Search NASA⌕ Search

SEARCH · Search NASA

Results for “BOX 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 1,153 records · Page 64
All sourcesNASA NTRSDOE OSTI
NASA NTRS1990

Estimation of area-averaged rainfall over tropical oceans from microwave radiometry - A single channel approach

This paper presents a new simple retrieval algorithm for estimating area-time averaged rain rates over tropical oceans by using single channel microwave measurements from satellites. The algorithm was tested by using the Nimbus-5 Electrically Scanning Microwave Radiometer and a simple microwave radiative transfer model to retrieve seasonal 5-deg x 5-deg area averaged rainrate over the tropical Atlantic and Pacific from December 1973 to November 1974. The brightness temperatures were collected and analyzed into histograms for each season and in each grid box from December 1973 to November 1974. The histograms suggest a normal distribution of background noise plus a skewed rain distribution at the higher brightness temperatures. By using a statistical estimation procedure based upon normally distributed background noise, the rain distribution was separated from the raw histogram. The radiative transfer model was applied to the rain-only distribution to retrieve area-time averaged rainrates throughout the tropics. Despite limitations of single channel information, the retrieved seasonal rain rates agree well in the open ocean with expectations based upon previous estimates of tropical rainfall over the oceans.

Shin, Kyung-Sup↗
DOE OSTI2013

Simulating higher-dimensional geometries in GADRAS using approximate one-dimensional solutions.

The Gamma Detector Response and Analysis Software (GADRAS) software package is capable of simulating the radiation transport physics for one-dimensional models. Spherical shells are naturally one-dimensional, and have been the focus of development and benchmarking. However, some objects are not spherical in shape, such as cylinders and boxes. These are not one-dimensional. Simulating the radiation transport in two or three dimensions is unattractive because of the extra computation time required. To maintain computational efficiency, higher-dimensional geometries require approximations to simulate them in one-dimension. This report summarizes the theory behind these approximations, tests the theory against other simulations, and compares the results to experimental data. Based on the results, it is recommended that GADRAS users always attempt to approximate reality using spherical shells. However, if fissile material is present, it is imperative that the shape of the one-dimensional model matches the fissile material, including the use of slab and cylinder geometry.

Thoreson, Gregory G.↗
DOE OSTI2024

Probabilistic neural networks for improved analyses with phenomenological R -matrix

Here we present a method for measurement analyses based on probabilistic deep neural networks that provide several advantages over conventional analyses with phenomenological models. These include predicting physical quantities directly from data, the rapid generation of statistically robust uncertainties, and the ability to bypass some parameters that may induce ambiguities and complications in data analysis. As deep learning methods make predictions through “black boxes,” the uncertainty quantification is typically challenging. We use a probabilistic framework that provides thorough uncertainty quantification and is straightforward to follow in practice. With the network architecture based on the Transformer, we demonstrate the current method for predicting nuclear resonance parameters from scattering data using the phenomenological R-matrix model.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2019

Systematic Process Intensification involving Zeotropic Distillation

Distillation is commonly used for separating homogeneous fluid mixtures. While it is based on a simple thermally driven vapor-liquid equilibrium phenomenon at the two-phase region, different variants of distillation systems exist. These include simple distillation, dividing wall columns, pressure swing distillation, thermally coupled distillation and reactive distillation. Complex separation may also involve multiple distillation in a sequence to achieve the desired products. These sequences can be constructed as direct sequence, indirect sequence or thermally coupled columns. Even when a sequence is fixed, each column in the sequence can have a plethora of design alternatives to consider. All these make the design and intensification of distillation systems a challenging task. The large number of plausible design alternatives increases the combinatorial complexity for screening. Although significant efforts have been made in the past, we still lack a method for systematic identification of novel pathways for process intensification of distillation system. This is due to the lack of a generalized representation of all plausible alternatives in a single modeling and optimization framework. In this work, we first present a general representation of thermally driven vapor-equilibrium phenomenon using abstract building blocks. Specifically, each tray in a distillation column is represented by a liquid block and a vapor block with a shared semi-restricted boundary representing the interface. The arrangement of these building blocks in a two-dimensional grid gives rise to all plausible variants of distillation systems. These systems range from simple distillation, dividing wall columns to task-integrated columns. This enables us to systematically identify both classic and out-of-the-box intensified designs at the column and flowsheet levels. The overall distillation-based separation process synthesis problem is formulated as a single mixed-integer nonlinear programming (MINLP) model. In conclusion, the proposed approach is applied to multi-component zeotropic systems to show the capability of identifying promising intensified alternatives.

