Search NASA⌕ Search

SEARCH · Search NASA

Results for “learning based systems”

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 793 records · Page 44

Enhancing Lattice Kinetic Schemes for Fluid Dynamics with Lattice-Equivariant Neural Networks

A new class of equivariant neural networks is presented, hereby dubbed lattice-equivariant neural networks (LENNs), designed to satisfy local symmetries of a lattice structure. The approach develops within a recently introduced framework aimed at learning neural network-based surrogate models’ lattice Boltzmann collision operators. Whenever neural networks are employed to model physical systems, respecting symmetries and equivariance properties has been shown to be key for accuracy, numerical stability, and performance. Here, hinging on ideas from group representation theory, trainable layers are defined whose algebraic structure is equivariant with respect to the symmetries of the lattice cell. In this work, the presented method naturally allows for efficient implementations, in terms of both memory usage and computational costs, supporting scalable training/testing for lattices in two spatial dimensions and higher (in which the size of symmetry group grows). The approach is validated and tested considering 2D and 3D flowing dynamics, both in laminar and turbulent regimes. It is compared with group-averaged-based symmetric networks and with plain, nonsymmetric, networks, showing how the presented approach unlocks the (a posteriori) accuracy and training stability of the former models and the train/inference speed of the latter networks. (LENNs are about one order of magnitude faster than group-averaged networks in 3D.) The work in this paper opens toward practical use of machine learning-augmented lattice Boltzmann CFD in real-world simulations.

97 MATHEMATICS AND COMPUTING↗

Learning continuous models for continuous physics

Abstract Dynamical systems that evolve continuously over time are ubiquitous throughout science and engineering. Machine learning (ML) provides data-driven approaches to model and predict the dynamics of such systems. A core issue with this approach is that ML models are typically trained on discrete data, using ML methodologies that are not aware of underlying continuity properties. This results in models that often do not capture any underlying continuous dynamics—either of the system of interest, or indeed of any related system. To address this challenge, we develop a convergence test based on numerical analysis theory. Our test verifies whether a model has learned a function that accurately approximates an underlying continuous dynamics. Models that fail this test fail to capture relevant dynamics, rendering them of limited utility for many scientific prediction tasks; while models that pass this test enable both better interpolation and better extrapolation in multiple ways. Our results illustrate how principled numerical analysis methods can be coupled with existing ML training/testing methodologies to validate models for science and engineering applications.

97 MATHEMATICS AND COMPUTING↗

Secure Collaborative Environment for Seamless Sharing of Scientific Knowledge

In a secure collaborative environment, tera-bytes of data generated from powerful scientific instruments are used to train secure machine learning (ML) models on exascale computing systems, which are then securely shared with internal or external collaborators as cloud-based services. Devising such a secure platform is necessary for seamless scientific knowledge sharing without compromising individual, or institute-level, intellectual property and privacy details. By enabling new computing opportunities with sensitive data, we envision a secure collaborative environment that will play a significant role in accelerating scientific discovery. Several recent technological advancements have made it possible to realize these capabilities. In this paper, we present our efforts at ORNL toward developing a secure computation platform. We present a use case where scientific data generated from complex instruments, like those at the Spallation Neutron Source (SNS), are used to train a differential privacy enabled deep learning (DL) network on Summit, which is then hosted as a secure multi-party computation (MPC) service on ORNL’s Compute and Data Environment for Science (CADES) cloud computing platform for third-party inference. In this feasibility study, we discuss the challenges involved, elaborate on leveraged technologies, analyze relevant performance results and present the future vision of our work to establish secure collaboration capabilities within and outside of ORNL.

