Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Forward Modeling”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 865 records · Page 48

The Structure of Pectin

At the beginning of this project we hypothesized that pectin, which is a major polysaccharide in primary plant cell walls, is composed of various distinct structural regions covalently linked together into a high molecular weight complex polymer. We also hypothesized that a considerable portion of xyloglucan, the major hemicellulose in most primary cell walls, is linked to the pectin. Our goal was to determine if these interconnections exist and to characterize the exact nature of the interactions. It seems imperative that we have a complete knowledge of the structure of pectin to be able to propose realistic models of cell walls. There is a lot of interest in the biosynthesis of pectin. I do not think it will be possible to completely understand the biosynthesis of pectin without knowing the structure of pectin and thus the sequence of reactions needed to put each sugar or ester in its correct position in the polymer. We made considerable progress in determining the detailed structure of pectin and within a year or so will be able to put forward a comprehensive model of it.

59 BASIC BIOLOGICAL SCIENCES↗

SU-C-207-04: Reconstruction Artifact Reduction in X-Ray Cone Beam CT Using a Treatment Couch Model

Purpose: to mitigate artifacts induced by the presence of the RT treatment couch in on-board CBCT and improve image quality Methods: a model of a Varian IGRT couch is constructed using a CBCT scan of the couch in air. The model is used to generate a set of forward projections (FP) of the treatment couch at specified gantry angles. The model couch forward projections are then used to process CBCT scan projections which contain the couch in addition to the scan object (Catphan phantom), in order to remove the attenuation component of the couch at any given gantry angle. Prior to pre-processing with the model FP, the Catphan projection data is normalized to an air scan with bowtie filter. The filtered Catphan projections are used to reconstruct the CBCT with an in-house FDK algorithm. The artifact reduction in the processed CBCT scan is assessed visually, and the image quality improvement is measured with the CNR over a few selected ROIs of the Catphan modules. Results: Sufficient match between the forward projected data and the x-ray projections is achieved to allow filtering in attenuation space. Visual improvement of the couch induced artifacts is achieved, with a moderate expense of CNR. Conclusion: Couch model-based correction of CBCT projection data has a potential for qualitative improvement of clinical CBCT scans, without requiring position specific correction data. The technique could be used to produce models of other artifact inducing devices, such as immobilization boards, and reduce their impact on patient CBCT images.

60 APPLIED LIFE SCIENCES↗

Conditional Pseudo-Reversible Normalizing Flow for Surrogate Modeling in Quantifying Uncertainty Propagation

We introduce a conditional pseudo-reversible normalizing flow (PR-NF) that directly learns conditional probability distributions from noisy physical models to efficiently quantify both forward and inverse uncertainty propagation. Traditional surrogate modeling approaches approximate only the deterministic component of physical models, requiring separate noise characterization and computationally expensive sampling methods for inverse problems. Here, in this work, we develop the conditional PR-NF model to directly learn and efficiently generate samples from the conditional probability density functions (PDFs). The training process utilizes dataset consisting of input-output pairs without requiring prior knowledge about the noise and the function. Once trained, our model efficiently generates samples from conditional PDFs for any input within the training domain. Moreover, the pseudo-reversibility feature allows for the use of fully connected neural network architectures, which simplifies the implementation and enables theoretical analysis. We provide a rigorous convergence analysis of the conditional PR-NF model, showing its ability to converge to the target conditional PDF using the Kullback−Leibler divergence. To demonstrate the effectiveness of our method, we apply it to several benchmark tests and a real-world geologic carbon storage problem.

97 MATHEMATICS AND COMPUTING↗

Post-Fukushima Research and Development Strategy for MELCOR

Numerous MELCOR modeling improvements and analyses have been performed in the time since the severe accidents at Fukushima Daiichi Nuclear Power Station that occurred in March 2011. This report briefly summarizes the related accident reconstruction and uncertainty analysis efforts. It further discusses a number of potential pursuits to further advance MELCOR modeling and analysis of the severe accidents at Fukushima Daiichi and severe accident modeling in general. Proposed paths forward include further enhancements to identified MELCOR models primarily impacting core degradation calculations, and continued application of uncertainty analysis methods to improve model performance and a develop deeper understanding of severe accident progression.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

