Search NASA⌕ Search

SEARCH · Search NASA

Results for “machine”

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 451 records · Page 25
All sourcesNASA NTRSDOE OSTI
NASA NTRS2003

Intelligent machines in the twenty-first century: foundations of inference and inquiry

The last century saw the application of Boolean algebra to the construction of computing machines, which work by applying logical transformations to information contained in their memory. The development of information theory and the generalization of Boolean algebra to Bayesian inference have enabled these computing machines, in the last quarter of the twentieth century, to be endowed with the ability to learn by making inferences from data. This revolution is just beginning as new computational techniques continue to make difficult problems more accessible. Recent advances in our understanding of the foundations of probability theory have revealed implications for areas other than logic. Of relevance to intelligent machines, we recently identified the algebra of questions as the free distributive algebra, which will now allow us to work with questions in a way analogous to that which Boolean algebra enables us to work with logical statements. In this paper, we examine the foundations of inference and inquiry. We begin with a history of inferential reasoning, highlighting key concepts that have led to the automation of inference in modern machine-learning systems. We then discuss the foundations of inference in more detail using a modern viewpoint that relies on the mathematics of partially ordered sets and the scaffolding of lattice theory. This new viewpoint allows us to develop the logic of inquiry and introduce a measure describing the relevance of a proposed question to an unresolved issue. Last, we will demonstrate the automation of inference, and discuss how this new logic of inquiry will enable intelligent machines to ask questions. Automation of both inference and inquiry promises to allow robots to perform science in the far reaches of our solar system and in other star systems by enabling them not only to make inferences from data, but also to decide which question to ask, which experiment to perform, or which measurement to take given what they have learned and what they are designed to understand.

Review↗
NASA NTRS2004

Design of Ultra-High-Power-Density Machine Optimized for Future Aircraft

The NASA Glenn Research Center's Structural Mechanics and Dynamics Branch is developing a compact, nonpolluting, bearingless electric machine with electric power supplied by fuel cells for future "more-electric" aircraft with specific power in the projected range of 50 hp/lb, whereas conventional electric machines generate usually 0.2 hp/lb. The use of such electric drives for propulsive fans or propellers depends on the successful development of ultra-high-power-density machines. One possible candidate for such ultra-high-power-density machines, a round-rotor synchronous machine with an engineering current density as high as 20,000 A/sq cm, was selected to investigate how much torque and power can be produced.

Choi, Benjamin B.↗
NASA NTRS2004

Fast Query-Optimized Kernel-Machine Classification

A recently developed algorithm performs kernel-machine classification via incremental approximate nearest support vectors. The algorithm implements support-vector machines (SVMs) at speeds 10 to 100 times those attainable by use of conventional SVM algorithms. The algorithm offers potential benefits for classification of images, recognition of speech, recognition of handwriting, and diverse other applications in which there are requirements to discern patterns in large sets of data. SVMs constitute a subset of kernel machines (KMs), which have become popular as models for machine learning and, more specifically, for automated classification of input data on the basis of labeled training data. While similar in many ways to k-nearest-neighbors (k-NN) models and artificial neural networks (ANNs), SVMs tend to be more accurate. Using representations that scale only linearly in the numbers of training examples, while exploring nonlinear (kernelized) feature spaces that are exponentially larger than the original input dimensionality, KMs elegantly and practically overcome the classic curse of dimensionality. However, the price that one must pay for the power of KMs is that query-time complexity scales linearly with the number of training examples, making KMs often orders of magnitude more computationally expensive than are ANNs, decision trees, and other popular machine learning alternatives. The present algorithm treats an SVM classifier as a special form of a k-NN. The algorithm is based partly on an empirical observation that one can often achieve the same classification as that of an exact KM by using only small fraction of the nearest support vectors (SVs) of a query. The exact KM output is a weighted sum over the kernel values between the query and the SVs. In this algorithm, the KM output is approximated with a k-NN classifier, the output of which is a weighted sum only over the kernel values involving k selected SVs. Before query time, there are gathered statistics about how misleading the output of the k-NN model can be, relative to the outputs of the exact KM for a representative set of examples, for each possible k from 1 to the total number of SVs. From these statistics, there are derived upper and lower thresholds for each step k. These thresholds identify output levels for which the particular variant of the k-NN model already leans so strongly positively or negatively that a reversal in sign is unlikely, given the weaker SV neighbors still remaining. At query time, the partial output of each query is incrementally updated, stopping as soon as it exceeds the predetermined statistical thresholds of the current step. For an easy query, stopping can occur as early as step k = 1. For more difficult queries, stopping might not occur until nearly all SVs are touched. A key empirical observation is that this approach can tolerate very approximate nearest-neighbor orderings. In experiments, SVs and queries were projected to a subspace comprising the top few principal- component dimensions and neighbor orderings were computed in that subspace. This approach ensured that the overhead of the nearest-neighbor computations was insignificant, relative to that of the exact KM computation.

