Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “adaptive learning”

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 1,225 records · Page 68

On-the-fly active learning of interpretable Bayesian force fields for atomistic rare events

Abstract Machine learned force fields typically require manual construction of training sets consisting of thousands of first principles calculations, which can result in low training efficiency and unpredictable errors when applied to structures not represented in the training set of the model. This severely limits the practical application of these models in systems with dynamics governed by important rare events, such as chemical reactions and diffusion. We present an adaptive Bayesian inference method for automating the training of interpretable, low-dimensional, and multi-element interatomic force fields using structures drawn on the fly from molecular dynamics simulations. Within an active learning framework, the internal uncertainty of a Gaussian process regression model is used to decide whether to accept the model prediction or to perform a first principles calculation to augment the training set of the model. The method is applied to a range of single- and multi-element systems and shown to achieve a favorable balance of accuracy and computational efficiency, while requiring a minimal amount of ab initio training data. We provide a fully open-source implementation of our method, as well as a procedure to map trained models to computationally efficient tabulated force fields.

36 MATERIALS SCIENCE↗

Multi-Objective Reinforcement Learning-based Deep Neural Networks for Cognitive Space Communications

Future communication subsystems of space exploration missions can potentially benefit from software-defined radios (SDRs) controlled by machine learning algorithms. In this paper, we propose a novel hybrid radio resource allocation management control algorithm that integrates multi-objective reinforcement learning and deep artificial neural networks. The objective is to efficiently manage communications system resources by monitoring performance functions with common dependent variables that result in conflicting goals. The uncertainty in the performance of thousands of different possible combinations of radio parameters makes the trade-off between exploration and exploitation in reinforcement learning (RL) much more challenging for future critical space-based missions. Thus, the system should spend as little time as possible on exploring actions, and whenever it explores an action, it should perform at acceptable levels most of the time. The proposed approach enables on-line learning by interactions with the environment and restricts poor resource allocation performance through virtual environment exploration. Improvements in the multiobjective performance can be achieved via transmitter parameter adaptation on a packet-basis, with poorly predicted performance promptly resulting in rejected decisions. Simulations presented in this work considered the DVB-S2 standard adaptive transmitter parameters and additional ones expected to be present in future adaptive radio systems. Performance results are provided by analysis of the proposed hybrid algorithm when operating across a satellite communication channel from Earth to GEO orbit during clear sky conditions. The proposed approach constitutes part of the core cognitive engine proof-of-concept to be delivered to the NASA Glenn Research Center SCaN Testbed located onboard the International Space Station.

space archtiecture↗

Near-Net Forging Technology Demonstration Program

Significant advantages in specific mechanical properties, when compared to conventional aluminum (Al) alloys, make aluminum-lithium (Al-Li) alloys attractive candidate materials for use in cryogenic propellant tanks and dry bay structures. However, the cost of Al-Li alloys is typically five times that of 2219 aluminum. If conventional fabrication processes are employed to fabricate launch vehicle structure, the material costs will restrict their utilization. In order to fully exploit the potential cost and performance benefits of Al-Li alloys, it is necessary that near-net manufacturing methods be developed to off-set or reduce raw material costs. Near-net forging is an advanced manufacturing method that uses elevated temperature metal movement (forging) to fabricate a single piece, near-net shape, structure. This process is termed 'near-net' because only a minimal amount of post-forge machining is required. The near-net forging process was developed to reduce the material scrap rate (buy-to-fly ratio) and fabrication costs associated with conventional manufacturing methods. The goal for the near-net forging process, when mature, is to achieve an overall cost reduction of approximately 50 percent compared with conventional manufacturing options for producing structures fabricated from Al-Li alloys. This NASA Marshall Space Flight Center (MSFC) sponsored program has been a part of a unique government / industry partnership, coordinated to develop and demonstrate near-net forging technology. The objective of this program was to demonstrate scale-up of the near-net forging process. This objective was successfully achieved by fabricating four integrally stiffened, 170- inch diameter by 20-inch tall, Al-Li alloy 2195, Y-ring adapters. Initially, two 2195 Al-Li ingots were converted and back extruded to produce four cylindrical blockers. Conventional ring rolling of the blockers was performed to produce ring preforms, which were then contour ring rolled to produce 'contour preforms'. All of the contour preforms on this first-of-a-kind effort were imperfect, and the ingot used to fabricate two of the preforms was of an earlier vintage. As lessons were learned throughout the program, the tooling and procedures evolved, and hence the preform quality. Two of the best contour preforms were near- net forged to produce a process pathfinder Y-ring adapter and a 'mechanical properties pathfinder' Y-ring adapter. At this point, Lockheed Martin Astronautics elected to procure additional 2195 aluminum-lithium ingot of the latest vintage, produce two additional preforms, and substitute them for older vintage material non-perfectly filled preforms already produced on this contract. The existing preforms could have been used to fulfill the requirements of the contract.

