Search NASA⌕ Search

SEARCH · Search NASA

Results for “parameter optimization”

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

Homotopy approach to optimal, linear quadratic, fixed architecture compensation

Optimal linear quadratic Gaussian compensators with constrained architecture are a sensible way to generate good multivariable feedback systems meeting strict implementation requirements. The optimality conditions obtained from the constrained linear quadratic Gaussian are a set of highly coupled matrix equations that cannot be solved algebraically except when the compensator is centralized and full order. An alternative to the use of general parameter optimization methods for solving the problem is to use homotopy. The benefit of the method is that it uses the solution to a simplified problem as a starting point and the final solution is then obtained by solving a simple differential equation. This paper investigates the convergence properties and the limitation of such an approach and sheds some light on the nature and the number of solutions of the constrained linear quadratic Gaussian problem. It also demonstrates the usefulness of homotopy on an example of an optimal decentralized compensator.

Mercadal, Mathieu↗

Selection and generation of waveforms for differential mobility spectrometry

Devices based on differential mobility spectrometry (DMS) are used in a number of ways, including applications as ion prefilters for API-MS systems, as detectors or selectors in hybrid instruments (GC-DMS, DMS-IMS), and in standalone systems for chemical detection and identification. DMS ion separation is based on the relative difference between high field and low field ion mobility known as the alpha dependence, and requires the application of an intense asymmetric electric field known as the DMS separation field, typically in the megahertz frequency range. DMS performance depends on the waveform and on the magnitude of this separation field. In this paper, we analyze the relationship between separation waveform and DMS resolution and consider feasible separation field generators. We examine ideal and practical DMS separation field waveforms and discuss separation field generator circuit types and their implementations. To facilitate optimization of the generator designs, we present a set of relations that connect ion alpha dependence to DMS separation fields. Using these relationships we evaluate the DMS separation power of common generator types as a function of their waveform parameters. Optimal waveforms for the major types of DMS separation generators are determined for ions with various alpha dependences. These calculations are validated by comparison with experimental data.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

GPU code optimization using abstract kernel emulation and sensitivity analysis

In this paper, we develop a novel approach to GPU kernel optimization by focusing on identification of bottleneck resources and determining optimization parameters that can alleviate the bottleneck. Performance modeling for GPUs is done by abstract kernel emulation along with latency/gap modeling of resources. Sensitivity analysis with respect to resource latency/gap parameters is used to predict the bottleneck resource for a given kernel's execution. Here, the utility of the bottleneck analysis is demonstrated in two contexts: 1) Enhancing the OpenTuner auto-tuner with the new bottleneck-driven optimization strategy. Experimental results on all kernels from the Rodinia suite and GPU tensor contraction kernels from the NWChem computational chemistry suite demonstrate effectiveness. 2) Manual code optimization: two case studies illustrate the use of the bottleneck analysis to iteratively improve performance of code from state-of-the-art DSL code generators.

97 MATHEMATICS AND COMPUTING↗

Time-drift Aware RF Optimization with Machine Learning Techniques

The Fermilab Linac delivers 400 MeV H- beam to the rest of the accelerator chain. Providing stable intensity, energy, and emittance is key since it directly affects downstream machines. To operate high current beam, accelerators must minimize uncontrolled particle loss; this can be accomplished by minimizing beam longitudinal emittance via RF parameter optimization. However, RF tuning is required daily since the resonance frequency of the accelerating cavities is affected by ambient temperature and humidity variations and thus drifts with time. In addition, the energy and phase space distribution of particles emerging from the ion source are subject to fluctuations. Such drift is not unique to Fermilab, but rather affects most laboratories. We are exploring machine learning (ML) algorithms for automated RF tuning for 2 objectives: optimization of Linac output energy and phase oscillation correction, with an emphasis on time-drift aware modeling that can account for conditions changing over time.

