Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Nonlinear reduction”

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 253 records · Page 14
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Nonlinear multiobjective and dynamic real-time predictive optimization for optimal operation of baseload power plants under variable renewable energy

Considering the increase of disruptive variable renewable energy penetration into the power grid, this article focuses on the investigation of a multiobjective and dynamic real-time optimization framework to address the cycling of large-scale power plants under renewable penetration. In this framework, a parallelized particle swarm optimization step is first performed to generate feasible initial points. Then, a multiobjective and dynamic real-time optimization formulation generates optimal trajectories. Further, the benefit of predictive capability is investigated for the dynamic component, which introduces the novel nonlinear multiobjective and dynamic real-time predictive optimization approach. Two multiobjective formulations to obtain Pareto front optimal in real time are explored: the modified Tchebycheff-based weighted metric and ϵ-constraint methods. Economic and environmental objectives are considered in this study. A novel topical discussion on the intersection of dynamic real-time optimization with model predictive control is also presented. The developed framework is successfully applied to a baseload coal-fired power plant with postcombustion CO 2 capture. Results indicate that the approach can be deployed for a large-scale system if automatic differentiation, model reduction, and parallelization are adopted to improve computational tractability, with computational improvement up to 120-folds after performing these steps. Finally, market and carbon policies showed an impact on the optimal compromise between the objectives with an additional 63 ton of CO 2 captured under favorable market conditions.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2024

Sparsified time-dependent Fourier neural operators for fusion simulations

This paper presents a sparsified Fourier neural operator for coupled time-dependent partial differential equations (ST-FNO) as an efficient machine learning surrogate for fluid and particle-based fusion codes such as NIMROD (Non-Ideal Magnetohydrodynamics with Rotation - Open Discussion) and GTC (Gyrokinetic Toroidal Code). ST-FNO leverages the structures in the governing equations and utilizes neural operators to represent Green's function-like numerical operators in the corresponding numerical solvers. Once trained, ST-FNO can rapidly and accurately predict dynamics in fusion devices compared with first-principle numerical algorithms. In general, ST-FNO represents an efficient and accurate machine learning surrogate for numerical simulators for multi-variable nonlinear time-dependent partial differential equations, with the proposed architectures and loss functions. The efficacy of ST-FNO has been demonstrated using quiescent H-mode simulation data from NIMROD and kink-mode simulation data from GTC. The ST-FNO H-mode results show orders of magnitude reduction in memory and central processing unit usage in comparison with the numerical solvers in NIMROD when computing fields over a selected poloidal plane. The ST-FNO kink-mode results achieve a factor of 2 reduction in the number of parameters compared to baseline FNO models without accuracy loss.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2021

Constrained Block Nonlinear Neural Dynamical Models

Neural network modules conditioned by known priors can be effectively trained and combined to represent systems with nonlinear dynamics. This work explores a novel formulation for data-efficient learning of deep control-oriented nonlinear dynamical models by embedding local model structure and constraints. The proposed method consists of neural network blocks that represent input, state, and output dynamics with constraints placed on the network weights and system variables. For handling partially observable dynamical systems, we utilize a state observer neural network to estimate the states of the system's latent dynamics. We evaluate the performance of the proposed architecture and training methods on system identification tasks for three nonlinear systems: a continuous stirred tank reactor, a two tank interacting system, and an aerodynamics body. Models optimized with a few thousand system state observations accurately represent system dynamics in open loop simulation over thousands of time steps from a single set of initial conditions. Experimental results demonstrate an order of magnitude reduction in open-loop simulation mean squared error for our constrained, block-structured neural models when compared to traditional unstructured and unconstrained neural network models.

