Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “error indicators”

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 1,189 records · Page 66
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

First test of the consistency relation for the large-scale structure using the anisotropic three-point correlation function of BOSS DR12 galaxies

ABSTRACT We present, for the first time, an observational test of the consistency relation for the large-scale structure (LSS) of the Universe through a joint analysis of the anisotropic two- and three-point correlation functions (2PCF and 3PCF) of galaxies. We parameterize the breakdown of the LSS consistency relation in the squeezed limit by Es, which represents the ratio of the coefficients of the shift terms in the second-order density and velocity fluctuations. Es ≠ 1 is a sufficient condition under which the LSS consistency relation is violated. A novel aspect of this work is that we constrain Es by obtaining information about the non-linear velocity field from the quadrupole component of the 3PCF without taking the squeezed limit. Using the galaxy catalogues in the Baryon Oscillation Spectroscopic Survey (BOSS) Data Release 12, we obtain $E_{\rm s} = -0.92_{-3.26}^{+3.13}$, indicating that there is no violation of the LSS consistency relation in our analysis within the statistical errors. Our parameterization is general enough that our constraint can be applied to a wide range of theories, such as multicomponent fluids, modified gravity theories, and their associated galaxy bias effects. Our analysis opens a new observational window to test the fundamental physics using the anisotropic higher-order correlation functions of galaxy clustering.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2019

Towards Native Execution of Deep Learning on a Leadership-Class HPC System

Large parallel machines generally offer the best parallel performance with "native execution" that is achieved using codes developed with the optimized compilers, communication libraries, and runtimes offered on the machines. In this paper, we report and analyze performance results from native execution of deep learning on a leadership-class high-performance computing (HPC) system. Using our new code called DeepEx, we present a study of the parallel speed up and convergence rates of learning achieved with native parallel execution. In the trade-off between computational parallelism and synchronized convergence, we first focus on maximizing parallelism while still obtaining convergence. Scaling results are reported from execution on up to 15,000 GPUs using two scientific data sets from atom microscopy and protein folding applications, and also using the popular ImageNet data set. In terms of the traditional measure of parallel speed up, excellent scaling is observed up to 12,000 GPUs. Additionally, accounting for convergence rates of deep learning accuracy or error, a deep learning-specific metric called "learning speed up" is also tracked. The performance results indicate the need to evaluate parallel deep learning execution in terms of learning speed up, and point to additional directions for improved exploitation of high-end HPC systems.

Yoginath, Srikanth↗
NASA NTRS1965

Heats of mixing in praseodymium-neodymium solid solutions.

The heats of solution in liquid tin of pure praseodymium, pure neodymium and nine praseodymium-neodymium binary alloys have been measured using liquid metal solution calorimetry. A plot of the measured heat effect for the alloys as a function of composition is linear, with the terminal points being the measured values for the pure components. This linear dependency indicates that the heat of mixing in the binary praseodymium-neodymium system is zero within experimental error, and therefore the binary alloys obey one of the criteria for an ideal solution.

Calorimetry↗
NASA NTRS1972

An auditory display in a dual-axis tracking task.

Results of a study in which subjects were presented concurrently with the primary task of controlling a second-order plant and the secondary task of controlling a first-order plant. The plant errors for the two tasks were shown on separate visual displays. An auditory display, whose output varied in frequency and volume with the error, was used to supplement the secondary task in half of the runs. To study the effects of the auditory display, two performance measures were obtained: (1) the integral of the squared error (ISE) and (2) the describing functions of the human operator. Statistical analysis of the ISE measures indicated that when the secondary task was supplemented with an auditory display, there was a significant improvement in performance on the secondary task. The performance on the primary task improved on the average, but not significantly. The variances of the ISE values decreased for both the tasks, indicating a more consistent behavior with the auditory display. The describing function analysis showed that supplementing the secondary task with the auditory display increased the low frequency gain of the human operator for this task. The describing functions for the primary task did not show any apparent changes.

Mirchandani, P. B.↗
NASA NTRS1973

Effect of out-of-planeness of membrane quadrilateral finite elements.

Investigation of the effect of using plane quadrilateral membrane elements for modeling nonplanar structures. The effect is assessed by analyzing a simplified finite element model with the aid of the structure network analysis program. The results obtained indicate that the use of planar quadrilateral membrane elements for modeling bending problems can lead to large errors if the four points that define the quadrilateral are not in the same plane.

