Least-square methods in abstract pattern recognition.
Least square or least mean square approximation methods in abstract pattern recognition, noting unknown probability density function
SEARCH · Engineering Papers
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.
Least square or least mean square approximation methods in abstract pattern recognition, noting unknown probability density function
The physical and geometric interpretation of a surface fitting technique, multiquadric equations (MQ), is discussed. The relation of this method and its reciprocal form to biharmonic and harmonic techniques respectively, is pointed out. In both forms the solutions may be viewed as being numerical approximations of an integral equation in which an unknown density function is the physical source for disturbing potential (reciprocal-MQ) or elastic displacement (MQ).
The gravitational collapse of isothermal, nonrotating magnetic gas clouds have been calculated numerically, including the effects of ambipolar diffusion. The fractional ionization in the clouds is approximated by a power-law function of the gas density, f = K/n to the q-power, where K and q are adjustable parameters. Eleven numerical experiments were run, and the results indicate that the asymptotic character of collapse is determined mainly by the value of q and is largely independent of the other parameters characterizing a cloud (e.g., K, cloud mass). In particular, there is nearly a one-to-one correspondence between q and the slope, x, of the central magnetic field strength-gas density relationship. If q is no more than 0.8, a cloud collapses asymptotically, as though the magnetic field were 'frozen' to the neutral matter. The magnetic field strength at the center of a collapsing cloud is strongly amplified during collapse even for values of q of about 1, despite extremely low values of fractional ionization. A discussion of the theoretical basis for this unexpected behavior is given. Possible implications of our results for the problems of magnetic braking of rotating protostars and star formation in general are also presented.
A method is presented for the construction of velocity and surface density profiles of infinitely thin, self-gravitating disks formed from the collapse of spherical clouds, under the condition that the angular momentum of each material parcel is conserved. Results are presented for initially uniformly rotating clouds with density distributions varying as R exp -n and n = 0, 1, 2. The resulting disk surface density distributions are well represented by power laws in the cylindrical radius over most of their extent. The form of the zeroth-order approximation provides a useful representation of the final surface density function near the center for all three cases, and for n = 2 this result holds throughout the disk. This result is not generally true for the velocity functions, however.
Brief progress reports of the Center for Modeling of Turbulence and Transition (CMOTT) research staff from May 1990 to May 1991 are given. The objectives of the CMOTT are to develop, validate, and implement the models for turbulence and boundary layer transition in the practical engineering flows. The flows of interest are three dimensional, incompressible, and compressible flows with chemistry. The schemes being studied include the two-equation and algebraic Reynolds stress models, the full Reynolds stress (or second moment closure) models, the probability density function models, the Renormalization Group Theory (RNG) and Interaction Approximation (DIA), the Large Eddy Simulation (LES) and Direct Numerical Simulation (DNS).
This research brief contains the progress reports of the research staff of the Center for Modeling of Turbulence and Transition (CMOTT) from June 1992 to July 1993. It is also an annual report to the Institute for Computational Mechanics in Propulsion located at Ohio Aerospace Institute and NASA Lewis Research Center. The main objectives of the research activities at CMOTT are to develop, validate, and implement turbulence and transition models for flows of interest in propulsion systems. Currently, our research covers eddy viscosity one- and two-equation models, Reynolds-stress algebraic equation models, Reynolds-stress transport equation models, nonequilibrium multiple-scale models, bypass transition models, joint scalar probability density function models, and Renormalization Group Theory and Direct Interaction Approximation methods. Some numerical simulations (LES and DNS) have also been carried out to support the development of turbulence modeling. Last year was CMOTT's third year in operation. During this period, in addition to the above mentioned research, CMOTT has also hosted the following programs: an eighteen-hour short course on 'Turbulence--Fundamentals and Computational Modeling (Part I)' given by CMOTT at the NASA Lewis Research Center; a productive summer visitor research program that has generated many encouraging results; collaborative programs with industry customers to help improve their turbulent flow calculations for propulsion system designs; a biweekly CMOTT seminar series with speakers from within and without the NASA Lewis Research Center including foreign speakers. In addition, CMOTT members have been actively involved in the national and international turbulence research activities. The current CMOTT roster and organization are listed in Appendix A. Listed in Appendix B are the abstracts of the biweekly CMOTT seminar. Appendix C lists the papers contributed by CMOTT members.