Yoginath, Srikanth↗

Transmission-distribution long-term volt-var planning considering reactive power support capability of distributed PV

High penetration of grid-edge, inverter-based photovoltaic (PV) can cause significant voltage fluctuations not only at the distribution but also at the sub-transmission levels due to PV output intermittency. Traditional reactive power planning approaches do not consider intermittency, nor the possibility of coordinating the control of existing and future volt-ampere reactive resources. This paper proposes a reactive power planning tool for sub-transmission systems to mitigate voltage violations and fluctuations caused by high PV penetration and intermittency with a minimum investment cost. The planning tool coordinates with an optimization-based volt-var operational tool for: a) modeling the coordination of all existing var assets in both sub-transmission and distribution systems to reduce the need of new equipment, and b)selecting a set of scenarios with voltage violations, derived from PV intermittency c) testing the final investment decision. The tool obtains an investment need for each intermittency scenario with a proposed optimal power-flow framework with efficient techniques to handle a high number of discrete variables. Two options are provided for final planning decision: i) a conservative direct combination of investment need solutions and ii) a machine learning-based selection of representative investment needs at most time steps. The final investment decision options are verified using a realistic large-scale sub-transmission system and 5-minute PV and load data. The results show a significant voltage performance improvement with a lower investment cost for additional var equipment compared to conventional approaches.

14 SOLAR ENERGY↗

SynthNotes: A Generator Framework for High-volume, High-fidelity Synthetic Mental Health Notes

One of the key, emerging challenges that connects the "Big Data" and the AI domain is the availability of sufficient volumes of training data for AI/Machine Learning tasks. SynthNotes is a framework for generating standards-compliant, realistic mental health progress report notes at the very large, population-level scale, and in a strict privacy-preserving manner. Our framework, inspired by the needs to explore, evaluate, and train computational methods for the emerging mental health crisis in the US, is useful for benchmarking, optimization, and training of biomedical natural language processing, information extraction, and machine learning systems intended to operate at "Big Data" scale (billions of notes). The free text notes generated by SynthNotes are based on the literature and public statistical models allowing for realistic, natural language representation of a patient, and his or her mental health characteristics. Additionally, SynthNotes can partially simulate stylistic, grammatical, and expressive characteristics of a licensed mental health professional. SynthNotes is modular and flexible, allowing for representation of variety of conditions, incorporation of alternative foundational models, and parametrization of the variability of the structure, content, and size of the synthetically generated corpus. In this paper, we report on the initial use and performance characteristics of our SynthNotes framework and on the ongoing work for inclusion of content planning and deep learning-based generative methods trained on real data.

Begoli, Edmon↗

Fast Frequency Response using Reinforcement Learning-Controlled Wind Turbines

To fulfill the auxiliary grid services such as load regulation, spin and non-spin reserve, and frequency support during emergencies, power system operators often require certain wind farms to operate in de-loaded modes. By leveraging the fast response capability of wind farms, the reserved power in deloaded modes can significantly enhance the stability and reliability of power grids. This paper presents a novel methodology that incorporates wind turbines into reinforcement learning-based solutions for frequency response. The proposed approach employs the state-of-the-art reinforcement learning algorithm, surrogategradient- based evolution strategy (GSES), for continuous control of the wind farm output. Our methodology is tested on a modified IEEE-39 bus system, and simulation outcomes demonstrate that the proposed approach can reliably support the frequency of the power system and prevent unnecessary load shedding.

Gao, Wei↗

Machine learning reveals factors that control ion mobility in anti-perovskite solid electrolytes