Mazzoni, Dominic↗
NASA NTRS2008

Review, Selection and Installation of a Rapid Prototype Machine

The objective of this paper is to impress upon the reader the benefits and advantages of investing in rapid prototyping (additive manufacturing) technology thru the procurement of one or two new rapid prototyping machines and the creation of a new Prototype and Model Lab at the Kennedy Space Center (KSC). This new resource will be available to all of United Space Alliance, LLC (USA), enabling engineers from around the company to pursue a more effective means of communication and design with our co-workers, and our customer, the National Aeronautics and Space Administration (NASA). The Rapid Protoyping/3D printing industry mirrors the transition the CAD industry made several years ago, when companies were trying to justify the expenditure of converting to a 3D based system from a 2D based system. The advantages of using a 3D system seemed to be outweighed by the cost it would take to convert not only legacy 2D drawings into 3D models but the training of personnel to use the 3D CAD software. But the reality was that when a 3D CAD system is employed, it gives engineers a much greater ability to conceive new designs and the ability to engineer new tools and products much more effectively. Rapid Prototyping (RP) is the name given to a host of related technologies that are used to fabricate physical objects directly from Computer Aided Design (CAD) data sources. These methods are generally similar to each other in that they add and bond materials in a layer wise-fashion to form objects, instead of machining away material. The machines used in Rapid Prototyping are also sometimes referred to as Rapid Manufacturing machines due to the fact that some of the parts fabricated in a RP machine can be used as the finished product. The name "Rapid Prototyping" is really a misnomer. It is much more than prototypes and it is not always rapid.

McEndree, Caryl↗
NASA NTRS2018

Creating a Multifunctional Composite Stator Slot Material System to Enable High Power Density Electric Machines for Electrified Aircraft Applications

Increasing the power density and efficiency of electric machines (motors and generators) is integral to bringing Electrified Aircraft (EA) to commercial realization. However, power density and efficiency are not qualities that can be developed independently. At the heart of any electric machine are the conductors (usually copper) that carry current and generate magnetic fields. Increased power density means increased current density and increased joule heating in a smaller volume. To increase efficiency at the wire level means minimizing electrical resistance and hence power lost to joule heating. There are fundamental challenges with concomitantly increasing both power density and efficiency since the copper resistivity is very temperature sensitive at common electric machine operating conditions. Simple calculations of the linear increase in resistivity of copper as a function of temperature, reveals that a one degree Celsius increase in temperature results in a 0.39% decrease in efficiency. Conversely, a 20 degree Celsius decrease in copper temperature produces a 7.8% increase in efficiency. Therefore, improved thermal management concepts for electric machine building blocks such as stator winding are a priority for improving efficiency and power density. This paper proposing changing the view of component materials in the stator slots from individual components with singular functionality to a composite system where the components take on a multifunctional roles. In the composite framework, achievable material development goals are defined that together have maximum system impact on the thermal environment inside of high power density electric machines for aerospace applications.

