Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “decision model”

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 1,171 records · Page 65
All sourcesNASA NTRSDOE OSTI
DOE OSTI2020

Optimal Sizing of PV and Energy Storage in an Electric Vehicle Extreme Fast Charging Station

This paper proposes an optimization model for the optimal sizing of photovoltaic (PV) and energy storage in an electric vehicle extreme fast charging station considering the coordinated charging strategy of the electric vehicles. The proposed model minimizes the annualized cost of the extreme fast charging station, including investment and maintenance cost of PV and energy storage, cost of purchasing energy from utility and demand charge. The decision variables are capacity of invested PV and the power and energy ratings of invested energy storage. To further reduce the annualized cost of the extreme fast charging station, the charging strategy of electric vehicles are integrated into the optimization model and coordinated with the power output of PV and charging/discharging of energy storage. Results of numerical simulations indicate that investment of PV and energy storage could help reduce the annualized cost of the extreme fast charging station significantly. Meanwhile, the impacts of various parameters on the optimal solution are investigated by sensitivity analysis.

Liu, Guodong↗
DOE OSTI2021

Distributional Deep Reinforcement Learning-Based Emergency Frequency Control

Emergency frequency control is one of the most critical approaches to maintain power system stability after major disturbances. With the increasing number of grid-connected renewable energy sources, existing model-based methods of frequency control are facing up with challenges of computational speed and scalability for large-scale systems. In this paper, the emergency frequency control problem is formulated as a Markov Decision Process (MDP) and solved through a novel Distributional Deep Reinforcement Learning (DDRL) method, namely the distributional soft actor critic (DSAC) method. Compared with other RL methods that only estimate the mean value, the proposed DSAC model estimates the distribution of value function over returns. This advancement can lead to more insights and knowledge for the agent, with the benefit of a much faster and more stable learning process, and the improved frequency control performance. Here, the simulation results on IEEE 39-bus and IEEE 118-bus systems demonstrate the effectiveness and robustness of proposed models, as well as the advantage compared to other state-of-the-art DRL algorithms.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2025

Evaluation of the Energy, Hygrothermal, and Thermal Capacity Performance of Cross-Laminated Timber

Cross-laminated timber (CLT) construction is gaining momentum in the US because it offers multiple advantages over traditional construction methods. Benefits that have received the most attention focus on constructability, the environment, and protection (e.g., blast resistance), although CLT construction is likely to offer other benefits, as well. Still, these have not been studied at length because such evaluations are costly, requiring long-term assessments in an actual building and specialized technical knowledge. Among the possible benefits, CLT construction likely provides a higher-performing building envelope. Using CLT panels to enclose a building means fewer joints in the opaque envelope than what is required in traditional stick-framed construction. Fewer joints mean fewer locations where the air- and water-resistive barrier (WRB) could be compromised; thus, a CLT building enclosure may require less maintenance and have a longer lifespan than a traditionally built structure because of fewer air and water leaks. In addition, CLT’s thermal mass moderates indoor temperatures, allowing the heating, ventilation, and air conditioning (HVAC) system to operate more efficiently during peak hours, reducing operational energy consumption throughout the lifetime of the CLT building (Salonvaara et al., 2022). Furthermore, more stable indoor temperatures can increase occupant comfort. The CLT’s thermal mass can also reduce energy costs by adjusting to utility time-of-use pricing without affecting occupant comfort. The ability of CLT buildings to bridge periods without HVAC operation prepares them for future grid interaction and provides a certain level of resilience against power outages. Researchers have attempted to quantify these benefits; however, their work is based on simplified simulations with numerous assumptions. To correctly understand the benefits, an actual building must be monitored. Therefore, information needs to be gathered on indoor and outdoor temperatures, HVAC energy consumption, thermostat setpoints, temperatures, and thermal transport in CLT components to comprehend how these parameters are affected by the CLT’s thermal mass. These data are needed to reduce the number of assumptions and calibrate simulation models to optimize HVAC controls to minimize overall energy consumption, reduce energy use and higher fees during peak demand, and maintain occupant comfort. Additionally, the calibrated simulation model allows the optimization exercise to be repeated in various US climates. Potential benefits can be tailored to buildings in various locations, and decisions can be made on where CLT construction could be most advantageous. Furthermore, monitoring and simulation results are needed to evaluate the durability of the CLT structures in different climates. This project’s researchers gathered information to help understand and quantify the benefits of CLT buildings concerning operational energy, moderated indoor temperatures, and comfort; the dynamic operation to provide grid services; and resilience in times of power outage. Through the corroboration of simulation models with real-world measurements, this study paves the way for extrapolating findings to other climatic zones and building typologies, thereby broadening the understanding of CLT’s multifaceted benefits and reinforcing its position as a material of choice in sustainable construction.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2023

