SEARCH · Engineering Papers
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.
Pancreas stereotactic body radiation therapy (SBRT) treatment planning requires planners to make sequential, time-consuming interactions with the treatment planning system to reach the optimal dose distribution. We sought to develop a reinforcement learning (RL)–based planning bot to systematically address complex tradeoffs and achieve high plan quality consistently and efficiently.
Buildings account for 35% of the global final energy demand. Efficiency improvements and advanced control strategies have a significant impact in the reduction of energy costs and CO2 emissions. Building energy simulation is widely used to help planners, contractors, and building owners analyse diverse options regarding the planning and management of energy consumption in buildings. Furthermore, recent advances in data processing and computing have led to the development of sophisticated machine learning algorithms that can learn from large datasets, e.g., sensor data from buildings, and use them to develop building-specific adaptive and automatic energy controllers. Control algorithms, such as deep reinforcement learning can tune themselves, are model-free, and economical to implement. In this paper, we introduce an integrated simulation environment that combines CitySim, a fast building energy simulator, and TensorFlow, a platform for efficient implementation of advanced machine learning algorithms. The integration is achieved via Keras - an API for TensorFlow - and a set of text and csv files for data transfer between the applications. This new environment will allow researchers to investigate novel learning control algorithms, and demonstrate their robustness and potential for diverse applications in the built environment. We present two case studies for energy savings and demand response, respectively.
In 2017, Westinghouse performed a full system chemical decontamination of Units 1 and 2 of the Bohunice V1 Nuclear Power Plant. Bohunice V1 is a VVER 440 design with six loops. The application was performed two loops at a time, with the reactor core bypassed. The Pressurizer and Pressurizer Relief Tank (Bubble Tank) were decontaminated as well, in a separate application. This was a recovery decontamination that had been attempted by another company prior to Westinghouse. Several modifications to equipment and design of process were required to make the decontamination successful. The project was initiated to perform pre-dismantling chemical decontamination of primary circuits of both V1 units using Decontamination for Decommissioning (DfD) to meet the following objectives (see References 1 and 2 for process descriptions of DfD): - Remove activity from the primary circuit components; - Reduce dose rates around piping and components; - Simplify access to the installations, making it possible to use hands-on techniques for dismantling, rather than the more expensive use of robotics or manipulators; - Minimize the potential for spreading contamination during decommissioning activities - Reduce the activity of components that they may be disposed of at a lower classification, therefore more economical, waste disposal category. Several techniques were employed that were unique, or first of a kind, to aid in the decontamination, including resin regeneration and bypass of the reactor core, while still providing enough flow to to allow sufficient dispersal through all the steam generator tubes. Equipment that was already available on site was evaluated and, when possible, combined with Westinghouse equipment to minimize application cost and schedule. The lessons learned from the first unit (Unit 2) were applied to the second unit (Unit 1) application. The partial relocation of equipment allowed for the most efficient schedule to complete the second decontamination. The paper will describe in more detail the results and lessons learned from the full system chemical decontaminations performed, as well as updates on another recent decontamination project. (authors)
The objective of the SMART Initiative, i.e., Science-informed Machine Learning (ML) for Accelerating Real-Time Decisions in Subsurface Applications, is to show how the utilization of ML can significantly improve efficiency and effectiveness of field-scale commercial carbon storage operations in three main areas: real-time visualization, virtual learning, and real-time forecasting. This presentation reports the status of SMART initiative for demonstrating: (a) virtual learning during the pre-injection permitting phase, and (b) ML-assisted operational decision making and visualization.
Bayesian inference with a simple Gaussian error model is used to efficiently compute prediction variances for energies, forces, and stresses in the linear SNAP interatomic potential. Here, the prediction variance is shown to have a strong correlation with the absolute error over approximately 24 orders of magnitude. Using this prediction variance, an active learning algorithm is constructed to iteratively train a potential by selecting the structures with the most uncertain properties from a pool of candidate structures. The relative importance of the energy, force, and stress errors in the objective function is shown to have a strong impact upon the trajectory of their respective net error metrics when running the active learning algorithm. Batched training of different batch sizes is also tested against singular structure updates, and it is found that batches can be used to significantly reduce the number of retraining steps required with only minor impact on the active learning trajectory.
