Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “performance 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 901 records · Page 50

Development of Data-Driven Models for Performance Prediction and Chemical Dosing of a Full-Scale Controlled Phosphorus Precipitation Reactor

This study evaluated the use of data-driven models to improve control of a struvite precipitation reactor that removes phosphorus from wastewater while producing a fertilizer product. The researchers developed predictive models for influent orthophosphate concentration, effluent orthophosphate concentration, and phosphorus removal using operational data from a full-scale MagPrex™ reactor at a water resource recovery facility in Denver, Colorado. Model predictions were used to recommend magnesium chloride dosing adjustments needed to achieve a target effluent phosphorus concentration. Several machine learning approaches were tested, with ridge regression providing the best predictions for influent orthophosphate concentration and phosphorus removal, and XGBoost providing the best predictions for effluent orthophosphate concentration. Simulation results indicated that the decision-support approach could correctly identify dosing adjustments in most cases and reduce chemical use. Full-scale implementation achieved lower accuracy due to changing operating conditions and limited historical data in some operating ranges. Here, the results demonstrate the potential of data-driven tools to support phosphorus recovery process control while also identifying practical limitations that affect deployment in full-scale systems.

42 ENGINEERING↗

Software design space exploration for exascale combustion co-design

The design of hardware for next-generation exascale computing systems will require a deep understanding of how software optimizations impact hardware design trade-offs. In order to characterize how co-tuning hardware and software parameters affects the performance of combustion simulation codes, we created ExaSAT, a compiler-driven static analysis and performance modeling framework. Our framework can evaluate hundreds of hardware/software configurations in seconds, providing an essential speed advantage over simulators and dynamic analysis techniques during the co-design process. Our analytic performance model shows that advanced code transformations, such as cache blocking and loop fusion, can have a significant impact on choices for cache and memory architecture. Our modeling helped us identify tuned configurations that achieve a 90% reduction in memory traffic, which could significantly improve performance and reduce energy consumption. These techniques will also be useful for the development of advanced programming models and runtimes, which must reason about these optimizations to deliver better performance and energy efficiency.

97 MATHEMATICS AND COMPUTING↗

A two-equation turbulence model for compressible flows

A k-epsilon model of turbulence that accounts for pressure gradients and compressibility effects is presented. To correct the degradation of k-epsilon model performance when pressure gradients are introduced, an extra term is added to the dissipation equation. By including this term and by adding terms to account for compressibility effects, a new k-epsilon model is developed for shear flows. The results show that the new model performs better than the existing k-epsilon models.

Abid, Ridha↗

Systems, methods and computer-readable media to model kinetic performance of rechargeable electrochemical devices

A system includes an electrochemical cell, monitoring hardware, and a computing system. The monitoring hardware samples performance characteristics of the electrochemical cell. The computing system determines cell information from the performance characteristics. The computing system also analyzes the cell information of the electrochemical cell with a Butler-Volmer (BV) expression modified to determine exchange current density of the electrochemical cell by including kinetic performance information related to pulse-time dependence, electrode surface availability, or a combination thereof. A set of sigmoid-based expressions may be included with the modified-BV expression to determine kinetic performance as a function of pulse time. The determined exchange current density may be used with the modified-BV expression, with or without the sigmoid expressions, to analyze other characteristics of the electrochemical cell. Model parameters can be defined in terms of cell aging, making the overall kinetics model amenable to predictive estimates of cell kinetic performance along the aging timeline.