On the feasibility of using physics-informed machine learning for underground reservoir pressure management

In this work, we evaluate the feasibility of using physics-informed machine learning (PIML) for underground energy-related pressure management. To this end, we develop a PIML framework to manage underground reservoir pressures by training neural networks to determine fluid extraction rates for dedicated extraction wells during fluid injection operations given a range of reservoir conditions (e.g., transmissivity and storativity). We implement an automatically-differentiable analytical physics model of fluid flow in porous media within the PIML framework as a proxy for more complicated models. This allows us to execute a sufficient number of training scenarios to fully evaluate the feasibility of using PIML to support pressure management activities. We quantify the number of physics-model parameters required for automatic differentiation to become more efficient than finite-difference gradient calculations. We use a simple scenario with a single injector, extractor, and critical location for our feasibility analysis. We evaluate the effect of the size of the training dataset (i.e., the number of reservoir condition samples) on the accuracy and efficiency of the PIML framework. For an equivalent number of model evaluations, the larger training dataset took less time to train and produced a neural network that was able to more accurately manage reservoir pressures. We also evaluate the effect of the training dataset batch size (i.e., number of reservoir condition samples used to update the neural network coefficients during training; i.e., how the training dataset is partitioned). While training ran faster with larger batch sizes, they produced neural networks that managed pressures less accurately. We demonstrate the approach on a more complex scenario involving 10 injectors, 10 extractors, and 4 critical locations (a relatively high well density of 20 wells/km2). We provide the number of forward and adjoint model evaluations required in each case as an indication of the feasibility of using PIML for pressure management when more complicated physics models with longer execution times are used.

54 ENVIRONMENTAL SCIENCES↗

Measuring the fourth-generation b{yields}s quadrangle at the LHC

We show that simultaneous precision measurements of the CP-violating phase in time-dependent B{sub s}{yields}J/{psi}{phi} study and the B{sub s}{yields}{mu}{sup +}{mu}{sup -} rate, together with measuring m{sub t'} by direct search at the LHC, would determine V{sub t's}*V{sub t'b} and therefore the b{yields}s quadrangle in the four-generation standard model. The forward-backward asymmetry in B{yields}K*l{sup +}l{sup -} provides further discrimination.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Field Demonstration of The Krauklis Seismic Wave In A Novel Mva Method For Geologic CO 2 Storage

