our theoretical results indicate an accuracy improvement when using few Monte Carlo histories per random sample with respect to configurations with an equivalent computational cost. These theoretical results are numerically illustrated for a simple synthetic example and two configurations of a one-dimensional radiation transport problem in which a slab is represented by means of materials with uncertain cross sections.

Abstract not provided.

Abstract not provided.

Explosively driven ferroelectric generators (FEG) are used as pulsed power sources in many applications that require a compact design that delivers a short high-voltage and high-current pulse. A mechanical shock applied to ferroelectrics releases bound electrical charge through a combination of piezoelectric, domain reorientation, and phase transformation effects. Lead-zirconate-titanate (PZT) 95/5 lies near the ferroelectric (FE)-antiferroelectric (AF) phase boundary and readily transforms to AF phase under compression because AF has a smaller unit volume. This makes it a popular choice for FEGs as the FE-AF transformation completely releases all the stored dipole charge. The complexity of piezoelectric, domain reorientation, and phase transformation behaviors under high deviatoric stress makes modeling this FE to AF transformation and the accompanying charge release challenging. The mode and direction of domain reorientation and phase transformation varies significantly with different deviatoric and hydrostatic stress states. Microstructure changes due to domain reorientation and phase alter the piezoelectric properties of the material. Inaccuracies in modeling any one of these phenomena can result in inaccurate electrical response. This work demonstrates a material model that accurately captures the linear piezoelectric, domain reorientation and phase transformation phenomena by using a micromechanical approach to approximate the changes in domain-structure.

Abstract not provided.

To meet the extreme compute demands for deep learning across commercial and scientific applications, dataflow accelerators are becoming increasingly popular. While these “domain-specific” accelerators are not fully programmable like CPUs and GPUs, they retain varying levels of flexibility with respect to data orchestration, i.e., dataflow and tiling optimizations to enhance efficiency. There are several challenges when designing new algorithms and mapping approaches to execute the algorithms for a target problem on new hardware. Previous works have addressed these challenges individually. To address this challenge as a whole, in this work, we present a HW-SW codesign ecosystem for spatial accelerators called Union within the popular MLIR compiler infrastructure. Our framework allows exploring different algorithms and their mappings on several accelerator cost models. Union also includes a plug-and-play library of accelerator cost models and mappers which can easily be extended. The algorithms and accelerator cost models are connected via a novel mapping abstraction that captures the map space of spatial accelerators which can be systematically pruned based on constraints from the hardware, workload, and mapper. We demonstrate the value of Union for the community with several case studies which examine offloading different tensor operations (CONV/GEMM/Tensor Contraction) on diverse accelerator architectures using different mapping schemes.

In this study, we present spectral equivalence results for high-order tensor product edge- and face-based finite elements for the H(curl) and H(div) function spaces. Specifically, we show for certain choices of shape functions that the mass and stiffness matrices of the high-order elements are spectrally equivalent to those for an assembly of low-order elements on the associated Gauss-Lobatto-Legendre mesh. Based on this equivalence, efficient preconditioners can be designed with favorable computational complexity. Numerical results are presented which confirm the theory and demonstrate the benefits of the equivalence results for overlapping Schwarz preconditioners.

The Quantum Scientific Computing Open User Testbed (QSCOUT) at Sandia National Laboratories is a trapped-ion qubit system designed to evaluate the potential of near-term quantum hardware in scientific computing applications for the U.S. Department of Energy and its Advanced Scientific Computing Research program. Similar to commercially available platforms, it offers quantum hardware that researchers can use to perform quantum algorithms, investigate noise properties unique to quantum systems, and test novel ideas that will be useful for larger and more powerful systems in the future. However, unlike most other quantum computing testbeds, the QSCOUT allows both quantum circuit and low-level pulse control access to study new modes of programming and optimization. The purpose of this article is to provide users and the general community with details of the QSCOUT hardware and its interface, enabling them to take maximum advantage of its capabilities.

Typical Type-4 wind turbines use DC-link inverters to couple the electrical machine to the power grid. Each wind turbine has two power conversion steps. Therefore, an N-turbine farm will have 2N power converters. This work presents a DC bus collection system for a type-4 wind farm that reduces the overall required number of converters and minimizes the energy storage system (ESS) requirements. This approach requires one conversion step per turbine, one converter for the ESS and a single grid coupling converter, which leads to N + 2 converters for the wind farm which will result in significant cost savings. However, one of the trade-offs for a DC collection system is the need for increased energy storage to filter the power variations and improve power quality to the grid. This paper presents a novel approach to an effective DC bus collection system design. The DC collection for the wind farm implements a power phasing control method between turbines that filter the variations and improves power quality while minimizing the need for added energy storage system hardware and improved power quality. The phasing control takes advantage of a novel power packet network concept with nonlinear power flow control design techniques that guarantees both stable and enhanced dynamic performance. This paper presents the theoretical design of the DC collection and phasing control. To demonstrate the efficacy of this approach detailed numerical simulation examples are presented.

In the Republic of Uganda, it is estimated that nearly 28 million people lack access to clean and safe drinking water [1]. The authors developed a model by performing multiple linear regression using predictors air temperature, irradiance, and wind speeds in two rural areas to determine the suitability of a proposed system to assist. The system consists of a commercial grade vertical axis wind turbine (VAWT) with embedded solar cells capable of providing water pumping and small-scale electricity generation. A suite of equations was used alongside the regression models to determine how to adjust the mechanical properties of the turbine such that the solar and wind energy function as mutually redundant drivers for both a small-scale electricity generator and water pumping system. The results are consistent with the following: solar output variation of 20% (from 4.5 to 5.5 W/m2) [2], [3]; 3.7 m/s to 6m/sec require to operate the turbine; and water pumped at the rate of 3748.54 ft-lbs to 42314.72 ft-lbs per hour. The primary researcher for this project has applied for and received a provisional patent to advance the VAWT/SES technology. This year a utility patent was filed to move the complete energy renewal system toward commercialization.

Abstract not provided.

This paper reports on the development of a self-synchronizing underwater acoustic network developed for remote monitoring of mooring loads in Wave Energy Converters (WECs). This network uses Time Division Multiple Access and operates self-contained with the ability for users to remotely transmit commands to the network as needed. Each node is a self-contained unit, consisting of a protocol adaptor board, an FAU-DPAM underwater acoustic modem and a battery pack. A node can be connected to a load cell, to a topside user or to the WEC. Every node is swapable. The protocol adaptor board, named Protocol Adaptor for Digital LOad Cell (PADLOC) supports a variety of digital load cell message formats (CAN, MODBUS, custom ASCII) and underwater acoustic modem serial formats. PADLOC enables topside users to connect to separate load cells through a user-specific command. This is especially important if the user is monitoring multiple load cells during deployment or maintenance, when the primary data system may be offline. Each PADLOC board handles formatting, buffering and has a one-on-one serial connection with each pair (node) of a digital load cell and acoustic modem. In addition, each PADLOC board handles the timekeeping and power saving features for each node. The only limitation is the data bit rate and delay limitations associated with the underwater acoustic modem. A four node self-synchronizing network has been developed to demonstrate the load cell monitoring capability using the PADLOC technology on the CalWave WEC.

Both the data science and scientific computing communities are embracing GPU acceleration for their most demanding workloads. For scientific computing applications, the massive volume of code and diversity of hardware platforms at supercomputing centers has motivated a strong effort toward performance portability. This property of a program, denoting its ability to perform well on multiple architectures and varied datasets, is heavily dependent on the choice of parallel programming model and which features of the programming model are used. In this paper, we evaluate performance portability in the context of a data science workload in contrast to a scientific computing workload, evaluating the same sparse matrix kernel on both. Among our implementations of the kernel in different performance-portable programming models, we find that many struggle to consistently achieve performance improvements using the GPU compared to simple one-line OpenMP parallelization on high-end multicore CPUs. We show one that does, and its performance approaches and sometimes even matches that of vendor-provided GPU math libraries.

