Search NASA⌕ Search

SEARCH · Search NASA

Results for “Robust 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 487 records · Page 27

Loss Landscape Analysis for Reliable Quantized ML Models for Scientific Sensing

In this paper, we propose a method to perform empirical analysis of the loss landscape of machine learning (ML) models. The method is applied to two ML models for scientific sensing, which necessitates quantization to be deployed and are subject to noise and perturbations due to experimental conditions. Our method allows assessing the robustness of ML models to such effects as a function of quantization precision and under different regularization techniques -- two crucial concerns that remained underexplored so far. By investigating the interplay between performance, efficiency, and robustness by means of loss landscape analysis, we both established a strong correlation between gently-shaped landscapes and robustness to input and weight perturbations and observed other intriguing and non-obvious phenomena. Our method allows a systematic exploration of such trade-offs a priori, i.e., without training and testing multiple models, leading to more efficient development workflows. This work also highlights the importance of incorporating robustness into the Pareto optimization of ML models, enabling more reliable and adaptive scientific sensing systems.

Baldi, Tommaso [Pisa, Scuola Normale Superiore]↗

Bell inequalities for continuous-variable measurements

Tests of local hidden-variable theories using measurements with continuous-variable (CV) outcomes are developed, and a comparison of different methods is presented. As examples, we focus on multipartite entangled Greenberger-Horne-Zeilinger and cluster states. We suggest a physical process that produces the states proposed here, and investigate experiments both with and without binning of the continuous variable. In the former case, the Mermin-Klyshko inequalities can be used directly. For unbinned outcomes, the moment-based Cavalcanti-Foster-Reid-Drummond inequalities are extended to functional inequalities by consideration of arbitrary functions of the measurements at each site. By optimizing these functions, we obtain more robust violations of local hidden-variable theories than with either binning or moments. Recent inequalities based on the algebra of quaternions and octonions are compared with these methods. Since the prime advantage of CV experiments is to provide a route to highly efficient detection via homodyne measurements, we analyze the effect of noise and detection losses in both binned and unbinned cases. The CV moment inequalities with an optimal function have greater robustness to both loss and noise. This could permit a loophole-free test of Bell inequalities.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Efficient energy transfer in light-harvesting systems: Quantum-classical comparison, flux network, and robustness analysis

Following the calculation of optimal energy transfer in thermal environment in our first paper [J. L. Wu, F. Liu, Y. Shen, J. S. Cao, and R. J. Silbey, New J. Phys. 12, 105012 (2010)], full quantum dynamics and leading-order 'classical' hopping kinetics are compared in the seven-site Fenna-Matthews-Olson (FMO) protein complex. The difference between these two dynamic descriptions is due to higher-order quantum corrections. Two thermal bath models, classical white noise (the Haken-Strobl-Reineker (HSR) model) and quantum Debye model, are considered. In the seven-site FMO model, we observe that higher-order corrections lead to negligible changes in the trapping time or in energy transfer efficiency around the optimal and physiological conditions (2% in the HSR model and 0.1% in the quantum Debye model for the initial site at BChl 1). However, using the concept of integrated flux, we can identify significant differences in branching probabilities of the energy transfer network between hopping kinetics and quantum dynamics (26% in the HSR model and 32% in the quantum Debye model for the initial site at BChl 1). This observation indicates that the quantum coherence can significantly change the distribution of energy transfer pathways in the flux network with the efficiency nearly the same. The quantum-classical comparison of the average trapping time with the removal of the bottleneck site, BChl 4, demonstrates the robustness of the efficient energy transfer by the mechanism of multi-site quantum coherence. To reconcile with the latest eight-site FMO model which is also investigated in the third paper [J. Moix, J. L. Wu, P. F. Huo, D. F. Coker, and J. S. Cao, J. Phys. Chem. Lett. 2, 3045 (2011)], the quantum-classical comparison with the flux network analysis is summarized in Appendix C. The eight-site FMO model yields similar trapping time and network structure as the seven-site FMO model but leads to a more disperse distribution of energy transfer pathways.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Robust implementation of generative modeling with parametrized quantum circuits

