Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Models”

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 271 records · Page 15

Combined selection of the dynamic model and modeling error in nonlinear aeroelastic systems using Bayesian Inference

Here, we report a Bayesian framework for concurrent selection of physics-based models and (modeling) error models. We investigate the use of colored noise to capture the mismatch between the predictions of calibrated models and observational data that cannot be explained by measurement error alone within the context of Bayesian estimation for stochastic ordinary differential equations. Proposed models are characterized by the average data-fit, a measure of how well a model fits the measurements, and the model complexity measured using the Kullback–Leibler divergence. The use of a more complex error models increases the average data-fit but also increases the complexity of the combined model, possibly over-fitting the data. Bayesian model selection is used to find the optimal physical model as well as the optimal error model. The optimal model is defined using the evidence, where the average data-fit is balanced by the complexity of the model. The effect of colored noise process is illustrated using a nonlinear aeroelastic oscillator representing a rigid NACA0012 airfoil undergoing limit cycle oscillations due to complex fluid–structure interactions. Several quasi-steady and unsteady aerodynamic models are proposed with colored noise or white noise for the model error. The use of colored noise improves the predictive capabilities of simpler models.

42 ENGINEERING↗

Intercomparison of model simulations of mixed-phase clouds observed during the ARM Mixed-Phase Arctic Cloud Experiment. I: Single layer cloud

Results are presented from an intercomparison of single-column and cloud-resolving model simulations of a cold-air outbreak mixed-phase stratocumulus cloud observed during the ARM Mixed-Phase Arctic Cloud Experiment. The observed cloud occurred in a well-mixed boundary layer with a cloud top temperature of –15°C. While the cloud was water dominated, ice precipitation appears to have lowered the liquid water path to about 2/3 of the adiabatic value. The simulations, which were performed by seventeen single column and nine cloud-resolving models, generally underestimate the liquid water path with the median single-column and cloud-resolving model liquid water path a factor of 3 smaller than observed. While the simulated ice water path is in general agreement with the observed values, results from a sensitivity study in which models removed ice microphysics indicate that in many models the interaction between liquid and ice phase microphysics is responsible for the strong model underestimate of liquid water path. Although no single factor is found to lead to a good simulation, these results emphasize the need for care in the model treatment of mixed-phase microphysics. This case study, which has been well observed from both aircraft and ground-based remote sensors, could be benchmark for model simulations of mixed-phase clouds.

54 ENVIRONMENTAL SCIENCES↗

Development and Validation of SAM Multi-dimensional Flow Model for Thermal Mixing and Stratification Modeling

Thermal mixing and stratification in large pools or enclosures are very important phenomena that are critical to nuclear reactor safety. Because of the wide ranges of time and length scales associated with such phenomena, accurate modeling and simulation of thermal mixing and stratification remain as the key unresolved, challenging problems for reactor transient analyses. In traditional system analysis codes, simplified zero-dimensional (0-D) models are widely used for their high numerical efficiency, but they generally suffer from very limited prediction accuracies or range of applicability. Like traditional system analysis codes, the current version of the SAM code has implemented such simplified 0-D and one-dimensional mixing models. On the other hand, high-resolution Computational Fluid Dynamics (CFD) tools are often used to model complex thermal mixing and stratification phenomena. They are, however, generally numerically expensive, and they require large amounts of computational resources. It is therefore desirable to implement advanced and efficient thermal mixing and stratification modeling capabilities embedded in a system analysis code. This approach will improve the accuracy of reactor safety analyses when thermal mixing and stratification are involved, and also avoid using the large computational resources needed for high-resolution CFD analysis. Currently, with the support of the U.S. DOE Office of Nuclear Energy’s Nuclear Energy Advanced Modeling and Simulation program, an effort has been launched to develop and implement a multi-dimensional flow model in the system analysis code SAM, and demonstrate its applications to model thermal mixing and stratification phenomena in large enclosures. The main outcomes of this research and development activity are summarized in this report, which presents an attempt to include a built-in advanced multi-dimensional flow model in a system analysis code with the focus on overcoming the simulation challenges of thermal mixing and stratification phenomena. In this report, we start with the introduction of existing SAM code capabilities to simulate thermal mixing and stratification phenomena, which is followed by a short summary of the multi-dimensional model implemented in the SAM code, including both the physical model and the Finite Element Method code implementation. In this study, two options were implemented in the SAM code to model turbulent flows: a relatively simple built-in turbulence model and an interface to accept externally computed turbulent viscosities (e.g., from a high- resolution CFD simulation). Code validation studies on this newly added capability were then carried out to compare SAM simulation results with experimental data from the SUPERCAVNA facility, which was designed to study the complex flow recirculation and thermal stratification phenomena relevant to sodium fast reactor designs. In this study, one transient and two steady-state test cases were used for code validation. Different approaches have been used to model the complex turbulence flow fields in the SUPERCAVNA facility. A highly simplified zero-equation turbulence model was first used, but it was determined that it is too simple to capture the complex turbulence flow fields in these test cases. Subsequently, the code validation continued with the use of turbulent viscosity data from high-resolution STAR-CCM+ CFD simulations to improve the accuracy of the results. Using this approach, the SAM simulation results showed very good agreement with both the SUPERCAVNA experimental data and STAR-CCM+ simulation results. In this report, we demonstrate the development, implementation, and successful validation of a multi-dimensional flow model in the SAM code, which aims to improve the simulation accuracy for complex thermal mixing and stratification phenomena. Lessons have also been learned, including that in cases where the flow fields are not well predicted by the zero-equation model, the iinput of turbulent viscosities from an external source can enhance the overall predictive capabilities needed to accurately capture complex thermal-hydraulic phenomena. Therefore, future research will be needed to further improve the code’s capabilities, such as by developing a more efficient and robust approach to capture the turbulence effects in the SAM code.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Renewable Energy and Efficiency Modeling Analysis Partnership: An Analysis of How Different Energy Models Addressed a Common High Renewable Energy Penetration Scenario in 2025

