Search NASA⌕ Search

SEARCH · Search NASA

Results for “Predictive modeling”

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 955 records · Page 53

Development of a Mechanistic Model for Predicting Sliding Vapor Bubble Growth

A new mechanistic model for predicting the growth of sliding vapor bubble in subcooled boiling flows has been developed. This model employs a physical base that the fundamental mechanisms applied to the bubble growth at a nucleation site can also be applied to the growth of sliding bubble. The heat transfer mechanisms considered are microlayer evaporation, evaporation of superheated liquid, and condensation. Modifications were made to the physical assumptions/reasoning applied to the existing growth models to reflect the intrinsic physics associated with the growth of sliding bubble. The present model has been successfully validated against a broad set of subcooled flow boiling data generated by three separate research groups using different working fluids and heater surfaces. The physical dependency of bubble growth during sliding on the relevant flow boiling parameters (i.e., mass flux, wall superheat, and subcooling level) observed in these experiments was well reproduced by the present model.

42 ENGINEERING↗

Chemical composition based machine learning model to predict defect formation in additive manufacturing

With a goal of exploiting additive manufacturing to improve the manufacturing of existing reactor materials, we developed a chemical composition-based machine learning model to predict the printability of any given alloy in laser powder bed fusion (L-PBF) using experimental data from peer-reviewed literature. We defined printability as the ability to avoid defects like cracking, balling, porosity, and lack of fusion, that are caused by thermal stresses (during solidification or liquation), molten pool disintegration into disconnected small beads or lack of heat input respectively. Our models predict the tendency of balling defect formation and porosity percentage for a given composition, under a given set of processing conditions. To predict the likelihood of balling defect, three models: a random forest classifier, a gradient boost regressor and a neural network were trained on a dataset containing both traditional alloys and high entropy alloys. The neural network model showed the highest accuracy of 92.3 % in predicting the balling defect formation. A random forest regressor, gradient boost regressor and neural network were trained and tested on a dataset of various alloys to predict porosity. The random forest regressor showed the best predictions with an R 2 score of 0.97. The models also revealed the relative importance of the input descriptors on defect-formation tendency. Of particular significance was the identification of carbon as an important element in determining the occurrence of balling and percent porosity in alloys like steel, as well as being moderately important to the percentage porosity in other alloys as well as steel. Manganese was also identified as a key descriptor for the percentage of porosity in steel and other alloys. Manganese’s low thermal conductivity and consistent presence in the dataset is the likely cause for its contribution. Carbon’s role is attributable to its relatively high specific heat and high melting temperature. In conclusion, our model serves as a swift, chemistry-based tool to design experiments and find modified compositions better suited for additive manufacturing.

36 MATERIALS SCIENCE↗

A Screening Model to Predict Microalgae Biomass Growth in Photobioreactors and Raceway Ponds

A microalgae biomass growth model was developed for screening novel strains for their potential to exhibit high biomass productivities under nutrient-replete conditions in photobioreactors or outdoor ponds. Growth is modeled by first estimating the light attenuation by biomass according to Beer-Lambert’s law, and then calculating the specific growth rate in discretized culture volume slices that receive declining light intensities due to attenuation. The model requires only two physical and two species-specific biological input parameters, all of which are relatively easy to determine: incident light intensity, culture depth, as well as the biomass light absorption coefficient and the specific growth rate as a function of light intensity. Roux bottle culture experiments were performed with Nannochloropsis salina at constant temperature (23 °C) at six different incident light intensities (5, 10, 25, 50, 100, 250, and 850 μmol/m2∙ sec) to determine both the specific growth rate under non-shading conditions and the biomass light absorption coefficient as a function of light intensity. The model was successful in predicting the biomass growth rate in these Roux bottle cultures during the light-limited linear phase at different incident light intensities. Model predictions were moderately sensitive to minor variations in the values of input parameters. The model was also successful in predicting the growth performance of Chlorella sp. cultured in LED-lighted 800 L raceway ponds operated at constant temperature (30 °C) and constant light intensity (1650 μmol/m2∙ sec). Measurements of oxygen concentrations as a function of time demonstrated that following exposure to darkness, it takes at least 5 seconds for cells to initiate dark respiration. As a result, biomass loss due to dark respiration in the aphotic zone of a culture is unlikely to occur in highly mixed small-scale photobioreactors where cells move rapidly in and out of the light. By contrast, as supported also by the growth model, biomass loss due to dark respiration occurs in the dark zones of the relatively less well mixed pond cultures. In addition to screening novel microalgae strains for high biomass productivities, the model can also be used for optimizing the pond design and operation. Additional research is needed to validate the biomass growth model for other microalgae species and for the more realistic case of fluctuating temperatures and light intensities observed in outdoor pond cultures.