Although the performance of hybrid quantum-classical algorithms is highly dependent on the selection of the classical optimizer and the circuit ansätze, a robust and thorough assessment on-hardware of such features has been missing to date. From the optimizer perspective, the primary challenge lies in the solver’s stochastic nature, and their significant variance over the random initialization. Therefore, a robust comparison requires one to perform several training curves for each solver before one can reach conclusions about their typical performance. Since each of the training curves requires the execution of thousands of quantum circuits in the quantum computer, such a robust study remained a steep challenge for most hybrid platforms available today. In this work, we leverage on Rigetti’s Quantum Cloud Services (QCS™) to overcome this implementation barrier, and we study the on-hardware performance of the data-driven quantum circuit learning (DDQCL) for three different state-of-the-art classical solvers, and on two-different circuit ansätze associated to different entangling connectivity graphs for the same task. Additionally, we assess the gains in performance from varying circuit depths. To evaluate the typical performance associated with each of these settings in this benchmark study, we use at least five independent runs of DDQCL towards the generation of quantum generative models capable of capturing the patterns of the canonical Bars and Stripes dataset. In this experimental benchmarking, the gradient-free optimization algorithms show an outstanding performance compared to the gradient-based solver. In particular, one of them had better performance when handling the unavoidable noisy objective function to be minimized under experimental conditions.

97 MATHEMATICS AND COMPUTING↗

Development and application of a machine learning based multi-objective optimization workflow for CO 2 -EOR projects

Carbon dioxide-Enhanced Oil Recovery (CO 2 -EOR) is known as one of techniques for hydrocarbon production improvement as wells as an important candidate to reduce greenhouse gas emissions. Thus, an ideal development strategy for a CO 2 -EOR project would consider multiple objectives including to maximize oil recovery, CO 2 storage volume and project economic outcomes. This work proposes a robust computational framework that couples artificial neural network (ANN) and multi-objective optimizers to optimize the aforementioned objectives in CO 2 -EOR processes simultaneously. Expert ANN systems are trained and employed as surrogate models of the high-fidelity compositional simulator in the optimization workflow. The robustness of the development optimization protocol is confirmed via a synthetic injection-pattern-base case study. Afterward a field implementation to Morrow-B formation to optimize the tertiary recovery stage of the field development is discussed. This work compares the optimum solution found using an aggregate objective function and the solution repository (Pareto front) generated by the multi-objective optimization process. The comparison indicates the existence of potential multi-solutions satisfying certain criteria in a CO 2 -EOR project designing, which cannot be found using traditional weighted sum method. The optimization results provide significant insight into the decision-making process of CO 2 -EOR project when multiple objective functions are considered.

02 PETROLEUM↗

Sensitivity of joint contagiousness and susceptibility-based dynamic optimal control strategies for HIV prevention

Predicting the population-level effects of an infectious disease intervention that incorporate multiple modes of intervention is complicated by the joint non-linear dynamics of both infection transmission and the intervention itself. In this paper, we consider the sensitivity of Dynamic Optimal Control Profiles (DOCPs) for the optimal joint investment in both a contagiousness and susceptibility-based control of HIV to bio-behavioral, economic, and programmatic assumptions. The DOCP is calculated using recently developed numerical algorithms that allow controls to be represented by a set of piecewise constant functions that maintain a constant yearly budget. Our transmission model assumes multiple stages of HIV infection corresponding to acute and chronic infection and both within- and between-individual behavioral heterogeneity. We parameterize a baseline scenario from a longitudinal study of sexual behavior in MSM and consider sensitivity of the DOCPs to deviations from that baseline scenario. In the baseline scenario, the primary determinant of the dominant control were programmatic factors, regardless of budget. In sensitivity analyses, the qualitative aspects of the optimal control policy were often robust to significant deviation in assumptions regarding transmission dynamics. In addition, we found several conditions in which long-term joint investment in both interventions was optimal. Our results suggest that modeling in the service of decision support for intervention design can improve population-level effects of a limited set of economic resources. We found that economic and programmatic factors were as important as the inherent transmission dynamics in determining population-level intervention effects. Given our finding that the DOCPs were robust to alternative biological and behavioral assumptions it may be possible to identify DOCPs even when the data are not sufficient to identify a transmission model.