97 MATHEMATICS AND COMPUTING↗

Incorporating microbial dormancy dynamics into soil decomposition models to improve quantification of soil carbon dynamics of northern temperate forests

Abstract Soil carbon dynamics of terrestrial ecosystems play a significant role in the global carbon cycle. Microbial‐based decomposition models have seen much growth recently for quantifying this role, yet dormancy as a common strategy used by microorganisms has not usually been represented and tested in these models against field observations. Here we developed an explicit microbial‐enzyme decomposition model and examined model performance with and without representation of microbial dormancy at six temperate forest sites of different forest types. We then extrapolated the model to global temperate forest ecosystems to investigate biogeochemical controls on soil heterotrophic respiration and microbial dormancy dynamics at different temporal‐spatial scales. The dormancy model consistently produced better match with field‐observed heterotrophic soil CO 2 efflux ( R H ) than the no dormancy model. Our regional modeling results further indicated that models with dormancy were able to produce more realistic magnitude of microbial biomass (<2% of soil organic carbon) and soil R H (7.5 ± 2.4 Pg C yr −1 ). Spatial correlation analysis showed that soil organic carbon content was the dominating factor (correlation coefficient = 0.4–0.6) in the simulated spatial pattern of soil R H with both models. In contrast to strong temporal and local controls of soil temperature and moisture on microbial dormancy, our modeling results showed that soil carbon‐to‐nitrogen ratio (C:N) was a major regulating factor at regional scales (correlation coefficient = −0.43 to −0.58), indicating scale‐dependent biogeochemical controls on microbial dynamics. Our findings suggest that incorporating microbial dormancy could improve the realism of microbial‐based decomposition models and enhance the integration of soil experiments and mechanistically based modeling.

He, Yujie↗

Machine Learning Surrogates of a Fuel Matrix Degradation Process Model for Performance Assessment of a Nuclear Waste Repository

Spent nuclear fuel repository simulations are currently not able to incorporate detailed fuel matrix degradation (FMD) process models due to their computational cost, especially when large numbers of waste packages breach. The current paper uses machine learning to develop artificial neural network and k-nearest neighbor regression surrogate models that approximate the detailed FMD process model while being computationally much faster to evaluate. Further, using fuel cask temperature, dose rate, and the environmental concentrations of CO 3 2- , O 2 , Fe 2+ , and H 2 as inputs, these surrogates show good agreement with the FMD process model predictions of the UO 2 degradation rate for conditions within the range of the training data. A demonstration in a full-scale shale repository reference case simulation shows that the incorporation of the surrogate models captures local and temporal environmental effects on fuel degradation rates while retaining good computational efficiency.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Evaluation of 2 1-D cloud models for the analysis of VAS soundings

Evaluation of the satellite Visual Infrared Spin Scan Radiometer Atmospheric Sounder (VISSR) has begun to document several of its critical shortcomings as far as numerical cloud models are concerned: excessive smoothing of thermal inversions; imprecise measurement of boundary layer moisture; and tendency to exaggerate atmospheric stability. The sensitivity of 1-D cloud models to their required inputs is stressed with special attention to those parameters obtained from atmospheric soundings taken by the VAS or rawinsonde. In addition to performing model experiments using temperature and moisture profiles having the general characteristics of VAS soundings, standard input sensitivity tests were made and 1-D model performance was compared with observations and the results of a 2-D model experiment using AVE/VAS data (Atmospheric Variability Experiment). Although very encouraging, the results are not sufficient to make any specific conclusions. In general, the VAS soundings are likely to be inadequate to provide the cloud base (and subcloud layer) information needed for inputs to current cumulus models. Above cloud base, the tendency to exaggerate the stability of the atmosphere requires solution before meaningful model experiments are run.

Emmitt, G. D.↗

Final Technical Report: Increasing Prediction Accuracy.

PV performance models are used to quantify the value of PV plants in a given location. They combine the performance characteristics of the system, the measured or predicted irradiance and weather at a site, and the system configuration and design into a prediction of the amount of energy that will be produced by a PV system. These predictions must be as accurate as possible in order for finance charges to be minimized. Higher accuracy equals lower project risk. The Increasing Prediction Accuracy project at Sandia focuses on quantifying and reducing uncertainties in PV system performance models.

