Search NASA⌕ Search

SEARCH · Search NASA

Results for “gradient methods”

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 505 records · Page 28

Phase-field modeling of dislocation–interstitial interactions

The mechanical behavior of body-centered cubic (BCC) materials can be dramatically affected by the presence of interstitial solute atoms. Here we present a new phase-field dislocation dynamics formulation to include the diffusion of interstitials. Short-range interactions are accounted for by a concentration-dependent lattice energy, and long-range interactions are accounted for by modifications to the elastic energy. The interstitial diffusion law introduces gradients that require methods for minimizing Gibbs oscillations, which is done via a modified Green’s function. The formulation is general to any solute-solvent system and is applied here to Nb-O as a model system, whose interstitial parameters are obtained from ab initio calculations. The effect of O on the core structures of Nb edge and screw dislocations is calculated. The diffusion of O to form interstitial atmospheres around dislocation cores is simulated, as well as the critical stresses required for dislocations to break away or cross slip from these atmospheres. Future applications of the method to simulate complex interstitial embrittlement mechanisms are discussed.

36 MATERIALS SCIENCE↗

Identifying common stored product insects using automated deep learning methods

Monitoring stored product insect pests is a common practice for post-harvest management of stored grain and grain-based commodities, which helps ensure product quality from harvest to final consumer. Current methods of sampling and monitoring can be time-consuming, labor-intensive, expensive and require expertise in insect identification. Therefore, this study aims to develop an image-based automated identification system for common stored product insect species using deep-learning methods. Top-down images of the common stored product adult insect species of Rhyzopertha dominica, Cryptolestes ferrugineus, Tribolium castaneum, Sitophilus oryzae, and Oryzaephilus surinamensis were acquired and analyzed. Deep learning-based, state-of-the-art Convolutional Neural Networks (CNN) models (ResNet-50, MobileNet-v2, DarkNet-53, and EfficientNet-b0) were fine-tuned with a transfer learning approach to classify the insect species. All models were able to correctly identify the insect species with at least 96% accuracy and with few misclassifications. One issue with trained CNNs is that they do not explain the reasoning for the classification and are often called a “black box”. Therefore, visualization methods called Gradient-weighted Class Activation Mapping (Grad-CAM) were implemented to explore the black box network. The Grad-CAM uses heat maps to highlight the major image features that the network focused on to make insect species predictions. The Grad-CAM verifies the network's prediction and also helps improve network performance. This study contributes to the overall goal of developing a camera-based system for monitoring stored grain insects. As a result, the developed system would empower warehouse, flour mills, and other food facilities with a tool to quickly and accurately identify insect species in stored product environments and could be implemented as part of a close to real-time monitoring system.

60 APPLIED LIFE SCIENCES↗

In-depth quantitative analysis of the microstructures produced by Surface Mechanical Attrition Treatment (SMAT)

Electron BackScattered Diffraction (EBSD) maps are used to characterize quantitatively the graded microstructure formed by Surface Mechanical Attrition Treatment (SMAT) and applied here to the 316L stainless steel. In particular, the analysis of GNDs – coupled with relevant and reliable criteria – was used to depict the thickness of each zone identified in the SMAT-affected layers: (i) the “ultrafine grain” (UFG) zone present at the extreme top surface, (ii), the “transition zone” where grains were fragmented under the heavy plastic deformation and, finally, (iii) the “deformed zone” where initial grains are simply deformed. The interest of this procedure is illustrated through the comparative analysis of the effect of some SMAT processing parameters (amplitude of vibration and treatment duration). The UFG and transition zones are more significantly modified than the overall affected thickness under our tested conditions. - Highlights: • EBSD maps are used to characterize quantitatively the microstructure of SMAT treated samples. • Calculation of the GND density to quantify strain gradients • A new method to depict the different zone thicknesses in the SMAT affected layer • Effects of SMAT processing parameters on the surface microstructure evolution.

