Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “machine learning and control”

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 829 records · Page 46

Traffic Signal Optimization by Integrating Reinforcement Learning and Digital Twins

Machine learning (ML) methods, especially reinforcement learning (RL), have been widely considered for traffic signal optimization in intelligent transportation systems. Most of these ML methods are centralized, lacking in scalability and adaptability in large traffic networks. Further, it is challenging to train such ML models due to the lack of training platforms and/or the cost of deploying and training in a real traffic networks. This paper presents an approach for the integration of decentralized graph-based multi-agent reinforcement learning (DGMARL) with a Digital Twin (DT) to optimize traffic signals for the reduction of traffic congestion and network-wide fuel consumption related to stopping. Specifically, the DGMARL agents learn traffic state patterns and make decisions regarding traffic signal control with assistance from a Digital Twin module, which simulates and replicates the traffic behaviors of a real traffic network. The proposed approach was evaluated using PTV-Vissim [1], a microscopic traffic simulation platform. PTV-Vissim is also the simulation engine of the DT, enabling emulation and optimization of the traffic signals on the MLK Smart Corridor in Chattanooga, Tennessee. Compared to an actuated signal control baseline approach, experiment results show that Eco_PI, a developed performance measure capturing the impact of stops on fuel consumption, was reduced by 44.27% in a 24-hour and an average of 29.88% in a PM-peak-hour scenario.

Multi-Agent Reinforcement Learning, Digital Twin, ↗

Deep learning with mixup augmentation for improved pore detection during additive manufacturing

In additive manufacturing (AM), process defects such as keyhole pores are difficult to anticipate, affecting the quality and integrity of the AM-produced materials. Hence, considerable efforts have aimed to predict these process defects by training machine learning (ML) models using passive measurements such as acoustic emissions. This work considered a dataset in which keyhole pores of a laser powder bed fusion (LPBF) experiment were identified using X-ray radiography and then registered both in space and time to acoustic measurements recorded during the LPBF experiment. Due to AM’s intrinsic process controls, where a pore-forming event is relatively rare, the acoustic datasets collected during monitoring include more non-pores than pores. In other words, the dataset for ML model development is imbalanced. Moreover, this imbalanced and sparse data phenomenon remains ubiquitous across many AM monitoring schemes since training data is nontrivial to collect. Hence, we propose a machine learning approach to improve this dataset imbalance and enhance the prediction accuracy of pore-labeled data. Specifically, we investigate how data augmentation helps predict pores and non-pores better. This imbalance is improved using recent advances in data augmentation called Mixup, a weak-supervised learning method. Convolutional neural networks (CNNs) are trained on original and augmented datasets, and an appreciable increase in performance is reported when testing on five different experimental trials. When ML models are trained on original and augmented datasets, they achieve an accuracy of 95% and 99% on test datasets, respectively. We also provide information on how dataset size affects model performance. Lastly, we investigate the optimal Mixup parameters for augmentation in the context of CNN performance.

36 MATERIALS SCIENCE↗

Deep learning with mixup augmentation for improved pore detection during additive manufacturing

In additive manufacturing (AM), process defects such as keyhole pores are difficult to anticipate, affecting the quality and integrity of the AM-produced materials. Hence, considerable efforts have aimed to predict these process defects by training machine learning (ML) models using passive measurements such as acoustic emissions. This work considered a dataset in which keyhole pores of a laser powder bed fusion (LPBF) experiment were identified using X-ray radiography and then registered both in space and time to acoustic measurements recorded during the LPBF experiment. Due to AM’s intrinsic process controls, where a pore-forming event is relatively rare, the acoustic datasets collected during monitoring include more non-pores than pores. In other words, the dataset for ML model development is imbalanced. Moreover, this imbalanced and sparse data phenomenon remains ubiquitous across many AM monitoring schemes since training data is nontrivial to collect. Hence, we propose a machine learning approach to improve this dataset imbalance and enhance the prediction accuracy of pore-labeled data. Specifically, we investigate how data augmentation helps predict pores and non-pores better. This imbalance is improved using recent advances in data augmentation called Mixup, a weak-supervised learning method. Convolutional neural networks (CNNs) are trained on original and augmented datasets, and an appreciable increase in performance is reported when testing on five different experimental trials. When ML models are trained on original and augmented datasets, they achieve an accuracy of 95% and 99% on test datasets, respectively. We also provide information on how dataset size affects model performance. Lastly, we investigate the optimal Mixup parameters for augmentation in the context of CNN performance.

