Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “machine learning potentials”

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 1,585 records · Page 88

Load Disaggregation (Modeling Individual Appliance Power Generation & Consumption in Real-Time)

We developed a machine learning-based load disaggregation method to estimate the real-time output of individual appliances from the whole-house measurements. We first learn the important features associated with each type of appliances using the ground truth consumption data of individual appliances potentially available for a small set of houses equipped with submeters. The learned features are then be used to estimate the power generation/consumption of the appliances from the whole-house consumption. This developed load disaggregation software includes two steps. The first step is to identify the on/off status of different appliances using a classification method, and the second step is to estimate the appliance output using a regression method.

Hao, Jun↗

Computational Exploration of High Entropy Alloys as Promising Materials for Future Beam Windows

With the ever-increasing demand for high beam power, the currently used beam-intercepting devices (BIDs) such as targets, and beam windows may not be able to handle the high power required for future accelerator complexes or the lifetime may be reduced drastically. As beam power increases, the damage incurred by BIDs, including thermal shock, fatigue, and irradiation damage, also rises. Therefore, it is imperative to design materials that can withstand high beam power for longer lifetimes. High entropy alloys (HEAs) have emerged as potential alternative materials for designing next-generation BIDs. In this study, we primarily focus on materials for developing beam windows for next-generation accelerator complexes. We propose an integrated approach that combines various computational techniques to study and design new materials. Specifically, we use CALPHAD, density functional theory (DFT), and molecular dynamics (MD) to comprehensively investigate the defect properties of suitable HEAs, offering potential alternatives for future beam windows. We begin by scanning the extensive phase space provided by Cr-Mn-V-Ti-Al-Co HEAs, selecting 8 compositions after evaluating approximately 120,000 unique compositions using CALPHAD. We, then employ DFT-informed machine learning techniques to develop force-field parameters. Finally, MD simulations using these developed force-field parameters will be used to study the effects of radiation damage on the defect and mechanical properties of the selected alloys. This research explains the use of the CALPHAD approach and shows how critical modeling (DFT and MD) is in developing novel material such as HEAs. It also highlights the promising role of machine learning in this field. The results from this study will greatly improve the novel materials development to be used in next-generation accelerator components, leading to higher beam power and longer operational times of BIDs.

43 PARTICLE ACCELERATORS↗

A Fortran–Python interface for integrating machine learning parameterization into earth system models

Abstract. Parameterizations in earth system models (ESMs) are subject to biases and uncertainties arising from subjective empirical assumptions and incomplete understanding of the underlying physical processes. Recently, the growing representational capability of machine learning (ML) in solving complex problems has spawned immense interests in climate science applications. Specifically, ML-based parameterizations have been developed to represent convection, radiation, and microphysics processes in ESMs by learning from observations or high-resolution simulations, which have the potential to improve the accuracies and alleviate the uncertainties. Previous works have developed some surrogate models for these processes using ML. These surrogate models need to be coupled with the dynamical core of ESMs to investigate the effectiveness and their performance in a coupled system. In this study, we present a novel Fortran–Python interface designed to seamlessly integrate ML parameterizations into ESMs. This interface showcases high versatility by supporting popular ML frameworks like PyTorch, TensorFlow, and scikit-learn. We demonstrate the interface's modularity and reusability through two cases: an ML trigger function for convection parameterization and an ML wildfire model. We conduct a comprehensive evaluation of memory usage and computational overhead resulting from the integration of Python codes into the Fortran ESMs. By leveraging this flexible interface, ML parameterizations can be effectively developed, tested, and integrated into ESMs.

54 ENVIRONMENTAL SCIENCES↗

A Fortran-Python Interface for Integrating Machine Learning Parameterization into Earth System Models

Parameterizations in Earth System Models (ESMs) are subject to biases and uncertainties arising from subjective empirical assumptions and incomplete understanding of the underlying physical processes. Recently, the growing representational capability of machine learning (ML) in solving complex problems has spawned immense interests in climate science applications. Specifically, ML-based parameterizations have been developed to represent convection, radiation and microphysics processes in ESMs by learning from observations or high-resolution simulations, which have the potential to improve the accuracies and alleviate the uncertainties. Previous works have developed some surrogate models for these processes using ML. These surrogate models need to be coupled with the dynamical core of ESMs to investigate the effectiveness and their performance in a coupled system. In this study, we present a novel Fortran-Python interface designed to seamlessly integrate ML parameterizations into ESMs. This interface showcases high versatility by supporting popular ML frameworks like PyTorch, TensorFlow, and Scikit-learn. We demonstrate the interface's modularity and reusability through two cases: a ML trigger function for convection parameterization and a ML wildfire model. We conduct a comprehensive evaluation of memory usage and computational overhead resulting from the integration of Python codes into the Fortran ESMs. By leveraging this flexible interface, ML parameterizations can be effectively developed, tested, and integrated into ESMs.