The Renewable Energy and Efficiency Modeling and Analysis Partnership (REMAP) sponsors ongoing workshops to discuss individual 'renewable' technologies, energy/economic modeling, and - to some extent - policy issues related to renewable energy. Since 2002, the group has organized seven workshops, each focusing on a different renewable technology (geothermal, solar, wind, etc.). These workshops originated and continue to be run under an informal partnership of the Environmental Protection Agency (EPA), the Department of Energy's (DOE) Office of Energy Efficiency and Renewable Energy (EERE), the National Renewable Energy Laboratory (NREL), and the American Council on Renewable Energy (ACORE). EPA originally funded the activities, but support is now shared between EPA and EERE. REMAP has a wide range of participating analysts and models/modelers that come from government, the private sector, and academia. Modelers include staff from the Energy Information Administration (EIA), the American Council for an Energy-Efficient Economy (ACEEE), NREL, EPA, Resources for the Future (RFF), Argonne National Laboratory (ANL), Northeast States for Coordinated Air Use Management (NESCAUM), Regional Economic Models Inc. (REMI), ICF International, OnLocation Inc., and Boston University. The working group has more than 40 members, which also includes representatives from DOE, Lawrence Berkeley National Laboratory (LBNL), Union of Concerned Scientists (UCS), Massachusetts Renewable Energy Trust, Federal Energy Regulatory Commission (FERC), and ACORE. This report summarizes the activities and findings of the REMAP activity that started in late 2006 with a kickoff meeting, and concluded in mid-2008 with presentations of final results. As the project evolved, the group compared results across models and across technologies rather than just examining a specific technology or activity. The overall goal was to better understand how and why different energy models give similar and/or different answers in response to a set of focused energy-related questions. The focus was on understanding reasons for model differences, not on policy implications, even though a policy of high renewable penetration was used for the analysis. A group process was used to identify the potential question (or questions) to be addressed through the project. In late 2006, increasing renewable energy penetration in the electricity sector was chosen from among several options as the general policy to model. From this framework, the analysts chose a renewable portfolio standard (RPS) as the way to implement the required renewable energy market penetration in the models. An RPS was chosen because it was (i) of interest and represented the group's consensus choice, and (ii) tractable and not too burdensome for the modelers. Because the modelers and analysts were largely using their own resources, it was important to consider the degree of effort required. In fact, several of the modelers who started this process had to discontinue participation because of other demands on their time. Federal and state RPS policy is an area of active political interest and debate. Recognizing this, participants used this exercise to gain insight into energy model structure and performance. The results are not intended to provide any particular insight into policy design or be used for policy advocacy, and participants are not expected to form a policy stance based on the outcomes of the modeling. The goals of this REMAP project - in terms of the main topic of renewable penetration - were to: (1) Compare models and understand why they may give different results to the same question, (2) Improve the rigor and consistency of assumptions used across models, and (3) Evaluate the ability of models to measure the impacts of high renewable-penetration scenarios.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

