Derivative-Based Global Sensitivity Analysis for Modelswith High-Dimensional Inputs and Functional Output
Explore the source record for details and available documents.
Cleaves, Helen Larrabee↗
SEARCH · Search NASA
Results for “Input Sensitivity”
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.
Exploring Sensitivity of Infrasound Signal Predictions to Atmospheric Inputs
Explore the source record for details and available documents.
Schaible, Loring Pratt [Sandia National Laboratori↗
Modeling Streamflow Sensitivity to Climate Warming and Surface Water inputs in a Montane Catchment
Explore the source record for details and available documents.
K E Hale↗
The role of ion and electron-scale turbulence in setting heat and particle transport in the DIII-D ITER baseline scenario
In this work, the heat and particle transport in a DIII-D ITER Baseline Scenario (IBS) discharge has been investigated using both linear and nonlinear gyrokinetic simulations performed with the CGYRO code [J. Candy et al: Journal Comp. Phys. (2016)].These simulations were used to investigate the role that ion-scale ($k_θp_s$ < 1:0) and electron-scale ($k_θp_s$ > 1:0) turbulence play in determining heat and particle transport in the core of conditions believed to directly extrapolate to ITER operation. This investigation spans over nearly half of the plasma minor radius, from ρ = 0:45 - 0:85. To probe the nature of the transport and turbulence in these conditions and to validate the model against experimental fluxes, scans of $a/L_{T_i}$, $a/L_{T_e}$, and $a/L_n$ were performed at all radial locations. Long wavelength turbulence is found to be dominated by ITG modes with strong non-adiabatic electron effects and appears unable to reproduce ion and electron heat fluxes and electron particle fluxes simultaneously at most radial positions when single parameter scans are performed. To investigate the potential role of short wavelength turbulence, a series of electron-scale simulations are performed that indicate that experimentally relevant levels of electron heat flux could arise at sub-ion scales. Quantitative comparison of the simulated fluxes obtained from ion and electron-scale simulations with experimental levels along with the sensitivity of simulations results to changes in inputs is presented. Through extensive sensitivity scans and comparison with multiple transport channels, this work demonstrates a clear need for self-consistent modification of multiple inputs and suggests multi-scale interactions play a significant role at many of the radial locations studied. The results of this analysis help shape our understanding of the model fidelity needed to predict turbulence and transport reactor-relevant conditions and have important implications for the prediction of future fusion devices.
70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
MCNP6 Status and plans for criticality safety applications - 14368
The first production version of MCNP6 was released in 2013, with an update in 2014. MCNP6 provides all of the standard methodology for criticality calculations available in previous versions and new capabilities for adjoint-based sensitivity profiles and perturbations. This paper reviews the current status of MCNP6, including new features, performance improvements, verification testing, and nuclear data libraries. For criticality safety support, a new statistical package called Whisper was developed and first applied in 2014 for validation and determining USLs. Whisper includes input files and sensitivity profiles for over 1000 ICSBEP benchmarks and will be distributed with MCNP6. For the long term, an initiative called MCNP 2020 is underway to improve performance, strengthen the code foundation, and prepare for future computer systems. (authors)
42 ENGINEERING↗
Patterning of somatosympathetic reflexes
In a previous study, we reported that vestibular nerve stimulation in the cat elicits a specific pattern of sympathetic nerve activation, such that responses are particularly large in the renal nerve. This patterning of vestibulosympathetic reflexes was the same in anesthetized and decerebrate preparations. In the present study, we report that inputs from skin and muscle also elicit a specific patterning of sympathetic outflow, which is distinct from that produced by vestibular stimulation. Renal, superior mesenteric, and lumbar colonic nerves respond most strongly to forelimb and hindlimb nerve stimulation (approximately 60% of maximal nerve activation), whereas external carotid and hypogastric nerves were least sensitive to these inputs (approximately 20% of maximal nerve activation). In contrast to vestibulosympathetic reflexes, the expression of responses to skin and muscle afferent activation differs in decerebrate and anesthetized animals. In baroreceptor-intact animals, somatosympathetic responses were strongly attenuated (to <20% of control in every nerve) by increasing blood pressure levels to >150 mmHg. These findings demonstrate that different types of somatic inputs elicit specific patterns of sympathetic nerve activation, presumably generated through distinct neural circuits.