biomass growth model↗

Performance and power modeling and prediction using MuMMI and 10 machine learning methods

Energy-efficient scientific applications require insight into how high performance computing system features impact the applications' power and performance. This insight can result from the development of performance and power models. Here, in this article, we use the modeling and prediction tool MuMMI (Multiple Metrics Modeling Infrastructure) and 10 machine learning methods to model and predict performance and power consumption and compare their prediction error rates. We use an algorithm-based fault-tolerant linear algebra code and a multilevel checkpointing fault-tolerant heat distribution code to conduct our modeling and prediction study on the Cray XC40 Theta and IBM BG/Q Mira at Argonne National Laboratory and the Intel Haswell cluster Shepard at Sandia National Laboratories. Our experimental results show that the prediction error rates in performance and power using MuMMI are less than 10% for most cases. By utilizing the models for runtime, node power, CPU power, and memory power, we identify the most significant performance counters for potential application optimizations, and we predict theoretical outcomes of the optimizations. Based on two collected datasets, we analyze and compare the prediction accuracy in performance and power consumption using MuMMI and 10 machine learning methods.

97 MATHEMATICS AND COMPUTING↗

Measurements of WW and WZ Production in W+jets Final States in pp̄ Collisions

We study WW and WZ production with lνqq (l=e,μ) final states using data collected by the D0 detector at the Fermilab Tevatron Collider corresponding to 4.3 fb⁻¹ of integrated luminosity from pp̄ collisions at √s=1.96 TeV. Assuming the ratio between the production cross sections σ(WW) and σ(WZ) as predicted by the standard model, we measure the total WV (V=W,Z) cross section to be σ(WV)=19.6 +3.2 -3.0 pb and reject the background-only hypothesis at a level of 7.9 standard deviations. We also use b-jet discrimination to separate the WZ component from the dominant WW component. Simultaneously fitting WW and WZ contributions, we measure σ(WW)=15.9 +3.7 -3.2 pb and σ(WZ)=3.3 +4.1 -3.3 pb, which is consistent with the standard model predictions.

Abazov, V. M.↗

Building a Wide-Area File Transfer Performance Predictor: An Empirical Study

Wide-area data transfer is central to geographically distributed scientific workflows. Faster delivery of data is important for these workflows. Predictability is equally (or even more) important. With the goal of providing a reasonably accurate estimate of data transfer time to improve resource allocation & scheduling for workflows and enable end-to-end data transfer optimization, we apply machine learning methods to develop predictive models for data transfer times over a variety of wide area networks. To build and evaluate these models, we use 201,388 transfers, involving 759 million files totaling 9 PB transferred, over 115 heavily used source-destination pairs (“edges”) between 135 unique endpoints. We evaluate the models for different retraining frequencies and different window size of history data. In the best case, the resulting models have a median prediction error of ≤21% for 50% of the edges, and ≤32% for 75% of the edges. We present a detailed analysis of these results that provides insights into the cause of some of the high errors. We envision that the performance predictor will be informative for scheduling geo-distributed workflows. The insights also suggest obvious directions for both further analysis and transfer service optimization.

Liu, Zhengchun↗

Strategies for Energy-Efficient Resource Management of Hybrid Programming Models

Many scientific applications are programmed using hybrid programming models that use both message passing and shared memory, due to the increasing prevalence of large-scale systems with multicore, multisocket nodes. Previous work has shown that energy efficiency can be improved using software-controlled execution schemes that consider both the programming model and the power-aware execution capabilities of the system. However, such approaches have focused on identifying optimal resource utilization for one programming model, either shared memory or message passing, in isolation. The potential solution space, thus the challenge, increases substantially when optimizing hybrid models since the possible resource configurations increase exponentially. Nonetheless, with the accelerating adoption of hybrid programming models, we increasingly need improved energy efficiency in hybrid parallel applications on large-scale systems. In this work, we present new software-controlled execution schemes that consider the effects of dynamic concurrency throttling (DCT) and dynamic voltage and frequency scaling (DVFS) in the context of hybrid programming models. Specifically, we present predictive models and novel algorithms based on statistical analysis that anticipate application power and time requirements under different concurrency and frequency configurations. Furthermore, we apply our models and methods to the NPB MZ benchmarks and selected applications from the ASC Sequoia codes. Overall, we achieve substantial energy savings (8.74 percent on average and up to 13.8 percent) with some performance gain (up to 7.5 percent) or negligible performance loss.