42 ENGINEERING↗

Neural networks as a control methodology

While conventional computers must be programmed in a logical fashion by a person who thoroughly understands the task to be performed, the motivation behind neural networks is to develop machines which can train themselves to perform tasks, using available information about desired system behavior and learning from experience. There are three goals of this fellowship program: (1) to evaluate various neural net methods and generate computer software to implement those deemed most promising on a personal computer equipped with Matlab; (2) to evaluate methods currently in the professional literature for system control using neural nets to choose those most applicable to control of flexible structures; and (3) to apply the control strategies chosen in (2) to a computer simulation of a test article, the Control Structures Interaction Suitcase Demonstrator, which is a portable system consisting of a small flexible beam driven by a torque motor and mounted on springs tuned to the first flexible mode of the beam. Results of each are discussed.

Mccullough, Claire L.↗

Adaptive Power System Emergency Control using Deep Reinforcement Learning

Power system emergency control is generally regarded as the final safety net for grid security and resiliency. Existing emergency control schemes are usually designed off-line based on either the conceived “worst” case scenarios or a few typical operation scenarios. These schemes are facing significant adaptiveness and robustness issues as increasing uncertainties and variations occur in modern electrical grids. To address these challenges, for the first time, this paper proposes a novel adaptive emergency control scheme using deep reinforcement learning (DRL), by leveraging the high-dimensional feature extraction and non-linear generalization capabilities DRL has for complex systems with high-dimensional variations. Furthermore, an open-source platform named DeepGrid has been designed for the first time to assist the DRL development and benchmarking processes in power system emergency control. Details of the platform, DRL, and emergency control schemes that use dynamic braking or under-voltage load shedding are presented. Extensive case studies performed in both two-area four-machine system and IEEE 39-Bus system have demonstrated the excellent performance and robustness of the proposed schemes.

Deep reinforcement learning, emergency control, lo↗

Coupling High-Throughput Experiments and Regression Algorithms to Optimize PGM-Free ORR Electrocatalyst Synthesis

Over the past decades, significant improvement has been achieved in the performance of platinum group metal-free (PGM-free) materials as an alternative to Pt-based electrocatalysts for oxygen reduction reaction (ORR). However, further progress in ORR activity requires evaluation of precursors and synthesis approaches. In response to this challenge, we generated a first of its kind experimental data set of 36 samples using high-throughput synthesis and activity measurements. Several control parameters (e.g., Fe precursor identity, the precursor content, and pyrolysis temperature) were varied. We then developed several state-of-the-art machine learning (ML) based regression models to predict ORR activity, dependent on selected synthesis variables. Through an iterative algorithm, higher prediction accuracy (smaller root-mean-square error) was achieved. We identified that gradient boosting regression (GBR) and support vector regression (SVR), among several methods, work best for this data set. Aided by our ML-based surrogate models, we decided to alter catalyst synthesis conditions, which resulted in a 36% increase in measured ORR activity in comparison to the maximum ORR mass activity value of 21.9 A/g catalyst in the original data set. Overall, this combined experiment and machine learning approach represents a promising path forward toward developing highly efficient next-generation ORR electrocatalysts and, more generally, functional materials.