The Energy & Environmental Research Center (EERC) conducted a feasibility study of a new, low-impact geophysics monitoring method designed to incrementally track the saturation front of CO 2 injected into a reservoir. The method leverages a new way of transmitting energy from the surface to the reservoir and employs a new subsurface signal called the Krauklis wave (K-wave) and other guided waves for seismic monitoring. The EERC partnered with Seismos, Inc., the company that holds the exclusive technology license for this method, to advance development of K-wave technology, and Denbury Onshore, LLC, to field-test the system. The system was field-tested at Denbury’s Bell Creek oil field, currently undergoing CO 2 -based enhanced oil recovery (EOR) in southeastern Montana. The K-wave system uses a wellhead-mounted source to propagate energy down the wellbore in the form of a tube wave and through the well perforations into the reservoir. Part of the energy transmitted through the perforations into the reservoir converts to slow moving K-waves and/or other guided waves that are expected to propagate through the formation to neighboring wells. The propagated signal would induce tube waves in the neighboring wells that return to the surface to be recorded with a wellhead-mounted receiver. The propagation time of the signal is directly affected by the presence of CO 2 between the wells, so the difference in transit time of the returning signal between pre- and postinjection data sets could be used to help define the lateral extent of a CO 2 saturation front. Objectives of the study were to demonstrate, validate, and evaluate the K-wave method’s ability to monitor the morphology of injected CO 2 and estimate saturation distribution. Results from the demonstration were to be compared to conventional 4-D surface seismic methods, dynamic reservoir simulations, and CO 2 breakthrough data to validate the technology and provide useful information that may result in improvements. Prior to mobilizing for the full-scale K-wave baseline survey, a field test was conducted to better understand wellhead hardware configurations and test the response from the Bell Creek reservoir using three wells. Useful operational information resulted but signal from the first-generation gas release/gas injection source was not recorded in neighboring wells. After the initial field test, numerical modeling of the K-wave method was initiated using parameters specific to the Bell Creek Field. Modeling results indicated that the fundamental concepts behind the K-wave method exhibited problems and revealed a significantly greater attenuation of the signal at the wellbore perforations than expected, totaling on the order of 10-8 from source well to receiver well over a 400-m well spacing. Modeling also showed that greater source frequency bandwidth was needed to produce guided waves within a 30-foot-thick reservoir layer such as that at Bell Creek and that reservoir layer seismic velocity would need to be slower than the velocities of the bounding layers in order to act as a wave guide opposite to what exists at Bell Creek. Modeling results prompted a second field test with a stronger, second-generation displacement source developed in the interim. Concurrent with the second field test, a 3-D seismic survey was acquired to serve as a baseline for project validation and to serve as a monitor data set for the portion of the project area where CO 2 injection had already commenced. The new source was tested using a four-well geometry and failed to produce a detectable signal, which appeared to confirm modeling results and indicated that the Bell Creek Field is not an ideal site for the K-wave test. It was concluded that a source with sufficient strength and bandwidth could not be built to overcome the factors discovered by the modeling within a reasonable amount of time. Several other sites undergoing CO 2 -based EOR were also evaluated; however, none of these sites was amenable because of geology, limitations of source and receiver hardware, or the advanced state of the CO 2 flood. In addition, two other configurations of the source and receiver layout were considered; however, these design modifications either changed the scope of the project or did not meet the goals of the Carbon Storage Program. In consultation with the U.S. Department of Energy (DOE) Project Manager and Technology Manager, a no-go decision was invoked to forego additional expenditures as the technology would not meet current project objectives. Project activities to date have moved the technology forward by developing modeling equations that can be applied elsewhere, advanced the state of wellhead sensors and sources, and spurred improvements to field operations. The 3-D seismic survey in the project area and the dynamic reservoir simulations done as part of this study have provided value for ongoing and future synergistic activities at Bell Creek related to DOE carbon storage goals.

54 ENVIRONMENTAL SCIENCES↗

SU-F-I-49: Vendor-Independent, Model-Based Iterative Reconstruction On a Rotating Grid with Coordinate-Descent Optimization for CT Imaging Investigations

Purpose: Iterative reconstruction methods show promise for improving image quality and lowering the dose in helical CT. We aim to develop a novel model-based reconstruction method that offers potential for dose reduction with reasonable computation speed and storage requirements for vendor-independent reconstruction from clinical data on a normal desktop computer. Methods: In 2012, Xu proposed reconstructing on rotating slices to exploit helical symmetry and reduce the storage requirements for the CT system matrix. Inspired by this concept, we have developed a novel reconstruction method incorporating the stored-system-matrix approach together with iterative coordinate-descent (ICD) optimization. A penalized-least-squares objective function with a quadratic penalty term is solved analytically voxel-by-voxel, sequentially iterating along the axial direction first, followed by the transaxial direction. 8 in-plane (transaxial) neighbors are used for the ICD algorithm. The forward problem is modeled via a unique approach that combines the principle of Joseph’s method with trilinear B-spline interpolation to enable accurate reconstruction with low storage requirements. Iterations are accelerated with multi-CPU OpenMP libraries. For preliminary evaluations, we reconstructed (1) a simulated 3D ellipse phantom and (2) an ACR accreditation phantom dataset exported from a clinical scanner (Definition AS, Siemens Healthcare). Image quality was evaluated in the resolution module. Results: Image quality was excellent for the ellipse phantom. For the ACR phantom, image quality was comparable to clinical reconstructions and reconstructions using open-source FreeCT-wFBP software. Also, we did not observe any deleterious impact associated with the utilization of rotating slices. The system matrix storage requirement was only 4.5GB, and reconstruction time was 50 seconds per iteration. Conclusion: Our reconstruction method shows potential for furthering research in low-dose helical CT, in particular as part of our ongoing development of an acquisition/reconstruction pipeline for generating images under a wide range of conditions. Our algorithm will be made available open-source as “FreeCT-ICD”. NIH U01 CA181156; Disclosures (McNitt-Gray): Institutional research agreement, Siemens Healthcare; Past recipient, research grant support, Siemens Healthcare; Consultant, Toshiba America Medical Systems; Consultant, Samsung Electronics.

