Search NASA⌕ Search

SEARCH · Search NASA

Results for “regularization”

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 667 records · Page 37
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Efficient calculation of self magnetic field, self-force, and self-inductance for electromagnetic coils with rectangular cross-section

Abstract For designing high-field electromagnets, the Lorentz force on coils needs to be computed in order to design suitable support structures, and the inductance should be computed to evaluate the stored energy and dynamics. Also, the magnetic field and its variation inside the conductor is of interest for computing stress and strain, and due to superconducting quench limits. For these force, inductance, energy, and internal field calculations, the coils cannot be naively approximated as infinitesimally thin filaments due to divergences when the source and evaluation points coincide, so more computationally demanding calculations are usually required, resolving the finite cross-section of the conductors. Here, we present a new alternative method that enables the internal magnetic field vector, self-force, and self-inductance to be computed rapidly and accurately within a 1D filament model. The method is applicable to coils for which the curve center-line can have general noncircular shape, as long as the conductor width is small compared to the radius of curvature. This paper extends a previous calculation for circular-cross-section conductors (Hurwitz et al 2024 IEEE Trans. Magn. ) to consider the case of rectangular cross-section. The reduced model is derived by rigorous analysis of the singularity, regularizing the filament integrals such that they match the true high-dimensional integrals at high coil aspect ratio. The new filament model exactly recovers analytic results for a circular coil, and is shown to accurately reproduce full finite-cross-section calculations for a non-planar coil of a stellarator magnetic fusion device. Due to the efficiency of the model here, it is well suited for use inside design optimization.

Landreman, Matt (ORCID:000000027233577X)↗
DOE OSTI2025

A framework for discrete optimization of stellarator coils

Designing magnets for three-dimensional plasma confinement is a key task for advancing the stellarator as a fusion reactor concept. Stellarator magnets must produce an accurate field while leaving adequate room for other components and being reasonably simple to construct and assemble. In this paper, a framework for coil design and optimization is introduced that enables the attainment of sparse magnet solutions with arbitrary restrictions on where coils may be located. The solution space is formulated as a 'wireframe' consisting of a mesh of interconnected wire segments enclosing the plasma. Two methods are developed for optimizing the current distribution on a wireframe: Regularized Constrained Least Squares, which uses a linear least-squares approach to optimize the currents in each segment, and Greedy Stellarator Coil Optimization, a fully discrete procedure in which loops of current are added to the mesh one by one to achieve the desired magnetic field on the plasma boundary. Examples are presented of solutions obtainable with each method, some of which achieve high field accuracy while obeying spatial constraints that permit easy assembly.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2025

Physics of beam-driven ion cyclotron emission in the large plasma device

Abstract Ion cyclotron emission (ICE) is widely observed from spatially localised minority energetic ion populations in toroidal magnetically confined fusion (MCF) plasmas, both tokamaks and stellarators. Its spectral structure is typically regular with narrow suprathermal peaks, whose frequency separation matches a local energetic ion cyclotron frequency. Here we report the first computational (fully nonlinear self-consistent kinetic particle-in-cell code) and analytical (linear magnetoacoustic cyclotron instability (MCI)) studies of ICE observations from cylindrical plasmas contained in the Large Plasma Device (LAPD). Because LAPD is cylindrical, the plasma physics giving rise to the observed ICE spectrum necessarily excludes toroidal effects. Our approach, previously successful for toroidal plasma ICE, assumes slab geometry and hence is well adapted to LAPD. ICE from LAPD is strongly electrostatic, as distinct from electromagnetic, and is driven by 15 keV beam ions for which the ratio of perpendicular speed to the local Alfven speed, v ⊥ / v A , is 0.15, lower than in MCF plasmas from which beam-driven ICE has previously been observed. Our results are in good agreement with these observations. There is congruence between simulated ICE spectra, obtained in the saturated nonlinear regime of our computations, and observed ICE spectra; the underlying physics is essentially the same as in toroidal plasmas; and there is alignment with linear analytical theory where appropriate. The present work establishes a mapping from the cylindrical LAPD ICE observations to toroidal MCF ICE observations. The LAPD spectra are instances of beam-driven MCI-type ICE in its sub-Alfvenic, predominantly electrostatic manifestation, which has precedents in MCF stretching back to the 1990s. An interesting corollary is that, for many purposes, ICE in toroidal MCF plasmas ‘might as well’ be occurring in a cylinder.

