Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “entropy functional”

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 37 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2020

Conservative discontinuous Galerkin scheme of a gyro-averaged Dougherty collision operator

Here, a conservative discontinuous Galerkin scheme for a non-linear Dougherty collision operator in full-f long-wavelength gyrokinetics is presented. Analytically this model operator has the advective-diffusive form of Fokker-Planck operators, it has a non-decreasing entropy functional, and conserves particles, momentum and energy. Discretely these conservative properties are maintained exactly as well, independent of numerical resolution. In this work the phase space discretization is performed using a novel version of the discontinuous Galerkin scheme, carefully constructed using concepts of weak equality and recovery. Discrete time advancement is carried out with an explicit time-stepping algorithm, whose stability limits we explore. The formulation and implementation within the long-wavelength gyrokinetic solver of Gkeyll are validated with relaxation tests, collisional Landau-damping benchmarks and the study of 5D gyrokinetic turbulence on helical, open field lines.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2026

Working Fluid Characterization and Performance Assessment of Subcritical Organic Rankine Cycles Based on the Lee–Kesler Approach for Energy Recovery

Here, a generalized model using the Lee–Kesler approach based on the corresponding states principle is developed to assess the performance of subcritical Organic Rankine Cycles operating with different working fluids. Each fluid is characterized by five parameters: the acentric factor, critical temperature, critical pressure, molar mass, and the ideal-gas ratio of specific heats at the critical temperature. The model was developed using the compressibility factor modified version of the Benedict–Webb–Rubin equation proposed by Lee and Kesler and the enthalpy and entropy functions to calculate thermodynamic state properties. The model was validated by comparing the results calculated with the model and working fluid thermodynamic properties obtained with the CoolProp database. This comparison was conducted for 91 working fluids, obtaining a relative error below 5% for 88 out of the 91 fluids (∼97%). A generalized parametric study was conducted to determine the influence of the pinch point and each fluid parameter on the performance of Organic Rankine Cycle (ORC) systems. It was found that efficiency increases with critical temperature, ideal-gas ratio of specific heats at the critical temperature, and acentric factor, reaching up to 13%. The developed model enables the evaluation of ORC system performance for existing working fluids. It also allows the formulation and evaluation of new fluids to enhance the performance of the ORC while retrieving energy from any kind of source; and likewise, the methodology can be applied to other power generation cycles.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2023

An Analysis of FPGA LUT Bias and Entropy for Physical Unclonable Functions

Process variations within Field Programmable Gate Arrays (FPGAs) provide a rich source of entropy and are therefore well-suited for the implementation of Physical Unclonable Functions (PUFs). However, careful considerations must be given to the design of the PUF architecture as a means of avoiding undesirable localized bias effects that adversely impact randomness, an important statistical quality characteristic of a PUF. Here in this paper, we investigate a ring-oscillator (RO) PUF that leverages localized entropy from individual look-up table (LUT) primitives. A novel RO construction is presented that enables the individual paths through the LUT primitive to be measured and isolated at high precision, and an analysis is presented that demonstrates significant levels of localized design bias. The analysis demonstrates that delay-based PUFs that utilize LUTs as a source of entropy should avoid using FPGA primitives that are localized to specific regions of the FPGA, and instead, a more robust PUF architecture can be constructed by distributing path delay components over a wider region of the FPGA fabric. Compact RO PUF architectures that utilize multiple configurations within a small group of LUTs are particularly susceptible to these types of design-level bias effects. The analysis is carried out on data collected from a set of identically designed, hard macro instantiations of the RO implemented on 30 copies of a Zynq 7010 SoC.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Temperature-dependent solvent reorganization entropies, free energies, and transition dipole strengths for the photoexcitation of Reichardt’s dye B30

Absorption spectra of the solvatochromic dye 2,6-diphenyl-4-2,4,6-triphenyl-1-pyridinophenolate (B30) were measured in seven solvents of varying polarity over temperature ranging from each solvent’s freezing point to 300 K. The excitation energies and their variances allowed calculations of the solvent reorganization energies, reorganization free energies and reorganization entropies as functions of temperature. The entropies of solvent packing around the chromophore are found to make major contributions to the reorganization free energies. The variances of the excitation energies depend only weakly on temperature, in disagreement with an expression that is often used for solvent reorganization free energies. Polar solvents reduce the transition dipole strength of B30’s long-wavelength absorption band, probably because interactions with the solvent enhance the charge-transfer character of the transition. The dipole strength drops further at low temperatures.