54 ENVIRONMENTAL SCIENCES↗

Beyond Human Vision: Exploring Materials with Machine Intelligence

Machine intelligence has the potential to revolutionize materials science, enabling autonomous synthesis, self-driving characterization, and accelerated modeling. However, despite the promise, successful implementation of these methods in day-to-day research remains a challenge. This talk will delve into the reasons behind this, exploring how truly intelligent experiments are hindered by opaque experiment control, a lack of domain-specific models, and human-centric design. Through a focus on the characterization of next-generation microelectronics and energy storage materials, I will share insights from both successful and failed attempts to implement machine intelligence. We will then explore the next steps necessary to unlock the full potential of machine intelligence in materials science, creating a future where intelligent systems work seamlessly alongside researchers to drive innovation and discovery.

artificial intelligence↗

Deep Generative Models for Materials Discovery and Machine Learning-Accelerated Innovation

Machine learning and artificial intelligence (AI/ML) methods are beginning to have significant impact in chemistry and condensed matter physics. For example, deep learning methods have demonstrated new capabilities for high-throughput virtual screening, and global optimization approaches for inverse design of materials. Recently, a relatively new branch of AI/ML, deep generative models (GMs), provide additional promise as they encode material structure and/or properties into a latent space, and through exploration and manipulation of the latent space can generate new materials. These approaches learn representations of a material structure and its corresponding chemistry or physics to accelerate materials discovery, which differs from traditional AI/ML methods that use statistical and combinatorial screening of existing materials via distinct structure-property relationships. However, application of GMs to inorganic materials has been notably harder than organic molecules because inorganic structure is often more complex to encode. In this work we review recent innovations that have enabled GMs to accelerate inorganic materials discovery. We focus on different representations of material structure, their impact on inverse design strategies using variational autoencoders or generative adversarial networks, and highlight the potential of these approaches for discovering materials with targeted properties needed for technological innovation.

36 MATERIALS SCIENCE↗

The Effects of Compounded Model Size Reductions on Adversarial Robustness

Recent advances in Edge AI and Tiny Machine Learning (TinyML) have enabled the deployment of machine learning models on resource-constrained environments. However, deploying these models on edge devices, such as micro-controllers, requires significant model footprint reduction through a variety of techniques such as quantization, pruning, and clustering. While these optimization methods offer considerable advantages, they potentially introduce AI-related security vulnerabilities, particularly concerning model robustness with respect to adversarial AI attacks. Prior research has extensively examined the impact of quantization on adversarial robustness; however, the effects of alternative reduction techniques and their combinations remain understudied. This paper investigates the impact of model size reduction techniques on adversarial robustness, when applied individually and combined. We utilized Fast Gradient Sign Method (FGSM) and Projected Gradient Descent (PGD) attacks to generate adversarial perturbations for both training and testing data, and then evaluated the models' accuracy under adversarial training conditions. Our findings revealed that reduction techniques generally diminished robustness; although, combining techniques was not found to make robustness any worse than when applied individually. Moreover, specific techniques can potentially enhance resistance to small size perturbations. This research provides insights into the trade-offs between model size reduction and security, establishing a foundation for future investigations into improving adversarial training techniques and methodologies for maintaining robustness while preserving memory footprint benefits.

