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.
Fatigue damage estimation using neural networks is described in the paper. Attention is focused on the method of data generation for both the training and test data used by radial basis function (RBF), backpropagation, and CID4 algorithms used in this study. The performance results of the three neural algorithms are analyzed in terms of their strengths and weaknesses in training.
Metal-organic frameworks (MOFs) exhibit great promise for CO 2 capture. However, finding the best performing materials poses computational and experimental grand challenges in view of the vast chemical space of potential building blocks. Here, we introduce GHP-MOFassemble, a generative artificial intelligence (AI), high performance framework for the rational and accelerated design of MOFs with high CO 2 adsorption capacity and synthesizable linkers. GHP-MOFassemble generates novel linkers, assembled with one of three pre-selected metal nodes (Cu paddlewheel, Zn paddlewheel, Zn tetramer) into MOFs in a primitive cubic topology. GHP-MOFassemble screens and validates AI-generated MOFs for uniqueness, synthesizability, structural validity, uses molecular dynamics simulations to study their stability and chemical consistency, and crystal graph neural networks and Grand Canonical Monte Carlo simulations to quantify their CO 2 adsorption capacities. We present the top six AI-generated MOFs with CO 2 capacities greater than 2m mol g -1 , i.e., higher than 96.9% of structures in the hypothetical MOF dataset.
Abstract Progress towards the energy breakthroughs needed to combat climate change can be significantly accelerated through the efficient simulation of atomistic systems. However, simulation techniques based on first principles, such as density functional theory (DFT), are limited in their practical use due to their high computational expense. Machine learning approaches have the potential to approximate DFT in a computationally efficient manner, which could dramatically increase the impact of computational simulations on real-world problems. However, they are limited by their accuracy and the cost of generating labeled data. Here, we present an online active learning framework for accelerating the simulation of atomic systems efficiently and accurately by incorporating prior physical information learned by large-scale pre-trained graph neural network models from the Open Catalyst Project. Accelerating these simulations enables useful data to be generated more cheaply, allowing better models to be trained and more atomistic systems to be screened. We also present a method of comparing local optimization techniques on the basis of both their speed and accuracy. Experiments on 30 benchmark adsorbate-catalyst systems show that our method of transfer learning to incorporate prior information from pre-trained models accelerates simulations by reducing the number of DFT calculations by 91%, while meeting an accuracy threshold of 0.02 eV 93% of the time. Finally, we demonstrate a technique for leveraging the interactive functionality built in to Vienna ab initio Simulation Package (VASP) to efficiently compute single point calculations within our online active learning framework without the significant startup costs. This allows VASP to work in tandem with our framework while requiring 75% fewer self-consistent cycles than conventional single point calculations. The online active learning implementation, and examples using the VASP interactive code, are available in the open source FINETUNA package on Github.
The NASA Engine Performance Program (NEPP) can configure and analyze almost any type of gas turbine engine that can be generated through the interconnection of a set of standard physical components. In addition, the code can optimize engine performance by changing adjustable variables under a set of constraints. However, for engine cycle problems at certain operating points, the NEPP code can encounter difficulties: nonconvergence in the currently implemented Powell's optimization algorithm and deficiencies in the Newton-Raphson solver during engine balancing. A project was undertaken to correct these deficiencies. Nonconvergence was avoided through a cascade optimization strategy, and deficiencies associated with engine balancing were eliminated through neural network and linear regression methods. An approximation-interspersed cascade strategy was used to optimize the engine's operation over its flight envelope. Replacement of Powell's algorithm by the cascade strategy improved the optimization segment of the NEPP code. The performance of the linear regression and neural network methods as alternative engine analyzers was found to be satisfactory. This report considers two examples-a supersonic mixed-flow turbofan engine and a subsonic waverotor-topped engine-to illustrate the results, and it discusses insights gained from the improved version of the NEPP code.
Abstract Equations of State model relations between thermodynamic variables and are ubiquitous in scientific modelling, appearing in modern day applications ranging from Astrophysics to Climate Science. The three desired properties of a general Equation of State model are adherence to the Laws of Thermodynamics, incorporation of phase transitions, and multiscale accuracy. Analytic models that adhere to all three are hard to develop and cumbersome to work with, often resulting in sacrificing one of these elements for the sake of efficiency. In this work, two deep-learning methods are proposed that provably satisfy the first and second conditions on a large-enough region of thermodynamic variable space. The first is based on learning the generating function (thermodynamic potential) while the second is based on structure-preserving, symplectic neural networks, respectively allowing modifications near or on phase transition regions. They can be used either “from scratch” to learn a full Equation of State, or in conjunction with a pre-existing consistent model, functioning as a modification that better adheres to experimental data. We formulate the theory and provide several computational examples to justify both approaches, highlighting their advantages and shortcomings.