Emulating a System Dynamics Model with Agent-Based Models: A Methodological Case Study in Simulation of Diabetes Progression

An agent-based simulation model hierarchy emulating disease states and behaviors critical to progression of diabetes type 2 was designed and implemented in the DEVS framework. The models are translations of basic elements of an established system dynamics model of diabetes. In this model hierarchy, which mimics diabetes progression over an aggregated U.S. population, was dis-aggregated and reconstructed bottom-up at the individual (agent) level. Four levels of model complexity were defined in order to systematically evaluate which parameters are needed to mimic outputs of the system dynamics model. Moreover, the four estimated models attempted to replicate stock counts representing disease states in the system dynamics model, while estimating impacts of an elderliness factor, obesity factor and health-related behavioral parameters. Health-related behavior was modeled as a simple realization of the Theory of Planned Behavior, a joint function of individual attitude and diffusion of social norms that spread over each agent s social network. Although the most complex agent-based simulation model contained 31 adjustable parameters, all models were considerably less complex than the system dynamics model which required numerous time series inputs to make its predictions. In all three elaborations of the baseline model provided significantly improved fits to the output of the system dynamics model. The performances of the baseline agent-based model and its extensions illustrate a promising approach to translate complex system dynamics models into agent-based model alternatives that are both conceptually simpler and capable of capturing main effects of complex local agent-agent interactions.

59 BASIC BIOLOGICAL SCIENCES↗

On the modeling of a single-stage, entrained-flow gasifier using Aspen Custom Modeler (ACM)

Coal-fired gasifiers are the centerpiece of integrated gasification combined cycle (IGCC) power plants. The gasifier produces synthesis gas that is subsequently converted into electricity through combustion in a gas turbine. Several mathematical models have been developed to study the physical and chemical processes taking place inside the gasifier. Such models range from simple one-dimensional (1D) steady-state models to sophisticated dynamic 3D computational fluid dynamics (CFD) models that incorporate turbulence effects in the reactor. The practical operation of the gasifier is dynamic in nature but most 1D and some higher-dimensional models are often steady state. On the other hand, many higher order CFD-based models are dynamic in nature, but are too computationally expensive to be used directly in operability and controllability dynamic studies. They are also difficult to incorporate in the framework of process simulation software such as Aspen Plus Dynamics. Thus lower-dimensional dynamic models are still useful in these types of studies. In the current study, a 1D dynamic model for a single-stage, downward-firing, entrained-flow GE-type gasifier is developed using Aspen Custom Modeler{reg_sign} (ACM), which is a commercial equation-based simulator for creating, editing, and re-using models of process units. The gasifier model is based on mass, momentum, and energy balances for the solid and gas phases. The physical and chemical reactions considered in the model are drying, devolatilization/pyrolysis, gasification, combustion, and the homogeneous gas phase reactions. The dynamic gasifier model is being developed for use in a plant-wide dynamic model of an IGCC power plant. For dynamic simulation, the resulting highly nonlinear system of partial differential algebraic equations (PDAE) is solved in ACM using the well-known Method of Lines (MoL) approach. The MoL discretizes the space domain and leaves the time domain continuous, thereby converting the PDAE to a differential algebraic equation (DAE) system with respect to time. The DAE system is solved using a variable-step implicit Euler integrator. For steady-state simulations, the set of nonlinear algebraic equations are solved using a Newton-type method. In this presentation, preliminary results from the steady-state non-isothermal gasifier model will be reported. Comparisons of the results from the gasifier model to available pilot plant data, industrial gasifier data, and other published models will be made. Sensitivity studies will be presented for different types of coal and operating conditions.

