Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Continual learning”

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 811 records · Page 45

Machine Learning-Assisted High-Temperature Reservoir Thermal Energy Storage Optimization: Numerical Modeling and Machine Learning Input and Output Files

This data set includes the numerical modeling input files and output files used to synthesize data, and the reduced-order machine learning models trained from the synthesized data for reservoir thermal energy storage site identification. In this study, a machine-learning-assisted computational framework is presented to identify High-Temperature Reservoir Thermal Energy Storage (HT-RTES) site with optimal performance metrics by combining physics-based simulation with stochastic hydrogeologic formation and thermal energy storage operation parameters, artificial neural network regression of the simulation data, and genetic algorithm-enabled multi-objective optimization. A doublet well configuration with a layered (aquitard-aquifer-aquitard) generic reservoir is simulated for cases of continuous operation and seasonal-cycle operation scenarios. Neural network-based surrogate models are developed for the two scenarios and applied to generate the Pareto fronts of the HT-RTES performance for four potential HT-RTES sites. The developed Pareto optimal solutions indicate the performance of HT-RTES is operation-scenario (i.e., fluid cycle) and reservoir-site dependent, and the performance metrics have competing effects for a given site and a given fluid cycle. The developed neural network models can be applied to identify suitable sites for HT-RTES, and the proposed framework sheds light on the design of resilient HT-RTES systems. All the simulations and the neural network model were done by Idaho National Laboratory. A detailed description of the work was reported in publication linked below.

15 GEOTHERMAL ENERGY↗

A Learning Framework for Control-Oriented Modeling of Buildings

Buildings consume a significant amount of energy worldwide. Several building optimization and control use cases require models of energy consumption which are control oriented, have high predictive capability, imposes minimal data pre-processing requirements, and have the ability to be adapted continuously to account for changing conditions as new data becomes available. Data driven modeling techniques, that have been investigated so far, while promising in the context of buildings, have been unable to simultaneously satisfy all the requirements mentioned above. In this context, deep learning techniques such as Recurrent Neural Networks (RNNs) hold promise, empowered by advanced computational capabilities and big data opportunities. In this paper, we propose a deep learning based methodology for the development of control oriented models for building energy management and test in on data from a real building. Results show that the proposed methodology outperforms other data driven modeling techniques significantly. We perform a detailed analysis of the proposed methodology along dimensions such as topology, sensitivity, and downsampling. Lastly, we conclude by envisioning a building analytics suite empowered by the proposed deep framework, that can drive several use cases related to building energy management.

buildings↗

Code for the manuscript titled "Blackout Diffusion: Generative Diffusion Models in Discrete-State Spaces"

We would like to disclose two scripts, written in Jupyter notebook, in which we implement the "Blackout Diffusion Process" described in the manuscript "Blackout Diffusion: Generative Diffusion Models in Discrete-State Spaces" (LA-UR-23-20509), to be submitted to the International Conference of Machine Learning (ICML). The abstract of the manuscript is append below. == Typical generative diffusion models rely on a Gaussian diffusion process for training the backward transformations, which can then be used generate samples from Gaussian noise. However, real world data often takes place in discrete-state spaces, which occur in many scientific applications. Here, we develop a theoretical formulation for arbitrary discrete-state Markov processes in the forward diffusion process. We relate the theory to the existing continuous-state Gaussian diffusion and identify the corresponding reverse-time stochastic process and score function in the continuous-time setting, and the reverse-time mapping in the discrete-time setting. As an example of this framework, we introduce "Blackout Diffusion", which learns to produce samples from an empty image instead of from noise. Numerical experiments on the CIFAR-10 dataset confirm the feasibility of generative diffusion modeling in a discrete space. Generalizing from specific (Gaussian) forward processes to a more general framework also sheds light on how to interpret generative diffusion models and their mathematical structure, which we comment on.