97 MATHEMATICS AND COMPUTING↗

Co-Evolving Climate Models under Uncertainty to Improve Predictive Skill

Focal Area: Predictive modeling through the use of AI techniques and AI-derived model components; the use of AI and other tools to design a prediction system comprising of a hierarchy of models Science Challenge: Advances in modeling of climate and improved observational capabilities have led to great improvements in understanding large-scale historical climate effects. It however remains a challenge to reliably predict the risk of fine-scale regional climate events on human systems even as such risks are on the rise because warmer and wetter climates are more prone to hydrometeorological extremes,

54 ENVIRONMENTAL SCIENCES↗

Status Report on Improved Understanding of Creep‐fatigue Damage in Advanced Materials

This report provides an update on the materials performance criteria and methodology relevant to liquid metal reactors (LMRs), in particular, sodium cooled fast reactors. The report is the first deliverable (Level 3) in FY11 (M3A11AN04030303) under the work package A-11AN040303 “Materials Performance Criteria and Methodology” as part of Advanced Structural Materials Program for the Advanced Reactor Concepts. The overall objective of the Advanced Materials Performance Criteria and Methodology project is to evaluate the key requirements for the ASME Code qualification and the Nuclear Regulatory Commission (NRC) approval of advanced structural materials in support of the design and licensing of the liquid metal fast reactors. Advanced materials are a critical element in the development of fast reactor technologies. Enhanced materials performance not only improves safety margins and provides design flexibility but also is essential for the economics of future advanced fast reactors. Qualification and licensing of advanced materials are prominent needs for the development and implementation of advanced fast reactor technologies. Nuclear structural component designs in the U.S. comply with the ASME Boiler and Pressure Vessel (B&PV) Code Section III (Rules for Construction of Nuclear Facility Components) and the NRC grants licensing. As the LMR will operate at higher temperatures than the current light water reactors (LWRs), the design of elevated-temperature components must comply with ASME Section III Subsection NH (Class 1 Components in Elevated Temperature Service). A number of technical issues relevant to materials performance criteria and high temperature design methodology in the LMR were identified and presented in earlier reports [Natesan et al. 2008, 2009]. A viable approach to resolve these issues and the R&D priority were also recommended. The development of mechanistically based creep-fatigue interaction models for life prediction and reliable data extrapolation was chosen to be the central focus in near-term efforts. Our current efforts focus on the creep-fatigue damage issue in high-strength ferritic-martensitic steels for two primary reasons. First, the current ASME design rule of bilinear damage summation puts severe limits of fatigue and creep loads for mod.9Cr-1Mo (G91) ferritic-martensitic steel, the lead structural material for fast reactors; secondly, the ferritic-martensitic steels behave fundamentally differently from austenitic stainless steels, for which the current ASME creep-fatigue design rules were developed. The unique deformation and damage characteristics in G91 steel, e.g. cyclic softening, degradation of creep and rupture strength during cyclic service, demands a new creep-fatigue design procedure that explicitly accounts for the material’s unique creep-fatigue behavior. To support the development of predictive models and to resolve the over-conservative issue with the ASME design rule for G91 steel, we recovered stress relaxation data from thirteen creep-fatigue tests conducted at Oak Ridge National Laboratory in late 1980s and early 1990s, and conducted extensive data analysis. Based on this limited database and available literature data, we have developed a Cyclic Softening Model and a Stress Relaxation Model specifically applied to G91 steel. The Cyclic Softening Model incorporates strain hardening and creep deformation, and predicts the cyclic stress variation as a function of cycle number during creep-fatigue loading. The Stress Relaxation Model predicts the stress relaxation curve during the hold time of cyclic loading, based on the creep properties of the material. By incorporating the newly developed Cyclic Softening Model and Stress Relaxation Model with the Bilinear Damage Model, we have evaluated the creep-fatigue damage in G91 steel using experimental data without any safety margins. The analysis provided a more realistic assessment of creep-fatigue damage and identified critical factors that need to be considered in future advanced models. We have found that due to the significant cyclic softening in G91 steel, the unit creep damage in each cycle decreased continuously with increasing number of cycles. This continuous change in creep damage with cycling has significant implication in the evaluation of the total creep-fatigue damage. The total creep damage can be one order of magnitude lower when the unit creep damage at half-life was used in the evaluation in comparison with the calculations using the unit creep damage at cycle 10. An important conclusion is the accurate assessment of the creep-fatigue damage in G91 steel must consider the cycling history. Future work will refine the Cyclic Softening Model and Stress Relaxation Model and integrate these two models with the Damage Rate Model for life prediction under creep-fatigue loading. An accelerated creep-fatigue testing methodology is proposed for the design of model-oriented experiments to guide and validate model development. Key material parameters in the model will be determined by creep and fatigue data, obtained under symmetric continuous cycling (fast-fast and slow-slow) and asymmetric cycling (slow-fast and fast-slow) which often lead to accelerated creep-fatigue damage. The established model, based on creep and fatigue properties, will then be used to predict the life under creep-fatigue loading with hold-time in either tension or compression or combined tension and compression. With the success of this mechanistically based life prediction model for G91 steel, the model will be extended to advanced ferritic-martensitic steels, such as NF616.

