Search NASA⌕ Search

SEARCH · Search NASA

Results for “manifold optimization”

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 19 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Optimization on Manifolds via Graph Gaussian Processes

This paper integrates manifold learning techniques within a Gaussian process upper confidence bound algorithm to optimize an objective function on a manifold. Our approach is motivated by applications where a full representation of the manifold is not available and querying the objective is expensive. We rely on a point cloud of manifold samples to define a graph Gaussian process surrogate model for the objective. Query points are sequentially chosen using the posterior distribution of the surrogate model given all previous queries. We establish regret bounds in terms of the number of queries and the size of the point cloud. Several numerical examples complement the theory and illustrate the performance of our method.

Bayesian optimization↗
DOE OSTI2025

Compressing Vision Transformers in Geospatial Transfer Learning with Manifold-Constrained Optimization

Deploying geospatial foundation models on resource-constrained edge devices demands compact architectures that maintain high downstream performance. However, their large parameter counts and the accuracy loss often induced by compression limit practical adoption.In this work, we leverage manifold-constrained optimization framework DLRT to compress large vision transformer–based geospatial foundation models during transfer learning. By enforcing structured low-dimensional parameterizations aligned with downstream objectives, this approach achieves strong compression while preserving task-specific accuracy. We show that the method outperforms of-the-shelf low-rank methods as LoRA. Experiments on diverse geospatial benchmarks confirm substantial parameter reduction with minimal accuracy loss, enabling high-performing, on-device geospatial models.

Snyder, Thomas [Yale University]↗
DOE OSTI2024

CMLM (Co-Optimized Machine-Learned Manifolds) [SWR-23-41]

Co-optimized Machine-Learned Manifolds (CMLM) is a data-driven approach for developing reduced-order manifold models for high-dimensional chemically reacting systems. It involves a specially designed neural network, the training of which simultaneously optimizes linear combinations of species that define the manifold, nonlinear mapping to outputs of interest such as reaction rates, and (optionally) subfilter closure for large eddy simulation. This software package provides an implementation of the CMLM approach in Python using the PyTorch machine learning library. A few example cases are included, showing how the tool can be applied to different types of data from 0D and 1D reacting simulations performed using Cantera. The neural networks can be saved in a format that is readable by the Pele suite of combustion solvers for use in reacting computational fluid dynamics simulations. This software repository contains several python scripts to perform various tasks associated with the Co-optimized Machine Learned Manifolds (CMLM) model, which is described in Perry, Henry de Frahan, and Yellapantula, CNF, 2022 (https://doi.org/10.1016/j.combustflame.2022.112286). This includes not only the code that defines the CMLM model, but also scripts to generate suitable training data, scripts to pre-process the data, scripts to train the CMLM model, and scripts to plot the output, as well as various other helper files. The scripts depend on several commonly used python libraries for data analysis and chemical reaction computations. The trained models that result from this tool are designed to work with the an interface being implemented in the Pele suite of reacting flow solvers (https://github.com/AMReX-Combustion).

Perry, Bruce↗
DOE OSTI2024

Exploiting Power Flow Manifold to Solve AC Optimal Power Flow

AC optimal power flow has proven difficult to solve with interior point methods on GPUs. This is largely due to challenging linear algebra problems that current state of the art massively parallel linear solvers struggle with. However, the advent of Riemannian optimization techniques and the fact that the power flow equations form a smooth manifold present an alternative approach. In this talk, we present the basics of Riemannian optimization techniques in which optimization is done directly on a manifold. Then we present computational results showing that Riemannian techniques are capable of producing solutions of comparable quality as interior point methods.

AC optimal power flow↗
DOE OSTI2025

Power Flow Geometry and Approximation

Here, the power flow equations are important in numerous power systems problems of practical interest which consider alternating current power flow (ACPF) physics. Perhaps the most well studied being the alternating current optimal power flow problem (ACOPF), seeking to optimize the operation of an electric power system. Due to their non-linearity, problems which include the power flow equations are typically challenging, particularly in optimization. Interestingly, the set of solutions to the power flow equations forms a smooth manifold. As a result, differential geometry can be used to describe and analyze this set of equations. This approach has proven effective in several engineering applications (e.g., solving ACOPF and analyzing the solution space boundary). Central to the success of this approach is an understanding of the power flow manifold's geometry. In this work, we develop the geometric and topological properties of this manifold using concepts from differential geometry. After demonstrating the convenience of this manifold's representation as a function's graph, computational methods are emphasized: we develop retractions, error bounds for linear approximation, and formulas for evaluating the Riemannian metric (including associated objects such as geodesics and the curvature tensor). Scalar curvature and the second fundamental form play a new role in quantifying the quality of linear approximations, like the popular direct current approximation. All functions are implemented in Julia and available in an online repository. Proofs are included for completeness.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2024