Haftka, R. T.↗
NASA NTRS1974

Modification of a three-dimensional supersonic nozzle analysis and comparison with experimental data

A computer program previously developed to analyze three-dimensional supersonic nozzles by the method of characteristics has been modified to study less restrictive nozzle geometries and nonuniform inlet conditions. An example indicates that a one-dimensional calculation that uses an averaged initial profile may be significantly in error. A comparison between the analysis and the data from a three-dimensional experiment shows generally good agreement between the two.

Bishop, A. R.↗
NASA NTRS1973

Delta modulation

The overshoot suppression algorithm has been more extensively studied. Computer generated test-pictures show a radical improvement due to the overshoot suppression algorithm. Considering the delta modulator link as a nonlinear digital filter, a formula that relates the minimum rise time that can be handled for given filter parameters and voltage swings has been developed. The settling time has been calculated for the case of overshoot suppression as well as when no suppression is employed. The results indicate a significant decrease in settling time when overshoot suppression is used. An algorithm for correcting channel errors has been developed. It is shown that pulse stuffing PCM words in the DM bit stream results in a significant reduction in error length.

Schilling, D. L.↗
NASA NTRS1977

Orbit determination with the tracking data relay satellite system

The possibility of employing the tracking data relay satellite system to satisfy the orbit determination demands of future applications missions is investigated. It is shown that when the relay satellites are continuously and independently tracked from ground stations it is possible, using six hour data arcs, to recover user satellite state with an average error of about 25 m radially, 260 m along track, and 20 m cross track. For this arc length, range sum data and range sum rate data are equally useful in determining orbits. For shorter arc lengths (20 min), range sum rate data is more useful than range sum data. When relay satellites are not continuously tracked, user satellite state can be recovered with an average error of about 140 m radially, 515 m along track, and 110 m cross track. These results indicate that the TDRS system can be employed to satisfy the orbit determination demands of applications missions, such as the MAGSAT and potential gradiometer missions, provided the relay satellites are continuously and independently tracked.

Argentiero, P.↗
NASA NTRS1977

Monopulse receiver

A time division multiplexing and a quadraphase combining dual channel system were analyzed, designed, and tested. Analyses performed include the following: (1) boresight error as a function of the error channel bandwidth; (2) error channel interference with the sum channel functions; (3) threshold performance; and (4) error channel crosstalk, linearity, and drift as a function of signal level, Doppler, and environment. Test results indicate that the time division multiplexing system meets the design goals of the program. However, careful selection and alignment of all gain controlled amplifiers are required along with temperature compensation of the angle channel gain. The quadraphase system crosstalk performance is comparatively poor (-15 dB) with respect to the -30 dB requirements and this consequently affects gain tracking performance. Gain tracking was found to be + or - 20 percent rather than the + or - 5 percent specification. Data for the two systems is compared and recommendations are presented.

Source record↗
NASA NTRS1978

Absolute flux density calibrations: Receiver saturation effects

The effect of receiver saturation was examined for a total power radiometer which uses an ambient load for calibration. Extension to other calibration schemes is indicated. The analysis shows that a monotonic receiver saturation characteristic could cause either positive or negative measurement errors, with polarity depending upon operating conditions. A realistic model of the receiver was made by using a linear-cubic voltage transfer characteristic. The evaluation of measurement error for this model provided a means for correcting radio source measurements.

Freiley, A. J.↗
NASA NTRS1980

Precision of NO2 photolysis rate measurements

Gas-phase kinetic calculations indicate that actinometric measurements of the rate of solar photolysis of NO2 in nitrogen can be in error by +5 to -15% or more, unless odd oxygen N2O5 chemistry, and trace oxygen content are considered. A field experiment confirms that the effect of the quartz tube on the measured photolysis rate is less than 1.7%.

Dickerson, R. R.↗
NASA NTRS1981

An analysis of LANDSAT MSS scene-to-scene registration accuracy

Measurements were made for 12 registrations done by ERL for 8 registrations done by SRS. The results indicate that the ERL method is significantly more accurate in five of the eight comparison. The difference between the two methods are not significant in the other three cases. There are two possible reasons for the differences. First, the ERL model is a piecewise linear model and the EDITOR model is a cubic polynomial model. Second, the ERL program resamples using bilinear interpolation while the EDITOR software uses a nearest neighbor resampling. This study did not indicate how much of the difference is attributable to each factor. The average of all merged scene error values for ERL was 31.6 meters and the average for the eight common areas was 32.6 meters. The average of the eight merged scene error values for SRS was 40.1 meters.

