Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Real Time Optimization”

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

Piloted simulation of an on-board trajectory optimization algorithm

This paper will describe a real time piloted simulation of algorithms designed for on-board computation of time-optimal intercept trajectories for an F-8 aircraft. The algorithms, which were derived using singular perturbation theory, generate commands that are displayed to the pilot on flight director needles on the 8-ball. By flying the airplane so as to zero the horizontal and vertical needles, the pilot flies an approximation to a time-optimal intercept trajectory. The various display and computation modes that are available will be described and results will be presented illustrating the performance of the algorithms with a pilot in the loop.

Price, D. B.↗

Data and Communications in Basic Energy Sciences: Creating a Pathway for Scientific Discovery

This report is based on the Department of Energy (DOE) Workshop on “Data and Communications in Basic Energy Sciences: Creating a Pathway for Scientific Discovery” that was held at the Bethesda Marriott in Maryland on October 24-25, 2011. The workshop brought together leading researchers from the Basic Energy Sciences (BES) facilities and Advanced Scientific Computing Research (ASCR). The workshop was co-sponsored by these two Offices to identify opportunities and needs for data analysis, ownership, storage, mining, provenance and data transfer at light sources, neutron sources, microscopy centers and other facilities. Their charge was to identify current and anticipated issues in the acquisition, analysis, communication and storage of experimental data that could impact the progress of scientific discovery, ascertain what knowledge, methods and tools are needed to mitigate present and projected shortcomings and to create the foundation for information exchanges and collaboration between ASCR and BES supported researchers and facilities. The workshop was organized in the context of the impending data tsunami that will be produced by DOE’s BES facilities. Current facilities, like SLAC National Accelerator Laboratory’s Linac Coherent Light Source, can produce up to 18 terabytes (TB) per day, while upgraded detectors at Lawrence Berkeley National Laboratory’s Advanced Light Source will generate ~10TB per hour. The expectation is that these rates will increase by over an order of magnitude in the coming decade. The urgency to develop new strategies and methods in order to stay ahead of this deluge and extract the most science from these facilities was recognized by all. The four focus areas addressed in this workshop were: Workflow Management - Experiment to Science: Identifying and managing the data path from experiment to publication. Theory and Algorithms: Recognizing the need for new tools for computation at scale, supporting large data sets and realistic theoretical models. Visualization and Analysis: Supporting near-real-time feedback for experiment optimization and new ways to extract and communicate critical information from large data sets. Data Processing and Management: Outlining needs in computational and communication approaches and infrastructure needed to handle unprecedented data volume and information content. It should be noted that almost all participants recognized that there were unlikely to be any turn-key solutions available due to the unique, diverse nature of the BES community, where research at adjacent beamlines at a given light source facility often span everything from biology to materials science to chemistry using scattering, imaging and/or spectroscopy. However, it was also noted that advances supported by other programs in data research, methodologies, and tool development could be implemented on reasonable time scales with modest effort. Adapting available standard file formats, robust workflows, and in-situ analysis tools for user facility needs could pay long-term dividends. Workshop participants assessed current requirements as well as future challenges and made the following recommendations in order to achieve the ultimate goal of enabling transformative science in current and future BES facilities: Theory and analysis components should be integrated seamlessly within experimental workflow. Develop new algorithms for data analysis based on common data formats and toolsets. Move analysis closer to experiment. Move the analysis closer to the experiment to enable real-time (in-situ) streaming capabilities, live visualization of the experiment and an increase of the overall experimental efficiency. Match data management access and capabilities with advancements in detectors and sources. Remove bottlenecks, provide interoperability across different facilities/beamlines and apply forefront mathematical techniques to more efficiently extract science from the experiments. This workshop report examines and reviews the status of several BES facilities and highlights the successes and shortcomings of the current data and communication pathways for scientific discovery. It then ascertains what methods and tools are needed to mitigate present and projected data bottlenecks to science over the next 10 years. The goal of this report is to create the foundation for information exchanges and collaborations among ASCR and BES supported researchers, the BES scientific user facilities, and ASCR computing and networking facilities. To jumpstart these activities, there was a strong desire to see a joint effort between ASCR and BES along the lines of the highly successful Scientific Discovery through Advanced Computing (SciDAC) program in which integrated teams of engineers, scientists and computer scientists were engaged to tackle a complete end-to-end workflow solution at one or more beamlines, to ascertain what challenges will need to be addressed in order to handle future increases in data

