Codesign Performance Prediction for Computational Physics 3rd Year Review Overview talk
This report describes the various models encompassed by the Performance Prediction Toolkit.
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.
This report describes the various models encompassed by the Performance Prediction Toolkit.
Selected and annotated bibliography of human performance prediction in man machine systems - Vol. 3
Computerized thermal model simulating environment control system, crew and vehicle structure in performance prediction for Apollo lunar module
The sample case presented in this volume is an asymmetrical eight sector thermal gradient performance prediction for the solid rocket motor. This motor is the TC-227A-75 grain design and the initial grain geometry is assumed to be symmetrical about the motors longitudinal axis.
The author has identified the following significant results. A new technique for image processing system performance prediction and product quality evaluation was developed. It was entirely objective, quantitative, and general, and should prove useful in system design and quality control. The technique and its application to determination of quality control procedures for the Earth Resources Technology Satellite NASA Data Processing Facility are described.
This paper presents a significant improvement in accuracy for photovoltaic (PV) module performance prediction using a new approach based on IEC-61853 test data. The method is based on the concept of a single diode model and calibrates its performance across a broad set of operating conditions. The new method reduces irradiance-weighted error in maximum power point (MPP) power relative to the standard 6-parameter single diode model, on average across 20 different module types, by 88% across the entire range of measured irradiance and temperature conditions, while retaining a continuous mathematical representation of the full current-voltage characteristic that allows for accurate off-MPP module modeling.
Celestial X ray source positions from rotating modulation collimator, predicting performance of collimator
A key step limiting how fast batteries can be deployed is the time necessary to provide evaluation and validation of performance. Using data analysis approaches, such as machine learning, the validation process can be accelerated. However, questions on the validity of projecting models trained on limited data or simple cycling profiles, such as constant current cycling, to real-world scenarios with complex loads remains. Here, we present the ability to predict performance with less than 1.2% mean absolute percent error when trained on cells aged using complex electric vehicle discharge profiles, and either AC Level 2 charge or DC Fast charge profiles, using only the first 45 cycles, namely 5% of the total testing time. While error is low across the projections, this study also highlights that battery lifetime analysis using only cycling data may not extrapolate safely to certain real-world conditions due to the impact of calendar degradation.
Abstract Large-scale high performance computing (HPC) systems typically consist of many thousands of CPUs and storage units used by hundreds to thousands of users simultaneously. Applications from large numbers of users have diverse characteristics, such as varying computation, communication, memory, and I/O intensity. A good understanding of the performance characteristics of each user application is important for job scheduling and resource provisioning. Among these performance characteristics, I/O performance is becoming increasingly important as data sizes rapidly increase and large-scale applications, such as simulation and model training, are widely adopted. However, predicting I/O performance is difficult because I/O systems are shared among all users and involve many layers of software and hardware stack, including the application, network interconnect, operating system, file system, and storage devices. Furthermore, updates to these layers and changes in system management policy can significantly alter the I/O behavior of applications and the entire system. To improve the prediction of the I/O performance on HPC systems, we propose integrating information from several different system logs and developing a regression-based approach to predict the I/O performance. Our proposed scheme can dynamically select the most relevant features from the log entries using various feature selection algorithms and scoring functions, and can automatically select the regression algorithm with the best accuracy for the prediction task. The evaluation results show that our proposed scheme can predict the write performance with up to 90% prediction accuracy and the read performance with up to 99% prediction accuracy using the real logs from the Cori supercomputer system at NERSC.
This report details the accomplishments from the ASCR funded project “Integrated End-to-end Performance Prediction and Diagnosis for Extreme Scientific Workflows” under the award numbers FWP-66406 and DE-SC0012630, with a focus on the UC San Diego (Award No. DE-SC0012630) part of the accomplishments. We refer to the project as IPPD. The main activities of IPPD were centered on the development and integration of provenance information to capture empirically workflow information and identify the sources of bottlenecks as well as variability, and modeling and simulation to provide insights and predictively explore multiple scenarios for development of advanced techniques for optimization of resources and workflow execution. After a Phase I of the project, a Phase II effort focused on three major aspects: a) observe how data is generated, distributed, and used; b) analyze how data is (repeatedly) consumed with a focus both on repeated patterns and anomalies; and c) explore how to optimize data motion. The project leveraged and extended our existing tools with new research and demonstrated our work on the Belle II workflow suite as well as on workflows from NSLS-II.
A computer program predicts the performances of solid-state lasers that operate at wavelengths from ultraviolet through mid-infrared and that comprise various combinations of stable and unstable resonators, optical parametric oscillators (OPOs), and sum-frequency generators (SFGs), including second-harmonic generators (SHGs). The input to the program describes the signal, idler, and pump beams; the SFG and OPO crystals; and the laser geometry. The program calculates the electric fields of the idler, pump, and output beams at three locations (inside the laser resonator, just outside the input mirror, and just outside the output mirror) as functions of time for the duration of the pump beam. For each beam, the electric field is used to calculate the fluence at the output mirror, plus summary parameters that include the centroid location, the radius of curvature of the wavefront leaving through the output mirror, the location and size of the beam waist, and a quantity known, variously, as a propagation constant or beam-quality factor. The program provides a typical Windows interface for entering data and selecting files. The program can include as many as six plot windows, each containing four graphs.