Riemannian Optimization Applied to AC Optimal Power Flow: Preprint

The nonlinear, nonconvex AC optimal power flow problem is of growing importance as the nature of the power grid evolves. This problem can be difficult to solve for interior point methods. However, the advent of optimization algorithms over smooth Riemannian manifolds presents an alternative approach. The nonlinear, nonconvex constraints in the AC power flow problem form an embedded submanifold of Euclidean space. In this paper, the authors explore the performance of Riemannian optimization algorithms for the ACOPF problem where the optimization is performed directly on the AC power flow manifold. They demonstrate that these are viable computational alternatives to interior point methods. This is done by using Julia and the packages PowerModels.jl and Manopt.jl.

manifold optimization↗
DOE OSTI2024

Riemannian Optimization Applied to AC Optimal Power Flow

The nonlinear, nonconvex AC optimal power flow problem is of growing importance as the nature of the power grid evolves. This problem can be difficult to solve for interior point methods. However, the advent of optimization algorithms over smooth Riemannian manifolds presents an alternative approach. The nonlinear, nonconvex constraints in the AC power flow problem form an embedded submanifold of Euclidean space. In this paper, the authors explore the performance of Riemannian optimization algorithms for the ACOPF problem where the optimization is performed directly on the AC power flow manifold. This is done by using the Julia programming language and the Julia packages PowerModels.jl and Manopt.jl.

AC optimal power flow↗
DOE OSTI2025

Data-Driven Compositional Optimization in Misspecified Regimes

With a manifold growth in the scale and intricacy of systems, the challenges of parametric misspecification become pronounced. These concerns are further exacerbated in compositional settings, which emerge in problems complicated by modeling risk and robustness. In “Data-Driven Compositional Optimization in Misspecified Regimes,” the authors consider the resolution of compositional stochastic optimization problems, plagued by parametric misspecification. In considering settings where such misspecification may be resolved via a parallel learning process, the authors develop schemes that can contend with diverse forms of risk, dynamics, and nonconvexity. They provide asymptotic and rate guarantees for unaccelerated and accelerated schemes for convex, strongly convex, and nonconvex problems in a two-level regime with extensions to the multilevel setting. Surprisingly, the nonasymptotic rate guarantees show no degradation from the rate statements obtained in a correctly specified regime and the schemes achieve optimal (or near-optimal) sample complexities for general T-level strongly convex and nonconvex compositional problems.

Business & Economics↗
DOE OSTI2025

Using "AI Poincare" to analyze non-linear integrable optics

This study dives into the applicability of using automated discovery of conserved quantities in dynamical systems relevant to accelerator physics. Specifically, we explore the performance of AI Poincaré in analyzing numerical trajectory data obtained using the McMillan system of non-linear integrable optics. A comprehensive evaluation of the algorithm's performance is conducted through diverse methodologies. These include the analysis of the estimated number of conserved quantities embedded in a dataset and the deviation of interpolated points on the inferred manifold with respect to points in actually in the dataset. the investigation identifies an optimal range of perturbation distances where the underlying manifold extraction algorithm inside AI Poincaré exhibits optimal performance. Additionally, an improved neural network architecture is proposed based on the observed results. Finally, we apply the algorithm to preliminary experimental data from the Integrable Optics Test Accelerator at Fermilab to successfully infer the number of conserved quantities even in the presence of fast decoherence of the measured signal.