For the past year, the HEP.TrkX project has been investigating machine learning solutions to LHC particle track reconstruction problems. A variety of models were studied that drew inspiration from computer vision applications and operated on an image-like representation of tracking detector data. While these approaches have shown some promise, image-based methods face challenges in scaling up to realistic HL-LHC data due to high dimensionality and sparsity. In contrast, models that can operate on the spacepoint representation of track measurements ("hits") can exploit the structure of the data to solve tasks efficiently. In this paper we will show two sets of new deep learning models for reconstructing tracks using space-point data arranged as sequences or connected graphs. In the first set of models, Recurrent Neural Networks (RNNs) are used to extrapolate, build, and evaluate track candidates akin to Kalman Filter algorithms. Such models can express their own uncertainty when trained with an appropriate likelihood loss function. The second set of models use Graph Neural Networks (GNNs) for the tasks of hit classification and segment classification. These models read a graph of connected hits and compute features on the nodes and edges. They adaptively learn which hit connections are important and which are spurious. The models are scaleable with simple architecture and relatively few parameters. Results for all models will be presented on ACTS generic detector simulated data.
Abstract Silica is an abundant and technologically attractive material. Due to the structural complexities of silica polymorphs coupled with subtle differences in Si–O bonding characteristics, the development of accurate models to predict the structure, energetics and properties of silica polymorphs remain challenging. Current models for silica range from computationally efficient Buckingham formalisms (BKS, CHIK, Soules) to reactive (ReaxFF) and more recent machine-learned potentials that are flexible but computationally costly. Here, we introduce an improved formalism and parameterization of BKS model via a multireward reinforcement learning (RL) using an experimental training dataset. Our model concurrently captures the structure, energetics, density, equation of state, and elastic constants of quartz (equilibrium) as well as 20 other metastable silica polymorphs. We also assess its ability in capturing amorphous properties and highlight the limitations of the BKS-type functional forms in simultaneously capturing crystal and amorphous properties. We demonstrate ways to improve model flexibility and introduce a flexible formalism, machine-learned ML-BKS, that outperforms existing empirical models and is on-par with the recently developed 50 to 100 times more expensive Gaussian approximation potential (GAP) in capturing the experimental structure and properties of silica polymorphs and amorphous silica.
Enzymes are proteins that catalyze specific biochemical reactions in cells. Enzyme Commission (EC) numbers are used to annotate enzymes in a four-level hierarchy that classifies enzymes based on the specific chemical reactions they catalyze. Accurate EC number prediction is essential for understanding enzyme functions. Despite the availability of numerous methods for predicting EC numbers from protein sequences, there is no unified framework for evaluating and studying such methods systematically. This gap limits the ability of the community to identify the most effective approaches for enzyme annotation. We introduce EC-Bench, a benchmark for EC number prediction, consisting of 1) an initial representative set of existing methods (including homology-based, deep learning, contrastive learning, and language model methods), 2) existing and novel accuracy and efficiency performance metrics, and 3) selected datasets to allow for comprehensive comparative study. EC-Bench is open-source and provides a framework for researchers to not only compare among existing methods objectively under uniform conditions, but also to introduce and effectively evaluate performance of new methods in a comparative framework. To demonstrate the utility of EC-Bench, we perform extensive experimentation to compare the existing EC number prediction methods and establish their advantages and disadvantages in a variety of prediction tasks, namely “exact EC number prediction”, “EC number completion” and (partial or additional) “EC number recommendation”. We find wide variation in the performance of different methods, but also subtle but potentially useful differences in the performance of different methods across tasks and for different parts of the EC hierarchy.