Chemistry↗
DOE OSTI2020

Kinetic entropy-based measures of distribution function non-Maxwellianity: theory and simulations

We investigate kinetic entropy-based measures of the non-Maxwellianity of distribution functions in plasmas, i.e. entropy-based measures of the departure of a local distribution function from an associated Maxwellian distribution function with the same density, bulk flow and temperature as the local distribution. First, we consider a form previously employed by Kaufmann & Paterson (J. Geophys. Res., vol. 114, 2009, A00D04), assessing its properties and deriving equivalent forms. To provide a quantitative understanding of it, we derive analytical expressions for three common non-Maxwellian plasma distribution functions. We show that there are undesirable features of this non-Maxwellianity measure including that it can diverge in various physical limits and elucidate the reason for the divergence. We then introduce a new kinetic entropy-based non-Maxwellianity measure based on the velocity-space kinetic entropy density, which has a meaningful physical interpretation and does not diverge. We use collisionless particle-in-cell simulations of two-dimensional anti-parallel magnetic reconnection to assess the kinetic entropy-based non-Maxwellianity measures. We show that regions of non-zero non-Maxwellianity are linked to kinetic processes occurring during magnetic reconnection. We also show the simulated non-Maxwellianity agrees reasonably well with predictions for distributions resembling those calculated analytically. These results can be important for applications, as non-Maxwellianity can be used to identify regions of kinetic-scale physics or increased dissipation in plasmas.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2021

Train Like a (Var)Pro: Efficient Training of Neural Networks with Variable Projection

Deep neural networks (DNNs) have achieved state-of-the-art performance across a variety of traditional machine learning tasks, e.g., speech recognition, image classification, and segmentation. The ability of DNNs to efficiently approximate high-dimensional functions has also motivated their use in scientific applications, e.g., to solve partial differential equations and to generate surrogate models. In this paper, we consider the supervised training of DNNs, which arises in many of the above applications. We focus on the central problem of optimizing the weights of the given DNN such that it accurately approximates the relation between observed input and target data. Devising effective solvers for this optimization problem is notoriously challenging due to the large number of weights, nonconvexity, data sparsity, and nontrivial choice of hyperparameters. To solve the optimization problem more efficiently, we propose the use of variable projection (VarPro), a method originally designed for separable nonlinear least-squares problems. Our main contribution is the Gauss--Newton VarPro method (GNvpro) that extends the reach of the VarPro idea to nonquadratic objective functions, most notably cross-entropy loss functions arising in classification. These extensions make GNvpro applicable to all training problems that involve a DNN whose last layer is an affine mapping, which is common in many state-of-the-art architectures. In our four numerical experiments from surrogate modeling, segmentation, and classification, GNvpro solves the optimization problem more efficiently than commonly used stochastic gradient descent (SGD) schemes. Finally, GNvpro finds solutions that generalize well, and in all but one example better than well-tuned SGD methods, to unseen data points.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Understanding the heat generation mechanisms and the interplay between joule heat and entropy effects as a function of state of charge in lithium-ion batteries

The thermal performance of lithium-ion battery cells is critical for ensuring their safe and reliable operation across various applications. In this study, we employed an isothermal calorimetry method to investigate the heat generation of commercial 18650 lithium-ion battery fresh cells during charge and discharge at different current rates, ranging from 0.05C to 0.5C, and across various temperatures: 20 °C, 30 °C, 40 °C, and 50 °C. Our findings revealed a direct correlation between heat generation and current rates, indicating that higher current rates lead to increased heat generation within the cells. Conversely, we observed that heat generation remained relatively stable as the temperature rose, suggesting that temperature changes within this range may not significantly impact the heat generation of fresh cells during typical operations. Furthermore, our study explored irreversible heat generation, which depends on the applied current and overpotential, using the galvanostatic intermittent titration technique at 0.05C–0.5C and 30 °C. Additionally, electrochemical impedance spectroscopy was performed on the same cells during charge and discharge at 20 °C, 30 °C, and 40 °C to analyze cell impedance. Finally, our results indicated a consistent dependence of impedance on the state of charge and depth of discharge, with a significant increase in impedance observed at the end of the discharge process.