59 BASIC BIOLOGICAL SCIENCES↗

A Scalable Interior‐Point Gauss–Newton Method for PDE‐Constrained Optimization With Bound Constraints

Here, we present a scalable approach to solve a class of partial differential equation (PDE)‐constrained optimization problems with bound constraints. This approach utilizes a robust full‐space interior‐point (IP)‐Gauss–Newton optimization method. To cope with the poorly‐conditioned IP‐Gauss–Newton saddle‐point linear systems that need to be solved approximately, once per optimization step, we propose two spectrally related preconditioners. These preconditioners leverage the limited informativeness of data in regularized PDE‐constrained optimization problems. A block Gauss–Seidel preconditioner is proposed for the GMRES‐based solution of the IP‐Gauss–Newton linear systems. It is shown, for a large‐class of PDE‐ and bound‐constrained optimization problems, that the spectrum of the block Gauss–Seidel preconditioned IP‐Gauss–Newton matrix is asymptotically independent of discretization and is not impacted by the ill‐conditioning that notoriously plagues interior‐point methods. We exploit symmetry of the IP‐Gauss–Newton linear systems and propose a regularization and log‐barrier Hessian preconditioner for the preconditioned conjugate gradient (PCG)‐based solution of the equivalent IP‐Gauss–Newton–Schur complement linear systems. The eigenvalues of the block Gauss–Seidel preconditioned IP‐Gauss–Newton matrix, that are not equal to one, are identical to the eigenvalues of the regularization and log‐barrier Hessian preconditioned Schur complement matrix. The scalability of the approach is demonstrated on two example problems. The numerical solution of these optimization problems is shown to require a discretization independent number of IP‐Gauss–Newton linear solves. Furthermore, the linear systems are solved in a discretization and IP ill‐conditioning independent number of preconditioned Krylov subspace iterations. The parallel scalability of the preconditioner, achieved via algebraic multigrid component solvers when applicable, and the aforementioned algorithmic scalability permits a parallel scalable means to compute solutions of a large class of PDE‐ and bound‐constrained problems.

PDE-constrained optimization↗

A New Shape Optimization Approach for Lightweighting Electric Machines Inspired by Additive Manufacturing

Minimizing the mass in electric machines while maintaining superior performance is a new requirement for the advancement of drivetrains used in wind energy and electric mobility. Topology optimization (TO) for lightweighting electric machines using traditional approaches typically explores a restricted design space allowed by standard parametrizable geometry and manufacturing, while advanced methods, such as cell-based density approaches, suffer from a lack of robust manufacturability constraints during the optimization process. To overcome these drawbacks, we explore a grid-independent, boundary optimization where the outer shape of the magnet is parameterized using Bezier curves. We conduct a design of experiments (DOE) to study the effect of different magnet shapes on machine performance by varying the control points on the Bezier curves. A machine-learning-based surrogate model is constructed using the data from the DOE to quantify the relationship between the control points, air-gap torque, and mass. The control points are then optimized to maximize the torque density. The approach is used for minimizing electrical steel mass in the International Energy Agency (IEA) 15-MW radial flux direct-drive wind turbine generator. The new approach to shape optimization resulted in smooth and concise shapes that can be easily additively manufactured with up to a 20-ton reduction in electrical steel mass.

additive manufacturing↗

A Boundary Topology Optimization Approach for Lightweighting Electric Machines Inspired by Additive Manufacturing

Minimizing the mass in electric machines while maintaining superior performance has become a new requirement for the advancement of drivetrains used in wind energy and electric mobility. Topology optimization for light-weighting electric machines using traditional approaches typically explore a restricted design space allowed by standard parametrizable geometry and manufacturing, while advanced methods like cell-based density approaches suffer from a lack of robust manufacturability constraints during the optimization process. To overcome these drawbacks, we explore a grid-independent, boundary TO where the outer shape of the magnet is parameterized using Bezier curves. We conduct a design of experiments (DOE) to study the effect of different magnet shapes on machine performance by varying the control points on the Bezier curves. A machine-learning based surrogate model is constructed using the data from the DOE to quantify the relationship between the control points, air-gap torque and mass. The control points are then optimized to maximize the torque density. The approach is used for minimizing electrical steel mass in the IEA-15MW radial flux direct-drive wind turbine generator. The free-from boundary TO resulted in smooth and concise shapes that can be easily additively manufactured with upto 20-ton reduction in electrical steel mass.

additive manufacturing↗

A New Shape Optimization Approach for Lightweighting Electric Machines Inspired by Additive Manufacturing: Preprint

Minimizing the mass in electric machines while maintaining superior performance is a new requirement for the advancement of drivetrains used in wind energy and electric mobility. Topology optimization (TO) for lightweighting electric machines using traditional approaches typically explores a restricted design space allowed by standard parametrizable geometry and manufacturing, while advanced methods, such as cell-based density approaches, suffer from a lack of robust manufacturability constraints during the optimization process. To overcome these drawbacks, we explore a grid-independent, boundary TO where the outer shape of the magnet is parameterized using Bezier curves. We conduct a design of experiments (DOE) to study the effect of different magnet shapes on machine performance by varying the control points on the Bezier curves. A machine-learning-based surrogate model is constructed using the data from the DOE to quantify the relationship between the control points, air-gap torque, and mass. The control points are then optimized to maximize the torque density. The approach is used for minimizing electrical steel mass in the International Energy Agency (IEA) 15-MW radial flux direct-drive wind turbine generator. The free-form boundary TO resulted in smooth and concise shapes that can be easily additively manufactured with up to a 20-ton reduction in electrical steel mass.