Osmanov, Lazare [Free U. Tbilisi]↗
DOE OSTI2024

Active learning of ternary alloy structures and energies

Abstract Machine learning models with uncertainty quantification have recently emerged as attractive tools to accelerate the navigation of catalyst design spaces in a data-efficient manner. Here, we combine active learning with a dropout graph convolutional network (dGCN) as a surrogate model to explore the complex materials space of high-entropy alloys (HEAs). We train the dGCN on the formation energies of disordered binary alloy structures in the Pd-Pt-Sn ternary alloy system and improve predictions on ternary structures by performing reduced optimization of the formation free energy, the target property that determines HEA stability, over ensembles of ternary structures constructed based on two coordinate systems: (a) a physics-informed ternary composition space, and (b) data-driven coordinates discovered by the Diffusion Maps manifold learning scheme. Both reduced optimization techniques improve predictions of the formation free energy in the ternary alloy space with a significantly reduced number of DFT calculations compared to a high-fidelity model. The physics-based scheme converges to the target property in a manner akin to a depth-first strategy, whereas the data-driven scheme appears more akin to a breadth-first approach. Both sampling schemes, coupled with our acquisition function, successfully exploit a database of DFT-calculated binary alloy structures and energies, augmented with a relatively small number of ternary alloy calculations, to identify stable ternary HEA compositions and structures. This generalized framework can be extended to incorporate more complex bulk and surface structural motifs, and the results demonstrate that significant dimensionality reduction is possible in thermodynamic sampling problems when suitable active learning schemes are employed.

Chemistry↗
DOE OSTI2026

Metal additively manufactured wavy fin cold-plate architecture for improved thermal-hydraulic performance

Rapid growth in artificial intelligence and data center workloads demands high-performance liquid cooling to manage increasing chip power. This study presents two metal-additive-manufactured cold plates with sinusoidal fins, constant-amplitude wavy fins and linearly variable-amplitude wavy fins and compares them against metal-additive-manufactured straight fins using experiments conducted at 1 kW heat dissipation as well as high-fidelity 3D conjugate computational fluid dynamic simulations. The cold plates were printed in AlSi10Mg material and underwent design using a Python-automated workflow prior to manufacture and testing. The experiments show that wavy fins reduce the normalized thermal resistance by 35 to 45 % at water flow rates from 1 to 4 LPM. At a fixed 20 kPa pressure drop, the variable-waviness design lowered peak surface temperature by 9 °C and thermal resistance by 51 %, while edge-channel maldistribution in the constant wavy fin design limited gains. A thermal resistance breakdown revealed that 55–63 % of the total thermal resistance in wavy designs comes from base heat conduction, 27–33 % from fin heat conduction, and 9–13 % from fin heat convection, indicating the need to address conduction bottlenecks. Parametric sweeps identify a 3 mm fin pitch as optimal, and that horizontal inlet/outlet manifolds further reduce pressure drop by 30–60 % and thermal resistance by 9–16 % relative to vertical inlet-outlet manifolds. The results yield comprehensive guidelines for fin geometry, manifold alignment, material selection and additive-manufacturing constraints to realize high-performance liquid-cooled cold plates for power-dense electronics.

3d printing↗
DOE OSTI2024

Finite elements for Matérn-type random fields: Uncertainty in computational mechanics and design optimization

This work highlights an approach for incorporating realistic uncertainties into scientific computing workflows based on finite elements, focusing on prevalent applications in computational mechanics and design optimization. We leverage Matérn-type Gaussian random fields (GRFs) generated using the SPDE method to model aleatoric uncertainties, including environmental influences, variating material properties, and geometric ambiguities. Our focus lies on delivering practical GRF realizations that accurately capture imperfections and variations and understanding how they impact the predictions of computational models as well as the shape and topology of optimized designs. Here we describe a numerical algorithm based on solving a generalized SPDE to sample GRFs on arbitrary meshed domains. The algorithm leverages established techniques and integrates seamlessly with the open-source finite element library MFEM and associated scientific computing workflows, like those found in industrial and national laboratory settings. Our solver scales efficiently for large-scale problems and supports various domain types, including surfaces and embedded manifolds. We showcase its versatility through biomechanics and topology optimization applications, emphasizing the potential to influence these domains. The flexibility and efficiency of SPDE-based GRF generation empowers us to run large-scale optimization problems on 2D and 3D domains, including finding optimized designs on embedded surfaces, and to generate design features and topologies beyond the reach of conventional techniques. Moreover, these capabilities allow us to model and quantify geometric uncertainties on reconstructed submanifolds, such as the interpolated surfaces of cerebral aneurysms provided by postprocessing CT scans. In addition to offering benefits in these specific domains, the proposed techniques transcend specific applications and generalize to arbitrary forward and backward problems in uncertainty quantification involving finite elements.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Deep kernel methods learn better: from cards to process optimization