Austria, Phillipe [ORNL] (ORCID:0000000236223973)↗

Covalent integration of polymers and porous organic frameworks

Covalent integration of polymers and porous organic frameworks (POFs), including metal-organic frameworks (MOFs), covalent organic frameworks (COFs) and hydrogen-bonded organic frameworks (HOFs), represent a promising strategy for overcoming the existing limitations of traditional porous materials. This integration allows for the combination of the advantages of polymers, i.e., flexibility, processability and chemical versatility etc., and the superiority of POFs, like the structural integrity, tunable porosity and the high surface area, creating a type of hybrid materials. These resulting polymer-POF hybrid materials exhibit enhanced mechanical strength, chemical stability and functional diversity, thus opening up new opportunities for applications across a large variety of fields, such as gas separation, catalysis, biomedical applications, environmental remediation and energy storage. In this review, an overview of synthetic routes and strategies on how to covalently integrate different polymers with various POFs is discussed, especially with a particular focus on methods like polymerization within, on and among POF structures. To investigate the unique properties and functions of these resultant hybrid materials, the characterization techniques, including nuclear magnetic resonance spectroscopy (NMR), Fourier transform infrared spectroscopy (FTIR), X-ray diffraction (XRD), thermogravimetric analysis (TGA), transmission electron microscopy (TEM) and scanning electron microscopy (SEM), gas adsorption analysis (BET) and computational modeling and machine learning, are also presented. The ability of polymer-POFs to manipulate the pore environments at the molecular level affords these materials a wide range of applications, providing a versatile platform for future advancements in material science. Looking forward, to fully realize the potential of these hybrid materials, the authors highlight the scalability, green synthesis methods, and potential for stimuli-responsive polymer-POF materials as critical areas for future research.

Hossain, Md Amjad↗

Geomancy: Automated Performance Enhancement through Data Layout Optimization

Large distributed storage systems such as high- performance computing (HPC) systems used by national or international laboratories require sufficient performance and scale for demanding scientific workloads and must handle shifting workloads with ease. Ideally, data is placed in locations to optimize performance, but the size and complexity of large storage systems inhibit rapid effective restructuring of data layouts to maintain performance as workloads shift. To address these issues, we have developed Geomancy, a tool that models the placement of data within a distributed storage system and reacts to drops in performance. Using a combination of machine learning techniques suitable for temporal modeling, Geomancy determines when and where a bottleneck may happen due to changing workloads and suggests changes in the layout that mitigate or prevent them. Our approach to optimizing throughput offers benefits for storage systems such as avoiding potential bottlenecks and increasing overall I/O throughput from 11% to 30%.

Bel, Oceane M.↗

Forecasting influenza-like illness dynamics for military populations using neural networks and social media

This work is the first to take advantage of recurrent neural networks to predict influenza-like-illness (ILI) dynamics from various linguistic signals extracted from social media data. Unlike other approaches that rely on timeseries analysis of historical ILI data [1, 2] and the state-of-the-art machine learning models [3, 4], we build and evaluate the predictive power of Long Short Term Memory (LSTMs) architectures capable of nowcasting (predicting in \real-time") and forecasting (predicting the future) ILI dynamics in the 2011 { 2014 influenza seasons. To build our models we integrate information people post in social media e.g., topics, stylistic and syntactic patterns, emotions and opinions, and communication behavior. We then quantitatively evaluate the predictive power of different social media signals and contrast the performance of the-state-of-the-art regression models with neural networks. Finally, we combine ILI and social media signals to build joint neural network models for ILI dynamics prediction. Unlike the majority of the existing work, we specifically focus on developing models for local rather than national ILI surveillance [1], specifically for military rather than general populations [3] in 26 U.S. and six international locations. Our approach demonstrates several advantages: (a) Neural network models learned from social media data yield the best performance compared to previously used regression models. (b) Previously under-explored language and communication behavior features are more predictive of ILI dynamics than syntactic and stylistic signals expressed in social media. (c) Neural network models learned exclusively from social media signals yield comparable or better performance to the models learned from ILI historical data, thus, signals from social media can be potentially used to accurately forecast ILI dynamics for the regions where ILI historical data is not available. (d) Neural network models learned from combined ILI and social media signals significantly outperform models that rely solely on ILI historical data, which adds to a great potential of alternative public sources for ILI dynamics prediction. (e) Location-specific models outperform previously used location-independent models e.g., U.S. only. (f) Prediction results significantly vary across geolocations depending on the amount of social media data available and ILI activity patterns.