Hall, I. Keith↗

Multi-Objective Reinforcement Learning-Based Deep Neural Networks for Cognitive Space Communications

Future communication subsystems of space exploration missions can potentially benefit from software-defined radios (SDRs) controlled by machine learning algorithms. In this paper, we propose a novel hybrid radio resource allocation management control algorithm that integrates multi-objective reinforcement learning and deep artificial neural networks. The objective is to efficiently manage communications system resources by monitoring performance functions with common dependent variables that result in conflicting goals. The uncertainty in the performance of thousands of different possible combinations of radio parameters makes the trade-off between exploration and exploitation in reinforcement learning (RL) much more challenging for future critical space-based missions. Thus, the system should spend as little time as possible on exploring actions, and whenever it explores an action, it should perform at acceptable levels most of the time. The proposed approach enables on-line learning by interactions with the environment and restricts poor resource allocation performance through virtual environment exploration. Improvements in the multiobjective performance can be achieved via transmitter parameter adaptation on a packet-basis, with poorly predicted performance promptly resulting in rejected decisions. Simulations presented in this work considered the DVB-S2 standard adaptive transmitter parameters and additional ones expected to be present in future adaptive radio systems. Performance results are provided by analysis of the proposed hybrid algorithm when operating across a satellite communication channel from Earth to GEO orbit during clear sky conditions. The proposed approach constitutes part of the core cognitive engine proof-of-concept to be delivered to the NASA Glenn Research Center SCaN Testbed located onboard the International Space Station.

space archtiecture↗

Hardware-accelerated inference for real-time gravitational-wave astronomy

The field of transient astronomy has seen a revolution with the first gravitational-wave detections and the arrival of multi-messenger observations they enabled. Transformed by the first detection of binary black hole and binary neutron star mergers, computational demands in gravitational-wave astronomy are expected to grow by at least a factor of two over the next five years as the global network of kilometer-scale interferometers are brought to design sensitivity. With the increase in detector sensitivity, real-time delivery of gravitational-wave alerts will become increasingly important as an enabler of multi-messenger followup. In this work, we report a novel implementation and deployment of deep learning inference for real-time gravitational-wave data denoising and astrophysical source identification. This is accomplished using a generic Inference-as-a-Service model that is capable of adapting to the future needs of gravitational-wave data analysis. Overall, our implementation allows seamless incorporation of hardware accelerators and also enables the use of commercial or private (dedicated) as-a-service computing. Based on our results, we propose a paradigm shift in low-latency and offline computing in gravitational-wave astronomy. Such a shift can address key challenges in peak-usage, scalability and reliability, and provide a data analysis platform particularly optimized for deep learning applications. The achieved sub-millisecond scale latency will also be relevant for any machine learning-based real-time control systems that may be invoked in the operation of near-future and next generation ground-based laser interferometers, as well as the front-end collection, distribution and processing of data from such instruments.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