Language Model Inventory v1.0 [Slides]

This presentation is an inventory of the latest large language model (LLM) offerings. It lists some of the most prominent open- and closed-source LLMs. Given the rapid pace of progress in this space, it is not guaranteed to be a comprehensive list. This resource is intended to help inform decision-making about LLM capabilities at NREL.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Adaptively driven X-ray diffraction guided by machine learning for autonomous phase identification

Machine learning (ML) has become a valuable tool to assist and improve materials characterization, enabling automated interpretation of experimental results with techniques such as X-ray diffraction (XRD) and electron microscopy. Because ML models are fast once trained, there is a key opportunity to bring interpretation in-line with experiments and make on-the-fly decisions to achieve optimal measurement effectiveness, which creates broad opportunities for rapid learning and information extraction from experiments. Here, we demonstrate such a capability with the development of autonomous and adaptive XRD. By coupling an ML algorithm with a physical diffractometer, this method integrates diffraction and analysis such that early experimental information is leveraged to steer measurements toward features that improve the confidence of a model trained to identify crystalline phases. We validate the effectiveness of an adaptive approach by showing that ML-driven XRD can accurately detect trace amounts of materials in multi-phase mixtures with short measurement times. The improved speed of phase detection also enables in situ identification of short-lived intermediate phases formed during solid-state reactions using a standard in-house diffractometer. Our findings showcase the advantages of in-line ML for materials characterization and point to the possibility of more general approaches for adaptive experimentation.

36 MATERIALS SCIENCE↗
DOE OSTI2026

Cross-sectoral synergies for household energy savings: the role of electric vehicles, solar photovoltaics, and remote work

Over the past two decades, new technologies and behavioral shifts – such as electric vehicles, solar photovoltaics, and increased work-from-home practices – have reshaped residential electricity consumption and cost. However, their combined or synergistic impact on a household’s electricity cost remains unexplored. Leveraging data from the 2020 Residential Energy Consumption Survey, this study uses a structural equation model to unravel the extent to which the bundled adoption of EV-PV and stay-at-home decisions impact the total electricity cost of households. Results indicate that adopting both EVs and PV reduces electricity costs by 31% despite a 16% rise in consumption. When combined with stay-at-home practices, households still experience a 13% cost reduction, even with a 25% increase in electricity consumption. In conclusion, these findings suggest that financial savings are not merely a byproduct of adopting new technologies or behavioral changes but could serve as a key consideration for household contemplating engagement with multiple modern energy solutions simultaneously.

14 SOLAR ENERGY↗
DOE OSTI2024

Pipeline Hydrogen Decarbonization and Repurposing Analyzers (P-HyDRAs)

The Pipeline Hydrogen Decarbonization and Repurposing Analyzers (P-HyDRAs) are a set of prototype computational tools for simulating and optimizing midstream natural gas pipeline system operations subject to location and time-dependent hydrogen blending. The models can accurately resolve dynamic gas flows through large-scale pipeline networks using non-ideal gas equations of state. The codes can be used as decision support for planning and design decisions involving intra-day energy flow schedules as well as spatiotemporal economic values of natural gas, hydrogen, and net energy delivered to consumers while ensuring that pipeline hydraulic limitations, gas compressor station constraints, operational factors, and pre-existing shipping contracts are satisfied. The inputs to the codes are a model of the pipeline system as well as time-series data that specify boundary conditions on the network. For optimization, the code module requires price and quantity offers for natural gas and hydrogen and price and quantity bids for energy, which are used as time-dependent constraints in an optimal control problem. The outputs are time-series data that provide a predictive simulation of gas flows, mass fractions, and pressures, or with additional degrees of freedom give an approximately optimal solution for gas injections/withdrawals, compressor settings, and sensitivities to the objective function that provide locational values of energy.