60 APPLIED LIFE SCIENCES↗

Investigating the Dynamic Compaction Behavior of Brittle Powders

The compression of porous materials to full density is a complex physical and mechanical process. Modeling efforts to capture the compaction behavior of distended solids (e.g., porous materials or powders) have produced a variety of compaction models, each functionalized to address specific compaction mechanisms. Of the available models, those suited for high-strain-rate deformation associated with shock loading have generally been developed for ductile materials. Compaction models for brittle materials have only been developed for quasistatic loading conditions. No predictive models currently exist that successfully capture the physical processes and mechanics for shock compaction of brittle distended materials, such as sand and oxide powders. The primary hindrance in producing a dynamic compaction model for brittle distended materials is the lack of availability of high-fidelity, high-precision shock compaction data for validation. The proposed work is aimed at alleviating this issue by developing a high-precision testing method in combination with high-fidelity analysis methods to accurately and precisely determine the shocked state of an ideal test material, ceria powder, under uniaxial strain dynamic compaction. The measured shock compaction data from these brittle powder experiments is then used in initial computations to design a cylindrically converging shock compaction experiment within situ time-resolved shocked density measurements, employing proton radiography. The preliminary experimental and computational results obtained from these uniaxial strain dynamic compaction and cylindrically converging shock compaction experiments are presented. An experimental and computational path forwards to predictive model development is then proposed.

36 MATERIALS SCIENCE↗

In the Face of Cybersecurity: How the Common Information Model Can Be Used

Efforts are underway to combine smart grid information, devices, networking, and emergency response information to create messages that are not dependent on specific standards development organizations (SDOs). This supports a future-proof approach of allowing changes in the canonical data models (CDMs) going forward without having to perform forklift replacements of solutions that use the messages. This also allows end users (electric utilities) to upgrade individual components of a larger system while keeping the message payload definitions intact. The goal is to enable public and private information sharing securely in a standards-based approach that can be integrated into existing operations. We provide an example architecture that could benefit from this multi-SDO, secure message approach. This article also describes how to improve message security

capability maturity model↗

Plasma Surface Presented as Array of Virtual Ports

Calculations aimed at representing the plasma-vacuum boundary condition in real space are presented. Some ideas about how this representation can be used for power coupling analysis and antenna modeling are put forward. The relation between tangential electric and magnetic fields on the plasma surface is linear and customary expressed in the spectral domain for every partial wave via 2x2 plasma impedance matrix - Z{sup p}(k{sub y},K{sub z}). In this work, the tangential fields E-vector{sub t} and B-vector{sub t} on the plasma surface are projected on a suitably chosen set of basis functions. Then the linear operator relating the components E-vector{sub t} and B-vector{sub t} expressed in such a way is derived from the known matrix Z{sup p}(k{sub y},K{sub z}). It is shown that in this representation the plasma surface has properties similar to the array of ports of a microwave system and can be formally described by a corresponding impedance matrix in the way it is done in microwave analysis. The obtained representation can be exploited for deeper understanding of the RF power coupling to the plasmas, antenna modeling, measurements of plasma electromagnetic properties and for plasma modeling with suitably designed structures or metamaterials.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

The ARM Radar Network: At the Leading-edge of Cloud and Precipitation Observations