Skomski, Elliott↗
DOE OSTI2026

Interpretable and flexible non-intrusive reduced-order models using reproducing kernel Hilbert spaces

This paper develops an interpretable, non-intrusive reduced-order modeling technique using regularized kernel interpolation. Existing non-intrusive approaches approximate the dynamics of a reduced-order model (ROM) by solving a data-driven least-squares regression problem for low-dimensional matrix operators. Our approach instead leverages regularized kernel interpolation, which yields an optimal approximation of the ROM dynamics from a user-defined reproducing kernel Hilbert space. We show that our kernel-based approach can produce interpretable ROMs whose structure mirrors full-order model structure by embedding judiciously chosen feature maps into the kernel. The approach is flexible and allows a combination of informed structure through feature maps and closure terms via more general nonlinear terms in the kernel. We also derive a computable a posteriori error bound that combines standard error estimates for intrusive projection-based ROMs and kernel interpolants. In conclusion, the approach is demonstrated in several numerical experiments that include comparisons to operator inference using both proper orthogonal decomposition and quadratic manifold dimension reduction.

Data-driven model reduction↗
DOE OSTI2021

Nonlinear dynamics of quadratic gravity in spherical symmetry

We present the first numerically stable nonlinear evolution for the leading-order gravitational effective field theory (quadratic gravity) in the spherically-symmetric sector. The formulation relies on (i) the harmonic gauge to cast the evolution system into quasilinear form (ii) the Cartoon method to reduce to spherical symmetry in keeping with the harmonic gauge, and (iii) order reduction to first order (in time) by means of introducing auxiliary variables. The well posedness of the respective initial-value problem is numerically confirmed by evolving randomly perturbed flat-space and black-hole initial data. In conclusion, our study serves as a proof-of-principle for the possibility of stable numerical evolution in the presence of higher derivatives.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2022

Parallel-in-Time Integration for Nonlinear Hyperbolic Problems (Final Report)

The work for the subcontract was situated in the area of parallel-in-time integration for hyperbolic partial differential equations (PDEs). Parallel-in-time integration is an active area of research due to its ability to enable faster numerical simulations for applications throughout many areas of science. Over the past two decades, much progress has been made in this area; however, this progress has largely been limited to diffusion-dominated PDEs, with some recent success in scalar linear hyperbolic PDEs. Given the ubiquity of numerical simulations of hyperbolic PDEs throughout the sciences, in particular, nonlinear hyperbolic systems, there is a strong need to develop efficient parallel-in-time techniques for hyperbolic problems beyond simple scalar and linear cases, which is the main topic of this subcontract. The main focus of the work was to further develop and perfect coarse-grid operators for the Multigrid Reduction-inTime (MGRIT) method applied to hyperbolic PDEs that were recently proposed in PhD thesis, based on a modified semi-Lagrangian approach.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

Limits on Mode Coherence in Pulsating DA White Dwarfs Due to a Nonstatic Convection Zone

The standard theory of pulsations deals with the frequencies and growth rates of infinitesimal perturbations in a stellar model. Modes which are calculated to be linearly driven should increase their amplitudes exponentially with time; the fact that nearly constant amplitudes are usually observed is evidence that nonlinear mechanisms inhibit the growth of finite amplitude pulsations. Models predict that the mass of convection zones in pulsating hydrogen-atmosphere (DAV) white dwarfs is very sensitive to temperature (i.e., M CZ α T –90 eff ), leading to the possibility that even low-amplitude pulsators may experience significant nonlinear effects. In particular, the outer turning point of finite-amplitude g-mode pulsations can vary with the local surface temperature, producing a reflected wave that is out of phase with what is required for a standing wave. This can lead to a lack of coherence of the mode and a reduction in its global amplitude. In this paper we show that: (1) whether a mode is calculated to propagate to the base of the convection zone is an accurate predictor of its width in the Fourier spectrum, (2) the phase shifts produced by reflection from the outer turning point are large enough to produce significant damping, and (3) amplitudes and periods are predicted to increase from the blue edge to the middle of the instability strip, and subsequently decrease as the red edge is approached. Furthermore, this amplitude decrease is in agreement with the observational data while the period decrease has not yet been systematically studied.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2022