Zlotnik, Anatoly↗
DOE OSTI2020

Site-wide occupancy assessment using camera traps for seven mammalian species at Los Alamos National Laboratory

Los Alamos National Laboratory (LANL or Laboratory) is committed to solving national security challenges through scientific excellence and has been serving the nation and northern New Mexico for over 70 years. Being located on the Pajarito Plateau in the eastern flanks of the Jemez Mountains, the Laboratory is surrounded by a rich diversity of plants and animals. It is common to see many different species of wildlife on Laboratory property; however, sometimes interactions with wildlife can be negative. Vehicle accidents with wildlife have become a common occurrence. With the current and ongoing expansion of the Laboratory on the Pajarito Plateau, it has the potential to further impact wildlife movement including large game species. Local agencies and tribal Pueblos rely on large game species and do not want these species to be restricted from moving across property boundaries. Temporal and spatial aspects of where wildlife occur on the site is a phenomena that is either not well understood in uncommon species or needs periodic reevaluation for common species. Estimating the distribution of multiple species across the landscape provides wildlife biologists with crucial information for monitoring and conserving animal populations in a particular area. Utilizing motion activated wildlife cameras, also known as camera traps, to monitor wildlife populations has become an essential tool for biologists. Camera traps are non-invasive and cost-effective and can document multiple elusive or uncommon wildlife species, such as carnivores, simultaneously. Occupancy modeling provides a flexible framework for the analysis of the distribution for multiple wildlife species. It explicitly recognizes whether a species is spatially common or rare (occupancy = ψ) and if that species is easy or hard to detect (detection probability = p ). Multispecies and multi-season occupancy models can detect trends in species occupancy because individual species may vary in seasonal movements, detection probability, and transition rates between habitats. In this study, we assessed the site as a whole to ascertain when and where medium and large mammal species are present. Understanding wildlife patterns at the Laboratory will better inform future management decisions regarding land use and development strategies. We placed motion activated wildlife cameras in a random systematic sampling design and used these data to create occupancy models. We tested for differences in single-species occupancy and detection probability by season of mammal species captured on 20 camera traps placed across the Laboratory in a 40 mi² (103 km²) area. We focus the interpretation of our findings on seven mammal species found during this study. They are Rocky Mountain elk ( Cervus canadensis nelsoni ; hereafter “elk”), mule deer ( Odocoileus hemionus ; hereafter “deer”), mountain lion ( Puma concolor ; hereafter “lion”), American black bear ( Ursus americanus ; hereafter “bear”), coyote ( Canis latrans ), bobcat ( Lynx rufus ), and gray fox ( Urocyon cinereoargenteus ; hereafter “fox”).

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2026

A human-in-the-loop explanation framework for morphologically transparent AI predictions from whole-slide images

Deep learning models enable the prediction of clinical endpoints from whole-slide images (WSIs), but many such models function as “black boxes”, lacking transparency about whether and which histomorphological patterns drive their predictions, hindering interpretability and clinical adoption. Here we propose a human-in-the-loop explanation framework, MorphoXAI, which provides both local and global interpretability for deep learning models by incorporating human-expert interpretations. At the global level, it reveals the histomorphological patterns on which the model consistently relies to distinguish between classes of WSIs, as well as the patterns associated with confusion between classes. At the local level, it indicates which of these patterns are used in the prediction of an individual WSI and which regions within the slide correspond to such patterns. We validated our method across multiple deep learning–based WSI analysis tasks spanning different tissue types. The results show that our framework generates explanations that accurately reflect the histomorphology underlying the model’s predictions at both global and local levels. For interpretability and clinical utility in diagnostic contexts, human evaluation results showed that our explanations were easy to interpret, rich in diagnostic features, and directly helpful for diagnostic decision-making, thereby enhancing pathologist-AI collaboration. Our work highlights that unifying global and local explanations and grounding them in expert-interpreted morphology enhances the interpretability and verifiability of deep learning models, thereby facilitating the transparent deployment of such models in clinical practice.

