Search NASA⌕ Search

SEARCH · Search NASA

Results for “Operator 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 1,045 records · Page 58
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Progress of Gas Injection EOR Surveillance in the Bakken Unconventional Play—Technical Review and Machine Learning Study

Although considerable laboratory and modeling activities were performed to investigate the enhanced oil recovery (EOR) mechanisms and potential in unconventional reservoirs, only limited research has been reported to investigate actual EOR implementations and their surveillance in fields. Eleven EOR pilot tests that used CO2, rich gas, surfactant, water, etc., have been conducted in the Bakken unconventional play since 2008. Gas injection was involved in eight of these pilots with huff ‘n’ puff, flooding, and injectivity operations. Surveillance data, including daily production/injection rates, bottomhole injection pressure, gas composition, well logs, and tracer testing, were collected from these tests to generate time-series plots or analytics that can inform operators of downhole conditions. A technical review showed that pressure buildup, conformance issues, and timely gas breakthrough detection were some of the main challenges because of the interconnected fractures between injection and offset wells. The latest operation of co-injecting gas, water, and surfactant through the same injection well showed that these challenges could be mitigated by careful EOR design and continuous reservoir monitoring. Reservoir simulation and machine learning were then conducted for operators to rapidly predict EOR performance and take control actions to improve EOR outcomes in unconventional reservoirs.

Energy & Fuels↗
DOE OSTI2021

InversionNet3D: Efficient and Scalable Learning for 3-D Full-Waveform Inversion

Seismic full-waveform inversion (FWI) techniques aim to find a high-resolution subsurface geophysical model provided with waveform data. Some recent effort in data-driven FWI has shown some encouraging results in obtaining 2-D velocity maps. However, due to high computational complexity and large memory consumption, the reconstruction of 3-D high-resolution velocity maps via deep networks is still a great challenge. Here, in this article, we present InversionNet3D (InvNet3D), an efficient and scalable encoder–decoder network for 3-D FWI. The proposed method employs group convolution in the encoder to establish an effective hierarchy for learning information from multiple sources while cutting down unnecessary parameters and operations at the same time. The introduction of invertible layers further reduces the memory consumption of intermediate features during training and, thus, enables the development of deeper networks with more layers and higher capacity as required by different application scenarios. Experiments on the 3-D Kimberlina dataset demonstrate that InvNet3D achieves state-of-the-art reconstruction performance with lower computational cost and lower memory footprint compared to the baseline.

58 GEOSCIENCES↗
DOE OSTI2021

Hybrid Analytics Solution to Improve Coal Power Plant Operations

This project focused on developing advanced methods for thermal performance monitoring of a coal-fueled power plant. The specific goal was to develop and demonstrate a new thermal performance monitoring approach using a hybrid model that integrates a physics-based heat balance model with a machine learning-based pattern recognition model. The hybrid model enables increased accuracy and scope of the thermal analysis and an improved ability to monitor and detect changes in plant operation. This new approach takes full advantage of the individual model capabilities and creates an important new set of capabilities not previously possible using the two types of models separately. Using the heat balance model, a rich set of derived parameters (virtual sensors) are calculated from the measured plant operating data at each time point. The combined measured and derived data values are used by machine learning algorithms to create pattern recognition models over the range of normal unit operation. To create the monitoring models, historical data from normal operation of the plant is first processed by the heat balance model to compute the derived parameter data. The result is a greatly expanded set of normal operating data that can be used as input to create the pattern recognition model. Once the models are calibrated for normal operation, the hybrid model is suitable for use in continuous online monitoring. During online monitoring, new plant operating data is processed first by the heat balance model and then by the pattern recognition model. Results from the pattern recognition model quantify the deviation of each measured or derived parameter from its expected value in normal operation. The hybrid models can detect abnormal changes in plant operating data with very high accuracy and sensitivity. When abnormal behavior is detected, alerts are generated automatically for evaluation by the plant monitoring staff. The new hybrid solution product was developed and verified in the performance of the project. The hybrid solution was tested first in a simulation environment that mimicked the plant data systems and infrastructure used by U.S. power generating plants and utilities. The hybrid solution was then deployed for real-time, online monitoring of an operating coal-fueled power plant at a field test site. Field testing demonstrated that all hybrid solution development objectives were accomplished. The project work was based on combining the capabilities of two existing software products to create the new hybrid solution product. One of these was the existing MapEx® heat balance product and the other was the existing SureSense® advanced pattern recognition product. Each of these separate products was assessed to be at a Technology Readiness Level (TRL) of 9 at the start of the effort. The hybrid solution product was assessed to be at a TRL of 2 at the start of the project based on early feasibility work by the project team. At completion of the field testing performed in the project, the hybrid solution product was assessed to be at a TRL of 7. The project team expects that the hybrid solution product will be deployed commercially and will achieve a TRL of 9 within one year after completion of the project.