Lin, Yen Ting↗

BayesCLUMPY: BAYESIAN INFERENCE WITH CLUMPY DUSTY TORUS MODELS

Our aim is to present a fast and general Bayesian inference framework based on the synergy between machine learning techniques and standard sampling methods and apply it to infer the physical properties of clumpy dusty torus using infrared photometric high spatial resolution observations of active galactic nuclei. We make use of the Metropolis-Hastings Markov Chain Monte Carlo algorithm for sampling the posterior distribution function. Such distribution results from combining all a priori knowledge about the parameters of the model and the information introduced by the observations. The main difficulty resides in the fact that the model used to explain the observations is computationally demanding and the sampling is very time consuming. For this reason, we apply a set of artificial neural networks that are used to approximate and interpolate a database of models. As a consequence, models not present in the original database can be computed ensuring continuity. We focus on the application of this solution scheme to the recently developed public database of clumpy dusty torus models. The machine learning scheme used in this paper allows us to generate any model from the database using only a factor of 10{sup -4} of the original size of the database and a factor of 10{sup -3} in computing time. The posterior distribution obtained for each model parameter allows us to investigate how the observations constrain the parameters and which ones remain partially or completely undetermined, providing statistically relevant confidence intervals. As an example, the application to the nuclear region of Centaurus A shows that the optical depth of the clouds, the total number of clouds, and the radial extent of the cloud distribution zone are well constrained using only six filters. The code is freely available from the authors.

79 ASTRONOMY AND ASTROPHYSICS↗

deadtrees.earth — An open-access and interactive database for centimeter-scale aerial imagery to uncover global tree mortality dynamics

Excessive tree mortality is a global concern and remains poorly understood as it is a complex phenomenon. We lack global and temporally continuous coverage on tree mortality data. Ground-based observations on tree mortality, e.g., derived from national inventories, are very sparse, and may not be standardized or spatially explicit. Earth observation data, combined with supervised machine learning, offer a promising approach to map overstory tree mortality in a consistent manner over space and time. However, global-scale machine learning requires broad training data covering a wide range of environmental settings and forest types. Low altitude observation platforms (e.g., drones or airplanes) provide a cost-effective source of training data by capturing high-resolution orthophotos of overstory tree mortality events at centimeter-scale resolution. Here, we introduce deadtrees.earth, an open-access platform hosting more than two thousand centimeter-resolution orthophotos, covering more than 1,000,000 ha, of which more than 58,000 ha are manually annotated with live/dead tree classifications. This community-sourced and rigorously curated dataset can serve as a comprehensive reference dataset to uncover tree mortality patterns from local to global scales using space-based Earth observation data and machine learning models. This will provide the basis to attribute tree mortality patterns to environmental changes or project tree mortality dynamics to the future. The open nature of deadtrees.earth, together with its curation of high-quality, spatially representative, and ecologically diverse data will continuously increase our capacity to uncover and understand tree mortality dynamics.

Citizen science↗

A representation-independent electronic charge density database for crystalline materials

Abstract In addition to being the core quantity in density-functional theory, the charge density can be used in many tertiary analyses in materials sciences from bonding to assigning charge to specific atoms. The charge density is data-rich since it contains information about all the electrons in the system. With the increasing prevalence of machine-learning tools in materials sciences, a data-rich object like the charge density can be utilized in a wide range of applications. The database presented here provides a modern and user-friendly interface for a large and continuously updated collection of charge densities as part of the Materials Project. In addition to the charge density data, we provide the theory and code for changing the representation of the charge density which should enable more advanced machine-learning studies for the broader community.

36 MATERIALS SCIENCE↗

A representation-independent electronic charge density database for crystalline materials

In addition to being the core quantity in density functional theory, the charge density can be used in many tertiary analyses in materials sciences from bonding to assigning charge to specific atoms. The charge density is data-rich since it contains information about all the electrons in the system. With increasing utilization of machine-learning tools in materials sciences, a data-rich object like the charge density can be utilized in a wide range of applications. The database presented here provides a modern and user-friendly interface for a large and continuously updated collection of charge densities as part of the Materials Project. In addition to the charge density data, we provide the theory and code for changing the representation of the charge density which should enable more advanced machine-learning studies for the broader community.