State Predictor of Classification Cognitive Engine Applied to Channel Fading

This study presents the application of machine learning (ML) to a space-to-ground communication link, showing how ML can be used to detect the presence of detrimental channel fading. Using this channel state information, the communication link can be used more efficiently by reducing the amount of lost data during fading. The motivation for this work is based on channel fading observed during on-orbit operations with NASA's Space Communication and Navigation (SCaN) testbed on the International Space Station (ISS). This paper presents the process to extract a target concept (fading and not-fading) from the raw data. The pre-processing and data exploration effort is explained in detail, with a list of assumptions made for parsing and labelling the dataset. The model selection process is explained, specifically emphasizing the benefits of using an ensemble of algorithms with majority voting for binary classification of the channel state. Experimental results are shown, highlighting how an end-to-end communication system can utilize knowledge of the channel fading status to identity fading and take appropriate action. With a laboratory testbed to emulate channel fading, the overall performance is compared to standard adaptive methods without fading knowledge, such as adaptive coding and modulation.

Fading↗

Best Practice for the Commercial Deployment of Carbon Dioxide Geologic Storage: Adaptive Management Approach

The Plains CO 2 Reduction (PCOR) Partnership is one of seven regional partnerships formed as part of the U.S. Department of Energy’s National Energy Technology Laboratory Regional Carbon Sequestration Partnerships (RCSP) Initiative. The RCSP Initiative is focused on the safe and long-term storage of CO 2 to support the commercial deployment of carbon capture and storage (CCS). To that end, the PCOR Partnership has spent over 10 years developing, testing, and validating the best methods and technologies to conduct the geologic storage of CO 2 (hereafter referred to as CO 2 storage). Through this effort, the PCOR Partnership has formalized an adaptive management approach (AMA) for the commercial development of CO 2 storage projects. The use of this approach, which draws upon the collective experience and lessons learned from the PCOR Partnership, represents best practice for advancing CO 2 storage projects toward commercial deployment. At the heart of the AMA are four technical elements necessary for any successful CO 2 storage project: 1) site characterization; 2) modeling and simulation; 3) risk assessment; and 4) monitoring, verification, and accounting (MVA). Each of these elements plays a key role in gathering and assessing site-specific data that provide a fundamental understanding of the storage complex and its performance. While each of the four technical elements can provide useful data independently, integrating them through the AMA yields a streamlined, fit-for-purpose strategy for the commercial deployment of CO 2 storage. Key to this integration and resulting best practice are feedback loops that allow the results of each element to serve as inputs to the others. Each iteration of the AMA creates an improved understanding of the storage complex and thus more targeted and efficient applications of the technical elements. For the purpose of establishing an adaptive management framework, hard lines have been drawn between the technical elements of the AMA. However, in practice, the rapid and seamless interaction between the elements can blur these lines. For example, to aid in the analysis and interpretation of site characterization data, a static geocellular model is often required. While this model development is part of the technical element, modeling and simulation, it is an integral part of the site characterization effort. Likewise, much of the monitoring data collected as part of the MVA technical element can be used to inform site characterization. This back-and-forth flow of data and use of models between the technical elements continues throughout the project.

54 ENVIRONMENTAL SCIENCES↗

Gait Training Improves Performance in Healthy Adults Exposed to Novel Discordant Conditions

After they return to Earth, astronauts experience sensorimotor disturbances that disrupt their ability to walk. We have previously shown that training with a variety of sensorimotor adaptive challenges enhances the capability of adapting to novel sensorimotor conditions. We are currently developing a sensorimotor adaptability (SA) training program designed to facilitate recovery of function after gravitational transitions. The purpose of this study was to determine whether trained subjects could transfer learned skills from one discordant visuo-proprioceptive environment to another. During three sessions, subjects walked at 2.5 km/h on a treadmill mounted on a motion base platform. Ten subjects trained with a combination of lateral treadmill translation and superimposed sinusoidal lateral optic flow that was presented on a large screen positioned in front of them. Ten controls completed the same training schedule while viewing only the forward optic flow with no visual or physical oscillation. Twenty minutes after the final training session, all subjects completed a 2-minute trial with a novel combination of visual and treadmill roll perturbations not previously experienced during the training (Transfer Test). Compared to the untrained group, participants who received SA training showed faster reaction times and, based on a composite score derived from stride frequency, heart rate, and reaction time, an overall enhanced performance. Our results showed that an SA training program can improve overall walking performance when subjects are exposed to novel incongruent sensory environments. This training has application for both enhancing adaptive responses in astronauts and reducing fall and injury risk in the elderly.