The accuracy of various methods used to predict tilt rotor hover performance was established by comparing predictions with large-scale experimental data. A wide range of analytical approaches were examined. Blade lift was predicted with a lifting line analysis, two lifting surface analyses, and by a finite-difference solution of the full potential equation. Blade profile drag was predicted with two different types of airfoil tables and an integral boundary layer analysis. The inflow at the rotor was predicted using momentum theory, two types of prescribed wakes, and two free wake analyses. All of the analyses were accurate at moderate thrust coefficients. The accuracy of the analyses at high thrust coefficients was dependent upon their treatment of high sectional angles of attack on the inboard sections of the rotor blade. The analyses which allowed sectional lift coefficients on the inboard stations of the blade to exceed the maximum observed in two-dimensional wind tunnel tests provided better accuracy at high thrust coefficients than those which limited lift to the maximum two-dimensional value. These results provide tilt rotor aircraft designers guidance on which analytical approaches provide the best results, and the level of accuracy which can be expected from the best analyses.
The Geoscience Laser System (GLRS) will be a high-precision distance-measuring instrument planned for deployment on the EOS-B platform. Its primary objectives are to perform ranging measurements to ground targets to monitor crustal deformation and tectonic plate motions, and nadir-looking altimetry to determine ice sheet thicknesses, surface topography, and vertical profiles of clouds and aerosols. The system uses a mode-locked, 3-color Nd:YAG laser source, a Microchannel Plate-PMT for absolute time-of-flight (TOF) measurement (at 532 nm), a streak camera for TOF 2-color dispersion measurement (532 nm and 355 nm), and a Si avalanche photodiode for altimeter waveform detection (1064 nm). The performance goals are to make ranging measurements to ground targets with about 1 cm accuracy, and altimetry height measurements over ice with 10 cm accuracy. This paper presents an overview of the design concept developed during a phase B study. System engineering issues and trade studies are discussed, with particular attention to error budgets and performance predictions.
Focused ion beams are indispensable tools in the semiconductor industry because of their ability to image and modify structures at the nanometer length scale. Here, we report on performance predictions of a new type of focused ion beam based on photo-ionization of a laser cooled and compressed atomic beam. Particle tracing simulations are performed to investigate the effects of disorder-induced heating after ionization in a large electric field. They lead to a constraint on this electric field strength which is used as input for an analytical model which predicts the minimum attainable spot size as a function of, amongst others, the flux density of the atomic beam, the temperature of this beam, and the total current. At low currents (I < 10 pA), the spot size will be limited by a combination of spherical aberration and brightness, while at higher currents, this is a combination of chromatic aberration and brightness. It is expected that a nanometer size spot is possible at a current of 1 pA. The analytical model was verified with particle tracing simulations of a complete focused ion beam setup. A genetic algorithm was used to find the optimum acceleration electric field as a function of the current. At low currents, the result agrees well with the analytical model, while at higher currents, the spot sizes found are even lower due to effects that are not taken into account in the analytical model.
Two continuous operation 18 K/20 K sorption coolers are being developed by JPL for the ESA Planck mission to provide about 200 mW of cooling at 18 K and 1.4 W at 20 K. A detailed performance prediction model has been developed to support the design process and to evaluate the results of a prototype testing.
The lithium vapor cave is a detached divertor design that uses a single private flux region baffle to contain a dense cloud of lithium vapor to dissipate heat flux. Plasma flows are created via fuel gas puffing in order to minimize lithium contamination of the main plasma. Significant modeling using the 2D edge code SOLPS-ITER has already been performed, predicting that sufficient target heat flux reductions (q$^{max}_{Target}$ MW m −2 ) with ( $n$ Li /$n$ e ) LCFS < 0.05 is possible in a case with 90 MW m −2 unmitigated heat flux in NSTX-U. Low heat flux and low upstream concentration was found with a variety of combinations of divertor geometries, target recycling coefficients, upstream plasma parameters, lithium evaporation locations and deuterium fueling locations, with variations in performance found for each design choice. However, the most universal uncertainty of SOLPS-ITER simulations has until now remained unaddressed systematically, namely the cross-field anomalous particle and heat diffusivities. This article aims to bound the uncertainty in lithium concentration prediction as a result of the assumed deuterium cross-field transport and recycling. For the simulations presented here, a factor of 2.1 increase in the upstream lithium density prediction across a factor of four decrease to the assumed deuterium radial particle diffusivity is found. This result is compared across different assumed deuterium recycling coefficients, known to reduce with lithium injection. Upstream lithium density is found to vary by a factor of 2.4 across a feasible range of recycling coefficients at similar $n$$^{OMP,sep}_{e}$.
The new generation of satellite-borne earth resources scanners, the Thematic Mapper, is being built for launch on the Landsat-D spacecraft. It will gather data for applications such as crop inventory, land use planning, forest management, and geology. This paper gives an overall design description, further discussion of principal design features, performance achievements where data are available, and system performance predictions.
Large data analysis problems often involve a large number of variables, and the corresponding analysis algorithms may examine all variable combinations to find the optimal solution. For example, to model the time required to complete a scientific workflow, we need to consider the impact of dozens of parameters. To reduce the model building time and reduce the likelihood of overfitting, we look to variable selection methods to identify the critical variables for the performance model. In this work, we create a combination of variable selection and performance prediction methods that is as effective as the exhaustive search approach when the exhaustive search could be completed in a reasonable amount of time. To handle the cases where the exhaustive search is too time consuming, we develop the parallelized variable selection algorithm. Additionally, we develop a parallel grouping mechanism that further reduces the variable selection time by 70%.As a case study, we exercise the variable selection technique with the performance measurement data from the Palomar Transient Factory (PTF) workflow. The application scientists have determined that about 50 variables and parameters are important to the performance of the workflows. Our tests show that the Sequential Backward Selection algorithm is able to approximate the optimal subset relatively quickly. By reducing the number of variables used to build the model from 50 to 4, we are able to maintain the prediction quality while reducing the model building time by a factor of 6. Using the parallelization and grouping techniques we developed in this work, the variable selection process was reduced from over 18 hours to 15 minutes while ending up with the same variable subset.