SEARCH · Engineering Papers

Results for “Distributed model predictive control”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 757 records · Page 42

All sources NASA NTRS DOE OSTI

DOE OSTI2018

Mechanisms for Methane Transport and Hydrate Accumulation in Coarse-Grained Reservoirs

The distribution of natural methane hydrate within high saturation reservoirs and mechanisms that control the migration of methane are an active area of research. In this project, we studied methane migration mechanisms and associated hydrate accumulation rates in coarse-grained sands of the Terrebonne Basin, located in Walker Ridge Block 313 (WR313), northern Gulf of Mexico. Hydrate in this area is distributed heterogeneously within ~900 m of methane hydrate stability zone, in both thick (10-25 m) and thin (< 3 m) sand layers, and in units of subvertical hydrate-filled fractures. We investigated hydrate formation from diffusively and advectively supplied methane using one-, two-, and three-dimensional basin modeling with inputs from well log and seismic data. The hydrate accumulations at WR313 can mostly be explained by short, diffusive migration of methane as well as short-range advective transport of methane from clays into neighboring coarser-grained layers. This is likely enhanced by small-scale lithologic variations, including local microbial methanogenesis in interbedded clays. Furthermore, our work shows overpressure in clay sediments surrounding sand reservoirs may enhance short-range diffusive migration. Long-range migration of dissolved methane and gas mainly influence hydrate saturation right at the base of the hydrate stability zone, although the dissolved phase can be channeled upwards along sands by compaction-driven flow. Capillary effects can allow gas bubbles to migrate appreciable distances above the predicted base of gas hydrate stability, possibly reaching a few tens of meters into the hydrate stability zone. Recycling methane from hydrate buried beneath the base of the hydrate stability zone can concentrate hydrate at large saturations within a few tens of meters of the base of hydrate stability.

58 GEOSCIENCES↗

DOE OSTI2021

Topographical Controls on Hillslope-Scale Hydrology Drive Shrub Distributions on the Seward Peninsula, Alaska

Observations indicate shrubs are expanding across the Arctic tundra, mainly on hillslopes and primarily in response to climate warming. However, the impact topography exerts on hydrology, nutrient dynamics, and plant growth can make untangling the mechanisms behind shrub expansion difficult. We examined the role topography plays in determining shrub expansion by applying a coupled transect version of a mechanistic ecosystem model (ecosys) in a tundra hillslope site in the Seward Peninsula, Alaska. Modeled biomass of the dominant plant functional types agreed well with field measurements (R 2 = 0.89) and accurately represented shrub expansion over the past 30 years inferred from satellite observations. In the well-drained crest position, canopy water potential and plant nitrogen (N) uptake was modeled to be low from plant and microbial water stress. Intermediate soil water content in the mid-slope position enhanced mineralization and plant N uptake, increasing shrub biomass. The deciduous shrub growth in the mid-slope position was further enhanced by symbiotic N 2 fixation primed by increased root carbon allocation. The gentle slope in the poorly drained lower-slope position resulted in saturated soil conditions that reduced soil O 2 concentrations, leading to lower root O 2 uptake and lower nutrient uptake and plant biomass. Additionally, a simulation that removed topographical interconnectivity between grid cells resulted in (1) a 28% underestimate of mean shrub biomass and (2) over or underestimated shrub productivity at the various hillslope positions. Our results indicate that land models need to account for hillslope-scale coupled surface and subsurface hydrology to accurately predict current plant distributions and future trajectories in Arctic ecosystems.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2016

A spatially distributed model for assessment of the effects of changing land use and climate on urban stream quality: Development of a Spatially Distributed Urban Water Quality Model