97 MATHEMATICS AND COMPUTING↗

Integration of the Full Tokamak Reference Model with the Complex Model for ITER Neutronic Analysis

The ITER International Organization has developed a number of reference Monte Carlo N-Particle (MCNP) models including the tokamak machine C-model, the Tokamak Complex model, and the neutral beam injection (NBI) systems model. The Tokamak Complex model primarily describes building structures beyond the bioshield. Representation of the tokamak and its systems are not included in this model. The Oak Ridge National Laboratory Radiation Transport Group has conducted two ITER neutronic analysis model integrations: (1) integration of the tokamak C-model with the Tokamak Complex model for shutdown dose rate characterization in Port Cell 16 at level B1, and (2) integration of the NBI model with the Tokamak Complex model for estimating the spatial distribution of biological dose rate at levels L1, L2, and L3 of the Tokamak Complex. The integrated models were further extended to include models of system components that are essential to the neutronic analyses. Furthermore, this paper presents the approach and computer tools used to integrate existing reference models, describes the additional design details implemented in the integrated models, and provides representative neutronic calculations based on the extended models.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Implications of Model Structure and Detail for Utility Planning. Scenario Case Studies using the Resource Planning Model

We examine how model investment decisions change under different model configurations and assumptions related to renewable capacity credit, the inclusion or exclusion of operating reserves, dispatch period sampling, transmission power flow modeling, renewable spur line costs, and the ability of a planning region to import and export power. For all modeled scenarios, we find that under market conditions where new renewable deployment is predominantly driven by renewable portfolio standards, model representations of wind and solar capacity credit and interactions between balancing areas are most influential in avoiding model investments in excess thermal capacity. We also compare computation time between configurations to evaluate tradeoffs between computational burden and model accuracy. From this analysis, we find that certain advanced dispatch representations (e.g., DC optimal power flow) can have dramatic adverse effects on computation time but can be largely inconsequential to model investment outcomes, at least at the renewable penetration levels modeled. Finally, we find that certain underappreciated aspects of new capacity investment decisions and model representations thereof, such as spur lines for new renewable capacity, can influence model outcomes particularly in the renewable technology and location chosen by the model. Though this analysis is not comprehensive and results are specific to the model region, input assumptions, and optimization-modeling framework employed, the findings are intended to provide a guide for model improvement opportunities.

24 POWER TRANSMISSION AND DISTRIBUTION↗

An integrated two-dimensional modeling method for predicting ductility of thin-walled die cast magnesium

In this paper, a two-dimensional (2D) microstructure-based modeling method is developed in order to predict the ductility of a thin-walled high pressure die cast magnesium (Mg) by considering the three-dimensional (3D) thru-thickness pore distributions. For this purpose, a series of 3D synthetic microstructure-based finite element models and the corresponding 2D models are first generated with various pore volume fractions, pore size distributions and pore shapes. The input material properties for the 2D models are determined based on the 3D cubic model with a spherical pore and the generalized Neuber’s rule. Based on the resulting ductility of the 3D and 2D models, a 3D/2D ductility correlation curve is developed as a function of the characteristics/shape of the input fracture strain curve used in the 3D model. The validity of the ductility correlation curve is examined with casting samples with actual microstructures. Here, the results show that the suggested 2D modeling methodology can be used together with the ductility correlation curve in predicting the ductility of thin-walled Mg castings.

36 MATERIALS SCIENCE↗

Grain-Size Based Additivity Models for Scaling Multi-rate Uranyl Surface Complexation in Subsurface Sediments

This study statistically analyzed a grain-size based additivity model that has been proposed to scale reaction rates and parameters from laboratory to field. The additivity model assumed that reaction properties in a sediment including surface area, reactive site concentration, reaction rate, and extent can be predicted from field-scale grain size distribution by linearly adding reaction properties for individual grain size fractions. This study focused on the statistical analysis of the additivity model with respect to reaction rate constants using multi-rate uranyl (U(VI)) surface complexation reactions in a contaminated sediment as an example. Experimental data of rate-limited U(VI) desorption in a stirred flow-cell reactor were used to estimate the statistical properties of multi-rate parameters for individual grain size fractions. The statistical properties of the rate constants for the individual grain size fractions were then used to analyze the statistical properties of the additivity model to predict rate-limited U(VI) desorption in the composite sediment, and to evaluate the relative importance of individual grain size fractions to the overall U(VI) desorption. The result indicated that the additivity model provided a good prediction of the U(VI) desorption in the composite sediment. However, the rate constants were not directly scalable using the additivity model, and U(VI) desorption in individual grain size fractions have to be simulated in order to apply the additivity model. An approximate additivity model for directly scaling rate constants was subsequently proposed and evaluated. The result found that the approximate model provided a good prediction of the experimental results within statistical uncertainty. This study also found that a gravel size fraction (2-8mm), which is often ignored in modeling U(VI) sorption and desorption, is statistically significant to the U(VI) desorption in the sediment.

