On the Bayesian treed multivariate Gaussian process with linear model of coregionalization
Not Available
SEARCH · Search NASA
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.
Not Available
With recent developments in computational techniques, Data-driven Machine Learning Models (DMLs) have shown great potential in simulating streamflow and capturing the rainfall-runoff relationship in given watersheds, which are traditionally fulfilled by Process-based Hydrologic Models (PHMs). There are debates on whether the DMLs can outperform and possibly replace the classical PHMs for streamflow simulation and river forecasting, but no clear conclusions have been made. This study aims to investigate whether the newer DMLs have any potential in further improving the simulation accuracy of classical PHMs, and vice versa. To do this, we compared a few popular PHMs and DMLs over four watersheds across the Continental US (CONUS) that are associated with different input, climate, and regional conditions. A total of five hydrologic models were chosen, including (1) two classical lumped models, i.e., the Sacramento Soil Moisture Accounting (SAC-SMA) and Xinanjiang (XAJ); (2) one modern distributed model, termed Coupled Routing and Excess Storage (CREST); (3) and two DMLs including an Artificial Neural Networks (ANN) and a deep learning model, termed Long Short Term Memory (LSTM). Our results demonstrated that the DMLs still significantly biased when using the baseline input scenario with the PHMs. However, the DMLs fed with delayed input scenarios had great potential and can reach high simulation accuracy. The DMLs, especially the ANN, outperformed other employed models under the rainfall-runoff relationship in which rainfall dominantly drives. Furthermore, the DMLs also showed better performance in the high-flow regime, while the PHMs had a better performance for the low-flow regime, implying both PHMs and DMLs have their own merits and are worthy of joint development. In general, our study indicated a great potential of using DMLs to simulate streamflow, but further studies are still needed to verify the transferability and scalability of DMLs in large-scale experiments, such as the Distributed Model Intercomparison Projects 1&2 conducted by National Weather Services but to compare modern DMLs and PHMs.
A model is presented which can be used to determine the appropriate values of the process variables for filament winding a cylinder. The model provides the cylinder temperature, viscosity, degree of cure, fiber position and fiber tension as functions of position and time during the filament winding and subsequent cure, and the residual stresses and strains within the cylinder during and after the cure. A computer code was developed to obtain quantitative results. Sample results are given which illustrate the information that can be generated with this code.
This work presents a machine learning (ML) framework that model the anisotropic actuation responses in a shape memory alloy. A Gaussian Process Regression (GPR) based ML model is trained on a set of different crystal orientations subjected to different actuation conditions. The training employed thermo-mechanical responses from a crystal-plasticity model that captures phase-transformation, stress-induced plasticity, and transformation-induced plasticity. Further, on training the GPR-ML model at fixed stress level for different orientations, it captured the thermo-mechanical responses accounting for the anisotropy, and predicted responses for new orientations with good accuracy. The GPR-ML model is able to capture the transformation temperature variations even when trained using multiple stress levels, and the transformation strain showed significant deviations. The developed GPR-ML model gave reasonable predictions for an unexplored sample set of orientations and loading conditions.
The NASA s Evolutionary Xenon Thruster (NEXT) project is developing an advanced ion propulsion system for future NASA missions for solar system exploration. An engineering model (EM) power processing unit (PPU) for the NEXT project was designed and fabricated by L-3 Communications under contract with NASA Glenn Research Center (GRC). This modular PPU is capable of processing up from 0.5 to 7.0 kW of output power for the NEXT ion thruster. Its design includes many significant improvements for better performance over the state-of-the-art PPU. The most significant difference is the beam supply which is comprised of six modules and capable of very efficient operation through a wide voltage range because of innovative features like dual controls, module addressing, and a high current mode. The low voltage power supplies are based on elements of the previously validated NASA Solar Electric Propulsion Technology Application Readiness (NSTAR) PPU. The highly modular construction of the PPU resulted in improved manufacturability, simpler scalability, and lower cost. This paper describes the design of the EM PPU and the results of the bench-top performance tests.
Unscheduled shutdown of nuclear power facilities for recalibration and replacement of faulty sensors can be expensive and disruptive to grid management. In this work, we present virtual (software) sensors that can replace a faulty physical sensor for a short duration thus allowing recalibration to be safely deferred to a later time. The virtual sensor model uses a Gaussian process model to process input data from redundant and other nearby sensors. Predicted data includes uncertainty bounds including spatial association uncertainty and measurement noise and error. Using data from an instrumented cooling water flow loop test bed, the virtual sensor model has predicted correct sensor measurements and the associated error corresponding to a faulty sensor. (authors)
Unscheduled shutdown of nuclear power facilities for recalibration and replacement of faulty sensors can be expensive and disruptive to grid management. In this work, we present virtual (software) sensors that can replace a faulty physical sensor for a short duration thus allowing recalibration to be safely deferred to a later time. The virtual sensor model uses a Gaussian process model to process input data from redundant and other nearby sensors. Predicted data includes uncertainty bounds including spatial association uncertainty and measurement noise and error. Using data from an instrumented cooling water flow loop testbed, the virtual sensor model has predicted correct sensor measurements and the associated error corresponding to a faulty sensor.