Onymous early‐life performance degradation analysis of recent photovoltaic module technologies

Abstract The cost of photovoltaic (PV) modules has declined by 85% since 2010. To achieve this reduction, manufacturers altered module designs and bill of materials; changes that could affect module durability and reliability. To determine if these changes have affected module durability, we measured the performance degradation of 834 fielded PV modules representing 13 module types from 7 manufacturers in 3 climates over 5 years. Degradation rates ( Rd ) are highly nonlinear over time, and seasonal variations are present in some module types. Mean and median degradation rate values of −0.62%/year and −0.58%/year, respectively, are consistent with rates measured for older modules. Of the 23 systems studied, 6 have degradation rates that will exceed the warranty limits in the future, whereas 13 systems demonstrate the potential of achieving lifetimes beyond 30 years, assuming Rd trends have stabilized.

14 SOLAR ENERGY↗
DOE OSTI2022

MULTI-TECHNIQUE CHARACTERIZATION OF SUPERCONDUCTING MATERIALS FOR PARTICLE ACCELERATOR APPLICATIONS

We investigated the performance limitations of superconducting radio-frequency (SRF) cavities and materials using multiple experimental techniques. In particular, this study focuses on understanding the surface properties of nitrogen-doped Nb cavities and super?conducting thin films with higher Tc such as Nb3Sn. The main goal of this work is to use different techniques to better understand each aspect of the complex loss mechanism in superconductors to further improve the already highly efficient SRF cavities. Nitrogen doping applied to a Nb SRF cavity significantly improves the quality factor Q0 compared to a conventional Nb cavity, at an expense of reduced maximum accelerating gradient. The early quench mechanism was analyzed by using temperature maps before and during the quenching. The temperature maps revealed insignificant heating before the quench, and we concluded that nitrogen doping reduces critical magnetic fields in local regions, leading to premature quenching. To understand the origin of the increasing Q0 with the rf field, the density of states (DOS) of cold spots from nitrogen-doped and standard cavities were measured and analyzed using scanning tunneling microscopy. The results suggested that nitrogen doping reduces the spatial inhomogeneity of superconducting properties and shrinks the metallic suboxide layers, which tunes the DOS in such a way as to produce the field-induced reduction in the surface resistance. To characterize SRF thin films, an experimental setup for measuring a coplanar waveguide (CPW) resonator was developed and tested. A surface impedance measurement of the Nb film showed good agreement with the BCS calculation. The preliminary results from measurements of Nb3Sn and NbTiN films are also presented here. The nonlinear Meissner effect was investigated in Nb3Sn film CPW resonators by mea?suring the resonant frequency as a function of a parallel magnetic field. Contrary to a conventional quadratic dependence of the penetration depth ?(B) on the applied magnetic field B, as expected in s-wave superconductors, nearly a linear increase of ?(B) with B was observed. It was concluded that this behavior of ?(B) is due to weakly linked grain bound?aries on the polycrystalline Nb3Sn films, which can mimic the NLME expected in a clean d-wave superconductor.

Makita, Junki↗
DOE OSTI2023

Improving snow albedo modeling in the E3SM land model (version 2.0) and assessing its impacts on snow and surface fluxes over the Tibetan Plateau