Li, Meimei↗

Forward and inverse predictive transient models of TREAT using surrogate reactivity models

In this work, we present two novel approaches for predicting the power evolution and control rod height of the Transient Reactor Test Facility (TREAT) to support experiment modeling; specifically, transient analysis of the NASA-sponsored Sirius series of experiments. These approaches utilize steady-state Monte Carlo model, point kinetics model, and surrogate models to predict power evolution and control rod axial position during transient experiments. Both approaches were tested and validated against several Sirius experiments that were performed in the TREAT facility at different power levels. The validation test results show very good agreement with the experimental data, and the models were able to accurately predict the power evolution and the axial control rod position with an average error within 3.0%. Finally, this indicates that these approaches will help the reactor engineering team of the TREAT facility in preparing and predicting the power and temperature of the experiment.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

A predictive standard model for heavy electron systems

We propose a predictive standard model for heavy electron systems based on a detailed phenomenological two-fluid description of existing experimental data. It leads to a new phase diagram that replaces the Doniach picture, describes the emergent anomalous scaling behavior of the heavy electron (Kondo) liquid measured below the lattice coherence temperature, T*, seen by many different experimental probes, that marks the onset of collective hybridization, and enables one to obtain important information on quantum criticality and the superconducting/antiferromagnetic states at low temperatures. Because T* is {approx} J{sup 2} {rho}/2, the nearest neighbor RKKY interaction, a knowledge of the single-ion Kondo coupling, J, to the background conduction electron density of states, {rho}, makes it possible to predict Kondo liquid behavior, and to estimate its maximum superconducting transition temperature in both existing and newly discovered heavy electron families.

36 MATERIALS SCIENCE↗

A newly proposed isotherm model to predict Cs exchange with crystalline silicotitanate in tank waste simulants

The Zheng Anthony Miller (ZAM) computer model, a multicomponent ion exchange model used to predict the exchange of Group I metals onto crystalline silicotitanate (CST), has historically been used to predict Cs distribution coefficients from Hanford and Savannah River Site (SRS) tank waste simulants. Comparison of experimentally determined Cs distribution coefficients from tank waste simulants with ZAM isotherm model predictions indicate overprediction of Cs and K distribution coefficients for simple and complex simulants with the engineered form of CST. Additionally, recent changes in chemical composition/manufacturing of IONSIV TM R9140-B have resulted in increased Cs capacity from high-salt, highly alkaline solutions. Here, this work served to assess different isotherm models and refine equilibrium parameters to develop a model that can be applied to Hanford and SRS tank waste Cs removal efforts. Toward this goal, the Campbell Westesen Peterson (CWP) model was developed. This model utilized the experimentally determined Cs capacity, and simplified ZAM equilibria expressions to include only the binary substitution of Cs + or K + on the Na + sites. Equilibrium constants for these equations were refined using experimentally determined distribution coefficients. Overall, the CWP model significantly improved our ability to predict both Cs and K loading capacity from complex matrices.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Data-Enabled Predictive Control for Building HVAC Systems

Model predictive control is widely used as a control technology for the computation of optimal control inputs of building heating, ventilating, and air conditioning (HVAC) systems. However, both the benefits and widespread adoption of model predictive control (MPC) are hindered by the effort of model creation, calibration, and accuracy of the predictions. In this paper, we apply the data-enabled predictive control (DeePC) algorithm for designing controls for building HVAC systems. The algorithm solely depends on input/output data from the system to predict future state trajectories without the need for system identification. The algorithm relies on the idea that a vector space of all input–output trajectories of a discrete-time linear time-invariant (LTI) system is spanned by time-shifts of a single measured trajectory, given the input signal is persistently exciting. Closed-loop simulations using EnergyPlus are performed to demonstrate the approach. The simulated building modeled in EnergyPlus is a modified commercial large office prototype building served by an air handling unit-variable air volume HVAC system. Temperature setpoints of zones are used as control variables to minimize the HVAC energy cost of the building considering a time-of-use electricity rate structure. Furthermore, sensitivity analysis is conducted to gain insights into the effect of parameter tuning on DeePC performance. Simulation results are used to illustrate the performance of the algorithm and compare the algorithm with model-based MPC and occupancy-based setpoint controller. Overall, DeePC achieves similar performance compared to MPC for lower engineering effort.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Zonal-Based Emission Source Term Model for Predicting Particulate Emission Factors in Wildfire Simulations