While the effects of land use change in urban areas have been widely examined, the combined effects of climate and land use change on the quality of urban and urbanizing streams have received much less attention. We describe a modeling framework that is applicable to the evaluation of potential changes in urban water quality and associated hydrologic changes in response to ongoing climate and landscape alteration. The grid-based spatially distributed model, DHSVM-WQ, is an outgrowth of the Distributed Hydrology-Soil-Vegetation Model (DHSVM) that incorporates modules for assessing hydrology and water quality in urbanized watersheds at a high spatial and temporal resolution. DHSVM-WQ simulates surface runoff quality and in-stream processes that control the transport of nonpoint-source (NPS) pollutants into urban streams. We configure DHSVM-WQ for three partially urbanized catchments in the Puget Sound region to evaluate the water quality responses to current conditions and projected changes in climate and/or land use over the next century. Here we focus on total suspended solids (TSS) and total phosphorus (TP) from nonpoint sources (runoff), as well as stream temperature. The projection of future land use is characterized by a combination of densification in existing urban or partially urban areas, and expansion of the urban footprint. The climate change scenarios consist of individual and concurrent changes in temperature and precipitation. Future precipitation is projected to increase in winter and decrease in summer, while future temperature is projected to increase throughout the year. Our results show that urbanization has a much greater effect than climate change on both the magnitude and seasonal variability of streamflow, TSS and TP loads largely due to substantially increased streamflow, and particularly winter flow peaks. Water temperature is more sensitive to climate warming scenarios than to urbanization and precipitation changes. Future urbanization and climate change together are predicted to significantly increase annual mean streamflow (up to 55%), water temperature (up to 1.9 ºC), TSS load (up to 182%), and TP load (up to 74%).

Sun, Ning↗

DOE OSTI2015

Novel Sensor for the In Situ Measurement of Uranium Fluxes

The goal of this project was to develop a sensor that incorporates the field-tested concepts of the passive flux meter to provide direct in situ measures of flux for uranium and groundwater in porous media. Measurable contaminant fluxes [J] are essentially the product of concentration [C] and groundwater flux or specific discharge [q ]. The sensor measures [J] and [q] by changes in contaminant and tracer amounts respectively on a sorbent. By using measurement rather than inference from static parameters, the sensor can directly advance conceptual and computational models for field scale simulations. The sensor was deployed in conjunction with DOE in obtaining field-scale quantification of subsurface processes affecting uranium transport (e.g., advection) and transformation (e.g., uranium attenuation) at the Rifle IFRC Site in Rifle, Colorado. Project results have expanded our current understanding of how field-scale spatial variations in fluxes of uranium, groundwater and salient electron donor/acceptors are coupled to spatial variations in measured microbial biomass/community composition, effective field-scale uranium mass balances, attenuation, and stability. The coupling between uranium, various nutrients and micro flora can be used to estimate field-scale rates of uranium attenuation and field-scale transitions in microbial communities. This research focuses on uranium (VI), but the sensor principles and design are applicable to field-scale fate and transport of other radionuclides. Laboratory studies focused on sorbent selection and calibration, along with sensor development and validation under controlled conditions. Field studies were conducted at the Rifle IFRC Site in Rifle, Colorado. These studies were closely coordinated with existing SBR (formerly ERSP) projects to complement data collection. Small field tests were conducted during the first two years that focused on evaluating field-scale deployment procedures and validating sensor performance under controlled field conditions. In the third and fourth year a suite of larger field studies were conducted. For these studies, the uranium flux sensor was used with uranium speciation measurements and molecular-biological tools to characterize microbial community and active biomass at synonymous wells distributed in a large grid. These field efforts quantified spatial changes in uranium flux and field-scale rates of uranium attenuation (ambient and stimulated), uranium stability, and quantitatively assessed how fluxes and effective reaction rates were coupled to spatial variations in microbial community and active biomass. Analyses of data from these field experiments were used to generate estimates of Monod kinetic parameters that are ‘effective’ in nature and optimal for modeling uranium fate and transport at the field-scale. This project provided the opportunity to develop the first sensor that provides direct measures of both uranium (VI) and groundwater flux. A multidisciplinary team was assembled to include two geochemists, a microbiologist, and two quantitative contaminant hydrologists. Now that the project is complete, the sensor can be deployed at DOE sites to evaluate field-scale uranium attenuation, source behavior, the efficacy of remediation, and off-site risk. Because the sensor requires no power, it can be deployed at remote sites for periods of days to months. The fundamental science derived from this project can be used to advance the development of predictive models for various transport and attenuation processes in aquifers. Proper development of these models is critical for long-term stewardship of contaminated sites in the context of predicting uranium source behavior, remediation performance, and off-site risk.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2012

NEAMS Software Licensing, Release, and Distribution: Implications for FY2013 Work Package Planning