25 ENERGY STORAGE↗
DOE OSTI2023

Resolving Atomic-Scale Structure and Chemical Coordination in High-Entropy Alloy Electrocatalysts for Structure–Function Relationship Elucidation

The recent breakthrough in confining five or more atomic species in nanocatalysts, referred to as high-entropy alloy nanocatalysts (HEAs), has revealed the possibilities of multielemental interactions that can surpass the limitations of binary and ternary electrocatalysts. The wide range of potential surface configurations in HEAs, however, presents a significant challenge in resolving active structural motifs, preventing the establishment of structure-function relationships for rational catalyst design and optimization. Here, we present a methodology for creating sub-5 nm HEAs using an aqueous-based peptide-directed route. Using a combination of pair distribution function and X-ray absorption spectroscopy, HEA structure models are constructed from reverse Monte Carlo modeling of experimental data sets and showcase a clear peptide-induced influence on atomic-structure and chemical miscibility. Coordination analysis of our structure models facilitated the construction of structure-function correlations applied to electrochemical methanol oxidation reactions, revealing the complex interplay between multiple metals that leads to improved catalytic properties. Our results showcase a viable strategy for elucidating structure-function relationships in HEAs, prospectively providing a pathway for future materials design.

36 MATERIALS SCIENCE↗
DOE OSTI2021

Uncertainty Quantification and Error Propagation in the Enthalpy and Entropy of Surface Reactions Arising from a Single DFT Functional

Accounting for parametric uncertainty in models is essential for quantifying the models’ predictive ability. Recently, approaches have been introduced to estimate parametric uncertainty in kinetic models while accounting for correlations among energy parameters. However, correlations have been estimated indirectly and correlations in entropies have not been accounted for. For surface-catalyzed microkinetic models of >C2 (more than two carbon-containing) molecules, which consist of thousands of reaction steps and intermediate surface species, first-principles density functional theory (DFT) is costly, and thus, estimation of thermochemistry and reaction barriers requires surrogate methods of DFT, such as group additivity and Brønsted–Evans–Polanyi relationships, respectively. For such parametrization, model uncertainty is unclear. This work develops a framework to overcome these gaps using group additivity and a single DFT functional. We estimate correlations in parameters of kinetic models and quantify uncertainty for thermochemistry, reaction barriers, reaction paths, and ultimately reaction rates, accounting also for the contribution of entropic uncertainty. Furthermore, the approach is illustrated on propane combustion and ethane oxidative dehydrogenation reactions.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

Functional composites by programming entropy-driven nanosheet growth

Here, nanomaterials must be systematically designed to be technologically viable. Driven by optimizing intermolecular interactions, current designs are too rigid to plug in new chemical functionalities and cannot mitigate condition differences during integration. Despite extensive optimization of building blocks and treatments, accessing nanostructures with the required feature sizes and chemistries is difficult. Programming their growth across the nano-to-macro hierarchy also remains challenging, if not impossible. To address these limitations, we should shift to entropy-driven assemblies to gain design flexibility, as seen in high-entropy alloys, and program nanomaterial growth to kinetically match target feature sizes to the mobility of the system during processing. In tis report, following a micro-then-nano growth sequence in ternary composite blends composed of block-copolymer-based supramolecules, small molecules and nanoparticles, we successfully fabricate high-performance barrier materials composed of more than 200 stacked nanosheets (125 nm sheet thickness) with a defect density less than 0.056 µm -2 and about 98% efficiency in controlling the defect type. Contrary to common perception, polymer-chain entanglements are advantageous to realize long-range order, accelerate the fabrication process (<30 min) and satisfy specific requirements to advance multilayered film technology. This study showcases the feasibility, necessity and unlimited opportunities to transform laboratory nanoscience into nanotechnology through systems engineering of self-assembly.