36 MATERIALS SCIENCE↗

Utility and Industry Perceptions of Control Room Modernization Over the Last 10 Years

Within the current United States (U.S.) nuclear power fleet, main control room modernization (CRM) is an important step towards cost savings. In recent decades, plants have been engaged in upgrades to varying degrees. This process requires a nuanced, balanced, and timely approach that ensures continued safety and long-term sustainability. In 2012 a survey was issued to individuals from the nuclear industry to learn their perspectives on a range of CRM issues. The survey targeted the benefits and challenges for utilities undertaking this process, including the main drivers and barriers to technology upgrades, regulatory compliance, and the effects these factors have on concepts of operations, strategic approaches, and staffing. In 2022, the survey was issued again to understand whether CRM perceptions had changed in the last 10 years. Our findings identify changes in industry thinking from a decade ago. Here we reveal perspective shifts that represent increased optimism and, in some instances, increased doubt regarding the opportunities and challenges inherent in CRM and implementation. We also report nuanced differences in CRM perspectives between utility and surrounding nuclear industry respondents.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

STag. II. Classification of Serendipitous Supernovae Observed by Galaxy Redshift Surveys

With the number of supernovae observed expected to drastically increase thanks to large-scale surveys like the Dark Energy Spectroscopic Instrument (DESI), it is necessary that the tools we use to classify these objects keep up with this increase. We previously created Supernova Tagging and Classification (STag) to address this problem by employing machine learning techniques alongside logistic regression in order to assign “tags” to spectra based on spectral features. STag II is a continuation of this work, which now makes use of model supernova spectra combined with real DESI spectra in order to train STag to better deal with realistic data. Furthermore, we also make use of the rlap score as a trustworthiness cut, making for a more robust and accurate supernova classifier than before.

Astrostatistics techniques↗

Probabilistic Wind Power Ramp Forecasting Based on a Scenario Generation Method

Wind power ramps (WPRs) are particularly important in the management and dispatch of wind power and currently drawing the attention of balancing authorities. With the aim to reduce the impact of WPRs for power system operations, this paper develops a probabilistic ramp forecasting method based on a large number of simulated scenarios. An ensemble machine learning technique is first adopted to forecast the basic wind power forecasting scenario and calculate the historical forecasting errors. A continuous Gaussian mixture model (GMM) is used to fit the probability distribution function (PDF) of forecasting errors. The cumulative distribution function (CDF) is analytically deduced. The inverse transform method based on Monte Carlo sampling and the CDF is used to generate a massive number of forecasting error scenarios. An optimized swinging door algorithm is adopted to extract all the WPRs from the complete set of wind power forecasting scenarios. The probabilistic forecasting results of ramp duration and start-time are generated based on all scenarios. Numerical simulations on publicly available wind power data show that within a predefined tolerance level, the developed probabilistic wind power ramp forecasting method is able to predict WPRs with a high level of sharpness and accuracy.

17 WIND ENERGY↗

Inertial Confinement Fusion R&D and Nuclear Proliferation

In a few months, or a few years, the National Ignition Facility (NIF) at the Lawrence Livermore National Laboratory may achieve fusion gain using 192 powerful lasers to generate x-rays that will compress and heat a small target containing isotopes of hydrogen. This event would mark a major milestone after decades of research on inertial confinement fusion (ICF). It might also mark the beginning of an accelerated global effort to harness fusion energy based on this science and technology. Unlike magnetic confinement fusion (ITER, 2011), in which hot fusion fuel is confined continuously by strong magnetic fields, inertial confinement fusion involves repetitive fusion explosions, taking advantage of some aspects of the science learned from the design and testing of hydrogen bombs. The NIF was built primarily because of the information it would provide on weapons physics, helping the United States to steward its stockpile of nuclear weapons without further underground testing. The U.S. National Academies' National Research Council is now hosting a study to assess the prospects for energy from inertial confinement fusion. While this study has a classified sub-panel on target physics, it has not been charged with examining the potential nuclear proliferation risks associated with ICF R&D. We argue here that this question urgently requires direct and transparent examination, so that means to mitigate risks can be assessed, and the potential residual risks can be balanced against the potential benefits, now being assessed by the NRC. This concern is not new (Holdren, 1978), but its urgency is now higher than ever before.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Fast, Dynamic, and Scalable Algorithms for Large-Scale Constrained Optimization