thermal management↗
NASA NTRS2019

Machine Learning Prototype App For Recognition of Fruits

As the incidence of obesity and associated negative health consequences is rising, it becomes crucial to monitor the dietary choices of individuals. Unfortunately, traditional methods to collect this information involve collecting food frequency questionnaires from individuals using paper. Electronic food trackers have been developed to collect food data, but they require participants to manually label and describe the content of their meals, and which may be difficult for researchers to interpret in a standardized fashion. Machine learning, however, provides an easy and efficient method for both participants and researchers to label food items with standardized descriptions. This project aims to create a prototype phone application that can identify and label photos of apples. This is done by making a machine learning model through Turicreate, a python module, which is then implemented into an iOS app through Xcode and Swift. The modules used in Swift include CoreML and AVFoundation. This machine learning application will be incorporated with a MealLogger phone app that is also under development. The MealLogger app will be used to keep track of participants' calorie intake and other personal details throughout the sleep study. The machine learning model will present several potential identities of the foods found in the photo, and the user will only need to select the correct option. This will be a user-friendly method for participants to easily log their food consumption without the hard work of manually inputting each and every description. Some limitations to this project include the wide variety of food, including those within different cultures. To deal with this, the model will include the most generic food categories, which the participant may select, and produce a drop-down menu of more specific dishes under that specified category, with the option of self-input. Additional questionnaires may be implemented according to the food type selected This will allow the process to be quick and easy, but also specific for the purpose of analysis. The release of the application will require a much longer process, but the machine learning prototype presents a first step toward an application that may change data analysis for researchers interested in collecting food intake from individuals living in the real world.

Food tracker↗
NASA NTRS2019

Reusing Data and Metadata to Create New Metadata Through Machine-Learning & Other Programmatic Methods

Recent improvements in natural language processing (NLP) enable metadata to be created programmatically from reused original metadata or even the dataset itself. Transfer-learning applied to NLP has greatly improved performance and reduced training data requirements. In this talk, we’ll compare machine-generated metadata to human-generated metadata and discuss characteristics of metadata and data archives that affect suitability for machine-learning reuse of metadata. Where as human-generated metadata is often populated once, populated from the perspective of data supplier, populated by many individuals with different words for the same thing, and limited in length, machine-generated metadata can be updated any number of times, generated from the perspective of any user, constrained to a standardized set of terms that can be evolved over time, and be any length required. Machine-learning generated metadata offers benefits but also additional needs in terms of version control, process transparency, human-computer interaction, and IT requirements. As a successful example, we’ll discuss how a dataset of abstracts and associated human-tagged keywords from a standardized list of several thousand keywords were used to create a machine-learning model that predicted keyword metadata for open-source code projects on code.nasa.gov. We’ll also discuss a less successful example from data.nasa.gov to show how data archive architecture and characteristics of initial metadata can be strong controls on how easy it is to leverage programmatic methods to reuse metadata to create additional metadata.

Gosses, Justin↗
NASA NTRS2020

Machine Learning-Based Predictive Analytics for Aircraft Engine Conceptual Design

Big data and artificial intelligence/machine learning are transforming the global business environment. Data is now the most valuable asset for enterprises in every industry. Companies are using data-driven insights for competitive advantage. With that, the adoption of machine learning-based data analytics is rapidly taking hold across various industries, producing autonomous systems that support human decision-making. This work explored the application of machine learning to aircraft engine conceptual design. Supervised machine-learning algorithms for regression and classification were employed to study patterns in an existing, open-source database of production and research turbofan engines, and resulting in predictive analytics for use in predicting performance of new turbofan designs. Specifically, the author developed machine learning-based analytics to predict cruise thrust specific fuel consumption (TSFC) and core sizes of high-efficiency turbofan engines, using engine design parameters as the input. The predictive analytics were trained and deployed in Keras, an open-source neural networks application program interface (API) written in Python, with Google’s TensorFlow (an open source library for numerical computation) serving as the backend engine. The promising results of the predictive analytics show that machine-learning techniques merit further exploration for application in aircraft engine conceptual design.

