Search NASA⌕ Search

SEARCH · Search NASA

Results for “Incremental Computing”

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 91 records · Page 5

A Computational/Experimental Study of Two Optimized Supersonic Transport Designs and the Reference H Baseline

Two supersonic transport configurations designed by use of non-linear aerodynamic optimization methods are compared with a linearly designed baseline configuration. One optimized configuration, designated Ames 7-04, was designed at NASA Ames Research Center using an Euler flow solver, and the other, designated Boeing W27, was designed at Boeing using a full-potential method. The two optimized configurations and the baseline were tested in the NASA Langley Unitary Plan Supersonic Wind Tunnel to evaluate the non-linear design optimization methodologies. In addition, the experimental results are compared with computational predictions for each of the three configurations from the Enter flow solver, AIRPLANE. The computational and experimental results both indicate moderate to substantial performance gains for the optimized configurations over the baseline configuration. The computed performance changes with and without diverters and nacelles were in excellent agreement with experiment for all three models. Comparisons of the computational and experimental cruise drag increments for the optimized configurations relative to the baseline show excellent agreement for the model designed by the Euler method, but poorer comparisons were found for the configuration designed by the full-potential code.

Cliff, Susan E.↗

Mitigating Catastrophic Forgetting in Deep Learning in a Streaming Setting Using Historical Summary

Recent advancements in scientific equipment and the adaptation of electronics and the Internet of Things (IoT) in our everyday lives resulted in large and complex data production at a high rate. Making meaningful and timely knowledge discovery at a modest cost from this big data is difficult for computing power and storage limitations. Training deep learning models incrementally in a streaming setting can help us with overcoming these limitations. However, in a well-known phenomenon named catastrophic forgetting, incrementally trained models increasingly perform poorly on the past data. To mitigate catastrophic forgetting in training in a streaming setting, we propose constructing a historical summary over time and use the summary with newly arrived data during incremental training. We propose various data summarization techniques such as random sampling, micro clustering, coreset computation, and Auto Encoders to counteract catastrophic forgetting. We built a pipeline for incremental training with a historical summary for training deep learning models for streaming data. We demonstrate the effectiveness of historical summary in mitigating catastrophic forgetting using three case studies involving three different deep learning applications: an Artificial Neural Network (ANN) for classification task on MNIST dataset, a language model (RNN-LM) on the WikiText2 dataset, and a Convolutional Neural Network (CNN), ResNet50 to classify the ImageNet dataset. Through the training of the models, we observe that catastrophic forgetting is evident in ANN and CNN but not in an RNN. For the first task, our method recovers up to 47.9% lost accuracy due to catastrophic forgetting. For the third task, the historical summary recovers classification accuracy by up to 25%. For the second task, though there is not proof of catastrophic forgetting, the training performance (PPL) improves by up to 26% with historical summary.

Dash, Sajal↗

Temperature Humidity and Sea Level Pressure Increments Induced by 1DVAR Analysis of GPS Refractivity

The Global Positioning System (GPS) transmitted signals are affected by the atmosphere. Using the radio occultation technique, where a receiver is placed on a low-Earth-orbiting platform. it is possible to perform soundings. by measuring the bending angles of the rays. The information can be converted into atmospheric refractivity. We have developed a one dimensional variational (1DVAR) analysis that uses GPS/MET 1995 refractivity and 6-hour FVDAS (Finite Volume Data Assimilation System) forecasts as background information to constrain the retrievals. The analysis increments are defined as 1DVAR minus background temperature, humidity and sea level pressure. Before assimilating the 1DVAR profiles into the FVDAS. the increments need to be understood. First, some bias could be induced in the retrievals when confronted with actual biased data: second. bias in the back-round could create undesired bias in the retrievals. Anv bias in the analyses will ultimately change the climatology of the model the retrievals will be assimilated into. We relate the increments to the reduction of the difference between observed minus computed refractivity profiles. We also point out the difference in the mean increments using backgrounds which have assimilated either NESDIS TIROS Operational Vertical Sounder (TOVS) operational retrievals or Data Assimilation Office (DAO) TOVS interactive retrievals. The climatology of the model in terms of refractivity is significantly different and this impacts the GPS 1DVAR increments. This testifies that changing the basic load of assimilated data has an influence on the impact the GPS data may have in a DAS.

