Search NASA⌕ Search

SEARCH · Search NASA

Results for “differentiable programming”

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.

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

At least 361 records · Page 20
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

CO2 Sorption in Ionic Liquid Crystals

Ionic liquid crystals (ILCs) have an affinity for certain polarizable gases such as CO2, due to their similarity to ionic liquids. We investigated three ILCs in the [1-alkyl-3-methylimidazolium+] family: n=12,14 with [BF4-] and [PF6-]: liquid crystalline analogues to ionic liquids with moderate (e.g., 1-2 mol%) CO2 solubility at atmospheric conditions: [1-butyl-3-methylimidazolium+] with [BF4-] and [PF6-]. While ionic liquids show high CO2 solubility, regenerating the CO2 is a high-energy process. Liquid crystals show low CO2 solubility but have a much lower regeneration energy requirement. Will ionic liquid crystals uptake CO2? What are the energy requirements of regenerating CO2? Conclusions: 1. C12mim BF4- shows the highest sorption at 0.12 wt% CO2 in the isotropic phase vs. C14mim BF4- with 0.097 wt% in the smectic phase. We hypothesize that the increase in chain length affects the free volume of the smectic vs. isotropic phase of C14mim BF4-, increasing the latter. 2. The change in anion from BF4- to PF6- decreased the sorption to an insignificant level more analogous to a physical adsorption onto the material in all phases. We hypothesize that the change in anion to the larger, less charge dense PF6- decreased the attractive forces between CO2 and the anion. 3. 0.12 wt% of CO2 in C12mim BF4- is small but significant. This in combination with the room temperature release of CO2 after only requiring refrigeration temperatures to occlude the CO2, making ionic liquid crystals promising materials for future.

carbon dioxide sorption↗
DOE OSTI2015

Toward the credibility of Northeast United States summer precipitation projections in CMIP5 and NARCCAP simulations

Abstract Precipitation projections for the northeast United States and nearby Canada (Northeast) are examined for 15 Fifth Phase of the Coupled Model Intercomparison Project (CMIP5) models. A process‐based evaluation of atmospheric circulation features associated with wet Northeast summers is performed to examine whether credibility can be differentiated within the multimodel ensemble. Based on these evaluations, and an analysis of the interannual statistical properties of area‐averaged precipitation, model subsets were formed. Multimodel precipitation projections from each subset were compared to the multimodel projection from all of the models. Higher‐resolution North American Regional Climate Change Assessment Program (NARCCAP) regional climate models (RCMs) were subjected to a similar evaluation, grouping into subsets, and examination of future projections. CMIP5 models adequately simulate most large‐scale circulation features associated with wet Northeast summers, though all have errors in simulating observed sea level pressure and moisture divergence anomalies in the western tropical Atlantic/Gulf of Mexico. Relevant large‐scale processes simulated by the RCMs resemble those of their driving global climate models (GCMs), which are not always realistic. Future RCM studies could benefit from a process analysis of potential driving GCMs prior to dynamical downscaling. No CMIP5 or NARCCAP models were identified as clearly more credible, but six GCMs and four RCMs performed consistently better. Among the “Better” models, there is no consistency in the direction of future summer precipitation change. CMIP5 projections suggest that the Northeast precipitation response depends on the dynamics of the North Atlantic anticyclone and associated circulation and moisture convergence patterns, which vary among “Better” models. Even when model credibility cannot be clearly differentiated, examination of simulated processes provides important insights into their evolution under greenhouse warming.

Thibeault, Jeanne M.↗
DOE OSTI2016

Local Petrovskii lacunas close to parabolic singular points of the wavefronts of strictly hyperbolic partial differential equations

We enumerate the local Petrovskii lacunas (that is, the domains of local regularity of the principal fundamental solutions of strictly hyperbolic PDEs with constant coefficients in R{sup N}) close to parabolic singular points of their wavefronts (that is, at the points of types P{sub 8}{sup 1}, P{sub 8}{sup 2}, ±X{sub 9}, X{sub 9}{sup 1}, X{sub 9}{sup 2}, J{sub 10}{sup 1} and J{sub 10}{sup 3}). These points form the next most difficult family of classes in the natural classification of singular points after the so-called simple singularities A{sub k}, D{sub k}, E{sub 6}, E{sub 7} and E{sub 8}, which have been investigated previously. Also we present a computer program which counts the topologically distinct morsifications of critical points of smooth functions, and hence also the local components of the complement of a generic wavefront at its singular points. Bibliography: 22 titles. (paper)