Zhang, Xiaoying↗

Merging a mechanistic enzymatic model of soil heterotrophic respiration into an ecosystem model in two AmeriFlux sites of northeastern USA

Heterotrophic respiration (Rh), microbial processing of soil organic matter to carbon dioxide (CO 2 ), is a major, yet highly uncertain, carbon (C) flux from terrestrial systems to the atmosphere. Temperature sensitivity of Rh is often represented with a simple Q 10 function in ecosystem models and earth system models (ESMs), sometimes accompanied by an empirical soil moisture modifier. More explicit representation of the effects of soil moisture, substrate supply, and their interactions with temperature has been proposed as a way to disentangle the confounding factors of apparent temperature sensitivity of Rh and improve the performance of ecosystem models and ESMs. The objective of this work was to insert into an ecosystem model a more mechanistic, but still parsimonious, model of environmental factors controlling Rh and evaluate the model performance in terms of soil and ecosystem respiration. The Dual Arrhenius and Michaelis-Menten (DAMM) model simulates Rh using Michaelis-Menten, Arrhenius, and diffusion functions. Soil moisture affects Rh and its apparent temperature sensitivity in DAMM by regulating the diffusion of oxygen, soluble C substrates, and extracellular enzymes to the enzymatic reaction site. Here, we merged the DAMM soil flux model with a parsimonious ecosystem flux model, FöBAAR (Forest Biomass, Assimilation, Allocation and Respiration). We used high-frequency soil flux data from automated soil chambers and landscape-scale ecosystem fluxes from eddy covariance towers at two AmeriFlux sites (Harvard Forest, MA and Howland Forest, ME) in the northeastern USA to estimate parameters, validate the merged model, and to quantify the uncertainties in a multiple constraints approach. The optimized DAMM-FöBAAR model better captured the seasonal and inter-annual dynamics of soil respiration (Soil R) compared to the FöBAAR-only model for the Harvard Forest, where higher frequency and duration of drying events significantly regulate substrate supply to heterotrophs. However, DAMM-FöBAAR showed improvement over FöBAAR-only at the boreal transition Howland Forest only in unusually dry years. The frequency of synoptic-scale dry periods is lower at Howland, resulting in only brief water limitation of Rh in some years. At both sites, the declining trend of soil R during drying events was captured by the DAMM-FöBAAR model; however, model performance was also contingent on site conditions, climate, and the temporal scale of interest. While the DAMM functions require a few more parameters than a simple Q10 function, we have demonstrated that they can be included in an ecosystem model and reduce the model-data mismatch. Moreover, the mechanistic structure of the soil moisture effects using DAMM functions should be more generalizable than the wide variety of empirical functions that are commonly used, and these DAMM functions could be readily incorporated into other ecosystem models and ESMs.

54 ENVIRONMENTAL SCIENCES↗

Subtropical marine low stratiform cloud deck spatial errors in the E3SMv1 Atmosphere Model

Marine low-level clouds continue to be poorly simulated in models despite many studies and field experiments devoted to their improvement. Here we focus on the spatial errors in the cloud decks in the Department of Energy Earth system model (E3SM) relative to the satellite climatology by calculating centroid distances, area ratios, and overlap ratios. Since model dynamics is better simulated than clouds, these errors are attributed primarily to the model physics. To gain additional insight, we performed a sensitivity run in which model winds were nudged to those of reanalysis. This results in a large change (but not necessarily an improvement) in the simulated cloud decks. These differences between simulations are mainly due to the interactions between model dynamics and physics. These results suggest that both model physics (widely recognized) and its interaction with dynamics (less recognized) are important to model improvement in simulating these low-level clouds.