Lou, Peiliang↗
DOE OSTI2023

A General Framework for Human-in-the-Loop Cognitive Digital Twins

Modelling and analysis of systems that are equipped with sensors and connected to the Internet are becoming more automated and less human-dependent. However, bringing expert knowledge into the loop along with data obtained from Internet of Thing (IoT) devices minimizes the risk of making poor and unexplainable decisions and helps to assess the impact of different strategies before applying them in reality. While Digital Twins are more of a data-driven simulation of the physical system, Cognitive Digital Twins bring the human dimension into the modelling and simulation. In this paper, we aim to emphasize the crucial role of explainability and the underlying rationale behind automated or interactive decision-making processes. Furthermore, we propose an initial framework that delineates the specific points within the feedback loop of a cognitive digital twin where human involvement can be incorporated.

Niloofar, Parisa↗
DOE OSTI2022

System Analysis Modeling and Intermodal Transportation for Commercial Spent Nuclear Fuel

The United States Department of Energy (DOE) Office of Nuclear Energy is applying knowledge and understanding in the areas of systems level engineering, analysis, and decision making to better inform the waste management pathways for U.S. spent nuclear fuel (SNF). Currently there are 93 operating and 23 shutdown commercial nuclear reactors in the United States. SNF at most of these locations is being stored in spent fuel pools, dry storage, or both [1, 2]. This paper initially goes over the basics of the agent-based simulation tool known as the Next Generation System Analysis Model (NGSAM) that has the capability to model interaction and movements of individual components or groups like casks, railcars etc. The next section covers some of the methods implemented in the Java Transportation Operations Model (JTOM) to facilitate the movement of assemblies, casks, railcars etc. and the various intermodal transfer options. This section also covers the current cask loading and intermodal transfer times implemented in NGSAM. The last section goes over the proposed values for intermodal transportation, and cask transfer/loading operations.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2021

Nuclear Safeguards: Feature Extraction for Machine Learning Enrichment Analysis of UF 6 Cylinders

With the increasing international interest in using nuclear material as sources of energy, comes the growing concern that we may see an increase in proliferation threats. Uranium hexafluoride (UF 6 ) is used in the nuclear fuel cycle for uranium enrichment. Inspectors from the International Atomic Energy Agency (IAEA) monitor the enrichment levels of UF 6 stored in transportation cylinders, to ensure the enrichments match that of a facility operator’s declarations. However, only a characteristic subset of UF 6 cylinders can be measured by inspectors from the IAEA during these inspections. In turn, the inspector may not identify a cylinder whose enrichment levels do not match the facility declarations. Therefore, our team attempts to develop a machine learning network that could determine the enrichment percentage of cylinders as they enter and exit facilities. The machine learning model must be robust against spectral variations due to factors that are internal and external to the UF 6 cylinder. Such factors include but are not limited to the speed and distance of the moving vehicle, cylinder type, cylinder orientation, and fill level. Many features in the spectral continuum can be used to identify and correct for some of these variations. As a consequence, the model must extract numerous features in the continuum. We investigate new approaches in continuum subtraction to verify enrichment percentage: interpolation and extrapolation of lines at the notable characteristic spectral peaks. It was determined that there was no clear answer on which method proved superior, therefore the decision was made to implement both new methods into the current codebase. We currently rely on synthetic training and testing data to analyze the results and fine tune our models but hope to test the system on measured data soon.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗
DOE OSTI2020

Adaptive Recovery Model: Designing Systems for Testing Tracing and Vaccination to Support COVID-19 Recovery Planning.