Samant, O. (ORCID:0000000226055363)↗
DOE OSTI2025

Metrics and extrapolation of resonant magnetic perturbation thresholds for ELM suppression

This large database study of resonant magnetic perturbation (RMP) edge localized mode (ELM) suppression thresholds in the AUG, DIII-D, EAST, and KSTAR tokamaks details the key strengths and weaknesses of RMP metrics. The RMP ELM suppression database used for this work contains plasma information at the time of transition from ELMing to ELM suppressed states where a clear experimental threshold is identified. The experimental threshold distributions are compared for five metrics: (1) the island overlap width, (2) pedestal top Chirikov overlap, (3) peeling edge displacement, (4) pedestal top resonant drive, and (5) edge dominant mode overlap. The distributions, the regularity of the dependence on RMP coil currents, and the sensitivities of a given metric to equilibrium reconstruction details are compared. The overlap metric proves to be a good compromise between including the appropriate plasma response physics and maintaining a numerical robustness. This quantity does not exhibit clear power-law scalings for projection, but machine learning can assist in predicting thresholds within the existing parameter ranges and providing uncertainty quantification of those predictions. Two new first-principles models, one utilizing a threshold from the non-linear Modified Rutherford equation evaluated at the pedestal top and one utilizing the SLAYER code to calculate the linear tearing threshold from torque balance, offer possible paths to extrapolation beyond the existing database parameter space.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2025

Neural network-based classification and regression of magnetohydrodynamic modes in tokamaks

We present a machine learning-based magnetohydrodynamic (MHD) classifier and regressor that utilizes real or complex-valued 3D magnetic sensor array data to determine neoclassical tearing mode (NTM) onset times in tokamaks with millisecond accuracy. The input dataset consists of poloidal profiles of complex Fourier amplitudes with an n = 1 toroidal mode number from 144 human-labeled ITER Baseline Scenario discharges in the DIII-D tokamak, spanning both tearing-dominated and sawtooth-dominated regimes. Since m, n = 2,1 NTMs frequently emerge alongside sawteeth at the same frequency in this scenario, the focus is on isolating the m = 1 and m = 2 components of the n = 1 MHD mode near the tearing onset. To improve model regularization and prediction stability, singular value decomposition was applied to balance the sawtooth and tearing datasets. The enriched datasets facilitated training neural networks that learn the key distinguishing features of sawtooth and tearing modes in the poloidal profiles of their magnetic amplitude and phase. When the modes occur independently, the networks achieve perfect classification due to the modes’ distinct characteristics and low measurement noise. In the more experimentally relevant case where both modes coexist, the networks maintain exceptional performance across key metrics. Tests on synthetic data with known ground truth demonstrate the superior accuracy of the neural network trained on complex-valued input compared to models using real amplitude, phase, or pseudo-complex data, achieving both a mean time delay and standard deviation below 1 ms. Notably, standard linear regression methods fitting the dominant singular modes to the data closely match the neural network’s performance. Applying these methods across a broad range of H-mode scenarios will enable future studies to systematically identify dominant NTM triggers as scenario-specific variables, paving the way for more effective tearing mode avoidance strategies in future fusion reactor designs.

machine learning↗
DOE OSTI2026

Overview of the MAST Upgrade physics programme: testing novel concepts at low aspect ratio to inform future devices

The research programme performed on the Mega Amp Spherical Tokamak (MAST) Upgrade experiment has made significant advances in developing the physics understanding of low aspect ratio tokamaks in support of the operation of ITER and design of fusion powerplants. High performance plasma scenarios have been developed to facilitate a broad programme of experiments, in which confinement is constrained by the presence of m/n = 2/1 modes that cause substantial losses of fast ions. The onset of these modes coincides with the q = 2 surface residing in a local minimum in the toroidal current density profile. The maximum electron temperature at the pedestal top, T e,ped is limited with gas fuelling to ∼350 eV to maintain regular ELMs; higher T e,ped results in a transition to a non-stationary ELM-free regime. The operational space of spherical tokamaks has been expanded into small and ELM-free regimes. Strong shaping of the last closed flux surface can induce a transition from large to small ELMs, and ELM suppression with resonant magnetic perturbations has been observed for the first time in a low aspect ratio tokamak. Negative triangularity shaping has induced a transition from ELMy H-mode to a high-performance L-mode regime for the first time in a low aspect ratio tokamak. In studies of fast ion confinement, losses of fast particles due to Global Alfvén Eigenmodes have been identified. Interactions between fast ions generated by off-axis neutral beam injection and thermal neutrals can result in significant losses of fast ions. Experiments with on- and off-axis neutral beam injection exhibit a flux pumping mechanism, where the central safety factor is held to ∼1 in the absence of sawteeth. In studies of pedestal physics, it has been found that elevated main chamber neutral pressures result in an increase in the electron density and reduction in the temperature at the pedestal top. Advances in understanding plasma exhaust include the integration of a high-performance plasma core with detached outer divertors in the X-point target configuration. A newly commissioned lower divertor cryopump reduces the lower divertor neutral pressure by up to 50%, with minimal effect on the main chamber or upper divertor. New measurements and SOLPS-ITER simulations emphasise the importance of plasma–neutral interactions on divertor detachment in the conditions accessible in experiments. Real-time control of the ionisation front location in both divertor chambers independently has been demonstrated in double null experiments, enabled by the tightly baffled divertor chambers.