Solid-state batteries are projected to exhibit improved energy densities and safety compared to liquid-electrolyte-based systems. Consequently, the development of solid electrolytes that can conduct cations at rates comparable to that of liquid electrolytes is an active area of research. To facilitate this search it is desirable to understand the chemical and structural features that control ion transport through a crystalline lattice. Here, machine learning (ML) is used to identify features that influence ion mobility in solids, quantify the relative importance of those features, and demonstrate how the variation of individual features alters ion mobility. Adopting the anti-perovskite lattice as a model system, ML algorithms were trained to predict migration barriers from a data set containing hundreds of barriers calculated using density functional theory. In this study, more than 10 6 feature sets of non-redundant descriptors were evaluated for their ability to predict migration barriers. The present analysis correctly reiterates the importance of several features that are known to influence mobility – for example, anion polarizability – but also identifies others whose importance is not widely recognized. Lattice properties such as hopping distance and channel width exhibit the greatest influence on cation mobility. Individual conditional expectation analyses performed on these features shows that barriers decrease as hopping distance decreases and as channel width increases. The defect formation energy – which is commonly associated with the concentration of defects, but not with their mobility – is identified as an important feature for interstitial migration. In sum, this analysis aids in the design of optimal solid electrolytes by simplifying the multi-dimensional design space to a sub-set of properties that are the most important.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Bayesian stability and force modeling for uncertain machining processes

Accurately simulating machining operations requires knowledge of the cutting force model and system frequency response. However, this data is collected using specialized instruments in an ex-situ manner. Bayesian statistical methods instead learn the system parameters using cutting test data, but to date, these approaches have only considered milling stability. This paper presents a physics-based Bayesian framework which incorporates both spindle power and milling stability. Initial probabilistic descriptions of the system parameters are propagated through a set of physics functions to form probabilistic predictions about the milling process. The system parameters are then updated using automatically selected cutting tests to reduce parameter uncertainty and identify more productive cutting conditions, where spindle power measurements are used to learn the cutting force model. The framework is demonstrated through both numerical and experimental case studies. Results show that the approach accurately identifies both the system natural frequency and cutting force model.

42 ENGINEERING↗

Machine-learning Kohn–Sham potential from dynamics in time-dependent Kohn–Sham systems

Abstract The construction of a better exchange-correlation potential in time-dependent density functional theory (TDDFT) can improve the accuracy of TDDFT calculations and provide more accurate predictions of the properties of many-electron systems. Here, we propose a machine learning method to develop the energy functional and the Kohn–Sham potential of a time-dependent Kohn–Sham (TDKS) system is proposed. The method is based on the dynamics of the Kohn–Sham system and does not require any data on the exact Kohn–Sham potential for training the model. We demonstrate the results of our method with a 1D harmonic oscillator example and a 1D two-electron example. We show that the machine-learned Kohn–Sham potential matches the exact Kohn–Sham potential in the absence of memory effect. Our method can still capture the dynamics of the Kohn–Sham system in the presence of memory effects. The machine learning method developed in this article provides insight into making better approximations of the energy functional and the Kohn–Sham potential in the TDKS system.

97 MATHEMATICS AND COMPUTING↗

Estimating Lossy Compressibility of Scientific Data Using Deep Neural Networks

Simulation based scientific applications generate increasingly large amounts of data on high-performance computing (HPC) systems. To allow data to be stored and analyzed efficiently, data compression is often utilized to reduce the volume and velocity of data. However, a question often raised by domain scientists is the level of compression that can be expected so that they can make more informed decisions, balancing between accuracy and performance. In this letter, we propose a deep neural network based approach for estimating the compressibility of scientific data. To train the neural network, we build both general features as well as compressor-specific features so that the characteristics of both data and lossy compressors are captured in training. Our approach is demonstrated to outperform a prior analytical model as well as a sampling based approach in the case of a biased estimation, i.e., for SZ. However, for the unbiased estimation (i.e., ZFP), the sampling based approach yields the best accuracy, despite the high overhead involved in sampling the target dataset.

