Search NASA⌕ Search

SEARCH · Search NASA

Results for “algorithmic differentiation”

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 145 records · Page 8
All sourcesNASA NTRSDOE OSTI
DOE OSTI2011

Gaussian beam decomposition of high frequency wave fields using expectation-maximization

A new numerical method for approximating highly oscillatory wave fields as a superposition of Gaussian beams is presented. The method estimates the number of beams and their parameters automatically. This is achieved by an expectation-maximization algorithm that fits real, positive Gaussians to the energy of the highly oscillatory wave fields and its Fourier transform. Beam parameters are further refined by an optimization procedure that minimizes the difference between the Gaussian beam superposition and the highly oscillatory wave field in the energy norm.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2014

Residual Monte Carlo high-order solver for Moment-Based Accelerated Thermal Radiative Transfer equations

In this article we explore the possibility of replacing Standard Monte Carlo (SMC) transport sweeps within a Moment-Based Accelerated Thermal Radiative Transfer (TRT) algorithm with a Residual Monte Carlo (RMC) formulation. Previous Moment-Based Accelerated TRT implementations have encountered trouble when stochastic noise from SMC transport sweeps accumulates over several iterations and pollutes the low-order system. With RMC we hope to significantly lower the build-up of statistical error at a much lower cost. First, we display encouraging results for a zero-dimensional test problem. Then, we demonstrate that we can achieve a lower degree of error in two one-dimensional test problems by employing an RMC transport sweep with multiple orders of magnitude fewer particles per sweep. We find that by reformulating the high-order problem, we can compute more accurate solutions at a fraction of the cost.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2024

An ion specific continuum model on ionic surfactant's binary phase diagram in aqueous solution

Here, a continuum aggregation model is proposed to account for specific ion effects, enabling accurate phase diagrams calculations for the micellar, cylindrical and lamellar aggregates of sodium/potassium carboxylate surfactants in aqueous solution across a range of temperatures. Three groups of concentrations at distinctive temperatures are fitted to build empirical temperature dependence given the limited availability of relevant experimental measurements. The specific ion effects are manifested in the aggregates' surface tension as well as in the distributions of counter-ions' concentrations in the vicinity of the aggregates. The aggregates' geometric sizes are well-reproduced. The differential evolution algorithm is applied to address boundary conditions of the electrostatic potential, aggregate size optimization as well as the equilibrium of monomers transferring between the aggregate and aqueous region.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2011

Multi-bump solutions for the nonlinear Schroedinger-Poisson system

In this paper, we study a kind of nonlinear Schroedinger-Poisson system with a parameter {epsilon}. For any positive integer m, we prove that there exists {epsilon}(m) > 0 such that, for 0 < {epsilon} < {epsilon}(m), the equation has an m-bump positive solution under some suitable conditions. As a consequence, the equation has more and more multi-bump positive solutions as {epsilon}{yields} 0.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2011

Nonlinear dynamical system approaches towards neural prosthesis

An asynchronous discrete-state spiking neurons is a wired system of shift registers that can mimic nonlinear dynamics of an ODE-based neuron model. The control parameter of the neuron is the wiring pattern among the registers and thus they are suitable for on-chip learning. In this paper an asynchronous discrete-state spiking neuron is introduced and its typical nonlinear phenomena are demonstrated. Also, a learning algorithm for a set of neurons is presented and it is demonstrated that the algorithm enables the set of neurons to reconstruct nonlinear dynamics of another set of neurons with unknown parameter values. The learning function is validated by FPGA experiments.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2013

Periodically multilayered planar optical concentrator for photovoltaic solar cells

A planar optical concentrator comprising a periodic multilayered isotropic dielectric material backed by a metallic surface-relief grating was theoretically examined for silicon photovoltaics. The concentrator was optimized using a differential evolution algorithm for solar-spectrum-integrated power-flux density. Further optimization was carried out for tolerance to variations in the incidence angle, spatial dimensions, and dielectric properties. The average electron-hole pair density in a silicon solar cell can be doubled, and the material costs substantially diminished by this concentrator, whose efficacy is due to the excitation of waveguide modes and multiple surface-plasmon-polariton waves in a broad spectral regime.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2015