The ARM observatories include radars that span a range of frequencies and capabilities to characterize the three-dimensional structure and properties of atmospheric hydrometeors. Improving our ability to predict future weather and climate conditions is strongly linked to achieving significant advancements in our understanding of cloud and precipitation processes. Observations are critical to making these advancements because they both improve our understanding of these processes and provide constraints on numerical models. Historically, instruments for observing cloud properties have limited cloud-aerosol investigations to a small subset of cloud-process interactions. To address these challenges, the last decade has seen the U.S. DOE ARM facility significantly upgrade and expand its surveillance radar capabilities towards providing holistic and multi-scale observations of clouds and precipitation. These upgrades include radars that operate at four frequency bands covering a wide range of scattering regimes, improving upon the information contained in earlier ARM observations. The traditional ARM emphasis on the vertical column is maintained, providing more comprehensive, calibrated, and multiparametric measurements of clouds and precipitation. Furthermore, the ARM radar network now features multiple scanning dual-polarization Doppler radars to exploit polarimetric and multi-Doppler capabilities that provide a wealth of information on storm microphysics and dynamics under a wide range of conditions. While the diversity in wavelengths and detection capabilities are unprecedented, there is still considerable work ahead before the full potential of these radar advancements is realized. This includes synergy with other observations, improved forward and inverse modeling methods, and well-designed data-model integration methods. The overarching goal is to offer a comprehensive characterization of a complete volume of the cloudy atmosphere and to act as a natural laboratory for the study of cloud processes.

54 ENVIRONMENTAL SCIENCES↗

Accuracy of Some Approximate Gaussian Filters for the Navier--Stokes Equation in the Presence of Model Error

Bayesian state estimation of a dynamical system from a stream of noisy measurements is important in many geophysical and engineering applications where high dimensionality of the state space, sparse observations, and model error pose key challenges. In this work, three computationally feasible, approximate Gaussian data assimilation/filtering algorithms are considered in various regimes of turbulent 2D Navier--Stokes dynamics in the presence of model error. The first source of error arises from the necessary use of reduced models for the forward dynamics of the filters, while a particular type of representation error arises from the finite resolution of observations which mix up information about resolved and unresolved dynamics. Two stochastically parameterized filtering algorithms, referred to as cSPEKF and GCF, are compared with 3DVAR---a prototypical time-sequential algorithm known to be accurate for filtering dissipative systems for a suitably inflated “background” covariance. We provide the first evidence that the stochastically parameterized algorithms, which do not rely on detailed knowledge of the underlying dynamics and do not require covariance inflation, can compete with or outperform an optimally tuned 3DVAR algorithm, and they can overcome competing sources of error in a range of dynamical scenarios.

97 MATHEMATICS AND COMPUTING↗

Unraveling electronic absorption spectra using nuclear quantum effects: Photoactive yellow protein and green fluorescent protein chromophores in water

Many physical phenomena must be accounted for to accurately model solution-phase optical spectra line shapes, from the sampling of chromophore-solvent configurations to the electronic-vibrational transitions leading to vibronic fine structure. Here we thoroughly explore the role of nuclear quantum effects, direct and indirect solvent effects, and vibronic effects in the computation of the optical spectrum of the aqueously solvated anionic chromophores of green fluorescent protein (GFP) and photoactive yellow protein (PYP). By analyzing the chromophore and solvent configurations, the distributions of vertical excitation energies, the absorption spectra computed within the ensemble approach, and the absorption spectra computed within the ensemble plus zero-temperature Franck-Condon approach, we introduce how solvent, nuclear quantum effects, and vibronic transitions alter the optical absorption spectra. We discover that including nuclear quantum effects in the sampling of chromophore-solvent configurations using ab initio path integral molecular dynamics simulations leads to improved spectral shapes through three mechanisms. The three mechanisms that lead to line shape broadening and a better description of the high-energy tail are softening of heavy atom bonds in the chromophore that couple to the optically bright state, widening the distribution of vertical excitation energies from more diverse solvation environments, and redistributing spectral weight from the 0-0 vibronic transition to higher energy vibronic transitions when computing the Franck-Condon spectrum in a frozen solvent pocket. The absorption spectra computed using the combined ensemble plus zero-temperature Franck-Condon approach yield significant improvements in spectral shape and width compared to spectra computed with the ensemble approach. Using the combined approach with configurations sampled from path integral molecular dynamics trajectories presents a significant step forward in accurately modeling the absorption spectra of aqueously solvated chromophores.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

A Dynamic Photovoltaic Model Incorporating Capacitive and Reverse-Bias Characteristics

