Search NASA⌕ Search

SEARCH · Search NASA

Results for “augmented random search”

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.

9 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Physics Informed Reinforcement Learning for Power Grid Control using Augmented Random Search

Wide adoption of deep reinforcement learning need to overcome several challenges in energy system domain, including scalability, learning from limited samples, and high-dimensional continuous state and action spaces. In this paper, we integrated physics-based information from the normal generator operation state formula in the reinforcement learning agent's neural network loss function, and applied an augmented random search agent to optimize the generator control under dynamic contingency. Simulation results demonstrated the reliability performance improvements in training speed, reward convergence, sampling efficiency, scalability, and transferability.

physics informed ML, Physics Informed Neural Netwo↗
DOE OSTI2023

Efficient learning of power grid voltage control strategies via model-based deep reinforcement learning

Here this article proposes a model-based deep reinforcement learning (DRL) method to design emergency control strategies for short-term voltage stability problems in power systems. Recent advances show promising results for model-free DRL-based methods in power systems control problems. But in power systems applications, these model-free methods have certain issues related to training time (clock time) and sample efficiency; both are critical for making state-of-the-art DRL algorithms practically applicable. DRL-agent learns an optimal policy via a trial-and-error method while interacting with the real-world environment. It is also desirable to minimize the direct interaction of the DRL agent with the real-world power grid due to its safety-critical nature. Additionally, the state-of-the-art DRL-based policies are mostly trained using a physics-based grid simulator where dynamic simulation is computationally intensive, lowering the training efficiency. We propose a novel model-based DRL framework where a deep neural network (DNN)-based dynamic surrogate model (SM), instead of a real-world power grid or physics-based simulation, is utilized within the policy learning framework, making the process faster and more sample efficient. However, having stable training in model-based DRL is challenging because of the complex system dynamics of large-scale power systems. We addressed these issues by incorporating imitation learning to have a warm start in policy learning, reward-shaping, and multi-step loss in surrogate model training. Finally, we achieved 97.5% reduction in samples and 87.7% reduction in training time for an application to the IEEE 300-bus test system.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2025

Application of Machine Learning and Data Augmentation Algorithms in the Discovery of Metal Hydrides for Hydrogen Storage

The development of efficient and sustainable hydrogen storage materials is a key challenge for realizing hydrogen as a clean and flexible energy carrier. Among various options, metal hydrides offer high volumetric storage density and operational safety, yet their application is limited by thermodynamic, kinetic, and compositional constraints. In this work, we investigate the potential of machine learning (ML) to predict key thermodynamic properties—equilibrium plateau pressure, enthalpy, and entropy of hydride formation—based solely on alloy composition using Magpie-generated descriptors. We significantly expand an existing experimental dataset from ~400 to 806 entries and assess the impact of dataset size and data augmentation, using the PADRE algorithm, on model performance. Models including Support Vector Machines and Gradient Boosted Random Forests were trained and optimized via grid search and cross-validation. Results show a marked improvement in predictive accuracy with increased dataset size, while data augmentation benefits are limited to smaller datasets and do not improve accuracy in underrepresented pressure regimes. Furthermore, clustering and cross-validation analyses highlight the limited generalizability of models across different material classes, though high accuracy is achieved when training and testing within a single hydride family (e.g., AB2). The study demonstrates the viability and limitations of ML for accelerating hydride discovery, emphasizing the importance of dataset diversity and representation for robust property prediction.

augmentation↗
DOE OSTI2022

An adaptive stochastic sequential quadratic programming with differentiable exact augmented lagrangians

In this study, we consider solving nonlinear optimization problems with a stochastic objective and deterministic equality constraints. We assume for the objective that its evaluation, gradient, and Hessian are inaccessible, while one can compute their stochastic estimates by, for example, subsampling. We propose a stochastic algorithm based on sequential quadratic programming (SQP) that uses a differentiable exact augmented Lagrangian as the merit function. To motivate our algorithm design, we first revisit and simplify an old SQP method Lucidi developed for solving deterministic problems, which serves as the skeleton of our stochastic algorithm. Based on the simplified deterministic algorithm, we then propose a non-adaptive SQP for dealing with stochastic objective, where the gradient and Hessian are replaced by stochastic estimates but the stepsizes are deterministic and prespecified. Finally, we incorporate a recent stochastic line search procedure Paquette and Scheinberg into the non-adaptive stochastic SQP to adaptively select the random stepsizes, which leads to an adaptive stochastic SQP. The global "almost sure" convergence for both non-adaptive and adaptive SQP methods is established. Numerical experiments on nonlinear problems in CUTEst test set demonstrate the superiority of the adaptive algorithm.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

Multifunctional electrochemical memory stabilized by phase coexistence

Our growing computing needs, especially in applications that heavily rely on artificial intelligence (AI), motivate a search for new components that could substantially augment the performance of general-purpose digital computers. Beyond ON/OFF switching, new components with linear multistate analog resistive tuning, nonlinear volatile switching, spiking, oscillatory, stochastic and other complex functionalities could enable highly efficient neuromorphic computing schemes for AI information processing. Compared to the extreme multifunctionality of biological neurons, realizing all the above characteristics in a single, scalable analog component remains a grand challenge. Here we investigate electrochemical gating combined with localized thermal activation to program and switch a single, vertically integrated and dimensionally scaled electrothermal chemical random access memory (ETCRAM) with a channel and reservoir composed of phase-separated vanadium oxide. Closely related to electrochemical RAM (ECRAM), ETCRAM uses an integrated gate-heater electrode to overcome kinetic barriers that help retain states at ambient temperatures. In addition to synapse-like stable and programmable analog resistance states arising from redox-tunable phase coexistence, a single component exhibits neuron-like nonlinear conductance switching with a tunable threshold and self-driven dynamics owing to the thermally driven metal-insulator phase transition in vanadium dioxide. More broadly, we demonstrate that electrochemically stabilized phase coexistence could unlock analog electronics with novel functionality, stability, reconfigurability, and scalability.