Mathematical model of the metal mould surface temperature optimization

The article is focused on the problem of generating a uniform temperature field on the inner surface of shell metal moulds. Such moulds are used e.g. in the automotive industry for artificial leather production. To produce artificial leather with uniform surface structure and colour shade the temperature on the inner surface of the mould has to be as homogeneous as possible. The heating of the mould is realized by infrared heaters located above the outer mould surface. The conceived mathematical model allows us to optimize the locations of infrared heaters over the mould, so that approximately uniform heat radiation intensity is generated. A version of differential evolution algorithm programmed in Matlab development environment was created by the authors for the optimization process. For temperate calculations software system ANSYS was used. A practical example of optimization of heaters locations and calculation of the temperature of the mould is included at the end of the article.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2024

Machine learning without a processor: Emergent learning in a nonlinear analog network

Standard deep learning algorithms require differentiating large nonlinear networks, a process that is slow and power-hungry. Electronic contrastive local learning networks (CLLNs) offer potentially fast, efficient, and fault-tolerant hardware for analog machine learning, but existing implementations are linear, severely limiting their capabilities. These systems differ significantly from artificial neural networks as well as the brain, so the feasibility and utility of incorporating nonlinear elements have not been explored. Here, we introduce a nonlinear CLLN—an analog electronic network made of self-adjusting nonlinear resistive elements based on transistors. We demonstrate that the system learns tasks unachievable in linear systems, including XOR (exclusive or) and nonlinear regression, without a computer. We find our decentralized system reduces modes of training error in order (mean, slope, curvature), similar to spectral bias in artificial neural networks. The circuitry is robust to damage, retrainable in seconds, and performs learned tasks in microseconds while dissipating only picojoules of energy across each transistor. This suggests enormous potential for fast, low-power computing in edge systems like sensors, robotic controllers, and medical devices, as well as manufacturability at scale for performing and studying emergent learning.

Science & Technology - Other Topics↗
DOE OSTI2020

Thermoeconomic cost optimization of superconducting magnets for proton therapy gantries

A compact gantry delivering 70-220 MeV protons with fixed field in the superconducting magnets could reduce the cost and improve the adoption of proton therapy. While a number of magnet and cryogenics designs have been proposed, the combined capital and operating costs of state-of-the-art superconducting materials have not been analyzed. In response, we develop a thermoeconomic model of a multi-stage, conduction cooled gantry lattice and analyze the cryocooler operating cost, cryocooler capital cost and conductor capital cost for Nb-Ti, Nb 3 Sn, REBCO and Bi-2223 over a continuous range of magnet temperatures, and a differential evolution algorithm is used to identify the optimal combination of thermal intercept temperatures. Although Nb3Sn yields the lowest Net Present Value (NPV) of $111.7k at a magnet temperature of 9.4 K, the optimized Bi-2223 design at 12.8 K approaches the realm of commercial feasibility by offering improved thermal stability and forgoing the need for costly conductor heat treatment and magnet quench training. Furthermore, it was found that Nb3Sn was more cost effective than Nb-Ti and that REBCO was not economically viable for the parameters of this investigation. Overall, the thermoeconomic model developed herein can optimize conductor choices, magnet temperatures and thermal staging which has value for any conduction-cooled superconducting magnet.

BSCCO↗
DOE OSTI2017

Accurate force field for molybdenum by machine learning large materials data

In this work, we present a highly accurate spectral neighbor analysis potential (SNAP) model for molybdenum (Mo) developed through the rigorous application of machine learning techniques on large materials data sets. Despite Mo's importance as a structural metal, existing force fields for Mo based on the embedded atom and modified embedded atom methods still do not provide satisfactory accuracy on many properties. We will show that by fitting to the energies, forces and stress tensors of a large density functional theory (DFT)-computed dataset on a diverse set of Mo structures, a Mo SNAP model can be developed that achieves close to DFT accuracy in the prediction of a broad range of properties, including energies, forces, stresses, elastic constants, melting point, phonon spectra, surface energies, grain boundary energies, etc. We will outline a systematic model development process, which includes a rigorous approach to structural selection based on principal component analysis, as well as a differential evolution algorithm for optimizing the hyperparameters in the model fitting so that both the model error and the property prediction error can be simultaneously lowered. We expect that this newly developed Mo SNAP model will find broad applications in large-scale, long-time scale simulations.