Photovoltaics (PVs) are typically modeled only for their forward-biased dc characteristics, as in the commonly used single-diode model. While this approach accurately models the I-V curve under steady forward bias, it lacks dynamic and reverse-bias characteristics. The dynamic characteristics, primarily parallel capacitance and series inductance, affect operation when a PV cell or string interacts with switching converters or experiences sudden transients. Reverse-bias characteristics are often ignored because PV devices are not intended to operate in the reverse-biased region. However, when partial shading occurs on a string of PVs, the shaded cell can become reverse biased and develop into a hot spot that permanently degrades the cell. To fully examine PV behavior under hot spots and various other faults, reverse-bias characteristics must also be modeled. This study develops a comprehensive mathematical PV model based on circuit components that accounts for forward bias, reverse bias, and dynamic characteristics. Using a series of three experimental tests on an unilluminated PV cell, all required model parameters are determined. The model is implemented in MATLAB Simulink and accurately models the measured data.

Kim, KA↗

Modelling of pathologies of the nervous system by the example of computational and electronic models of elementary nervous systems

The paper puts forward principles of action of devices operating similarly to the nervous system and the brain of biological systems. We propose an alternative method of studying diseases of the nervous system, which may significantly influence prevention, medical treatment, or at least retardation of development of these diseases. This alternative is to use computational and electronic models of the nervous system. Within this approach, we represent the brain in the form of a huge electrical circuit composed of active units, namely, neuron-like units and connections between them. As a result, we created computational and electronic models of elementary nervous systems, which are based on the principles of functioning of biological nervous systems that we have put forward. Our models demonstrate reactions to external stimuli and their change similarly to the behavior of simplest biological organisms. The models possess the ability of self-training and retraining in real time without human intervention and switching operation/training modes. In our models, training and memorization take place constantly under the influence of stimuli on the organism. Training is without any interruption and switching operation modes. Training and formation of new reflexes occur by means of formation of new connections between excited neurons, between which formation of connections is physically possible. Connections are formed without external influence. They are formed under the influence of local causes. Connections are formed between outputs and inputs of two neurons, when the difference between output and input potentials of excited neurons exceeds a value sufficient to form a new connection. On these grounds, we suggest that the proposed principles truly reflect mechanisms of functioning of biological nervous systems and the brain. In order to confirm the correspondence of the proposed principles to biological nature, we carry out experiments for the study of processes of formation of connections between neurons in simplest biological objects. Based on the correspondence of function of the created models to function of biological nervous systems we suggest the use of computational and electronic models of the brain for the study of its function under normal and pathological conditions, because operating principles of the models are built on principles imitating the function of biological nervous systems and the brain.

60 APPLIED LIFE SCIENCES↗

Grain boundary strain as a determinant of localized sink efficiency

The opportunity to achieve radiation tolerance in crystalline materials hinges on understanding the structure and response of grain boundary sinks to irradiation. A common descriptor of grain boundary efficiency as a defect sink is the denuded zone, which is a defect free zone adjacent to the grain boundary dictated by its ability to absorb radiation induced defects. This descriptor is often used at the mesoscale, which requires an averaging of absorption events. In this paper, we resolve the defect sink efficiency as a function of interfacial strain with respect to grain boundary character, and correlate high levels of grain boundary strain to an enhanced absorption efficiency. Here, we also introduce a key relationship between localized strain in proximity with the grain boundary sink and the variation absorption efficiency associated with these regions, revealing the pitfalls of averaging absorption events along a grain boundary, and presenting a path forward toward improved models for denuded zones and localized grain boundary absorption phenomena.

36 MATERIALS SCIENCE↗

The Coupled Adjoint-State Equation in forward and inverse linear elasticity: Incompressible plane stress

A persistent challenge present in inverse or parameter estimation problems with interior data is how to deal with uncertainty in the boundary conditions employed in the forward or state model. In this work we focus on a linear plane stress inverse elasticity problem with measured displacement data where one component of the measured displacement field is known with considerably greater precision than the other. This situation is commonly encountered when the displacement field is measured using ultrasound or optical coherence tomography. We present a novel computational formulation in which no displacement or traction boundary conditions are assumed. The formulation results in coupling the state and adjoint equations, that are typically uncoupled when a well-posed state model is available. Two variants of residual-based stabilization are added. Our approach is applied to a simulated data set and experimental data from an ultrasound phantom.

42 ENGINEERING↗