Batson, Crystal D.↗

Learning to Simulate Aerosol Dynamics with Graph Neural Networks

Aerosol effects on climate, weather, and air quality depend on characteristics of individual particles, which are tremendously diverse and change in time. Particle-resolved models are the only models able to capture this diversity in particle physiochemical properties, and these models are computationally expensive. As a strategy for accelerating particle-resolved microphysics models, we introduce Graph-based Learning of Aerosol Dynamics (GLAD) and use this model to train a surrogate of the particle-resolved model PartMC-MOSAIC. GLAD implements a Graph Network-based Simulator (GNS), a machine learning framework that has been used to simulate particle-based fluid dynamics models. In GLAD, each particle is represented as a node in a graph, and the evolution of the particle population over time is simulated through learned message passing. Here, we demonstrate our GNS approach on a simple aerosol system that includes condensation of sulfuric acid onto particles composed of sulfate, black carbon, organic carbon, and water. A graph with particles as nodes is constructed, and a graph neural network (GNN) is then trained using the model output from PartMC-MOSAIC. The trained GNN can then be used for simulating and predicting aerosol dynamics over time. Results demonstrate the framework's ability to accurately learn chemical dynamics and generalize across different scenarios, achieving efficient training and prediction times. We evaluate the performance across four scenarios, highlighting the framework's robustness and adaptability in modeling aerosol microphysics and chemistry.

aerosol chemistry dynamics↗

Deep Learning for Spectral Filling in Radio Frequency Applications

Due to the Internet of Things (IoT) proliferation, Radio Frequency (RF) channels are increasingly congested with new kinds of devices, which carry unique and diverse communication needs. This poses complex challenges in modern digital communications, and calls for the development of technological innovations that (i) optimize capacity (bitrate) in limited bandwidth environments, (ii) integrate cooperatively with already-deployed RF protocols, and (iii) are adaptive to the ever-changing demands in modern digital communications. In this paper we present methods for applying deep neural networks for spectral filling. Given an RF channel transmitting digital messages with a pre-established modulation scheme, we automatically learn novel modulation schemes for sending extra information, in the form of additional messages, “around” the fixed-modulation signals (i.e., without interfering with them). In so doing, we effectively increase channel capacity without increasing bandwidth. We further demonstrate the ability to generate signals that closely resemble the original modulations, such that the presence of extra messages is undetectable to third-party listeners. We present three computational experiments demonstrating the efficacy of our methods, and conclude by discussing the implications of our results for modern RF applications.

Setzler, Matthew D.↗

Analyzing inference workloads for spatiotemporal modeling

Ensuring power grid resiliency, forecasting climate conditions, and optimization of transportation infrastructure are some of the many application areas where data is collected in both space and time. Spatiotemporal modeling is about modeling those patterns for forecasting future trends and carrying out critical decision-making by leveraging machine learning/deep learning. Once trained offline, field deployment of trained models for near real-time inference could be challenging because performance can vary significantly depending on the environment, available compute resources and tolerance to ambiguity in results. Users deploying spatiotemporal models for solving complex problems can benefit from analytical studies considering a plethora of system adaptations to understand the associated performance-quality trade-offs. To facilitate the co-design of next-generation hardware architectures for field deployment of trained models, it is critical to characterize the workloads of these deep learning (DL) applications during inference and assess their computational patterns at different levels of the execution stack. In this paper, we develop several variants of deep learning applications that use spatiotemporal data from dynamical systems. We study the associated computational patterns for inference workloads at different levels, considering relevant models (Long short-term Memory, Convolutional Neural Network and Spatio-Temporal Graph Convolution Network), DL frameworks (Tensorflow and PyTorch), precision (FP16, FP32, AMP, INT16 and INT8), inference runtime (ONNX and AI Template), post-training quantization (TensorRT) and platforms (Nvidia DGX A100 and Sambanova SN10 RDU). Overall, our findings indicate that although there is potential in mixed-precision models and post-training quantization for spatiotemporal modeling, extracting efficiency from contemporary GPU systems might be challenging. Instead, co-designing custom accelerators by leveraging optimized High Level Synthesis frameworks (such as SODA High-Level Synthesizer for customized FPGA/ASIC targets) can make workload-specific adjustments to enhance the efficiency.