In this work, we show that reduced communication algorithms for distributed stochastic gradient descent improve the time per epoch and strong scaling for the Generalized Canonical Polyadic (GCP) tensor decomposition, but with a cost, achieving convergence becomes more difficult. The implementation, based on MPI, shows that while one-sided algorithms offer a path to asynchronous execution, the performance benefits of optimized allreduce are difficult to best.

Gate set tomography (GST) is a protocol for detailed, predictive characterization of logic operations (gates) on quantum computing processors. Early versions of GST emerged around 2012-13, and since then it has been refined, demonstrated, and used in a large number of experiments. This paper presents the foundations of GST in comprehensive detail. The most important feature of GST, compared to older state and process tomography protocols, is that it is calibration-free. GST does not rely on pre-calibrated state preparations and measurements. Instead, it characterizes all the operations in a gate set simultaneously and self-consistently, relative to each other. Long sequence GST can estimate gates with very high precision and efficiency, achieving Heisenberg scaling in regimes of practical interest. In this paper, we cover GST’s intellectual history, the techniques and experiments used to achieve its intended purpose, data analysis, gauge freedom and fixing, error bars, and the interpretation of gauge-fixed estimates of gate sets. Our focus is fundamental mathematical aspects of GST, rather than implementation details, but we touch on some of the foundational algorithmic tricks used in the pyGSTi implementation.

On high-performance computing (HPC) systems, job allocation strategies control the placement of a job among available nodes. As the placement changes a job's communication performance, allocation can significantly affects execution times of many HPC applications. Existing allocation strategies typically make decisions based on resource limit, network topology, communication patterns, etc. However, system network performance at runtime is seldom consulted in allocation, even though it significantly affects job execution times.In this work, we demonstrate using monitoring data to improve HPC systems' performance by proposing a NetworkData-Driven (NeDD) job allocation framework, which monitors the network performance of an HPC system at runtime and allocates resources based on both network performance and job characteristics. NeDD characterizes system network performance by collecting the network traffic statistics on each router link, and it characterizes a job's sensitivity to network congestion by collecting Message Passing Interface (MPI) statistics. During allocation, NeDD pairs network-sensitive (network-insensitive) jobs with nodes whose parent routers have low (high) network traffic. Through experiments on a large HPC system, we demonstrate that NeDD reduces the execution time of parallel applications by 11% on average and up to 34%.

The ability to handle data is critical at all stages of marine energy development. The Marine and Hydrokinetic Toolkit (MHKiT) is an open-source marine energy software, which includes modules for ingesting, applying quality control, processing, visualizing, and managing data. MHKiT-Python and MHKiT-MATLAB provide robust and verified functions that are needed by the marine energy community to standardize data processing. Calculations and visualizations adhere to International Electrotechnical Commission technical specifications and other guidelines. A resource assessment of National Data Buoy Center buoy 46050 near PACWAVE is performed using MHKiT and we discuss comparisons to the resource assessment provided performed by Dunkle et al. (2020).

The Domain Name System (DNS) has been frequently abused for distributed denial-of-service (DDoS) attacks and cache poisoning because it relies on the User Datagram Protocol (UDP). Since UDP is connection-less, it is trivial for an attacker to spoof the source of a DNS query or response. While other secure transport mechanisms provide identity management, such as the Transmission Control Protocol (TCP) and DNS Cookies, there is currently no method for a client to state that they only use a given protocol. This paper presents a new method to allow protocol enforcement: DNS Protocol Advertisement Records (DPAR). Advertisement records allow Internet Protocol (IP) address subnets to post a public record in the reverse DNS zone stating which DNS mechanisms are used by their clients. DNS servers may then look up this record and require a client to use the stated mechanism, in turn preventing an attacker from sending spoofed messages over UDP. In this paper, we define the specification for DNS Protocol Advertisement Records, considerations that were made, and comparisons to alternative approaches. We additionally estimate the effectiveness of advertisements in preventing DDoS attacks and the expected burden to DNS servers.

Industrial and security communities leverage x-ray computed tomography for several applications in non-destructive evaluation such as material detection and metrology. Many of these applications ultimately reach a limit as most x-ray systems have a nonlinear mathematical operator due to the Bremsstrahlung radiation emitted from the x-ray source. This work proposes a design of a multi-metal pattered anode coupled with a hyperspectral X-ray detector to improve spatial resolution, absorption signal, and overall data quality for various quantitative. The union of a multi-metal pattered anode x-ray source with an energy-resolved photon counting detector permits the generation and detection of a preferential set of X-ray energy peaks. When photons about the peaks are detected, while rejecting photons outside this neighborhood, the overall quality of the image is improved by linearizing the operator that defines the image formation. Additionally, the effective X-ray focal spot size allows for further improvement of the image quality by increasing resolution. Previous works use machine learning techniques to analyze the hyperspectral computed tomography signal and reliably identify and discriminate a wide range of materials based on a material's composition, improving data quality through a multi-material pattern anode will further enhance these identification and classification methods. This work presents initial investigations of a multi-metal patterned anode along with a hyperspectral detector using a general-purpose Monte Carlo particle transport code known as PHITS version 3.24. If successful, these results will have tremendous impact on several nondestructive evaluation applications in industry, security, and medicine.

This study analyzes the ability of various processing techniques to reduce leakage current in vertical GaN MOS devices. Careful analysis is required to determine suitable gate dielectric materials in vertical GaN MOSFET devices since they are largely responsible for determination of threshold voltage, gate leakage reduction, and semiconductor/dielectric interface traps. SiO2, Al2 O3, and HfO2 films were deposited by Atomic Layer Deposition (ALD) and subjected to treatments nominally identical to those in a vertical GaN MOSFET fabrication sequence. This work determines mechanisms for reducing gate leakage by reduction of surface contaminants and interface traps using pre-deposition cleans, elevated temperature depositions, and post-deposition anneals. Breakdown measurements indicate that ALD Al2O3 is an ideal candidate for a MOSFET gate dielectric, with a breakdown electric field near 7.5 MV/cm with no high temperature annealing required to increase breakdown strength. SiO2 ALD films treated with a post deposition anneal at 850 °C for 30 minutes show significant reduction in leakage current while maintaining breakdown at 5.5 MV/cm. HfO2 films show breakdown nominally identical to annealed SiO2 films, but with significantly higher leakage. Additionally, HfO2 films show more sensitivity to high temperature annealing suggesting that more research into surface cleans is necessary to improving these films for MOSFET gate applications.

This paper focuses on the role of the Marine Renewable Energy (MRE) Software Knowledge Hub on the Portal and Repository for Information on Marine Renewable Energy (PRIMRE). The MRE Software Knowledge Hub provides online services for MRE software users and developers, and seeks to develop assessments and recommendations for improving MRE software in the future. Online software discovery platforms, known as the Code Hub and the Code Catalog, are provided. The Code Hub is a collection of open-source MRE software that includes a landing page with search functionality, linked to files hosted on the MRE Code Hub GitHub organization. The Code Catalog is a searchable online platform for discovery of useful (open-source or commercial) software packages, tools, codes, and other software products. To gather information about the existing MRE software landscape, a software survey is being performed, the preliminary results of which are presented herein. Initially, the data collected in the MRE software survey will be used to populate the MRE Software knowledge hub on PRIMRE, and future work will use data from the survey to perform a gap analysis and develop a vision for future software development. Additionally, as one of PRIMRE’s roles is to support development of MRE software within project partners, a silo of knowledge relating to best practices has been gathered. An early draft of new guidance developed from this knowledge is presented.