Seyfarth, B. R.↗
NASA NTRS1981

Measurement of solar radius changes

Results of daily photometric measurements of the solar radius from Mt. Wilson over the past seven years are reported. Reduction of the full disk magnetograms yields a formal error of 0.1 arcsec in the boustrophedonic scans in the 5250.2 A FeI line. 150 scan lines comprise each observation; 1,412 observations were made from 1974-1981. Measurement procedures, determination of the scattered light of the optics and the atmosphere, and error calculations are described, noting that days of poor atmospheric visibility are omitted from the data. The horizontal diameter of the sun remains visually fixed while the vertical component changes due to atmospheric diffraction; error accounting for thermal effects, telescope aberrations, and instrument calibration are discussed, and results, within instrument accuracy, indicate no change in the solar radius over the last seven years.

Labonte, B. J.↗
NASA NTRS1991

Stellar evolution and period changes in RR Lyrae stars

The observed secular period changes of RR Lyrae stars in five globular clusters have been compared with those predicted by synthetic models of the horizontal branch (HB). It is shown that most of the observed number distributions of the period change rate beta could be attributed to evolutionary effects, if the random observational error is of order + or - 0.07 days/Myr in beta as suggested by the observers. The model calculations indicate that the mean rate of period change in RR Lyrae stars in globular clusters depends sensitively on their HB type, as is the case for the period shift at a given effective temperature. The distinct bias toward positive period changes in M15 and Omega Cen is evidence that most HB stars in the Oosterhoff group II clusters pass through the instability strip from blue to red toward the end of their core helium burning phase. If confirmed by future observations, this is strong support for the Lee et al. (1990) explanation of the Sandage (1957) period-shift effect.

Lee, Young-Wook↗
NASA NTRS1992

Analysis of space environment effects on active fiber optic links orbited aboard the LDEF

The interim analysis correlates the results of the 'Preliminary Analysis of WL Experiment no. 701, Space Environment Effects on Operating Optic Systems' (NASA Report CP-3134) with space simulated post retrieval terrestrial studies performed on the M0004 experiment. Temperature cycling measurements were performed on the active optical data links for the purpose of assessing link signal to noise ratio and bit error rate performance some 69 months following the experiment deployment in low earth orbit. The early results indicate a high correlation between pre-orbit, orbit recorded, and post orbit functionality of the first known and longest space demonstration of operating optic fibers.

Monarski, T. W.↗
NASA NTRS1990

Acoustic results of the Boeing model 360 whirl tower test

An evaluation is presented for whirl tower test results of the Model 360 helicopter's advanced, high-performance four-bladed composite rotor system intended to facilitate over-200-knot flight. During these performance measurements, acoustic data were acquired by seven microphones. A comparison of whirl-tower tests with theory indicate that theoretical prediction accuracies vary with both microphone position and the inclusion of ground reflection. Prediction errors varied from 0 to 40 percent of the measured signal-to-peak amplitude.

Watts, Michael E.↗
NASA NTRS1993

Analyses of space environment effects on active fiber optic links orbited aboard the LDEF

The results of the 'Preliminary Analysis of WL Experiment no. 701, Space Environment Effects on Operating Fiber Optic Systems,' is correlated with space simulated post retrieval terrestrial studies performed on the M0004 experiment. Temperature cycling measurements were performed on the active optical data links for the purpose of assessing link signal to noise ratio and bit error rate performance some 69 months following the experiment deployment in low Earth orbit. The early results indicate a high correlation between pre-orbit, orbit, and post-orbit functionality of the first known and longest space demonstration of operating fiber optic systems.

Taylor, Edward W.↗
NASA NTRS1993

Kinematics and control of a fully parallel force-reflecting hand controller for manipulator teleoperation

Design of a parallel force-reflecting hand controller that implements a friction- and inertia canceling control loop about the entire mechanism based on wrench sensing in the mechanism handgrip is discussed. Kinematics of the controller under consideration is analyzed and results are presented using a closed-form solution for the inverse kinematics and Newton-Raphson's method for the forward kinematics. Results indicate that the force control scheme based on a handgrip force sensor provides smaller steady-state errors than the scheme without a handigrip sensor.

Bryfogle, Mark D.↗