A physics/chemistry-based numerical model for predicting the emission of fine particles from wildfires is proposed. This model implements the fundamental mechanisms of soot formation in a combustion environment: soot nucleation, surface growth, agglomeration, oxidation, and particle fragmentation. These mechanisms occur on a scale too fine for the discretization of most wildfire models, which need to simulate landscape-scale dynamics. As a result this model implements a zonal approach, where the computed soot particle distribution is partitioned into process zones within a single resolved grid cell. These process zones include: an inception zone (for nucleation), a heating zone (for coagulation, surface growth, and fragmentation), a reaction zone (for oxidation), and a quenched zone (for atmospheric processes). Governing mechanisms are applied to the appropriate zones to predict total particle growth and emission. Additionally, the proposed model is implemented into HIGRAD/FIRETEC, a physics-based wildfire simulation code which couples interactions between fire, fuels, atmosphere, and topography on a landscape scale. Fire simulations among grasslands and conifer forests are performed and compared against experimental data for emission factors.

54 ENVIRONMENTAL SCIENCES↗

Nuclear imaging of the fuel assembly in ignition experiments

First results from the analysis of neutron image data collected on implosions of cryogenically layered deuterium-tritium capsules during the 2011-2012 National Ignition Campaign are reported. The data span a variety of experimental designs aimed at increasing the stagnation pressure of the central hotspot and areal density of the surrounding fuel assembly. Images of neutrons produced by deuterium–tritium fusion reactions in the hotspot are presented, as well as images of neutrons that scatter in the surrounding dense fuel assembly. The image data are compared with 1D and 2D model predictions, and consistency checked using other diagnostic data. The results indicate that the size of the fusing hotspot is consistent with the model predictions, as well as other imaging data, while the overall size of the fuel assembly, inferred from the scattered neutron images, is systematically smaller than models' prediction. Preliminary studies indicate these differences are consistent with a significant fraction (20%–25%) of the initial deuterium-tritium fuel mass outside the compact fuel assembly, due either to low mode mass asymmetry or high mode 3D mix effects at the ablator-ice interface.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Mathematical Analysis of the Solidification Behavior of Plain Steel Based on Solute- and Heat-Transfer Equations in the Liquid–Solid Zone

An analytical approximate solution to non-linear solute- and heat-transfer equations in the unsteady-state mushy zone of Fe-C plain steel has been obtained, assuming a linear relationship between the solid fraction and the temperature of the mushy zone. The heat transfer equations for both the solid and liquid zone along with the boundary conditions have been linked with the equations to solve the whole equations. The model predictions (e.g., the solidification constants and the effective partition ratio) agree with the generally accepted values and with a separately performed numerical analysis. The solidus temperature predicted by the model is in the intermediate range of the reported formulas. The model and Neuman’s solution are consistent in the low carbon range. A conventional numerical heat analysis (i.e., an equivalent specific heat method using the solidus temperature predicted by the model) is consistent with the model predictions for Fe-C plain steels. The model presented herein simplifies the computations to solve the solute- and heat-transfer simultaneous equations while searching for a solidus temperature as a part of the solution. Thus, this model can reduce the complexity of analyses considering the heat- and solute-transfer phenomena in the mushy zone.

36 MATERIALS SCIENCE↗

Dynamic pulse propagation modelling for predictive femtosecond-laser-microbonding of transparent materials

A dynamic pulse propagation modeling for femtosecond laser bonding of Borofloat glass is presented. The temperature evolution and internal modifications are predicted by incorporating the nonlinear electron dynamics along with temperature dependent thermal properties. The modelling predicts the spatial and temporal distribution of absorption coefficient and plasma density that gives quantitative estimations of the heat affected zone and weld geometry. The impact of focusing condition on heat affected zone and weld geometry is investigated, which for the first time to our knowledge allows to numerically determine the desired relative position between the geometrical focus of a femtosecond-laser-pulse and the interface of the two substrates to be welded. The prediction of the modelling on the offset distance is applied to weld Borofloat glass plates having optical contact and can be applied to other dielectric solids.

Sahoo, Pankaj K. (ORCID:0000000209437319)↗