Container escapes enable the adversary to execute code on the host from inside an isolated container. These high severity escape vulnerabilities originate from three sources: (1) container profile misconfigurations, (2) Linux kernel bugs, and (3) container runtime vulnerabilities. While the first two cases have been studied in the literature, no works have investigated the impact of container runtime vulnerabilities. In this paper, to fill this gap, we study 59 CVEs for 11 different container runtimes. As a result of our study, we found that five of the 11 runtimes had nine publicly available PoC container escape exploits covering 13 CVEs. Our further analysis revealed all nine exploits are the result of a host component leaked into the container. We apply a user namespace container defense to prevent the adversary from leveraging leaked host components and demonstrate that the defense stops seven of the nine container escape exploits.

This work provides the first demonstration of vertical GaN Junction Barrier Schottky (JBS) rectifiers fabricated by etch and regrowth of p-GaN. A reverse blocking voltage near 1500 V was achieved at 1 mA reverse leakage, with a sub 1 V turn-on and a specific on-resistance of 10 mΩ-cm2. This result is compared to other reported JBS devices in the literature and our device demonstrates the lowest leakage slope at high reverse bias. A large initial leakage current is present near zero-bias which is attributed to a combination of inadequate etch-damage removal and passivation induced leakage current.

This paper presents an isolated bidirectional dc/dc converter for battery energy storage applications. Two main features of the proposed circuit topology are high voltage-conversion ratio and reduced battery current ripple. The primary side circuit is a quasi-switched-capacitor circuit with reduced voltage stress on switching devices and a 3:1 voltage step down ratio, which reduces the turns ratio of the transformer to 6:1:1. The secondary side circuit has an interleaved operation by utilizing the split magnetizing inductance of the transformer, which not only helps to increase the step down ratio but also reduces the battery current ripple. Similar to the dual-active-bridge circuit, the phase shift control is implemented to regulate the operation power of the circuit. A 1-kW, 300-kHz, 380-420 V/20-33 V GaN-based circuit prototype is currently under fabrication. The preliminary test results are presented.

This study presents a numerical model of a WEC array. The model will be used in subsequent work to study the ability of data assimilation to support power prediction from WEC arrays and WEC array design. In this study, we focus on design, modeling, and control of the WEC array. A case study is performed for a small remote Alaskan town. Using an efficient method for modeling the linear interactions within a homogeneous array, we produce a model and predictionless feedback controllers for the devices within the array. The model is applied to study the effects of spectral wave forecast errors on power output. The results of this analysis show that the power performance of the WEC array will be most strongly affected by errors in prediction of the spectral period, but that reductions in performance can realistically be limited to less than 10% based on typical data assimilation based spectral forecasting accuracy levels.

To meet the extreme compute demands for deep learning across commercial and scientific applications, dataflow accelerators are becoming increasingly popular. While these “domain-specific” accelerators are not fully programmable like CPUs and GPUs, they retain varying levels of flexibility with respect to data orchestration, i.e., dataflow and tiling optimizations to enhance efficiency. There are several challenges when designing new algorithms and mapping approaches to execute the algorithms for a target problem on new hardware. Previous works have addressed these challenges individually. To address this challenge as a whole, in this work, we present a HW-SW codesign ecosystem for spatial accelerators called Union within the popular MLIR compiler infrastructure. Our framework allows exploring different algorithms and their mappings on several accelerator cost models. Union also includes a plug-and-play library of accelerator cost models and mappers which can easily be extended. The algorithms and accelerator cost models are connected via a novel mapping abstraction that captures the map space of spatial accelerators which can be systematically pruned based on constraints from the hardware, workload, and mapper. We demonstrate the value of Union for the community with several case studies which examine offloading different tensor operations (CONV/GEMM/Tensor Contraction) on diverse accelerator architectures using different mapping schemes.

Ultrasonic transducers can be leveraged to transmit power and data through metallic enclosures such as Faraday cages for which standard electromagnetic methods are infeasible. The design of these systems features a number of variables that must be carefully tweaked for optimal data and power transfer rate and efficiency. The objective of this work is to present a toolkit, COMET, standing for Computational Optimization of Mechanical Energy Transduction, in which the design process and analysis of such transducer systems is streamlined. The toolkit features flexible tools for introducing an arbitrary number of backing/bonding layers, material libraries, parameter sweeps, and optimization.

Several applications, such as underwater vehicles or waste containers, require the ability to transfer data from transducers enclosed by metallic structures. In these cases, Faraday shielding makes electromagnetic transmission highly inefficient, and suggests the employment of ultrasonic transmission as a promising alternative. While ultrasonic data transmission by piezoelectric transduction provides a practical solution, the amplitude of the transmitted signal strongly depends on acoustic resonances of the transmission line, which limits the bandwidth over which signals are sent and the rate of data transmission. The objective of this work is to investigate piezoelectric acoustic transducer configurations that enable data transmission at a relatively constant amplitude over large frequency bands. This is achieved through structural modifications of the transmission line, which includes layering of the transducers, as well as the introduction of electric circuits connected to both transmitting and receiving transducers. Both strategies lead to strong enhancements in the available bandwidth and show promising directions for the design of effective acoustic transmission across metallic barriers.

For systems that require complete metallic enclosures, it is impossible to power and communicate with interior electronics using conventional electromagnetic techniques. Instead, pairs of ultrasonic transducers can be used to send and receive elastic waves through the enclosure, forming an equivalent electrical transmission line that bypasses the Faraday cage effect. These mechanical communication systems introduce the possibility for electromechanical crosstalk between channels on the same barrier, in which receivers output erroneous electrical signals due to ultrasonic guided waves generated by transmitters in adjacent communication channels. To minimize this crosstalk, this work investigates the use of a phononic crystal/metamaterial machined into the barrier via periodic grooving. Barriers with simultaneous ultrasonic power and data transfer are fabricated and tested to measure the effect of grooving on crosstalk between channels.

Both the data science and scientific computing communities are embracing GPU acceleration for their most demanding workloads. For scientific computing applications, the massive volume of code and diversity of hardware platforms at supercomputing centers has motivated a strong effort toward performance portability. This property of a program, denoting its ability to perform well on multiple architectures and varied datasets, is heavily dependent on the choice of parallel programming model and which features of the programming model are used. In this paper, we evaluate performance portability in the context of a data science workload in contrast to a scientific computing workload, evaluating the same sparse matrix kernel on both. Among our implementations of the kernel in different performance-portable programming models, we find that many struggle to consistently achieve performance improvements using the GPU compared to simple one-line OpenMP parallelization on high-end multicore CPUs. We show one that does, and its performance approaches and sometimes even matches that of vendor-provided GPU math libraries.

Poisson Tensor Factorization (PTF) is an important data analysis method for analyzing patterns and relationships in multiway count data. In this work, we consider several algorithms for computing a low-rank PTF of tensors with sparse count data values via maximum likelihood estimation. Such an approach reduces to solving a nonlinear, non-convex optimization problem, which can leverage considerable parallel computation due to the structure of the problem. However, since the maximum likelihood estimator corresponds to the global minimizer of this optimization problem, it is important to consider how effective methods are at both leveraging this inherent parallelism as well as computing a good approximation to the global minimizer. In this work we present comparisons of multiple methods for PTF that illustrate the tradeoffs in computational efficiency and accurately computing the maximum likelihood estimator. We present results using synthetic and real-world data tensors to demonstrate some of the challenges when choosing a method for a given tensor.

This paper reports results from an ongoing investigation on potential ways to utilize small wave energy devices that can be transported in, and deployed from, torpedo tubes. The devices are designed to perform designated ocean measurement operations and thus need to convert enough energy to power onboard sensors, while storing any excess energy to support vehicle recharging operations. Examined in this paper is a traditional tubular oscillating water column device, and particular interest here is in designs that lead to optimization of power converted from shorter wind sea waves. A two step design procedure is investigated here, wherein a more approximate two-degree-of-freedom model is first used to identify relative dimensions (of device elements) that optimize power conversion from relative oscillations between the device elements. A more rigorous mathematical model based on the hydrodynamics of oscillating pressure distributions within solid oscillators is then used to provide the hydrodynamic coefficients, forces, and flow rates for the device. These results provide a quick but rigorous way to estimate the energy conversion performance of the device in various wave climates, while enabling more accurate design of the power takeoff and energy storage systems.