42 ENGINEERING↗
DOE OSTI2013

KINETIC MODELING OF PARTICLE ACCELERATION IN A SOLAR NULL-POINT RECONNECTION REGION

The primary focus of this paper is on the particle acceleration mechanism in solar coronal three-dimensional reconnection null-point regions. Starting from a potential field extrapolation of a Solar and Heliospheric Observatory (SOHO) magnetogram taken on 2002 November 16, we first performed magnetohydrodynamics (MHD) simulations with horizontal motions observed by SOHO applied to the photospheric boundary of the computational box. After a build-up of electric current in the fan plane of the null point, a sub-section of the evolved MHD data was used as initial and boundary conditions for a kinetic particle-in-cell model of the plasma. We find that sub-relativistic electron acceleration is mainly driven by a systematic electric field in the current sheet. A non-thermal population of electrons with a power-law distribution in energy forms in the simulated pre-flare phase, featuring a power-law index of about -1.78. This work provides a first step toward bridging the gap between macroscopic scales on the order of hundreds of Mm and kinetic scales on the order of centimeter in the solar corona, and explains how to achieve such a cross-scale coupling by utilizing either physical modifications or (equivalent) modifications of the constants of nature. With their exceptionally high resolution-up to 135 billion particles and 3.5 billion grid cells of size 17.5 km-these simulations offer a new opportunity to study particle acceleration in solar-like settings.

79 ASTRONOMY AND ASTROPHYSICS↗
NASA NTRS2003

Analysis and Test Correlation of Proof of Concept Box for Blended Wing Body-Low Speed Vehicle

The Low Speed Vehicle (LSV) is a 14.2% scale remotely piloted vehicle of the revolutionary Blended Wing Body concept. The design of the LSV includes an all composite airframe. Due to internal manufacturing capability restrictions, room temperature layups were necessary. An extensive materials testing and manufacturing process development effort was underwent to establish a process that would achieve the high modulus/low weight properties required to meet the design requirements. The analysis process involved a loads development effort that incorporated aero loads to determine internal forces that could be applied to a traditional FEM of the vehicle and to conduct detailed component analyses. A new tool, Hypersizer, was added to the design process to address various composite failure modes and to optimize the skin panel thickness of the upper and lower skins for the vehicle. The analysis required an iterative approach as material properties were continually changing. As a part of the material characterization effort, test articles, including a proof of concept wing box and a full-scale wing, were fabricated. The proof of concept box was fabricated based on very preliminary material studies and tested in bending, torsion, and shear. The box was then tested to failure under shear. The proof of concept box was also analyzed using Nastran and Hypersizer. The results of both analyses were scaled to determine the predicted failure load. The test results were compared to both the Nastran and Hypersizer analytical predictions. The actual failure occurred at 899 lbs. The failure was predicted at 1167 lbs based on the Nastran analysis. The Hypersizer analysis predicted a lower failure load of 960 lbs. The Nastran analysis alone was not sufficient to predict the failure load because it does not identify local composite failure modes. This analysis has traditionally been done using closed form solutions. Although Hypersizer is typically used as an optimizer for the design process, the failure prediction was used to help gain acceptance and confidence in this new tool. The correlated models and process were to be used to analyze the full BWB-LSV airframe design. The analysis and correlation with test results of the proof of concept box is presented here, including the comparison of the Nastran and Hypersizer results.

Spellman, Regina L.↗
DOE OSTI2026

Gradient-informed Hamiltonian Monte Carlo for multicomponent CALPHAD model optimization and uncertainty quantification