36 MATERIALS SCIENCE↗

Spontaneous imbibition in a square tube with corner films: theoretical model and numerical simulation

Spontaneous imbibition in an angular tube with corner films is a fundamental problem in many scientific and engineering processes. In this paper, a modified interacting capillary bundle model is developed to describe the liquid imbibition dynamics in a square tube with corner films. The square tube is decomposed into several interacting sub-capillaries and the local capillary pressure in each sub-capillary is derived based on the specific shape of its meniscus. The conductance of each sub-capillary is calculated using single-phase lattice Boltzmann simulation. The modified interacting capillary bundle model and color-gradient lattice Boltzmann method are used to simulate the liquid imbibition dynamics in the square tube with different fluid properties. The predictions by the modified interacting capillary bundle model match well with the lattice Boltzmann simulation results for different conditions, demonstrating the accuracy and robustness of the interacting capillary bundle model to describe the imbibition dynamics with corner films. In addition, the interacting capillary bundle model is helpful to investigate the mechanisms during spontaneous imbibition and the influences of fluid viscosity, surface tension, wetting phase contact angle and gravity on imbibition dynamics. Finally, a universal scaling law of imbibition dynamics for the main meniscus is developed and the scaling law for arc meniscus is also analyzed.

42 ENGINEERING↗

Localized Opto-Mechanical Control of Protein Adsorption onto Carbon Nanotubes

Chemical reactions can be described by an energy diagram along a reaction coordinate in which an activation barrier limits the rate at which reactants can be transformed into products. This reaction impedance can be overcome by reducing the magnitude of the barrier through the use of catalysis, increasing the thermal energy of the system, or through macroscopic mechanical processes. Here, we demonstrate direct molecular-scale control of a reaction through the precise application of opto-mechanical work. The method uses optical gradient forces generated in the evanescent field surrounding hybrid photonic-plasmonic structures to drive an otherwise unlikely adsorption reaction between proteins and carbon nanotubes. The adsorption of immunoglobulins on carbon nanotubes is used as a model reaction and investigated with an extended DLVO theory. The technique is also used to force a Förster resonance energy transfer between fluorophores on mismatched immunoglobulin proteins and is expected to lead to novel forms of chemical synthesis.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

High-speed adaptive contact-mode atomic force microscopy imaging with near-minimum-force

In this paper, an adaptive contact-mode imaging approach is proposed to replace the traditional contact-mode imaging by addressing the major concerns in both the speed and the force exerted to the sample. The speed of the traditional contact-mode imaging is largely limited by the need to maintain precision tracking of the sample topography over the entire imaged sample surface, while large image distortion and excessive probe-sample interaction force occur during high-speed imaging. In this work, first, the image distortion caused by the topography tracking error is accounted for in the topography quantification. Second, the quantified sample topography is utilized in a gradient-based optimization method to adjust the cantilever deflection set-point for each scanline closely around the minimal level needed for maintaining stable probe-sample contact, and a data-driven iterative feedforward control that utilizes a prediction of the next-line topography is integrated to the topography feeedback loop to enhance the sample topography tracking. The proposed approach is demonstrated and evaluated through imaging a calibration sample of square pitches at both high speeds (e.g., scan rate of 75 Hz and 130 Hz) and large sizes (e.g., scan size of 30 μm and 80 μm). The experimental results show that compared to the traditional constant-force contact-mode imaging, the imaging speed can be increased by over 30 folds (with the scanning speed at 13 mm/s), and the probe-sample interaction force can be reduced by more than 15% while maintaining the same image quality.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

First order nonadiabatic coupling matrix elements between excited states: Implementation and application at the TD-DFT and pp-TDA levels