The Sandia Critical Experiments (SCX) Program provides a specialized facility for performing water moderated and reflected critical experiments with UO₂ fuel rod arrays. A history of safe reactor operations and flexibility in reactor core configuration has resulted in the completion of several benchmark critical experiment evaluations that are published in the International Criticality Safety Benchmark Evaluation Project (ICSBEP) Handbook. The LEUCOMP-THERM-078 and LEU-COMP-THERM-080 evaluations from the handbook provide similar cases for reference. The set of experiments described here were performed using the Seven Percent Critical Experiment (7uPCX) fuel to measure the effects of decreasing the fuel-to-water volume ratio on the critical array size. This was accomplished by using fuel loading patterns to effectively increase the pitch of the fuel arrays in the assembly. The fuel rod pitch variations provided assembly configurations that ranged from strongly undermoderated to slightly overmoderated.

In this work, we show that reduced communication algorithms for distributed stochastic gradient descent improve the time per epoch and strong scaling for the Generalized Canonical Polyadic (GCP) tensor decomposition, but with a cost, achieving convergence becomes more difficult. The implementation, based on MPI, shows that while one-sided algorithms offer a path to asynchronous execution, the performance benefits of optimized allreduce are difficult to best.

On high-performance computing (HPC) systems, job allocation strategies control the placement of a job among available nodes. As the placement changes a job's communication performance, allocation can significantly affects execution times of many HPC applications. Existing allocation strategies typically make decisions based on resource limit, network topology, communication patterns, etc. However, system network performance at runtime is seldom consulted in allocation, even though it significantly affects job execution times.In this work, we demonstrate using monitoring data to improve HPC systems' performance by proposing a NetworkData-Driven (NeDD) job allocation framework, which monitors the network performance of an HPC system at runtime and allocates resources based on both network performance and job characteristics. NeDD characterizes system network performance by collecting the network traffic statistics on each router link, and it characterizes a job's sensitivity to network congestion by collecting Message Passing Interface (MPI) statistics. During allocation, NeDD pairs network-sensitive (network-insensitive) jobs with nodes whose parent routers have low (high) network traffic. Through experiments on a large HPC system, we demonstrate that NeDD reduces the execution time of parallel applications by 11% on average and up to 34%.

We present StressBench, a network benchmarking framework written for testing MPI operations and file I/O concurrently. It is designed specifically to execute MPI communication and file access patterns that are representative of real-world scientific applications. Existing tools consider either the worst case congestion with small abstract patterns or peak performance with simplistic patterns. StressBench allows for a richer study of congestion by allowing orchestration of network load scenarios that are representative of those typically seen at HPC centres, something that is difficult to achieve with existing tools. We demonstrate the versatility of the framework from micro benchmarks through to finely controlled congested runs across a cluster. Validation of the results using four proxy application communication schemes within StressBench against parent applications shows a maximum difference of 15%. Using the I/O modeling capabilities of StressBench, we are able to quantify the impact of file I/O on application traffic showing how it can be used in procurement and performance studies.

Significant costs can be related to losing circulation of drilling fluids in geothermal drilling. This paper is the second of four case studies of geothermal fields operated by Ormat Technologies, directed at forming a comprehensive strategy to characterize and address lost circulation in varying conditions, and examines the geologic context of and common responses to lost circulation in the loosely consolidated, shallow sedimentary reservoir of the Don A. Campbell geothermal field. The Don A. Campbell Geothermal Field is in the SW portion of Gabbs Valley in NV, along the eastern margin of the Central Walker Lane shear zone. The reservoir here is shallow and primarily in the basin fill, which is hydrothermally altered along fault zones. Wells in this reservoir are highly productive (250-315 L/s) with moderate temperatures (120-125 °C) and were drilled to an average depth of ~1500 ft (450 m). Lost circulation is frequently reported beginning at depths of about 800 ft, slightly shallower than the average casing shoe depth of 900- 1000 ft (275-305 m). Reports of lost circulation frequently coincide with drilling through silicified basin fill. Strategies to address lost circulation differ above and below the cased interval; bentonite chips were used at shallow depths and aerated, gelled drilling fluids were used in the production intervals. Further study of this and other areas will contribute to developing a systematic understanding of geologic contextual-informed lost circulation mitigation strategies.

Resilience has been defined as a priority for the US critical infrastructure. This paper presents a process for incorporating resiliency-derived metrics into security system evaluations. To support this analysis, we used a multi-layer network model (MLN) reflecting the defined security system of a hypothetical nuclear power plant to define what metrics would be useful in understanding a system's ability to absorb perturbation (i.e., system resilience). We defined measures focusing on the system's criticality, rapidity, diversity, and confidence at each network layer, simulated adversary path, and the system as a basis for understanding the system's resilience. For this hypothetical system, our metrics indicated the importance of physical infrastructure to overall system criticality, the relative confidence of physical sensors, and the lack of diversity in assessment activities (i.e., dependence on human evaluations). Refined model design and data outputs will enable more nuanced evaluations into temporal, geospatial, and human behavior considerations. Future studies can also extend these methodologies to capture respond and recover aspects of resilience, further supporting the protection of critical infrastructure.

Abstract not provided.

The Domain Name System (DNS) has been frequently abused for Distributed Denial of Service (DDoS) attacks and cache poisoning because it relies on the User Datagram Protocol (UDP). Since UDP is connection-less, it is trivial for an attacker to spoof the source of a DNS query or response. DNS Cookies, a protocol standardized in 2016, add pseudo-random values to DNS packets to provide identity management and prevent spoofing attacks. In this paper, we present the first study measuring the deployment of DNS Cookies in nearly all aspects of the DNS architecture. We also provide an analysis of the current benefits of DNS Cookies and the next steps for stricter deployment. Our findings show that cookie use is limited to less than 30% of servers and 10% of recursive clients. We also find several configuration issues that could lead to substantial problems if cookies were strictly required. Overall, DNS Cookies provide limited benefit in a majority of situations, and, given current deployment, do not prevent DDoS or cache poisoning attacks.

We present a novel graph convolutional layer that is conceptually simple, fast, and provides high accuracy with reduced overfitting. Based on pseudo-differential operators, our layer operates on graphs with relative position information available for each pair of connected nodes. Our layer represents a generalization of parameterized differential operators (previously shown effective for shape correspondence, image segmentation, and dimensionality reduction tasks) to a larger class of graphs. We evaluate our method on a variety of supervised learning tasks, including 2D graph classification using the MNIST and CIFAR-100 datasets and 3D node correspondence using the FAUST dataset. We also introduce a superpixel graph version of the lesion classification task using the ISIC 2016 challenge dataset and evaluate our layer versus other state-of-the-art graph convolutional network architectures.The new layer outperforms multiple recent architectures on graph classification tasks using the MNIST and CIFAR-100 superpixel datasets. For the ISIC dataset, we outperform all other graph neural networks examined as well as all of the submissions to the original ISIC challenge despite the best of those models having more than 200 times as many parameters as our model.

Rechargeable alkaline Zn–MnO2 (RAM) batteries are a promising candidate for grid-scale energy storage owing to their high theoretical energy density rivaling lithium-ion systems (∼400 Wh/L), relatively safe aqueous electrolyte, established supply chain, and projected costs below $100/kWh at scale. In practice, however, many fundamental chemical and physical processes at both electrodes make it difficult to achieve commercially competitive energy density and cycle life. This review presents a detailed and timely analysis of the constituent materials, current commercial status, electrode processes, and performance-limiting factors of RAM batteries. We also examine recently reported strategies in RAM and related systems to address these issues through additives and modifications to the electrode materials and electrolyte, special ion-selective separators and/or coatings, and unconventional cycling protocols. We conclude with a critical summary of these developments and discussion of how future studies should be focused toward the goal of energy-dense, scalable, and cost-effective RAM systems.