36 MATERIALS SCIENCE↗
DOE OSTI2015

SU-C-BRD-03: Closing the Loop On Virtual IMRT QA

Purpose: To develop an algorithm that predicts a priori IMRT QA passing rates. Methods: 416 IMRT plans from all treatment sites were planned in Eclipse version 11 and delivered using a dynamic sliding window technique on Clinac iX or TrueBeam linacs (Varian Medical Systems, Palo Alto, CA). The 3%/3mm and 2%/2mm local distance to agreement (DTA) were recorded during clinical operations using a commercial 2D diode array (MapCHECK 2, Sun Nuclear, Melbourne, FL). Each plan was characterized by 37 metrics that describe different failure modes between the calculated and measured dose. Machine-learning algorithms (MLAs) were trained to learn the relation between the plan characteristics and each passing rate. Minimization of the cross validated error, together with maximum a posteriori estimation (MAP), were used to choose the model parameters. Results: 3%/3mm local DTA can be predicted with an error smaller than 3% for 98% of the plans. For the remaining 2% of plans, the residual error was within 5%. For 2%/2mm local DTA passing rates, 96% percent of the plans were successfully predicted with an error smaller than 5%. All high-risk plans that failed the 2%/2mm local criteria were correctly identified by the algorithm. The most important metric to describe the passing rates was determined to be the MU per Gray (modulation factor). Conclusions: Logs files and independent dose calculations have been suggested as possible substitutes for measurement based IMRT QA. However, none of these methods answer the fundamental question of whether a plan can be delivered with a clinically acceptable error given the limitations of the linacs and the treatment planning system. Predicting the IMRT QA passing rates a priori closes that loop. For additional robustness, virtual IMRT QA can be combined with Linac QA and log file analysis to confirm appropriate delivery.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2011

Basic principles of the surface harmonics method: Flat geometry

The basic principles of the surface harmonics method are described. A one-dimensional problem is used to exemplify the specific features of the method and the algorithms for construction of finite-difference equations. The objective of this study is to popularize the surface harmonics method among specialists.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2018

PyBioNetFit

Systems biology models written in the BioNetGen language often contain parameters that must be fit based on experimental data. PyBioNetFit is a software tool designed to fit parameters of BioNetGen models. The user is expected to provide a BioNetGen model that contains unknown parameters, along with one or more text files containing experimental data. PyBioNetFit will perform a series of calls to the BioNetGen simulator using different parameter sets, to identify the parameter set that best fits the data. The parameter search is dictated by the user's choice of one of several metaheuristic optimization algorithms, including differential evolution, particle swarm optimization, scatter search, and simulated annealing. The software supports parallelized computing on multi-core computers and on computing clusters. We expect this software to be a useful tool for researchers designing BioNetGen models in collaboration with experimentalists.

Mitra, Eshan↗
DOE OSTI2017

Stan : A Probabilistic Programming Language

Stan is a probabilistic programming language for specifying statistical models. A Stan program imperatively defines a log probability function over parameters conditioned on specified data and constants. As of version 2.14.0, Stan provides full Bayesian inference for continuous-variable models through Markov chain Monte Carlo methods such as the No-U-Turn sampler, an adaptive form of Hamiltonian Monte Carlo sampling. Penalized maximum likelihood estimates are calculated using optimization methods such as the limited memory Broyden-Fletcher-Goldfarb-Shanno algorithm. Stan is also a platform for computing log densities and their gradients and Hessians, which can be used in alternative algorithms such as variational Bayes, expectation propagation, and marginal inference using approximate integration. To this end, Stan is set up so that the densities, gradients, and Hessians, along with intermediate quantities of the algorithm such as acceptance probabilities, are easily accessible. Stan can also be called from the command line using the cmdstan package, through R using the rstan package, and through Python using the pystan package. All three interfaces support sampling and optimization-based inference with diagnostics and posterior analysis. rstan and pystan also provide access to log probabilities, gradients, Hessians, parameter transforms, and specialized plotting.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2019