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2019

A moment approach for entropy solutions to nonlinear hyperbolic PDEs

We propose to solve hyperbolic partial differential equations(PDEs) with polynomial flux using a convex optimization strategy. This approach is based on a very weak notion of solution of the nonlinear equation, namely the measure-valued (mv) solution, satisfying a linear equation in the space of Borel measures. The aim of this paper is, first, to provide the conditions that ensure the equivalence between the two formulations and, second, to introduce a method which approximates the infinite-dimensional linear problem by a hierarchy of convex, finite-dimensional, semidefinite programming problems. This result is then illustrated on the celebrated Burgers equation. Here, we also compare our results with an existing numerical scheme, namely the Godunov scheme.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Model-based and Model-free Designs for an Extended Continuous-time LQR with Exogenous Inputs

We present an extended linear quadratic regulator (LQR) design for continuous-time linear time-invariant (LTI) systems in the presence of exogenous inputs. We first propose a model-based solution with cost minimization guarantees for states and inputs using dynamic programming (DP). The control law consists of a combination of the optimal state feedback and an additional optimal term dependent on the exogenous inputs. The control gains for the two components are obtained by solving a set of matrix differential equations. We provide these solutions for both finite horizons and steady-state cases. In the second part of the paper, we formulate a reinforcement learning (RL) based algorithm which does not need any model information except the input matrix, and can compute an approximate steady-state LQR gain using measurements of the states, the control inputs, and the exogenous inputs. Both model-based and data-driven optimal control algorithms are tested with a numerical example under different exogenous inputs showcasing the effectiveness of the designs.

Mukherjee, Sayak↗
DOE OSTI2019

NEK5000 Assessment Milestone Report: Single-Phase Natural Circulation using Direct Numerical Simulation (DNS) & Large Eddy Simulation (LES) Methods

This study evaluates the capabilities of the NEK5000 open-source code as an alternative to the commercially available software that is currently heavily relied upon. The first three case studies initiated the author with (1) the syntax and flow of the program, (2) varying initial and boundary conditions and (3) geometries of greater complexity & dimension. Case 1 studies transient conjugate heat transfer in the entry region of a pipe. Cases 2 & 3 simulate the natural convection flow in an enclosed differentially heated cavity (DHC) of two and three dimensions, respectively. The final case study, Case 4, produces three-dimensional single-phase natural convective flow using the Boussinesq approximation within a simplified representation of an aqueous homogeneous reactor system. The Boussinesq flow in Case 4 clearly depicts both Rayleigh-Taylor and Kelvin-Helmholtz instabilities qualitatively during post-processing analysis. Further analysis of Case 4, up to the point of a quasi-steady state (i.e. thermal-equilibrium condition), is done using the LES solver method in NEK5000. The volume-averaged solution temperatures, when compared between solvers, showed no significant difference. Due to its minimal compromise in fidelity of solution resolution and inexpensive computing cost, compared to the DNS solver, the LES solver is recommended for the current application. Methodology development for the four case studies are reported and a preliminary optimization study with different polynomial resolutions for Case 4 is addressed as a resource when performing future simulations using NEK5000. Hence, this report can be used as a reference starting point to the end user of NEK5000 for natural circulation phenomena in nuclear reactor safety applications.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2013

Uncertainty of Personnel Dosimetry Systems Using Performance Data Testing

While accreditation programs require a study of uncertainty for each dosimeter model for which accreditation is sought (Torres 2005, DOE 1986), no one method is specifically recommended by accrediting bodies. Typically, the methods used to determine uncertainty have included the partial differentiation method described in the U.S. Guide to Uncertainty in Measurements (ANSI 1997) or the use of Monte Carlo techniques and probability distribution functions to generate simulated dose results. Each of these techniques has particular strengths and should be employed when the areas of uncertainty are required to be understood in detail. However, the uncertainty of dosimeter results can also be determined using a Model II One-Way Analysis of Variance technique and accreditation testing data (Cummings and Flood 2010). The strengths of this method are (1) the method is based in robust and well-accepted statistical techniques; (2) the method is straightforward and easily applied to dose algorithms, which are so complex as to render the use of other methods unwieldy; and (3) the method uses dosimeter performance data obtained during accreditation. The general method and criteria of inclusion of radiation types and fields are discussed.