Dimension reduction techniques have frequently been used to summarize information from high dimensional hyperspectral data, usually done in effort to classify or visualize the materials contained in the hyperspectral image. The main challenge in applying these techniques to Hyperspectral Computed Tomography (HCT) data is that if the materials in the field of view are of similar composition then it can be difficult for a visualization of the hyperspectral image to differentiate between the materials. We propose novel alternative methods of preprocessing and summarizing HCT data in a single colorized image and novel measures to assess desired qualities in the resultant colored image, such as the contrast between different materials and the consistency of color within the same object. Proposed processes in this work include a new majority-voting method for multi-level thresholding, binary erosion, median filters, PAM clustering for grouping pixels into objects (of homogeneous materials) and mean/median assignment along the spectral dimension for representing the underlying signature, UMAP or GLMs to assign colors, and quantitative coloring assessment with developed measures. Strengths and weaknesses of various combinations of methods are discussed. These results have the potential to create more robust material identification methods from HCT data that has wide use in industrial, medical, and security-based applications for detection and quantification, including visualization methods to assist with rapid human interpretability of these complex hyperspectral signatures.

Special protection schemes (SPSs) safeguard the grid by detecting predefined abnormal conditions and deploying predefined corrective actions. Utilities leverage SPSs to maintain stability, acceptable voltages, and loading limits during disturbances. However, traditional SPSs cannot defend against unpredictable disturbances. Events such as cyber attacks, extreme weather, and electromagnetic pulses have unpredictable trajectories and require adaptive response. Therefore, we propose a harmonized automatic relay mitigation of nefarious intentional events (HARMONIE)-SPS that learns system conditions, mitigates cyber-physical consequences, and preserves grid operation during both predictable and unpredictable disturbances. In this paper, we define the HARMONIE-SPS approach, detail progress on its development, and provide initial results using a WSCC 9-bus system.

I/O performance in a multi-user environment is difficult to predict. Users do not know what I/O performance to expect when running and tuning applications. We propose to use the IO500 benchmark as a way to guide user expectations on their application's performance and to aid identifying root causes of their I/O problems that might come from the system. Our experiments describe how we manage user expectation with IO500 and provide a mechanism for system fault identification. This work also provides us with information of the tail latency problem that needs to be addressed and granular information about the impact of I/O technique choices (POSIX and MPI-IO).

In high-performance computing (HPC), scientific applications often manage a massive amount of data using I/O libraries. These libraries provide convenient data model abstractions, help ensure data portability, and, most important, empower end users to improve I/O performance by tuning configurations across multiple layers of the HPC I/O stack. We propose SCTuner, an autotuner integrated within the I/O library itself to dynamically tune both the I/O library and the underlying I/O stack at application runtime. To this end, we introduce a statistical benchmarking method to profile the behaviors of individual supercomputer I/O subsystems with varied configurations across I/O layers. We use the benchmarking results as the built-in knowledge in SCTuner, implement an I/O pattern extractor, and plan to implement an online performance tuner as the SCTuner runtime. We conducted a benchmarking analysis on the Summit supercomputer and its GPFS file system Alpine. The preliminary results show that our method can effectively extract the consistent I/O behaviors of the target system under production load, building the base for I/O autotuning at application runtime.

Conditional Point Sampling (CoPS) is a recently developed stochastic media transport algorithm that has demonstrated a high degree of accuracy in 1D and 3D simulations implemented for the CPU in Python. However, it is increasingly important that modern, production-level transport codes like CoPS be adapted for use on next-generation computing architectures. In this project, we describe the creation of a fast and accurate variant of CoPS implemented for the GPU in C++. As an initial test, we performed a code-to-code verification using single-history cohorts, which showed that the GPU implementation matched the original CPU implementation to within statistical uncertainty, while improving the speed by over a factor of 4000. We then tested the GPU implementation for cohorts up to size 64 and compared three variants of CoPS based on how the particle histories are grouped into cohorts: successive, simultaneous, and a successive-simultaneous hybrid. We examined the accuracy-efficiency tradeoff of each variant for 9 different benchmarks, measuring the reflectance and transmittance in a cubic geometry with reflecting boundary conditions on the four non-transmissive or reflective faces. Successive cohorts were found to be far more accurate than simultaneous cohorts for both reflectance (4.3 times) and transmittance (5.9 times), although simultaneous cohorts run more than twice as fast as successive cohorts, especially for larger cohorts. The hybrid cohorts demonstrated speed and accuracy behavior most similar to that of simultaneous cohorts. Overall, successive cohorts were found to be more suitable for the GPU due to their greater accuracy and reproducibility, although simultaneous and hybrid cohorts present an enticing prospect for future research.

Performance measures commonly used in systems security engineering tend to be static, linear, and have limited utility in addressing challenges to security performance from increasingly complex risk environments, adversary innovation, and disruptive technologies. Leveraging key concepts from resilience science offers an opportunity to advance next-generation systems security engineering to better describe the complexities, dynamism, and non-linearity observed in security performance—particularly in response to these challenges. This article introduces a multilayer network model and modified Continuous Time Markov Chain model that explicitly captures interdependencies in systems security engineering. The results and insights from a multilayer network model of security for a hypothetical nuclear power plant introduce how network-based metrics can incorporate resilience concepts into performance metrics for next generation systems security engineering.

Network modeling is a powerful tool to enable rapid analysis of complex systems that can be challenging to study directly using physical testing. Two approaches are considered: emulation and simulation. The former runs real software on virtualized hardware, while the latter mimics the behavior of network components and their interactions in software. Although emulation provides an accurate representation of physical networks, this approach alone cannot guarantee the characterization of the system under realistic operative conditions. Operative conditions for physical networks are often characterized by intrinsic variability (payload size, packet latency, etc.) or a lack of precise knowledge regarding the network configuration (bandwidth, delays, etc.); therefore uncertainty quantification (UQ) strategies should be also employed. UQ strategies require multiple evaluations of the system with a number of evaluation instances that roughly increases with the problem dimensionality, i.e., the number of uncertain parameters. It follows that a typical UQ workflow for network modeling based on emulation can easily become unattainable due to its prohibitive computational cost. In this paper, a multifidelity sampling approach is discussed and applied to network modeling problems. The main idea is to optimally fuse information coming from simulations, which are a low-fidelity version of the emulation problem of interest, in order to decrease the estimator variance. By reducing the estimator variance in a sampling approach it is usually possible to obtain more reliable statistics and therefore a more reliable system characterization. Several network problems of increasing difficulty are presented. For each of them, the performance of the multifidelity estimator is compared with respect to the single fidelity counterpart, namely, Monte Carlo sampling. For all the test problems studied in this work, the multifidelity estimator demonstrated an increased efficiency with respect to MC.

The US DOE Office of Electricity's Energy Storage Program's joint RD work with the Cordova Electric Cooperative (CEC) has deployed several advanced grid technologies that are providing benefits today to Cordova Alaska's electricity users. Advanced grid technologies deployed through DoE co-funded RD include a 1MW Battery Energy Storage System (BESS), and enhanced monitoring including Phasor Measurement Units (PMU's) to help better understand the operational impacts of the added BESS. This paper will highlight key accomplishments to-date in deploying and using advanced grid technologies, and then outline the next phase of work that will use these technologies to implement an operating scheme to reconfigure the utility's distribution system and utility resources including BESS to provide emergency back-up power to a critical load: The Cordova Community Medical Center (CCMC). This paper will include additional insights on the use of utility resources to support critical loads via case study examples by the National Rural Electric Cooperative Assoc. (NRECA).