Poli, Paul↗

An evaluation of point-insertion sequences for incremental Delaunay tessellations

Currently, incremental algorithms may be seen as the lowest-cost computational methods to generate Delaunay tessellations in several point distributions. In this work, eight point-insertion sequences in incremental algorithms for generating Delaunay tessellations are evaluated. More specifically, four point-insertion sequences in incremental algorithms for generating Delaunay tessellations are proposed: with orders given by the red–black tree with in-order and level-order traversals, spiral ordering, and H-indexing. These four incremental algorithms with such sequences are compared with four incremental algorithms with point-insertion orders given by the following sequences: the Hilbert and Lebesgue curves, cut-longest-edge kd-tree, and random order. Six 2-D and seven 3-D point distributions are tested, with sets ranging from 25,000 to 8,000,000 points. The results of computational and storage costs of these eight algorithms are analyzed. It follows that the incremental algorithm with a point-insertion sequence in the order given by the cut-longest-edge kd-tree shows the lowest computational and storage costs of the sequences tested.

97 MATHEMATICS AND COMPUTING↗

Resilience Evaluation and Enhancement for Island City Integrated Energy Systems

Extreme natural hazards, such as hurricanes or earthquakes, have a high probability of threatening energy supply security and causing high-order contingencies to island city-integrated energy systems (IC-IESs). To better evaluate and enhance resilience, a novel approach is proposed in this work for IC-IESs. The resilience of an IC-IES is analyzed from both the system level and the component level. At the system level, the impacts of extreme natural disasters are quantified. At the component level, the importance of individual components is analyzed through pre-failure and post-failure indices. The pre-failure index identifies the system’s weak links before an energy interruption, and the post-failure index determines the optimal repair strategy to restore the service. The proposed indices are solved by the impact increment method (IIM), which significantly improves computational efficiency without much affecting result accuracy. Numerical simulation studies are conducted on the modified Barry Island IES and IES E123-G48-H32 test systems. Furthermore, the results validate the effectiveness of the proposed approach.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Optimizing Management of Persistent Data Structures in High-Performance Analytics

Large-scale data analytics workflows ingest massive input data into various data structures, including graphs and key-value datastores. These data structures undergo multiple transformations and computations and are typically reused in incremental and iterative analytics workflows. Persisting in-memory views of these data structures enables reusing them beyond the scope of a single program run while avoiding repetitive raw data ingestion overheads. Memory-mapped I/O enables persisting in-memory data structures without data serialization and deserialization overheads. However, memory-mapped I/O lacks the key feature of persisting consistent snapshots of these data structures for incremental ingestion and processing. The obstacles to efficient virtual memory snapshots using memory-mapped I/O include background writebacks outside the application’s control, and the significantly high storage footprint of such snapshots. To address these limitations, we present Privateer, a memory and storage management tool that enables storage-efficient virtual memory snapshotting while also optimizing snapshot I/O performance. Here, we integrated Privateer into Metall, a state-of-the-art persistent memory allocator for C++, and the Lightning Memory-Mapped Database (LMDB), a widely-used key-value datastore in data analytics and machine learning. Privateer optimized application performance by 1.22× when storing data structure snapshots to node-local storage, and up to 16.7× when storing snapshots to a parallel file system. Privateer also optimizes storage efficiency of incremental data structure snapshots by up to 11× using data deduplication and compression.

Computer science↗

Analog computer implementation of four instantaneous distortion indices

Dynamic distortion data obtained from inlet engine compatibility test on a J85-GE-13 engine are used to establish the feasibility of using on-line generated indices as control signals. These data are analyzed over time increments of 1.3 seconds on an analog computer. The analog program of four instantaneous distortion indices and their results are presented. A modified approach in determining the extent of distortion is also presented.

Costakis, W. G.↗

VISCEL: A general-purpose computer program for analysis of linear viscoelastic structures (user's manual), volume 1

This program, an extension of the linear equilibrium problem solver ELAS, is an updated and extended version of its earlier form (written in FORTRAN 2 for the IBM 7094 computer). A synchronized material property concept utilizing incremental time steps and the finite element matrix displacement approach has been adopted for the current analysis. A special option enables employment of constant time steps in the logarithmic scale, thereby reducing computational efforts resulting from accumulative material memory effects. A wide variety of structures with elastic or viscoelastic material properties can be analyzed by VISCEL. The program is written in FORTRAN 5 language for the Univac 1108 computer operating under the EXEC 8 system. Dynamic storage allocation is automatically effected by the program, and the user may request up to 195K core memory in a 260K Univac 1108/EXEC 8 machine. The physical program VISCEL, consisting of about 7200 instructions, has four distinct links (segments), and the compiled program occupies a maximum of about 11700 words decimal of core storage.