additive manufacturing↗

Physics and material science of ultra-high quality factor superconducting resonator

The nitrogen doping into niobium superconducting radio frequency cavity walls aiming to improve the fundamental mode quality factor is the subject of the research in the given work. Quantitative nitrogen diffusion into niobium model calculating the concentration profile was developed. The model estimations were confirmed with secondary ion mass spectrometry technique measurements. The model made controlled nitrogen doping recipe optimization possible. As a result the robust reproducible recipe for SRF cavity walls treatment with nitrogen doping was developed. The cavities produced with optimized recipe met LCLS–II requirements on quality factor of 2.7 ∙ 10 10 at acceleration field of 16 MV/m. The microscopic effects of nitrogen doping on superconducting niobium properties were studied with low energy muon spin rotation technique and magnetometer measurements. No significant effect of nitrogen on the following features was found: electron mean free path, magnetic field penetration depth, and upper and surface critical magnetic fields. It was detected that for nitrogen doped niobium samples magnetic flux starts to penetrate inside the superconductor at lower external magnetic field value compared to the low temperature baked niobium ones. This explains lower quench field of SRF cavities treated with nitrogen. Quality factor improvement of fundamental mode forced to analyze the high order mode (HOM) impact on the particle beam dynamics. Both resonant and cumulative effects caused by monopole and dipole HOMs respectively are found to be negligible within the requirements for LCLS–II.

36 MATERIALS SCIENCE↗

Steps towards ‘drop-in’ biofuels: focusing on metabolic pathways

The past decade has witnessed rapid advance in microbial production of ‘drop-in’ biofuels from renewable resources. Various biosynthetic pathways have been constructed to produce biofuels with diverse structures, and multiple metabolic engineering strategies have been developed to increase biofuel titers, yields, productivities and system robustness. In this review, we intend to give a brief but comprehensive overview of the most recent progresses on four essential pathways leading to ‘drop-in’ biofuel production, with an emphasis on the metabolic pathway efficiencies and biofuel structures. Furthermore, we also provide an insightful discussion on optimization strategies to improve the robustness of the microbial platforms for biofuel production.

09 BIOMASS FUELS↗

An Antibody Cocktail-Based Immunoaffinity-LC-MS Method Enabled Ultra-Sensitive and Robust Quantification of Circulating Proinsulin Proteoforms and C-Peptide

Accurately measuring circulating proinsulin proteoforms is crucial for clinical investigation of diabetes, but was previously not feasible owing to limited assay specificity/sensitivity. Here, in this study, we devised a highly sensitive LC-MS-based strategy to quantify intact proinsulin, des-31,32 and des-64,65 proinsulin, and C-peptide in circulation. The method involves: (i) quantitative, robust affinity capture using an optimized antibody cocktail, eliminating the severe quantitative bias across multiple proteoforms typically introduced when using a single antibody; (ii) Lys-C digestion producing unique signature peptides for each proteoform, and (iii) trapping-nano-LC coupled with FAIMS/dCV-MS for an ultrasensitive analysis. The selective trapping/delivery ensured sensitive/selective analysis of the targets while achieving excellent analytical robustness that is critical for clinical assays, and the FAIMS/dCV substantially reduces baseline noise/interferences, further enhancing S/N. The assay achieved exceptional sensitivity, with serum LOQs of 1.7, 2.3, and 3.6 pg/mL respectively for intact-proinsulin, des-31,32 and des-64,65, representing the first assay capable of sensitively quantifying these major circulating proinsulin proteoforms. We applied this assay to 78 subjects, including autoantibody positive (n = 20) and new-onset type 1 diabetes (T1D, n = 19) with respective age/sex/BMI-matched controls, enabling the first accurate profiling of proinsulin proteoforms in clinical groups. The assay results demonstrated a clear separation of control and new-onset T1D groups that a parallel total-proinsulin ELISA assay fails to capture. Furthermore, distinct expression patterns in relative abundance ratios among proteoforms were observed across clinical groups. This assay may provide valuable insights into the β-cell functions and the onset/progression of diabetes and other associated conditions. Moreover, the strategy is broadly applicable to targeted measurement of other biomarker proteoforms.