MAST Upgrade↗
DOE OSTI2024

Robust wind farm layout optimization

Wake interactions in wind farms cause losses in annual energy production (AEP) on the order of 10%. Wind farm designers optimize the layout of the farm to mitigate wake losses, especially in the dominant site-specific wind directions. As wind turbines and wind farms grow in scale, optimization becomes more complex. Offshore wind farms regularly comprise more than 100 wind turbines and are characterized by complex boundaries due to shipping lanes, neighboring wind farms, and other constraints. Layout optimization methods are broadly split between gradient-based and gradient-free approaches. Gradient-based approaches can converge quickly and perform well for smaller, academic problems but are often sensitive to initial conditions and tuning parameters and require expert knowledge to use. On the other hand, gradient-free approaches can be more robust to problem complexities. We present a robust layout optimization approach based on a random search algorithm. The algorithm is intended for those who are not optimization experts and has few tuning parameters that need specification to achieve satisfactory results. Unlike off-the-shelf methods, which use generally available, non-domain-specific optimization routines that accept as inputs an optimization function and constraint definitions, this approach takes advantage of the relative computational costs of the different evaluations by evaluating cheaper computations first (boundary and minimum distance constraints) and running expensive AEP evaluations only if all other checks pass. Moreover, an outer genetic algorithm allows multiple solutions to evolve in parallel, enabling rapid solution development on high-performance computers. We discuss the relative ease of selecting necessary tuning parameters and demonstrate the efficacy of the genetic random search on a complex layout problem consisting of placing 70 turbines in a nonconvex and unconnected boundary region.

17 WIND ENERGY↗
DOE OSTI2024

Integrated floating wind farm layout design and mooring system optimization to increase annual energy production

As we cluster wind turbines in wind farms to gain energy from sites with high wind speeds, wake losses occur within the wind farm. Wake loss is a term used to describe the lower energy production of a downwind turbine that is totally or partially in the wake of an upwind turbine. To decrease wake losses inside the wind farm, the wind farm’s layout is optimized. However, a variety of factors constrain the wind farm layout optimization, such as the size of the lease area relative to the number of turbines to be placed, or the shape of the lease area. Therefore, many wind farms end up with a regular grid layout, such as the Horns Rev 1 wind farm in the North Sea. The ability of a floating offshore wind turbine (FOWT) to change its position based on the wind direction and its mooring system design presents an opportunity to further decrease wake losses in floating wind farms. In this work, we integrate the design of the FOWT mooring systems with the floating wind farm layout design with the goal of increasing the farm’s annual energy production. We use the Horns Rev 1 wind farm as a case study to demonstrate our method. The results show that allowing the FOWT to relocate can decrease wake losses up to 18%. Moreover, the newly developed mooring systems are less stiff and therefore allow larger motion of the FOWT; hence, the material cost of the mooring system decreases by an estimated 17%.

17 WIND ENERGY↗
DOE OSTI2025

Recent activities of National Synchrotron Light Source-II Insertion Device Group

Insertion Device (ID) Group at the National Synchrotron Light Source-II (NSLS-II) has been involved in the following activities: Installation and maintenance of the HEX-Superconducting Wiggler (SCW), Laboratory Directed Research and Development (LDRD) for SC adaptive gap undulator (AGU), NSLS-II Experimental Tools II & III (NEXT-II & III) project IDs. Improvement of in-vacuum flip coil bench, development of in-vacuum pulsed wire bench and upgrade of a regular flip coil bench. Development of rotating coil bench for the lattice magnets for future upgrade. This paper describes the specifics of each activity.