CALPHAD model parameter optimization is inherently challenging due to non-smooth objective functions, high-dimensional parameter spaces, and the need for uncertainty quantification (UQ). Traditional weighted nonlinear least squares approaches are computationally efficient but local, whereas black-box global optimizers and ensemble Markov Chain Monte Carlo (MCMC) methods provide broader exploration at substantial computational cost. The objective of this work is to combine the global exploration capability of gradient-informed Hamiltonian Monte Carlo – specifically the No-U-Turn Sampler (NUTS) – with local deterministic refinement using BFGS to efficiently optimize multicomponent CALPHAD models with minimal manual intervention. Analytic gradients are computed via the Jansson derivative framework. The methodology is demonstrated on the Cr—Fe binary system and extended to the Cr—Fe—Ni ternary system with 32 degrees of freedom. For Cr—Fe, NUTS achieves comparable or superior optimality relative to ensemble MCMC while requiring over an order-of-magnitude fewer likelihood evaluations. Parameter uncertainties are quantified through NUTS sampling and propagated to thermodynamic observables using local expansion, demonstrating a novel modular approach that combines binary and ternary parameter subsets without requiring global relaxation. These results establish gradient-informed exploration as a scalable strategy for multicomponent CALPHAD optimization and provide a practical route towards efficient higher-order database development with quantified uncertainty.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Exploring Gradient-Based Multi-directional Controls in GANs

Generative Adversarial Networks (GANs) have been widely applied in modeling diverse image distributions. However, despite its impressive applications, the structure of the latent space in GANs largely remains as a black-box, leaving its controllable generation an open problem, especially when spurious correlations between different semantic attributes exist in the image distributions. To address this problem, previous methods typically learn linear directions or individual channels that control semantic attributes in the image space. However, they often suffer from imperfect disentanglement, or are unable to obtain multi-directional controls. Furthermore, in this work, in light of the above challenges, we propose a novel approach that discovers nonlinear controls, which enables multi-directional manipulation as well as effective disentanglement, based on gradient information in the learned GAN latent space. More specifically, we first learn interpolation directions by following the gradients from classification networks trained separately on the attributes, and then navigate the latent space by exclusively controlling channels activated for the target attribute in the learned directions. Empirically, with small training data, our approach is able to gain fine-grained controls over a diverse set of bi-directional and multi-directional attributes, and we showcase its ability to achieve disentanglement significantly better than state-of-the-art methods both qualitatively and quantitatively.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Uncertainty quantification for Multiphase-CFD simulations of bubbly flows: a machine learning-based Bayesian approach supported by high-resolution experiments

In this paper, we developed a machine learning-based Bayesian approach to inversely quantify and reduce the uncertainties of multiphase computational fluid dynamics (MCFD) simulations for bubbly flows. The proposed approach is supported by high-resolution two-phase flow measurements, including those by double-sensor conductivity probes, high-speed imaging, and particle image velocimetry. Local distributions of key physical quantities of interest (QoIs), including the void fraction and phasic velocities, are obtained to support the Bayesian inference. In the process, the epistemic uncertainties of the closure relations are inversely quantified while the aleatory uncertainties from stochastic fluctuations of the system are evaluated based on experimental uncertainty analysis. The combined uncertainties are then propagated through the MCFD solver to obtain uncertainties of the QoIs, based on which probability-boxes are constructed for validation. The proposed approach relies on three machine learning methods: feedforward neural networks and principal component analysis for surrogate modeling, and Gaussian processes for model form uncertainty modeling. The whole process is implemented within the framework of an open-source deep learning library PyTorch with graphics processing unit (GPU) acceleration, thus ensuring the efficiency of the computation. The results demonstrate that with the support of high-resolution data, the uncertainties of MCFD simulations can be significantly reduced. The proposed approach has the potential for other applications that involve numerical models with empirical parameters.

42 ENGINEERING↗
DOE OSTI2014

Poster — Thur Eve — 22: A water calorimeter for low-energy particle beams

In this work, the feasibility of absolute dose to water measurements in low-energy electron beams using a water calorimeter specifically developed for shallow measurements is established. The calorimeter design consists of a cylindrical glass vessel encased in a block of expanded polystyrene. The vessel has a front window thickness of 1.1 mm, a 4 cm radius, and is 2.5 cm in depth. The vessel-block assembly sits inside a thermally-insulated box and is air-cooled to an operating temperature of 4 °C. Radiation-induced thermal gradients were simulated in a geometric model of the calorimeter using a finite element analysis software package. 52 absorbed dose to water measurements were performed in a 6 and 8 MeV electron beam (z{sub max} of 1.32 and 1.76 cm, respectively) for 60 seconds at a repetition rate of 400 MU/min and an SSD of 120 cm. Within the vessel, the depth of measurement was set to 1.08 cm relative to the inner front window. The average measured dose to water was 59.6 ± 0.2 cGy/100 MU (6 MeV), and 63.7 ± 0.3 cGy/100 MU (8 MeV). The associated heat transfer corrections were determined to be 1.026 ± 0.003 and 1.017 ± 0.004 for the 6 and 8 MeV beams, respectively. The most significant source of uncertainty in this study was the repeatability (type A, 0.42%). It is expected that performing fewer consecutive measurements under higher dose rate conditions will improve the stability of the thermal background, thereby improving the repeatability and reducing the overall standard uncertainty.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2021