97 MATHEMATICS AND COMPUTING↗

How Well Does Kohn–Sham Regularizer Work for Weakly Correlated Systems?

Kohn–Sham regularizer (KSR) is a differentiable machine learning approach to finding the exchange-correlation functional in Kohn–Sham density functional theory that works for strongly correlated systems. Here we test KSR for a weak correlation. We propose spin-adapted KSR (sKSR) with trainable local, semilocal, and nonlocal approximations found by minimizing density and total energy loss. We assess the atoms-to-molecules generalizability by training on one-dimensional (1D) H, He, Li, Be, and Be 2+ and testing on 1D hydrogen chains, LiH, BeH 2 , and helium hydride complexes. The generalization error from our semilocal approximation is comparable to other differentiable approaches, but our nonlocal functional outperforms any existing machine learning functionals, predicting ground-state energies of test systems with a mean absolute error of 2.7 mH.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Adaptive Hierarchical Cyber Attack Detection and Localization in Active Distribution Systems

Development of a cyber security strategy for the active distribution systems is challenging due to the inclusion of distributed renewable energy generations. Here this paper proposes an adaptive hierarchical cyber attack detection and localization framework for distributed active distribution systems via analyzing electrical waveforms. Cyber attack detection is based on a sequential deep learning model, via which even minor cyber attacks can be identified. The two-stage cyber attack localization algorithm first estimates the cyber attack sub-region, and then localize the specified cyber attack within the estimated subregion. We propose a modified spectral clustering-based network partitioning method for the hierarchical cyber attack ‘coarse’ localization. Next, to further narrow down the cyber attack location, a normalized impact score based on waveform statistical metrics is proposed to obtain a ‘fine’ cyber attack location by characterizing different waveform properties. Finally, compared with classical and state-of-art methods, a comprehensive quantitative evaluation with two case studies shows promising estimation results of the proposed framework.

42 ENGINEERING↗

Crew Roles and Interactions in Scientific Space Exploration

Future piloted space exploration missions will focus more on science than engineering, a change which will challenge existing concepts for flight crew tasking and demand that participants with contrasting skills, values, and backgrounds learn to cooperate as equals. In terrestrial space flight analogs such as Desert Research And Technology Studies, engineers, pilots, and scientists can practice working together, taking advantage of the full breadth of all team members training to produce harmonious, effective missions that maximize the time and attention the crew can devote to science. This paper presents, in a format usable as a reference by participants in the field, a successfully tested crew interaction model for such missions. The model builds upon the basic framework of a scientific field expedition by adding proven concepts from aviation and human spaceflight, including expeditionary behavior and cockpit resource management, cooperative crew tasking and adaptive leadership and followership, formal techniques for radio communication, and increased attention to operational considerations. The crews of future spaceflight analogs can use this model to demonstrate effective techniques, learn from each other, develop positive working relationships, and make their expeditions more successful, even if they have limited time to train together beforehand. This model can also inform the preparation and execution of actual future spaceflights.