14 SOLAR ENERGY↗

The Continual Intercomparison of Radiation Codes: Results from Phase I

We present results from Phase I of the Continual Intercomparison of Radiation Codes (CIRC), intended as an evolving and regularly updated reference source for evaluation of radiative transfer (RT) codes used in Global Climate Models. CIRC differs from previous intercomparisons in that it relies on an observationally validated catalogue of cases. The seven CIRC Phase I baseline cases, five cloud-free, and two with overcast liquid clouds, are built around observations by the Atmospheric Radiation Measurements (ARM) program that satisfy the goals of Phase I, namely to examine radiative transfer (RT) model performance in realistic, yet not overly complex, atmospheric conditions. In addition to the seven baseline cases, additional idealized "subcases" are also examined to facilitate intrepretation of the causes of model errors. In addition to summarizing individual model performance with respect to reference line-by-line calculations and inter-model differences, we also highlight RT model behavior for conditions of doubled CO2, aspects of utilizing a spectral specification of surface albedo, and the impact of the inclusion of scattering in the thermal infrared. Our analysis suggests that RT models should work towards improving their calculation of diffuse shortwave flux, shortwave absorption, treatment of spectral surface albedo, and shortwave CO2 forcing. On the other hand, LW calculations appear to be significantly closer to the reference results. By enhancing the range of conditions under which participating codes are tested, future CIRC phases will hopefully allow even more rigorous examination of RT code performance.

radiative transfer, climate models, radiative forc↗

A model-data intercomparison of CO2 exchange across North America: Results from the North American Carbon Program site synthesis

Our current understanding of terrestrial carbon processes is represented in various models used to integrate and scale measurements of CO2 exchange from remote sensing and other spatiotemporal data. Yet assessments are rarely conducted to determine how well models simulate carbon processes across vegetation types and environmental conditions. Using standardized data from the North American Carbon Program we compare observed and simulated monthly CO2 exchange from 44 eddy covariance flux towers in North America and 22 terrestrial biosphere models. The analysis period spans 220 site-years, 10 biomes, and includes two large-scale drought events, providing a natural experiment to evaluate model skill as a function of drought and seasonality. We evaluate models' ability to simulate the seasonal cycle of CO2 exchange using multiple model skill metrics and analyze links between model characteristics, site history, and model skill. Overall model performance was poor; the difference between observations and simulations was 10 times observational uncertainty, with forested ecosystems better predicted than nonforested. Model-data agreement was highest in summer and in temperate evergreen forests. In contrast, model performance declined in spring and fall, especially in ecosystems with large deciduous components, and in dry periods during the growing season. Models used across multiple biomes and sites, the mean model ensemble, and a model using assimilated parameter values showed high consistency with observations. Models with the highest skill across all biomes all used prescribed canopy phenology, calculated NEE as the difference between GPP and ecosystem respiration, and did not use a daily time step.

0793↗

Carbon and energy fluxes in cropland ecosystems: a model-data comparison

Croplands are extremely productive ecosystems that contribute to land–atmosphere exchange of carbon, energy, and water during their short growing seasons. We evaluated and compared net ecosystem exchange (NEE), latent heat flux (LE), and sensible heat flux (H) simulated by a suite of ecosystem models at five agricultural eddy covariance flux tower sites in the central United States as part of the North American Carbon Program Site Synthesis project. Most of the models overestimated H and underestimated LE during the growing season, leading to overall higher Bowen ratios compared to the observations. Most models systematically under predicted NEE, especially at rain-fed sites. Certain crop-specific models that were developed considering the high productivity and associated physiological changes in specific crops better predicted the NEE and LE at both rain-fed and irrigated sites. Models with specific parameterization for different crops better simulated the inter-annual variability of NEE for maize-soybean rotation compared to those models with a single generic crop type. Stratification according to basic model formulation and phenological methodology did not explain significant variation in model performance across these sites and crops. The under prediction of NEE and LE and over prediction of H by most of the models suggests that models developed and parameterized for natural ecosystems cannot accurately predict the more robust physiology of highly bred and intensively managed crop ecosystems. When coupled in Earth System Models, it is likely that the excessive physiological stress simulated in many land surface component models leads to overestimation of temperature and atmospheric boundary layer depth, and underestimation of humidity and CO 2 seasonal uptake over agricultural regions.