Work on radiation transport in stochastic media has tended to focus on binary mixing with Markovian mixing statistics. However, although some real-world applications involve only two materials, others involve three or more. Therefore, we seek to provide a foundation for ongoing theoretical and numerical work with “N-ary” stochastic media comprised of discrete material phases with spatially homogenous Markovian mixing statistics. To accomplish this goal, we first describe a set of parameters and relationships that are useful to characterize such media. In doing so, we make a noteworthy observation: media that are frequently called Poisson media only comprise a subset of those that have Markovian mixing statistics. Since the concept of correlation length (as it has been used in stochastic media transport literature) and the hyperplane realization generation method are both tied to the Poisson property of the media, we argue that not all media with Markovian mixing statistics have a correlation length in this sense or are realizable with the traditional hyperplane generation method. Second, we describe methods for generating realizations of N-ary media with Markovian mixing. We generalize the chord- and hyperplane-based sampling methods from binary to N-ary mixing and propose a novel recursive hyperplane method that can generate a broader class of material structures than the traditional, non-recursive hyperplane method. Finally, we perform numerical studies that provide validation to the proposed N-ary relationships and generation methods in which statistical quantities are observed from realizations of ternary and quaternary media and are shown to agree with predicted values.

We extend the Monte Carlo Chord Length Sampling (CLS) and Local Realization Preserving (LRP) algorithms to the N-ary stochastic medium case using two recently developed uniform and volume fraction models that follow a Markov-chain process for N-ary problems in one-dimensional, Markovian-mixed media. We use the Lawrence Livermore National Laboratory Mercury Monte Carlo particle transport code to compute CLS and LRP reflection and transmission leakage values and material scalar flux distributions for one-dimensional, Markovian-mixed quaternary stochastic media based on the two N-ary stochastic medium models. We conduct accuracy comparisons against benchmark results produced with the Sandia National Laboratories PlaybookMC stochastic media transport research code. We show that CLS and LRP produce exact results for purely absorbing N-ary stochastic medium problems and find that LRP is generally more accurate than CLS for problems with scattering.

Sobol' sensitivity indices (SI) provide robust and accurate measures of how much uncertainty in output quantities is caused by different uncertain input parameters. These allow analysts to prioritize future work to either reduce or better quantify the effects of the most important uncertain parameters. One of the most common approaches to computing SI requires Monte Carlo (MC) sampling of uncertain parameters and full physics code runs to compute the response for each of these samples. In the case that the physics code is a MC radiation transport code, this traditional approach to computing SI presents a workflow in which the MC transport calculation must be sufficiently resolved for each MC uncertain parameter sample. This process can be prohibitively expensive, especially since thousands or more particle histories are often required on each of thousands or so uncertain parameter samples. We propose a process for computing SI in which only a few MC radiation transport histories are simulated before sampling new uncertain parameter values. We use Embedded Variance Deconvolution (EVADE) to parse the desired parametric variance from the MC transport variance on each uncertain parameter sample. To provide a relevant benchmark, we propose a new radiation transport benchmark problem and derive analytic solutions for its outputs, including SI. The new EVADE-based approach is found to converge with MC convergence behavior and be at least an order of magnitude more precise for the same computational cost than the traditional approach for several SI on our test problem.

In this paper, we introduce and analyze a new class of optimal control problems constrained by elliptic equations with uncertain fractional exponents. We utilize risk measures to formulate the resulting optimization problem. We develop a functional analytic framework, study the existence of solution, and rigorously derive the first-order optimality conditions. Additionally, we employ a sample-based approximation for the uncertain exponent and the finite element method to discretize in space. We prove the rate of convergence for the optimal risk neutral controls when using quadrature approximation for the uncertain exponent and conclude with illustrative examples.

Container escapes enable the adversary to execute code on the host from inside an isolated container. These high severity escape vulnerabilities originate from three sources: (1) container profile misconfigurations, (2) Linux kernel bugs, and (3) container runtime vulnerabilities. While the first two cases have been studied in the literature, no works have investigated the impact of container runtime vulnerabilities. In this paper, to fill this gap, we study 59 CVEs for 11 different container runtimes. As a result of our study, we found that five of the 11 runtimes had nine publicly available PoC container escape exploits covering 13 CVEs. Our further analysis revealed all nine exploits are the result of a host component leaked into the container. We apply a user namespace container defense to prevent the adversary from leveraging leaked host components and demonstrate that the defense stops seven of the nine container escape exploits.

Estimated parameters in Battery Energy Storage Systems (BESSs) may be vulnerable to cyber-attacks such as False Data Injection Attacks (FDIAs). FDIAs, which typically evade bad data detectors, could damage or degrade Battery Energy Storage Systems (BESSs). This paper will investigate methods to detect small magnitude FDIA using battery equivalent circuit models, an Extended Kalman Filter (EKF), and a Cumulative Sum (CUSUM) algorithm. A priori error residual data estimated by the EKF was used in the CUSUM algorithm to find the lowest detectable FDIA for this battery equivalent model. The algorithm described in this paper was able to detect attacks as low as 1 mV, with no false positives. The CUSUM algorithm was compared to a chi-squared based FDIA detector. In this study the CUSUM was found to detect attacks of smaller magnitudes than the conventional chi-squared detector.

This paper describes a co-simulation environment used to investigate how high penetrations of electric vehicles (EV s) impact a distribution feeder during a resilience event. As EV adoption and EV supply equipment (EVSE) technology advance, possible impacts to the electric grid increase. Additionally, as weather related resilience events become more common, the need to understand possible challenges associated with EV charging during such events becomes more important. Software designed to simulate vehicle travel patterns, EV charging characteristics, and the associated electric demand can be integrated with power system software using co-simulation to provide more realistic results. The work in progress described here will simulate varying EV loading and location over time to provide insights about EVSE characteristics for maximum benefit and allow for general sizing of possible micro grids to supply EVs and critical loads.

Advanced solar PV inverter control settings may not be reported to utilities or may be changed without notice. This paper develops an estimation method for determining a fixed power factor control setting of a behind-the-meter (BTM) solar PV smart inverter. The estimation is achieved using linear regression methods with historical net load advanced metering infrastructure (AMI) data. Notably, the BTM PV power factor setting may be unknown or uncertain to a distribution engineer, and cannot be trivially estimated from the historical AMI data due to the influence of the native load on the measurements. To solve this, we use a simple percentile-based approach for filtering the measurements. A physics-based linear sensitivity model is then used to determine the fixed power factor control setting from the sensitivity in the complex power plane. This sensitivity parameter characterizes the control setting hidden in the aggregate data. We compare several loss functions, and verify the models developed by conducting experiments on 250 datasets based on real smart meter data. The data are augmented with synthetic quasi-static-timeseries (QSTS) simulations of BTM PV that simulate utility-observed aggregate measurements at the load. The simulations demonstrate the reactive power sensitivity of a BTM PV smart inverter can be recovered efficiently from the net load data after applying the filtering approach.

Due to the existence of DC-DC converters, fast-tripping fault location in DC power systems is of particular importance to ensure the reliable operation of DC systems. Traveling wave (TW) protection is one of the promising approaches to accommodate fast detection and location of faults in DC systems. This paper proposes a numerical approach for a DC system fault location using the concept of TWs. The proposed approach is based on multiresolution analysis to calculate the TW signal's wavelet coefficients for different frequency ranges, and then, the Parseval theorem is used to calculate the energy of wavelet coefficients. A curve-fitting approach is used to find the best curve that fits the Parseval energy as a function of fault location for a set of curve-fitting datapoints. The identified Parseval energy curves are then utilized to estimate the fault location when a new fault is applied on a DC cable. A DC test system simulated in PSCAD/EMTDC is used to verify the performance of the proposed fault location algorithm.