Love, Stanley G.↗

An adaptive, data-driven multiscale approach for dense granular flows

The accuracy of coarse-grained continuum models of dense granular flows is limited by the lack of high-fidelity closure models for granular rheology. One approach to addressing this issue, referred to as the hierarchical multiscale method, is to use a high-fidelity fine-grained model to compute the closure terms needed by the coarse-grained model. The difficulty with this approach is that the overall model can become computationally intractable due to the high computational cost of the high-fidelity model. In this work, we describe a multiscale modeling approach for dense granular flows that utilizes neural networks trained using high-fidelity discrete element method (DEM) simulations to approximate the constitutive granular rheology for a continuum incompressible flow model. Our approach leverages an ensemble of neural networks to estimate predictive uncertainty that allows us to determine whether the rheology at a given point is accurately represented by the neural network model. Additional DEM simulations are only performed when needed, minimizing the number of additional DEM simulations required when updating the rheology. This adaptive coupling significantly reduces the overall computational cost of the approach while controlling the error. In addition, the neural networks are customized to learn regularized rheological behavior to ensure well-posedness of the continuum solution. We first validate the approach using two-dimensional steady-state and decelerating inclined flows. We then demonstrate the efficiency of our approach by modeling three-dimensional sub-aerial granular column collapse for varying initial column aspect ratios, where our multiscale method compares well with the computationally expensive computational fluid dynamics (CFD)-DEM simulation.

Dense granular flows↗

First responder training: Supporting commercialization of hydrogen and fuel cell technologies

A properly trained first responder community is critical to the successful introduction of hydrogen fuel cell applications and their transformation in how we use energy. Providing resources with accurate information and current knowledge is essential to the delivery of effective hydrogen and fuel cell-related first responder training. Furthermore, the California Fuel Cell Partnership and the Pacific Northwest National Laboratory have over 15 years of experience in developing and delivering hydrogen safety-related first responder training materials and programs. A National Hydrogen and Fuel Cell Emergency Response Training Resource was recently released (http://h2tools.org/fr/nt/). This training resource serves the delivery of a variety of training regimens. Associated materials are adaptable for different training formats, ranging from high-level overview presentations to more comprehensive classroom training. Our paper presents what has been learned from the development and delivery of hydrogen safety-related first responder training programs (online, classroom, hands-on) by the respective organizations. We discussed the collaborative strategy being developed for enhancing training materials and methods for greater accessibility based on stakeholder input.

25 ENERGY STORAGE↗

Artificial intelligence techniques for integrative structural biology of intrinsically disordered proteins

Here, we outline recent developments in artificial intelligence (AI) and machine learning (ML) techniques for integrative structural biology of intrinsically disordered proteins (IDP) ensembles. IDPs challenge the traditional protein structure-function paradigm by adapting their conformations in response to specific binding partners leading them to mediate diverse, and often complex cellular functions such as biological signaling, self-organization and compartmentalization. Obtaining mechanistic insights into their function can therefore be challenging for traditional structural determination techniques. Often, scientists have to rely on piecemeal evidence drawn from diverse experimental techniques to characterize their functional mechanisms. Multiscale simulations can help bridge critical knowledge gaps about IDP structure-function relationships - however, these techniques also face challenges in resolving emergent phenomena within IDP conformational ensembles. We posit that scalable statistical inference techniques can effectively integrate information gleaned from multiple experimental techniques as well as from simulations, thus providing access to atomistic details of these emergent phenomena.

59 BASIC BIOLOGICAL SCIENCES↗

New Era Towards Autonomous Additive Manufacturing: A Review of Recent Trends and Future Perspectives

Abstract The Additive Manufacturing (AM) landscape has significantly transformed in alignment with Industry 4.0 principles, primarily driven by the integration of Artificial Intelligence (AI) and Digital Twin (DT). However, current Intelligent Additive Manufacturing (IAM) systems face limitations such as fragmented AI tool usage and suboptimal human-machine interaction (HMI). This paper reviews existing IAM solutions, emphasizing control, monitoring, process autonomy, and end-to-end integration, and identifies key limitations, such as the absence of a high-level controller for global decision-making. To address these gaps, we propose a transition from IAM to Autonomous Additive Manufacturing (AAM), featuring a hierarchical framework with four integrated layers: knowledge, generative solution, operational, and cognitive. In the cognitive layer, AI agents notably enable machines to independently observe, analyze, plan, and execute operations that traditionally require human intervention. These capabilities streamline production processes and expand the possibilities for innovation, particularly in sectors like in-space manufacturing (ISM). Additionally, this paper discusses the role of AI in self-optimization and lifelong learning, positing that the future of AM will be characterized by a symbiotic relationship between human expertise and advanced autonomy, fostering a more adaptive, resilient manufacturing ecosystem.

Fan, Haolin↗