47 OTHER INSTRUMENTATION↗
DOE OSTI2026

OC7 Project Phase II: Code Comparison and Experimental Validation of Hydroelastic Effects and Member-Level Loads in Floating Structures

This paper presents results from Phase II of the OC7 project, focusing on hydroelastic modeling and member-level load prediction for a flexible floating offshore structure. Numerical predictions from 11 academic and industrial partners are validated against experimental measurements obtained from a 1:70-scale test of the VolturnUS-S semisubmersible platform. A comprehensive set of verification and validation cases is examined. The results demonstrate that hydrodynamic added mass has a significant impact on predicted elastic natural frequencies. Under regular wave excitation, the numerical models reproduce platform motions and mooring line tensions with good accuracy. Member-level loads are also predicted with reasonable accuracy. Some discrepancies are observed for potential-flow models not accounting for higher-order effects associated with the instantaneous wetted surface.

17 WIND ENERGY↗
DOE OSTI2024

The CMOS Pseudo-Thyristor: a zero-static current discriminator circuit

We report a very low power discriminator circuit for pixelized detectors, called the Pseudo-Thyristor is described in this document. It is a positive feedback topology using regular PMOS and NMOS field-effect transistors (FET's) with zero static current. When a small charge is injected into the circuit, it flips rapidly due to the positive feedback and outputs a logic transition for further digitization. Simulation shows that in a 65 nm process, it is possible to achieve a detecting threshold below 5 fC while maintain the average power consumption below 10 micro-Watts when the hit occupancy is <10% for 40 MHz operation.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2024

Paired autoencoders for likelihood-free estimation in inverse problems

Abstract We consider the solution of nonlinear inverse problems where the forward problem is a discretization of a partial differential equation. Such problems are notoriously difficult to solve in practice and require minimizing a combination of a data-fit term and a regularization term. The main computational bottleneck of typical algorithms is the direct estimation of the data misfit. Therefore, likelihood-free approaches have become appealing alternatives. Nonetheless, difficulties in generalization and limitations in accuracy have hindered their broader utility and applicability. In this work, we use a paired autoencoder framework as a likelihood-free estimator (LFE) for inverse problems. We show that the use of such an architecture allows us to construct a solution efficiently and to overcome some known open problems when using LFEs. In particular, our framework can assess the quality of the solution and improve on it if needed. We demonstrate the viability of our approach using examples from full waveform inversion and inverse electromagnetic imaging.

Chung, Matthias (ORCID:0000000178224539)↗
DOE OSTI2025

SymbolNet: neural symbolic regression with adaptive dynamic pruning for compression

Abstract Compact symbolic expressions have been shown to be more efficient than neural network (NN) models in terms of resource consumption and inference speed when implemented on custom hardware such as field-programmable gate arrays (FPGAs), while maintaining comparable accuracy (Tsoi et al 2024 EPJ Web Conf. 295 09036). These capabilities are highly valuable in environments with stringent computational resource constraints, such as high-energy physics experiments at the CERN Large Hadron Collider. However, finding compact expressions for high-dimensional datasets remains challenging due to the inherent limitations of genetic programming (GP), the search algorithm of most symbolic regression (SR) methods. Contrary to GP, the NN approach to SR offers scalability to high-dimensional inputs and leverages gradient methods for faster equation searching. Common ways of constraining expression complexity often involve multistage pruning with fine-tuning, which can result in significant performance loss. In this work, we propose S y m b o l N e t , a NN approach to SR specifically designed as a model compression technique, aimed at enabling low-latency inference for high-dimensional inputs on custom hardware such as FPGAs. This framework allows dynamic pruning of model weights, input features, and mathematical operators in a single training process, where both training loss and expression complexity are optimized simultaneously. We introduce a sparsity regularization term for each pruning type, which can adaptively adjust its strength, leading to convergence at a target sparsity ratio. Unlike most existing SR methods that struggle with datasets containing more than O ( 10 ) inputs, we demonstrate the effectiveness of our model on the LHC jet tagging task (16 inputs), MNIST (784 inputs), and SVHN (3072 inputs).