43 PARTICLE ACCELERATORS↗

A modeling framework for optimizing current density in four-terminal tandem solar cells: A case study on GaAs/Si tandem

Here we present a modeling framework to quantify current density losses and determine optimization steps for four-terminal tandem solar cells, which consist of a high-bandgap thin-film top cell and an industrial-type Si wafer bottom cell. We demonstrate the framework on the example of a 21.3% efficient stacked four-terminal GaAs/Si tandem solar cell. In a first step, we develop an optical model for stacked tandem configuration and carry out the current loss analysis for this tandem device. Photon absorption in the tandem is broken down into different channels. All parasitic absorptions that do not contribute to current generation are considered part of a loss channel. In a second step, we quantify the potential current gains from single-parameter optimizations, the redistribution of photon losses into other channels, and the synergistic gains that can be achieved by co-optimization. Finally, using the simulation model, we are able to determine the best sequence of optimizing the short-circuit current density in the GaAs/Si tandem solar cell to possibly achieve 37.8 mA/cm 2 with experimentally feasible parameters.

14 SOLAR ENERGY↗

Transient Growth Theory Prediction of Optimal Placing of Passive and Active Flow Control Devices for Separation Delay in LPT Airfoils

An analysis of the non-modal growth of perturbations in a boundary layer in the presence of a streamwise pressure gradient is presented. The analysis is based on PSE equations for an incompressible fluid. Examples with Falkner-Skan profiles indicate that a favorable pressure gradient decreases the non-modal growth while an unfavorable pressure gradient leads to an increase of the amplification. It is suggested that the transient growth mechanism be utilized to choose optimal parameters of tripping elements on a low-pressure turbine (LPT) airfoil. As an example, a boundary layer flow with a streamwise pressure gradient corresponding to the pressure distribution over a LPT airfoil is considered. It is shown that there is an optimal spacing of the tripping elements and that the transient growth effect depends on the starting point. At very low Reynolds numbers, there is a possibility to enhance the transient energy growth by means of wall cooling.

Tumin, Anatoli↗

Optimal disturbances in boundary layers subject to streamwise pressure gradient

An analysis of the optimal non-modal growth of perturbations in a boundary layer in the presence of a streamwise pressure gradient is presented. The analysis is based on PSE equations for an incompressible fluid. Examples with Falkner-Scan profiles indicate that a favorable pressure gradient decreases the non-modal growth, while an unfavorable pressure gradient leads to an increase of the amplification. It is suggested that the transient growth mechanism be utilized to choose optimal parameters of tripping elements on a low-pressure turbine (LPT) airfoil. As an example, a boundary layer flow with a streamwise pressure gradient corresponding to the pressure distribution over a LPT airfoil is considered. It is shown that there is an optimal spacing of the tripping elements and that the transient growth effect depends on the starting point.

Ashpis, David E.↗

Optimization of Binning Parameters in the Feynman Variance-to-Mean Method for Delayed Neutron Reinterrogation

The focus of this work is to determine the optimal input parameters for analyzing active interrogation data that utilizes the physics of the delayed neutron re-interrogation technique. This work compares the signature from an active neutron measurement of a highly enriched uranium (HEU) system to an active background measurement (i.e., without the HEU item in place). The item being measured represents a 21.8 kg metal sphere of HEU (93 weight percent) assembled from a subset of the Rocky Flats (RF) uranium hemishells. The detection system consisted of two NoMAD neutron multiplicity detectors, a SNAP III neutron detector, and a Thermo Fisher P211 pulsed 14 MeV neutron source. In addition, varying analyses of the NoMAD list-mode detector data are performed to determine whether it is preferential to use a particular analysis bin size for producing Feynman histograms, selection of pulses to prevent discrepancies from buildup, or window of time during the pulsed neutron interrogation. This work is still currently ongoing.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Pyridine Catalysis of Anhydride Hydrolysis within Carbodiimide‐Driven Reaction Networks