54 ENVIRONMENTAL SCIENCES↗

Final report for “Extreme-scale Algorithms and Solver Resilience”

This is a joint project with principal investigators at Oak Ridge National Laboratory, Sandia National Laboratories, the University of California at Berkeley, and the University of Tennessee. Our part of the project involves developing performance models for highly scalable algorithms and the development of latency tolerant iterative methods. During this project, we extended our performance models for the Multigrid method for solving large systems of linear equations and conducted experiments with highly scalable variants of conjugate gradient methods that avoid blocking synchronization. In addition, we worked with the other members of the project on alternative techniques for resilience and reproducibility. We also presented an alternative approach for reproducible dot-products in parallel computations that performs almost as well as the conventional approach by separating the order of computation from the details of the decomposition of vectors across the processes.

97 MATHEMATICS AND COMPUTING↗

Models for Facilitating Government-Funded Activities in the Post-ISS LEO Ecosystem

The International Space Station National Lab (ISSNL) was established through Congressional authority in 2005 and has been managed by a nonprofit called the Center for the Advancement of Science in Space (CASIS) since 2011 to support research, development, education, and commercial activities in low-Earth orbit (LEO). NASA leadership is considering next steps for microgravity and orbital activities in LEO in preparation for the retirement of the International Space Station (ISS) and transition to Commercial LEO Destinations (CLDs) in the 2030 timeframe. Specifically, our study team within the NASA Office of Technology, Policy, and Strategy addressed the question: What are potential models for an ISS National Lab facilitating government-funded or subsidized activities on a commercial LEO platform after the transition of the ISS to one or more private platforms? To inform the trade space of potential models, we conducted over 40 discussions with stakeholders internal and external to NASA and reviewed over 35 documents related to LEO legislation, commercialization, models, and activities. Following the extensive literature review and discussions with diverse stakeholders, we defined and evaluated six models, in addition to the current ISSNL model. We assessed each model across three possible future scenarios (dynamic, steady, and limited growth of the commercial LEO market and activities) and across five stakeholder-driven model evaluation criteria (ability to meet NASA’s needs, adaptability, opportunity for collaboration, market sustainability, and equity and accessibility). The six models in the future trade space were: • Government Research Broker: Customizable research efforts on transport vehicles and CLDs • Innovation Campus: Modern terrestrial campus with workforce focus • Anchor Tenant: Long-term, substantial agreement for leasing space on a single CLD • Fee for Service: Free market approach with NASA-sponsored grants and service/data buys • Matchmaker: Neutral third-party connecting users to platforms • Institute Network: Network of separate but related efforts to enable commercial scaling The Government Research Broker model performed best across the scenarios, followed by Innovation Campus, Anchor Tenant, and Fee for Service. While Matchmaker and Institute Network exhibited positive aspects, the two models performed most favorably in future scenarios with well-established communities and markets. The six models presented in this study are representative of the potential trade space and are illustrative examples. NASA leadership can adjust models as desired to align to their priorities more closely by using combinations of the unique model features provided in our analysis. The optimal model is likely a combination of features from multiple models. Building upon this work to best position NASA and the diverse stakeholders of LEO activities for success post-ISS, further work will explore the near-term impact of preparing for the transition of the ISS to CLDs by identifying opportunities for modifying the current ISSNL-CASIS partnership. ACKNOWLEDGEMENTS We thank the many reviewers and discussants, internal and external to NASA, that contributed to the work.

