Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Distributed Computing Resources”

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 505 records · Page 28
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Wholesale Electricity Analysis via Simulation & Learning Experiments (WEASLE): Platform Development and Pilot Competition

This document reports the development of the Wholesale Electricity Analysis via Simulation and Learning Experiments (WEASLE) platform and the pilot competition that was conducted to test the platform. Due to the increasing reliance on variable renewable energy resources for bulk power, the pilot competition, called the Energy Storage Participation Algorithm Competition (ESPA-Comp), was used to test the effect of various market designs on storage utilization and market efficiency. Basic details of the platform are provided, including an overview of the market clearing engine, the battery dispatch and degradation models, electric grid topology and resource mix, and software architecture. Two market designs were tested: a two-settlement market analogous to typical ISO design today, and a multi-settlement market that allows additional forward-trading periods during the real-time market. Results from the pilot competition show that the storage bidding problem is nontrivial and is well suited for future challenges. We find that: 1) all four teams utilized different approaches to the bidding problem, 2) different methodological approaches led to substantially different offer behaviors, 3) resource profits are clustered by team and methodological approach, 4) simulated offers reduced market surplus by about 0.5%, 5) substantially different prices between two-settlement and multi-settlement markets albeit minimal difference in overall market surplus.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2023

SCUC-DER Integration Report: Integrating Distributed Energy Resources (DER) Using Advanced Unit Commitment Models and DER Aggregation Methodologies

Distributed energy resources (DERs) are continuing to grow due to regulatory, policy, and market shifts, and it needs to be ensured that small DERs are given a level playing field with traditional resources. Legacy market processes such as unit-commitment problems were designed for a power grid consisting largely of centralized power plants. In contrast, DERs consist of many small devices with distinct operating characteristics that may or may not be connected at the transmission interconnection points, limiting their visibility to the independent system operators (ISOs) who operate wholesale electricity markets. This report details the development and initial results from a simulation platform that integrates state-of-the-art security-constrained unit commitment software, detailed feeder models, and a DER aggregator model to quantify potential DER integration issues. Quantitative results to date illustrate potential infeasible scheduling solutions from SCUC when the DERA includes aggregations of energy-limited energy storage resources. Likewise, if aggregations are not penalized for dispatch deviations, they may have incentives to deviate from the SCUC-determined resource schedules. Assumptions about the amount of aggregated demand response resources (DRRs) and the ability of DERAs to follow profit incentives have an important impact – DRRs have significant flexibility and can typically feasibly meet their SCUC schedule, but on the other hand, their profit incentives can cause unscheduled increases in load before and after DRR dispatch. Computation time results on the SCUC solver and simulation platform only show a modest increase in SCUC solution time as the number of DERAs is increased, but results to date only reflect the RTS-GMLC test system; results may show more significant solver slowdown in larger transmission systems. The largest contribution to simulation time is attributed to the DERA offer generation method, which is suggested for future improvements.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2023

A Model Calibration Method for Grid-Forming Inverters Using Iterative Bayesian Optimization

As inverter-based resources (IBRs) are rapidly deployed, especially at the distribution and microgrid levels, the need to include their accurate representations in power systems models increases. With well-calibrated models, IBR-interactions can be examined, preventing any stability and operational issues, especially in islanded systems. This paper proposes a method to calibrate generic inverter models using Bayesian optimization, but with a parameter-grouping approach to improve speed and accuracy. This approach is illustrated for a grid-forming inverter using synthetic data and field measurements from a simple IBRbased microgrid. Tests show that the calibration algorithm has modest computation cost and is robust to noisy measurements..

Biswas, Shuchismita↗
DOE OSTI2024

Efficient Distributed Sequence Parallelism for Transformer-Based Image Segmentation

We introduce an efficient distributed sequence parallel approach for training transformer-based deep learning image segmentation models. The neural network models are comprised of a combination of a Vision Transformer encoder with a convolutional decoder to provide image segmentation mappings. The utility of the distributed sequence parallel approach is especially useful in cases where the tokenized embedding representation of image data are too large to fit into standard computing hardware memory. To demonstrate the performance and characteristics of our models trained in sequence parallel fashion compared to standard models, we evaluate our approach using a 3D MRI brain tumor segmentation dataset. We show that training with a sequence parallel approach can match standard sequential model training in terms of convergence. Furthermore, we show that our sequence parallel approach has the capability to support training of models that would not be possible on standard computing resources.