Optical Metrology tools, especially for short wavelength (EUV and X-Ray), must cover a wide range of spatial frequencies from the very low, which affects figure, to the important mid-spatial frequencies and the high spatial frequency range, which produces undesirable flair. A major difficulty in using surface profilometers arises due to the unknown Modulation Transfer Function (MTF) of the instruments. Therefore, accurate calibration of profilometers, the understanding of their MTF limitations, and cross calibration between tools represents a considerable challenge for quantitative optical metrology. In previous work the instrumental MTF of a surface profiler was precisely measured using reference test surfaces based on binary pseudo-random (BPR) gratings. Here, they present results of fabricating and using two-dimensional (2D) BPR arrays that allow for a direct 2D calibration of the instrumental MTF. BPR sequences are widely used in engineering and communication applications such as Global Position System, and wireless communication protocol. The ideal BPR pattern has a flat 'white noise' response over the entire range of spatial frequencies of interest. The BPR array used here is based on the Uniformly Redundant Array prescription initially used for x-ray and gamma ray astronomy applications. The URA's superior imaging capability originates from the fact that its cyclical autocorrelation function very closely approximates a delta function, which produces a flat Power Spectrum Density (PSD). Three different BPR array patterns were fabricated by electron beam lithography and ICP etching of silicon. The basic size unit was 200 nm, 400 nm, and 600 nm. Figure 1 shows the fabrication sequence. The 2D BPR arrays were used as standard test surfaces for MTF calibration of the MicroMap{trademark}-570 interferometric microscope with all available objectives. Figure 2 shows representative scanning probe height data for the 400 nm BPR sample. Figure 2 shows the raw Power Spectral Density for 5 different objectives. They demonstrate that the two dimensional BPR array is a very effective calibration standard. However, departures from ideal, such as square sidewall, and uniform etch depth ultimately can limit the accuracy of the calibration. The effects of fabrication imperfections on the efficiency of calibration will be discussed.
Very accurate non-Born-Oppenheimer quantum-mechanical calculations are performed to determine the average values of the interparticle distances and the proton-deuteron density function for the rovibrationally excited HD{sup +} ion. The states corresponding to excitations to all bound vibrational states (v = 0, …, 22) and simultaneously excited to the first excited rotational state (N = 1) are considered. To describe each state up to 8000 explicitly correlated all-particle Gaussian functions are used. The nonlinear parameters of the Gaussians are variationally optimized using a procedure that employs the analytical energy gradient determined with respect to these parameters. The results show an increasing asymmetry in the electron distribution with the vibrational excitation as the electron density shifts towards deuteron and away from the proton.
We study the scale and redshift dependence of the power spectra for density perturbations and peculiar velocities, and the evolution of a coarse grained phase space density for warm dark matter (WDM) particles that decoupled during the radiation dominated stage. The WDM corrections are obtained in a perturbative expansion valid in the range of redshifts at which N-body simulations set up initial conditions, and for a wide range of scales. The redshift dependence is determined by the kurtosis {beta}{sub 2} of the distribution function at decoupling. At large redshift there is an enhancement of peculiar velocities for {beta}{sub 2}>1 that contributes to free streaming and leads to further suppression of the matter power spectrum and an enhancement of the peculiar velocity autocorrelation function at scales smaller than the free streaming scale. Statistical fluctuations of peculiar velocities are also suppressed on these scales by the same effect. In the linearized approximation, the coarse grained phase space density features redshift dependent WDM corrections from gravitational perturbations determined by the power spectrum of density perturbations and {beta}{sub 2}. For {beta}{sub 2}>25/21 it grows logarithmically with the scale factor as a consequence of the suppression of statistical fluctuations. Two specific models for WDM are studied in detail. The WDM corrections relax the bounds on the mass of the WDM particle candidate.