Non-NASA Center↗
Identification of Bare-Airframe Dynamics from Closed-Loop Data Using Multisine Inputs and Frequency Responses
Amethod is presented for computing multiple-input multiple-output frequency responses of bare-airframe dynamics for systems excited using orthogonal phase-optimized multisines and including correlated data arising from control mixing or feedback control. The estimation was posed as the solution to an underdetermined system of linear equations, for which additional information was supplied using interpolation of the frequency responses. A simulation model of the NASA T-2 aircraft having two inputs and two outputs was used to investigate the method in the open-loop configuration and under closed-loop control. The method was also applied to flight test data from the X-56A aeroelastic demonstrator having five inputs and ten outputs and flying under closed-loop control with additional control allocation mixing. Results demonstrated that the proposed method accurately estimates the bare airframe frequency responses in the presence of correlated data from control mixing and feedback control. Results also agreed with estimates obtained using different methods that are less sensitive to correlated inputs.
Grauer, Jared A.↗
ILI Code
This R code consists of three scripts, described below. 1. ILI_LHC_sample.R The first script generate 10,000 samples from the total parameter space formed by the ranges of the following five parameters inputted into the ILI model: transmission rate (beta), per capita rate of progress from exposed to infectious state (gamma_1), per capita rate of progress through initial infectious state (gamma_2), per capita rate of progress through hospitalized state (gamma_3), and per capita rate of progress through non-hospitalized infectious state (gamma_4). The output is LHS_samples.csv, for input into the global sensitivity script, ILI_SEIR. 2. ILI_SEIR.R The second script solves a standard Susceptible-Exposed-Infected- Recovered (SEIR) model (a system of Ordinary Differential Equations (ODEs)), runs global sensitivity analyses, and generates plots showing the impact of the input variables on the response variables for each virus. 3. virus_sims.R The third script solves a standard Susceptible-Exposed-Infected- Recovered (SEIR) model (a system of Ordinary Differential Equations (ODEs)), aimed at modeling outbreaks for five common upper respiratory viruses within a contained population. The output of this code is an epidemic curve to show timing and magnitude of the epidemic peak for each virus. Note: for all three scripts, the parameter values must be hard-coded by the user or the output will not be accurate.
Spencer, Julie↗
Power Prediction of Airborne Wind Energy Systems Using Multivariate Machine Learning
Kites can be used to harvest wind energy at higher altitudes while using only a fraction of the material required for conventional wind turbines. In this work, we present the kite system of Kyushu University and demonstrate how experimental data can be used to train machine learning regression models. The system is designed for 7 kW traction power and comprises an inflatable wing with suspended kite control unit that is either tethered to a fixed ground anchor or to a towing vehicle to produce a controlled relative flow environment. A measurement unit was attached to the kite for data acquisition. To predict the generated tether force, we collected input–output samples from a set of well-designed experimental runs to act as our labeled training data in a supervised machine learning setting. We then identified a set of key input parameters which were found to be consistent with our sensitivity analysis using Pearson input–output correlation metrics. Finally, we designed and tested the accuracy of a neural network, among other multivariate regression models. The quality metrics of our models show great promise in accurately predicting the tether force for new input/feature combinations and potentially guide new designs for optimal power generation.
17 WIND ENERGY↗
(U) LNK3DNT Capability in SENSMG Extended
The SENSMG multigroup neutron sensitivity code can read materials and geometry from redoin and lnk3dnt files written by the PARTISN multigroup discrete ordinates code or the MCNP6 Monte Carlo code. This report describes the user interface for this capability and the code verification that was done. Reference 5 described an initial version of this capability. Two important extensions have been made: 1) Reaction-rate ratios are now allowed as responses with redoin/lnk3dnt inputs; and 2) Sensitivities of reaction-rate ratios to mass densities and interface locations are now calculated (for all problems, not just redoin/lnk3dnt problems).