Brunke, Michael A.↗

Current data are consistent with flat spatial hypersurfaces in the Λ CDM cosmological model but favor more lensing than the model predicts

Here, we study the performance of three pairs of tilted, and a pair of untilited, ΛCDM cosmological models, with three of these four pairs allowing for non-flat spatial hypersurfaces, against cosmic microwave background (CMB) temperature and polarization power spectrum data (P18), measurements of the Planck 2018 lensing potential power spectrum (lensing), and a large compilation of non-CMB data (non-CMB). For the eight models, we measure cosmological parameters and study whether or not pairs of the data sets (as well as subsets of them) are mutually consistent in these models. Half of these models allow the lensing consistency parameter A L , which re-scales the gravitational potential power spectrum, to be an additional free parameter to be determined from data, while the other three have A L = 1 which is the theoretically expected value. The pair of untilted non-flat ΛCDM models are incompatible with P18 data. The tilted spatially-flat models assume the usual primordial spatial inhomogeneity power spectrum that is a power law in wave number. The tilted non-flat models assume either the primordial power spectrum used in the Planck group anal yses [Planck P(q)], that has recently been numerically shown to be a good approximation to what is quantum-mechanically generated from a particular choice of closed inflation model initial conditions, or a recently computed power spectrum [new P(q)] that quantum-mechanically follows from a different set of non-flat inflation model initial conditions. In the tilted non-flat models with A L = 1 we find differences between P18 data and non-CMB data cosmological parameter constraints, which are large enough to rule out the Planck P(q) model at 3σ but not the new P(q) model. No significant differences are found when cosmological parameter constraints obtained with two different data sets are compared within the standard tilted flat ΛCDM model. While both P18 data and non-CMB data separately favor a closed geometry, with spatial curvature density parameter Ω k < 0, when P18+non-CMB data are jointly analyzed the evidence in favor of non-flat hypersurfaces subsides. Differences between P18 data and non-CMB data cosmological constraints subside when A L is allowed to vary. From the most restrictive P18+lensing+non-CMB data combination we get almost model-independent constraints on the cosmological parameters and find that the A L > 1 option is preferred over the Ω k < 0 one, with the A L parameter, for all models, being larger than unity by ~ 2.5σ. According to the deviance information criterion, in the P18+lensing+non-CMB analysis, the varying A L option is on the verge of being strongly favored over the A L = 1 one, which could indicate a problem for the standard tilted flat ΛCDM model. These data are consistent with flat spatial hypersurfaces but more and better data could improve the constraints on Ω k and might alter this conclusion. Error bars on some cosmological parameters are significantly reduced when non-CMB data are used jointly with P18+lensing data. For example, in the tilted flat ΛCDM model for P18+lensing+non-CMB data the Hubble constant H 0 = 68.09 ± 0.38 km s -1 Mpc -1 , which is consistent with that from a median statistics analysis of a large compilation of H 0 measurements as well as with a number of local measurements of the cosmological expansion rate. This H 0 error bar is 31% smaller than that from P18+lensing data alone.

79 ASTRONOMY AND ASTROPHYSICS↗

Many but not all deep neural network audio models capture brain responses and exhibit correspondence between model stages and brain regions

Models that predict brain responses to stimuli provide one measure of understanding of a sensory system and have many potential applications in science and engineering. Deep artificial neural networks have emerged as the leading such predictive models of the visual system but are less explored in audition. Prior work provided examples of audio-trained neural networks that produced good predictions of auditory cortical fMRI responses and exhibited correspondence between model stages and brain regions, but left it unclear whether these results generalize to other neural network models and, thus, how to further improve models in this domain. We evaluated model-brain correspondence for publicly available audio neural network models along with in-house models trained on 4 different tasks. Most tested models outpredicted standard spectromporal filter-bank models of auditory cortex and exhibited systematic model-brain correspondence: Middle stages best predicted primary auditory cortex, while deep stages best predicted non-primary cortex. However, some state-of-the-art models produced substantially worse brain predictions. Models trained to recognize speech in background noise produced better brain predictions than models trained to recognize speech in quiet, potentially because hearing in noise imposes constraints on biological auditory representations. The training task influenced the prediction quality for specific cortical tuning properties, with best overall predictions resulting from models trained on multiple tasks. The results generally support the promise of deep neural networks as models of audition, though they also indicate that current models do not explain auditory cortical responses in their entirety.