Shen, Qingqing [University at Buffalo, NY (United ↗

Black-box optimization of CT acquisition and reconstruction parameters: a reinforcement learning approach

Protocol optimization is critical in Computed Tomography (CT) for achieving desired diagnostic image quality while minimizing radiation dose. Due to the inter-effect of influencing CT parameters, traditional optimization methods rely on the testing of exhaustive combinations of these parameters. This poses a notable limitation due to the impracticality of exhaustive parameter testing. This study introduces a novel methodology leveraging Virtual Imaging Trials (VITs) and reinforcement learning to more efficiently optimize CT protocols. Computational phantoms with liver lesions were imaged using a validated CT simulator and reconstructed with a novel CT reconstruction Toolkit. The optimization parameter space included tube voltage, tube current, reconstruction kernel, slice thickness, and pixel size. The optimization process was done using a Proximal Policy Optimization (PPO) agent which was trained to maximize the Detectability Index (d’) of the liver lesion for each reconstructed image. Results showed that our reinforcement learning approach found the absolute maximum d’ across the test cases while requiring 79.7% fewer steps compared to an exhaustive search, demonstrating both accuracy and computational efficiency, offering a efficient and robust framework for CT protocol optimization. The flexibility of the proposed technique allows for use of varying image quality metrics as the objective metric to maximize for. Our findings highlight the advantages of combining VIT and reinforcement learning for CT protocol management.

Fenwick, David [Duke University Medical Center]↗

DEVELOPMENT OF A SOFTWARE DESIGN TOOL FOR HYBRID SOLAR-GEOTHERMAL HEAT PUMP SYSTEMS IN HEATING- AND COOLING-DOMINATED BUILDINGS

This project provides an easy-to-use, menu-driven, software tool for designing hybrid solar-geothermal heat pump systems (GHP) for both heating- and cooling-dominated buildings. No such design tool currently exists. In heating-dominated buildings, the design approach takes advantage of glazed solar collectors to effectively balance the annual thermal loads on the ground with renewable solar energy. In cooling-dominated climates, the design approach takes advantage of relatively low-cost, unglazed solar collectors as the heat rejecting component. The primary benefit of hybrid GHPs is the reduced initial cost of the ground heat exchanger (GHX). Furthermore, solar thermal collectors can be used to balance the ground loads over the annual cycle, thus making the GHX fully sustainable; in heating-dominated buildings, the hybrid energy source (i.e., solar) is renewable, in contrast to a typical fossil fuel boiler or electric resistance as the hybrid component; in cooling-dominated buildings, use of unglazed solar collectors as a heat rejecter allows for passive heat rejection, in contrast to a cooling tower that consumes a significant amount of energy to operate, and hybrid GHPs can expand the market by allowing reduced GHX footprint in both heating- and cooling-dominated climates. The design tool allows for the straight-forward design of innovative GHP systems that currently pose a significant design challenge. The project lays the foundations for proper and reliable design of hybrid GHP systems, overcoming a series of difficult and cumbersome steps without the use of a system simulation approach, and without an automated optimization scheme. As new technologies and design concepts emerge, sophisticated design tools and methodologies must accompany them and be made usable for practitioners. Lack of reliable design tools results in reluctance of practitioners to implement more complex systems. A menu-driven software tool for the design of hybrid solar GHP systems is provided that is based on mathematically robust, validated models. An automated optimization tool is used to balance ground loads and incorporated into the simulation engine. With knowledge of the building loads, thermal properties of the ground, the borehole heat exchanger configuration, the heat pump peak hourly and seasonal COP for heating and cooling, the critical heat pump design entering fluid temperature, and the thermal performance of a solar collector, the total GHX length can be calculated along with the area of a supplemental solar collector array and the corresponding reduced GHX length. An economic analysis module allows for the calculation of the lowest capital cost combination of solar collector area and GHX length. ACKNOWLEDGMENTS This project was funded by the United States Department of Energy DOE-DE-FOA-0000116, Recovery Act Geothermal Technologies Program: Ground Source Heat Pumps. The lead contractor, The University of Hartford, was supported by The University of Dayton, and the Oak Ridge National Laboratories. All funding and support for this project as well as contributions of graduate and undergraduate students from the contributing institutions are gratefully acknowledged.

14 SOLAR ENERGY↗