deep-learning↗
NASA NTRSDate not supplied

ImageLabler: Labeling and Managing Image Data for Machine Learning in the Earth Sciences

While machine learning techniques for image classification have been around for a long time, storing and managing the vast number of images required as training data is still a problem for scientists. This is especially true for the field of Earth science, where only recently have experts begun using machine learning techniques for image-based phenomena classification. Image Labeler, a fast and scalable cloud-based tagging platform for Earth science images, seeks to improve upon existing methods of managing images and associated metadata, such as maintaining categorized folders of images on a local machine, a process that can be cumbersome and difficult to scale. The platform facilitates rapid development of image-based Earth science phenomena training datasets by allowing scientists to upload their existing imagery as well as extract new samples from open satellite imagery services made available through NASA’s Global Imagery Browse Service (GIBS). Image Labeler also supports GeoTIFF data, with capabilities such as displaying GeoTIFFs on an interactive map, drawing shapefiles over them, and tagging them with additional metadata. This allows scientists to perform spatiotemporal subsetting with geographic information and develop training data more quickly. Built using modern web technologies, Image Labeler includes additional capabilities such as team collaboration for large-scale image tagging projects. Users can download their data in a machine-learning-ready format, allowing scientists to spend time on experimentation rather than on the collection of training data. In this presentation, we demonstrate how Image Labeler seeks to become a one-stop image data management solution for machine learning applications in Earth science.

Ashish Acharya↗
NASA NTRSDate not supplied

Using Machine Learning to Infer Material Properties of Debris Fragments from X-ray Images in the DebriSat Project

The DebriSat project is a collaboration effort with the NASA Orbital Debris Program Office, the U.S. Space Force Space Systems Command Center, The Aerospace Corporation, and the University of Florida. To date, over 200,000 fragments from this ground-based, hypervelocity impact experiment have been collected, and processing is underway to determine their physical characteristics, such as material, shape, color, characteristic length, and average cross-sectional area. The x-ray process is primarily used to identify the location of the fragments and estimated size for extraction, so that these physical characteristics can be assessed. This paper proposes a machine learning-based approach to characterize materials from x-ray images of debris fragments embedded in soft-catch foam used in the DebriSat project. The novel methodology discussed in this paper will highlight the use of x-ray imagery data to characterize these fragments without extraction or a human-in-the-loop. Both supervised and unsupervised machine learning techniques are utilized with this approach to infer the physical parameters of the fragments embedded in the soft-catch foam panels used in the impact experiment based on x-ray images of the foam panels. Additionally, 3D reconstructions of the extracted fragments are created with images taken from two different angles using the structure from motion (SfM) method. The characteristic lengths and shape from the 3D reconstruction, alongside the physical characteristics of the debris, are used in the inference of the material type. To develop and test the approach, a dataset of x-ray images of debris fragments of varying sizes and materials is collected. Supervised learning methods such as convolutional neural networks (CNNs), support vector machines (SVM), decision trees, and random forest classifiers are used due to the high-dimensional feature spaces of the debris and nonlinear decision boundaries for material categorization. Given the limited pre-labeled data of embedded debris materials smaller than 10 mm, unsupervised machine learning techniques such as clustering algorithms and autoencoders are used, in addition to supervised learning methods. The clustering algorithms group similar fragments together based on their physical properties, and autoencoders reduce the dimensionality of the x ray images and extract relevant features. The performance of the proposed approach's is analyzed using a range of statistical methods, including confusion matrices, receiver operating characteristic curves, and precision-recall curves. The results are compared with those obtained using a baseline approach that relies on manual identification and classification of debris fragments. To evaluate the effectiveness of different machine learning methods, statistical tests such as t-tests, ANOVA, and cross-validation are performed, comparing the performance of CNNs, SVMs, clustering algorithms, and autoencoders. Additional analysis needs to be conducted to identify any sources of bias or variability that may affect the results, such as variations in imaging conditions or fragmentation patterns. Other topics explored are limitations, refinements, and the potential use of semi-supervised learning techniques, such as self-training to label unlabeled datasets and co-training using x-ray images taken from two different angles as two different models.