01 COAL, LIGNITE, AND PEAT↗
DOE OSTI2026

Bridging Cloud and Edge Computing at NREL Using CONNECT: Cloud Optimized Networking for Next-Gen Edge Computing Technologies [Slides]

CONNECT is an innovative on-premise hardware and software solution that integrates edge and cloud computing infrastructure at NREL. Built on the AWS Greengrass middleware and leveraging the MQTT protocol, CONNECT enables real-time data streaming from IoT devices and gateways to both cloud and local services, empowering researchers to rapidly capture, analyze, and act upon edge-generated data while leveraging cloud capabilities. The platform addresses research infrastructure challenges by providing a pre-approved platform which is already configured with the correct networking and cybersecurity baselines thus eliminating procurement delays and enabling on-demand availability. CONNECT's hybrid architecture efficiently manages burstable workloads, allowing research teams to dynamically scale computational capacity, handle peak data loads, and reduce operational bottlenecks. Advanced capabilities include built-in GPU support for executing machine learning models which enables low-latency inference at the edge from models trained in the cloud. This architecture supports real-time analytics and filtering, providing a mechanism to allow only transmitting and processing high-value data. Cloud-based configuration management permits engineers to manage on-premise systems remotely, optimizing operational efficiency. By bridging edge and cloud computing, CONNECT provides NREL researchers with a flexible, scalable platform that accelerates scientific discovery while maintaining robust security and performance standards.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

A CHIL Validation of Machine Learning-Assisted Methods for Real-Time Controls of Solar PV for Grid Services: Preprint

Recent research has highlighted the potential for solar to act as a zero-marginal-cost and zero-emission flexibility resource on the bulk power system when operated with advanced control systems. To increase the performance of these systems, leading technologies, including machine learning (ML) and hierarchical inverter set point allocation, have been proposed; however, these technologies lack comprehensive validation under real-world application scenarios. This paper addresses this gap by designing and developing a controller-hardware-in-the-loop framework to evaluate the performance of different flexible solar technologies in responding to automatic generation control signals in a closed-loop fashion. Simulation results indicate the superior performance of an ML-based approach compared to the conventional reference-control grouping-based approach, showcasing its potential to support grid stability and operational efficiency.

closed-loop validation↗
DOE OSTI2022

Constructing A New CHF Look-Up Table Based on the Domain Knowledge Informed Machine Learning Methodology

Accurate prediction of CHF under various fluid flow conditions continues to be required for design, operation and safety analysis of light water reactor rod bundles. Due to the lack of in-depth physical understanding as well as limited high-resolution data in the micro-scale flow and heat transfer, the existing models feature a sub-optimal uncertainty band. In this study, driven by the prior domain knowledge information obtained, an improved CHF look-up table is developed through unified machine learning algorithms for the vertical flow conditions within tube and annulus geometry. The Groeneveld 2006 look-up table is used as the domain knowledge to train machine learning process against tube and annulus CHF data for both DNB and DO type. The new look-up table shows improved accuracy for conditions relevant to PWRs and BWRs. In addition, its domain knowledge informed nature ensures that a rationale prediction can be made, thus accounting for previous valuable information in the machine learning model training process.