Carbodiimide-fueled reaction networks offer a versatile platform for nonequilibrium chemical systems. Typically, the carbodiimide converts a carboxylic acid to its anhydride, called “activation,” which subsequently undergoes hydrolysis, called “deactivation.” Here, we investigate pyridines with variable nucleophilicity as catalysts to control deactivation (pyridine, 4-methylpyridine, 4-methoxypyridine, and 4-dimethylaminopyridine). Reactions have been monitored by NMR spectroscopy. Although this reaction network is simple, determination of well-defined rate constants from kinetic modeling is challenging because of correlation between the parameters. This issue can be addressed by analyzing the anhydride hydrolysis independently. The rate of attack of the pyridines on the anhydride follows expected nucleophilicity trends, although this is offset by increased protonation of more-nucleophilic pyridines at typical pH's. The optimized parameters can be used to model the full carbodiimide-driven process, although the presence of the common carbodiimide EDC has unanticipated effects on the anhydride hydrolysis rate. The results offer context for controlling carbodiimide-fueled reaction networks through the choice of suitable catalysts and pH.

Anhydrides↗

Optimal Disturbances in Boundary Layers Subject to Streamwise Pressure Gradient

An analysis of the non-modal growth of perturbations in a boundary layer in the presence of a streamwise pressure gradient is presented. The analysis is based on PSE equations for an incompressible fluid. Examples with Falkner- Skan profiles indicate that a favorable pressure gradient decreases the non-modal growth while an unfavorable pressure gradient leads to an increase of the amplification. It is suggested that the transient growth mechanism be utilized to choose optimal parameters of tripping elements on a low-pressure turbine (LPT) airfoil. As an example, a boundary-layer flow with a streamwise pressure gradient corresponding to the pressure distribution over a LPT airfoil is considered. It is shown that there is an optimal spacing of the tripping elements and that the transient growth effect depends on the starting point. The amplification is found to be small at the LPT s very low Reynolds numbers, but there is a possibility to enhance the transient energy growth by means of wall cooling.

Ashpis, David E.↗

Parameterization of interatomic potential by genetic algorithms: A case study

A framework for Genetic Algorithm based methodology is developed to systematically obtain and optimize parameters for interatomic force field functions for MD simulations by fitting to a reference data base. This methodology is applied to the fitting of ThO{sub 2} (CaF{sub 2} prototype) – a representative of ceramic based potential fuel for nuclear applications. The resulting GA optimized parameterization of ThO{sub 2} is able to capture basic structural, mechanical, thermo-physical properties and also describes defect structures within the permissible range.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

AladynPi – Adaptive Neural Network Molecular Dynamics Simulation Code with Physically Informed Potential: Computational Materials Mini-Application

This report provides an overview and description of commands used in the Computational Materials mini-application, AladynPi. AladynPi is an extension of a previously released mini-application, Aladyn (https://github.com/nasa/aladyn; Yamakov, V.I., and Glaessgen, E.H., NASA/TM-2018-220104). Aladyn and AladynPi are basic molecular dynamics codes written in FORTRAN 2003, which are designed to demonstrate the use of adaptive neural networks (ANNs) in atomistic simulations. The role of ANNs is to efficiently reproduce the very complex energy landscape resulting from the atomic interactions in materials with the accuracy of the more expensive quantum mechanics-based calculations. The ANN is trained on a large set of atomic structures calculated using the density functional theory method. An input for the ANN is a set of structure coefficients, characterizing the local atomic environment of each atom, for which the atomic energy is obtained in the ANN inference process. In Aladyn, the ANN gives directly the energy of interatomic interactions. In AladynPi, the ANN gives optimized parameters for a predefined empirical function, known as bond-order-potential (BOP). The parameterized BOP function is then used to calculate the energy. AladynPi code is being released to serve as a training testbed for students and professors in academia to explore possible optimization algorithms for parallel computing on multicore central processing unit (CPU) computers or computers utilizing manycore architectures based on graphic processing units (GPUs). The effort is supported by the High Performance Computing incubator (HPCi) project at NASA Langley Research Center.