Protoplanetary Disk Rings as Sites for Planetesimal Formation

Axisymmetric dust rings are a ubiquitous feature of young protoplanetary disks. These rings are likely caused by pressure bumps in the gas profile; a small bump can induce a traffic-jam-like pattern in the dust density, while a large bump may halt radial dust drift entirely. The resulting increase in dust concentration may trigger planetesimal formation by the streaming instability (SI), as the SI itself requires some initial concentration of dust. Here we present the first 3D simulations of planetesimal formation in the presence of a pressure bump modeled specifically after those seen by Atacama Large Millimeter/submillimeter Array. We place a pressure bump at the center of a large 3D shearing box, along with an initial solid-to-gas ratio of Z = 0.01, and we include both particle back-reaction and particle self-gravity. We consider millimeter-sized and centimeter-sized particles separately. For simulations with centimeter-sized particles, we find that even a small pressure bump leads to the formation of planetesimals via the SI; a pressure bump does not need to fully halt radial particle drift for the SI to become efficient. Furthermore, pure gravitational collapse via concentration in pressure bumps (such as would occur at sufficiently high concentrations and without the SI) is not responsible for planetesimal formation. For millimeter-sized particles, we find tentative evidence that planetesimal formation does not occur. If this result is confirmed at higher resolution, it could put strong constraints on where planetesimals can form. Ultimately, our results show that for centimeter-sized particles planetesimal formation in pressure bumps is extremely robust.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
NASA NTRS1984

Finite-amplitude models of convection in the early mantle

Models of finite-amplitude, time-dependent mantle convection at Reynolds number ten million are calculated. The models are based on the Boussinesq approximation for convection at infinite Prandtl number and treat convection in a two-dimensional, Cartesian coordinate box with an aspect ratio of 1.4. The initial conditions consist of a static fluid with a purely conductive temperature profile, while the boundary conditions are taken to be free slip, with impenetrable insulating side walls. The initial conditions are found to result in an initial burst of convection that reverses the horizontally averaged temperature gradient, making the middle layers stable with respect to convection. With bottom heating and cooling of the top, two distinct layers of convective cells occur; with top cooling only, the upper layer alone convects.

Boss, A. P.↗
NASA NTRS1997

Demonstration of Oxygen and Carbon Monoxide Propellants for Mars Missions

Currently, proposed planetary exploration missions must be small, with low costs and a short development time. Relatively high-risk technologies are being accepted for such missions if they meet these guidelines. For a Mars sample-return mission, one of the higher risk technologies is the use of return propellants produced from indigenous materials such as the Martian atmosphere. This consists of 96 percent carbon dioxide, which can be processed into oxygen and carbon monoxide. This year, the NASA Lewis Research Center completed the experimental evaluation and subscale technology development of an oxygen/carbon monoxide propellant combination. Previous research included ignition characterization, combustion performance, and heat transfer characterization with gaseous propellants at room temperature. In this year s tests, we studied the ignition characteristics and combustion of oxygen and carbon monoxide at near liquid temperatures. The mixture ratio boundaries for oxygen and carbon monoxide were determined as a function of propellant temperature in a spark torch igniter. With both propellants at room temperature, the ignition range was between 0.50 and 1.44; and with both propellants chilled to near-liquid temperatures, it was between 2.4 and 3.1. Statistical analysis of the mean value of the ignition boundaries provided models that describe the combination of oxygen temperature, carbon monoxide temperature, and mixture ratio that resulted in ignition. This range is the larger boxed area shown in the figure. The smaller boxed area indicates the range at which there is a 90-percent confidence that ignition will occur. The relatively small range at only 90-percent confidence indicates that using the oxygen/carbon monoxide combination as its own ignition source may not be the best design for a remote engine operating on Mars. Tests also were performed in a simulated small rocket engine that used oxygen/hydrogen combustion gases as the ignition source for oxygen/carbon monoxide. In these experiments, the oxygen/carbon monoxide was successfully ignited in eight of eight tests at a mixture ratio of 0.52. In addition, the oxygen/carbon monoxide maintained steady combustion after the oxygen/hydrogen ignition source was removed, verifying that the oxygen/carbon monoxide rocket engine should continue to be included in mission plans as return propulsion from Mars.