97 MATHEMATICS AND COMPUTING↗

Data-driven particle dynamics: Structure-preserving coarse-graining for emergent behavior in non-equilibrium systems

Multiscale systems are ubiquitous in science and technology, but are notoriously challenging to simulate as short spatiotemporal scales must be appropriately linked to emergent bulk physics. When expensive high-dimensional dynamical systems are coarse-grained into low-dimensional models, the entropic loss of information leads to emergent physics which are dissipative, history-dependent, and stochastic. To machine learn coarse-grained dynamics from time-series observations of particle trajectories, we propose a framework using the metriplectic bracket formalism that preserves these properties by construction; most notably, the framework guarantees discrete notions of the first and second laws of thermodynamics, conservation of momentum, and a discrete fluctuation-dissipation balance crucial for capturing non-equilibrium statistics. We introduce the mathematical framework abstractly before specializing to a particle discretization. As labels are generally unavailable for entropic state variables, we introduce a novel self-supervised learning strategy to identify emergent structural variables. We validate the method on benchmark systems and demonstrate its utility on two challenging examples: (1) coarse-graining star polymers at challenging levels of coarse-graining while preserving non-equilibrium statistics, and (2) learning models from high-speed video of colloidal suspensions that capture coupling between local rearrangement events and emergent stochastic dynamics. We provide open-source implementations in both PyTorch and LAMMPS, enabling large-scale inference and extensibility to diverse particle-based systems.

Computational Engineering, Finance, and Science (c↗

Autonomous Sensor System for Wind Turbine Blade Collision Detection

This paper presents an automated blade collision detection system for use on wind turbines, toward the goal of supporting monitoring and quantitative assessment of wind energy impacts on wildlife. A wireless, multisensor module mounted at the blade root measures surface vibrations, and a blade-mounted camera provides image capture of colliding objects. Using sensor data recorded during field testing of the system on an operational wind turbine, we present the development, training, and testing of automated detection algorithms for collision detection using machine-learning approaches. In particular, we compare the use of a new two-step, anomaly-based classification algorithm with conventional adaptive boosting and amplitude-based detection techniques, where the two-step approach improves average precision for the experimental data set. This integrated sensor and classification systems demonstrates a new approach for automated, on-blade collision detection for wind turbines, with broad utility across structural health monitoring applications.

17 WIND ENERGY↗

Modeling I/O Performance Variability Using Conditional Variational Auto Encoders

Storage system performance modeling is crucial for efficient use of heterogeneous shared resources on leadership class computers. Variability in application performance, particularly variability arising from concurrent applications sharing I/O resources, is a major hurdle in the development of accurate performance models. We adopt a deep learning approach based on conditional variational auto encoders (CVAE) for I/O performance modeling, and use it to quantify performance variability. We illustrate our approach using the data collected on Edison, a production supercomputing system at the National Energy Research Scientific Computing Center (NERSC). The CVAE approach is investigated by comparing it to a previously proposed sensitivity-based Gaussian process (GP) model. We find that the CVAE model performs slightly better than the GP model in cases where training and testing data come from different applications, since CVAE can inherently leverage the whole data from multiple applications whereas GP partitions the data and builds separate models for each partition. Hence, the CVAE offers an alternative modeling approach that does not need preprocessing; it has enough flexibility to handle data from a wide variety of applications without changing the inference approach.

I/O perfornance variability↗

Early Fault Detection in Particle Accelerator Power Electronics Using Ensemble Learning

Early fault detection and fault prognosis are crucial to ensure efficient and safe operations of complex engineering systems such as the Spallation Neutron Source (SNS) and its power electronics (high voltage converter modulators). Following an advanced experimental facility setup that mimics SNS operating conditions, the authors successfully conducted 21 early fault detection experiments, where fault precursors are introduced in the system to a degree enough to cause degradation in the waveform signals, but not enough to reach a real fault. Nine different machine learning techniques based on ensemble trees, convolutional neural networks, support vector machines, and hierarchical voting ensembles are proposed to detect the fault precursors. Although all 9 models have shown a perfect and identical performance during the training and testing phase, the performance of most models has decreased in the next test phase once they got exposed to realworld data from the 21 experiments. The hierarchical voting ensemble, which features multiple layers of diverse models, maintains a distinguished performance in early detection of the fault precursors with 95% success rate (20/21 tests), followed by adaboost and extremely randomized trees with 52% and 48% success rates, respectively. The support vector machine models were the worst with only 24% success rate (5/21 tests). The study concluded that a successful implementation of machine learning in the SNS or particle accelerator power systems would require a major upgrade in the controller and the data acquisition system to facilitate streaming and handling big data for the machine learning models. In addition, this study shows that the best performing models were diverse and based on the ensemble concept to reduce the bias and hyperparameter sensitivity of individual models.