Particle physics theories, such as those which explain neutrino flavor mixing, arise from a vast landscape of model-building possibilities. A model’s construction typically relies on the intuition of theorists. It also requires considerable effort to identify appropriate symmetry groups, assign field representations, and extract predictions for comparison with experimental data. We develop Autonomous Model Builder (AMBer), a framework in which a reinforcement learning agent interacts with a streamlined physics software pipeline to search these spaces efficiently. AMBer selects symmetry groups, particle content, and group representation assignments to construct models while minimizing the number of free parameters introduced. We validate our approach in well-studied regions of theory space and extend the exploration to a previously unexamined symmetry group. While demonstrated in the context of neutrino flavor theories, this approach of reinforcement learning with physics software feedback may be extended to other theoretical model-building problems in the future.
A third-integer resonant slow extraction system is being developed for Fermilab's Delivery Ring to deliver protons to the upcoming Mu2e experiment. The timescale of the extraction (or spill) duration is 43 milliseconds, which is extremely short and unprecedented. Additionally, the experiment's strict and challenging requirements on the quality of the spill at this time scale has led to the development of a new Spill Regulation System (SRS) design. The SRS primarily consists of three components - slow regulation, fast regulation, and harmonic content suppressor. Contributions to the first two components of the SRS, i.e., Slow Regulation and Fast Regulation subsystems, will be presented in which new adaptive learning algorithm schemes for the slow regulation of the spill -- validated using particle tracking simulations -- shall be described. In addition to these novel methods for the enhancement of the spill regulation system, results of employing Machine Learning in enhancing the performance of the resonant extraction are also presented. At the forefront of applying ML techniques to solve non-linear accelerator control problems, this work includes optimizing the PID gains as well as the replacement of the traditional PID controller using Recurrent Neural Networks and Gated Recurrent Unit (GRU) ML models to achieve efficiencies greater than a PID controller. Cutting-edge on-going Reinforcement Learning efforts, including an actor-critic family of learning algorithms, to regulate the spill rate will be reviewed, as well as present analytical calculations pertaining the transit time of particles in a third-integer resonant extraction. Detailed numerical investigations and validations of such calculations, the model of which could be exported and reliably used in future analytical modeling of any resonant extraction, are discussed.
Recent years have seen a lot of progress in algorithms for learning parameters of spreading dynamics from both full and partial data. Some of the remaining challenges include model selection under the scenarios of unknown network structure, noisy data, missing observations in time, as well as an efficient incorporation of prior information to minimize the number of samples required for an accurate learning. Here, in this work, we introduce a universal learning method based on a scalable dynamic message-passing technique that addresses these challenges often encountered in real data. The algorithm leverages available prior knowledge on the model and on the data, and reconstructs both network structure and parameters of a spreading model. We show that a linear computational complexity of the method with the key model parameters makes the algorithm scalable to large network instances.
Abstract Through advanced mechanistic modeling and the generation of large high-quality datasets, machine learning is becoming an integral part of understanding and engineering living systems. Here we show that mechanistic and machine learning models can be combined to enable accurate genotype-to-phenotype predictions. We use a genome-scale model to pinpoint engineering targets, efficient library construction of metabolic pathway designs, and high-throughput biosensor-enabled screening for training diverse machine learning algorithms. From a single data-generation cycle, this enables successful forward engineering of complex aromatic amino acid metabolism in yeast, with the best machine learning-guided design recommendations improving tryptophan titer and productivity by up to 74 and 43%, respectively, compared to the best designs used for algorithm training. Thus, this study highlights the power of combining mechanistic and machine learning models to effectively direct metabolic engineering efforts.