Jin, Yue↗
DOE OSTI2022

Two-Stage Reinforcement Learning Policy Search for Grid-Interactive Building Control

This paper develops an intelligent grid-interactive building controller, which optimizes building operation during both normal hours and demand response (DR) events. To avoid costly on-demand computation and to adapt to non-linear building models, the controller utilizes reinforcement learning (RL) and makes real-time decisions based on a near-optimal control policy. Learning such a policy typically amounts to solving a hard non-convex optimization problem. We propose to address this problem with a novel global-local policy search method. In the first stage, an RL algorithm based on zero-order gradient estimation is leveraged to search for the optimal policy globally, due to its scalability and the potential to escape some poor performing local optima. The obtained policy is then fine-tuned locally to bring the first-stage solution closer to that of the original unsmoothed problem. Experiments on a simulated five-zone commercial building demonstrate the advantages of the proposed method over existing learning approaches. They also show that the learned control policy outperforms a pragmatic linear model predictive controller (MPC) and approaches the performance of an oracle MPC in testing scenarios. Using a state-of-the-art advanced computing system, we demonstrate that the controller can be learned and deployed within hours of training.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2023

Extending Parsimonious Bayesian Inference

Parsimonious Bayesian inference is a theoretical framework for efficient data assimilation that seeks to balance increased consistency between predictions and training data against corresponding increases in model complexity. Within this framework, over-training is understood as optimization that encodes excessive information within model parameters while only achieving small improvements between predictions and training data. This project aims to develop practical methods of limiting excess model information during optimization. One key observation is that practical heuristics for parsimonious learning in high-dimensions must balance expressivity, i.e. the ability of the model to capture diverse predictions with only a few non-zero parameters, against discoverability, i.e. the ability to train the model with gradient-based optimization and drive parameters to low information states. As such, we developed logical activation functions that are able to adaptively approximate arbitrary truth tables that define Boolean logic operations within a probabilistic framework. These functions have demonstrated the ability to learn exclusive disjunction (XOR) and conditioned disjunction (if [condition] then [result_if_true] else [result_if_false]) within a single layer of a neural network. To efficiently exploit these activation functions to drive parsimonious learning required several other advances within the domain of variational inference. The most efficient form of complexity suppression is structured sparsification, driving most model parameters to zero while achieving the structural coherence among nonzeros needed for bandwidth reduction. Such models are not only far more efficient at suppressing information-theoretic complexity, they also reduce the other forms of complexity (computations, communication, storage, and the number of dependencies needed to evaluate predictions). Aiming to support enhanced sparsification, this project examined new approaches to high-dimensional variational inference that allow us to calibrate and control parameter uncertainty during optimization. By identifying which parameters can sustain sparsifying perturbations with little impact on prediction quality, we can develop better pruning strategies by framing them as approximate Bayesian inference. These advances also open paths to mitigate concerns with deploying advanced learning methods in resource-constrained environments, such as running models on power-limited or communication-limited devices.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Design-Technology Co-Optimization for NVM-based Neuromorphic Processing Elements