Erica Rodgers↗

Using Measured Plane-of-Array Data Directly in Photovoltaic Modeling: Methodology and Validation

Measured plane-of-array (POA) irradiance may provide a lower-cost alternative to standard irradiance component data for photovoltaic (PV) system performance modeling without loss of accuracy. Previous work has shown that transposition models typically used by PV models to calculate POA irradiance from horizontal data introduce error into the POA irradiance estimates, and that measured POA data can correlate better to measured performance data. However, popular PV modeling tools historically have not directly used input POA data. This paper introduces a new capability in NREL's System Advisor Model (SAM) to directly use POA data in PV modeling, and compares SAM results from both POA irradiance and irradiance components inputs against measured performance data for eight operating PV systems.

14 SOLAR ENERGY↗

Improving performance of two-group interfacial area transport equation models using genetic algorithms and high resolution experimental data

In the two-fluid transport model, the coupling of mass, momentum and energy transfer between phases is highly dependent on interfacial area transfer terms. Several research efforts in the past have been focused on the development of an interfacial area transport equation model (IATE), in order to eliminate the drawbacks of static flow regime maps currently used in best-estimate thermal-hydraulic system codes. The IATE attempts to model the dynamics that are involved in two phase flows by accounting for the different interaction mechanisms affecting bubble transport in the flow. In general, current IATE models perform poorly in high void fraction flow regimes. As both models depend on closure relations that require empirically determined coefficients, a genetic algorithm is implemented to study their optimization. The preliminary study indicates that improvements are achievable by application of the genetic algorithm. Optimization of coefficients for an entire set of high-resolution wire mesh sensor (WMS) data provided a single set of optimal coefficients. The application of WMS-optimized coefficients to external tests improved IATE performance for similar hydraulic diameters and provided a neutral change in performance for significantly smaller/larger hydraulic diameters. This is a crucial outcome as an optimized set of coefficients that provide significant improvement for a subset of experimental data - is shown to benefit performance in external data for which the coefficients are not specifically optimized. (authors)

42 ENGINEERING↗

Predictive Battery Lifetime Modeling at NREL [Slides]

Battery lifetime models are used to extrapolate data from accelerated aging tests to simulate degradation in real-world applications such as electric vehicles and battery energy storage systems. Methods developed at NREL utilize both expert domain-knowledge and machine-learning to identify models, using statistical methods such as cross-validation and bootstrap resampling to interrogate model performance and quantify uncertainty. These models can be utilized in systems level simulations to predict battery performance or technoeconomic models to estimate the lifetime cost of battery systems.

25 ENERGY STORAGE↗

Quantum-Assisted Variational Segmentation for Image-to-Image Wildfire Detection Using Satellite Data