25 ENERGY STORAGE↗

Regional variability in the environmental controls of precipitation regimes in the tropics

The environmental factors that control precipitation regimes in Boreal winter rainfall in the tropics and their regional variabilities are examined using a simple statistical analysis and a machine learning model. Radar-derived precipitation from field campaigns at Darwin Australia, Manaus Brazil, and the Equatorial Indian Ocean, along with the corresponding large-scale environmental variables from ERA5 reanalysis, are used. The dependence of marginal distributions of frequencies of these regimes on five environmental variables are calculated. The variables are hourly column integrated precipitable water (PW), convective available potential energy (CAPE), convective inhibition (CIN), and lower and upper tropospheric wind shear. The simple machine learning model that predicts the probability of transition from suppressed to an active regime as a function of the environmental variables is designed and optimized for a potential application as a trigger function for convection parameterizations. To the first order the analysis shows an abrupt increase in the probability of an active regime near PW > 60 mm and CIN of <100 J kg -1 . The key differences in the frequencies of active regimes among the regions are found to be related to the fact that over Darwin there is strong variability in PW while CIN is generally low. Over Amazon, on the other hand, both PW and CIN are quite variable. Over the DYNAMO domain the comparatively frequent low PW is compensated for by consistently low CIN resulting in a moderate frequency of an active regime.

54 ENVIRONMENTAL SCIENCES↗

Physics-constrained, low-dimensional models for magnetohydrodynamics: First-principles and data-driven approaches

Plasmas are highly nonlinear and multiscale, motivating a hierarchy of models to understand and describe their behavior. However, there is a scarcity of plasma models of lower fidelity than magnetohydrodynamics (MHD), although these reduced models hold promise for understanding key physical mechanisms, efficient computation, and real-time optimization and control. Galerkin models, obtained by projection of the MHD equations onto a truncated modal basis, and data-driven models, obtained by modern machine learning and system identification, can furnish this gap in the lower levels of the model hierarchy. This work develops a reduced-order modeling framework for compressible plasmas, leveraging decades of progress in projection-based and data-driven modeling of fluids. We begin by formalizing projection-based model reduction for nonlinear MHD systems. To avoid separate modal decompositions for the magnetic, velocity, and pressure fields, we introduce an energy inner product to synthesize all of the fields into a dimensionally consistent, reduced-order basis. Next, we obtain an analytic model by Galerkin projection of the Hall-MHD equations onto these modes. We illustrate how global conservation laws constrain the model parameters, revealing symmetries that can be enforced in data-driven models, directly connecting these models to the underlying physics. We demonstrate the effectiveness of this approach on data from high-fidelity numerical simulations of a three-dimensional spheromak experiment. Finally, this manuscript builds a bridge to the extensive Galerkin literature in fluid mechanics and facilitates future principled development of projection-based and data-driven models for plasmas.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Driver Identification Midyear Report

First, we create a profile for each authorized driver based on their existing driving data. We then train a machine learning model on the driving data from this profile, yielding an individualized model for each driver. Finally during a drive, we pass the Controller Area Network (CAN) data to the model and authen ticate the driver’s identity in real-time. This verification or lack thereof could be used to alert supervisors of threats to their drivers or transported materials. Deviations from their normal driving behavior could indicate high-risk situations, medical events, or even insider threats.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗

