Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “algorithmic efficiency”

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 379 records · Page 21

Hybrid Model Based Approaches for Systems Health Management and Prognostics

This is a previously approved and published presentation. To accurately predict the future state of any system, it is required to possess knowledge of its current health state and future operational conditions. Present achievements of data-driven algorithms in regression of complex nonlinear functions and classification tasks have generated a growing interest in artificial intelligence for industrial applications. Complex multi-physics models as well as digital twins, once purely built on physics and corresponding simplified lumped parameter iterations, can now benefit from machine learning algorithms to mitigate the lack of understanding of some complex behavior. Given models of the current and future system behavior, a general approach of model-based prognostics can solve the prediction problem and further decision-making. In principle, data-driven approaches can replace expensive experimental test-setups as well as reduce the number of simulations needed to explore, e.g., the parametric space of a multi-parameter model. Nonetheless, the limitations of pure data-driven methods came to light rather quickly, at least for some industries. In many industrial applications, data acquisition is costly, and the volume of data that can be collected does not satisfy the requirements for effective model training and cross-validation. Therefore, some recent works in the area of machine learning is focusing on blending physics with data-driven algorithms, thus mitigating the drawbacks of the two approaches and emphasizing respective advantages. Partial physical knowledge of the problem can aid the learning process by “guiding” the algorithm towards efficient solutions that satisfy the physics driving the system behavior. The result is a hybrid modeling approach combining physical knowledge as well data-driven methods to develop a unified hybrid approach. A hybrid framework for fusing information from physics-based performance models along with deep learning algorithms for prognostics of complex safety-critical systems is presented. In this framework, physics-based performance models infer unobservable model parameters related to the system's components health solving a calibration problem in the deep learning approach.

Prognostics↗

Testing Linear Temporal Logic Formulae on Finite Execution Traces

We present an algorithm for efficiently testing Linear Temporal Logic (LTL) formulae on finite execution traces. The standard models of LTL are infinite traces, reflecting the behavior of reactive and concurrent systems which conceptually may be continuously alive. In most past applications of LTL. theorem provers and model checkers have been used to formally prove that down-scaled models satisfy such LTL specifications. Our goal is instead to use LTL for up-scaled testing of real software applications. Such tests correspond to analyzing the conformance of finite traces against LTL formulae. We first describe what it means for a finite trace to satisfy an LTL property. We then suggest an optimized algorithm based on transforming LTL formulae. The work is done using the Maude rewriting system. which turns out to provide a perfect notation and an efficient rewriting engine for performing these experiments.

Havelund, Klaus↗

Monitoring Programs Using Rewriting

We present a rewriting algorithm for efficiently testing future time Linear Temporal Logic (LTL) formulae on finite execution traces, The standard models of LTL are infinite traces, reflecting the behavior of reactive and concurrent systems which conceptually may be continuously alive in most past applications of LTL, theorem provers and model checkers have been used to formally prove that down-scaled models satisfy such LTL specifications. Our goal is instead to use LTL for up-scaled testing of real software applications, corresponding to analyzing the conformance of finite traces against LTL formulae. We first describe what it means for a finite trace to satisfy an LTL property end then suggest an optimized algorithm based on transforming LTL formulae. We use the Maude rewriting logic, which turns out to be a good notation and being supported by an efficient rewriting engine for performing these experiments. The work constitutes part of the Java PathExplorer (JPAX) project, the purpose of which is to develop a flexible tool for monitoring Java program executions.

Havelund, Klaus↗

Analysis and an image recovery algorithm for ultrasonic tomography system

The problem of an ultrasonic reflectivity tomography is similar to that of a spotlight-mode aircraft Synthetic Aperture Radar (SAR) system. The analysis for a circular path spotlight mode SAR in this paper leads to the insight of the system characteristics. It indicates that such a system when operated in a wide bandwidth is capable of achieving the ultimate resolution; one quarter of the wavelength of the carrier frequency. An efficient processing algorithm based on the exact two dimensional spectrum is presented. The results of simulation indicate that the impulse responses meet the predicted resolution performance. Compared to an algorithm previously developed for the ultrasonic reflectivity tomography, the throughput rate of this algorithm is about ten times higher.