With the highest albedo of the land surface, snow plays a vital role in Earth's surface energy budget and water cycle. Snow albedo is primarily controlled by snow grain properties (e.g., size and shape) and light-absorbing particles (LAPs) such as black carbon (BC) and dust. The mixing state of LAPs in snow also has impacts on LAP-induced snow albedo reduction and surface radiative forcing (RF). However, most land surface models assume that snow grain shape is spherical and LAPs are externally mixed with the snow grains. This study improves the snow radiative transfer model in the Energy Exascale Earth System Model version 2.0 (E3SM v2.0) Land Model (ELM v2.0) by considering non-spherical snow grain shapes (i.e., spheroid, hexagonal plate, and Koch snowflake) and internal mixing of dust–snow, and it systematically evaluates the impacts on the surface energy budget and water cycle over the Tibetan Plateau (TP). A series of ELM simulations with different treatments of snow grain shape, mixing state of BC–snow and dust–snow, and sub-grid topographic effects (TOP) on solar radiation are performed. Compared with two remote sensing snow products derived from the Moderate Resolution Imaging Spectroradiometer, the control ELM simulation (ELM_Control) with the default configurations of spherical snow grain shape, internal mixing of BC–snow, external mixing of dust–snow, and without TOP as well as the ELM simulation with new model features (ELM_New) can both capture the overall snow distribution reasonably. Additionally, ELM_New overall shows smaller biases in snow cover fraction than ELM_Control in spring when snowmelt is important for water management. The estimated LAP-induced RF in ELM_New ranges from 0 to 19.3 W m -2 with the area-weighted average value of 1.5 W m -2 that is comparable to the reported values in existing studies. The Koch snowflake shape, among other non-spherical shapes, shows the largest difference from the spherical shape in spring when snow processes related to the surface energy budget and water cycle have high importance. The impacts of the mixing state of LAP in snow are smaller than the shape effects and depend on snow grain shape. Compared to external mixing, internal mixing of LAP–snow can lead to larger snow albedo reduction and snowmelt, which further affect the surface energy budget and water cycle. The individual contributions of non-spherical snow shape, mixing state of LAP–snow, and local topography impacts on the snow and surface fluxes have different signs and magnitudes, and their combined effects may be negative or positive due to complex and nonlinear interactions among the factors. Overall, the changes in net solar radiation in spring due to individual and combined effects range from -28.6 to 16.9 W m -2 and -29.7 to 12.2 W m -2 , respectively. This study advances understanding of the role of snow grain shape and mixing state of LAP–snow in land surface processes and offers guidance for improving snow simulations and RF estimates in Earth system models under climate change.

58 GEOSCIENCES↗
DOE OSTI2020

UQpy: A general purpose Python package and development environment for uncertainty quantification

In this paper, we present the UQpy software toolbox, an open-source Python package for general uncertainty quantification (UQ) in mathematical and physical systems. The software serves as both a user-ready toolbox that includes many of the latest methods for UQ in computational modeling and a convenient development environment for Python programmers advancing the field of UQ. The paper presents an introduction to the software's architecture and existing capabilities, divided in the code in a set of modules centered around different UQ tasks such as sampling methods, generation of random processes and random fields, probabilistic inverse modeling, reliability analysis, surrogate modeling, and active learning. The paper also highlights the importance of the RunModel module, which is used to drive simulations in the uncertainty analyses performed in UQpy. This module conveniently allows the user to define computational models directly in Python, or to run simulations from a third-party software in serial or in parallel. To illustrate the various capabilities, two examples are tracked throughout the paper and analyzed repeatedly for various UQ tasks. The first is a Python model solving a nonlinear structural dynamics problem, used to illustrate UQpy's capabilities in sampling and forward propagation of high dimensional random vectors (stochastic processes), and probabilistic inference. The second model is a third-party Abaqus finite element model solving the thermomechanical response of a beam structure. This example is used to illustrate UQpy's capabilities in variance reduction sampling techniques, reliability analysis, surrogate modeling and active learning techniques.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

Distributed Aerodynamic Control using Active Trailing-Edge Flaps for Large Wind Turbines

This work presents a numerical framework to investigate distributed aerodynamic control devices for application in large wind turbines. Tool capabilities were extended to facilitate multiple aerodynamic polar tables. The airfoil aerodynamics characteristics were automatically determined, and blade-pitch, generator-torque, and trailing-edge-flap controllers were tuned in-the-loop according to a specific blade design. This automated workflow allows analysis and optimization of trailing-edge flaps, enabling codesign studies. Results targeted reductions of root-flap-bending moment derivatives. The applied trailing-edge-flap control reduced the standard deviation of root-flap-bending moments by more than 6% and benefit related parameters, e.g., reduce blade-tip deflections, by up to 8%. Because of varying thrust distributions along the blade span, different flap designs have nonlinear characteristics in terms of the control objective and show best performance when located at the radial position with maximum thrust. In general, larger flaps provide a greater influence to reduce the target control objective.