Crosscutting Recombination Metrology for Expediting VOC Engineering

Work carried out under this grant addressed the need for improved materials and device characterization of solar cell designs. This experimental program aimed to develop joint electrical and optical studies of solar cells based on the group-II-VI semiconductor CdTe and the chalcogenide semiconductor Cu(InGa)Se 2 . Electrical measurements focused on temperature (T) and illumination intensity (I) dependence of the open-circuit voltage (VOC), a quantity that plays an important role in conversion efficiency. The VOC depends on temperature and illumination intensity through the energy gap and positions of the quasiFermi levels (QFLs). Optical measurements exploited the illumination and temperature dependence of absolute photoluminescence (PL) intensity (PL-I), a quantity that also depends on the QFL positions. The QFL splitting is a quantity that also depends on recombination rates within the solar cell, including radiative and non-radiative processes. Other factors affect the electrical and optical properties. These include band offsets at both the front and back interfaces of the device active region where electrical contacts are formed. Understanding, locating, and quantifying the recombination properties of the solar cells was the principal goal of this research. Custom simulations were developed to solve continuity and Poisson equations selfconsistently to obtain free-electron and free-hole concentrations. The measurements of temperature and intensity dependent VOC and PL-I allowed the extraction of a number of physical parameters including the conduction band offset, the shunt constant, and the effective capture coefficient for recombination centers, both within the absorber and at the absorber-buffer interface. In addition, the back-contact barrier height is determined from electrical properties. Each of these is discussed in the report. The illumination intensity dependence of VOC and PL-I do not overlap in experiments primarily because PL is not observed in the low-intensity regime where VOC is studied. Absolute PL-I measurements enabled development of an approach directly comparing it with VOC. This in turn allowed development of a simulation capability that would consistently use both measurements, as functions of T and I, for producing a minimization scheme using a differential evolution algorithm. The approach was applied to both CdTe and Cu(InGa)Se 2 solar cells each prepared under various conditions. Shunt conductivity, QFL splitting, recombination were measured. Also extensively studied are current-voltage and frequency-dependent capacitance voltage. Work completed under this grant included studies of magnesium zinc oxide (MZO) buffered CdTe solar cells including the effects of hydroiodic acid (HI) treatment. From these studies the recombination rates were quantified at the buffer/absorber (front) interface. The effect of back-contact barrier height was investigated. CdTe solar cells were prepared with CdSe and CdS buffer layers. Lower grain-boundary potential fluctuation, higher back-contact barrier height, and greater bulk recombination were observed in early devices with CdSe buffer layer. Investigations were also carried out into CdS/Cu(InGa)Se 2 solar cells. A metal-insulatorsemiconductor capacitance-voltage technique was developed to extract recombinationrelevant properties of the absorber interface. From this work the density of interface states D it , their capture cross-section σ, fixed charge at the interface Q f , and surface recombination velocity were all extracted.

14 SOLAR ENERGY↗
DOE OSTI2022

Real-Time Drilling Optimization System for Improved Overall Rate of Penetration and Reduced Cost Per Foot in Geothermal Drilling