Tsoi, Ho Fung (ORCID:0000000225502184)↗
DOE OSTI2025

Overcoming sparse datasets with multi-task learning as applied to high entropy alloys

Abstract The design of novel High Entropy Alloys for use in high-temperature applications is an area of active interest due to their potential to provide exceptional properties compared to conventional alloys. Since the increased popularity of machine learning, an important cog in the design process has been training surrogate models on alloy properties. However, these Single-Task models are trained on individual mechanical properties and do not take advantage of the relatedness between properties. Multi-Task models can capture the interdependencies between tasks, leading to potentially more accurate predictions for all tasks. In this paper, we investigate if Multi-Task models can show improvement over Single-Task models when used for predicting the mechanical properties of these alloys. To ensure fair evaluation between the models, we apply L 0 regularization and skip connections to the models, which allows them to adjust the number of model parameters and depth for optimal performance. We find that the Multi-Task models can leverage task relationships to perform better than Single-Task models, especially for high amounts of missing data in the tasks. Furthermore, adding simple auxiliary targets can boost Multi-Task performance even further despite not being effective as input descriptors to single-task models themselves. We anticipate that the proposed strategies can achieve more accurate predictions and consequently enable better design capabilities for such data-constrained domains without incurring much additional computational cost.

Debnath, Arindam (ORCID:0000000194274499)↗
DOE OSTI2025

Physics-guided dual implicit neural representations for source separation

Significant challenges exist in efficient data analysis of most advanced experimental and observational techniques because the collected signals often include unwanted contributions, such as background and signal distortions, that can obscure the physically relevant information of interest. To address this, we have developed a self-supervised machine-learning approach for source separation using a dual implicit neural representation framework that jointly trains two neural networks: one for approximating distortions of the physical signal of interest and the other for learning the effective background contribution. Our method learns directly from the raw data by minimizing a reconstruction-based loss function without requiring labeled data or pre-defined dictionaries. We demonstrate the effectiveness of our framework by considering a challenging case study involving large-scale simulated, as well as experimental, momentum-energy-dependent inelastic neutron scattering data in a four-dimensional parameter space, characterized by heterogeneous background contributions and unknown distortions to the target signal. The method is found to successfully separate physically meaningful signals from a complex or structured background even when the signal characteristics vary across all four dimensions of the parameter space. An analytical approach that informs the choice of the regularization parameter is presented. Our method offers a versatile framework for addressing source separation problems across diverse domains, ranging from superimposed signals in astronomical measurements to structural features in biomedical image reconstructions.

47 OTHER INSTRUMENTATION↗
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 OSTI2025

LATTE: open-source, high-performance traveltime computation, tomography and source location in acoustic and elastic media

Traveltime-based tomography and source location are fundamental approaches for imaging subsurface structures and understanding the spatiotemporal distribution of seismicity from local to global scales. We present an open-source, high-performance framework integrating eikonal equation solvers and adjoint-state theory for traveltime computation, velocity tomography, source location and joint tomography-location in 2-D/3-D acoustic and elastic media. We introduce novel regularization schemes based on total generalized p-variation, structural similarity and multitask machine learning to enhance the fidelity and interpretability of inverted models and source locations. Key features of our implementation also include the ability to leverage both absolute-difference and double-difference traveltime misfits for high-fidelity velocity tomography and source parameter estimation; support for traveltime computation and inversion in diverse 2-D/3-D scenarios with arbitrary source and receiver distributions; and a perturbation-based optimal step-size estimation method to reduce computational costs. In addition, our implementation employs shared-memory and distributed-memory parallelization to provide an efficient solution for traveltime computation, tomography, and source location. In conclusion, we validate the efficacy and accuracy of our approach through multiple synthetic data examples.

58 GEOSCIENCES↗
DOE OSTI2024

Random Forest Prediction of Crystal Structure from Electron Diffraction Patterns

Transmission electron microscopy (TEM) diffraction patterns are regularly used to determine the structure of crystalline materials. Electron diffraction is the most common method to solve for unknown or partially known crystal structures, as it provides direct and interpretable feedback on the orientation of crystal grains under the beam [1]. However, it remains a challenge to determine the crystal structure of a new material or even a new phase of an existing material. Analysis of such materials commonly requires manual exploration and comparison with simulated diffraction patterns. This is often a time consuming process with no obvious start point when many similar structures are possible, and this method cannot be used to determine crystal structure or orientation from structures not included in the diffraction libraries. Therefore, we have developed a machine learning model to determine the crystal structure of a material from its electron diffraction pattern.

36 MATERIALS SCIENCE↗