The vision of the NEAMS program is to bring truly predictive modeling and simulation (M&S) capabilities to the nuclear engineering community in order to enable a new approach to the analysis of nuclear systems. NEAMS anticipates issuing in FY 2018 a full release of its computational 'Fermi Toolkit' aimed at advanced reactor and fuel cycles. The NEAMS toolkit involves extensive software development activities, some of which have already been underway for several years, however, the Advanced Modeling and Simulation Office (AMSO), which sponsors the NEAMS program, has not yet issued any official guidance regarding software licensing, release, and distribution policies. This motivated an FY12 task in the Capability Transfer work package to develop and recommend an appropriate set of policies. The current preliminary report is intended to provide awareness of issues with implications for work package planning for FY13. We anticipate a small amount of effort associated with putting into place formal licenses and contributor agreements for NEAMS software which doesn't already have them. We do not anticipate any additional effort or costs associated with software release procedures or schedules beyond those dictated by the quality expectations for the software. The largest potential costs we anticipate would be associated with the setup and maintenance of shared code repositories for development and early access to NEAMS software products. We also anticipate an opportunity, with modest associated costs, to work with the Radiation Safety Information Computational Center (RSICC) to clarify export control assessment policies for software under development.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

DOE OSTI2017

Adaptive RAS/SPS System Settings for Improving Grid Reliability and Asset Utilization through Predictive Simulation and Controls

Remedial Action Schemes (RAS) or Special Protection Schemes (SPS) are used throughout the bulk transmission system as a non-wires method of increasing transmission transfer capability. However, as the population grows within the Western Interconnection, more power needs to be transferred with less ability to build lines. This leads to a greater number of RAS, and the subsequent interaction between these disparate schemes can have unintended consequences. While the large number of RAS within the Western Electricity Coordinating Council (WECC) protect for different things, which can range from the violation of thermal limitations to transient stability issues, they can be described by common methods of operation. In order to understand these methods, a survey was sent out to members of the WECC Modeling & Validation Working Group (MVWG). The response received provides a necessary base in which to understand and build models that can be used within WECC. This report provides information on the survey results regarding RAS/SPS modeling practices for planning and operational studies within the WECC footprint, which is foundational for on-going and future work on RAS. Pacific Northwest National Laboratory (PNNL) and its industry partners are developing innovative methods for adaptively setting RAS parameters based on realistic and near real-time operational conditions to improve power grid reliability and grid asset utilization.

24 POWER TRANSMISSION AND DISTRIBUTION↗

DOE OSTI2013

DARK MATTER HALO PROFILES OF MASSIVE CLUSTERS: THEORY VERSUS OBSERVATIONS

Dark-matter-dominated cluster-scale halos act as an important cosmological probe and provide a key testing ground for structure formation theory. Focusing on their mass profiles, we have carried out (gravity-only) simulations of the concordance {Lambda}CDM cosmology, covering a mass range of 2 Multiplication-Sign 10{sup 12} to 2 Multiplication-Sign 10{sup 15} h {sup -1} M{sub Sun} and a redshift range of z = 0-2, while satisfying the associated requirements of resolution and statistical control. When fitting to the Navarro-Frenk-White profile, our concentration-mass (c-M) relation differs in normalization and shape in comparison to previous studies that have limited statistics in the upper end of the mass range. We show that the flattening of the c-M relation with redshift is naturally expressed if c is viewed as a function of the peak height parameter, {nu}. Unlike the c-M relation, the slope of the c-{nu} relation is effectively constant over the redshift range z = 0-2, while the amplitude varies by {approx}30% for massive clusters. This relation is, however, not universal: using a simulation suite covering the allowed wCDM parameter space, we show that the c-{nu} relation varies by about {+-}20% as cosmological parameters are varied. At fixed mass, the c(M) distribution is well fit by a Gaussian with {sigma}{sub c}/(c) {approx_equal} 1/3, independent of the radius at which the concentration is defined, the halo dynamical state, and the underlying cosmology. We compare the {Lambda}CDM predictions with observations of halo concentrations from strong lensing, weak lensing, galaxy kinematics, and X-ray data, finding good agreement for massive clusters (M{sub vir} > 4 Multiplication-Sign 10{sup 14} h {sup -1} M{sub Sun }), but with some disagreements at lower masses. Because of uncertainty in observational systematics and modeling of baryonic physics, the significance of these discrepancies remains unclear.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2014