Scientists and engineers constantly aim to optimize an objective subject to physical, environmental, and/or resource constraints. The technique of using mathematical models to formulate and find real solutions of such problems is known as mathematical optimization, a process that has become invaluable for design and discovery in numerous scientific fields. This project has focused on the design, analysis, and implementation of high-performance computing algorithms for solving cutting-edge optimization problems. These include problems that involve (1) data uncertainties, such as those in the future supply, demand, and capacity of a given power system, (2) extreme numbers of alternatives, such as in the design of electrical power grids to avoid network vulnerabilities, and (3) real-time decisions, such as in the control of chemical reactors. The key features of the new algorithms are that they are fast, dynamic, and scalable to meet the computational requirements of scientists and researchers working to optimize large-scale, complex systems. The algorithms that have been designed, analyzed, and tested as part of the project can be characterized based on the types of optimization problems they have been designed to solve. Much of the work has focused on algorithms for solving large-scale constrained optimization problems. Contributions have also been made in the design of algorithms for unconstrained optimization, with some specialized methods for solving problems that have inherent nonsmoothness. Contributions have also been made in areas that have grown to prominence in the field over the five-year award period. These include the design of algorithms for solving large-scale stochastic optimization problems, such as those that arise in large-scale machine learning, and in complexity analyses for solving nonconvex problems, which has been a focal point of much recent work in continuous optimization. Overall, the project has supported work that has led to almost 30 articles that have been published or accepted to top journals in the field of mathematical optimization.

97 MATHEMATICS AND COMPUTING↗

Hydrogen Filling Station

Hydrogen is an environmentally attractive transportation fuel that has the potential to displace fossil fuels. The Freedom CAR and Freedom FUEL initiatives emphasize the importance of hydrogen as a future transportation fuel. Presently, Las Vegas has one hydrogen fueling station powered by natural gas. However, the use of traditional sources of energy to produce hydrogen does not maximize the benefit. The hydrogen fueling station developed under this grant used electrolysis units and solar energy to produce hydrogen fuel. Water and electricity are furnished to the unit and the output is hydrogen and oxygen. Three vehicles were converted to utilize the hydrogen produced at the station. The vehicles were all equipped with different types of technologies. The vehicles were used in the day-to-day operation of the Las Vegas Valley Water District and monitoring was performed on efficiency, reliability and maintenance requirements. The research and demonstration utilized for the reconfiguration of these vehicles could lead to new technologies in vehicle development that could make hydrogen-fueled vehicles more cost effective, economical, efficient and more widely used. In order to advance the development of a hydrogen future in Southern Nevada, project partners recognized a need to bring various entities involved in hydrogen development and deployment together as a means of sharing knowledge and eliminating duplication of efforts. A road-mapping session was held in Las Vegas in June 2006. The Nevada State Energy Office, representatives from DOE, DOE contractors and LANL, NETL, NREL were present. Leadership from the National hydrogen Association Board of Directors also attended. As a result of this session, a roadmap for hydrogen development was created. This roadmap has the ability to become a tool for use by other road-mapping efforts in the hydrogen community. It could also become a standard template for other states or even countries to approach planning for a hydrogen future. Project partners also conducted a workshop on hydrogen safety and permitting. This provided an opportunity for the various permitting agencies and end users to gather to share experiences and knowledge. As a result of this workshop, the permitting process for the hydrogen filling station on the Las Vegas Valley Water District’s land was done more efficiently and those who would be responsible for the operation were better educated on the safety and reliability of hydrogen production and storage. The lessons learned in permitting the filling station and conducting this workshop provided a basis for future hydrogen projects in the region. Continuing efforts to increase the working pressure of electrolysis and efficiency have been pursued. Research was also performed on improving the cost, efficiency and durability of Proton Exchange Membrane (PEM) hydrogen technology. Research elements focused upon PEM membranes, electrodes/catalysts, membrane-electrode assemblies, seals, bipolar plates, utilization of renewable power, reliability issues, scale, and advanced conversion topics. Additionally, direct solar-to-hydrogen conversion research to demonstrate stable and efficient photoelectrochemistry (PEC) hydrogen production systems based on a number of optional concepts was performed. Candidate PEC concepts included technical obstacles such as inefficient photocatalysis, inadequate photocurrent due to non-optimal material band gap energies, rapid electron-hole recombination, reduced hole mobility and diminished operational lifetimes of surface materials exposed to electrolytes. Project Objective 1: Design, build, operate hydrogen filling station Project Objective 2: Perform research and development for utilizing solar technologies on the hydrogen filling station and convert two utility vehicles for use by the station operators Project Objective 3: Increase capacity of hydrogen filling station; add additional vehicle; conduct safety workshop; develop a roadmap for hydrogen development; accelerate the development of photovoltaic components Project Objective 4: Perform research on the Proton Exchange membrane