73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
Approach for Input Uncertainty Propagation and Robust Design in CFD Using Sensitivity Derivatives
An implementation of the approximate statistical moment method for uncertainty propagation and robust optimization for quasi 3-D Euler CFD code is presented. Given uncertainties in statistically independent, random, normally distributed input variables, first- and second-order statistical moment procedures are performed to approximate the uncertainty in the CFD output. Efficient calculation of both first- and second-order sensitivity derivatives is required. In order to assess the validity of the approximations, these moments are compared with statistical moments generated through Monte Carlo simulations. The uncertainties in the CFD input variables are also incorporated into a robust optimization procedure. For this optimization, statistical moments involving first-order sensitivity derivatives appear in the objective function and system constraints. Second-order sensitivity derivatives are used in a gradient-based search to successfully execute a robust optimization. The approximate methods used throughout the analyses are found to be valid when considering robustness about input parameter mean values.
Putko, Michele M.↗
Plant roots stimulate the decomposition of complex, but not simple, soil carbon
Roots release carbon into soil and can alleviate energy limitation of microbial organic matter decomposition. We know little about the effects of roots on microbial decomposition of different organic matter substrates, despite the importance for soil carbon stocks and turnover. Through implementing root-microbe interactions, the Carbon, Organisms, Rhizosphere, and Protection in the Soil Environment (CORPSE) model was previously shown to represent dynamics of total soil carbon in temperate forest field experiments. However, the model permits alternative hypotheses concerning microbial-substrate affinity. We investigated how root inputs affect decomposition of soil organic carbon (SOC) with variable decomposability. We simulated SOC stocks in CORPSE and compared microbial degradation of two substrates types with varying root-microbe interactions under two alternative hypotheses that varied in microbial-substrate affinity. We compared our modeled hypotheses to a forest field experiment where we quantified decomposition of isotopically-labeled starch and leaf tissues in soils with manipulated root access to microbes. Here, we tested the hypothesis that decomposition of leaves would be more sensitive to root inputs than decomposition of starch, corresponding to the alternative model hypothesis. In the field study, leaf decomposition increased with root density while starch decomposition was unchanged by root density. Microbial biomass and enzyme activity consistently increased with root inputs in CORPSE and the field study. Our field experiment supported the CORPSE simulations with high microbial-substrate affinity. Roots stimulated microbial growth and enzyme production, which increased degradation of more complex substrates such as leaf tissues. Substrates that were easily decomposed, such as starch, may already be degrading at a maximum rate in the absence of rhizosphere influence because their decomposition rate was unchanged by root inputs. We found that the degree to which roots stimulate microbial decomposition depends on the substrate being decomposed, and that root-microbe interactions influenced SOC stocks in both our model and field experiment. Environmental changes that alter root-microbe interactions could, therefore, alter soil C stocks and biogeochemical cycling, and models of these interactions should incorporate differential influence of rhizosphere inputs on different substrates.
59 BASIC BIOLOGICAL SCIENCES↗
COSP for Windows: Strategies for Rapid Analyses of Cyclic Oxidation Behavior
COSP is a publicly available computer program that models the cyclic oxidation weight gain and spallation process. Inputs to the model include the selection of an oxidation growth law and a spalling geometry, plus oxide phase, growth rate, spall constant, and cycle duration parameters. Output includes weight change, the amounts of retained and spalled oxide, the total oxygen and metal consumed, and the terminal rates of weight loss and metal consumption. The present version is Windows based and can accordingly be operated conveniently while other applications remain open for importing experimental weight change data, storing model output data, or plotting model curves. Point-and-click operating features include multiple drop-down menus for input parameters, data importing, and quick, on-screen plots showing one selection of the six output parameters for up to 10 models. A run summary text lists various characteristic parameters that are helpful in describing cyclic behavior, such as the maximum weight change, the number of cycles to reach the maximum weight gain or zero weight change, the ratio of these, and the final rate of weight loss. The program includes save and print options as well as a help file. Families of model curves readily show the sensitivity to various input parameters. The cyclic behaviors of nickel aluminide (NiAl) and a complex superalloy are shown to be properly fitted by model curves. However, caution is always advised regarding the uniqueness claimed for any specific set of input parameters,
Smialek, James L.↗
Probabilistic Assessment and Uncertainty Analysis of CO2 Storage Capacity of the Morrow B Sandstone—Farnsworth Field Unit
This paper presents probabilistic methods to estimate the quantity of carbon dioxide (CO2) that can be stored in a mature oil reservoir and analyzes the uncertainties associated with the estimation. This work uses data from the Farnsworth Field Unit (FWU), Ochiltree County, Texas, which is currently undergoing a tertiary recovery process. The input parameters are determined from seismic, core, and fluid analyses. The results of the estimation of the CO2 storage capacity of the reservoir are presented with both expectation curve and log probability plot. The expectation curve provides a range of possible outcomes such as the P90, P50, and P10. The deterministic value is calculated as the statistical mean of the storage capacity. The coefficient of variation and the uncertainty index, P10/P90, is used to analyze the overall uncertainty of the estimations. A relative impact plot is developed to analyze the sensitivity of the input parameters towards the total uncertainty and compared with Monte Carlo. In comparison to the Monte Carlo method, the results are practically the same. The probabilistic technique presented in this paper can be applied in different geological settings as well as other engineering applications.