43 PARTICLE ACCELERATORS↗

Accelerate Performance (Final Technical Report)

Encouraging commercial building owners to set measurable energy goals before design begins can drive design and contractor teams to develop innovative energy efficiency solutions within conventional building budgets. A federal building owner piloted this technique and created a large-scale zero-energy office building on a budget comparable to local market construction rates. The successes and lessons learned from this performance-based project formed the basis of Accelerate Performance, an expanded program using utilities as the primary outreach channel to replicate the approach. Four utilities partnered with Slipstream, a not-for-profit program implementer, to develop pilot incentive offerings to focus building owners on the whole building rather than on individual building components and systems. The original performance-based procurement process and its successful implementation in a federal building project informed all these efforts. The experiences and lessons learned in different new construction markets led to further process adaptations as well as the development of templates and tools to streamline outreach and technical support efforts. The approach can be adapted to utility new construction programs and is a powerful tool for engaging many building owners early in the project planning process as they make critical energy decisions that impact the performance of the building. The approach has led to real energy savings in building projects by encouraging cost-effective, low-risk, and creative design solutions. The performance-based approach was tested in a variety of applications including different building types and owner types, with many owners adopting performance-based procurement techniques, some with variations, but all maintaining the same core principles. Over the course of the project, nearly 100 building projects were impacted and additional projects will continue to be supported through ongoing program offerings with three leading utilities.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

Final Report: Assessment of Combined Heat and Power Premium Power Applications in California

This report analyzes the current economic and environmental performance of combined heat and power (CHP) systems in power interruption intolerant commercial facilities. Through a series of three case studies, key trade-offs are analyzed with regard to the provision of black-out ridethrough capability with the CHP systems and the resutling ability to avoid the need for at least some diesel backup generator capacity located at the case study sites. Each of the selected sites currently have a CHP or combined heating, cooling, and power (CCHP) system in addition to diesel backup generators. In all cases the CHP/CCHP system have a small fraction of the electrical capacity of the diesel generators. Although none of the selected sites currently have the ability to run the CHP systems as emergency backup power, all could be retrofitted to provide this blackout ride-through capability, and new CHP systems can be installed with this capability. The following three sites/systems were used for this analysis: (1) Sierra Nevada Brewery - Using 1MW of installed Molten Carbonate Fuel Cells operating on a combination of digestor gas (from the beer brewing process) and natural gas, this facility can produce electricty and heat for the brewery and attached bottling plant. The major thermal load on-site is to keep the brewing tanks at appropriate temperatures. (2) NetApp Data Center - Using 1.125 MW of Hess Microgen natural gas fired reciprocating engine-generators, with exhaust gas and jacket water heat recovery attached to over 300 tons of of adsorption chillers, this combined cooling and power system provides electricity and cooling to a data center with a 1,200 kW peak electrical load. (3) Kaiser Permanente Hayward Hospital - With 180kW of Tecogen natural gas fired reciprocating engine-generators this CHP system generates steam for space heating, and hot water for a city hospital. For all sites, similar assumptions are made about the economic and technological constraints of the power generation system. Using the Distributed Energy Resource Customer Adoption Model (DER-CAM) developed at the Lawrence Berkeley National Laboratory, we model three representative scenarios and find the optimal operation scheduling, yearly energy cost, and energy technology investments for each scenario below: Scenario 1 - Diesel generators and CHP/CCHP equipment as installed in the current facility. Scenario 1 represents a baseline forced investment in currently installed energy equipment. Scenario 2 - Existing CHP equipment installed with blackout ride-through capability to replace approximately the same capacity of diesel generators. In Scenario 2 the cost of the replaced diesel units is saved, however additional capital cost for the controls and switchgear for blackout ride-through capability is necessary. Scenario 3 - Fully optimized site analysis, allowing DER-CAM to specify the number of diesel and CHP/CCHP units (with blackout ride-through capability) that should be installed ignoring any constraints on backup generation. Scenario 3 allows DER-CAM to optimize scheduling and number of generation units from the currently available technologies at a particular site. The results of this analysis, using real data to model the optimal schedulding of hypothetical and actual CHP systems for a brewery, data center, and hospital, lead to some interesting conclusions. First, facilities with high heating loads will typically prove to be the most appropriate for CHP installation from a purely economic standpoint. Second, absorption/adsorption cooling systems may only be economically feasible if the technology for these chillers can increase above current best system efficiency. At a coefficient of performance (COP) of 0.8, for instance, an adsorption chiller paired with a natural gas generator with waste heat recovery at a facility with large cooling loads, like a data center, will cost no less on a yearly basis than purchasing electricity and natural gas directly from a utility. Third, at marginal additional cost, if the reliability of CHP systems proves to be at least as high as diesel generators (which we expect to be the case), the CHP system could replace the diesel generator at little or no additional cost. This is true if the thermal to electric (relative) load of those facilities was already high enough to economically justify a CHP system. Last, in terms of greenhouse gas emissions, the modeled CHP and CCHP systems provide some degree of decreased emissions relative to systems with less CHP installed. The emission reduction can be up to 10% in the optimized case (Scenario 3) in the application with the highest relative thermal load, in this case the hospital. Although these results should be qualified because they are only based on the three case studies, the general results and lessons learned are expected to be applicable across a broad range of potential and existing CCHP systems.