08 HYDROGEN↗

Visual and near-IR spectrophotometry of asteroids

We have been continuing our studies of the spectral properties of dark asteroids in the solar system. From these studies we expect to learn about the distribution of volatile materials, such as water in clay materials (water of hydration) and how the asteroids may relate to the comets. Our most recent work has been concentrating on simultaneous visual and near infrared photometry near Earth, main belt, and trojan asteroids. We have made observations of some unusual asteroids such as Chiron, which has recently shown cometary activity, and 944 Hidalgo, which has a comet-like orbit. We have also begun studies of the small, dark satellites of Mars and Jupiter in order to understand better how they may relate to the steroids. Could they actually be captured asteroids or comets?

Lebofsky, Larry A.↗

The K-8 Aeronautics Internet Textbook

Efforts were focused on web site migration, from UC (University of California) Davis to the National Business Aviation Association's (NBAA) web site. K8AIT (K-8 Aeronautics Internet Textbook), which has remained an unadvertised web site, receives almost two million hits per month. Project continuation funding with the National Business Aviation Association is being pursued. A Memorandum of Understanding (MOU) between NASA Ames LTP (Learning Technologies Project) and Cislunar has been drafted and approved by NASA's legal department. Additional web content on space flight and the Wright brothers has been added in English and Spanish.

Source record↗

Project Galileo: completing Europa, preparing for Io

Galileo has completed the Europa leg of the Galileo Europa Mission, and is now pumping down the apojove in each succeeding orbit in preparation for the Io phase. Including three encounters earlier in the primary mission, the total of ten close passes by Europa have provided a wealth of interesting and provocative information about this intriguing body. The results presented include new and exciting information about Europa's interactions with Jupiter's magnetosphere, its interior structure, and its tantalizing surface features, which strongly hint at a watery subsurface layer. Additional data concerning Callisto, and its own outlook for a subsurface ocean are also presented. In addition the engineering aspects of operating the spacecraft during the past year are explored, as well as a brief examination of what will be the challenges to prepare for the Io encounters. The steadily increasing radiation dosage that the spacecraft is experiencing is well beyond the original design parameters, and is contributing to a number of spacecraft problems and concerns. The ability of the flight team to analyze and solve these problems, even at the reduced staffing levels of an extended mission, is a testament to their tenacity and loyalty to the mission. The engineering data being generated by these continuing radiation-induced anomalies will prove invaluable to designers of future spacecraft to Jupiter and its satellites. The lessons learned during this arduous process are presented. c 2000 International Astronautical Federation. Published by Elsevier Science Ltd. All rights reserved.

long duration↗

NASA Exploration Launch Projects Overview: The Crew Launch Vehicle and the Cargo Launch Vehicle Systems