Saik Anam Siam↗
NASA NTRS2023

Using Machine Learning to Infer Material Properties of Debris Fragments from X-ray Images in the DebriSat Project

The DebriSat project is a collaboration effort with the NASA Orbital Debris Program Office, the U.S. Space Force Space Systems Command Center, The Aerospace Corporation, and the University of Florida. To date, over 200,000 fragments from this ground-based, hypervelocity impact experiment have been collected, and processing is underway to determine their physical characteristics, such as material, shape, color, characteristic length, and average cross-sectional area. The x-ray process is primarily used to identify the location of the fragments and estimated size for extraction, so that these physical characteristics can be assessed. This paper proposes a machine learning-based approach to characterize materials from x-ray images of debris fragments embedded in soft-catch foam used in the DebriSat project. The novel methodology discussed in this paper will highlight the use of x-ray imagery data to characterize these fragments without extraction or a human-in-the-loop. Both supervised and unsupervised machine learning techniques are utilized with this approach to infer the physical parameters of the fragments embedded in the soft-catch foam panels used in the impact experiment based on x-ray images of the foam panels. Additionally, 3D reconstructions of the extracted fragments are created with images taken from two different angles using the structure from motion (SfM) method. The characteristic lengths and shape from the 3D reconstruction, alongside the physical characteristics of the debris, are used in the inference of the material type. To develop and test the approach, a dataset of x-ray images of debris fragments of varying sizes and materials is collected. Supervised learning methods such as convolutional neural networks (CNNs), support vector machines (SVM), decision trees, and random forest classifiers are used due to the high-dimensional feature spaces of the debris and nonlinear decision boundaries for material categorization. Given the limited pre-labeled data of embedded debris materials smaller than 10 mm, unsupervised machine learning techniques such as clustering algorithms and autoencoders are used, in addition to supervised learning methods. The clustering algorithms group similar fragments together based on their physical properties, and autoencoders reduce the dimensionality of the x ray images and extract relevant features. The performance of the proposed approach's is analyzed using a range of statistical methods, including confusion matrices, receiver operating characteristic curves, and precision-recall curves. The results are compared with those obtained using a baseline approach that relies on manual identification and classification of debris fragments. To evaluate the effectiveness of different machine learning methods, statistical tests such as t-tests, ANOVA, and cross-validation are performed, comparing the performance of CNNs, SVMs, clustering algorithms, and autoencoders. Additional analysis needs to be conducted to identify any sources of bias or variability that may affect the results, such as variations in imaging conditions or fragmentation patterns. Other topics explored are limitations, refinements, and the potential use of semi-supervised learning techniques, such as self-training to label unlabeled datasets and co-training using x-ray images taken from two different angles as two different models.

Saik Anam Siam↗
NASA NTRS2024

EdgeCortix SAKURA-I Machine-Learning, PCIe Accelerator SEE Heavy Ion Test Report

To enable autonomy in space, machine-learning and computer vision applications become invaluable for sensor processing. However, these algorithms are computationally complex and unfeasible for many embedded central processing units (CPUs) and usually require external coprocessors, such as graphics processing units (GPUs) or accelerators specific to the application, including application specific integrated circuits (ASICs). In power-constrained systems, GPUs tend to consume more power than is acceptable (>40W), so lower-power accelerators have shown promise to provide the performance needed under spacecraft constraints. For radiation engineers, developing methodologies that can properly test CPUs, GPUs, and accelerators, and enable comparisons between them remains a necessary complication to solve as the devices become more complex. The methodology in this test aims to be a start in developing a baseline single-event effect (SEE) test for client-device machine learning accelerators. This category of devices do not host their own operating system. This testing campaign is a continuation of a previous 200 MeV proton test performed in January 2024. This report covers two heavy ion tests of the SAKURA-I card: one in April 2024, and one in June 2024. Additional data was needed after the April test due to ion-range issues experienced at higher linear-energy transfers (LETs). These range issues are described in more detail in Section 8. This experiment characterizes SEEs and data error susceptibility of the EdgeCortix SAKURA-I machine-learning accelerator under heavy ions. The device was monitored for single event upsets (SEUs) and single event functional interrupts (SEFIs) at the Lawrence Berkeley National Laboratory’s 88-inch cyclotron. The SAKURA-I board accelerates machine-learning inference applications on a host computer through a PCIex16 connection. For the purposes of devising an end to end automated analysis workflow for this experiment, the YOLO-V5 and SSD300 objection-detection models, and the ResNet-50, EfficientNet, and MobileNetV2 image classification models were used as a representative suite of analytical machine-learning models.