Geoelectrical Measurement of Multi-Scale Mass Transfer Parameters

Mass transfer affects contaminant transport and is thought to control the efficiency of aquifer remediation at a number of sites within the Department of Energy (DOE) complex. An improved understanding of mass transfer is critical to meeting the enormous scientific and engineering challenges currently facing DOE. Informed design of site remedies and long-term stewardship of radionuclide-contaminated sites will require new cost-effective laboratory and field techniques to measure the parameters controlling mass transfer spatially and across a range of scales. In this project, we sought to capitalize on the geophysical signatures of mass transfer. Previous numerical modeling and pilot-scale field experiments suggested that mass transfer produces a geoelectrical signature—a hysteretic relation between sampled (mobile-domain) fluid conductivity and bulk (mobile + immobile) conductivity—over a range of scales relevant to aquifer remediation. In this work, we investigated the geoelectrical signature of mass transfer during tracer transport in a series of controlled experiments to determine the operation of controlling parameters, and also investigated the use of complex-resistivity (CR) as a means of quantifying mass transfer parameters in situ without tracer experiments. In an add-on component to our grant, we additionally considered nuclear magnetic resonance (NMR) to help parse mobile from immobile porosities. Including the NMR component, our revised study objectives were to: 1. Develop and demonstrate geophysical approaches to measure mass-transfer parameters spatially and over a range of scales, including the combination of electrical resistivity monitoring, tracer tests, complex resistivity, nuclear magnetic resonance, and materials characterization; and 2. Provide mass-transfer estimates for improved understanding of contaminant fate and transport at DOE sites, such as uranium transport at the Hanford 300 Area. To achieve our objectives, we implemented a 3-part research plan involving (1) development of computer codes and techniques to estimate mass-transfer parameters from time-lapse electrical data; (2) bench-scale experiments on synthetic materials and materials from cores from the Hanford 300 Area; and (3) field demonstration experiments at the DOE’s Hanford 300 Area. In a synergistic add-on to our workplan, we analyzed data from field experiments performed at the DOE Naturita Site under a separate DOE SBR grant, on which PI Day-Lewis served as co-PI. Techniques developed for application to Hanford datasets also were applied to data from Naturita. 1. Introduction The Department of Energy (DOE) faces enormous scientific and engineering challenges associated with the remediation of legacy contamination at former nuclear weapons production facilities. Selection, design and optimization of appropriate site remedies (e.g., pump-and-treat, biostimulation, or monitored natural attenuation) requires reliable predictive models of radionuclide fate and transport; however, our current modeling capabilities are limited by an incomplete understanding of multi-scale mass transfer—its rates, scales, and the heterogeneity of controlling parameters. At many DOE sites, long “tailing” behavior, concentration rebound, and slower-than-expected cleanup are observed; these observations are all consistent with multi-scale mass transfer [Haggerty and Gorelick, 1995; Haggerty et al., 2000; 2004], which renders pump-and-treat remediation and biotransformation inefficient and slow [Haggerty and Gorelick, 1994; Harvey et al., 1994; Wilson, 1997]. Despite the importance of mass transfer, there are significant uncertainties associated with controlling parameters, and the prevalence of mass transfer remains a point of debate [e.g., Hill et al., 2006; Molz et al., 2006] for lack of experimental methods to verify and measure it in situ or independently of tracer breakthrough. There is a critical need for new field-experimental techniques to measure mass transfer in-situ and estimate multi-scale and spatially variable mass-transfer parameters. The current lack of such techniques results in large parameter uncertainty, which in turn translates into enormous prediction uncertainty and cost to DOE. In this project, we considered three hydrogeophysical approaches for providing information about mass-transfer parameters: (1) the combination of electrical-resistivity tomography (ERT) and ionic tracer experiments to explore rates of exchange and relative mobile and immobile porosities; (2) complex resistivity (CR) measurements to infer the distribution of diffusive length scales active in a porous medium; and (3) nuclear magnetic resonance (NMR) to estimate mobile and immobile porosity.