This report documents a new approach to designing disease control policies that allocate scarce testing, contact tracing, and vaccination resources to better control community transmission of COVID19 or similar diseases. The Adaptive Recovery Model (ARM) combines a deterministic compartmental disease model with a stochastic network disease propagation model to enable us to simulate COVID-19 community spread through the lens of two complementary modeling motifs. ARM contact networks are derived from cell-phone location data that have been anonymized and interpreted as individual arrivals to specic public locations. Modeling disease spread over these networks allows us to identify locations within communities conducive to rapid disease spread. ARM applies this model- and data-derived abstractions of community transmission to evaluate the effectiveness of disease control measures including targeted social distancing, contact tracing, testing and vaccination. The architecture of ARM provides a unique capacity to help decision makers understand how best to deploy scarce testing, tracing and vaccination resources to minimize disease-spread potential in a community. This document details the novel mathematical formulations underlying ARM, presents a dynamical stability analysis of the deterministic model components, a sensitivity analysis of control parameters and network structure, and summarizes a process for deriving contact networks from cell-phone location data. An example use case steps through applying ARM to evaluate three targeted social distancing policies using Bernalillo County, New Mexico as an exemplar test locale. This step-by-step analysis demonstrates how ARM can be used to measure the relative performance of competing public health policies. Initial scenario tests of ARM shows that ARMs design focus on resource utilization rather than simple incidence prediction can provide decision makers with additional quantitative guidance for managing ongoing public health emergencies and planning future responses.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
DOE OSTI2020

VADER: A Tool for Criticality Safety Validation

The purpose of criticality safety is to prevent any inadvertent criticality from occurring during the handling or storage of fissile material. Calculations are frequently used to demonstrate that a sufficient subcritical margin exists. Validation is a key aspect of the evaluation process, establishing the suitability, accuracy, and associated uncertainty of the computational method and data to be used for the intended application. The validation process is performed by comparing the results of critical experiments with the calculated results from models of the experiments using the computational method to be validated. Laboratory critical experiments are controlled systems that achieve a k eff of approximately 1 in order to investigate the parameters at which such a critical condition is achieved. The validation parameters that are traditionally applied to safety analysis calculations are the bias and the bias uncertainty . The bias is the deviation of the average k eff of the validation suite from unity. The bias uncertainty accounts for the statistical uncertainty in the bias based on the standard deviation, sample size, and distribution of k eff values of the validation suite. The values of bias and bias uncertainty ensure that the systems predicted to be subcritical by the computational method will indeed be subcritical. The bias and bias uncertainty are often combined to determine an upper subcritical limit (USL) or computational margin that can then be applied to safety analysis calculations. Many methods have been developed by different organizations to calculate the bias and bias uncertainty for various types of criticality analyses. Each of these methods typically requires that the validity of various underpinning statistical assumptions be confirmed to demonstrate that the method is appropriate for the analysis of a given validation suite. An example of the validation decision making flow is shown in Fig.1. As shown in Fig. 1, the analyst performing the validation fits a trend line to the data and performs a test to determine if the trend was a statistically better representation of the data than if it were treated as an uncorrelated sample. If the trend line is a better representation of the data, then the analyst uses any one of a number of trending techniques to determine the bias and bias uncertainty. If a trend is not an appropriate representation of the data, then the analyst proceeds to perform a normality assessment for the data. If the normal assumption can be shown to be acceptable, then the analyst calculates the bias and bias uncertainty with the parametric technique. If the assumption of normality cannot be justified, then the nonparametric technique is used. Once the decision flow has been followed and the appropriate technique has been selected, the bias and bias uncertainty is typically combined with an administrative margin to determine a USL below which calculated values of k eff for safety analysis models can be considered subcritical. The calculations used in each decision are often performed with spreadsheets or with small programs available at various sites performing criticality analyses. Expertise in understanding and interpreting the results must be maintained to perform these calculations. This can often be an error-prone process. Oak Ridge National Laboratory (ORNL) is currently developing the Validation and Data Evaluation Resource (VADER) to simplify and automate the criticality safety validation process and to provide a software quality assurance pedigree to the calculational methods used. This paper discusses the use of the Fulcrum user interface with VADER, the anticipated initial capabilities of VADER to perform validation analyses, and the output from the code.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2020

Stochastic Analysis for Long Term Capital Structures, Systems, and Components Refurbishment and Replacement