42 ENGINEERING↗
DOE OSTI2023

Optimization of Dimensions of Smooth and Twisted-Tape-Inserted Tubes for Heat Transfer with NaCl/KCl/MgCl 2 Molten Salts by Principle of Entropy Generation Minimization

The entropy generation minimization principle is used as the criterion to optimize the flow and heat transfer of solar collectors and heat exchangers that use molten salts NaCl–KCl–MgCl 2 and KCl–MgCl 2 . The Gnielinski correlation for the Nusselt number versus Reynolds number, as well as the Moody friction factor given by Petukhov, was used for the calculation of the convective heat transfer coefficient and pressure loss due to friction in smooth tubes. For twisted-tap-inserted tube, equations of Nu and friction factor provided by Manglik and Bergles were used. The objective function, the entropy generation rate of the heat transfer system, was expressed as the function of Reynolds number, Prandtl number, heating flux, tube diameter, etc. As a result of the analysis, the optimum Reynolds number was determined and thereby to determine the optimum Nusselt number, convective heat transfer coefficient, friction factor, and tube diameter, which also allows the calculation of optimum flow velocity. The analysis was conducted in the fluid temperature range of 500–700 °C, which covers the operation temperature for supercritical CO 2 power cycles in concentrated solar power (CSP) system. Here, optimized results from the smooth tube and twisted-tap-inserted tube are compared, which is important to the design of solar receivers for CSP systems.

14 SOLAR ENERGY↗
DOE OSTI2024

Accurate and Data‐Efficient Micro X‐ray Diffraction Phase Identification Using Multitask Learning: Application to Hydrothermal Fluids

Traditional analysis of highly distorted micro X‐ray diffraction (μ‐XRD) patterns from hydrothermal fluid environments is a time‐consuming process, often requiring substantial data preprocessing and labeled experimental data. Herein, the potential of deep learning with a multitask learning (MTL) architecture to overcome these limitations is demonstrated. MTL models are trained to identify phase information in μ‐XRD patterns, minimizing the need for labeled experimental data and masking preprocessing steps. Notably, MTL models show superior accuracy compared to binary classification convolutional neural networks. Additionally, introducing a tailored cross‐entropy loss function improves MTL model performance. Most significantly, MTL models tuned to analyze raw and unmasked XRD patterns achieve close performance to models analyzing preprocessed data, with minimal accuracy differences. This work indicates that advanced deep learning architectures like MTL can automate arduous data handling tasks, streamline the analysis of distorted XRD patterns, and reduce the reliance on labor‐intensive experimental datasets.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

ResNet and CycleGAN for pulse shape discrimination of He-4 detector pulses: Recovering pulses conventional algorithms fail to label unanimously

Pulse shape discrimination (PSD) capable detectors, such as He-4, that respond to neutron and gamma-ray 7 interactions have a threshold deposited energy value below which n/γ discrimination vanishes when using 8 conventional PSD algorithms. Recent attempts in applying supervised learning based artificial neural 9 networks for PSD use the pulses in the separated regions to train the networks so they can be used to classify 10 another set of separated pulses. In doing so, pulses previously indistinguishable are not recovered for 11 classification, which would have increased the number of neutron and gamma-ray pulses that could be used 12 for further analysis. Assuming the reason why conventional PSD algorithms have unseparated regions is 13 because the parameter space of the algorithms fail to capture the intrinsic (but subtle) distinguishing 14 behavior of some of the neutron and gamma-ray pulses, a cycle-consistent generative adversarial network 15 (CycleGAN) was trained to amplify those differences and extract well separated neutron and gamma-ray 16 clusters. Results show that, once the network is trained with pulses from separated and unseparated regions, 17 it was able to transform the pulses in the unseparated region to improve the PSD. Subsequent n/γ 18 classification was performed using deep residual network (ResNet) that takes pulses with 512 data points 19 as an input. Two different ResNets were explored – simple ResNet and modified ResNet which takes 20 segmented pulse inputs in the first layer and the corresponding time axis values in the last hidden layer. 21 The later approach enables the network to extract time correlated pulse features to enhance its ability to 22 capture the pulse behaviors relevant for PSD. Although it achieves slightly lower accuracy, 99.41% versus 23 99.89%, based on simply counting the number of correct n/γ labels assigned, compared to the simple 24 ResNet, the modified ResNets architecture was able to decreases the cross-entropy loss function by half, 25 which implies that the correct n/γ labels assigned are less likely to be accidental. PSD parameter 26 distributions based on n/γ classification by ResNet before and after transforming unseparated pulses using 27 CycleGAN show that by enhancing the separation between neutrons and gamma-rays, the transformation 28 helps improve the performance of classifier networks that are trained using labeled dataset. The 29 enhancement of neutron and gamma-ray separation by the CycleGAN increased the PSD figure of merit 30 (FOM) by up to 70% in some regions. Here, the results show that, if a given detector achieves clear separation 31 between neutron and gamma-ray pulses in any energy region, such neural network approaches can help 32 lower the energy threshold for the separation and increasing the number of neutron and gamma-ray pulses 33 that can be used for further analysis.