The U.S. Vision for Space Exploration (January 2004) serves as the foundation for the National Aeronautics and Space Administration's (NASA) strategic goals and objectives. As the NASA Administrator outlined during his confirmation hearing in April 2005, these include: 1) Flying the Space Shuttle as safely as possible until its retirement, not later than 2010. 2) Bringing a new Crew Exploration Vehicle (CEV) into service as soon as possible after Shuttle retirement. 3) Developing a balanced overall program of science, exploration, and aeronautics at NASA, consistent with the redirection of the human space flight program to focus on exploration. 4) Completing the International Space Station (ISS) in a manner consistent with international partner commitments and the needs of human exploration. 5) Encouraging the pursuit of appropriate partnerships with the emerging commercial space sector. 6) Establishing a lunar return program having the maximum possible utility for later missions to Mars and other destinations. In spring 2005, the Agency commissioned a team of aerospace subject matter experts to perform the Exploration Systems Architecture Study (ESAS). The ESAS team performed in-depth evaluations of a number of space transportation architectures and provided recommendations based on their findings? The ESAS analysis focused on a human-rated Crew Launch Vehicle (CLV) for astronaut transport and a heavy lift Cargo Launch Vehicle (CaLV) to carry equipment, materials, and supplies for lunar missions and, later, the first human journeys to Mars. After several months of intense study utilizing safety and reliability, technical performance, budget, and schedule figures of merit in relation to design reference missions, the ESAS design options were unveiled in summer 2005. As part of NASA's systems engineering approach, these point of departure architectures have been refined through trade studies during the ongoing design phase leading to the development phase that begins in 2008. Comprehensive reviews of engineering data and business assessments by both internal and independent reviewers serve as decision gates to ensure that systems can fully meet customer and stakeholder requirements. This paper provides the current CLV and CaLV configuration designs and gives examples of the progress being made during the first year of this significant effort. Safe, reliable, cost-effective space transportation systems are a foundational piece of America s future in space and the next step in realizing the plan for revitalizing lunar capabilities on the passageway to the human exploration of Mars. While building on legacy knowledge and heritage hardware for risk reduction, NASA will apply lessons learned from developing these new launch vehicles to the growth path for future missions. The elements for mission success and continued U.S. leadership in space have been assembled over the past year. As NASA designs and develops these two new systems over the next dozen years, visible progress, such as that reported in this paper, may sustain the national will to stay the course across political administrations and weather the inevitable trials that will be experienced during this challenging endeavor.

Snoddy, Jimmy R.↗

Multisensor Arrays for Greater Reliability and Accuracy

Arrays of multiple, nominally identical sensors with sensor-output-processing electronic hardware and software are being developed in order to obtain accuracy, reliability, and lifetime greater than those of single sensors. The conceptual basis of this development lies in the statistical behavior of multiple sensors and a multisensor-array (MSA) algorithm that exploits that behavior. In addition, advances in microelectromechanical systems (MEMS) and integrated circuits are exploited. A typical sensor unit according to this concept includes multiple MEMS sensors and sensor-readout circuitry fabricated together on a single chip and packaged compactly with a microprocessor that performs several functions, including execution of the MSA algorithm. In the MSA algorithm, the readings from all the sensors in an array at a given instant of time are compared and the reliability of each sensor is quantified. This comparison of readings and quantification of reliabilities involves the calculation of the ratio between every sensor reading and every other sensor reading, plus calculation of the sum of all such ratios. Then one output reading for the given instant of time is computed as a weighted average of the readings of all the sensors. In this computation, the weight for each sensor is the aforementioned value used to quantify its reliability. In an optional variant of the MSA algorithm that can be implemented easily, a running sum of the reliability value for each sensor at previous time steps as well as at the present time step is used as the weight of the sensor in calculating the weighted average at the present time step. In this variant, the weight of a sensor that continually fails gradually decreases, so that eventually, its influence over the output reading becomes minimal: In effect, the sensor system "learns" which sensors to trust and which not to trust. The MSA algorithm incorporates a criterion for deciding whether there remain enough sensor readings that approximate each other sufficiently closely to constitute a majority for the purpose of quantifying reliability. This criterion is, simply, that if there do not exist at least three sensors having weights greater than a prescribed minimum acceptable value, then the array as a whole is deemed to have failed.

Immer, Christopher↗