97 MATHEMATICS AND COMPUTING↗

Designing Monte Carlo Simulation and an Optimal Machine Learning to Optimize and Model Space Missions

This paper investigates applying artificial intelligence (AI) algorithms to attitude control system of satellites to optimally tune the controller using high performance computing. This methodology is applied to the Virtual Telescope for X-ray Observation mission, which is a precise formation of two separate spacecraft observing multiple objects in the space in the X-ray domain. The mission is divided into phases based on the instrumentation and the mission goal. To reach an stable precise formation robust to stochastic slew and slew rate (i.e., Euler angles and angular velocities) in a minimal constrained time T , consumed energy of the attitude control system, denoted as E, and root-mean-square state error of attitude control system, denoted as e, are minimized. Monte-Carlo simulation is used for the sensitivity analysis of optimization and designing a controller. Deep neural networks (DNN), Gaussian processes (GP), and support vector regression (SVR) learn this optimization as a surrogate model, while their hyperparameters are optimized in a novel approach. THETA supercomputer at Argonne Leadership Computing Facility (ALCF) is used for optimizing the hyperparameters of DNN. The surrogate model meets the requirements of the mission, and it shows a better performance over the optimization and Monte-Carlo. The optimal DNN can satisfy the mission requirements e and T while reducing E for 90% compared to the other given methods.

42 ENGINEERING↗

Electrodeposition of Actinides on a Semiconductor Detector for Concentration Monitoring

Pyroprocessing is a promising technology for separation and recycle of actinides from spent nuclear fuel. It is being actively researched in the United States as well as the Republic of Korea for possible implementation into a closed nuclear fuel cycle. The features of the chemical technology behind pyro-processing, unfortunately, present difficulties for real time accounting of actinides. Development of optimal process monitoring technology and safeguards approaches are, thus, a high priority for completing the development of the overall process technology. Real time monitoring of actinide concentrations in the molten salt electro-refiners is particularly important to achieve because of the potential for multiple significant quantities of plutonium in the molten salt. To address this need, development of a semiconductor detector for real-time monitoring of actinides in molten chloride salt (LiCl-KCl) is being conducted. Towards this end, a novel alpha spectroscopy detection system is being developed using wide band-gap semiconductor material, such as GaN or SiC, with a Boron-Doped-Diamond (BDD) contact. The objective is to deposit a thin layer of actinides onto the BDD layer, which will be used to measure the alpha particle spectrum that will distinguish individual actinides and their concentration. In the first stage, the actinide electrodeposition onto a surrogate material for the detector has been undertaken. While this project is in its very early states, a couple of observations have been made. One is that the electrodeposition is a slow, diffusion-controlled process. Therefore the deposit should be as thin (≅1 m) as the alpha spectroscopy permits (the resolution gets inferior with lesser number of atoms). Also, applied potentials should not be too negative; otherwise the deposit does not adhere well to the surface. The application of this method to measuring actinides in molten salt appears to be feasible, and a considerable amount of further investigation is planned. (authors)

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Performance Comparison of Object Detection Networks for Shrapnel Identification in Ultrasound Images