Seth S Roffe↗
DOE OSTI2011

Machine-Induced Showers Entering the Atlas and CMS Detectors in the LHC

One source of experimental background in the LHC is showers induced by particles hitting the upstream collimators or particles that have been scattered on the residual gas. We estimate the flux and distribution of particles entering the ATLAS and CMS detectors through FLUKA simulations starting either in the tertiary collimators or with inelastic beam-gas interactions. Comparisons to MARS15 results are also presented. Our results can be used as a source term for further simulations of the machine-induced background in the experimental detectors. To ensure optimal performance of the LHC experimental detectors, it is important to understand the background, which can come fromseveral sources. In this article we discuss machine-induced background, caused either by nearby beam losses or interactions between beam particles and the residual gas inside the vacuum pipe. Beam losses outside the experimental interaction regions (IRs) are unavoidable during collider operation. The halo is continuously repopulated and has to be cleaned by the collimation system, so that the losses in the cold magnets are kept at a safe level. The collimation system is located in two dedicated insertions (IR3 and IR7) but a small leakage of secondary and tertiary halo is expected to escape. Some particles make it to the experimental IRs, where they are intercepted by tertiary collimators (TCTs) that are installed in order to protect the inner triplet magnets. Some parts of the induced high-energy shower can escape and propagate into the detectors. Another source of background is beam-gas interactions. Beam protons can scatter elastically or inelastically on residual gas molecules. If an inelastic interaction occurs close to the detector, it causes a shower that could reach the detector. Elastic interactions can scatter protons directly onto the TCTs without passing IR7, which has to be treated separately from the beam-halo losses discussed above. Machine-induced background can also originate from a cross-talk between different IPs. In this article we focus on beam-halo losses and inelastic beam-gas interactions two LHC experiments: ATLAS in IR1 and CMS in IR5. We compare also to previous results. We simulate the machine-induced showers propagating through the interaction region up to an interface plane between the machine and the detector, which is defined to be at 22.6m from the interaction point (IP) along the beam direction. The coordinates and momenta of the particles crossing this plane are recorded and can be used as a source term for further simulations of the detector itself. The interface plane extends to 30 m radially, although 99% of the total energy was found within a 2.5 m radius in the simulations.

43 PARTICLE ACCELERATORS↗
DOE OSTI2012

Methods, systems and apparatus for controlling operation of two alternating current (AC) machines

A system is provided for controlling two AC machines. The system comprises a DC input voltage source that provides a DC input voltage, a voltage boost command control module (VBCCM), a five-phase PWM inverter module coupled to the two AC machines, and a boost converter coupled to the inverter module and the DC input voltage source. The boost converter is designed to supply a new DC input voltage to the inverter module having a value that is greater than or equal to a value of the DC input voltage. The VBCCM generates a boost command signal (BCS) based on modulation indexes from the two AC machines. The BCS controls the boost converter such that the boost converter generates the new DC input voltage in response to the BCS. When the two AC machines require additional voltage that exceeds the DC input voltage required to meet a combined target mechanical power required by the two AC machines, the BCS controls the boost converter to drive the new DC input voltage generated by the boost converter to a value greater than the DC input voltage.