This paper discusses a solar photovoltaic (PV) DC microgrid system consisting of a PV array, a battery, DC-DC converters, and a load, where all these elements are simulated in MATLAB/Simulink environment. The design and testing entail the functions of a boost converter and a bidirectional converter and how they work together to maintain stable control of the DC bus voltage and its energy management. Furthermore, the boost converter operates under Maximum Power Point Tracking (MPPT) settings to maximize the power that the PV array can output. The control algorithm can successfully maintain the output power of the PV array at its maximum point and can respond well to changes in input irradiance. This is shown in detail in the results section.

Once Synthetic Aperture Radar (SAR) images are formed, they typically need to be stored in some file format which might restrict the dynamic range of what can be represented. Thereafter, for exploitation by human observers, the images might need to be displayed in a manner to reveal the subtle scene reflectivity characteristics the observer seeks, which generally requires further manipulation of dynamic range. Proper image scaling, for both storage and for display, to maximize the perceived dynamic range of interest to an observer depends on many factors, and an understanding of underlying data characteristics. While SAR images are typically rendered with grayscale, or at least monochromatic intensity variations, color might also be usefully employed in some cases. We analyze these and other issues pertaining to SAR image scaling, dynamic range, radiometric calibration, and display.

Distribution system topology identification has historically been accomplished by unencrypting the information that is received from the smart meters and then running a topology identification algorithm. Unencrypted smart meter data introduces privacy and security issues for utility companies and their customers. This paper introduces security aware machine learning algorithms to alleviate the privacy and security issues raised with un-encrypted smart meter data. The security aware machine learning algorithms use the information received from the Advanced Metering Infrastructure (AMI) and identifies the distribution systems topology without unencrypting the AMI data by using fully homomorphic NTRU and CKKS encryption. The encrypted smart meter data is then used by Linear Discriminant Analysis, Convolution Neural Network, and Support Vector Machine algorithms to predict the distribution systems real time topology. This method can leverage noisy voltage magnitude readings from smart meters to accurately identify distribution system reconfiguration between radial topologies during operation under changing loads.

This paper presents a new method for modeling rarefied gas flows based on hybridization of direct simulation Monte Carlo (DSMC) and discrete velocity method (DVM)-based quasi-particle representations of the velocity distribution function. It is aimed at improving the resolution of the tails of the distribution function (compared with DSMC) and computational efficiency (compared with DVM). Details of the method, such as the collision algorithm and the particle merging scheme, are discussed. The hybrid approach is applied to the study of noise in a Maxwellian distribution, computation of electron-impact ionization rate coefficient, as well as numerical simulation of a supersonic Couette flow. The hybrid-based solver is compared with pure DSMC and DVM approaches in terms of accuracy, computational speed, and memory use. It is shown that such a hybrid approach can provide a lower computational cost than a pure DVM approach, while being able to retain accuracy in modeling high-velocity tails of the distribution function. For problems where trace species have a significant impact on the flow physics, the proposed method is shown to be capable of providing better computational efficiency and accuracy compared with standard fixed-weight DSMC.

The American distributed energy resource (DER) interconnection standard, IEEE Std. 1547, was updated in 2018 to include standardized interoperability functionality. As state regulators begin ratifying these requirements, all DER - such as photovoltaic (PV) inverters, energy storage systems (ESSs), and synchronous generators - in those jurisdictions must include a standardized SunSpec Modbus, IEEE 2030.5, or IEEE 1815 (DNP3) communication interface. Utilities and authorized third parties will interact with these DER interfaces to read nameplate information, power measurements, and alarms as well as configure the DER settings and grid-support functionality. In 2020, the certification standard IEEE 1547.1 was revised with test procedures for evaluating the IEEE 1547-2018 interoperability requirements. In this work, we present an open-source framework to evaluate DER interoperability. To demonstrate this capability, we used four test devices: a SunSpec DER Simulator with a SunSpec Modbus interface, an EPRI-developed DER simulator with an IEEE 1815 interface, a Kitu Systems DER simulator with an IEEE 2030.5 interface, and an EPRI IEEE 2030.5-to-Modbus converter. By making this test platform openly available, DER vendors can validate their implementations, utilities can spot check communications to DER equipment, certification laboratories can conduct type testing, and research institutions can more easily research DER interoperability and cybersecurity. We indicate several limitations and ambiguities in the communication protocols, information models, and the IEEE 1547.1-2020 test protocol which were exposed in these evaluations in anticipation that the standards-development organizations will address these issues in the future.

Partitioned point-To-point communication and persistent collective communication were both recently standardized in MPI-4.0. Each offers performance and scalability advantages over MPI-3.1-based communication when planned transfers are feasible in an MPI application. Their merger into a generalized, persistent collective communication with partitions is a logical next step, with significant advantages for performance portability. Non-Trivial decisions about the syntax and semantics of such operations need to be addressed, including scope of knowledge of partitioning choices by members of the communicator's group(s). This paper introduces and motivates proposed interfaces for partitioned collective communication. Partitioned collectives will be particularly useful for multithreaded, accelerator-offloaded, and/or hardware-collective-enhanced MPI implementations driving suitable applications, as well as for pipelined collective communication (e.g., partitioned allreduce) with single consumers and producers per MPI process. These operations also provide load imbalance mitigation. Halo exchange codes arising from regular and irregular grid/mesh applications are a key candidate class of applications for this functionality. Generalizations of lightweight notification procedures MPI-Parrived and MPI-Pready are considered. Generalization of MPIX-Pbuf-prepare, a procedure proposed for MPI-4.1 for point-To-point partitioned communication, are also considered, shown in context of supporting ready-mode send semantics for the operations. The option of providing local and incomplete modes for initialization procedures is mentioned (which could also apply to persistent collective operations); these semantics interact with the MPIX-Pbuf-prepare concept and the progress rule. Last, future work is outlined, indicating prerequisites for formal consideration for the MPI-5 standard.

Compression analytics have gained recent interest for application in malware classification and digital forensics. This interest is due to the fact that compression analytics rely on measured similarity between byte sequences in datasets without requiring prior feature extraction; in other words, these methods are featureless. Being featureless makes compression analytics particularly appealing for computer security applications, where good static features are either unknown or easy to circumvent by adversaries. However, previous classification methods based on compression analytics relied on algorithms that scaled with the size of each labeled class and the number of classes. In this work, we introduce an approach that, in addition to being featureless, can perform fast and accurate inference that is independent of the size of each labeled class. Our method is based on calculating a representative sample, the Fréchet mean, for each labeled class and using it at inference time. We introduce a greedy algorithm for calculating the Fréchet mean and evaluate its utility for classification across a variety of computer security applications, including authorship attribution of source code, file fragment type detection, and malware classification.

Boolean functions and binary arithmetic operations are central to standard computing paradigms. Accordingly, many advances in computing have focused upon how to make these operations more efficient as well as exploring what they can compute. To best leverage the advantages of novel computing paradigms it is important to consider what unique computing approaches they offer. However, for any special-purpose co-processor, Boolean functions and binary arithmetic operations are useful for, among other things, avoiding unnecessary I/O on-and-off the co-processor by pre- and post-processing data on-device. This is especially true for spiking neuromorphic architectures where these basic operations are not fundamental low-level operations. Instead, these functions require specific implementation. Here we discuss the implications of an advantageous streaming binary encoding method as well as a handful of circuits designed to exactly compute elementary Boolean and binary operations.

As climate change and human migration accelerate globally, decision-makers are seeking tools that can deepen their understanding of the complex nexus between climate change and human migration. These tools can help to identify populations under pressure to migrate, and to explore proactive policy options and adaptive measures. Given the complexity of factors influencing migration, this article presents a system dynamics-based model that couples migration decision making and behavior with the interacting dynamics of economy, labor, population, violence, governance, water, food, and disease. The regional model is applied here to the test case of migration within and beyond Mali. The study explores potential systems impacts of a range of proactive policy solutions and shows that improving the effectiveness of governance and increasing foreign aid to urban areas have the highest potential of those investigated to reduce the necessity to migrate in the face of climate change.