Linne, Diane L.↗
DOE OSTI2025

Second-generation downscaled earth system model data using generative machine learning

The second-generation Sup3rCC dataset provides high-resolution meteorological data generated through the downscaling of multiple earth system models (ESMs) from the Coupled Model Intercomparison Project Phase 6 (CMIP6). This downscaling is performed through application of a generative machine learning approach called Super-Resolution for Renewable Resource Data (sup3r). This dataset builds on the first-generation Sup3rCC data by applying improved bias correction methods and adding downscaled precipitation to the output variables. As with the first Sup3rCC version, the data still include temperature, wind speed and direction at multiple heights, pressure, three components of downwelling solar radiation, and relative humidity—all at 4-kilometer (km) hourly resolution over the contiguous United States. This is a 25x spatial enhancement and 24x temporal enhancement of the source 100-km daily-average ESM data. This extension of the Sup3rCC dataset includes data from six ESMs from two shared socioeconomic pathways (SSPs) totaling 400 years of data with multiple future projections of changing meteorological conditions. The scenario selection was based on a structured evaluation of historical ESM skill and comprehensive representation of possible trajectories of future climate change in temperature, humidity, precipitation, solar irradiance, and near-surface wind speeds. The inclusion of multiple future projections is intended to enable users to assess key drivers of un 36 certainty and variability. All data are double-bias corrected, resulting in a product that can be used out-of-the-box for energy system analysis with minimal historical bias. The potential applications of Sup3rCC data extend to various topics in renewable energy resource assessment, energy systems modeling, and grid resilience studies. High-resolution future meteorological projections are critical for evaluating the effects of changing meteorological conditions on renewable energy generation, energy demand, and for optimizing energy storage and grid infrastructure. The 4-km hourly resolution of the downscaled data enables understanding of spatial and temporal variability at the scales necessary for energy system operational planning. In addition, the dataset can support risk assessments by providing detailed information on possible future extreme weather events and long-term meteorological variability at scales relevant to energy infrastructure. By offering an enhanced representation of possible future meteorological conditions, the second-generation Sup3rCC dataset enables more precise modeling of energy resilience and adaptation strategies in response to changing meteorological conditions.

24 POWER TRANSMISSION AND DISTRIBUTION↗
NASA NTRS1994

Real-time remote scientific model validation

This paper describes flight results from the use of a CLIPS-based validation facility to compare analyzed data from a space life sciences (SLS) experiment to an investigator's preflight model. The comparison, performed in real-time, either confirms or refutes the model and its predictions. This result then becomes the basis for continuing or modifying the investigator's experiment protocol. Typically, neither the astronaut crew in Spacelab nor the ground-based investigator team are able to react to their experiment data in real time. This facility, part of a larger science advisor system called Principal Investigator in a Box, was flown on the space shuttle in October, 1993. The software system aided the conduct of a human vestibular physiology experiment and was able to outperform humans in the tasks of data integrity assurance, data analysis, and scientific model validation. Of twelve preflight hypotheses associated with investigator's model, seven were confirmed and five were rejected or compromised.

Frainier, Richard↗
DOE OSTI2016

Institute for Sustained Performance, Energy, and Resilience (SuPER)