The recently proposed rigorous yet abstract theory of first order nonadiabatic coupling matrix elements (fo-NACME) between electronically excited states [Z. Li and W. Liu, J. Chem. Phys. 141, 014110 (2014)] is specified in detail for two widely used models: The time-dependent density functional theory and the particle-particle Tamm-Dancoff approximation. The actual implementation employs a Lagrangian formalism with atomic-orbital based direct algorithms, which makes the computation of fo-NACME very similar to that of excited-state gradients. Although the methods have great potential in investigating internal conversions and nonadiabatic dynamics between excited states of large molecules, only prototypical systems as a first pilot application are considered here to illustrate some conceptual aspects.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Ab-initio study of B{sub 2}-type technetium AB (A=Tc, B=Nb and Ta) intermetallic compounds

The structural, electronic and elastic properties of AB type (A = Tc, B = Nb and Ta) technetium intermetallic compounds are studied using full potential linearized plane wave (FP-LAPW) method within generalized gradient approximation (GGA). The calculated lattice parameters agree well with the experimental results. The elastic constants obey the stability criteria for cubic system. Ductility for these compounds has been analyzed using the Pugh’s rule and Cauchy’s pressure and found that all the compounds are ductile in nature. Bonding nature is discussed in terms of Fermi surface and band structures.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Electronic structure, elasticity, bonding features and mechanical behaviour of zinc intermetallics: A DFT study

The structural stability, electronic structure, elastic and mechanical properties of TiZn and ZrZn intermetallics have been studied using ab-initio full potential linearized augmented plane wave (FP-LAPW) method within generalized gradient approximation for exchange and correlation potentials. The various structural parameters, such as lattice constant (a{sub 0}), bulk modulus (B), and its pressure derivative (B’) are analysed and compared. The investigation of elastic constants affirm that both TiZn and ZrZn are elastically stable in CsCl (B{sub 2} phase) structure. The electronic structures have been analysed quantitatively from the band structure which reveals the metallic nature of these compounds. To better illustrate the nature of bonding and charge transfer, we have also studied the Fermi surfaces. The three well known criterion of ductility namely Pugh’s rule, Cauchy’s pressure and Frantsevich rule elucidate the ductile nature of these compounds.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

A first principles study of half-metallic ferromagnetism in In{sub 1-x}Ti{sub x}P (x = 0.06) diluted magnetic semiconductor

A first principles approach has been used to calculate the electronic and magnetic properties of In{sub 1-x}Ti{sub x}P (x = 0.06) diluted magnetic semiconductor (DMS) compound. The calculations have been carried out using the highly precise all electron full potential Linear Augmented Plane Wave (FPLAPW) method within generalized gradient approximation (GGA) as exchange-correlation (XC) potentials. The estimated results show that the Ti-doping generate robust half metallic ferromagnetism with the 100% spin polarization at Fermi level (E{sub F}) in InP. Due to this peculiar property, the resultant compound behaves as true half-metallic ferromagnet which is best suited for spintronic applications. The total magnetic moments of this compound are mainly due to Ti-d states present at E{sub F} with almost negligible contribution from other atoms.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

EMPIRICAL DECODING OF THE SHAPES OF SPECTRAL-LINE BISECTORS

The great variety of shapes found for bisectors of spectral lines seen across the cool half of the Hertzsprung-Russell diagram are explained and empirically decoded using a novel method. The velocity gradient through the photosphere, as delineated by the star's third-signature plot, sets the basic shape of the bisector, which is then modified by the asymmetry in the amount of light coming from rising versus falling material. With this concept in place, one can then derive the apparent deficit of flux from downward flowing material in the granulation structure, and this is done for nine stars. The deficits peak between 3 and 4 km s{sup -1} for dwarfs and between 4 and 5 km s{sup -1} for giants, and they span a range from {approx}1 to 10 km s{sup -1} for all stars. The amplitudes of the flux deficits increase with effective temperature and with luminosity and have values {approx}7%-10% for dwarfs and {approx}15%-20% for giants. Such measurements of the contrast between granules and dark lanes are independent of models of granulation.