We present recent results toward the quantification of spray characteristics at engine conditions for an eight-hole counter-bored (stepped) GDI injector – Spray G in the ECN denomination. This computational study is characterized by two novel features: the detailed description of a real injector's internal surfaces via tomographic reconstruction; and a general equation of state that represents the thermodynamic properties of homogeneous liquid-vapor mixtures. The combined level-set moment-of-fluid approach, coupled to an embedded boundary formulation for moving solid walls, makes it possible to seamlessly connect the injector's internal flow to the spray. The Large Eddy Simulation (LES) discussed here presents evidence of partial hydraulic flipping and, during the closing transient, string cavitation. Results are validated by measurements of spray density profiles and droplet size distribution.

The next generation of supercomputing resources is expected to greatly expand the scope of HPC environments, both in terms of more diverse workloads and user bases, as well as the integration of edge computing infrastructures. This will likely require new mechanisms and approaches at the Operating System level to support these broader classes of workloads along with their different security requirements. We claim that a key mechanism needed for these workloads is the ability to securely compartmentalize the system software executing on a given node. In this paper, we present initial efforts in exploring the integration of secure and trusted computing capabilities into an HPC system software stack. As part of this work we have ported the Kitten Lightweight Kernel (LWK) to the ARM64 architecture and integrated it with the Hafnium hypervisor, a reference implementation of a secure partition manager (SPM) that provides security isolation for virtual machines. By integrating Kitten with Hafnium, we are able to replace the commodity oriented Linux based resource management infrastructure and reduce the overheads introduced by using a full weight kernel (FWK) as the node-level resource scheduler. While our results are very preliminary, we are able to demonstrate measurable performance improvements on small scale ARM based SOC platforms.

We describe efforts in generating synthetic malware samples that have specified behaviors that can then be used to train a machine learning (ML) algorithm to detect behaviors in malware. The idea behind detecting behaviors is that a set of core behaviors exists that are often shared in many malware variants and that being able to detect behaviors will improve the detection of novel malware. However, empirically the multi-label task of detecting behaviors is significantly more difficult than malware classification, only achieving on average 84% accuracy across all behaviors as opposed to the greater than 95% multi-class or binary accuracy reported in many malware detection studies. One of the difficulties in identifying behaviors is that while there are ample malware samples, most data sources do not include behavioral labels, which means that generally there is insufficient training data for behavior identification. Inspired by the success of generative models in improving image processing techniques, we examine and extend a 1) conditional variational auto-encoder and 2) a flow-based generative model for malware generation with behavior labels. Initial experiments indicate that synthetic data is able to capture behavioral information and increase the recall of behaviors in novel malware from 32% to 45% without increasing false positives and to 52% with increased false positives.

Many individuals' mobility can be characterized by strong patterns of regular movements and is influenced by social relationships. Social networks are also often organized into overlapping communities which are associated in time or space. We develop a model that can generate the structure of a social network and attribute purpose to individuals' movements, based solely on records of individuals' locations over time. This model distinguishes the attributed purpose of check-ins based on temporal and spatial patterns in check-in data. Because a location-based social network dataset with authoritative ground-truth to test our entire model does not exist, we generate large scale datasets containing social networks and individual check-in data to test our model. We find that our model reliably assigns community purpose to social check-in data, and is robust over a variety of different situations.

This work reports an on-wafer study of avalanche behavior and failure analysis of in-house fabricated 1.3 kV GaN-on-GaN P-N diodes. DC breakdown is measured at different temperatures to confirm avalanche behavior. Diode's avalanche ruggedness is measured directly on-wafer using a modified unclamped inductive switching (UIS) test set-up with an integrated thermal chuck and high-speed CCD for real-time imaging during the test. The avalanche ruggedness of the GaN P-N diode is evaluated and compared with a commercial SiC Schottky diode of similar voltage and current rating. Failure analysis is done using SEM and optical microscopy to gain insight into the diode's failure mechanism during avalanche operation.

We propose primal-dual mesh optimization algorithms that overcome shortcomings of the standard algorithm while retaining some of its desirable features. “Hodge-Optimized Triangulations” defines the “HOT energy” as a bound on the discretization error of the diagonalized Delaunay Hodge star operator. HOT energy is a natural choice for an objective function, but unstable for both mathematical and algorithmic reasons: it has minima for collapsed edges, and its extrapolation to non-regular triangulations is inaccurate and has unbounded minima. We propose a different extrapolation with a stronger theoretical foundation. We propose new objectives, based on normalizations of the HOT energy, with barriers to edge collapses and other undesirable configurations. We propose mesh improvement algorithms coupling these. When HOT optimization nearly collapses an edge, we actually collapse the edge. Otherwise, we use the barrier objective to update positions and weights. By combining discrete connectivity changes with continuous optimization, we more fully explore the space of possible meshes and obtain higher quality solutions.

The reactions of carbonyl oxide Criegee intermediates with acids proceed predominantly by an insertion mechanism. We characterise the products from one of the simplest reactions of carbonyl oxides with inorganic acids, CH2OO + hydrogen chloride, which occurs via a 1,2-insertion in the H–Cl bond. Reactions of both HCl and DCl isotopologues yield product signal at the mass of the insertion product chloro(hydroperoxy)methane and a dissociative ionisation peak at the mass of the protonated (or deuteronated) Criegee intermediate. The isotopic composition of the insertion product has been measured for reaction mixtures where both HCl isotopologues are present, and the H/D ratio of the product is consistently higher (by a factor of 1.6 ± 0.3) than that of the reactants. This isotope selectivity in the products has smaller uncertainty than the ratio of measured rate coefficients and suggests a normal (k H > k D) kinetic isotope effect in the reaction. Theoretical kinetics calculations predict a small normal kinetic isotope effect for the overall reaction (k H / k D = 1.35 at 20 Torr N2 and k H / k D = 1.2 at 1 atm N2) but predict a substantial inverse kinetic isotope effect (k D > k H) for the stabilisation fraction, in disagreement with the experimental observation.

Global thinning of integrated circuits is a technique that enables backside failure analysis and radiation testing. Prior work also shows increased thresholds for single-event latchup and upset in thinned devices. We present impacts of global thinning on device performance and reliability of 28 nm node field programmable gate arrays (FPGA). Devices are thinned to values of 50, 10, and 3 microns using a micromachining and polishing method. Lattice damage, in the form of dislocations, extend about 1 micron below the machined surface. The damage layer is removed after polishing with colloidal SiO2 slurry. We create a 2D finite-element model with liner elasticity equations and flip-chip packaged device geometry to show that thinning increases compressive global stress in the Si, while C4 bumps increase stress locally. Measurements of stress using Raman spectroscopy qualitatively agree with our stress model but also reveal the need for more complex structural models to account for nonlinear effects occurring in devices thinned to 3 microns and after temperature cycling to 125 °C. Thermal imaging shows that increased local heating occurs with increased thinning but the maximum temperature difference across the 3-micron die is less than 2 °C. Ring oscillators (ROs) programmed throughout the FPGA fabric slow about 0.5% after thinning compared to full thickness values. Temperature cycling the devices to 125 °C further decreases RO frequency about 0.5%, which we attribute to stress changes in the Si.

In this work, we analyze the second and third harmonic signal from a dielectric metasurface in conjunction with polarization selection rules to unambiguously demonstrate the occurrence of cascaded second-order nonlinearities.

Abstract not provided.

We present an empirical methodology for thermally characterizing and determining absorption and scattering losses in released ring whisper gallery mode optical resonators. We used the methodology to deduce absorption and scattering contributions in Q = 308,000 silicon nitride resonators coupled to on-chip waveguides.

We present an optical wavelength division multiplexer enabled by a ring resonator tuned by MEMS electrostatic actuation. Analytical analysis, simulation and fabrication are discussed leading to results showing controlled tuning greater than one FSR.