Recent years have seen a phenomenal rise in the performance and applications of transformer neural networks. The family of transformer networks, including Bidirectional Encoder Representations from Transformer (BERT), Generative Pretrained Transformer (GPT) and Vision Transformer (ViT), have shown their effectiveness across Natural Language Processing (NLP) and Computer Vision (CV) domains. Transformer-based networks such as ChatGPT have impacted the lives of common men. However, the quest for high predictive performance has led to an exponential increase in transformers' memory and compute footprint. Researchers have proposed techniques to optimize transformer inference at all levels of abstraction. Further, this paper presents a comprehensive survey of techniques for optimizing the inference phase of transformer networks. We survey techniques such as knowledge distillation, pruning, quantization, neural architecture search and lightweight network design at the algorithmic level. We further review hardware-level optimization techniques and the design of novel hardware accelerators for transformers. We summarize the quantitative results on the number of parameters/FLOPs and the accuracy of several models/techniques to showcase the tradeoff exercised by them. We also outline future directions in this rapidly evolving field of research. We believe that this survey will educate both novice and seasoned researchers and also spark a plethora of research efforts in this field.
We have proposed previously a computational neural-network model by which the complex patterns of retinal image motion generated during locomotion (optic flow) can be processed by specialized detectors acting as templates for specific instances of self-motion. The detectors in this template model respond to global optic flow by sampling image motion over a large portion of the visual field through networks of local motion sensors with properties similar to those of neurons found in the middle temporal (MT) area of primate extrastriate visual cortex. These detectors, arranged within cortical-like maps, were designed to extract self-translation (heading) and self-rotation, as well as the scene layout (relative distances) ahead of a moving observer. We then postulated that heading from optic flow is directly encoded by individual neurons acting as heading detectors within the medial superior temporal (MST) area. Others have questioned whether individual MST neurons can perform this function because some of their receptive-field properties seem inconsistent with this role. To resolve this issue, we systematically compared MST responses with those of detectors from two different configurations of the model under matched stimulus conditions. We found that the characteristic physiological properties of MST neurons can be explained by the template model. We conclude that MST neurons are well suited to support self-motion estimation via a direct encoding of heading and that the template model provides an explicit set of testable hypotheses that can guide future exploration of MST and adjacent areas within the superior temporal sulcus.
The ground state electron density — obtainable using Kohn-Sham Density Functional Theory (KS-DFT) simulations — contains a wealth of material information, making its prediction via machine learning (ML) models attractive. However, the computational expense of KS-DFT scales cubically with system size which tends to stymie training data generation, making it difficult to develop quantifiably accurate ML models that are applicable across many scales and system configurations. Here, we address this fundamental challenge by employing transfer learning to leverage the multi-scale nature of the training data, while comprehensively sampling system configurations using thermalization. Our ML models are less reliant on heuristics, and being based on Bayesian neural networks, enable uncertainty quantification. We show that our models incur significantly lower data generation costs while allowing confident — and when verifiable, accurate — predictions for a wide variety of bulk systems well beyond training, including systems with defects, different alloy compositions, and at multi-million-atom scales. Moreover, such predictions can be carried out using only modest computational resources.
Abstract Implementations of neurons, delays, and synapse circuits are presented with simulations. These neural elements are used to create two small spiking neural networks, the Rate-Window and Order-Biased clusters, which are capable of detecting simple two-spike spatiotemporal patterns. A simple pattern detecting network (SPDN) is created by combining the Rate-Window and Order-Biased clusters, where clusters are small spiking neural networks, and its simple pattern detection ability is demonstrated in simulation. The SPDN is used to implement a complex pattern detecting network (CPDN) and its complex pattern detection ability is demonstrated in simulation. Methods for generating arbitrary spatiotemporal patterns are presented. The CPDN and spatiotemporal pattern generation methods are then used to implement a novel spatiotemporal computing paradigm based on detecting and responding to spatiotemporal symbols. A simulation of a spatiotemporal half adder is presented to demonstrate the computing paradigm.
HydroBoost is the most realistic revenue optimization tool for the hybridization of hydropower and battery energy storage systems to date. The innovative representation of how operators actually schedule hydropower in practice results in more realistic predictions of revenue and operations. Unlike other optimization tools, HydroBoost generates forecast energy prices with uncertainty to use in the optimization. This allows HydroBoost to give users a range of potential revenue with an upper bound using the perfect foresight pricing and a lower bound using a naive persistence forecast model. Additional forecast can be generated and used in the optimization, such as additive models, random forest, and neural networks to give further insight into potential revenue. HydroBoost has been designed to be applicable for both run-of-river and reservoir storage sites. The primary focus is on the day-ahead market and requires year-long data with an hour time-step. All time-series input and constraints are contained in an Excel worksheet for convince. The user will run the forecasting generation first with a Python script to give the optimization model the necessary requirements. Next the optimization is ran using Julia and results are generated and stored into a directory as csv files. HydroBoost includes an additional module to generate figures based on the results of the optimization simulation. The results help analyze the results and users to draw insights into how the hydro and battery systems are operated and the revenue each is producing. Additionally, the difference between the perfect foresight model and models that include forecast can easily be inspected.