59 BASIC BIOLOGICAL SCIENCES↗

Development of Explainable, Knowledge-Guided AI Models to Enhance the E3SM Land Model Development and Uncertainty Quantification

Focal Area(s): (2)Predictive modeling using AI techniques and AI-derived model components; use of AI and other tools to design a prediction system comprising of a hierarchy of models. (3) Insight gleaned from complex data (both observed and simulated) using AI, big data analytics, and other advanced methods, including explainable AI and physics- or knowledge- guided AI. Science Challenge: The Energy Exascale Earth System Model (E3SM) is a fully coupled, state-of-the-science Earth system model that uses code optimized for DOE's advanced computers to address the most critical scientific questions facing our nation and society (Golaz et al., 2019). The E3SM Land model (ELM) is designed to understand how the changes in terrestrial land surfaces will interact with other Earth system components and has been used to understand hydrologic cycles, biogeophysics, and ecosystem dynamics. In spite of great successes, the ELM has several known issues that restrain rapid improvements. For example, the ELM uses equilibrium models to simulate dynamic land-climate interactions and it requires long model spin-up time to identify suitable initial conditions for transient simulations. The ELM lacks built-in uncertainty mechanisms that can improve the robustness of model predictions. The ELM is a holistic, deterministic model system with a rigid design, and in many situations, it is hard to modify the ELM system to incorporate new theory/hypothesis and new data across scales to address emerging science problems (such as predicting the impacts of water cycle extremes). In addition, The ELM is technically optimized for traditional CPU-centric computers and it cannot fully utilize the current and incoming leadership computers for model simulations and uncertainty quantification (UQ). The success of artificial intelligence (AI) has inspired scientists to use AI models to discover intrinsic features from simulation data (Chattopadhyay et al., 2020) and observational data (Reichstein et al., 2019) to gain further process understanding of Earth science problems. However, autonomous AI model training through deep learning usually requires a huge amount of annotated data. To overcome the limitations from the data and computing resources, knowledge-guided AI models are necessary where human-knowledge is ingested in model construction (Banino et al., 2018) and training process (Silver et al., 2016) for efficient learning. Herein, we present a new way that leverages the process understanding from the ELM to guide AI model development for the ELM enhancement and UQ. We hope this study can inspire further Earth and environmental system model developments and transformations.

54 ENVIRONMENTAL SCIENCES↗

Assessing Alignment Between a Top-Down LCA Model and a Bottom-Up Supply Chain Model for Data Gap Identification and Prioritization

This work presents a method for evaluating the alignment between the results of bottom-up and top-down models and using the evaluation to enhance the bottom-up model and underlying data set. Disaggregation is used to enhance the embodied energy results of top-down models, which are then compared to bottom-up model results. Using the disaggregated top-down results makes it possible to identify specific data gaps and deficiencies within the bottom-up model and its data set. The top-down results will also provide information around which data gaps in the bottom-up model are contributing the most error to the results; in this way, data gaps can be identified as either critical or non-critical, and the effort needed to acquire and add data can be prioritized. Although the proposed method could be applied to any physical quantity or environmental impact that can be quantified with both bottom-up models (typically in physical units) and top-down models (typically in monetary units), embodied energy is the focus of this work. The disaggregated energy baseline is calculated from the throughput vector x, which gives the total economic activity required from each sector to provide a final demand. Using the direct requirements matrix, each element in x is disaggregated into a vector of inputs to the corresponding sector. Energy data from the interventions matrix B is then used to convert the monetary flows of the disaggregated x into physical energy flows. For each energy flow, the origin sector, the destination sector, and the energy carrier type will all be known. These energetic flows are analogous but not identical to direct energy inputs that would appear in a unit process used in life cycle analysis and other bottom-up methods, and thus can be used to evaluate the alignment between the top-down results and the results of a bottom-up model. To demonstrate and further develop the proposed method, embodied energy for chemicals assessed in the 2015 U.S. DOE bandwidth study on energy use within the chemical sector will be calculated using a top-down and a bottom-up model. The Materials Flows through Industry (MFI) modeling tool developed at the National Renewable Energy Laboratory is used as the bottom-up model, and the U.S. EEIO model, recently published by the U.S. Environmental Protection Agency, is used as the top-down model. U.S. EEIO and MFI results will be evaluated to determine how well the model results align. Where the two models are not in good alignment, the disaggregation method described above will be applied to the top-down model to generate a more detailed set of results that will aid in identifying data gaps in the MFI database. Using only MFI, data gaps can be identified in terms of which sectors have few or no unit processes in the database, but no information is available regarding the impact of the missing data on MFI results. The disaggregated top-down results will provide this information on which gaps in the MFI data are critical and have a significant impact on the results, and which are less important.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