With a single f-electron, Ce(III) is the simplest test case for benchmarking the thermodynamic and structural properties of hydrated Ln(III) against varying density functionals and reaction field models, in addition to determining the importance of multiconfigurational character in their wave functions. Here, the electronic structure of Ce(H₂O)x(H₂O)y 3+ (x =8, 9; y= 0, 12-14) has been examined using DFT and CASSCF calculations. The latter confirmed that the wave function of octa- and nona-aqua Ce(III) is well-described by a single configuration. Benchmarking was performed for density functionals, reaction field cavity types, and solvation reactions against the experimental free energy of hydration, ΔGhyd(Ce3+). The UA0, UAKS, Pauling, and UFF polarized continuum model cavities displayed different performance, depending on whether one or two hydration shells were examined, and as a function of the size of the metal basis set. These results were essentially independent of the density functional employed. Using these benchmarks, the free energy for water exchange between CN = 8 and CN = 9, for which no experimental data are available, was estimated to be approximately -4 kcal/mol.
Performing high accuracy hybrid functional calculations for condensed matter systems containing a large number of atoms is at present computationally very demanding or even out of reach if high quality basis sets are used. We present a highly optimized multiple graphics processing unit implementation of the exact exchange operator which allows one to perform fast hybrid functional density-functional theory (DFT) calculations with systematic basis sets without additional approximations for up to a thousand atoms. With this method hybrid DFT calculations of high quality become accessible on state-of-the-art supercomputers within a time-to-solution that is of the same order of magnitude as traditional semilocal-GGA functionals. The method is implemented in a portable open-source library.
Electron density measurements have been made in steady-state plasmas in a spherical inertial electrostatic confinement (IEC) discharge using microwave interferometry. Plasma cores interior to two cathodes, having diameters of 15 and 23 cm, respectively, were probed over a transverse range of 10 cm with a spatial resolution of about 1.4 cm for buffer gas pressures from 0.2 to 6 Pa in argon and deuterium. The transverse profiles are generally flat, in some cases with eccentric symmetric minima, and give mean densities of from approx. = 0.4 to 7x 10(exp 10)/cu cm, the density generally increasing with the neutral gas pressure. Numerical solutions of the 1-D Poisson equation for EC plasmas are reviewed and energy distribution functions are identified which give flat transverse profiles. These functions are used with the plasma approximation to obtain solutions which also give densities consistent with the measurements, and a double potential well solution is obtained which has minima qualitatively similar to those observed. Explicit consideration is given to the compatibility of the solutions interior and exterior to the cathode, and to grid transparency. Deuterium fusion neutron emission rates were also measured and found to be isotropic, to within the measurement error, over two simultaneous directions. Anisotropy was observed in residual emissions during operation with non-fusing hydrogen-1. The deuterium rates are consistent with predictions from the model.
We present a calculation of the properties of vibrational states in deformed, axially-symmetric even-even nuclei, within the framework of a fully self-consistent quasiparticle random phase approximation (QRPA). The same Skyrme energy density and density-dependent pairing functionals are used to calculate the mean field and the residual interaction in the particle-hole and particle-particle channels. We have tested our software in the case of spherical nuclei against fully self-consistent calculations published in the literature, finding excellent agreement. We investigate the consequences of neglecting the spin-orbit and Coulomb residual interactions in QRPA. Furthermore we discuss the improvement obtained in the QRPA result associated with the removal of spurious modes. Isoscalar and isovector responses in the deformed {sup 24-26}Mg, {sup 34}Mg isotopes are presented and compared to experimental findings.
The thermal neutron scattering law (TSL) for crystalline beryllium hydride (BeH 2 ) is developed from first-principles ab initio lattice dynamics calculations and the impact of neutron thermalization in this material on critical mass is estimated. BeH 2 has a body-centered orthorhombic crystal structure with 12 molecules per unit cell and a theoretical density of 0.755 g/cm 3 . The vibrational (phonon) densities of states for H and Be bound in BeH 2 are determined using VASP density functional theory and PHONON lattice dynamics calculations. The TSLs for H bound in BeH 2 , H(BeH 2 ), and Be bound in BeH 2 , Be(BeH 2 ), are then evaluated in the incoherent approximation from the calculated H and Be partial phonon density of states using FLASSH. Finally, critical mass as a function of 235 U loading density for bare and reflected BeH 2 moderated spheres is predicted from MC21 Monte Carlo neutron transport calculations using ENDF/B-VIII.0 cross sections and the H(BeH 2 ) and Be(BeH 2 ) TSL evaluations. Comparisons are made to water (H 2 O), polyethylene (CH 2 ), and beryllium oxide (BeO) as moderators. These critical mass predictions are a refinement upon the prior work by Rao and Srinivasan that neglected thermal neutron scattering effects. The minimum critical mass of a BeH 2 moderated assembly is estimated to be 0.207 kg 235 U for a 0.20 m thick BeO reflected sphere and 0.178 kg 235 U for a 0.40 m thick BeO reflected sphere.