25 ENERGY STORAGE↗

Three dimensional cluster analysis for atom probe tomography using Ripley’s K-function and machine learning

The size and structure of spatial molecular and atomic clustering can significantly impact material properties and is therefore important to accurately quantify. Ripley’s K-function (K(r)), a measure of spatial correlation, can be used to perform such quantification when the material system of interest can be represented as a marked point pattern. This work demonstrates how machine learning models based on K (r)-derived metrics can accurately estimate cluster size and intra-cluster density in simulated three dimensional (3D) point patterns containing spherical clusters of varying size; over 90% of model estimates for cluster size and intra-cluster density fall within 11% and 18% error of the true values, respectively. These K (r)-based size and density estimates are then applied to an experimental APT reconstruction to characterize MgZn clusters in a 7000 series aluminum alloy. Here we find that the estimates are more accurate, consistent, and robust to user interaction than estimates from the popular maximum separation algorithm. Using K (r) and machine learning to measure clustering is an accurate and repeatable way to quantify this important material attribute.

36 MATERIALS SCIENCE↗

Electronic specific heat capacities and entropies from density matrix quantum Monte Carlo using Gaussian process regression to find gradients of noisy data

In this work, we present a machine learning approach to calculating electronic specific heat capacities for a variety of benchmark molecular systems. Our models are based on data from density matrix quantum Monte Carlo, which is a stochastic method that can calculate the electronic energy at finite temperature. As these energies typically have noise, numerical derivatives of the energy can be challenging to find reliably. In order to circumvent this problem, we use Gaussian process regression to model the energy and use analytical derivatives to produce the specific heat capacity. From there, we also calculate the entropy by numerical integration. We compare our results to cubic splines and finite differences in a variety of molecules in which Hamiltonians can be diagonalized exactly with full configuration interaction. We finally apply this method to look at larger molecules where exact diagonalization is not possible and make comparisons with more approximate ways to calculate the specific heat capacity and entropy.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