38 RADIATION CHEMISTRY, RADIOCHEMISTRY, AND NUCLEA↗

DOE OSTI2022

Random Forest Regressor-Based Approach for Detecting Fault Location and Duration in Power Systems

Power system failures or outages due to short-circuits or “faults” can result in long service interruptions leading to significant socio-economic consequences. It is critical for electrical utilities to quickly ascertain fault characteristics, including location, type, and duration, to reduce the service time of an outage. Existing fault detection mechanisms (relays and digital fault recorders) are slow to communicate the fault characteristics upstream to the substations and control centers for action to be taken quickly. Fortunately, due to availability of high-resolution phasor measurement units (PMUs), more event-driven solutions can be captured in real time. In this paper, we propose a data-driven approach for determining fault characteristics using samples of fault trajectories. A random forest regressor (RFR)-based model is used to detect real-time fault location and its duration simultaneously. This model is based on combining multiple uncorrelated trees with state-of-the-art boosting and aggregating techniques in order to obtain robust generalizations and greater accuracy without overfitting or underfitting. Four cases were studied to evaluate the performance of RFR: 1. Detecting fault location (case 1), 2. Predicting fault duration (case 2), 3. Handling missing data (case 3), and 4. Identifying fault location and length in a real-time streaming environment (case 4). A comparative analysis was conducted between the RFR algorithm and state-of-the-art models, including deep neural network, Hoeffding tree, neural network, support vector machine, decision tree, naive Bayesian, and K-nearest neighborhood. Experiments revealed that RFR consistently outperformed the other models in detection accuracy, prediction error, and processing time.

24 POWER TRANSMISSION AND DISTRIBUTION↗

DOE OSTI2026

Measurements of TRACER pre-convective conditions and mesoscale circulations using small unmanned aircraft systems (sUAS)

Improved comprehension of the physical processes governing convective cloud formation and lifecycle are of critical importance for understanding and predicting future climate states. The influence of these clouds on the planetary energy budget, including on precipitation, is significant. Things are particularly complex in coastal regimes, where gradients in aerosol particle properties, localized circulations such as sea breezes, and large population centers are found. To date, numerical models struggle to accurately represent these critical clouds and are therefore challenged to provide a realistic view on the planetary energy budget. Through the proposed research, we deployed two uncrewed aircraft systems (UAS) equipped with a variety of instruments alongside sensors deployed by the US Department of Energy Atmospheric Radiation Measurement (ARM) program for the TRACER (Tracking Aerosol Convection Interactions Experiment) field campaign. The two small UAS platforms consisted of a CU RAAVEN fixed-wing airplane and an OU CopterSonde system, with the copter collecting frequent vertical profiles of thermodynamic and kinematic variables such as temperature, pressure, wind and humidity. At the same time, the fixed-wing captured horizontal gradients of these quantities and aerosol size distribution. These systems were deployed south of the Houston metro area, in an area that is impacted by the Gulf of Mexico sea breeze on a daily basis. These observations offer enhanced and complementary perspectives to those provided by the DOE ARM Mobile Facility (AMF) which is was deployed in southeast Houston, and an ancillary site in a more rural location west of the urban Houston area. Quality-controlled versions of the UAS data were collected and posted on the DOE ARM data archive after the conclusion of the campaign where they are accessible by the research community and general public. The UAS perspective offers revolutionary insight into key spatial and temporal effects that have not been evaluated previously.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2018

Initial Core Physics and Safety Analyses Supporting the Versatile Test Reactor

The Versatile Test Reactor (VTR) is a proposed integrated facility for investigations into the behavior of fuels and materials subjected to high, fast neutron fluxes prototypical of fast reactor environments. Detailed design of the reactor is currently being performed by a group of U.S. National Laboratories, including Argonne National Laboratory. The current design of the reactor is a 300 MW sodium-cooled pool-type fast reactor fueled with ternary U-20Pu-10Zr metallic fuel. This report documents reactor core physics and preliminary safety analyses that have been performed. The reference VTR core has 313 assembly positions, 66 of which contain fuel assemblies. The remaining assembly positions are loaded with reflector, shield, control rod duct, and test assemblies. Ten positions are available in the reference design for experimental or test assemblies, however the VTR will support test assemblies in any position in the core for alternate core loadings. The performance of the reference VTR core has been assessed for an assumed equilibrium cycle. The core power distribution and reactivity feedback coefficients have been calculated. Preliminary assessments of the reactivity coefficients have shown that each of them results in a negative reactivity change for increasing temperatures. Safety analysis of a VTR layout has been performed using the SAS4A/SASSYS-1 fast reactor safety analysis code. A model has been developed to represent the current core design and a reference heat transport system design. The model has been used to simulate and perform analysis of protected and unprotected loss of heat sink, station blackout, and transient overpower accident scenarios. These scenarios represent the three main ways to perturb the reactor core, through changes to the core inlet temperature, mass flow rate, or reactivity insertions and are credible bounding scenarios for the evaluated transient categories. The results of these simulations have been assessed against simplified criteria for Anticipated Operational Occurrences for the protected transients and for Beyond Design Basis Events for the unprotected transients. Because the primary heat transport system is able to transition quickly and effectively to natural circulation and because the shutdown heat removal system provides sufficient heat rejection capability, large margins for all criteria were predicted for all protected transients. Large margins were also predicted for the unprotected loss of heat sink and station blackout transients due to the strong negative reactivity feedbacks generated by the core. Because of the high power density in the reactor, smaller fuel melting margins are predicted for the unprotected transient overpower accident.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