An emerging use-case of machine learning (ML) is to train a model on a high-performance system and deploy the trained model on energy-constrained embedded systems. Neuromorphic hardware platforms, which operate on principles of the biological brain, can significantly lower the energy overhead of a machine learning inference task, making these platforms an attractive solution for embedded ML systems. In this paper, we present a design-technology tradeoff analysis to implement such inference tasks on the processing elements (PEs) of a Non-Volatile Memory (NVM)-based neuromorphic hardware. Through detailed circuit-level simulations at scaled process technology nodes, we show the negative impact of technology scaling on the information-processing latency, which impacts the quality-of-service (QoS) of an embedded ML system. At a finer granularity, the latency inside a PE depends on 1) the delay introduced by parasitic components on its current paths, and 2) the varying delay to sense different resistance states of its NVM cells. Based on these two observations, we make the following three contributions. First, on the technology front, we propose an optimization scheme where the NVM resistance state that takes the longest time to sense is set on current paths having the least delay, and vice versa, reducing the average PE latency, which improves the QoS. Second, on the architecture front, we introduce isolation transistors within each PE to partition it into regions that can be individually power-gated, reducing both latency and energy. Finally, on the system-software front, we propose a mechanism to leverage the proposed technological and architectural enhancements when implementing a machine-learning inference task on neuromorphic PEs of the hardware. Evaluations with a recent neuromorphic hardware architecture show that our proposed design-technology co-optimization approach improves both performance and energy efficiency of machine-learning inference tasks without incurring high cost-per-bit.

42 ENGINEERING↗
DOE OSTI2024

When ancient numerical demons meet physics-informed machine learning: adjoint-based gradients for implicit differentiable modeling

Recent advances in differentiable modeling, a genre of physics-informed machine learning that trains neural networks (NNs) together with process-based equations, have shown promise in enhancing hydrological models' accuracy, interpretability, and knowledge-discovery potential. Current differentiable models are efficient for NN-based parameter regionalization, but the simple explicit numerical schemes paired with sequential calculations (operator splitting) can incur numerical errors whose impacts on models' representation power and learned parameters are not clear. Implicit schemes, however, cannot rely on automatic differentiation to calculate gradients due to potential issues of gradient vanishing and memory demand. Here we propose a “discretize-then-optimize” adjoint method to enable differentiable implicit numerical schemes for the first time for large-scale hydrological modeling. The adjoint model demonstrates comprehensively improved performance, with Kling–Gupta efficiency coefficients, peak-flow and low-flow metrics, and evapotranspiration that moderately surpass the already-competitive explicit model. Therefore, the previous sequential-calculation approach had a detrimental impact on the model's ability to represent hydrological dynamics. Furthermore, with a structural update that describes capillary rise, the adjoint model can better describe baseflow in arid regions and also produce low flows that outperform even pure machine learning methods such as long short-term memory networks. The adjoint model rectified some parameter distortions but did not alter spatial parameter distributions, demonstrating the robustness of regionalized parameterization. Despite higher computational expenses and modest improvements, the adjoint model's success removes the barrier for complex implicit schemes to enrich differentiable modeling in hydrology.

58 GEOSCIENCES↗
DOE OSTI2015

Learning from Trending, Precursor Analysis, and System Failures

Models of reliability growth relate current system unreliability to currently accumulated experience. But “experience” comes in different forms. Looking back after a major accident, one is sometimes able to identify previous events or measurable performance trends that were, in some sense, signaling the potential for that major accident: potential that could have been recognized and acted upon, but was not recognized until the accident occurred. This could be a previously unrecognized cause of accidents, or underestimation of the likelihood that a recognized potential cause would actually operate. Despite improvements in the state of practice of modeling of risk and reliability, operational experience still has a great deal to teach us, and work has been going on in several industries to try to do a better job of learning from experience before major accidents occur. It is not enough to say that we should review operating experience; there is too much “experience” for such general advice to be considered practical. The paper discusses the following: 1. The challenge of deciding what to focus on in analysis of operating experience. 2. Comparing what different models of learning and reliability growth imply about trending and precursor analysis.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Enhancing Distribution System Resilience: A First-Order Meta-RL Algorithm for Critical Load Restoration