Jin, Michael Y.↗

Topography Modeling in Atmospheric Flows Using the Immersed Boundary Method

Numerical simulation of flow over complex geometry needs accurate and efficient computational methods. Different techniques are available to handle complex geometry. The unstructured grid and multi-block body-fitted grid techniques have been widely adopted for complex geometry in engineering applications. In atmospheric applications, terrain fitted single grid techniques have found common use. Although these are very effective techniques, their implementation, coupling with the flow algorithm, and efficient parallelization of the complete method are more involved than a Cartesian grid method. The grid generation can be tedious and one needs to pay special attention in numerics to handle skewed cells for conservation purposes. Researchers have long sought for alternative methods to ease the effort involved in simulating flow over complex geometry.

Ackerman, A. S.↗

A SAR correlation algorithm which accommodates large-range migration

This paper presents an algorithm for azimuth correlation of synthetic aperture radar (SAR) data with extraordinarily large-range migration which cannot be accommodated by the existing frequency domain interpolation approach used in current Seasat SAR processing. A mathematical model is provided for the SAR range-correlated point-target response both in the spatial and frequency domains. A simple and efficient processing algorithm derived from the exact two-dimensional correlation algorithm is given. This algorithm enables azimuth correlation by two cascaded one-dimensional correlation steps. The first step is to transform the range-correlated SAR data into the frequency domain in azimuth, followed by a newly developed range convolution filter to correct the range-dispersed spectrum of the range-correlated point-target response (RCPTR). The second step is the frequency-domain range migration correction approach for the azimuth compression. Simulation results provided here indicate that this processing algorithm yields a satisfactory compressed impulse response for SAR data with large-range migration whereas previous methods were subject to a broadening of the impulse response along range.

Jin, M. Y.↗

SAR correlation technique - An algorithm for processing data with large range walk

This paper presents an algorithm for synthetic aperture radar (SAR) azimuth correlation with extraneously large range migration effect which can not be accommodated by the existing frequency domain interpolation approach used in current SEASAT SAR processing. A mathematical model is first provided for the SAR point-target response in both the space (or time) and the frequency domain. A simple and efficient processing algorithm derived from the hybrid algorithm is then given. This processing algorithm enables azimuth correlation by two steps. The first step is a secondary range compression to handle the dispersion of the spectra of the azimuth response along range. The second step is the well-known frequency domain range migration correction approach for the azimuth compression. This secondary range compression can be processed simultaneously with range pulse compression. Simulation results provided here indicate that this processing algorithm yields a satisfactory compressed impulse response for SAR data with large range migration.

Jin, M.↗

Progress in the Simulation of Steady and Time-Dependent Flows with 3D Parallel Unstructured Cartesian Methods

The proposed paper will present recent extensions in the development of an efficient Euler solver for adaptively-refined Cartesian meshes with embedded boundaries. The paper will focus on extensions of the basic method to include solution adaptation, time-dependent flow simulation, and arbitrary rigid domain motion. The parallel multilevel method makes use of on-the-fly parallel domain decomposition to achieve extremely good scalability on large numbers of processors, and is coupled with an automatic coarse mesh generation algorithm for efficient processing by a multigrid smoother. Numerical results are presented demonstrating parallel speed-ups of up to 435 on 512 processors. Solution-based adaptation may be keyed off truncation error estimates using tau-extrapolation or a variety of feature detection based refinement parameters. The multigrid method is extended to for time-dependent flows through the use of a dual-time approach. The extension to rigid domain motion uses an Arbitrary Lagrangian-Eulerlarian (ALE) formulation, and results will be presented for a variety of two- and three-dimensional example problems with both simple and complex geometry.