The volatile organic compounds (VOCs) in a wide spectrum of categories were identified as biomarkers in aquatic environments, playing an important role in marine and freshwater ecology and global atmospheric chemistry. VOCs released from biofuel have also attracted increasing attention. Although the importance has been recognized, the portable detection and analysis methods of VOC in aquatic systems have not yet been well developed and understood. Here, in this work, we innovatively proposed an intelligent electrochemical sensing approach to classify and quantify VOCs in solution. Utilizing the cyclic voltammetry (CV) method with an ionic liquid (IL)-based electrolyte, we analyzed 50 μL samples of various VOC analytes, including acetic acid (AC), acetone, dimethylformamide (DMF), dimethyl sulfoxide (DMSO), ethanol, formaldehyde, formic acid, methanol, methyl formate (MF), toluene, and a formaldehyde-methanol mixture, along with deionized water (DI water). The generated voltammograms were subsequently analyzed using our uniquely designed and optimized 1-D convolutional neural network (1D-CNN). This deep-learning algorithm achieved a 99.09% accuracy in VOC classification validated through fivefold cross-validation and demonstrated an impressive 94.4% test accuracy for methanol detection within a 10 μL error range. For quantification, the system accurately categorized methanol volumes ranging from 0 to 50 μL in 10 μL increments, achieving a 98.18% accuracy. A notable linear correlation (R2 = 95.56%) was found between max current density at the oxidation peak and methanol volume, with the limit of detection (LOD) at 9.3 μL. Such a sensing method exhibits potential for portability, high accuracy, and generalization in the classification and quantification, ultimately reshaping the realm of VOC analysis in solution.
Widespread adoption of high-temperature electrochemical systems such as polymer electrolyte membrane fuel cells (HT-PEMFCs) requires models and computational tools for accurate optimization and guiding new materials for enhancing fuel cell performance and durability. Furthermore, while robust and better suited for extrapolation, knowledge-based modeling has limitations as it is time-consuming and requires information about the system that is not always available (e.g., material properties and interfacial behavior between different materials). Data-driven modeling, on the other hand, is easier to implement but often necessitates large datasets that could be difficult to obtain. In this contribution, knowledge-based modeling and data-driven modeling are combined by implementing a few-shot learning (FSL) approach. A knowledge-based model originally developed for a HT-PEMFCs was used to generate simulated data (887,735 points) and used to pretrain a neural network source model tuned via a genetic algorithm-based AutoML. Then, experimental datasets from HT-PEMFCs with different materials and operating conditions (~50 points each) were used to train six target models via FSL. Models for the unseen data reached high accuracies in all cases (rRMSE < 10%).
Active feedback control in magnetic confinement fusion devices is desirable to mitigate plasma instabilities and enable robust operation. Optical high-speed cameras provide a powerful, non-invasive diagnostic and can be suitable for these applications. Here, in this study, we process high-speed camera data, at rates exceeding 100 kfps, on in situ field-programmable gate array (FPGA) hardware to track magnetohydrodynamic (MHD) mode evolution and generate control signals in real time. Our system utilizes a convolutional neural network (CNN) model, which predicts the n = 1 MHD mode amplitude and phase using camera images with better accuracy than other tested non-deep-learning-based methods. By implementing this model directly within the standard FPGA readout hardware of the high-speed camera diagnostic, our mode tracking system achieves a total trigger-to-output latency of 17.6 μs and a throughput of up to 120 kfps. This study at the High Beta Tokamak-Extended Pulse (HBT-EP) experiment demonstrates an FPGA-based high-speed camera data acquisition and processing system, enabling application in real-time machine-learning-based tokamak diagnostic and control as well as potential applications in other scientific domains.
This study explores data-driven prognostics for nuclear power plant (NPP) condensers, focusing on tube fouling. We utilized the Asherah nuclear power plant simulator (ANS) to compare four methods: Random Forest (RF), Support Vector Regressor (SVR), Fully Connected Neural Network (FCNN), and Long Short-Term Memory Neural Network (LSTM). By simulating various fouling scenarios in the ANS, we generated data with different degradation rates under transient operations. The models were trained and tested on these data, with performance evaluated visually and numerically including uncertainty assessment. The LSTM model excelled, exhibiting minimal prediction noise and the most accurate remaining useful life estimates across all degradation levels. Its ability to capture long-term dependencies and produce cleaner outputs makes it a strong candidate, although accurate training data across the entire component lifespan are crucial. The RF model emerged as a robust alternative, providing reliable predictions with high confidence. The FCNN and SVR models, while less effective overall, showed potential under specific conditions. FCNN offers a less complex alternative to LSTM and might benefit from larger datasets. SVR excels in precision when the quality of the training data is high. Furthermore, this study highlights the operational benefits of advanced prognostics in the energy sector and emphasizes the need for further research in NPP condenser health management through real-life experiments.