DOE OSTI2022

DC-Link Current Minimization Control for Current Source Converter-Based Solid-State Transformer

This article proposes a fast predictive control method and a small DC-link inductor to minimize the DC-link current in current-source converter (CSC)-based solid-state transformer. The DC-link current minimization can significantly reduce power loss and improve efficiency. The challenge of this problem is on improving both steady-state and dynamic performance. PI control methods and large DC-link inductors are conventionally used in the CSC but have limited dynamic performance. A model predictive control (MPC) method is proposed to achieve switching-cycle-level settling time, and the DC-link inductor is sized for 40% ripple to enable fast current change. Importantly, this article also proposes to minimize the DC-link current by varying the current even within a line cycle under single-phase load to improve the steady-state performance, in contrast with the reduction to a constant value in the literature. The proposed MPC features a constant switching frequency without weighting factors. The MPC does not have a high computational burden and is implemented in a regular digital controller for a prototype of soft-switching solid-state transformer (S4T) with reduced conduction loss. The effectiveness of the proposed method has been experimentally verified on the SiC S4T prototype during steady-state and dynamics under different multiport power flow conditions up to 2 kV peak. Here, the DC-link current in the experiments is close to the minimum current with a short zero-vector duration, which further verifies the performance of the proposed method.

14 SOLAR ENERGY↗

DOE OSTI2019

GRB and blazar jets shining through their stripes

Black hole-driven relativistic astrophysical jets, such as blazars and gamma-ray bursts (GRBs), are powerful sources of electromagnetic radiation. Their emission is powered by some energy dissipation and particle acceleration mechanism operating in situ at large distances from the black hole. We propose that the formation of the dissipative structures in the jet is controlled by the time variability of the accretion disc. We argue that the open magnetic field lines through the black hole, which drive a strongly magnetized jet, may have their polarity reversing over time-scales related to the growth of the magnetorotational dynamo in the disc. Consequently, the jet is characterized by an alternating toroidal field polarity along its propagation axis, i.e. it is a ‘striped jet’. Magnetic reconnection in the current sheets that form between the stripes dissipates the alternating field energy and thus powers further jet acceleration. Here, we consider a jet with a broad distribution of stripe widths l > l min , above a dominant scale l min . We find that the bulk acceleration of the jet, driven by the annihilation of the stripes, is very gradual. The dissipation rate peaks at a distance $z_{\rm peak}\sim 10^6 R_\mathrm{ g}\, (\Gamma _\infty /30)^2\, (l_{\rm min}/1000R_\mathrm{ g})$| , where R g is the black hole’s gravitational radius and Γ ∞ the jet’s asymptotic Lorentz factor, and exhibits a very broad plateau extending by ~4–5 orders of magnitude in distance. The prolonged energy dissipation accounts for the flat-to-inverted long-wavelength spectra commonly observed in jets. The model can also account for the broad range of flaring time-scales of blazars and the fact that their bulk acceleration appears to continue out to ~100 pc scales. In GRB jets, the model predicts comparable power for the photopsheric and Thomson-thin emission components.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2020

A Novel Active Optimization Approach for Rapid and Efficient Design Space Exploration Using Ensemble Machine Learning