The application of neural network models to scientific machine learning tasks has proliferated in recent years. In particular, neural networks have proved to be adept at modeling processes with spatial–temporal complexity. Nevertheless, these highly parameterized models have garnered skepticism in their ability to produce outputs with quantified error bounds over the regimes of interest. Hence there is a need to find uncertainty quantification methods that are suitable for neural networks. In this work we present comparisons of the parametric uncertainty quantification of neural networks modeling complex spatial–temporal processes with Hamiltonian Monte Carlo and Stein variational gradient descent and its projected variant. Specifically we apply these methods to graph convolutional neural network models of evolving systems modeled with recurrent neural network and neural ordinary differential equations architectures. We show that Stein variational inference is a viable alternative to Monte Carlo methods with some clear advantages for complex neural network models. For our exemplars, Stein variational interference gave similar pushed forward uncertainty profiles through time compared to Hamiltonian Monte Carlo, albeit with generally more generous variance. As a result, projected Stein variational gradient descent also produced similar uncertainty profiles to the non-projected counterpart, but large reductions in the active weight space were confounded by the stability of the neural network predictions and the convoluted likelihood landscape.
Brown University is part of the Collaboratory on Mathematics for Mesoscopic Modeling of Materials (CM4), led by PNNL. The Collaboratory focuses on developing rigorous mathematical foundations for understanding and controlling fundamental mechanisms in mesoscale processes to enable scalable synthesis of complex materials, through the design of efficient modeling methods and corresponding scalable algorithms.
In this work, the production of Lithium hexafluorophosphate (LiPF 6 ) for Lithium-ion battery application is studied. Spreadsheet-based process models are developed to simulate three different production processes. These process models are then used to estimate and analyze the factors affecting cost of manufacturing, energy demand, and environmental impact due to greenhouse gas (GHG) emissions. Moreover, the results indicate that in a facility with a capacity of making 10,000 metric tons per year of LiPF 6 the cost of production is around $20 per kg of LiPF 6 , energy consumption is around 30 GWh per year, and the emission of greenhouse gases in CO 2- equivalent gases is around 80 metric tons per day. The impact of change in process and economic parameters on the cost of production, energy demand, and emissions is studied. In addition, a few insights on reducing the cost of production are presented. Lastly, the impact of varying LiPF 6 costs on the overall cost of a Li-ion battery ($ kWh -1 ) is presented.
Soil carbon (C) and nitrogen (N) cycles and their complex responses to hydro-climatic forcing have gained increasing attention. While the temperature effects have been intensively studied, soil moisture response functions (SMRFs) are not well documented for various microbial and enzymatic processes due to the difficulties in directly measuring and differentiating the moisture effects on various processes. In this work we extended our C-only Microbial-ENzyme Decomposition (MEND) model to the C-N coupled MEND model with flexible element stoichiometry. Our model calibration showed good agreement between simulated and observed C:N ratios in soil organic matter and microbial biomass, as well as the ammonium and nitrate concentrations. We show that the selection of SMRFs for specific biogeochemical processes could result in significant differences in model simulated microbial and C-N processes. In particular, it is essential to account for the soil moisture effects on microbial dormancy and resuscitation, as the changes in microbial physiology under favorable or stressful conditions will exert strong controls on soil C and N dynamics. We also advocate the utilization of dynamic (time-variant) data (e.g., litter input, N deposition, soil temperature and moisture), instead of time-invariant data, to drive model simulations and analyses. Dynamic forcing data (particularly dynamic soil moisture) better represent the real-world climate and environmental conditions, which could facilitate more realistic modeling and understanding of soil C and nutrient cycling in a changing world.
Clay/shale has been considered as potential host rock for geological disposal of high-level radioactive waste throughout the world. Coupled thermal, hydrological, mechanical, and chemical (THMC) processes have a significant impact on the long-term safety of a clay repository. For example, the excavation damaged zone (EDZ) near repository tunnels can modify local permeability (resulting from induced fractures), potentially leading to less confinement capability. This report documents results from three R&D activities: (1) implementation and validation of constitutive relationships, (2) development of a discrete fracture network (DFN) model for investigating coupled processes in the EDZ, and (3) development of a THM model for the FE tests at Mont Terri, Switzerland, for the purpose of model validation. The overall objective of these activities is to provide an improved understanding of EDZ evolution in clay repositories and the associated coupled processes, and to develop advanced relevant modeling capabilities.