Cummings, Frederick (Rick) M.↗
DOE OSTI2015

Optimization of protein cross-linking in bicomponent electrospun scaffolds for therapeutic use

Bio-instructive electrospun scaffolds based on the combination of synthetic polymers, such as PCL or PLLA, and natural polymers (e.g., collagen) have been extensively investigated as temporary extracellular matrix (ECM) analogues able to support cell proliferation and stem cell differentiation for the regeneration of several tissues. The growing use of natural polymers as carrier of bioactive molecules is introducing new ideas for the design of polymeric drug delivery systems based on electrospun fibers with improved bioavailability, therapeutic efficacy and programmed drug release. In particular, the release mechanism is driven by the use of water soluble proteins (i.e., collagen, gelatin) which fully degrade in in vitro microenvironment, thus delivering the active principles. However, these protein are generally rapidly digested by enzymes (i.e., collagenase) produced by many different cell types, both in vivo and in vitro with significant drawbacks in tissue engineering and controlled drug delivery. Here, we aim at investigating different chemical strategies to improve the in vitro stability and mechanical strength of scaffolds against enzymatic degradation, by modifying the biodegradation rates of proteins embedded in bicomponent fibers. By comparing scaffolds treated by different cross-linking agents (i.e., GC, EDC, BDDGE), we have provided an extensive morphological/chemical/physical characterization via SEM and TGA to identify the best conditions to control drug release via protein degradation from bicomponent fibers without compromising in vitro cell response.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2022

dfdjaxGP

A small python package to fit Gaussian processes using Jax and leveraging the automatic differentiation in Jax to predict arbitrary derivatives from the GP. This package is meant to supplement the scientific community use of Gaussian process prediction with derivatives. The package is designed to smoothly work standalone or be used with the numpyro probabilistic programming language.

Grosskopf, Micheal↗
DOE OSTI2014

Benchmark Evaluation of the HTR-PROTEUS Absorber Rod Worths (Core 4)

PROTEUS was a zero-power research reactor at the Paul Scherrer Institute (PSI) in Switzerland. The critical assembly was constructed from a large graphite annulus surrounding a central cylindrical cavity. Various experimental programs were investigated in PROTEUS; during the years 1992 through 1996, it was configured as a pebble-bed reactor and designated HTR-PROTEUS. Various critical configurations were assembled with each accompanied by an assortment of reactor physics experiments including differential and integral absorber rod measurements, kinetics, reaction rate distributions, water ingress effects, and small sample reactivity effects [1]. Four benchmark reports were previously prepared and included in the March 2013 edition of the International Handbook of Evaluated Reactor Physics Benchmark Experiments (IRPhEP Handbook) [2] evaluating eleven critical configurations. A summary of that effort was previously provided [3] and an analysis of absorber rod worth measurements for Cores 9 and 10 have been performed prior to this analysis and included in PROTEUS-GCR-EXP-004 [4]. In the current benchmark effort, absorber rod worths measured for Core Configuration 4, which was the only core with a randomly-packed pebble loading, have been evaluated for inclusion as a revision to the HTR-PROTEUS benchmark report PROTEUS-GCR-EXP-002.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2022

On the Convergence of Overlapping Schwarz Decomposition for Nonlinear Optimal Control