17 WIND ENERGY↗
DOE OSTI2025

Hybrid learning techniques for scientific data reduction with performance guarantees

The research initiatives supported by the U.S. Department of Energy (DOE) Grant DE-SC0022265 are fundamentally aimed at pioneering advanced machine learning (ML) techniques for scientific data compression within high-performance computing (HPC) environments. This comprehensive body of work addresses the critical challenge posed by the exponential growth of data generated by scientific simulations in domains such as fusion energy, climate modeling, and computational fluid dynamics (CFD). A core objective is to develop compression algorithms that achieve substantial data reduction—often by orders of magnitude—while rigorously ensuring the fidelity of both the primary data (PD) and scientifically crucial derived quantities of interest (QoI). The methodologies deployed under this grant integrate sophisticated deep learning architectures, prominently featuring autoencoders, advanced generative models like conditional diffusion, and hybrid learning techniques. Key innovations include the development of Guaranteed Autoencoders (GAE) and the Guaranteed Conditional Diffusion with Tensor Correction (GCDTC) framework, which provide explicit, instance-level error bounds on reconstructed data. Furthermore, specialized strategies such as nonlinear constraint satisfaction are employed to preserve the integrity of QoI, a vital requirement for the trustworthiness of downstream scientific analyses. This research also focuses on the design and implementation of scalable, GPU-accelerated software pipelines that seamlessly integrate into existing HPC workflows, ensuring both computational efficiency and practical applicability. The CAESAR framework, for example, unifies foundation and generative models to create an adaptive and efficient compression solution for spatio-temporal scientific data. Collectively, these efforts represent a significant advancement in mitigating the scientific data deluge, enabling more effective data management, accelerated scientific discovery, and optimized utilization of HPC resources.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Final report- UFL - RAPIDS2: A SciDAC Institute for Computer Science, Data, and Artificial Intelligence

The research initiatives supported by the U.S. Department of Energy (DOE) Grant DE-SC0022265 are fundamentally aimed at pioneering advanced machine learning (ML) techniques for scientific data compression within high-performance computing (HPC) environments. This comprehensive body of work addresses the critical challenge posed by the exponential growth of data generated by scientific simulations in domains such as fusion energy, climate modeling, and computational fluid dynamics (CFD). A core objective is to develop compression algorithms that achieve substantial data reduction—often by orders of magnitude—while rigorously ensuring the fidelity of both the primary data (PD) and scientifically crucial derived quantities of interest (QoI). The methodologies deployed under this grant integrate sophisticated deep learning architectures, prominently featuring autoencoders, advanced generative models like conditional diffusion, and hybrid learning techniques. Key innovations include the development of Guaranteed Autoencoders (GAE) and the Guaranteed Conditional Diffusion with Tensor Correction (GCDTC) framework, which provide explicit, instance-level error bounds on reconstructed data. Furthermore, specialized strategies such as nonlinear constraint satisfaction are employed to preserve the integrity of QoI, a vital requirement for the trustworthiness of downstream scientific analyses. This research also focuses on the design and implementation of scalable, GPU-accelerated software pipelines that seamlessly integrate into existing HPC workflows, ensuring both computational efficiency and practical applicability. The CAESAR framework, for example, unifies foundation and generative models to create an adaptive and efficient compression solution for spatio-temporal scientific data. Collectively, these efforts represent a significant advancement in mitigating the scientific data deluge, enabling more effective data management, accelerated scientific discovery, and optimized utilization of HPC resources.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Drift kinetic theory of the NTM magnetic islands in a finite beta general geometry tokamak plasma