Potential applications of microbial genomics in nuclear non-proliferation

As nuclear technology evolves in response to increased demand for diversification and decarbonization of the energy sector, new and innovative approaches are needed to effectively identify and deter the proliferation of nuclear arms, while ensuring safe development of global nuclear energy resources. Preventing the use of nuclear material and technology for unsanctioned development of nuclear weapons has been a long-standing challenge for the International Atomic Energy Agency and signatories of the Treaty on the Non-Proliferation of Nuclear Weapons. Environmental swipe sampling has proven to be an effective technique for characterizing clandestine proliferation activities within and around known locations of nuclear facilities and sites. However, limited tools and techniques exist for detecting nuclear proliferation in unknown locations beyond the boundaries of declared nuclear fuel cycle facilities, representing a critical gap in non-proliferation safeguards. Microbiomes, defined as “characteristic communities of microorganisms” found in specific habitats with distinct physical and chemical properties, can provide valuable information about the conditions and activities occurring in the surrounding environment. Microorganisms are known to inhabit radionuclide-contaminated sites, spent nuclear fuel storage pools, and cooling systems of water-cooled nuclear reactors, where they can cause radionuclide migration and corrosion of critical structures. Microbial transformation of radionuclides is a well-established process that has been documented in numerous field and laboratory studies. These studies helped to identify key bacterial taxa and microbially-mediated processes that directly and indirectly control the transformation, mobility, and fate of radionuclides in the environment. Expanding on this work, other studies have used microbial genomics integrated with machine learning models to successfully monitor and predict the occurrence of heavy metals, radionuclides, and other process wastes in the environment, indicating the potential role of nuclear activities in shaping microbial community structure and function. Results of this previous body of work suggest fundamental geochemical-microbial interactions occurring at nuclear fuel cycle facilities could give rise to microbiomes that are characteristic of nuclear activities. These microbiomes could provide valuable information for monitoring nuclear fuel cycle facilities, planning environmental sampling campaigns, and developing biosensor technology for the detection of undisclosed fuel cycle activities and proliferation concerns.

59 BASIC BIOLOGICAL SCIENCES↗

Accelerated CO2 Storage Optimization Using Multi-Resolution Fourier Neural Operator at the Illinois Basin Decatur Project (IBDP)

This paper presents a deep learning-based approach for optimizing CO2 injection in carbon capture and storage (CCS) operations. We developed a multi-resolution machine learning model to significantly reduce data generation costs. Utilizing this proxy model, we implemented a multi-objective genetic algorithm to optimize well control during the CO2 injection process. The proposed approach was applied to the Illinois Basin Decatur Project (IBDP), successfully optimizing the CO2 injection schedule based on three key objectives: maximizing the amount of CO2 stored, maximizing sweep efficiency, and minimizing pressure increase. The use of the proxy model accelerated the optimization workflow by two orders of magnitude, while the cost of data generation for the proxy model was reduced by 90% by utilizing a coarse-scale model.

accelerated CO2 storage optimization↗

Achieving Cyber-Resilience for Power Systems using a Learning, Model-Assisted Blockchain Framework

The secure integration and management of distributed energy resources (DER) and power aggregators in the electric grid requires secure communications and a physics-aware Command and Control (C2) strategy. A Blockchain (BC)-based overlay network was developed to provide a security layer for the existing power grid network that mitigates risks in current and legacy network and C2 protocols. By integrating a Model-Assisted Machine Learning (MAML) framework with a Secure Blockchain Overlay Network (SBON) a defense-in-depth strategy was achieved. In our approach, the MAML framework leveraged a smart contract framework to gather network data and learn the dynamics of DER to develop detection strategies for attacks targeting sensors and actuators used by DER. The MAML framework learned dynamical systems models for individual DERs to detect sensor attacks. For DER we utilized a Digital Twin (DT) to accelerate the learning process for a model resistant to stealthy attacks. The project created DT for PV inverters and BESS. The DTs were coupled with a model-assisted, data-driven learning of DER behavior. Specifically, we evaluated architectures for model-based learning with model-free fine-tuning. Additionally, differential privacy techniques were used to obfuscate data, while still allowing the computation of attack detection results based on obfuscated data. The SBON developed leverages a private permissioned blockchain network orchestrated with the Hyperledger Fabric framework. To connect the cyber world, which orchestrates the blockchain fabric, and the physical world where the power network resides, we developed a system implementation to enable the secure interaction of the physical world and the abstracted blockchain.