79 ASTRONOMY AND ASTROPHYSICS↗

High mass and halo resolution from fast low resolution simulations

Generating mocks for future sky surveys requires large volumes and high resolutions, which is computationally expensive even for fast simulations. Here we try to develop numerical schemes to calibrate various halo and matter statistics in fast low resolution simulations compared to high resolution N-body and hydrodynamic simulations. For the halos, we improve the initial condition resolution and develop a halo finder "relaxed-FoF", where we allow different linking lengths for different halo mass and velocity dispersions. We show that our relaxed-FoF halo finder improves the common statistics, such as halo bias, halo mass function, halo auto power spectrum, cross correlation coefficient with the reference halo catalog, and halo-matter cross power spectrum. We also calibrate small-scale velocities of small halos to improve the power spectrum in redshift space. For the matter statistics, we incorporate the potential gradient descent (PGD) method into fast simulations to improve the matter distribution at nonlinear scales. By building a lightcone output, we show that the PGD method significantly improves the weak lensing convergence tomographic power spectrum. With these improvements FastPM is comparable to the high resolution full N-body simulation of the same mass resolution, with two orders of magnitude fewer time steps. These techniques can be used to improve the halo and matter statistics of FastPM simulations for mock catalogs of future surveys such as DESI and LSST.

79 ASTRONOMY AND ASTROPHYSICS↗

Indigo: A Domain-Specific Language for Fast, Portable Image Reconstruction

© 2018 IEEE. Linear operators used in iterative methods like conjugate gradient have typically been implemented either as ''matrix-driven'' subroutines backed by explicit sparse or dense matrices, or as ''matrix-free'' subroutines that implement specific linear operations directly (e.g. FFTs). The matrix-driven approach is generally more portable because it can target widely-Available BLAS libraries, but it can be inefficient in terms of time and space complexity. In contrast, the matrix-free approach is more performant because it leverages structure in operations, but it requires each operator be re-implemented on each new platform. To increase performance and portability, we propose a hybrid approach that represents linear operators as expression trees. Leaf nodes in the tree are either matrix-free or matrix-driven operators, and interior nodes represent mathematical compositions (sums, products, transposes) or structural compositions (stacks, block diagonals, etc.) of the leaf operators. This representation enables expert-guided reordering and fusion transformations that can improve performance or reduce memory pressure. We implement our approach in a domain-specific language called Indigo. We assess Indigo on image reconstruction problems arising in four application areas: magnetic resonance imaging, ptychography, magnetic particle imaging, and fluorescent microscopy. We give performance results from vendor BLAS libraries, and we introduce specializations to Sparse BLAS routines that achieve near-Roofline performance on multi-core, many-core, and GPU systems.

Driscoll, Michael↗

Adaptive Quantum Generative Training using an Unbounded Loss Function

We propose a generative quantum learning algorithm using the Adaptive Derivative-Assembled Problem Tailored ansatz (ADAPT) framework in which the loss function to be minimized is the maximal quantum Rényi divergence of order two, an unbounded function that mitigates barren plateaus which inhibit training variational circuits. We benchmark this method against other state-of-the-art adaptive algorithms by learning random two-local thermal states. We perform numerical experiments of up to 12 qubits comparing our method learning algorithms that use linear objective functions and show that Rényi-ADAPT is capable of constructing shallow quantum circuits competitive with existing methods, while the gradients remain favorable resulting from the maximal Rényi divergence loss function.

quantum algorithms, quantum machine learning, quan↗

SU-C-18A-05: Registration Accuracy of MR-Based Images to On-Board Megavoltage Cone-Beam CT for Brain Patient Setup