47 OTHER INSTRUMENTATION↗
DOE OSTI2013

Method and apparatus for characterizing and enhancing the functional performance of machine tools

Disclosed are various systems and methods for assessing and improving the capability of a machine tool. The disclosure applies to machine tools having at least one slide configured to move along a motion axis. Various patterns of dynamic excitation commands are employed to drive the one or more slides, typically involving repetitive short distance displacements. A quantification of a measurable merit of machine tool response to the one or more patterns of dynamic excitation commands is typically derived for the machine tool. Examples of measurable merits of machine tool performance include workpiece surface finish, and the ability to generate chips of the desired length.

42 ENGINEERING↗
DOE OSTI2021

Design Space Exploration of Emerging Memory Technologies for Machine Learning Applications

Memory design space exploration methods study memory systems’ performances and limitations before implementation. The computer memory design space has grown exponentially because of the enormous growth of memory types, memory controllers, and application software. Computer simulators are commonly used for memory design space exploration. However, complex memory simulations take an enormous amount of time. Hence, in this paper, we proposed a machine learning-based design space exploration method for dynamic random-access memory and non-volatile memory systems. We applied our method to the CosmoGAN and LeNet applications to predict the following six memory response parameters: (i) bandwidth, (ii) power, (iii) average latency, (iv) average total latency, (v) memory reads, and (vi) memory writes. Our experimental results show that machine learning models can predict memory response parameter values faster than simulations. We used support vector machine, random forest, and gradient boosting machine learning models. We observed that the support vector machine provides better performance for bandwidth, average latency, and average total latency. The random forest model works better for memory reads and writes. The gradient boosting model provides superior prediction performance for power. We provide a detailed discussion on learning curve characteristics, error analysis, and memory type recommendation.

Hasan, S M Shamimul↗
DOE OSTI2021

Recent Advances in Analysis and Design of Axial Flux Permanent Magnet Electric Machines

Axial flux permanent magnet machines (AFPM) are popular for applications that benefit from high torque density and an axially compact form factor, such as in-wheel traction drives. Although the radial flux permanent magnet machine (RFPM) and the AFPM work based on the same underlying principle, the differences in their geometry introduce complexities in analysis of the AFPM. In this paper, the different AFPM design variants, their sizing approaches, computationally efficient design optimization techniques, and manufacturing techniques reported in literature are reviewed. In addition to classical AFPM machines, emerging variants and research opportunities with potential to push the boundaries of electric machine technology are reviewed. These include bearingless AFPMs, magnetically geared AFPMs, and combined radial-axial flux machines.

Nishanth, FNU↗
DOE OSTI2024

Emerging Jets Search, Triton Server Deployment, and Track Quality Development: Machine Learning Applications in High Energy Physics

Machine learning is becoming prevalent in high energy physics, with numerous applications in physics analyses and event reconstruction showing great improvements compared to traditional computing methods. This thesis studies three projects which each propose new avenues for machine learning applications within the high energy physics CMS experiment located at CERN. In the first project, a search for a dark matter signal called “emerging jets” is performed, using graph neural networks to greatly increase sensitivity to the signal’s signature within the data. The result of this dark matter search sets the most stringent exclusion limits to date on theoretical emerging jet models. Motivated by inefficiencies encountered when processing the emerging jet graph neural network at Fermi National Accelerator Laboratory’s computing centers, the second project re-optimizes the computing centers for machine learning inference. This re-optimization uses NVIDIA Triton Inference Servers to process users’ analysis code heterogeneously, therefore achieving high processing throughput and decreasing user time-to-insight. The last project focuses on an upgrade to the CMS experiment’s real-time event selection system which improves physics object reconstruction under harsh processing conditions. A boosted decision tree is used to quickly and efficiently quantify a reconstructed particle’s “track quality” in order to remove particle tracks reconstructed erroneously. In summary, this thesis will not only present examples of how high energy physics can greatly benefit by leveraging machine learning techniques for physics analysis and reconstruction, but will also provide guidance on how the field can prepare for the inevitable increase in machine learning applications.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