Gupta, K. K.↗

A NEW IMPLEMENTATION OF THE MAGNETOHYDRODYNAMICS-RELAXATION METHOD FOR NONLINEAR FORCE-FREE FIELD EXTRAPOLATION IN THE SOLAR CORONA

The magnetic field in the solar corona is usually extrapolated from a photospheric vector magnetogram using a nonlinear force-free field (NLFFF) model. NLFFF extrapolation needs considerable effort to be devoted to its numerical realization. In this paper, we present a new implementation of the magnetohydrodynamics (MHD) relaxation method for NLFFF extrapolation. The magnetofrictional approach, which is introduced for speeding the relaxation of the MHD system, is realized for the first time by the spacetime conservation-element and solution-element scheme. A magnetic field splitting method is used to further improve the computational accuracy. The bottom boundary condition is prescribed by incrementally changing the transverse field to match the magnetogram, and all other artificial boundaries of the computational box are simply fixed. We examine the code using two types of NLFFF benchmark tests, the Low and Lou semi-analytic force-free solutions and a more realistic solar-like case constructed by van Ballegooijen et al. The results show that our implementation is successful and versatile for extrapolations of either the relatively simple cases or the rather complex cases that need significant rebuilding of the magnetic topology, e.g., a flux rope. We also compute a suite of metrics to quantitatively analyze the results and demonstrate that the performance of our code in extrapolation accuracy basically reaches the same level of the present best-performing code, i.e., that developed by Wiegelmann.

79 ASTRONOMY AND ASTROPHYSICS↗

Discrete event modeling and massively parallel execution of epidemic outbreak phenomena

In complex phenomena such as epidemiological outbreaks, the intensity of inherent feedback effects and the significant role of transients in the dynamics make simulation the only effective method for proactive, reactive or post facto analysis. The spatial scale, runtime speed, and behavioral detail needed in detailed simulations of epidemic outbreaks cannot be supported by sequential or small-scale parallel execution, making it necessary to use large-scale parallel processing. Here, an optimistic parallel execution of a new discrete event formulation of a reaction–diffusion simulation model of epidemic propagation is presented to facilitate a dramatic increase in the fidelity and speed by which epidemiological simulations can be performed. Rollback support needed during optimistic parallel execution is achieved by combining reverse computation with a small amount of incremental state saving. Parallel speedup of over 5,500 and other runtime performance metrics of the system are observed with weak-scaling execution on a small (8,192-core) Blue Gene/P system, while scalability with a weak-scaling speedup of over 10,000 is demonstrated on 65,536 cores of a large Cray XT5 system. Scenarios representing large population sizes, with mobility and detailed state evolution modeled at the level of each individual, exceeding several hundreds of millions of individuals in the largest cases, are successfully exercised to verify model scalability.

Computer Science↗

Time-Efficient Qudit Gates through Incremental Pulse Re-seeding

Current efforts to build quantum computers focus mainly on the two-state qubit, which often involves suppressing readily-available higher states. In this work, we break this abstraction and synthesize short-duration control pulses for gates on generalized d-state qudits. We present Incremental Pulse Reseeding, a practical scheme to guide optimal control software to the lowest-duration pulse by iteratively seeding the optimizer with previous results. We find a near-linear relationship between Hilbert space dimension and gate duration through explicit pulse optimization for one- and two-qudit gates on transmons. Furthermore, our results suggest that qudit operations are much more efficient than previously expected in the practical regime of interest and have the potential to significantly increase the computational power of current hardware.

pulse synthesis↗

Towards Rapid Interactive Machine Learning: Evaluating Tradeoffs of Classification without Representation

Our contribution is the design and evaluation of an interac- tive machine learning interface that rapidly provides the user with model feedback after every interaction. To address visual scalability, this interface communicates with the user via a “tip of the iceberg” approach, where the user interacts with a small set of recommended instances for each class. To address computational scalability, we developed an O(n) classification algorithm that incorporates user feedback incrementally, and without consulting the underlying data’s representation matrix. Our computational evaluation showed that this algorithm has similar accuracy to several off-the-shelf classification algo- rithms with small amounts of labeled data. Empirical evalu- ation revealed that users performed better using our design compared to an equivalent active learning setup.