Aftosmis, M. J.↗

Efficiency of group implicit concurrent algorithms for transient finite element analysis

The performance of group implicit algorithms is assessed on actual concurrent computers. It is shown that, as the number of subdomains is increased, performance enhancements are derived from two sources: the increased parallelism in the computations; and a reduction in equation solving effort. Moreover, these two performance enhancements are synergistic, in the sense that the corresponding speed-ups are multiplied, rather than merely added. Simulations on a 32-node hypercube are presented for which the interprocessor communications efficiencies obtained are consistently in excess of 90 percent.

Ortiz, M.↗

Algorithm and code development for unsteady three-dimensional Navier-Stokes equations

A streamwise upwind algorithm for solving the unsteady 3-D Navier-Stokes equations was extended to handle the moving grid system. It is noted that the finite volume concept is essential to extend the algorithm. The resulting algorithm is conservative for any motion of the coordinate system. Two extensions to an implicit method were considered and the implicit extension that makes the algorithm computationally efficient is implemented into Ames's aeroelasticity code, ENSAERO. The new flow solver has been validated through the solution of test problems. Test cases include three-dimensional problems with fixed and moving grids. The first test case shown is an unsteady viscous flow over an F-5 wing, while the second test considers the motion of the leading edge vortex as well as the motion of the shock wave for a clipped delta wing. The resulting algorithm has been implemented into ENSAERO. The upwind version leads to higher accuracy in both steady and unsteady computations than the previously used central-difference method does, while the increase in the computational time is small.

Obayashi, Shigeru↗

On current compositing algorithms

Several techniques exist for compositing the multitemporal NOAA Advanced Very High Resolution Radiometer (AVHRR) data for vegetation studies. The major pixel selection criteria of these techniques rely on the characteristics of the NDVI (Normalized Difference Vegetative Index): appearance of clouds, poor atmospheric conditions, and off-nadir viewing geometries would depress the NDVI values. Consequently, selecting the pixels with the maximum value of NDVI would presumably eliminate these external perturbating effects. However, the maximum NDVI does not always correspond to these ideal conditions. In fact, the NDVI varies with these external factors in an unpredictable way. There was an indication that the maximum NDVI tended to favor the off-nadir view in the forward direction. The resultant composite product would be consequently affected. To improve the multitemporal data via compositing, therefore, both the pixel selection criteria and the classifier NDVI need to be modified or corrected for external factors. The current compositing algorithms were reviewed, and alternatives were proposed to use the combinations of the red and near infrared channels and biological characteristics of vegetation as second criteria in pixel selections. The traditional classifier NDVI was replaced with different vegetation indices. The approach was applied to an AVHRR data set over the HAPEX study site in Niger in 1992. The results showed that the approach improved the AVHRR time series quality and was promising towards the development of an efficient compositing algorithm.

Qi, J.↗

The Micro-Pulse Lidar Network (MPLNET): A Federated Network of Micro-pulse Lidars and AERONET Sunphotometers

We present the formation of a new global-ground based eye-safe lidar network, the NASA Micro-Pulse Lidar Network (MPLNET). The aim of MPLNET is to acquire long- term observations of aerosol and cloud vertical profiles at unique geographic sites within the NASA Aerosol Robotic Network (AERONET). MPLNET utilizes standard instrumentation and data processing algorithms for efficient network operations and direct comparison of data between each site. The micro-pulse lidar is eye-safe, compact, and commercially available, and most easily allows growth of the network without sacrificing standardized instrumentation goals. Network growth follows a federated approach, pioneered by AERONET, wherein independent research groups may join MPLNET with their own instrument and site. MPLNET sites produce not only vertical profile data, but also column-averaged products already available from AERONET (aerosol optical depth, sky radiance, size distributions). Algorithms are presented for each MPLNET data product. Real-time Level 1 data products (next-day) include daily lidar signal images from the surface to -2Okm, and Level 1.5 aerosol extinction profiles at times co-incident with AERONET observations. Quality assured Level 2 aerosol extinction profiles are generated after screening the Level 1.5 results and removing bad data. Level 3 products include continuous day/night aerosol extinction profiles, and are produced using Level 2 calibration data. Rigorous uncertainty calculations are presented for all data products. Analysis of MPLNET data show the MPL and our analysis routines are capable of successfully retrieving aerosol profiles, with the strenuous accounting of uncertainty necessary for accurate interpretation of the results.