The increasing frequency of extreme events and the integration of distributed energy resources (DERs) into modern grids have elevated the need for resilient and efficient critical load restoration strategies in distribution systems. However, the stochastic nature of renewable DERs, limited energy resource availability and the intricate nonlinearities inherent in complex grid control problem make the problem challenging. Although reinforcement learning (RL) and warm-start RL methods have shown promising results, their performance often falls short in rapidly adapting to new, unseen situations and typically requires exhaustive problem-specific tuning. To address these gaps, we propose a First-Order Meta-based RL (FOM-RL) algorithm within an online framework for adaptive and robust critical load restoration. By harnessing local DERs as the enabling technology, FOM-RL allows the RL agent to swiftly adapt to new unseen scenarios by leveraging previously acquired knowledge of different tasks. Experimental results provide evidence that proposed algorithm learns more efficiently and showcases generalization capabilities across diverse set of operational scenarios. Moreover, a rigorous theoretical analysis yields a tight sublinear regret bound, sensitive to temporal variability, with a task-averaged optimality gap bounded by O(VM+D*/(Tsquare root(M))). These results suggest that optimality improves with task similarity and an increased number of tasks M, reaffirming the efficacy and scalability of the proposed approach in addressing the complexities of critical load restoration in distribution systems.

complexity theory↗
DOE OSTI2023

Reconstruction of fast neutron direction in segmented organic detectors using deep learning

A method for reconstructing the direction of a fast neutron source using a segmented organic scintillator-based detector and deep learning model is proposed and analyzed. Here, the model is based on recurrent neural network, which can be trained by a sequence of data obtained from an event recorded in the detector and suitably pre-processed. The performance of deep learning-based model is compared with the conventional double-scatter detection algorithm in reconstructing the direction of a fast neutron source. With the deep learning model, the uncertainty in source direction of 0.301 rad is achieved with 100 neutron detection events in a segmented cubic organic scintillator detector with a side length of 46 mm. To reconstruct the source direction with the same angular resolution as the double-scatter algorithm, the deep learning method requires 75% fewer events. Application of this method could augment the operation of segmented detectors operated in the neutron scatter camera configuration for applications such as special nuclear material detection.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2023

Stochastic projective splitting

Here, we present a new, stochastic variant of the projective splitting (PS) family of algorithms for inclusion problems involving the sum of any finite number of maximal monotone operators. This new variant uses a stochastic oracle to evaluate one of the operators, which is assumed to be Lipschitz continuous, and (deterministic) resolvents to process the remaining operators. Our proposal is the first version of PS with such stochastic capabilities. We envision the primary application being machine learning (ML) problems, with the method’s stochastic features facilitating “mini-batch” sampling of datasets. Since it uses a monotone operator formulation, the method can handle not only Lipschitz-smooth loss minimization, but also min–max and noncooperative game formulations, with better convergence properties than the gradient descent-ascent methods commonly applied in such settings. The proposed method can handle any number of constraints and nonsmooth regularizers via projection and proximal operators. We prove almost-sure convergence of the iterates to a solution and a convergence rate result for the expected residual, and close with numerical experiments on a distributionally robust sparse logistic regression problem.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Cloud-Control of Legacy Building Automation System: A case study

As Internet of Things devices and cloud-based platforms become more mature, Energy Management and Information Systems (EMIS) are increasingly gaining momentum in the building industry. In large commercial buildings, Fault-Detection and Diagnostic (FDD) and energy information systems (EIS) are now established technologies with tens of providers and thousands of deployment sites across North America. The new frontier for the EMIS technology is now represented by control systems that use advanced system optimization (ASO) methods to improve the operations of the HVAC system. Given the complexity of the integration of such systems with the existing building automation systems (BAS) and the higher risk involved with direct control of the HVAC, these systems are still emerging in the market. This paper presents the results of a project in which a start-up company partnered with a research institution to develop a cloud-based software EMIS solution and deployed it in a university campus in California. The software system included advanced sensing, data acquisition, storage and advanced control and analytics applications developed on top of the native BAS. The new platform controls ten buildings on the campus and the FDD and the ASO applications deployed on this platform were able to generate energy savings of up to 35% and 25% in certain buildings for each functionality respectively. Where the platform did not save energy, it improved building service (air quality). Lessons learned include the importance of collaborating with and training the building operators and evaluating whether the legacy system can work reliably with the new technology.