Here, we study the convergence properties of an overlapping Schwarz decomposition algorithm for solving nonlinear optimal control problems (OCPs). The algorithm decomposes the time domain into a set of overlapping subdomains, and solves all subproblems defined over subdomains in parallel. The convergence is attained by updating primal-dual information at the boundaries of overlapping subdomains. We show that the algorithm exhibits local linear convergence, and that the convergence rate improves exponentially with the overlap size. We also establish global convergence results for a general quadratic programming, which enables the application of the Schwarz scheme inside second-order optimization algorithms (e.g., sequential quadratic programming). The theoretical foundation of our convergence analysis is a sensitivity result of nonlinear OCPs, which we call "exponential decay of sensitivity" (EDS). Intuitively, EDS states that the impact of perturbations at domain boundaries (i.e., initial and terminal time) on the solution decays exponentially as one moves into the domain. Here, we expand a previous analysis available in the literature by showing that EDS holds for both primal and dual solutions of nonlinear OCPs, under uniform second-order sufficient condition, controllability condition, and boundedness condition. We conduct experiments with a quadrotor motion planning problem and a partial differential equations (PDE) control problem to validate our theory, and show that the approach is significantly more efficient than alternating direction method of multipliers and as efficient as the centralized interior-point solver.

42 ENGINEERING↗
DOE OSTI2014

Chemical Imaging and Dynamical Studies of Reactivity and Emergent Behavior in Complex Interfacial Systems. Final Technical Report

This research program explored the efficacy of using molecular-level manipulation, imaging and scanning tunneling spectroscopy in conjunction with supersonic molecular beam gas-surface scattering to significantly enhance our understanding of chemical processes occurring on well-characterized interfaces. One program focus was on the spatially-resolved emergent behavior of complex reaction systems as a function of the local geometry and density of adsorbate-substrate systems under reaction conditions. Another focus was on elucidating the emergent electronic and related reactivity characteristics of intentionally constructed single and multicomponent atom- and nanoparticle-based materials. We also examined emergent chirality and self-organization in adsorbed molecular systems where collective interactions between adsorbates and the supporting interface lead to spatial symmetry breaking. In many of these studies we combined the advantages of scanning tunneling (STM) and atomic force (AFM) imaging, scanning tunneling local electronic spectroscopy (STS), and reactive supersonic molecular beams to elucidate precise details of interfacial reactivity that had not been observed by more traditional surface science methods. Using these methods, it was possible to examine, for example, the differential reactivity of molecules adsorbed at different bonding sites in conjunction with how reactivity is modified by the local configuration of nearby adsorbates. At the core of this effort was the goal of significantly extending our understanding of interfacial atomic-scale interactions to create, with intent, molecular assemblies and materials with advanced chemical and physical properties. This ambitious program addressed several key topics in DOE Grand Challenge Science, including emergent chemical and physical properties in condensed phase systems, novel uses of chemical imaging, and the development of advanced reactivity concepts in combustion and catalysis including carbon management. These activities directly benefitted national science objectives in the areas of chemical energy production and advanced materials development.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

Dual-Salt Electrolytes to Effectively Reduce Impedance Rise of High-Nickel Lithium-Ion Batteries

Simply mixing several lithium salts in one electrolyte to obtain blended salt electrolytes has been demonstrated as a promising strategy to formulate advanced electrolytes for lithium metal batteries (LMBs) and lithium-ion batteries (LIBs). In this study, we report the use of dual-salt electrolytes containing lithium hexafluorophosphate (LiPF6) and lithium difluorophosphate (LiDFP) in ethylene carbonate/ethyl methyl carbonate (EC/EMC) mixture and tested them in layered high-nickel LIB cells. LiNi 0.94 Co 0.06 O 2 was synthesized through a coprecipitation method and was used as a representative high-nickel cathode for the U.S. DOE realizing next-generation cathode (RNGC) deep dive program. The ionic conductivity of dual-salt electrolytes can be maintained by controlling the amount of LiDFP. Techniques including 1 H Nuclear Magnetic Resonance (NMR), X-ray photoelectron spectroscopy (XPS), inductively coupled plasma-mass spectrometry (ICP-MS), and differential voltage analysis (DVA) were used to understand the improved performance. Additionally, the multifaceted benefits of using the dual-salt electrolytes include (1) reduced transesterification, (2) formation of a stable cathode electrolyte interface, and (3) mitigation of cathode degradation at high voltages, especially stabilization of oxide particles during the H 2 <-> H 3 transformation.