Abstract The ability of deep learning methods to perform classification and regression tasks relies heavily on their capacity to uncover manifolds in high-dimensional data spaces and project them into low-dimensional representation spaces. In this study, we investigate the structure and character of the manifolds generated by classical variational autoencoder (VAE) approaches and deep kernel learning (DKL). In the former case, the structure of the latent space is determined by the properties of the input data alone, while in the latter, the latent manifold forms as a result of an active learning process that balances the data distribution and target functionalities. We show that DKL with active learning can produce a more compact and smooth latent space which is more conducive to optimization compared to previously reported methods, such as the VAE. We demonstrate this behavior using a simple cards dataset and extend it to the optimization of domain-generated trajectories in physical systems. Our findings suggest that latent manifolds constructed through active learning have a more beneficial structure for optimization problems, especially in feature-rich target-poor scenarios that are common in domain sciences, such as materials synthesis, energy storage, and molecular discovery. The Jupyter Notebooks that encapsulate the complete analysis accompany the article.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Multiscale Porous High-temperature Heat Exchanger Using Ceramic Co-extrusion

In this project, our MIT, Purdue, and GE team aims to design, model, fabricate, and test a novel high temperature, compact, and durable ceramic heat exchanger to be operated under high temperature and pressure conditions for aerospace applications. Our approach is grounded in introducing multiscale porosity, i.e., centimeter-scale channels embedded with micrometer-scale channels, into the ceramic heat exchanger to significantly improve its heat transfer performance and mechanical strength while maintaining minimal pressure losses. We first developed high-fidelity thermal-fluid-mechanical model capable of precisely capturing the heat transfer rate, temperature profile, pressure drop, and mechanical stress throughout the entire heat exchanger design. Guided by our model, we identified the optimal design parameters for the SiC heat exchanger body and manifolds. Then, we established a completed fabrication procedure to create multiscale features in the ceramic heat exchanger, including co-extrusion, lamination, burnout, and sintering. We fabricated multiple heat exchanger bodies consisting of 6 × 6 and 3 × 3 centimeter-scale channels where each individual centimeter-scale channel comprises 625 crack-free microchannels with 90 μm × 90 μm opening. Owing to the multiscale features, our fabricated heat exchanger bodies exhibited desirable mechanical strength with 156 MPa flexural strength under 1300 Celsius degree. Despite the demonstrated highly tailorable microscopic features and superior mechanical strength, we identified delamination due to the complex interaction among ceramic, polymer, and gas species can be a critical challenge to create fully defect-free heat exchanger, which requires further fundamental investigations in future works. To test the heat exchanger performance, we constructed a high-temperature and high-pressure experimental apparatus that can be safely operated under 400 Celsius degree and 4 bar. With insights gained from mechanistic modeling, material development, and detailed characterization, a cost model was finally developed to understand the market potential of the developed technology, where a cost of $43,000 Celsius degree/kW was envisioned. This project developed a transformative approach to high-performance heat exchanger design. The thermal-fluid-mechanical design approach developed in this project can serve as a generic tool to guide the design of various heat-exchangers operated under high-temperature and high-pressure conditions. The material fabrication approach established in this project can be a useful guide for ceramic processing at extreme length scales.

42 ENGINEERING↗
DOE OSTI2024

Mathematical methods for optimal polynomial recovery of high-dimensional systems from noisy data

The goal of our Early Career Research Project (ECRP) is to establish a modern mathematical foundation that will enable next-generation computational methods for polynomial approximation of high-dimensional systems, having a certain set of constraints, from a limited amount of noisy data. Such a foundation is critical to realizing the future potential of the DOE user facilities, and will ultimately empower scientists to address a fundamental question, namely, “how many realizations of a nonlinear manifold are required to recover the entire high-dimensional solution map, with optimal approximation guarantees and minimal computational cost?” The central theme of this effort aims to conquer this challenge by pioneering the development of extraordinarily innovative theoretical analysis and transformational non-intrusive computational methodologies. Such approaches will enable the reconstruction of the entire high-dimensional solution map, with accuracy comparable to the best approximation, while utilizing an optimal number of samples. During this reporting period we have made significant progress on four thrusts.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Optimal operations of a nuclear-based integrated energy system: A mixed integer program approach