Purpose: To quantify the difference in isocenter shifts when co-registering MR and MR-based pseudo CTs (pCT) with on-board megavoltage conebeam CT (CBCT) images. Methods: Fast Spoiled Gradient Echo MRs were used to generate pCTs (research version of Advantage Sim MD™, GE Healthcare) for ten patients who had prior brain radiotherapy. The planning CT (rCT) for each was co-registered with the MR, and the plan isocenter and two other reference points were transferred to the MR and pCT. CBCT images (with the machine isocenter) from a single treatment day were coregistered with the 3 test images (MR, pCT and rCT), by two observers and by an automated registration algorithm. The reference points were used to calculate patient shifts and rotations from the registrations. The shifts calculated from the test image registrations were compared to each other and to the shifts performed by the therapists who treated the patients on that day. Results: The average difference in absolute value between the isocenter shifts from the MR-, pCT- and rCT-CBCT registrations, and the therapist shifts, were 2.02, 3.01 and 0.89 mm (craniocaudal), 1.14, 1.34 and 0.46 mm (lateral), and 1.37, 3.43 and 1.43 mm (vertical), respectively. The MR- and pCT-CBCT registrations differed by 1.99, and 2.53 mm (craniocaudal), 1.36, and 1.37 mm (lateral), and 0.74 and 2.34 mm (vertical), respectively, from the average rCT-CBCT shifts. On average, differences of 2.39 (craniocaudal), 1.28 (lateral) and 2.84 mm (vertical) were seen between the MR and pCT shifts. Rotations relative to the CBCT coordinate system were on average <2° for the MR and rCT, and <6° for the pCT. Conclusion: In this study, FSPGR MR-CBCT registrations were more precise compared to the pCT-CBCT registrations. For improved accuracy, MR sequences that are optimal for bony anatomy visualization are necessary. GE healthcare has provided a research version of Advantage Sim MD to UCSF. No financial support was provided.

60 APPLIED LIFE SCIENCES↗

Robust Optimal Experimental Design of Infinite-Dimensional Bayesian Nonlinear Inverse Problems

Abstract. We consider robust optimal experimental design (ROED) for nonlinear Bayesian inverse problems governed by partial differential equations (PDEs). An optimal design is one that maximizes some utility quantifying the quality of the solution of an inverse problem. However, the optimal design is dependent on elements of the inverse problem such as the simulation model, the prior, or the measurement error model. ROED aims to produce an optimal design that is aware of the additional uncertainties encoded in the inverse problem and remains optimal even after variations in them. We follow a worst-case scenario approach to develop a new framework for robust optimal design of nonlinear Bayesian inverse problems. The proposed framework (a) is scalable and designed for infinite-dimensional Bayesian nonlinear inverse problems constrained by PDEs; (b) develops efficient approximations of the utility, namely the expected information gain; (c) employs eigenvalue sensitivity techniques to develop analytical forms and efficient evaluation methods of the gradient of the utility with respect to the uncertainties against which we wish to be robust; and (d) employs a probabilistic optimization paradigm that properly defines and efficiently solves the resulting combinatorial max-min optimization problem. The effectiveness of the proposed approach is illustrated for optimal sensor placement problem in an inverse problem governed by an elliptic PDE.

Chowdhary, Abhijit↗

Source-to-Source Automatic Differentiation of OpenMP Parallel Loops

This article presents our work toward correct and efficient automatic differentiation of OpenMP parallel worksharing loops in forward and reverse mode. Automatic differentiation is a method to obtain gradients of numerical programs, which are crucial in optimization, uncertainty quantification, and machine learning. The computational cost to compute gradients is a common bottleneck in practice. For applications that are parallelized for multicore CPUs or GPUs using OpenMP, one also wishes to compute the gradients in parallel. Here, we propose a framework to reason about the correctness of the generated derivative code, from which we justify our OpenMP extension to the differentiation model. We implement this model in the automatic differentiation tool Tapenade and present test cases that are differentiated following our extended differentiation procedure. Performance of the generated derivative programs in forward and reverse mode is better than sequential, although our reverse mode often scales worse than the input programs.

97 MATHEMATICS AND COMPUTING↗