Prakash, Anand Krishnan↗
DOE OSTI2021

Deciphering Degradation: Machine Learning on Real-World Performance Data (Final Report)

This project addresses a fundamental flaw in solar PV research and solar project financing; the assumed rate of degradation for solar plants. The solar industry currently relies on an out-dated report that observed a 0.5% degradation rate based on a small sample size of systems (~100). While the research conducted at the time was new and innovative, the solar community has not updated this research and universally applies this 0.5% degradation assumption in financial models. Our project updates this assumption by analyzing observed degradation from the industry’s largest dataset of operating solar assets (>10,000 systems) and creating the first machine-learning model based on these observed results to quantify and identify features that drive degradation. There are two strategic goals for this award: reduce the cost of capital (enable solar to attract more capital) and improve the reliability of solar itself. These dual goals are achieved by leveraging an industry dataset to observe system degradation on a large scale, deploying advanced data analysis and machine learning methods to quantify and predict system reliability, and engaging with industry stakeholders to help them accurately price degradation in financial models.

14 SOLAR ENERGY↗
DOE OSTI2022

An Analysis of Grid Operator Survey Responses: Inexperience, Workload and Fatigue in the Control Room

Although a wide array of tools and technologies have been developed over the last decade to support power grid operators, deployment of these tools has been less successful. One reason for unsuccessful deployment may be an inadequate understanding of the factors that contribute to operator error in the control room. An analysis of operators’ current vulnerabilities may provide the baseline understanding needed to inform new technology integration. In an attempt to learn more about these vulnerabilities and their perceived impact on human error we collected and analyzed survey data from 20 electric grid control room operators. We asked survey respondents to consider the various operator, technology and interaction vulnerabilities that may arise during work in the control room and record their attitudes and experiences toward each. Results suggest operator inexperience, high mental workload and fatigue are the most common vulnerabilities experienced during a shift. Survey results were analyzed to explore these vulnerabilities in greater depth.

Inexperience, Workload, Fatigue↗
DOE OSTI2023

Anomaly Detection in Liquid Sodium Cold Trap Operation with Multisensory Data Fusion Using Long Short-Term Memory Autoencoder

Sodium-cooled fast reactors (SFR), which use high temperature fluid near ambient pressure as coolant, are one of the most promising types of GEN IV reactors. One of the unique challenges of SFR operation is purification of high temperature liquid sodium with a cold trap to prevent corrosion and obstructing small orifices. We have developed a deep learning long short-term memory (LSTM) autoencoder for continuous monitoring of a cold trap and detection of operational anomaly. Transient data were obtained from the Mechanisms Engineering Test Loop (METL) liquid sodium facility at Argonne National Laboratory. The cold trap purification at METL is monitored with 31 variables, which are sensors measuring fluid temperatures, pressures and flow rates, and controller signals. Loss-of-coolant type anomaly in the cold trap operation was generated by temporarily choking one of the blowers, which resulted in temperature and flow rate spikes. The input layer of the autoencoder consisted of all the variables involved in monitoring the cold trap. The LSTM autoencoder was trained on the data corresponding to cold trap startup and normal operation regime, with the loss function calculated as the mean absolute error (MAE). The loss during training was determined to follow log-normal density distribution. During monitoring, we investigated a performance of the LSTM autoencoder for different loss threshold values, set at a progressively increasing number of standard deviations from the mean. The anomaly signal in the data was gradually attenuated, while preserving the noise of the original time series, so that the signal-to-noise ratio (SNR) averaged across all sensors decreased below unity. Results demonstrate detection of anomalies with sensor-averaged SNR < 1.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