Ultrasound imaging is a critical tool for triaging and diagnosing subjects but only if images can be properly interpreted. Unfortunately, in remote or military medicine situations, the expertise to interpret images can be lacking. Machine-learning image interpretation models that are explainable to the end user and deployable in real time with ultrasound equipment have the potential to solve this problem. We have previously shown how a YOLOv3 (You Only Look Once) object detection algorithm can be used for tracking shrapnel, artery, vein, and nerve fiber bundle features in a tissue phantom. However, real-time implementation of an object detection model requires optimizing model inference time. Here, we compare the performance of five different object detection deep-learning models with varying architectures and trainable parameters to determine which model is most suitable for this shrapnel-tracking ultrasound image application. We used a dataset of more than 16,000 ultrasound images from gelatin tissue phantoms containing artery, vein, nerve fiber, and shrapnel features for training and evaluating each model. Every object detection model surpassed 0.85 mean average precision except for the detection transformer model. Overall, the YOLOv7tiny model had the higher mean average precision and quickest inference time, making it the obvious model choice for this ultrasound imaging application. Other object detection models were overfitting the data as was determined by lower testing performance compared with higher training performance. In summary, the YOLOv7tiny object detection model had the best mean average precision and inference time and was selected as optimal for this application. Next steps will implement this object detection algorithm for real-time applications, an important next step in translating AI models for emergency and military medicine.

60 APPLIED LIFE SCIENCES↗

Enhanced accuracy through ensembling of randomly initialized auto-regressive models for dynamical systems

Computational mechanics simulations using traditional finite element methods (FEM) require prohibitively expensive computational resources for real-time engineering applications, design optimization, and digital twin implementations. While machine learning (ML) surrogate models offer significant computational speedups, autoregressive ML models for time-dependent mechanical systems suffer from error accumulation that compromises long-term prediction reliability - a critical concern for engineering applications where accuracy over extended time horizons is essential for safety and performance assessments. Here, we propose a deep ensemble framework specifically designed to address this challenge in computational mechanics applications, where multiple ML surrogate models with random weight initializations are trained in parallel and their predictions aggregated during inference. This approach leverages statistical diversity to maximize information gain from a fixed set of training data and to mitigate error propagation, while maintaining the computational efficiency that makes ML surrogates attractive for engineering practice. We validate the framework on three representative problems spanning critical areas of computational mechanics: stress field evolution in heterogeneous microstructures under complex loading (relevant to advanced materials design and composite analysis), planetary-scale shallow water dynamics (applicable to environmental and geotechnical engineering), and Gray-Scott reaction-diffusion systems (relevant to mass transport and chemical process engineering). Across all test cases, the ensemble approach demonstrates consistent error reduction of 15-33% compared to individual models. The codes for this work are available on GitHub (https://github.com/Graham-Brady-Research-Group/AutoregressiveEnsemble_SpatioTemporal_Evolution).

autoregressive prediction↗

L0 Trigger for the EMCal Detector of the ALICE Experiment

The ALICE experiment at the CERN Large Hadron Collider (LHC) accelerator was designed to study ultra-relativistic heavy-ion collisions. The ALICE Electromagnetic Calorimeter (EMCal) was built to provide measurement of photons, electrons, and jets, and trigger selection of hard-QCD events containing them. The EMCal single-shower L0 trigger, which triggers on large energy deposit within a 4 x 4 tower sliding window, became operational in 2010. The implementation of the real-time FPGA based algorithm optimized to provide a fast L0 decision is presented.

43 PARTICLE ACCELERATORS↗

Comparing Sensor Fusion and Multimodal Chemometric Models for Monitoring U(VI) in Complex Environments Representative of Irradiated Nuclear Fuel