60 APPLIED LIFE SCIENCES↗

Jacobian-scaled K-means clustering for physics-informed segmentation of reacting flows

This work introduces Jacobian-scaled K-means (JSK-means) clustering, which is a physicsinformed clustering strategy centered on the K-means framework. The method allows for the injection of underlying physical knowledge into the clustering procedure through a distance function modification: instead of leveraging conventional Euclidean distance vectors, the JSKmeans procedure operates on distance vectors scaled by matrices obtained from dynamical system Jacobians evaluated at the cluster centroids. The goal of this work is to show how the JSKmeans algorithm - without modifying the input dataset - produces clusters that capture regions of dynamical similarity, in that the clusters are redistributed towards high-sensitivity regions in phase space and are described by similarity in the source terms of samples instead of the samples themselves. The algorithm is demonstrated on a complex reacting flow simulation dataset (a channel detonation configuration), where the dynamics in the thermochemical composition space are known through the highly nonlinear and stiff Arrhenius-based chemical source terms. Interpretations of cluster partitions in both physical space and composition space reveal how JSK-means shifts clusters produced by standard K-means towards regions of high chemical sensitivity (e.g., towards regions of peak heat release rate near the detonation reaction zone). Furthermore, the findings presented here illustrate the benefits of utilizing Jacobian-scaled distances in clustering techniques, and the JSK-means method in particular displays promising potential for improving former partition-based modeling strategies in reacting flow (and other multi-physics) applications.

Clustering↗

Engine Combustion System Optimization Using Computational Fluid Dynamics and Machine Learning: A Methodological Approach

Gasoline compression ignition (GCI) engines are considered an attractive alternative to traditional spark-ignition and diesel engines. Here, a Machine Learning-Grid Gradient Ascent (ML-GGA) approach was developed to optimize the performance of internal combustion engines. ML offers a pathway to transform complex physical processes that occur in a combustion engine into compact informational processes. The developed ML-GGA model was compared with a recently developed Machine Learning-Genetic Algorithm (ML-GA). Detailed investigations of optimization solver parameters and variable limit extension were performed in the present ML-GGA model to improve the accuracy and robustness of the optimization process. Detailed descriptions of the different procedures, optimization tools, and criteria that must be followed for a successful output are provided here. The developed ML-GGA approach was used to optimize the operating conditions (case 1) and the piston bowl design (case 2) of a heavy-duty diesel engine running on a gasoline fuel with a research octane number (RON) of 80. The ML-GGA approach yielded >2% improvements in the merit function, compared with the optimum obtained from a thorough computational fluid dynamics (CFD) guided system optimization. The predictions from the ML-GGA approach were validated with engine CFD simulations. This study demonstrates the potential of ML-GGA to significantly reduce the time needed for optimization problems, without loss in accuracy compared with traditional approaches.

33 ADVANCED PROPULSION SYSTEMS↗

Towards intelligent emergency control for large-scale power systems: Convergence of learning, physics, computing and control

Here, this paper has delved into the pressing need for intelligent emergency control in large-scale power systems, which are experiencing significant transformations and are operating closer to their limits with more uncertainties. Learning-based control methods are promising and have shown effectiveness for intelligent power system control. However, when they are applied to large-scale power systems, there are multifaceted challenges such as scalability, adaptiveness, and security posed by the complex power system landscape, which demand comprehensive solutions. The paper first proposes and instantiates a convergence framework for integrating power systems physics, machine learning, advanced computing, and grid control to realize intelligent grid control at a large scale. Our developed methods and platform based on the convergence framework have been applied to a large (more than 3000 buses) Texas power system, and tested with 56 000 scenarios. Our work achieved a 26% reduction in load shedding on average and outperformed existing rule-based control in 99.7% of the test scenarios. The results demonstrated the potential of the proposed convergence framework and DRL-based intelligent control for the future grid.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Two-Step Hyperparameter Optimization Method: Accelerating Hyperparameter Search by Using a Fraction of a Training Dataset

Abstract Hyperparameter optimization (HPO) is an important step in machine learning (ML) model development, but common practices are archaic—primarily relying on manual or grid searches. This is partly because adopting advanced HPO algorithms introduces added complexity to the workflow, leading to longer computation times. This poses a notable challenge to ML applications, as suboptimal hyperparameter selections curtail the potential of ML model performance, ultimately obstructing the full exploitation of ML techniques. In this article, we present a two-step HPO method as a strategic solution to curbing computational demands and wait times, gleaned from practical experiences in applied ML parameterization work. The initial phase involves a preliminary evaluation of hyperparameters on a small subset of the training dataset, followed by a reevaluation of the top-performing candidate models postretraining with the entire training dataset. This two-step HPO method is universally applicable across HPO search algorithms, and we argue it has attractive efficiency gains. As a case study, we present our recent application of the two-step HPO method to the development of neural network emulators for aerosol activation. Although our primary use case is a data-rich limit with many millions of samples, we also find that using up to 0.0025% of the data—a few thousand samples—in the initial step is sufficient to find optimal hyperparameter configurations from much more extensive sampling, achieving up to 135× speedup. The benefits of this method materialize through an assessment of hyperparameters and model performance, revealing the minimal model complexity required to achieve the best performance. The assortment of top-performing models harvested from the HPO process allows us to choose a high-performing model with a low inference cost for efficient use in global climate models (GCMs).