Oh, Sangheon [Sandia National Lab. (SNL-CA), Liver↗
DOE OSTI2021

Stopping criteria for ending autonomous, single detector radiological source searches

While the localization of radiological sources has traditionally been handled with statistical algorithms, such a task can be augmented with advanced machine learning methodologies. The combination of deep and reinforcement learning has provided learning-based navigation to autonomous, single-detector, mobile systems. However, these approaches lacked the capacity to terminate a surveying/search task without outside influence of an operator or perfect knowledge of source location (defeating the purpose of such a system). Two stopping criteria are investigated in this work for a machine learning navigated system: one based upon Bayesian and maximum likelihood estimation (MLE) strategies commonly used in source localization, and a second providing the navigational machine learning network with a “stop search” action. A convolutional neural network was trained via reinforcement learning in a 10 m × 10 m simulated environment to navigate a randomly placed detector-agent to a randomly placed source of varied strength (stopping with perfect knowledge during training). The network agent could move in one of four directions (up, down, left, right) after taking a 1 s count measurement at the current location. During testing, the stopping criteria for this navigational algorithm was based upon a Bayesian likelihood estimation technique of source presence, updating this likelihood after each step, and terminating once the confidence of the source being in a single location exceeded 0.9. A second network was trained and tested with similar architecture as the previous but which contained a fifth action: for self-stopping. The accuracy and speed of localization with set detector and source initializations were compared over 50 trials of MLE-Bayesian approach and 1000 trials of the CNN with self-stopping. The statistical stopping condition yielded a median localization error of ~1.41 m and median localization speed of 12 steps. The machine learning stopping condition yielded a median localization error of 0 m and median localization speed of 17 steps. This work demonstrated two stopping criteria available to a machine learning guided, source localization system.

38 RADIATION CHEMISTRY, RADIOCHEMISTRY, AND NUCLEA↗
NASA NTRSDate not supplied

Augmenting Topic Finding in the NASA Aviation Safety Reporting System using Topic Modeling

Context: The NASA Aviation and Safety Reporting System (ASRS) provides various publications to the aviation community (including individual anonymous reports, Callback, Database Search Requests, Directline, and Alerting Messages). Key to these publications are the timely processing of new reports, which is currently done mostly manually by ASRS staff, and which the volume increases yearly. Aim: We investigate whether existing topic modelling techniques are suitable to ease some of the manual effort, and to enhance it with additional visual cues regarding the process of grouping, sense making and labeling incoming (and previous) reports. Method: We evaluate the applicability of WarpLDA topic modelling results combined with three visualization tools, the first two of which have been extended by us in this work for ASRS: Termite, TopicFlow, and LDAVis. Based on the identified limitations in these tools, we propose a methodology for improving them, and evaluate their outputs using ASRS as our test dataset. Results: The user interfaces of Termite, Topicflow and LDAVis were found insufficient for sense-making of the narratives. Moreover, concerns regarding the stability of results due to the inherent randomness of topic modelling, and the lack of a measurable approach for evaluation against the existing ASRS manual workflow were also noted. Conclusion: While many tools to topic modeling and visualization have been proposed, more work is necessary before they can be applied in practical situations to improve existing manual workflows. The methodology presented and applied in this work contribute towards this effort.

ASRS↗
DOE OSTI2020

Enhancing nuclear data validation analysis by using machine learning

In this work, we demonstrate how machine learning (ML) techniques can be used as an auxiliary tool for nuclear data validation analysis. The output of the ML analysis can inform evaluators and validators on the quality, or lack thereof, of specific nuclear data and benchmark experiments with respect to simulating these experimental benchmark values. To this end, measured and simulated effective neutron multiplication factors, k eff , of 875 selected ICSBEP critical assemblies and the latter's sensitivities with respect to nuclear data as well as benchmarks' features (e.g., material nuclide, core geometry) are used as input for a random forest (RF) regression model. The RF is used to encode the complex inter-dependencies between thousands of nuclear data features (i.e., sensitivity profiles and aspects describing the measurements) and 875 simulated and experimental k eff values in order to understand what nuclear data features are most informative for predicting bias. The complexity of relationships and high-dimensional space is difficult-to-impossible to search fully using simply expert judgment. As a first proof-of-concept—step, it is demonstrated that this technique is able to correctly trace large discrepancies between simulated and experimental k eff back to fabricated shortcomings in nuclear data that were perturbed to simulated k eff values. In a second, real-case scenario, step, the RF algorithm is used to validate the ENDF/B-VIII.0 library in comparison to ENDF/B-VII.1 nuclear data. One case is showcased where the chosen ML algorithms highlighted nuclear data (the 19 F(n,inl) cross section from 0.4–0.9 MeV) that are shown to be problematic by comparing them to associated differential experimental data and nuclear data from other libraries. In addition to that, it is shown that the RF results point towards poor benchmark experiments and associated underestimated uncertainties (e.g., the PU-SOL-THERM-028 series). However, using the RF algorithm for validating nuclear data with respect to k eff is currently limited to pinpointing groups of questionable nuclear data due to the inherent correlations between features introduced by the nuclear data themselves and how k eff is simulated. Due to this, we recommend that the ML methods presented be used to augment—rather than replace—the expert knowledge of evaluators and validators.

96 KNOWLEDGE MANAGEMENT AND PRESERVATION↗