25 ENERGY STORAGE↗
DOE OSTI2015

Advanced instrumentation and analysis methods for in-pile thermal and nuclear measurements: from out-of-pile studies to irradiation campaigns

Research and development on nuclear fuel behavior under irradiations and accelerated ageing of structure materials is a key issue for sustainable nuclear energy in order to meet specific needs by keeping the best level of safety. A new Material Testing Reactor (MTR), the Jules Horowitz Reactor (JHR) currently under construction in the South of France in the CEA Cadarache research centre will offer a real opportunity to perform R and D programs and hence will crucially contribute to the selection, optimization and qualification of innovative materials and fuels. To perform such programs advanced accurate and innovative experiments, irradiation devices that contain material and fuel samples are required to be set up inside or beside the reactor core. These experiments needs beforehand in situ and on line sophisticated measurements to accurately reach specific and determining parameters such as thermal and fast neutron fluxes, nuclear heating and temperature conditions to precisely monitor and control the conducted assays. Consequently, since 2009 CEA and Aix-Marseille University collaborate in order to design and develop a new multi-sensor device which will be dedicated to measuring profiles of such conditions inside the experimental channels of the JHR. These works are performed in the framework of two complementary joint research programs called MAHRI-BETHY and INCORE. These programs couple experimental studies carried out both out-of nuclear fluxes (in laboratory) and under irradiation conditions (in OSIRIS MTR reactor in France and MARIA MTR reactor in Poland) with numerical works realized by thermal simulations (CAST3M code) and Monte Carlo simulations (MCNP code). These programs deal with three main aims. The first one corresponds to the design and/or the test of new in-pile instrumentation. The second one concerns the development of advanced calibration procedures in particular in the case of one specific sensor: a differential calorimeter used to quantify nuclear heating. The last one consists in the development of accurate measurement and analysis methods. The paper will be dedicated to a complete review of the experimental and numerical works performed since 2009 thanks to two parts. The first part will detail a new thermal approach implemented to improve nuclear heating measurements by radiometric calorimeters. New experimental tools (calorimeter prototypes and set-ups such BETHY Bench) developed to perform preliminary out-of-pile studies under suitable conditions will be presented (temperature and velocity of the external cooling fluid, heat source localization and intensity inside the calorimetric cells). Then the response of two kinds of sensors, their calibrations curves and their thermal behaviors will be compared for various parameters. Finally validated numerical thermal and Monte Carlo works will be discussed to propose new improvements. The second parts of the paper will focus on works realized in order to design, develop and test the first prototype of the multi-sensor device called CARMEN [7-9]. The two mock-ups dedicated respectively to neutron measurements and photon measurements will be detailed. The results obtained during two irradiation campaigns inside the periphery of OSIRIS reactor will be shown. The new analysis method will be discussed. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
DOE OSTI2022

Random field optimization

Herein we present a new modeling paradigm for optimization that we call random field optimization. Random fields are a powerful modeling abstraction that aims to capture the behavior of random variables that live on infinite-dimensional spaces (e.g., space and time) such as stochastic processes (e.g., time series, Gaussian processes, and Markov processes), random matrices, and random spatial fields. This paradigm involves sophisticated mathematical objects (e.g., stochastic differential equations and space-time kernel functions) and has been widely used in neuroscience, geoscience, physics, civil engineering, and computer graphics. Despite of this, however, random fields have seen limited use in optimization; specifically, existing optimization paradigms that involve uncertainty (e.g., stochastic programming and robust optimization) mostly focus on the use of finite random variables. This trend is rapidly changing with the advent of statistical optimization (e.g., Bayesian optimization) and multi-scale optimization (e.g., integration of molecular sciences and process engineering). Our work extends a recently-proposed abstraction for infinite-dimensional optimization problems by capturing more general uncertainty representations. Moreover, we discuss solution paradigms for this new class of problems based on finite transformations and sampling, and identify open questions and challenges.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