97 MATHEMATICS AND COMPUTING↗

Going off topics to demix quark and gluon jets in α S extractions

Quantum chromodynamics is the theory of the strong interaction between quarks and gluons; the coupling strength of the interaction, α S , is the least precisely-known of all interactions in nature. An extraction of the strong coupling from the radiation pattern within jets would provide a complementary approach to conventional extractions from jet production rates and hadronic event shapes, and would be a key achievement of jet substructure at the Large Hadron Collider (LHC). Presently, the relative fraction of quark and gluon jets in a sample is the limiting factor in such extractions, as this fraction is degenerate with the value of αS for the most well-understood observables. To overcome this limitation, we apply recently proposed techniques to statistically demix multiple mixtures of jets and obtain purified quark and gluon distributions based on an operational definiton. We illustrate that studying quark and gluon jet substructure separately can significantly improve the sensitivity of such extractions of the strong coupling. We also discuss how using machine learning techniques or infrared- and collinear-unsafe information can improve the demixing performance without the loss of theoretical control. While theoretical research is required to connect the extract topics with the quark and gluon objects in cross section calculations, our study illustrates the potential of demixing to reduce the dominant uncertainty for the α S extraction from jet substructure at the LHC.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

ND 2 AV: N-dimensional data analysis and visualization analysis for the National Ignition Campaign

Here, one of the biggest challenges in high-energy physics is to analyze a complex mix of experimental and simulation data to gain new insights into the underlying physics. Currently, this analysis relies primarily on the intuition of trained experts often using nothing more sophisticated than default scatter plots. Many advanced analysis techniques are not easily accessible to scientists and not flexible enough to explore the potentially interesting hypotheses in an intuitive manner. Furthermore, results from individual techniques are often difficult to integrate, leading to a confusing patchwork of analysis snippets too cumbersome for data exploration. This paper presents a case study on how a combination of techniques from statistics, machine learning, topology, and visualization can have a significant impact in the field of inertial confinement fusion. We present the $\mathrm{ND}^2\mathrm{AV}$: N-dimensional data analysis and visualization framework, a user-friendly tool aimed at exploiting the intuition and current workflow of the target users. The system integrates traditional analysis approaches such as dimension reduction and clustering with state-of-the-art techniques such as neighborhood graphs and topological analysis, and custom capabilities such as defining combined metrics on the fly. All components are linked into an interactive environment that enables an intuitive exploration of a wide variety of hypotheses while relating the results to concepts familiar to the users, such as scatter plots. $\mathrm{ND}^2\mathrm{AV}$ uses a modular design providing easy extensibility and customization for different applications. $\mathrm{ND}^2\mathrm{AV}$ is being actively used in the National Ignition Campaign and has already led to a number of unexpected discoveries.

97 MATHEMATICS AND COMPUTING↗

Improving Prediction of Surface Solar Irradiance Variability by Integrating Observed Cloud Characteristics and Machine Learning

A 5-year, 1-minute resolution observational dataset of clouds and solar radiation was produced that includes two metrics of the variability in surface solar irradiance due to cloud type and fractional sky cover. Multiple regression models were trained to fit observations of surface solar irradiance variability from those two cloud property predictors. We found that ensemble tree-based methods, Random Forest and Gradient Boosting Machine, have the least overfitting issues and showed the best performance with an R2 of 0.42. While the observational data trained in this study was only from one site, the U.S. Department of Energy (DOE) Atmospheric Radiation Measurement (ARM) Southern Great Plains (SGP) site in Oklahoma, initial comparisons of the seasonality of the statistics suggest that these results are relatively weather regime independent; the generality of such a finding across sites will be tested in future work. The observational data and developed machine learning model are being used to create a numerical weather prediction model parameterization to enable day-ahead solar variability prediction in a computationally efficient way. This is a first step towards creating a new paradigm of predicting day-ahead variability with the potential to provide a new tool to improve grid operation, planning, and resilience.

Riihimaki, Laura↗