Producing low carbon Hydrogen at a competitive price is one of the challenges to hydrogen being part of the solution to reach net-zero emission targets set by the U.S. DOE by 2050. With projected near-term improvements in technology, hydrogen production via solid oxide electrolysis cell (SOEC) / high-temperature steam electrolysis (HTSE) integrated with existing light water reactor (LWR) Nuclear Power Plants (NPP-HTSE) can produce carbon-free hydrogen competitively. In the near term, a 10-year production tax credit (PTC) found in the Inflation Reduction Act (IRA) has been passed, which will catalyze the development and improvement of hydrogen production technology to be competitive. The “1-1-1” target set by the U.S. DOE is to reduce the cost of carbon-free hydrogen by 80% to $1 per kilogram in 1 decade. Several models are available to analyze the profitability, opportunity, and technical capability of NPP-HTSE systems. In order of complexity from most complex to least complex some of these models include: RAVEN/HERON, process models using Aspen HYSYS and capital expense estimations using Aspen Process Economic Analyzer (APEA) and levelized cost of hydrogen (LCOH) calculation using the H2A model (Hydrogen Analysis Model), and custom spread sheets built by the interested party. Though some of the more advanced existing models provide detailed analysis to complex grid integrated problems, they also can take considerable time to setup and run. These advanced models are well suited to complex grid integrated analysis and the consideration of flexibility and variability of regulated and de-regulated electricity price and advanced estimation of capital and operating expenses and heat and material balances.
Despite tremendous progress in algorithm development, computational efficiency and transition into operations over the past two decades, coastal modeling still lacks scientific rigor due to proliferation of many ‘gray’ areas related to various modeling choices made by modelers. Here, in this paper, we propose some guiding principles for the modeling community to improve performance, and we also debunk commonly held myths that make the coastal modeling lack rigor. Using our own experience in developing seamless cross-scale unstructured-grid based models for the past two decades, we describe in unprecedented detail the end-to-end modeling process (i.e., from digital elevation models (DEMs) to mesh generation to post analysis), and demonstrate that defensible modeling is within reach for any end user by following three guiding principles: (1) Bathymetry is a first order forcing in coastal domains and thus should be respected in all aspects of modeling; (2) Oceanographic processes are driven across multiple spatial scales and so models should enable appropriate resolution as needed; and (3) Model assessment should focus on physical processes. Through qualitative and quantitative model assessments, we demonstrate the fundamental role played by bathymetry/topography as embedded in DEMs in making the results defensible, which is unfortunately glossed over in many modeling studies. Focusing on process-based assessment simplifies the calibration process. A major conclusion of this work is that model developers and operators should maximize the scientific rigor for in silico oceanography by avoiding some common pitfalls that rely on error compensation at the expense of representation of physical system processes. We present some best practice procedures for defensive and trustworthy numerical modeling.
Multilayer (ML) plastic films are essential packaging materials that help protect products from diverse external factors; however, only 5% of all ML films are recycled in the United States. Solvent-based technologies are a promising alternative for recycling ML films because they enable recovery of constituent polymer resins. For example, the Solvent Targeted Recovery and Precipitation (STRAPTM) process sequentially dissolves and separates polymer components using a series of targeted solvent washes. A crucial design aspect of this process is the impact of selected solvents on human health and on the environment. Here, this work introduces a computational framework that integrates molecular modeling, process modeling, techno-economic analysis (TEA), and life-cycle analysis (LCA) to quickly screen green solvents for solvent-based ML recycling processes. Initial screening for solvents based on selectivity is performed by estimating temperature-dependent solubilities using molecular-scale models. Subsequent screening uses basic estimates of energy use and octanol-water partition coefficients (logP) as key measures of health, safety, and environmental hazards. Detailed process modeling, TEA, and LCA are used on a reduced set of promising solvents identified in early screening steps to more accurately determine how solvent selection and associated operating conditions impact overall economics and environmental impacts. The framework is used for the identification of green solvents (from a database of 1,000 solvents) that separate an industrial ML film composed of polyethylene (PE), ethylene vinyl alcohol (EVOH), and polyethylene terephthalate (PET). Our analysis shows the effectiveness of the framework and reveals fundamental trade-offs between solvent greenness, solubility, and economics. Our work emphasizes the importance of taking a holistic systems view during solvent design and aims to inform the development of new processes for ML film recycling and the identification of new ML films that are easier to recycle.