The University of Tennessee (UTK) and University of Texas at El Paso (UTEP) partnership supported the three main thrusts of the SUPER project---performance, energy, and resilience. The UTK-UTEP effort thus helped advance the main goal of SUPER, which was to ensure that DOE's computational scientists can successfully exploit the emerging generation of high performance computing (HPC) systems. This goal is being met by providing application scientists with strategies and tools to productively maximize performance, conserve energy, and attain resilience. The primary vehicle through which UTK provided performance measurement support to SUPER and the larger HPC community is the Performance Application Programming Interface (PAPI). PAPI is an ongoing project that provides a consistent interface and methodology for collecting hardware performance information from various hardware and software components, including most major CPUs, GPUs and accelerators, interconnects, I/O systems, and power interfaces, as well as virtual cloud environments. The PAPI software is widely used for performance modeling of scientific and engineering applications---for example, the HOMME (High Order Methods Modeling Environment) climate code, and the GAMESS and NWChem computational chemistry codes---on DOE supercomputers. PAPI is widely deployed as middleware for use by higher-level profiling, tracing, and sampling tools (e.g., CrayPat, HPCToolkit, Scalasca, Score-P, TAU, Vampir, PerfExpert), making it the de facto standard for hardware counter analysis. PAPI has established itself as fundamental software infrastructure in every application domain (spanning academia, government, and industry), where improving performance can be mission critical. Ultimately, as more application scientists migrate their applications to HPC platforms, they will benefit from the extended capabilities this grant brought to PAPI to analyze and optimize performance in these environments, whether they use PAPI directly, or via third-party performance tools. Capabilities added to PAPI through this grant include support for new architectures such as the lastest GPU and Xeon Phi accelerators, and advanced power measurement and management features. Another important topic for the UTK team was providing support for a rich ecosystem of different fault management strategies in the context of parallel computing. Our long term efforts have been oriented toward proposing flexible strategies and providing building boxes that application developers can use to build the most efficient fault management technique for their application. These efforts span across the entire software spectrum, from theoretical models of existing strategies to easily assess their performance, to algorithmic modifications to take advantage of specific mathematical properties for data redundancy and to extensions to widely used programming paradigms to empower the application developers to deal with all types of faults. We have also continued our tight collaborations with users to help them adopt these technologies to ensure their application always deliver meaningful scientific data. Large supercomputer systems are becoming more and more power and energy constrained, and future systems and applications running on them will need to be optimized to run under power caps and/or minimize energy consumption. The UTEP team contributed to the SUPER energy thrust by developing power modeling methodologies and investigating power management strategies. Scalability modeling results showed that some applications can scale better with respect to an increasing power budget than with respect to only the number of processors. Power management, in particular shifting power to processors on the critical path of an application execution, can reduce perturbation due to system noise and other sources of runtime variability, which are growing problems on large-scale power-constrained computer systems.

42 ENGINEERING↗
NASA NTRS2015

State-of-the-Art Instrumentation Package to Support Model Organism Research in Space

Hardware was developed for the Ames student Fruit-Fly Experiment (AFEx) to support fly growth and analysis during spaceflight. The hardware consists of a 1.5U vented aluminum box that houses an acrylic habitat, video camera, LED lighting, and environmental sensors. Power is provided via two USB connectors, one of which also supports data downlink. While the hardware was designed for use with fruit flies, it will house plants on an upcoming mission and could be adapted for use with other systems.

Fruit-Fly Experiment↗
DOE OSTI2026

Input specific neural networks

Neural networks have emerged as powerful tools for mapping between inputs and outputs. However, their black-box nature limits the ability to encode or impose specific structural relationships between inputs and outputs. Many scientific and engineering problems, such as constitutive modeling in solid mechanics, require networks that can enforce convexity, monotonicity, or other structural constraints to ensure physical consistency. Here, we introduce the Input Specific Neural Network (ISNN), a new architecture that enables multiple, distinct constraints to be imposed on different input subsets for scalar-valued outputs. This framework unifies convex, monotone–convex, monotone, and arbitrary mappings within a single network for the first time. Two ISNN architectures with analytical first- and second-order derivatives are developed. We demonstrate the performance on synthetic toy problems, inverse problems in isotropic hyperelasticity, and finite element simulations. ISNNs achieve improved extrapolation behavior, require fewer invariant inputs than standard input convex networks for polyconvex potentials, and enable significant computational savings via manual differentiation. We also show how ISNNs can be used to learn structural relationships between inputs and outputs via a binary gating mechanism. Particularly, ISNNs are employed to model a homogenized anisotropic free energy potential in a decoupled multiscale setting. The network learns whether or not the potential should be modeled as polyconvex and retains only the relevant layers while using the minimum number of inputs. ISNNs provide a flexible foundation for embedding structural priors into neural networks, enhancing both interpretability and stability. They are broadly applicable across computational mechanics and other scientific domains requiring constrained functional relationships.

Jadoon, Asghar A. [Univ. of Texas, Austin, TX (Uni↗