03 NATURAL GAS↗
A Case Study in the Application of TSUNAMI-3D - Part 1, Multigroup
The use of the sensitivity and uncertainty tool suite within SCALE, known as the TSUNAMI suite, has grown following its initial introduction more than a decade ago. Proper use of these tools provides a more complete understanding of system responses to perturbations, especially in the area of nuclear cross section uncertainties. Recent advances within TSUNAMI have added generalized perturbation theory to the k{sub eff} sensitivity capability that was included in TSUNAMI originally. The generation of sensitivity data is encouraged in the review process for new critical experiment requests sponsored by the US Department of Energy (DOE) Nuclear Criticality Safety Program (NCSP) to demonstrate that the proposed experiment will satisfy the identified data need. Many individuals participate in this process, but many analysts have not received proper training in the use of the TSUNAMI tools. The adjoint perturbation theory used to calculate sensitivities is not typically included in nuclear engineering curricula, leaving inexperienced engineers particularly at risk for incorrect use of the tools. The first step in the use of the TSUNAMI tools is the generation of a sensitivity data file (SDF), which contains all of the energy-dependent sensitivity data generated for a model. Most often, the SDF is generated from a three-dimensional (3D) model using either the KENO V.a or KENO-VI code within the TSUNAMI-3D sequence. The sensitivity data generated in TSUNAMI-3D must be accurate for correct quantification of system similarity or any other use of the sensitivity data. The sensitivities calculated by TSUNAMI can be affected by many of the input parameters, cross section processing models, and geometry modeling techniques used within the TSUNAMI-3D input. Fortunately, the total sensitivity of the most important nuclides can be confirmed simply by using direct perturbation (DP) calculations. Confirmation of the total sensitivity of these important nuclides provides confidence that the flux and importance function calculated within the TSUNAMI sequence will yield accurate sensitivities for all nuclides, reactions, and energies of interest. This process is essentially identical for the multigroup (MG) methods and the continuous-energy (CE) methods introduced in SCALE 6.2. This paper discusses the use of DP calculations and provides a case study to confirm that accurate sensitivities are generated by TSUNAMI-3D for a particular benchmark experiment. A companion paper provides a similar case study applying the CE TSUNAMI capability to the same critical experiment. (authors)
21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
A Case Study in the Application of TSUNAMI-3D - Part 2, Continuous Energy
The use of the sensitivity and uncertainty tool suite within SCALE, which is known as the TSUNAMI suite, has grown following its initial introduction more than a decade ago. Proper use of this tool provides a more complete understanding of system responses to perturbations, especially in the area of nuclear cross section uncertainties. Recent advances within TSUNAMI have added generalized perturbation theory to the k{sub eff} sensitivity capability that was included in TSUNAMI originally. The sensitivities can also be used to quantify the similarity of two systems. As an example, calculated sensitivities can be combined with cross section uncertainties to determine the similarity index ck. The generation of sensitivity data is encouraged in the review process for new critical experiment requests sponsored by the US Department of Energy (DOE) Nuclear Criticality Safety Program (NCSP) to demonstrate that the proposed experiment will satisfy the identified data need. Many analysts participate in this process, but many have not received proper training in the use of the TSUNAMI tools. The adjoint perturbation theory used to calculate sensitivities is not typically included in nuclear engineering curricula, leaving inexperienced engineers particularly at risk for incorrect use of the tools. The first step in the use of the TSUNAMI tools is the generation of a sensitivity data file (SDF), which contains all of the energy-dependent sensitivity data generated for a model. Most often, the SDF is generated from a three-dimensional (3D) model using either the KENO V.a or KENO-VI code within the TSUNAMI-3D sequence. The sensitivity data generated in TSUNAMI-3D must be accurate for correct quantification of system similarity or any other use of the sensitivity data. The sensitivities calculated by TSUNAMI can be affected by many of the input parameters, cross section