Welton, Ellsworth J.↗

The fast decoding of Reed-Solomon codes using number theoretic transforms

It is shown that Reed-Solomon (RS) codes can be encoded and decoded by using a fast Fourier transform (FFT) algorithm over finite fields. The arithmetic utilized to perform these transforms requires only integer additions, circular shifts and a minimum number of integer multiplications. The computing time of this transform encoder-decoder for RS codes is less than the time of the standard method for RS codes. More generally, the field GF(q) is also considered, where q is a prime of the form K x 2 to the nth power + 1 and K and n are integers. GF(q) can be used to decode very long RS codes by an efficient FFT algorithm with an improvement in the number of symbols. It is shown that a radix-8 FFT algorithm over GF(q squared) can be utilized to encode and decode very long RS codes with a large number of symbols. For eight symbols in GF(q squared), this transform over GF(q squared) can be made simpler than any other known number theoretic transform with a similar capability. Of special interest is the decoding of a 16-tuple RS code with four errors.

Reed, I. S.↗

PNS predictions for supersonic/hypersonic flows over finned missile configurations

Finned missile design entails accurate and computationally fast numerical techniques for predicting viscous flows over complex lifting configurations at small to moderate angles of attack and over Mach 3 to 15; these flows are often characterized by strong embedded shocks, so that numerical algorithms are also required to capture embedded shocks. The recent real-gas Flux Vector Splitting technique is here extended to investigate the Mach 3 flow over a typical finned missile configuration with/without side fin deflections. Elliptic grid-generation techniques for Mach 15 flows are shown to be inadequate for Mach 3 flows over finned configurations and need to be modified. Fin-deflection studies indicate that even small amounts of missile fin deflection can substantially modify vehicle aerodynamics. This 3D parabolized Navier-Stokes scheme is also extended into an efficient embedded algorithm for studying small axially separated flow regions due to strong fin and control surface deflections.

Bhutta, Bilal A.↗

Spray Combustion Modeling with VOF and Finite-Rate Chemistry

A spray atomization and combustion model is developed based on the volume-of-fluid (VOF) transport equation with finite-rate chemistry model. The gas-liquid interface mass, momentum and energy conservation laws are modeled by continuum surface force mechanisms. A new solution method is developed such that the present VOF model can be applied for all-speed range flows. The objectives of the present study are: (1) to develop and verify the fractional volume-of-fluid (VOF) cell partitioning approach into a predictor-corrector algorithm to deal with multiphase (gas-liquid) free surface flow problems; (2) to implement the developed unified algorithm in a general purpose computational fluid dynamics (CFD) code, Finite Difference Navier-Stokes (FDNS), with droplet dynamics and finite-rate chemistry models; and (3) to demonstrate the effectiveness of the present approach by simulating benchmark problems of jet breakup/spray atomization and combustion. Modeling multiphase fluid flows poses a significant challenge because a required boundary must be applied to a transient, irregular surface that is discontinuous, and the flow regimes considered can range from incompressible to highspeed compressible flows. The flow-process modeling is further complicated by surface tension, interfacial heat and mass transfer, spray formation and turbulence, and their interactions. The major contribution of the present method is to combine the novel feature of the Volume of Fluid (VOF) method and the Eulerian/Lagrangian method into a unified algorithm for efficient noniterative, time-accurate calculations of multiphase free surface flows valid at all speeds. The proposed method reformulated the VOF equation to strongly couple two distinct phases (liquid and gas), and tracks droplets on a Lagrangian frame when spray model is required, using a unified predictor-corrector technique to account for the non-linear linkages through the convective contributions of VOF. The discontinuities within the sharp interface will be modeled as a volume force to avoid stiffness. Formations of droplets, tracking of droplet dynamics and modeling of the droplet breakup/evaporation, are handled through the same unified predictor-corrector procedure. Thus the new algorithm is non-iterative and is flexible for general geometries with arbitrarily complex topology in free surfaces. The FDNS finite-difference Navier-Stokes code is employed as the baseline of the current development. Benchmark test cases of shear coaxial LOX/H2 liquid jet with atomization/combustion and impinging jet test cases are investigated in the present work. Preliminary data comparisons show good qualitative agreement between data and the present analysis. It is indicative from these results that the present method has great potential to become a general engineering design analysis and diagnostics tool for problems involving spray combustion.