First Principles Nonadiabatic Excited-State Molecular Dynamics in NWChem

Computational simulation of non-adiabatic molecular dynamics is an indispensable tool for understanding complex photoinduced processes such as internal conversion, energy transfer, charge separation, and spatial localization of excitons, to name a few. Here, we report an implementation of the fewest-switches surface hopping algorithm in the NWChem computational chemistry program. Here the surface hopping method is combined with linear-response time-dependent density functional theory calculations of adiabatic excited state potential energy surfaces. To treat quantum transitions between arbitrary electronic Born{Oppenheimer states, we have implemented both numerical and analytical differentiation schemes for derivative non-adiabatic couplings. A numerical approach for the time-derivative non-adiabatic couplings together with an analytical method for calculating non-adiabatic coupling vectors is an efficient combination for surface hopping approaches. Additionally, electronic decoherence schemes and a state reassigned unavoided crossings algorithm are also implemented to improve the accuracy of the simulated dynamics and to handle trivial unavoided crossings. We apply our code to study the ultrafast decay of photoexcited benzene, including a detailed analysis of the potential energy surface, population decay time scales, and vibrational coordinates coupled to the excitation dynamics. The development presented in this work is a baseline for future implementations of more sophisticated frameworks for simulating of non-adiabatic molecular dynamics in NWChem.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2019

Vacuum Sanitation Energy Efficiency Retrofits Project

The Alaska Native Tribal Health Consortium (ANTHC) is a non-profit Tribal health organization designed to meet the unique health needs of Alaska Native and American Indian people living in Alaska. In partnership with communities and health organizations across the state, ANTHC provides health services to over 175,000 Alaska Native and American Indian people. Part of ANTHC’s mission includes ensuring that Native Alaskan communities have access to clean and safe water and sewer services. ANTHC works with communities to provide safe drinking water and healthy wastewater services that directly supports the health of Alaska Native people. To ensure water and sewer facilities operate sustainably, ANTHC Alaska Rural Utility Collaborative (ARUC) and Tribal Utility Support (TUS) programs provide operations and maintenance services on rural sanitation facilities. A key component of many water and sanitation systems in rural Alaskan villages is the vacuum sewer system, which is typically installed in places where the topography is primarily flat, and the elevation differential is not enough for a gravity powered sewer system to function without freezing during the winter. The vacuum sewer system is also the most costly and energy intensive form of sewer system, due to many factors including: large horsepower vacuum pumps needed for operation, potential for frequent vacuum leaks which increase the pump’s run-times, manpower needed to maintain pipe grade throughout the system, and the amount of heat needed to be added to the above ground pipes to keep them from freezing. The Vacuum Sanitation Energy Efficiency Retrofits Project implemented strategies to reduce the operating costs for vacuum sewer systems located in the remote Alaskan communities of Alakanuk, Kotlik, and Noorvik. A variety of deep energy efficiency retrofits were installed to reduce the amount of diesel fuel and electricity required to operate these systems which use large amounts of energy to pump, heat, and collect sewage from the community.

20 FOSSIL-FUELED POWER PLANTS↗
DOE OSTI2026

Charged Pion Production Explorations Using the NOvA Near Detector

Measurements of neutrino-nucleus interactions in the few-GeV region provide crucial inputs to the neutrino oscillation program being carried out by currently running experiments (NOvA, SBN, T2K), and next-generation experiments (DUNE, 2HK). This poster presents the status of a measurement using the high-intensity NuMI beam and the NOvA Near Detector of the flux-integrated differential cross-section in pion kinetic energy for $\nu_\mu + N \rightarrow \mu^- + \pi^\pm + X$, where X does not include any charged pions. This signal is sensitive to inelastic processes with Delta resonance production, which dominate at neutrino energies relevant for NOvA and DUNE. To overcome the difficulty of reconstructing the energy of charged pions in a lower-spatial-resolution detector, a novel analysis technique is used that incorporates a multi-template fit to extract the number of events for two signal categories that differ from one another by their purity and resolution. The two signal categories are combined leveraging their complementary errors into a more informative cross section result.

Ewart, Erin [Indiana U.]↗