Arendt, Dustin L.↗

Circumventing mesh bias by r- and h-adaptive techniques for variational eigenfracture

This work introduces and compares mesh r- and h-adaptivity for the eigenfracture model originally proposed in Schmidt et al. (Multiscale Model Simul 7:1237–1266, 2009), Pandolfi and Ortiz (J Numer Methods Eng 92:694–714, 2012), with the goal of suppressing potential mesh bias due to the element deletion. In the r-adaptive approach, we compute the configurational force at each incremental step and move nodes near the crack tip parallel to the normalized configurational forces field such that the crack propagation direction can be captured more accurately within each incremental step. In the h-adapative approach, we introduce mesh refinement via a quad-tree algorithm to introduce more degrees of freedom within the nonlocal ϵ-neighborhood such that a more refined crack path can be reproduced with a higher mesh resolution. Our numerical examples indicate that the r-adaptive approach is able to replicate curved cracks and complex geometrical features, whereas the h-adaptive approach is advantageous in simulating sub-scale fracture when the nonlocal regions are smaller than the un-refined coarse mesh.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

On the Optimal Stabilization of an Integral Manifold

By the method of Lyapunov functions, we study the problem of optimal stabilization of an analytically defined integral manifold in the class of stochastic differential equations in the case where random perturbations belong to the class of processes with independent increments.

97 MATHEMATICS AND COMPUTING↗

Bayesian Tensor Decompositions for Scalable Supervised Learning of Scientific Data (Final Report)

In this document we highlight the detailed accomplishments and progress that we have made in this period. This progress seeks to address the three main objectives to provide new algorithms for quantifying uncertainty in low-multilinear-rank models and to leverage them for data analysis. These include: (1) develop probabilistic models for low-multilinear-rank functions; (2) develop a suite of Bayesian learning approaches to learn the probabilistic models from data; (3) apply the techniques on challenging problems arising in DOE-relevant applications.

97 MATHEMATICS AND COMPUTING↗

User's guide for analysis of finite elastoplastic deformation: The FIPDEF and FIPAX programs for the CDC 6600

Computer programs are presented which provide incremental finite-element analysis capability for problems of quasi-static, finite, elastoplastic deformation in two spatial dimensions (plane strain, plane stress, axisymmetric). Monotonic or cyclic loading of isotropic hardening materials is considered. The only restriction on the form of the stress-strain curve is that the rate of work hardening exceed some small positive value. The user's guide assumes familiarity with both finite-element analysis and FORTRAN IV programming for the CDC 6600. Sufficient information is provided to support problem solving ultization of the programs.

Osias, J. R.↗

The influence of crack closure and elasto-plastic flow on the bending of a cracked plate

The influence of crack closure and elastoplastic flow on the bending behavior of thin cracked plates is investigated using an incremental elastoplastic plate bending finite element computer program. The finite element program was developed using assumptions consistent with Kirchhoff fourth-order plate theory while the material property treatment permits general isotropic work hardening with local elastic unloading. This technique is applied to the problem of a large centrally through cracked plate subject to remote circular bending. Comparison is drawn between two cases of the bending problem. The first neglects the possibility of crack face interference with bending, and the second includes a kinematic prohibition against the crack face from passing through the symmetry plane. Results are reported which isolate the effects of elastoplastic flow and crack closure.

Jones, D. P.↗

Circular earth orbits attainable with fixed two-impulse expendable tug

The performance of solid rocket motors in expendable stages as an alternative concept for Shuttle interim upper stage applications is assessed from the standpoint of mission flexibility. The system comprises two stages and a payload, released from a circular Shuttle orbit at 160 naut miles altitude. The stages are sized to deliver a given payload mass to synchronous equatorial orbit. Only nonreturn missions are considered. A parametric sweep was made with small increments of pitch and yaw to compute the possible transfer orbits with an apogee greater than 160 naut miles. At several points on each transfer orbit the possibility of attaining a circular orbit with the second burn impulse of 6000 fps was assessed. Constraints on the vertical component of the second impulse and on the total velocity after firing determine the pitch angle for second burn. Results indicate a fixed two-impulse system designed to place a payload into synchronous equatorial orbit from the Shuttle orbit can deliver the payload mass into any lower circular orbit at any inclination within about 30 deg of the Shuttle orbit. The only apparent disadvantage of the system is an inability to achieve circular orbits above geosynchronous altitudes with the design payload.

Porter, R. F.↗