As commercial Nuclear Power Plants (NPPs) pursue extended plant operation in the form of Second License Renewal (SLR), opportunities exist for these plants to provide capital investments to ensure long-term safe and economic performance. At the current time, several utilities have announced an intention to pursue extended operation for one or more of their NPPs via SLR . The goal of this research is to develop a risk-informed approach to evaluate and prioritize plant capital investments made in preparation for, and during the period of, extended plant operations to support decisions for NPP operations. Since the capital investments are influenced by various factors, such as markets, safety and regulatory, the decision-making process of NPP operations should take into account relevant factors for balancing risks, costs and profits. The traditional method of capital budgeting is based on the priority list of candidate projects using economic measures such as benefit-investment ratio, net present value (NPV) and internal rate of return. In the literatures, the problem of capital budgeting or the variant can be represented by an appropriate knapsack problem. The knapsack approach to capital budgeting takes as input as investment, along with the cost and profit of each project. The objective of capital budgeting is to find the combination of the binary decisions for every investment such that the overall profit is as large as possible. The output is a collection of projects to be carried out, and we refer this selected collection of projects as a project portfolio. One limitation of traditional optimization models for capital budgeting is that they do not account for risk/uncertainty in profit and cost streams associated with individual projects, they do not account for risk in resource availability in future years [1,2,3]. Projects can incur cost over-runs, especially when projects are large, performed infrequently, and when there is risk regarding technical viability, external contractors, and/or suppliers of requisite parts and materials. Occasionally, projects are performed ahead of schedule and with cost savings. Planned budgets for capital improvements can be cut and key personnel may be lost. Or, there may be surprise windfalls in budgets for maintenance activities due to decreased costs for “unplanned” maintenance. In these cases, how should we resolve capital budgeting when we have risk forecasts for costs, profits and budgets? One approach we proposed in this summary is to re-solve the optimization models based on assumed statistical distributions of given parameters. If these distributions were not available, a two-stage stochastic optimization approach can be used to provide priority lists to decision-makers to support better risk-informed decisions [4, 5]. In this summary, we will only focus on the first approach.

42 ENGINEERING↗
DOE OSTI2025

Analyzing and Exploring Training Recipes for Large-Scale Transformer-Based Weather Prediction

Abstract The rapid rise of deep learning (DL) in numerical weather prediction (NWP) has led to a proliferation of models which forecast atmospheric variables with comparable or superior skill than traditional physics-based NWP. However, among these leading DL models, there is a wide variance in both the training settings and architecture used. Further, the lack of thorough ablation studies makes it hard to discern which components are most critical to success. In this work, we show that it is possible to attain high forecast skill even with relatively off-the-shelf architectures, simple training procedures, and moderate compute budgets. Specifically, we train a minimally modified Swin Transformer V2 (SwinV2) on ERA5 data and find that it attains superior skill in terms of mean-square errors of deterministic forecasts when compared against the European Centre for Medium-Range Weather Forecasts’ Integrated Forecasting System (IFS). Almost all DL–NWP systems share a core set of hyperparameters and design decisions. To aid and expedite future DL–NWP research, we present an in-depth, systematic exploration of different loss functions, model sizes and depths, patch sizes, and multistep training objectives. We also examine the model performance with metrics beyond the typical accuracy (ACC) and RMSE and investigate how the performance scales with model size. Through our open-source code, scoring pipelines, and models, we share our findings on key aspects of the training pipeline. These ablations reduce the necessity for expensive hyperparameter tuning and lower the barrier to entry for future DL–NWP research. Significance Statement This study investigates the potential of using large-scale transformer-based models for weather prediction, showing that it is possible to achieve high forecast accuracy with simpler, off-the-shelf architectures. By training a minimally modified SwinV2 transformer on ERA5 data, we show that the model achieves competitive forecast skill in terms of mean-square error for key variables, outperforming the European Centre for Medium-Range Weather Forecasts’ Integrated Forecasting System (IFS) at all lead times. Our findings suggest that effective training strategies, such as multistep fine-tuning and channel-weighted losses, significantly enhance the model’s performance. However, we also highlight that these improvements come with trade-offs in other areas, such as ensemble spread and high-frequency spatial detail. This work highlights the promise of deep learning in improving weather forecasts, which could lead to better preparedness and response to weather events, ultimately benefiting society by providing more reliable weather predictions.