Plant trait measurements are needed for evaluating ecological responses to environmental conditions and for ecosystem process model development, parameterization, and testing. Here, we present a standardized dataset integrating measurements from projects conducted by the Terrestrial Ecosystem Research and Regional Analysis- Pacific Northwest (TERRA-PNW) research group between 1999 and 2014 across Oregon and Northern California, where measurements were collected for scaling and modeling regional terrestrial carbon processes with models such as Biome-BGC and the Community Land Model. The dataset contains measurements of specific leaf area, leaf longevity, leaf carbon and nitrogen for 35 tree and shrub species derived from more than 1,200 branch samples collected from over 200 forest plots, including several AmeriFlux sites. The dataset also contains plot-level measurements of forest composition, structure (e.g., tree biomass), and productivity, as well as measurements of soil structure (e.g., bulk density) and chemistry (e.g., carbon). Publically-archiving regional datasets of standardized, co-located, and geo-referenced plant trait measurements will advance the ability of earth system models to capture species-level climate sensitivity at regional to global scales.
A recent advancement in the conversion of methanol to hydrocarbon fuels is the development of a Cu-modified beta zeolite catalyst (Cu/BEA) that exhibits a higher productivity and yield than the parent H-form BEA catalyst in the production of branched C4-C8 hydrocarbons. This Cu/BEA catalyst enables process improvements in the High-Octane Gasoline (HOG) pathway, which operates under lower severity conditions than traditional methanol-to-fuels processes, offering a lower coke formation rate, higher selectivity to C5+gasoline-range hydrocarbons, and subsequently, improved process economics based on a recent process model and techno-economic analysis (TEA). Here, a market-responsive biorefinery concept around methanol and the HOG pathway is presented, where the major C4 products, isobutane and isobutene, serve as versatile intermediates that can either be recycled to increase the yield of the HOG product or directed to jet-fuel range hydrocarbons. Based on simulated recycle experiments with 13C-labeled isobutane, the Cu/BEA catalyst achieved increased yield to the HOG products during DME homologation by activating the co-fed isobutane and incorporating it into the chain growth pathway, as evidenced by the presence of isotopically-labeled 13C in C5+ products. Olefin coupling of an isobutene-rich mixture was also investigated, and the fuel properties of the resulting synthetic kerosene product were within the specifications for a typical jet-fuel. These experimental results were incorporated into an updated Aspen-based process model and TEA to quantify their impact on a market-responsive biorefinery concept. The improvements offered by C4 recycling with the Cu/BEA catalyst result in a decrease in the methanol-to-fuel synthesis cost from 101 cents per gallon of gasoline equivalent (GGE) with the baseline BEA catalyst to 72 cents/GGE, due in part to a significantly increased HOG yield from 40 to 55 GGE/dry-ton biomass. Expanding the process model to direct C4 products to a jet-range hydrocarbon product demonstrated a yield of 42 GGE/dry-ton HOG with 12 GGE/dry-ton distillate, resulting in a minor increase in the fuel synthesis cost to 77 cents/GGE versus the HOG-only case. With continued development, these advancements in fuel yield and reductions in cost position a market-responsive biorefinery around the HOG pathway to be competitive against mature fuel-synthesis technologies from a variety of carbon sources through methanol.
Electrochemical reduction of CO 2 gas is a novel CO 2 utilization technique that has the potential to mitigate the global climate crisis caused by anthropogenic CO 2 emissions, and enable the large-scale storage of energy generated from renewable sources in the form of carbon-based chemicals and fuels. However, due to the complexity of the electrochemical reactions, the explicit first-principles models for CO2 reduction are not available yet, and there has been a limited effort to develop process modeling, optimization and control of CO 2 electrochemical reactors. To this end, a rotating cylinder electrode (RCE) reactor has been constructed at UCLA to understand the mass transfer and reaction kinetics effects separately on the productivity. In the RCE reactor, the applied potential strongly influences the reaction energetics and the electrode rotation speed affects the hydrodynamic boundary layer and modifies the film mass transfer coefficient, which involves convective and diffusive transport. Further, the present work aims to develop a multi-input multi-output (MIMO) control scheme for the RCE reactor that integrates techniques from artificial and recurrent neural network modeling, nonlinear optimization, and process controller design. Specifically, production rates of two products from the experimental reactor, ethylene and carbon monoxide, are controlled by manipulating two inputs, applied potential and catalyst rotation speed. Process dynamics and controllability are analyzed, a feedback control strategy is designed and the controllers are tuned accordingly. The experimental electrochemical cell is employed to gather data for process modeling and implement the multivariable control system. Finally, the experimental results are presented which demonstrate excellent closed-loop performance by the control system and regulation of the outputs at three different set-points including an economically-optimal set-point.