The quantum computing community has been searching for suitable applications to demonstrate the potential of near-term quantum devices. Quantum machine learning is a potential candidate, particularly using models that cannot be efficiently simulated with classical computers [1, 2]. This work focuses on a transition phase of quantum computers where the quantum machine learning model is still simulable classically but projected not to be simulable as the size of the model grows. Ultimately quantum computers may have advantages for high-dimensional real-world problems. Due to the limited number of qubits in current noisy intermediate-scale quantum (NISQ) devices, the direct application of quantum computers in high dimensional data is not feasible. To remedy this problem, an encoder-decoder architecture can be utilized. The encoder model would transform the high-dimensional data into a compact representation, to a level that small quantum computers can be used today (or in the near future), and the decoder would take the quantum processed outputs back to the high-dimensional space. Addressing the two challenges of quantum machine learning, this work investigates a hybrid supervised generative model with a quantum Ising Born machine embedded as the latent distribution. The model contains four main parts (Figure 1.a.): (1) a U-NET architecture responsible for learning segmentation flow, (2) a Prior network responsible for learning an encoded latent distribution of the input data, (3) a Born machine which represents the latent distribution, and (4) a Posterior network in charge of learning the joint encoded latent distribution of inputs and target data. The initial model, proposed by [3], is optimized by (1) maximizing the overlap of the prior and posterior latent distributions, and (2) minimizing the segmentation loss. The proposed model is designed to be investigated in a simulation environment applied to the real-world application of wildfire segmentation. Specifically, the model is designed to solve the patchy wildfire segmentations of Moderate Resolution Imaging Spectroradiometer (MODIS) by taking the MODIS observations and using Visible Infrared Imaging Radiometer Suite’s (VIIRS) consistent wildfire product as the target. The model solves patchy wildfire segmentations and provides insight into the epistemic errors sourced from model variation. The model utilizes the Born machine as a QUBO solver to represent the latent space as a Bernoulli distribution. The proposed configuration allows the variational segmentation model to leverage the true quantum probabilistic nature and derive a more expressive latent configuration, increasing the model performance in describing wildfire segmentations. The quantum probabilistic information of the Born machine is directly incorporated in the Kullback-Leibler divergence loss in the prior and posterior distributions, forcing the Bernoulli latent distribution to maximize the overlap of input and joint input-target distributions. The proposed model is then trained and compared with a baseline only consisting of direct Bernoulli latent distribution with no Born machine representing the latent space. The models are evaluated based on the segmentation metrics, such as precision, recall, intersect of union, with uncertainty boundaries accounting for the stochastic nature of the model. Our findings show that even in low latent-dimensional space (due to the limit in computational power of the classical quantum simulator), we are able to effectively capture the latent representation and hence the model performs better than the baseline. The findings are a projection for scaling the model into higher dimensional latent space with the Born machine surpassing the baseline performance. Figure 1. Sub-figure (a) demonstrates the architecture for the training phase. The model consists of a Prior and Posterior network that encode inputs and joint input-target data into compact representations, respectively. The Born machine represents the latent distribution, and the U-NET branch learns the segmentation patterns of the data. The stochasticity is introduced to the U-NET through its last layer to create meaningful but stochastic segmentations. Sub-figure (b) represents the inference phase where the model takes the stochastic behavior from the prior network and injects that into the U-NET. Each attempt of inference will generate different but similar segmentations from the same distribution of the wildfire event. REFERENCES [1] Coyle, B., Mills, D., Danos, V., & Kashefi, E. (2020). The Born supremacy: quantum advantage and training of an Ising Born machine. npj Quantum Information, 6(1), 1-11. [2] Liu, J. G., & Wang, L. (2018). Differentiable learning of quantum circuit born machines. Physical Review A, 98(6), 062324. [3] Kohl, S., Romera-Paredes, B., Meyer, C., De Fauw, J., Ledsam, J. R., Maier-Hein, K., ... & Ronneberger, O. (2018). A probabilistic u-net for segmentation of ambiguous images. Advances in neural information processing systems, 31.

quantum machine learning↗

A model-data intercomparison of CO2 exchange across North America: Results from the North American Carbon Program Site Synthesis

Our current understanding of terrestrial carbon processes is represented in various models used to integrate and scale measurements of CO{sub 2} exchange from remote sensing and other spatiotemporal data. Yet assessments are rarely conducted to determine how well models simulate carbon processes across vegetation types and environmental conditions. Using standardized data from the North American Carbon Program we compare observed and simulated monthly CO{sub 2} exchange from 44 eddy covariance flux towers in North America and 22 terrestrial biosphere models. The analysis period spans {approx}220 site-years, 10 biomes, and includes two large-scale drought events, providing a natural experiment to evaluate model skill as a function of drought and seasonality. We evaluate models' ability to simulate the seasonal cycle of CO{sub 2} exchange using multiple model skill metrics and analyze links between model characteristics, site history, and model skill. Overall model performance was poor; the difference between observations and simulations was {approx}10 times observational uncertainty, with forested ecosystems better predicted than nonforested. Model-data agreement was highest in summer and in temperate evergreen forests. In contrast, model performance declined in spring and fall, especially in ecosystems with large deciduous components, and in dry periods during the growing season. Models used across multiple biomes and sites, the mean model ensemble, and a model using assimilated parameter values showed high consistency with observations. Models with the highest skill across all biomes all used prescribed canopy phenology, calculated NEE as the difference between GPP and ecosystem respiration, and did not use a daily time step.

54 ENVIRONMENTAL SCIENCES↗