We study the nonuniform potential effect in a quasi-two-dimensional (2D) Bose gas system at a finite temperature by numerically solving the projected Gross-Pitaevskii equation. The gradual emergence of superfluidity is confirmed by calculating the moment of inertia as the temperature decreases. Reflecting the nonuniformity of the system, the long-distance decay of a phase-correlation function is characterized by a local effective exponent, {eta}(r), which depends on the local superfluid density. This 'local-density approximation' is endorsed by the 2D Gaussian field theory with the nonuniform coupling constant. Indeed, the effective exponent turns out to be close to 1/4 at the boundary of the phase-coherent region.
Within both dynamical and thermodynamical approaches using the equation of state for neutron-rich nuclear matter constrained by the recent isospin diffusion data from heavy-ion reactions in the same subsaturation density range as the neutron star crust, the density and pressure at the inner edge separating the liquid core from the solid crust of neutron stars are determined to be 0.040 fm{sup -3}{<=}{rho}{sub t}{<=}0.065 fm{sup -3} and 0.01 MeV/fm{sup 3}{<=}P{sub t}{<=}0.26 MeV/fm{sup 3}, respectively. These together with the observed minimum crustal fraction of the total moment of inertia allow us to set a new limit for the radius of the Vela pulsar significantly different from the previous estimate. It is further shown that the widely used parabolic approximation to the equation of state of asymmetric nuclear matter leads systematically to significantly higher core-crust transition densities and pressures, especially with stiffer symmetry energy functionals.
In the framework of time-dependent density functional theory (TDDFT), the exact exchange-correlation (xc) kernel f xc (n, q, ω) determines the ground-state energy, excited-state energies, lifetimes, and the time-dependent linear density response of any many-electron system. The recently developed MCP07 xc kernel f xc (n, q, ω) of A. Ruzsinszky et al. [Phys. Rev. B 101, 245135 (2020)] yields excellent uniform electron gas (UEG) ground-state energies and plausible plasmon lifetimes. As MCP07 is constructed to describe f xc of the UEG, it cannot capture optical properties of real materials. To verify this claim, we follow Nazarov et al. [Phys. Rev. Lett. 102, 113001 (2009)] to construct the long-range, dynamic xc kernel, lim q→0 f xc (n, q, ω) = -α(ω)e 2 /q 2 , of a weakly inhomogeneous electron gas, using MCP07 and other common xc kernels. The strong wavevector and frequency dependence of the “ultranonlocality” coefficient α(ω) is demonstrated for a variety of simple metals and semiconductors. We examine how imposing exact constraints on an approximate kernel shapes α(ω). Comparisons to kernels derived from correlated-wavefunction calculations are drawn.
Electron density measurements have been made in steady-state plasmas in a spherical inertial electrostatic confinement (IEC) discharge using microwave interferometry. Plasma cores interior to two cathodes, having diameters of 15 and 23 cm, respectively, were probed over a transverse range of 10 cm with a spatial resolution of about 1.4 cm for buffer gas pressures from 0.2 to 6 Pa in argon and deuterium. The transverse profiles are generally flat, in some cases with eccentric symmetric minima, and give mean densities of from approx. = 0.4 to 7 x 10(exp 10)/cu cm, the density generally increasing with the neutral gas pressure. Numerical solutions of the one-dimensional Poisson equation for IEC plasmas are reviewed and energy distribution functions are identified which give flat transverse profiles. These functions are used with the plasma approximation to obtain solutions which also give densities consistent with the measurements, and a double potential well solution is obtained which has minima qualitatively similar to those observed. Explicit consideration is given to the compatibility of the solutions interior and exterior to the cathode, and to grid transparency. Deuterium fusion neutron emission rates were also measured and found to be isotropic, to within the measurement error, over two simultaneous directions. Anisotropy was observed in residual emissions during operation with nonfusing hydrogen- 1. The deuterium rates are consistent with predictions from the model.