Lyngaas, Isaac↗
DOE OSTI2021

Multi-Period Active Distribution Network Planning Using Multi-Stage Stochastic Programming and Nested Decomposition by SDDIP

This paper presents a multi-period active distribution network planning (ADNP) with distributed generation (DG). The objective of the proposed ADNP is to minimize the total planning cost, subject to both investment and operation constraints. The paper proposes a multi-stage stochastic optimization model to address DG uncertainties over several periods, in which the decisions are made sequentially by only using the present-stage information. A nested decomposition method is proposed which applies the stochastic dual dynamic integer programming (SDDIP) method to address computational intractabilities of the proposed ADNP approach. The presented numerical results and discussions on a 33-bus distribution system and a large-scale 906-bus system verify the effectiveness of the proposed ADNP method and its solution method.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2023

NREL: Working Towards a Superlab 2.0 With ESnet and Other National Laboratories

We can currently link models and simulations across the national laboratory complex, but imagine connecting these models with equipment, devices, and scientists so they can work together as if side by side instead of thousands of miles apart. In a future like this, researchers could run experiments that leverage thousands of devices to model the millions of devices that make up our increasingly complex energy systems. With such resources, labs could simulate, for example, the electric grid of California and validate control strategies to best serve critical loads during outages caused by wildfires. The National Renewable Energy Laboratory's (NREL's) Advanced Research on Integrated Energy Systems (ARIES) platform is making that vision a reality, one step at a time.

ARIES↗
DOE OSTI2024

Disentangling the physics of the attractive Hubbard model as a fully interacting model of fermions via the accessible and symmetry-resolved entanglement entropies

The complicated ways in which electrons interact in many-body systems such as molecules and materials have long been viewed through the lens of local electron correlation and associated correlation functions. However, quantum information science has demonstrated that more global diagnostics of quantum states like the entanglement entropy can provide a complementary and clarifying lens on electronic behavior. One particularly useful measure that can be used to distinguish between quantum and classical sources of entanglement is the accessible entanglement, the entanglement available as a quantum resource for systems subject to conservation laws, such as fixed particle number, due to superselection rules. In this work, we introduce an algorithm and demonstrate how to compute accessible and symmetry-resolved entanglements for interacting fermion systems. This is accomplished by combining an incremental version of the swap algorithm with a recursive auxiliary field quantum Monte Carlo algorithm recently developed by the authors. We apply these tools to study the pairing and charge density waves exhibited in the paradigmatic attractive Hubbard model via entanglement. We find that the particle and spin symmetry-resolved entanglements and their related full probability distribution functions show very clear—and unique—signatures of the underlying electronic behavior even when those features are less pronounced in conventional correlation functions. Altogether, this work provides a systematic means of characterizing the entanglement within quantum systems that can grant a deeper understanding of the complicated electronic behavior that underlies quantum phase transitions and crossovers in many-body systems.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2025

A digital twin platform for building performance monitoring and optimization: Performance simulation and case studies

Advancements in sensor technology, data analytics, affordable compute, and communication infrastructure have paved the way for Digital Twin technology in optimizing building operations and controls. This study presents the development of an open and interoperable web-based Digital Twin platform for integrating diverse data streams and facilitating effective user interactions. The platform utilizes modern technologies for the web framework and time-series data management, ensuring scalability and responsiveness. The backend supports seamless integration of diverse data sources and emulators, incorporating data from building sensors and meters, external weather Application Programming Interfaces, and advanced EnergyPlus simulation models of the building and its energy systems including the Distributed Energy Resources that are formulated in Functional Mockup Units. A simulation case study was conducted with FlexLab, a test facility on Lawrence Berkeley National Laboratory campus. The case study includes normal operations, Distributed Energy Resource integration, and power outage scenarios, to illustrate the Digital Twin’s ability to provide critical insights into energy performance and thermal resilience. The results demonstrated the platform’s potential as a decision-support tool for optimizing building energy performance and enhancing resilience against extreme weather events. Future work will focus on deploying the Digital Twin platform to a real building for field validation, extending its capabilities to cover more scenarios such as bidirectional Electric Vehicle interactions, and enhancing user engagement.