ABSTRACTION OF INFORMATION FROM 2- AND 3-DIMENSIONAL PORFLOW MODELS INTO A 1-D GOLDSIM MODEL - 11404

The Savannah River National Laboratory has developed a 'hybrid' approach to Performance Assessment modeling which has been used for a number of Performance Assessments. This hybrid approach uses a multi-dimensional modeling platform (PorFlow) to develop deterministic flow fields and perform contaminant transport. The GoldSim modeling platform is used to develop the Sensitivity and Uncertainty analyses. Because these codes are performing complementary tasks, it is incumbent upon them that for the deterministic cases they produce very similar results. This paper discusses two very different waste forms, one with no engineered barriers and one with engineered barriers, each of which present different challenges to the abstraction of data. The hybrid approach to Performance Assessment modeling used at the SRNL uses a 2-D unsaturated zone (UZ) and a 3-D saturated zone (SZ) model in the PorFlow modeling platform. The UZ model consists of the waste zone and the unsaturated zoned between the waste zone and the water table. The SZ model consists of source cells beneath the waste form to the points of interest. Both models contain 'buffer' cells so that modeling domain boundaries do not adversely affect the calculation. The information pipeline between the two models is the contaminant flux. The domain contaminant flux, typically in units of moles (or Curies) per year from the UZ model is used as a boundary condition for the source cells in the SZ. The GoldSim modeling component of the hybrid approach is an integrated UZ-SZ model. The model is a 1-D representation of the SZ, typically 1-D in the UZ, but as discussed below, depending on the waste form being analyzed may contain pseudo-2-D elements. A waste form at the Savannah River Site (SRS) which has no engineered barriers is commonly referred to as a slit trench. A slit trench, as its name implies, is an unlined trench, typically 6 m deep, 6 m wide, and 200 m long. Low level waste consisting of soil, debris, rubble, wood, etc. is disposed within the trench which is then covered with soil and a cap. The filled trench resembles the surrounding soil, albeit with a higher porosity. As a result, the flow field through the trench is essentially 1-dimensional. This dimensionality makes the abstraction of information from this waste form fairly simple. Engineered waste forms present challenges not seen in the slit trench because of their higher dimensionality. Flow fields must conform to the barriers and are therefore subject to changes in direction. This paper will examine one case and show how this multidimensional flow field can be abstracted into a 1-dimensional flow field while retaining characteristics important to the transport of radioactive contaminants. One complication not addressed by the two preceding examples is that of multiple sources. This presents quite a challenge to the benchmarking exercise, but a fairly robust method has been developed to deal with it. While the PorFlow analyses can treat all sources as independent in space, this is not possible with a 1-D model. This problem has been addressed by constructing multiple 1-D models of the waste forms and using plume overlaps at the assessment points to address the multiple sources which can contribute to a distinct assessment point.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Error modeling for surrogates of dynamical systems using machine learning: Machine-learning-based error model for surrogates of dynamical systems

A machine learning–based framework for modeling the error introduced by surrogate models of parameterized dynamical systems is proposed. The framework entails the use of high-dimensional regression techniques (eg, random forests, and LASSO) to map a large set of inexpensively computed “error indicators” (ie, features) produced by the surrogate model at a given time instance to a prediction of the surrogate-model error in a quantity of interest (QoI). This eliminates the need for the user to hand-select a small number of informative features. The methodology requires a training set of parameter instances at which the time-dependent surrogate-model error is computed by simulating both the high-fidelity and surrogate models. Using these training data, the method first determines regression-model locality (via classification or clustering) and subsequently constructs a “local” regression model to predict the time-instantaneous error within each identified region of feature space. We consider 2 uses for the resulting error model: (1) as a correction to the surrogate-model QoI prediction at each time instance and (2) as a way to statistically model arbitrary functions of the time-dependent surrogate-model error (eg, time-integrated errors). We then apply the proposed framework to model errors in reduced-order models of nonlinear oil-water subsurface flow simulations, with time-varying well-control (bottom-hole pressure) parameters. The reduced-order models used in this work entail application of trajectory piecewise linearization in conjunction with proper orthogonal decomposition. Moreover, when the first use of the method is considered, numerical experiments demonstrate consistent improvement in accuracy in the time-instantaneous QoI prediction relative to the original surrogate model, across a large number of test cases. When the second use is considered, results show that the proposed method provides accurate statistical predictions of the time- and well-averaged errors.

97 MATHEMATICS AND COMPUTING↗