In this work, a novel design optimization technique based on active learning, which involves dynamic exploration and exploitation of the design space of interest using an ensemble of machine learning algorithms, is presented. In this approach, a hybrid methodology incorporating an explorative weak learner (regularized basis function model) that fits high-level information about the response surface and an exploitative strong learner (based on committee machine) that fits finer details around promising regions identified by the weak learner is employed. For each design iteration, an aristocratic approach is used to select a set of nominees, where points that meet a threshold merit value as predicted by the weak learner are selected for evaluation. In addition to these points, the global optimum as predicted by the strong learner is also evaluated to enable rapid convergence to the actual global optimum once the most promising region has been identified by the optimizer. Additionally, this methodology is first tested by applying it to the optimization of a two-dimensional multi-modal surface and, subsequently, to a complex internal combustion (IC) engine combustion optimization case with nine control parameters related to fuel injection, initial thermodynamic conditions, and in-cylinder flow. It is found that the new approach significantly lowers the number of function evaluations that are needed to reach the optimum design configuration (by up to 80%) when compared to conventional optimization techniques, such as particle swarm and genetic algorithm-based optimization techniques.

24 POWER TRANSMISSION AND DISTRIBUTION↗

DOE OSTI2022

Determination of the characteristic magnetic pre-sheath length at divertor surfaces using micro-engineered targets on DiMES at DIII-D

The magnetic pre-sheath (MPS) width, L MPS , is a critical parameter to define the sheath potential, which controls the ion trajectory of low-Z species (D, T, He, and C), as well as the prompt re-deposition of high-Z species. In this work, to determine L MPS , we fabricated micro-trenches (30×30×4 µm) via focused ion beam (FIB) milling on a silicon surface and exposed them to L-mode deuterium plasmas in DIII-D via the Divertor Material Evaluation System (DiMES) removable sample exposure probe. The areal distribution of impurity depositions, mainly consisting of carbon, was measured by energy-dispersive X-ray spectroscopy (EDS) to reveal the deuterium ion shadowing effect on the trench floors. The carbon deposition profiles showed that the erosion was maximized for the azimuthal direction of φ = -40° (referenced to the toroidal magnetic field direction) as well as the polar angle of θ = 80°. A Monte Carlo equation-of-motion model, based on a collisionless MPS, was used to calculate the azimuthal and polar deuterium ion angle distributions (IADs) for a range of L MPS = k × ρ i , where ρ i is the ion gyro radius and k = 0.5-4. Then, gross erosion profiles were calculated by a Monte Carlo micro-patterning and roughness (MPR) code for ion sputtering using as input the calculated azimuthal and polar IADs for each value of k . Good agreement with the experimental C deposition profiles was obtained for the case k = 2.5-3.5. This result is consistent with a previous kinetic modeling prediction of k ~ 3, as well as previous analytical investigations that predicted the L MPS to be several ion gyro radii. A validation of theoretical sheath models supports its applicability to ITER and pilot plant divertors to successfully predict plasma-materials interactions.

Erosion↗

DOE OSTI2022

Data-Centric Approach to Capture Non-Polynomial Nonlinear Dynamics

We propose an analytical construction of observable functions in the extended dynamic mode decomposition (EDMD) algorithm. EDMD is a numerical method for approximating the spectral properties of the Koopman operator. The choice of observable functions is fundamental for applying EDMD to nonlinear problems arising in systems and control. Existing methods either start from a set of dictionary functions and look for the subset that best fits the underlying nonlinear dynamics or rely on machine learning algorithms to “learn” observable functions. Conversely, in this paper, we start from the dynamical system model and lift it through the Lie derivatives, rendering it into a polynomial form. This proposed transformation into a polynomial form is exact and provides an adequate set of observable functions. The strength of the proposed approach is its applicability to a broader class of nonlinear dynamical systems, particularly those with nonpolynomial functions and compositions thereof. Moreover, it retains the physical interpretability of the underlying dynamical system and can be readily integrated into existing numerical libraries. We demonstrate the proposed approach with an application to electric power systems. The modeled system consists of a single generator connected to an infinite bus, where nonlinear terms include sine and cosine functions. The results demonstrate the effectiveness of the proposed procedure in off-attractor nonlinear dynamics for estimation and prediction; the observable functions obtained from the proposed construction outperform methods that use dictionary functions comprising monomials or radial basis functions.