With the increased prevalence of neural operators being used to provide rapid solutions to partial differential equations (PDEs), understanding the accuracy of model predictions and the associated error levels is necessary for deploying reliable surrogate models in scientific applications. Existing uncertainty quantification (UQ) frameworks employ ensembles or Bayesian methods, which can incur substantial computational costs during both training and inference. Here, we propose a lightweight predictive UQ method tailored for Deep operator networks (DeepONets) that also generalizes to other operator networks. Numerical experiments on linear and nonlinear PDEs demonstrate that the framework’s uncertainty estimates are unbiased and provide accurate out-of-distribution uncertainty predictions with a sufficiently large training dataset. Our framework provides fast inference and uncertainty estimates that can efficiently drive outer-loop analyses that would be prohibitively expensive with conventional solvers. We demonstrate how predictive uncertainties can be used in the context of Bayesian optimization and active learning problems to yield improvements in accuracy and data-efficiency for outer-loop optimization procedures. In the active learning setup, we extend the framework to Fourier Neural Operators (FNO) and describe a generalized method for other operator networks. To enable real-time deployment, we introduce an inference strategy based on precomputed trunk outputs and a sparse placement matrix, reducing evaluation time by more than a factor of five. Our method provides a practical route to uncertainty-aware operator learning in time-sensitive settings.
Ultrafast electron diffraction using MeV energy beams(MeV-UED) has enabled unprecedented scientific opportunities in the study of ultrafast structural dynamics in a variety of gas, liquid and solid state systems. Broad scientific applications usually pose different requirements for electron probe properties. Due to the complex, nonlinear and correlated nature of accelerator systems, electron beam property optimization is a time-taking process and often relies on extensive hand-tuning by experienced human operators. Algorithm based efficient online tuning strategies are highly desired. Here, we demonstrate multi-objective Bayesian active learning for speeding up online beam tuning at the SLAC MeV-UED facility. The multi-objective Bayesian optimization algorithm was used for efficiently searching the parameter space and mapping out the Pareto Fronts which give the trade-offs between key beam properties. Such scheme enables an unprecedented overview of the global behavior of the experimental system and takes a significantly smaller number of measurements compared with traditional methods such as a grid scan. This methodology can be applied in other experimental scenarios that require simultaneously optimizing multiple objectives by explorations in high dimensional, nonlinear and correlated systems.
The sharp increase of pressure at the edge of a high confinement mode (H-mode) plasma, the pedestal, strongly impacts overall plasma performance. Predicting the pedestal is a necessity to control and optimize tokamak operations. An experimental data-driven machine learning (ML) approach is presented that predicts the pedestal heights and widths of electron density (ne) and electron temperature (Te) profiles as well as the separatrix ne from externally controllable parameters such as the plasma shape, heating method and power, and gas puff rate and integrated gas puff. The OMFIT framework was used with DIII-D data to efficiently, robustly, and automatically build a database of pedestal parameters to train machine learning models. Database creation was enabled by the search engine tool for DIII-D data, TokSearch, which parallelizes data fetching, enabling fast searches through basic signals of thousands of DIII-D shots and selection of relevant time intervals. Principal Component Analysis (PCA) separated the database into three clusters that represent classes of plasma shapes that are regularly used in DIII-D. The most important parameters for setting the pedestal structure were plasma current (Ip), toroidal magnetic field (Bφ), neutral beam heating power (PNBI) and shaping quantities. The Deep Jointly Informed Neural Networks (DJINN) algorithm was applied to identify suitable neural network (NN) architectures that appropriately capture the features of the pedestal database. Separate NNs were implemented for each pedestal parameter, and ensembling methods were used to improve the prediction accuracy and allowed estimation of the prediction uncertainty. The pedestal predictions of the test dataset lie within the measurement uncertainties of the pedestal parameters. The NN outperformed simple Linear Regression (LR) analysis, indicating non-linear dependencies in the pedestal structure. The presented achievements illustrate a promising path for future research, using feature extraction to infer experimental trends and thereby improve pedestal models as well as deploying NN for a fast pedestal prediction in DIII-D scenario development.