Optical sensors and chemometric models were leveraged for the quantification of uranium(VI) (0–100 μg mL –1 ), europium (0–150 μg mL –1 ), samarium (0–250 μg mL –1 ), praseodymium (0–350 μg mL –1 ), neodymium (0–1000 μg mL –1 ), and HNO 3 (2–4 M) with varying corrosion product (iron, nickel, and chromium) levels using laser fluorescence, Raman scattering, and ultraviolet–visible–near-infrared absorption spectra. In this paper, an efficient approach to developing and evaluating tens of thousands of partial least-squares regression (PLSR) models, built from fused optical spectra or multimodal acquisitions, is discussed. Each PLSR model was optimized with unique preprocessing combinations, and features were selected using genetic algorithm filters. The 7-factor D-optimal design training set contained just 55 samples to minimize the number of samples. The performance of PLSR models was evaluated by using an automated latent variable selection script. PLS1 regression models tailored to each species outperformed a global PLS2 model. PLS1 models built using fused spectra data and a multimodal (i.e., analyzed separately) approach yielded similar information, resulting in percent root-mean-square error of prediction values of 0.9–5.7% for the seven factors. Further, the optical techniques and data processing strategies established in this study allow for the direct analysis of numerous species without measuring luminescence lifetimes or relying on a standard addition approach, making it optimal for near-real-time, in situ measurements. Nuclear reactor modeling helped bound training set conditions and identified elemental ratios of lanthanide fission products to characterize the burnup of irradiated nuclear fuel. Leveraging fluorescence, spectrophotometry, experimental design, and chemometrics can enable the remote quantification and characterization of complex systems with numerous species, monitor system performance, help identify the source of materials, and enable rapid high-throughput experiments in a variety of industrial processes and fundamental studies.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Predictive modeling of NSTX discharges with the updated multi-mode anomalous transport module

Abstract The objective of this study is twofold: firstly, to demonstrate the consistency between the anomalous transport results produced by updated Multi-Mode Model (MMM) version 9.0.4 and those obtained through gyrokinetic simulations; and secondly, to showcase MMM’s ability to predict electron and ion temperature profiles in low aspect ratio, high beta NSTX discharges. MMM encompasses a range of transport mechanisms driven by electron and ion temperature gradients, trapped electrons, kinetic ballooning, peeling, microtearing, and drift resistive inertial ballooning modes. These modes within MMM are being verified through corresponding gyrokinetic results. The modes that potentially contribute to ion thermal transport are stable in MMM, aligning with both experimental data and findings from linear CGYRO simulations. The isotope effects on these modes are also studied and higher mass is found to be stabilizing, consistent with the experimental trend. The electron thermal power across the flux surface is computed within MMM and compared to experimental measurements and nonlinear CGYRO simulation results. Specifically, the electron temperature gradient modes (ETGM) within MMM account for 2.0 MW of thermal power, consistent with experimental findings. It is noteworthy that the ETGM model requires approximately 5.0 ms of computation time on a standard desktop, while nonlinear CGYRO simulations necessitate 8.0 h on 8 K cores. MMM proves to be highly computationally efficient, a crucial attribute for various applications, including real-time control, tokamak scenario optimization, and uncertainty quantification of experimental data.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Developing the hybrid scenario in DIII-D and KSTAR for W-compatible long-pulse operation

Long-pulse, high-performance operation with tungsten plasma-facing components is a crucial challenge for achieving steady-state (SS) goals in ITER and future fusion devices. This paper presents a joint international effort between the DIII-D and KSTAR tokamaks aimed at demonstrating the compatibility of the high q 95 ~ 6 hybrid scenario with a tungsten divertor and superconducting coils. A DIII-D hybrid scenario recipe is successfully tailored to KSTAR’s operational constraints, which facilitates cross-device studies. The results show successful 30 s long-pulse operation in KSTAR at a normalized beta (β N = 2.4) with a benign m/n = 4/3 mode and effective mitigation of tungsten accumulation. Despite this success, the KSTAR scenario exhibits a greater than 25% reduction in thermal confinement compared to the DIII-D reference. To investigate this discrepancy, we performed a predictive TGYRO validation study, which suggests that the difference is likely caused by a weaker temperature pedestal in the KSTAR discharge due to a higher and narrower density pedestal. In addition, such lower thermal confinement leads to higher impurity accumulation, possibly due to a change in neoclassical transport with increased inward pinch, which further degrades global confinement. We also report that stationary, long-pulse operation with sustained thermal confinement requires consistent fueling and wall conditioning. These findings highlight the importance of optimized fueling and real-time wall conditioning in achieving a high temperature pedestal and strong global confinement during stable long-pulse operation. We propose potential approaches toward fully SS, long-pulse operation, including real-time wall conditioning using an impurity dropper in KSTAR and future tokamaks.