Understanding charge motion in a background of interacting quantum spins is a fundamental problem in quantum many-body physics. The most extensively studied model for this problem is the so-called t-t'-t''-J model, where the determination of the parameter t' in the context of cuprate superconductors is challenging. Here we present a theoretical study of the spectral functions of a mobile hole in the t-t'-t''-J model using two machine-learning techniques: K-nearest neighbor regression (KNN) and a feed-forward neural network (FFNN). We employ the self-consistent Born approximation to generate a dataset of about 1.3 x 10 5 spectral functions. Here we show that, for the forward problem, both methods allow for the accurate and efficient prediction of spectral functions, allowing, e.g., rapid searches through parameter space. Furthermore, we find that for the inverse problem (inferring Hamiltonian parameters from spectra), the FFNN can, but the KNN cannot, accurately predict the model parameters using merely the density of states. Our results suggest that it may be possible to use deep-learning methods to predict materials parameters from experimentally measured spectral functions.
In this paper, we introduce MARS, a new scheduling system for HPC-cloud infrastructures based on a cost-aware, flexible reinforcement learning approach, which serves as an intermediate layer for next generation HPC-cloud resource manager. MARS ensembles the pre-trained models from heuristic workloads and decides on the most cost-effective strategy for optimization. A whole workflow application would be split into several optimizable dependent sub-tasks, then based on the pre- defined resource management plan, a reward will be generated after executing a scheduled task. Lastly, MARS updates the Deep Neural Network (DNN) model based on the reward. MARS is designed to optimize the existing models through reinforcement mechanisms. MARS adapts to the dynamics of workflow applications, selects the most cost-effective scheduling solution among pre-built scheduling strategies (backfilling, SJF, etc.) and self- learning deep neural network model at run-time. We evaluate MARS with different real-world workflow traces. MARS can achieve 5%-60% increased performance compare to state-of-the- art approaches.
This work describes the development of a transported Livengood–Wu (L–W) integral model for computational fluid dynamics (CFD) simulation to predict autoignition and engine knock tendency. The currently employed L–W integral model considers both single-stage and two-stage ignition processes, thus can be generally applied to different fuels such as paraffin, olefin, aromatics, and alcohol. The model implementation is first validated in simulations of homogeneous charge compression ignition (HCCI) combustion for three different fuels, showing good accuracy in prediction of autoignition timing for fuels with either single-stage or two-stage ignition characteristics. Then, the L–W integral model is coupled with G-equation model to indicate end-gas autoignition and knock tendency in CFD simulations of a direct-injection spark-ignition engine. This modeling approach is about 10 times more efficient than the ones that based on detailed chemistry calculation and pressure oscillation analysis. Two fuels with same Research Octane Number (RON) but different octane sensitivity are studied, namely, Co-Optima alkylate and Co-Optima E30. Feed-forward neural network model in conjunction with multivariable minimization technique is used to generate fuel surrogates with targets of matched RON, octane sensitivity, and ethanol content. The CFD model is validated against experimental data in terms of pressure traces and heat release rate for both fuels under a wide range of operating conditions. The knock tendency—indicated by the fuel energy contained in the autoignited region—of the two fuels at different load conditions correlates well with the experimental results and the fuel octane sensitivity, implying the current knock modeling approach can capture the octane sensitivity effect and can be applied to further investigation on composition of octane sensitivity.
This tool takes a genome assembly as input and predicts per-site nucleosome occupancy as output. Trained on physical maps of nucleosome binding preferences across the fungal kingdom, NuclPred can be applied broadly across fungi (and other eukaryotes). This breadth, combined with its accuracy, means it could have both basic and applied biological implications, for example in understanding eukaryotic gene regulation and genetic engineering. Almost universally across eukaryotes, nucleosomes - each wrapping ~150 base pairs of DNA - serve to package DNA inside the nucleus, with major consequences on DNA access, gene activity and DNA integration. NuclPred was generated using a supervised deep learning approach combining convolutional and recurrent neural networks to take DNA features (nucleotides, GC content and structural information) as input, then use that information to predict the physical attractiveness DNA sequences might have for forming nucleosomes. With this information at hand, researchers can design more efficient CRISPR constructs, explore the interplay between DNA signatures and other regulators impact nucleosome locations, predict expression patterns, etc. This tool will be published as part of a manuscript currently under revision at iScience (draft attached).