Willard, Jared D. [Lawrence Berkeley National Labo↗
DOE OSTI2024

Learning model combining convolutional deep neural network with a self-attention mechanism for AC optimal power flow

Alternating current optimal power flow (OPF) analysis is critical for efficient and reliable operation of power systems. For large systems or repetitive computations, the traditional methods such as the direct and gradient methods, or non-traditional methods, such as the genetic algorithm and simulating annealing, are time-consuming and unsuitable for real-time computing. The work in this paper proposes a novel framework to obtain the optimal solution of power flow in real-time using a combination of convolutional neural networks and a self-attention mechanism. All parameters of the power networks are rearranged in an image-like shape of a multi-channel image where each channel is a two-dimensional matrix. The proposed approach is adaptive with every input size of power systems as well as frequent variations of network topologies without intervention to the framework core. The encompassment of all power system contexts in which all parameters of internal elements, generation costs, and topology information are included, contributes to the higher accuracy of inference compared to other current machine-learning-based OPF-solving methods. Besides, the proposed framework established on ubiquitous platforms is effortlessly integrated into current infrastructures of power systems, and the great efficiency along with the computation speed may serve as a critical point for practical implications, such as enabling faster decision-making during real-time operations, predicting system contingencies, and remedial actions based on an offline pre-trained model. Furthermore, this supervised learning process is applied to the dataset of four case studies of meshed power systems: the IEEE 5-bus system (IEEE-5), the IEEE 30-bus system (IEEE-30), the IEEE 39-bus system (IEEE-39), and the IEEE 57-bus system (IEEE-57) to prove the efficacy of the proposed method.

42 ENGINEERING↗
DOE OSTI2020

Agent-Based Coordination Scheme for PV Integration (ABC4PV)

Renewables and especially photovoltaics (PV) have benefitted significantly from a host of incentives and policies targeted toward enhanced integration and adoption of specific energy technologies. However, with the push to move forward into a subsidy-free market framework, behind-the-meter residential PV applications have generally struggled to retain their value (unlike utility scale and commercial projects) [1]. This project focused on developing control-theoretic solutions aimed at improving the integration and interaction of behind-the-meter residential PV with other distribution system assets (controllable and non-controllable) to enhance the integrated value of residential PV. To this end, a suite of decentralized control methodologies have been developed to enable effective coordination and control of behind-the-meter residential load customers’ PV, battery storage systems (BSS), controllable loads and other similar assets within a distribution feeder. This interaction aims at procuring energy savings and, thus, energy bill savings. The main source of savings is drawn from reducing the effect of demand charge pricing and is realized at the feeder level, assuming community level interaction and management among the aforementioned assets. Optimal control of the assets is implemented with a distributed optimization methodology, leveraging consensus-based algorithms. The results gathered from the optimal control simulations demonstrates that the savings can be duly achieved and the algorithm decision times (to dynamically control asset set points, for example) are fast. As for the overall efficiency of PV+BSS systems, to procure energy savings from curtailment of the demand charge pricing effects, the optimal control is set up so as to minimize the variance of the load for all customers, throughout a feeder and throughout time in a rolling horizon scheduling with model predictive control. The control takes into account inter-temporal electrochemical storage (battery) degradation costs: specifically, we have developed a long-term lifetime model for the BSS that weighs in the effect of the degradation factor in the dispatch formulations, thus, a considerable operating cost that affects energy decision making. The levelized cost of energy (LCOE – redefined for the purpose of quantifying asset integration effectiveness through the customers’ energy cost) is shown to be below the threshold set for the combined PV+BSS topology of $ 0.14/kWh for multiple cases of PV penetration all the way up to 50%, provided that a policy of shared ownership of and savings is in place. Further, the LCOE calculated for the case before the deployment PV+BSS systems is also achievable, i.e. the deployment of PV+BSS, if planned and scheduled optimally. will have no effect on customers’ energy costs. From the control methodology viewpoint, the developed consensus-based algorithms are shown to converge for a wide range of problem cases (spanning normal operating scenarios and contingencies), guaranteeing dispatch solutions under forecasting errors, communication break-downs and cyber-security attacks. The proposed control solutions are scalable and real-time implementable, with dispatch computations and device set-point updates converging in less than 2s in most practical instances of the above events.

14 SOLAR ENERGY↗