97 MATHEMATICS AND COMPUTING↗

Insights into twinning in $\mathrm{Mg AZ31: A}$ combined $\mathrm{EBSD}$ and machine learning study

To explore the driving forces behind deformation twinning in Mg AZ31, a machine learning framework is utilized to mine data obtained from electron backscatter diffraction (EBSD) scans in order to extract correlations in physical characteristics that cause twinning. The results are intended to inform physics-based models of twin nucleation and growth. A decision tree learning environment is selected to capture the relationships between microstructure and twin formation; this type of model effectively highlights the more influential characteristics of the local microstructure. Trees are assembled to analyze both twin nucleation in a given grain, and twin propagation across grain boundaries. Each model reveals a unique combination of crystallographic attributes that affect twinning in the Mg. Twin nucleation is found to be mostly controlled by a combination of grain size, basal Schmid factor, and bulk dislocation density while twin propagation is affected most by grain boundary length, basal Schmid factor, angle from grain boundary plane to the RD plane, and grain boundary misorientation. The machine learning framework can be readily adapted to investigate other relationships between microstructure and material response.

36 MATERIALS SCIENCE↗

Bayesian inverse modeling of the atmospheric transport and emissions of a controlled tracer release from a nuclear power plant

Probability distribution functions (PDFs) of model inputs that affect the transport and dispersion of a trace gas released from a coastal California nuclear power plant are quantified using ensemble simulations, machine-learning algorithms, and Bayesian inversion. The PDFs are constrained by observations of tracer concentrations and account for uncertainty in meteorology, transport, diffusion, and emissions. Meteorological uncertainty is calculated using an ensemble of simulations of the Weather Research and Forecasting (WRF) model that samples five categories of model inputs (initialization time, boundary layer physics, land surface model, nudging options, and reanalysis data). The WRF output is used to drive tens of thousands of FLEXPART dispersion simulations that sample a uniform distribution of six emissions inputs. Machine-learning algorithms are trained on the ensemble data and used to quantify the sources of ensemble variability and to infer, via inverse modeling, the values of the 11 model inputs most consistent with tracer measurements. We find a substantial ensemble spread in tracer concentrations (factors of 10 to 10 3 ), most of which is due to changing emissions inputs (about 80%), though the cumulative effects of meteorological variations are not negligible. The performance of the inverse method is verified using synthetic observations generated from arbitrarily selected simulations. When applied to measurements from a controlled tracer release experiment, the inverse method satisfactorily determines the location, start time, duration and amount. In a 2 km × 2 km area of possible locations, the actual location is determined to within 200m. The start time is determined to within 5min out of 2h, and the duration to within 50min out of 4h. Over a range of release amounts of 10 to 1000kg, the estimated amount exceeds the actual amount of 146kg by only 32kg. The inversion also estimates probabilities of different WRF configurations. In conclusion, to best match the tracer observations, the highest-probability cases in WRF are associated with using a late initialization time and specific reanalysis data products.

54 ENVIRONMENTAL SCIENCES↗

Feasibility of Adding Twitter Data to Aid Drought Depiction: Case Study in Colorado

The use of social media, such as Twitter, has changed the information landscape for citizens’ participation in crisis response and recovery activities. Given that drought progression is slow and also spatially extensive, an interesting set of questions arise, such as how the usage of Twitter by a large population may change during the development of a major drought alongside how the changing usage facilitates drought detection. For this reason, contemporary analysis of how social media data, in conjunction with meteorological records, was conducted towards improvement in the detection of drought and its progression. The research utilized machine learning techniques applied over satellite-derived drought conditions in Colorado. Three different machine learning techniques were examined: the generalized linear model, support vector machines and deep learning, each applied to test the integration of Twitter data with meteorological records as a predictor of drought development. It is found that the integration of data resources is viable given that the Twitter-based model outperformed the control run which did not include social media input. Eight of the ten models tested showed quantifiable improvements in the performance over the control run model, suggesting that the Twitter-based model was superior in predicting drought severity. Future work lies in expanding this method to depict drought in the western U.S.