The key to success in geothermal drilling is economic feasibility, and a major cost in the development of geothermal resources is the actual drilling of the wells. In this project, a real-time drilling optimization system for geothermal drilling was developed. The system couples three individual components while drilling. The first component is a drill stem vibration analysis model, the second is Mechanical Specific Energy (MSE) analyses, and the third is a detailed PDC Rate of Penetration (ROP) drill bit model for optimum RPM and WOB combinations. The benefit of the coupled system is that the range of WOB and RPM could be selected to avoid drill stem vibrations. Secondly, MSE is used as an efficiency measure and the detailed PDC drill bit model ensures the drill bit does not endure temperatures that exceed the temperature at which the PDC cutters experience accelerated wear. The new detailed PDC bit model is based on rock/bit interaction that physically tracks the PDC cutter wear flats as the bit drills ahead giving the capability to calculate the temperature being generated underneath the worn cutters to better advise on operational parameters to avoid accelerated cutter wear and failure and to ensure that operational parameters are applied so that overall ROP is maximized. By combining the drill stem vibrations and the detailed PDC bit cutter wear and “safe” non-accelerated cutter wear temperature and optimum ranges of operating parameters, it results in higher ROP and lower cost drilling. Single cutter PDC testing performed in different lithologies at Sandia was utilized to verify the PDC cutter forces and depth of cut for new and worn cutters. Based on single cutter PDC temperature modeling, verification using single cutter data from the testing done by National Oilwell Varco (NOV) was performed. Sandia’s Hard-Rock Drilling Facility (HRDF) was utilized to test different drill bit configurations with different cutter designs and wear status with different induced modes of vibration to obtain the critical bit RPM/WOB ranges resulting in ineffective drilling and low ROP. The collected test data were further used to verify and calibrate the full hole PDC ROP model that was developed based on single cutter interaction data. A full coupled drill stem vibration model was formulated and verified with geothermal field data from the Chocolate Mountain Aerial Gunnery Range (CMAGR). A graphical user interface (GUI) was developed using Tkinter library in the computer programming language Python, which integrates all the developed models in one system. The developed system consists mainly of the PDC ROP model, PDC bit wear model, PDC cutter temperature model, Mechanical Specific Energy (MSE) model, and drillstring vibration model integrated into one system. The developed system can be used for both, post well analysis and real-time optimization using different criteria such as ROP maximization or MSE minimization. The software uses Differential Evolution Algorithm (DEA) to find optimum values for operational parameters based on last foot drilled while avoiding the drillstring vibration and cutter temperature critical operating parameters.

15 GEOTHERMAL ENERGY↗
DOE OSTI2019

Identification of Stellar Flares Using Differential Evolution Template Optimization

In this paper, we explore methods for the identification of stellar flare events in irregularly sampled data of ground-based time domain surveys. In particular, we describe a new technique for identifying flaring stars, which we have implemented in a publicly available Python module called "PyVAN." The approach uses the Differential Evolution algorithm to optimize parameters of empirically derived light curve templates for different types of stars to fit a candidate light curve. The difference of the likelihoods that these best-fit templates produced the observed data is then used to delineate targets that are well-explained by a flare template but simultaneously poorly explained by templates of common contaminants. By testing on light curves of known identity and morphology, we show that our technique is capable of recovering flaring status in 69% of all light curves containing a flare event above thresholds drawn to include <1% of any contaminant population. By applying to Palomar Transient Factory data, we show consistency with prior samples of flaring stars, and identify a small selection of candidate flaring G-type stars for possible follow-up.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2022

Estimate of the Mass and Radial Profile of the Orphan–Chenab Stream's Dwarf-galaxy Progenitor Using MilkyWay@home

We fit the mass and radial profile of the Orphan–Chenab Stream's (OCS) dwarf-galaxy progenitor by using turnoff stars in the Sloan Digital Sky Survey and the Dark Energy Camera to constrain N-body simulations of the OCS progenitor falling into the Milky Way on the 1.5 PetaFLOPS MilkyWay@home distributed supercomputer. We infer the internal structure of the OCS's progenitor under the assumption that it was a spherically symmetric dwarf galaxy composed of a stellar system embedded in an extended dark matter halo. We optimize the evolution time, the baryonic and dark matter scale radii, and the baryonic and dark matter masses of the progenitor using a differential evolution algorithm. The likelihood score for each set of parameters is determined by comparing the simulated tidal stream to the angular distribution of OCS stars observed in the sky. We fit the total mass of the OCS's progenitor to (2.0 ± 0.3) × 10 7 M ⊙ with a mass-to-light ratio of γ = 73.5 ± 10.6 and (1.1 ± 0.2) × 10 6 M ⊙ within 300 pc of its center. Within the progenitor's half-light radius, we estimate a total mass of (4.0 ± 1.0) × 10 5 M ⊙ . We also fit the current sky position of the progenitor's remnant to be (α, δ) = ((166.0 ± 0.9)°, (–11.1 ± 2.5)°) and show that it is gravitationally unbound at the present time. The measured progenitor mass is on the low end of previous measurements and, if confirmed, lowers the mass range of ultrafaint dwarf galaxies. Our optimization assumes a fixed Milky Way potential, OCS orbit, and radial profile for the progenitor, ignoring the impact of the Large Magellanic Cloud.

79 ASTRONOMY AND ASTROPHYSICS↗