In (Imada et al 2019 Nucl. Fusion 59 046016 and references therein) a new 4D drift kinetic nonlinear theory, valid in the limit of a low beta, small inverse aspect ratio, circular cross section, toroidal geometry, to describe the plasma response to the neoclassical tearing mode (NTM) magnetic perturbation is derived. In (Dudkovskaia et al 2021 Plasma Phys. Control. Fusion 63 054001) this theory is reduced in a low collisionality limit, which allows a dimensionality reduction to a 3D problem to efficiently resolve the collisional dissipation layer in the vicinity of the trapped-passing boundary. (Dudkovskaia et al 2021 Plasma Phys. Control. Fusion 63 054001) adopts an improved model for the magnetic drift frequency, which reduces the threshold magnetic island half-width from $8.73\rho_{b i}$, where $\rho_{b i}$ is the trapped ion banana orbit width, to $1.46 \rho_{b i}$, making it in closer agreement with experimental observations for the large aspect ratio tokamak equilibrium. In the present paper, the theory is extended to a high beta, arbitrary tokamak geometry to capture the plasma shaping effects on the NTM threshold physics with the focus on the non-zero triangularity discharges that are known to have a strong impact on the plasma MHD stability. First, it is found that the higher triangularity plasma is more prone to NTMs which is in agreement with the 2/1 tearing mode onset relative frequency measurements in DIII-D. Second, the NTM threshold dependence on the tokamak inverse aspect ratio obtained in (Dudkovskaia et al 2021 Plasma Phys. Control. Fusion 63 054001) is refined and extended to a finite aspect ratio limit. Third, the NTM threshold dependence on poloidal beta is obtained and benchmarked against the EAST threshold island width measurements.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2020

Robust Learning with Implicit Residual Networks

In this effort, we propose a new deep architecture utilizing residual blocks inspired by implicit discretization schemes. As opposed to the standard feed-forward networks, the outputs of the proposed implicit residual blocks are defined as the fixed points of the appropriately chosen nonlinear transformations. We show that this choice leads to the improved stability of both forward and backward propagations, has a favorable impact on the generalization power, and allows for control the robustness of the network with only a few hyperparameters. In addition, the proposed reformulation of ResNet does not introduce new parameters and can potentially lead to a reduction in the number of required layers due to improved forward stability. Finally, we derive the memory-efficient training algorithm, propose a stochastic regularization technique, and provide numerical results in support of our findings.

Reshniak, Viktor↗
DOE OSTI2022

Optimal Operation of PV Sources in DC Grids for Improving Technical, Economical, and Environmental Conditions by Using Vortex Search Algorithm and a Matrix Hourly Power Flow

This document presents a master–slave methodology for solving the problem of optimal operation of photovoltaic (PV) distributed generators (DGs) in direct current (DC) networks. This problem was modeled using a nonlinear programming model (NLP) that considers the minimization of three different objective functions in a daily operation of the system. The first one corresponds to the minimization of the total operational cost of the system, including the energy purchasing cost to the conventional generators and maintenance costs of the PV sources; the second objective function corresponds to the reduction of the energy losses associated with the transport of energy in the network, and the third objective function is related to the minimization of the total emissions of CO2 by the conventional generators installed on the DC grid. The minimization of these objective functions is achieved by using a master–slave optimization approach through the application of the Vortex Search algorithm combined with a matrix hourly power flow. To evaluate the effectiveness and robustness of the proposed approach, two test scenarios were used, which correspond to a grid-connected and a standalone network located in two different regions of Colombia. The grid-connected system emulates the behavior of the solar resource and power demand of the city of Medellín-Antioquia, and the standalone network corresponds to an adaptation of the generation and demand curves for the municipality of Capurganá-Choco. A numerical comparison was performed with four optimization methodologies reported in the literature: particle swarm optimization, multiverse optimizer, crow search algorithm, and salp swarm algorithm. The results obtained demonstrate that the proposed optimization approach achieved excellent solutions in terms of response quality, repeatability, and processing times.

14 SOLAR ENERGY↗