Advanced high-temperature fluid reactors, such as sodium-cooled fast reactors (SFRs) and molten salt–cooled reactors (MSCRs), require coolant purification systems to prevent fluid contamination and local freezing that can lead to plugging. Liquid sodium purification can be achieved with a cold trap, where the sodium temperature is reduced to a near-freezing point to precipitate out impurities. Automation of monitoring of the cold trap performance with machine learning algorithms can aid in early detection of incipient anomalies. An efficient approach to loss-of-coolant–type anomaly detection in a cold trap monitored with more than two dozen thermal-hydraulic sensors consists of a long short-term memory (LSTM) autoencoder. This work develops the uncertainty quantification of the LSTM autoencoder performance for cold trap anomaly detection using the Monte Carlo (MC) dropout method. The MC dropout methodology creates a distribution of sister distributions that all slightly differ from each other because of random neurons being turned off for testing. The variances of the sister network distributions are used to make an uncertainty interval. Our analysis shows that the uncertainty in the autoencoder performance is largest near the peak of the anomaly signal. Using the MC dropout method, we investigate the uncertainty in the anomaly detection with missing sensor inputs. This capability allows the reactor operator to evaluate resilience of the anomaly detection system and to make informed decisions about continuity of operation in the event of sensor failure.
CUORE Upgrade with Particle IDentification (CUPID) is a foreseen ton-scale array of Li 2 MoO 4 (LMO) cryogenic calorimeters with double readout of heat and light signals. Its scientific goal is to fully explore the inverted hierarchy of neutrino masses in the search for neutrinoless double beta decay of 100 Mo. Pile-up of standard double beta decay of the candidate isotope is a relevant background. We generate pile-up heat events via injection of Joule heater pulses with a programmable waveform generator in a small array of LMO crystals operated underground in the Laboratori Nazionali del Gran Sasso, Italy. This allows to label pile-up pulses and control both time difference and underlying amplitudes of individual heat pulses in the data. We present the performance of supervised learning classifiers on data and the attained pile-up rejection efficiency.
Medical physics is learned in a combination of activities including classroom sessions, individual study, small-group collaborative problem solving, and direct experience in the laboratory or clinical environment. Each type of learning activity is characterized by its effectiveness in producing the desired knowledge for the learner and the cost in terms of resources and human effort required providing it. While learning and teaching is a human activity, modern technology provides a variety of tools that can be used to enhance human performance. The class or conference room is the common setting for educational sessions in both academic institutions and continuing education conferences and programs such as those sponsored by the AAPM. A major value of a class/conference room program is efficiency by bringing a group of learners together to share in a common learning experience under the guidance of one or more experienced learning facilitators (lecturers or presenters). A major challenge is that the class/conference room is separated from the real world of medical physics. The design of an educational activity needs to take into consideration the desired outcomes with respect to what the learners should be able to do. The distinction is that of being able to apply the knowledge to perform specific physics functions rather than just knowing and being able to recall facts, and perhaps do well on written examinations. These are different types of knowledge structures within the human brain and distinctly different learning activities to develop each. Much of medical physics education, especially at the post-graduate and continuing education level, is for the purpose of enhancing the ability of physicists and other related professionals to perform applied procedures and tasks and requires specific types of knowledge.In this session we will analyze various learning activity models, the values and limitations of each, and how they can be used in medical physics education. An example we will use is optimizing CT image quality and dose which is an important topic for medical physicists, radiologists and residents, along with technologists. The knowledge structure for this is best developed by a combination of learning activities including class/conference discussions, individual study and review, and direct observation and interaction in the clinical setting under the direction of a knowledgeable leader.The function of the human brain will be considered with respect to learning experiences that contribute to effective medical physics knowledge structures. The characteristics of various types of educational activities will be compared with respect to their effectiveness for producing desired outcomes along with their limitations. Emphasis will be given to the design of highly-effective classroom/conference presentations, and activities will be demonstrated with an emphasis on using technology to enhance human performance of both learners and the learning facilitators. Learning Objectives: Develop and provide highly effective medical physics educational sessions. Use technology to enhance human performance in the educational process. Identify and analyze various models of educational activities Select and use educational activities that contribute value to the medical physics profession.