4He↗
DOE OSTI2023

Robust second-order approximation of the compressible Euler equations with an arbitrary equation of state

Here, this paper is concerned with the approximation of the compressible Euler equations supplemented with an arbitrary or tabulated equation of state. The proposed approximation technique is robust, formally second-order accurate in space, invariant-domain preserving, and works for every equation of state, tabulated or analytic, provided the pressure is nonnegative. An entropy surrogate functional that grows across shocks is proposed. The numerical method is verified with novel analytical solutions and then validated with several computational benchmarks seen in the literature including problems with composite waves.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Data imbalance in drug response prediction: multi-objective optimization approach in deep learning setting

Abstract Drug response prediction (DRP) methods tackle the complex task of associating the effectiveness of small molecules with the specific genetic makeup of the patient. Anti-cancer DRP is a particularly challenging task requiring costly experiments as underlying pathogenic mechanisms are broad and associated with multiple genomic pathways. The scientific community has exerted significant efforts to generate public drug screening datasets, giving a path to various machine learning models that attempt to reason over complex data space of small compounds and biological characteristics of tumors. However, the data depth is still lacking compared to application domains like computer vision or natural language processing domains, limiting current learning capabilities. To combat this issue and improves the generalizability of the DRP models, we are exploring strategies that explicitly address the imbalance in the DRP datasets. We reframe the problem as a multi-objective optimization across multiple drugs to maximize deep learning model performance. We implement this approach by constructing Multi-Objective Optimization Regularized by Loss Entropy loss function and plugging it into a Deep Learning model. We demonstrate the utility of proposed drug discovery methods and make suggestions for further potential application of the work to achieve desirable outcomes in the healthcare field.

Biochemistry & Molecular Biology↗
DOE OSTI2024

Entropy-Infused Deep Learning Loss Function for Capturing Extreme Values in Wind Power Forecasting

Extreme scenarios in wind power generation occur with higher frequency and larger magnitude in the recent years due to the ever-increasing extreme meteorological factors. Accurate forecasting of the occurrence of extreme values in wind power generation is of great concern to ensure reliable power system operation. Recently, deep learning models have surged in popularity for wind power forecasting, with the mean squared error (MSE) loss function being commonly used. However, the MSE loss function, being sensitive to extreme values, disproportionately penalizes larger errors, cannot adequately capture the extreme values present in wind energy data, and novel loss functions have seldom been tailored for wind power forecasting. To this end, in this paper, we introduce a novel loss function specifically crafted to capture extreme values in wind power forecasting. The experimental results with four fundamental deep learning methods on open source wind power dataset validate that the new loss function is efficient and superior in all cases compared to MSE in capturing extreme values while maintaining forecasting performance.

17 WIND ENERGY↗
DOE OSTI2023

pnnl/mlxrd

A multi-task neural network for the simultaneous identification of multiple compounds from XRD data obtained in a synchrotron-based hydrothermal fluid system. To enhance the alignment between the training and test datasets, random noise is introduced into the synthetic training dataset, and minor labeled experimental data is incorporated. Additionally, we employ a weighted cross-entropy loss function to address data imbalances and ensure smoother model training.

Li, Ang↗