Yamakov, Vesselin I.↗

Reliable design of H-2 optimal reduced-order controllers via a homotopy algorithm

Due to control processor limitations, the design of reduced-order controllers is an active area of research. Suboptimal methods based on truncating the order of the corresponding linear-quadratic-Gaussian (LQG) compensator tend to fail if the requested controller dimension is sufficiently small and/or the requested controller authority is sufficiently high. Also, traditional parameter optimization approaches have only local convergence properties. This paper discusses a homotopy algorithm for optimal reduced-order control that has global convergence properties. The exposition is for discrete-time systems. The algorithm has been implemented in MATLAB and is applied to a benchmark problem.

Collins, Emmanuel G.↗

Passive damping augmentation of flexible beam-like lattice trusses for large space structures

A Timoshenko beam continuum model is developed to determine the transient response of a beam-like latticed large space structure, subjected to a unit impulse. It is demonstrated that an increase in diagonal stiffness, on account of the stiffness of the vertical girder, leads to a rise in the transverse shear rigidity. This results in higher natural frequencies and reduction in peak displacement. In addition, in an asymmetrical truss configuration, coupling between the extensional and shear modes raises the maximum peak displacement compared with that obtained for a symmetrical truss. The model is modified to investigate the introduction of passive damping in the form of several dynamic vibration absorbers. A unique absorber parameter optimization procedure, based on the classical steady-state criteria, is used to tune the absorbers, having a total allocated mass budget of 10%, to the first two frequencies of the latticed structure. It is found that inclusion of transverse shear rigidity as a design parameter in damping augmentation studies reduces settling time for predetermined maximum peak displacements.

Cohen, K.↗

Data compression and covariance matrix inspection: Cosmic shear

Covariance matrices are among the most difficult pieces of end-to-end cosmological analyses. In principle, for two-point functions, each component involves a four-point function, and the resulting covariance often has hundreds of thousands of elements. We investigate various compression mechanisms capable of vastly reducing the size of the covariance matrix in the context of cosmic shear statistics. This helps identify which of its parts are most crucial to parameter estimation. We start with simple compression methods, by isolating and “removing” 200 modes associated with the lowest eigenvalues, then those with the lowest signal-to-noise ratio, before moving on to more sophisticated schemes like compression at the tomographic level and, finally, with the massively optimized parameter estimation and data compression (MOPED). We find that, while most of these approaches prove useful for a few parameters of interest, like Ω m , the simplest yield a loss of constraining power on the intrinsic alignment (IA) parameters as well as S 8 . For the case considered—cosmic shear from the first year of data from the Dark Energy Survey—only MOPED was able to replicate the original constraints in the 16-parameter space. Finally, we apply a tolerance test to the elements of the compressed covariance matrix obtained with MOPED and confirm that the IA parameter A IA is the most susceptible to inaccuracies in the covariance matrix.

79 ASTRONOMY AND ASTROPHYSICS↗

An expert system to perform on-line controller tuning

An expert system which tunes a Proportional-Integral-Derivative (PID) controller on-line for a single-input-single-output multiple-lag process with dead time is described. The expert system examines features of the previous transient responses and their corresponding sets of controller parameters. It determines a new set of controller gains to obtain a more desirable time response. This technique can be used to determine and implement a different set of PID gains for each operating regime and, once in steady state, the system can be used to find optimal parameters for load disturbance rejection. The expert system can be applied to any system of the specified form (aerospace, industrial, etc.) and can be expanded to include additional process models.

Litt, Jonathan S.↗

An expert system to perform on-line controller tuning

An expert system which tunes a proportional-integral-derivative (PID) controller online for a single-input-single-output multiple-lag process with dead time is described. The expert system examines features of the previous transient responses and their corresponding sets of controller parameters. It determines a new set of controller gains to obtain a more desirable time response. This technique can be used to determine and implement a different set of PID gains for each operating regime and, once in steady state, the system can be used to find optimal parameters for load disturbance rejection. The expert system can be applied to any system of the specified form (aerospace, industrial, etc.) and can be expanded to include additional process models.

Litt, Jonathan↗