54 ENVIRONMENTAL SCIENCES↗

Machine learning unifies flexibility and efficiency of spinodal structure generation for stochastic biomaterial design

Abstract Porous biomaterials design for bone repair is still largely limited to regular structures (e.g. rod-based lattices), due to their easy parameterization and high controllability. The capability of designing stochastic structure can redefine the boundary of our explorable structure–property space for synthesizing next-generation biomaterials. We hereby propose a convolutional neural network (CNN) approach for efficient generation and design of spinodal structure—an intriguing structure with stochastic yet interconnected, smooth, and constant pore channel conducive to bio-transport. Our CNN-based approach simultaneously possesses the tremendous flexibility of physics-based model in generating various spinodal structures (e.g. periodic, anisotropic, gradient, and arbitrarily large ones) and comparable computational efficiency to mathematical approximation model. We thus successfully design spinodal bone structures with target anisotropic elasticity via high-throughput screening, and directly generate large spinodal orthopedic implants with desired gradient porosity. This work significantly advances stochastic biomaterials development by offering an optimal solution to spinodal structure generation and design.

59 BASIC BIOLOGICAL SCIENCES↗

Scalable Hybrid Classification-Regression Solution for High-Frequency Nonintrusive Load Monitoring

Residential buildings with the ability to monitor and control their net-load (sum of load and generation) can provide valuable flexibility to power grid operators. We present a novel multiclass nonintrusive load monitoring (NILM) approach that enables effective net-load monitoring capabilities at high-frequency with minimal additional equipment and cost. The proposed machine learning based solution provides accurate multiclass state predictions while operating at a faster timescale (able to provide a prediction for each 60- Hz ac cycle used in US power grid) without relying on event-detection techniques. We also introduce an innovative hybrid classification-regression method that allows for the prediction of not only load on/off states but also individual load operating power levels. A test bed with eight residential appliances is used for validating the NILM approach. Results show that the overall method has high accuracy, good scaling and generalization properties.

feature extraction↗

The disCO2ver Platform: Curating Data and Tools for Geologic Carbon Sequestration and Deep Subsurface Research Systems

The U.S. DOE National Energy Technology Laboratory has invested 12+ years of development into the data repository and digital laboratory, the Energy Data eXchange (EDX, edx.netl.doe.gov). Supporting a variety of research areas across the DOE Office of Fossil Energy and Carbon Management, the platform has successfully curated and preserved thousands of data products from DOE research. The Carbon Storage Program has successfully supported data curation, upload, and publishing of data products on EDX for many years, demonstrating a success story of how resources like EDX can effectively help with long term preservation and publishing of DOE data products. EDX continues to shift towards cloud-supported infrastructure, taking a hybrid approach combining on-premises compute and storage integrated with cloud-hosted services. The integration of cloud compute and hybrid architecture enables the development of EDX-hosted platforms that tailor the data and tools hosted on them to a specific community, enables implementation of machine learning tools for data discovery and filtering, and enables the hosting of virtual (online user interface) tools. Geologic carbon sequestration (GCS) research continues to scale up in response to the current administration goals to reduce greenhouse gas emissions and transition the energy economy. Over the last year, EDX’s disCO2ver platform has been developed in response to the need for access to data products and tools to support the scaling up of GCS research. disCO2ver provides access to data resources and tools, produced by DOE and outside authoritative external resources. The platform also provides a user-access control component for the virtualization and cloud hosting of tools. Tools that need to be virtualized, to eliminate the need for users to download the tool and use local compute resources, is essential to supporting big-data analysis and machine learning that is becoming common place in carbon storage modeling, risk analysis, and data publishing practices. This talk will review the EDX’s disCO2ver platform and the current work ongoing to curate data and tools to support GCS and deep subsurface systems research.

Morkner, Paige↗