DIII-D↗

Bayesian optimization algorithms for accelerator physics

Accelerator physics relies on numerical algorithms to solve optimization problems in online accelerator control and tasks such as experimental design and model calibration in simulations. The effectiveness of optimization algorithms in discovering ideal solutions for complex challenges with limited resources often determines the problem complexity these methods can address. The accelerator physics community has recognized the advantages of Bayesian optimization algorithms, which leverage statistical surrogate models of objective functions to effectively address complex optimization challenges, especially in the presence of noise during accelerator operation and in resource-intensive physics simulations. In this review article, we offer a conceptual overview of applying Bayesian optimization techniques toward solving optimization problems in accelerator physics. We begin by providing a straightforward explanation of the essential components that make up Bayesian optimization techniques. We then give an overview of current and previous work applying and modifying these techniques to solve accelerator physics challenges. Finally, we explore practical implementation strategies for Bayesian optimization algorithms to maximize their performance, enabling users to effectively address complex optimization challenges in real-time beam control and accelerator design. Published by the American Physical Society 2024

43 PARTICLE ACCELERATORS↗

Metric Learning to Accelerate Convergence of Operator Splitting Methods

Recent developments in machine learning have led to promising advances in accelerating the solution of constrained optimization problems. Increasing demand for real-time decision-making capabilities in applications such as artificial intelligence and optimal control has led to a variety of proposed strategies for learning to produce fast solutions to optimization problems. For example, recent works have shown that it is possible to accelerate the convergence of optimization algorithms by learning to select their parameters, such as gradient descent stepsizes. This work proposes a new approach, in which the underlying metric spaces of proximal operator splitting algorithms are learned to maximize convergence rate. While prior works in optimization theory have derived optimal metrics in simple cases, no such result exists for many practical problem forms including general Quadratic Programming (QP). This paper shows how differentiable optimization can enable the end-to-end learning of proximal metrics, enhancing the convergence of proximal algorithms for QP problems beyond what is possible based on known theory. Additionally, the results illustrate a strong connection between the learned proximal metrics and active constraints at the optima, leading to an interpretation in which the predicted proximal metrics can be viewed as a form of active set prediction.

King, Ethan [BATTELLE (PACIFIC NW LAB)]↗

Digital Twins for Data Centers

Fueled by an unprecedented adoption of AI (Artificial Intelligence), data centers are becoming the largest growing consumers of energy. Digital Twins provide living digital models of physical systems that enable data-driven analysis and application of AI to better manage selective aspects of the data center and drive efficiency for sustainability. Digital twins have emerged as a way to create virtual prototypes of physical artifacts, which may be used in a variety of contexts. Physical artifacts include airplanes, factories, or even static objects, such as bridges or dams. Digital twin helps monitor changes and assist in predicting planned or unplanned behaviors of physical objects. In this paper, we discuss digital twins for data centers.

97 MATHEMATICS AND COMPUTING↗

An Efficient Electrochemical Tanks-in-Series Model for Lithium Sulfur Batteries

This article applies and efficiently implements the Tanks-in-Series methodology ( J. Electrochem. Soc. , 167 , 013534 (2020)) to generate a computationally efficient electrochemical model for Lithium-Sulfur batteries. The original Tank model approach for Lithium-ion batteries is modified to account for porosity changes with time. In addition, an exponential scaling method is introduced that enables efficient simulation of the model equations to address the wide range of time constants present for different reactions in the Lithium-Sulfur system. The Tank Model achieves acceptable voltage error even for transport-limited discharged conditions. Predictions of internal electrochemical variables are examined, and electrochemical implications of the approximations discussed. This suggests significant potential for real-time applications such as optimal charging, cell-balancing, and estimation, and represents a step forward in efforts to incorporate detailed electrochemical models in advanced Battery Management Systems for Lithium-Sulfur batteries.

Parke, Caitlin D. (ORCID:0000000177324044)↗