Chen, Yen-Sen↗

Multi-Resolution Elevation Mapping and Safe Landing Site Detection with Applications to Planetary Rotorcraft

In this paper, we propose a resource-efficient approach to provide an autonomous UAV with an on-board perception method to detect safe, hazard-free landing sites during flights over complex 3D terrain. We aggregate 3D measurements acquired from a sequence of monocular images by a Structure-from-Motion approach into a local, robot-centric, multi-resolution elevation map of the overflown terrain, which fuses depth measurements according to their lateral surface resolution (pixel-footprint) in a probabilistic framework based on the concept of dynamic Level of Detail. Map aggregation only requires depth maps and the associated poses, which are obtained from an on-board Visual Odometry algorithm. An efficient landing site detection method then exploits the features of the underlying multi-resolution map to detect safe landing sites based on the slope, roughness, and quality of the reconstructed terrain surface. The evaluation of the performance of the mapping and landing site detection modules are analyzed independently and jointly in simulated and realworld experiments in order to establish the efficacy of the proposed approach.

Brockers, Roland↗

AND/OR graph representation of assembly plans

A compact representation of all possible assembly plans of a product using AND/OR graphs is presented as a basis for efficient planning algorithms that allow an intelligent robot to pick a course of action according to instantaneous conditions. The AND/OR graph is equivalent to a state transition graph but requires fewer nodes and simplifies the search for feasible plans. Three applications are discussed: (1) the preselection of the best assembly plan, (2) the recovery from execution errors, and (3) the opportunistic scheduling of tasks. An example of an assembly with four parts illustrates the use of the AND/OR graph representation in assembly-plan preselection, based on the weighting of operations according to complexity of manipulation and stability of subassemblies. A hypothetical error situation is discussed to show how a bottom-up search of the AND/OR graph leads to an efficient recovery.

Homem De Mello, Luiz S.↗

Adaptive Stress Testing: Using Reinforcement Learning to Find Failures in Safety-Critical Systems

Emerging applications in artificial intelligence, such as driverless cars and autonomous aircraft promise to be more efficient, cheaper to operate, and always available. However, ensuring the safety of these systems remains a major challenge to their certification and adoption. These autonomous systems are expected to routinely make safety-critical decisions where failures can have serious consequences including loss of life and property. Testing and validation techniques aim to identify and diagnose potential failures before the system is deployed. However, finding failure scenarios in autonomous systems can be very challenging due to high-dimensional and continuous state spaces, interaction with large environments over many time steps, and the rarity of failures. This talk presents Adaptive Stress Testing (AST), a simulation-based testing framework for finding the most likely path to a failure event of a safety-critical system. The key idea of AST is that stress testing can be formulated as a Partially Observable Markov Decision Process (POMDP), which enables reinforcement learning techniques to be used for finding failure events. Reinforcement learning algorithms can efficiently explore the search space and have been shown to scale to very large systems. We present applications of AST to find failures in various safety-critical systems including the aircraft collision avoidance systems, autonomous cars, and small unmanned aerial vehicles.

autonomous vehicles↗