processing models, and geometry modeling techniques used within the TSUNAMI-3D input. Fortunately, the total sensitivity of the most important nuclides can be confirmed simply by using direct perturbation (DP) calculations. Confirmation of the total sensitivity of these important nuclides provides confidence that the flux and importance function calculated within the TSUNAMI sequence will yield accurate sensitivities for all nuclides, reactions, and energies of interest. This process is essentially identical for multigroup (MG) and continuous-energy (CE) methods introduced in SCALE 6.2. This paper discusses the use of direct perturbation calculations and provides a case study in the use of DP calculations to confirm accurate sensitivities generated by the CE CLUTCH methods in TSUNAMI-3D for a particular benchmark experiment. A companion paper provides a similar case study applying the MG TSUNAMI capability to the same critical experiment. (authors)
21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
A Novel use of Direct Simulation Monte-Carlo to Model Dynamics of COVID-19 Pandemic Spread
In this report, we evaluate a novel method for modeling the spread of COVID-19 pandemic. In this new approach we leverage methods and algorithms developed for fully-kinetic plasma physics simulations using Particle-In-Cell (PIC) Direct Simulation Monte-Carlo (DSMC) models. This approach then leverages Sandia-unique simulation capabilities, and High-Performance Computer (HPC) resources and expertise in particle-particle interactions using stochastic processes. Our hypothesis is that this approach would provide a more efficient platform with assumptions based on physical data that would then enable the user to assess the impact of mitigation strategies and forecast different phases of infection. This work addresses key scientific questions related to the assumptions this new approach must make to model the interactions of people using algorithms typically used for modeling particle interactions in physics codes (kinetic plasma, gas dynamics). The model developed uses rational/physical inputs while also providing critical insight; the results could serve as inputs to, or alternatives for, existing models. The model work presented was developed over a four-week time frame, thus far showing promising results and many ways in which this model/approach could be improved. This work is aimed at providing a proof-of-concept for this new pandemic modeling approach, which could have an immediate impact on the COVID-19 pandemic modeling, while laying a basis to model future pandemic scenarios in a manner that is timely and efficient. Additionally, this new approach provides new visualization tools to help epidemiologists comprehend and articulate the spread of this and other pandemics as well as a more general tool to determine key parameters needed in order to better predict pandemic modeling in the future. In the report we describe our model for pandemic modeling, apply this model to COVID-19 data for New York City (NYC), assess model sensitivities to different inputs and parameters and , finally, propagate the model forward under different conditions to assess the effects of mitigation and associated timing. Finally, our approach will help understand the role of asymptomatic cases, and could be extended to elucidate the role of recovered individuals in the second round of the infection, which is currently being ignored.
59 BASIC BIOLOGICAL SCIENCES↗
On the sensitivity of a residual circulation model to differences in input temperature data
The residual mean circulation (RMC) formulation of zonally averaged transport in the middle atmosphere produces a circulation which depends on the distributions of net diabatic heating and temperature. Such circulations are from two temperature data sets, using the same radiative transfer code (Rosenfield et al. 1987). These circulations are then used to transport N2O in a photochemical model. The circulations and the resulting N2O distributions are notably different during the Northern Hemisphere winter, with that based on the NMC temperatures producing too much upward transport in the tropical stratosphere, as judged by comparison with the stratospheric and mesoscale sounder data. The experiment demonstrates that model calculations, in general, and perturbation assessments, in particular, are likely to be quite sensitive to the choice of input temperature data (where this is not computed self-consistently). It also reveals what appears to be a seasonally dependent bias in NMC zonally averaged temperatures with respect to those obtained from the LIMS instrument during 1978/1979.
Guthrie, Paul D.↗