A nuclear-based integrated energy system (IES), consisting of multiple carbon-free energy generation and conversion technologies to meet heterogeneous end-use demands, offers a promising approach to decarbonizing the U.S. economy. Operating such an IES is challenging due to its complexity and the diverse end-use demands, such as heating and electricity. This paper aims to address the optimal operation of an IES composed of a small modular reactor (SMR), a steam manifold, a balance of plant (BOP), a high-temperature steam electrolysis (HTSE) system, a district heating (DH) network, and electrical grids. We formulate the system’s operation as a mixed integer linear programming (MILP) problem to maximize net revenues from sales of electricity and hydrogen. To evaluate the efficacy of the proposed model, we conduct a 24-hour simulation considering day-ahead (DA) electricity prices from CAISO and a varying DH demand profile with hourly resolution. The simulation results show that our model effectively optimizes the operation by selling electricity during on-peak periods and purchasing electricity for hydrogen production during off-peak periods, while satisfying operating constraints within the IES.

08 HYDROGEN↗
DOE OSTI2025

Construction of 3D MHD pressure drop correlation and flow characterization in the contraction region of a fusion blanket manifold

Inlet and outlet manifolds are typical components of liquid metal (LM) blanket designs of a fusion power reactor to be used to distribute the LM flow into breeding channels and collect it at the exit of the blanket. High pressure loss in the magnetohydrodynamic (MHD) flows featuring abrupt geometrical changes is one of the main feasibility issues of such designs. Recently, optimization studies were conducted to construct 3D MHD pressure drop correlations for a LM flow in an electrically insulating manifold with gradual expansion. Here, the 3D computational approach developed in that study is applied to the outlet manifold featuring gradual contraction. A systematic analysis was performed with a total number of 135 flow cases computed with COMSOL Multiphysics for Hartmann numbers 1000 < Ha < 10,000, Reynolds numbers 100 < Re < 12,000, and contraction angles 45° < θ < 75° for a fixed contraction ratio of 4. The effects of Ha, Re and θ on the flow recirculation, development length and the total pressure drop were carefully examined. A linear regression analysis was used to determine the power rule of pressure drop coefficient k related to Ha and Re, demonstrating a good match with the Ludford layer theory. Eventually, a correlation for the 3D MHD pressure drop coefficient was constructed as a function of Ha, Re and θ. Further, the results were compared against the inlet manifold. It was found that the flow in the inlet manifold exhibits larger recirculation zones. In the investigated range of Ha, Re and θ, the pressure drop coefficient k of the LM MHD flow in the gradual contraction is only slightly lower (< 8 %) than that in the gradual expansion.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2025

Investigating the Combustion Performance of Dual Fuel Combustion with Diesel and Port Injected Hydrogen in a Large Bore Locomotive Engine

The heavy-duty transportation sector has primarily relied on conventional diesel combustion engines given their reliability and high thermal efficiency relative to spark ignition engines, but increased focus on reducing greenhouse gas emissions has led to investigation into alternative fuels. Gaseous hydrogen fuel has garnered a great deal of recent interest in the engine community given it has zero carbon, but hydrogen is not available at the scale and cost that petroleum fuels are currently available, and this is a barrier to adoption for industries that are looking to decarbonize their operations. Because of the fuel flexibility provided, dual fuel technology offers a pathway for some industries to adopt hydrogen as a fuel source while maintaining sufficient flexibility in times and locations where the new fuel is not yet available. This computational study investigates dual fuel combustion in a large bore locomotive engine architecture using direct injected diesel and port injected gaseous hydrogen fuel. With an optimal port fuel injection configuration from previous work, simulations of varying substitution ratio, compression ratio, manifold air temperature, diesel injection timing, and diesel injection pressure were performed to understand their effect on combustion performance. Results indicated that both increased substitution ratio and higher intake air temperature accelerates hydrogen flame propagation and can result in high peak cylinder pressures. Additionally, diesel injection timing and injection pressure were demonstrated as effective methods for controlling dual fuel combustion heat release rates.

ODonnell, Patrick Christopher↗