extended dynamic mode decomposition↗

DOE OSTI2017

Efficient Machine Learning Approach for Optimizing Scientific Computing Applications on Emerging HPC Architectures

Efficient parallel implementations of scientific applications on multi-core CPUs with accelerators such as GPUs and Xeon Phis is challenging. This requires - exploiting the data parallel architecture of the accelerator along with the vector pipelines of modern x86 CPU architectures, load balancing, and efficient memory transfer between different devices. It is relatively easy to meet these requirements for highly structured scientific applications. In contrast, a number of scientific and engineering applications are unstructured. Getting performance on accelerators for these applications is extremely challenging because many of these applications employ irregular algorithms which exhibit data-dependent control-ow and irregular memory accesses. Furthermore, these applications are often iterative with dependency between steps, and thus making it hard to parallelize across steps. As a result, parallelism in these applications is often limited to a single step. Numerical simulation of charged particles beam dynamics is one such application where the distribution of work and memory access pattern at each time step is irregular. Applications with these properties tend to present significant branch and memory divergence, load imbalance between different processor cores, and poor compute and memory utilization. Prior research on parallelizing such irregular applications have been focused around optimizing the irregular, data-dependent memory accesses and control-ow during a single step of the application independent of the other steps, with the assumption that these patterns are completely unpredictable. We observed that the structure of computation leading to control-ow divergence and irregular memory accesses in one step is similar to that in the next step. It is possible to predict this structure in the current step by observing the computation structure of previous steps. In this dissertation, we present novel machine learning based optimization techniques to address the parallel implementation challenges of such irregular applications on different HPC architectures. In particular, we use supervised learning to predict the computation structure and use it to address the control-ow and memory access irregularities in the parallel implementation of such applications on GPUs, Xeon Phis, and heterogeneous architectures composed of multi-core CPUs with GPUs or Xeon Phis. We use numerical simulation of charged particles beam dynamics simulation as a motivating example throughout the dissertation to present our new approach, though they should be equally applicable to a wide range of irregular applications. The machine learning approach presented here use predictive analytics and forecasting techniques to adaptively model and track the irregular memory access pattern at each time step of the simulation to anticipate the future memory access pattern. Access pattern forecasts can then be used to formulate optimization decisions during application execution which improves the performance of the application at a future time step based on the observations from earlier time steps. In heterogeneous architectures, forecasts can also be used to improve the memory performance and resource utilization of all the processing units to deliver a good aggregate performance. We used these optimization techniques and anticipation strategy to design a cache-aware, memory efficient parallel algorithm to address the irregularities in the parallel implementation of charged particles beam dynamics simulation on different HPC architectures. Experimental result using a diverse mix of HPC architectures shows that our approach in using anticipation strategy is effective in maximizing data reuse, ensuring workload balance, minimizing branch and memory divergence, and in improving resource utilization.

Arumugam, Kamesh↗

DOE OSTI2022

Demonstration of a fault impact reduction control module for wind turbines

Abstract. Traditionally, wind turbines in distributed applications make control decisions as isolated systems. They generally provide maximum power output during operation and manage internal faults with little consideration of the rest of the power system. Although fault detection and tolerance schemes are widely researched and implemented, controls to ameliorate such faults are uncommon in research and industry. The rapid shutdown of a wind turbine in a large transmission-connected wind plant will have a minimal impact on a large power system, but in a microgrid or isolated grid context the abrupt loss of a single wind turbine may cause grid instability and high stress on the system. This paper demonstrates a fault impact reduction control (FIRC) module for a wind turbine, which implements wider warning thresholds around fault thresholds. When the turbine crosses a warning threshold, the controller sends its predicted action to the grid controller, which facilitates the grid operator’s response to a potential wind turbine fault and then takes appropriate action to ameliorate the fault. Various test cases demonstrate the controller action under a variety of faults, and various scenarios demonstrate the grid benefit of an FIRC in both microgrid- and grid-connected contexts. The FIRC maximizes wind turbine generation and eases generation transition under a variety of fault scenarios. The FIRC module is easy to integrate with most existing controllers, just requiring derate capability, and can be easily modified to include the various warnings and thresholds that the user desires. This analysis is mainly performed in a MATLAB-Simulink-based research wind turbine model and is also implemented in the existing LabVIEW-based controller of the same research turbine at NREL.

17 WIND ENERGY↗