EnergyPlus↗
DOE OSTI2025

Fair Concurrent Training of Multiple Models in Federated Learning

Federated learning (FL) enables collaborative learning across multiple clients. In most FL work, all clients train a single learning task. However, the recent proliferation of FL applications may increasingly require multiple FL tasks to be trained simultaneously, sharing clients’ computing resources, which we call Multiple-Model Federated Learning (MMFL). Current MMFL algorithms use naïve average-based client-task allocation schemes that often lead to unfair performance when FL tasks have heterogeneous difficulty levels, as the more difficult tasks may need more client participation to train effectively. Furthermore, in the MMFL setting, we face a further challenge that some clients may prefer training specific tasks to others, and may not even be willing to train other tasks, e.g., due to high computational costs, which may exacerbate unfairness in training outcomes across tasks. We address both challenges by firstly designing FedFairMMFL, a difficulty-aware algorithm that dynamically allocates clients to tasks in each training round, based on the tasks’ current performance levels. We provide guarantees on the resulting task fairness and FedFairMMFL’s convergence rate. We then propose novel auction designs that incentivizes clients to train multiple tasks, so as to fairly distribute clients’ training efforts across the tasks, and extend our convergence guarantees to this setting. Here, we finally evaluate our algorithm with multiple sets of learning tasks on real world datasets, showing that our algorithm improves fairness by improving the final model accuracy and convergence speed of the worst performing tasks, while maintaining the average accuracy across tasks.

Federated learning↗
DOE OSTI2026

Regional inertia dynamics of U.S. interconnections: An event-based measurement approach

Power grid inertia plays a vital role in frequency stability following large disturbances, yet its distribution across the U.S. grid is highly uneven. While interconnection-wide inertia benchmarks are useful, they can mask regional variability driven by resource mix, network coupling, and geographic separation. This paper extends event-driven inertia estimation to the regional scale using field measurements from the Frequency Monitoring Network (FNET/GridEye). Starting from balancing authority and independent system operator footprints, candidate regions are refined using a composite coherency score that combines frequency-trajectory shape similarity, timing spread, and lead/lag behavior to ensure dynamic consistency. A filtered sliding difference method (FSDM) is then used to construct regional frequency trajectories, detect disturbance onset, and compute robust regional rate-of-change of frequency (RoCoF). Regional, local, and interconnection inertia are estimated by combining RoCoF with event power imbalance, and additional indicators (regional-to-system inertia ratio and inertial-support arrival time) quantify regional-to-interconnection coupling and relative regional contributions. The method is demonstrated on eleven regions across the Eastern Interconnection (EI) and the Western Electricity Coordinating Council (WECC), with the Electric Reliability Council of Texas (ERCOT) used for validation. In ERCOT, estimates compared against energy management system (EMS) values achieve a mean absolute percentage error of 17.94%. WECC exhibits consistently shorter inertial-support arrival times (0.15–0.3 s) than EI (0.7–1.1 s), highlighting contrasting coupling and disturbance-propagation behavior. Overall, the results reveal pronounced spatial heterogeneity in inertia and coupling, underscoring the value of regional monitoring for both operational decision-making and long-term system planning.

Disturbance events↗
DOE OSTI2023

Comparison of Shape Optimization Methods for Heat Exchanger Fins Using Computational Fluid Dynamics

Inverse design techniques are one way to leverage advances in 3D printing, artificial intelligence, and computational resources to achieve increased performance of heat exchangers. Two optimization techniques (genetic algorithm and particle swarm) and three geometry representations (binary level set, composite Bézier, and free form deformation) are used to increase the heat transfer and reduce the pressure drop of a heat exchanger fin. After running 210,810 OpenFOAM simulations, results indicate that a significant performance increase of the fin can be realized in less than 48 hrs, allowing for such a process to be integrated in traditional design processes. The best design increased the performance of the objective function, compared to the baseline rectangular geometry by 75%. A custom distributed infrastructure was built, allowing for all methods to reach 95% of the final objective values in a little over 4 hrs, handling 1674 OpenFoam simulations per hour.

42 ENGINEERING↗
DOE OSTI2022

Digital Twin of an ANPC inverter with integrated Design-For-Trust

Reducing costs and increasing manufacturing output while creating reliable products are three essential components needed to keep a business competitive. In the past, having an asset that could address these three challenges at the same time would be challenging. However, recent advances in computer processing capabilities coupled with increases to network speeds allowed such a system to become possible; and this is known as a Digital Twin (DT), which represents a virtual copy of a production system or a product. Here, this work exhibits a DT application of an Active Neutral Point Clamped (ANPC) inverter using a controller. The controller emulates the ANPC responses of a new Digital Signal Processing (DSP) firmware and validates this new firmware by using Design-For-Trust (DFTr) characteristics. If the new firmware is malicious, then it should be rejected, thus adding preemptive protection measures for inverter based distributed energy resources.

cybersecurity↗
DOE OSTI2026

Tula: Optimizing Time, Cost, and Generalization in Distributed Large-Batch Training

Distributed training increases the number of batches processed per iteration either by scaling-out (adding more nodes) or scaling-up (increasing the batch-size). However, the largest configuration does not necessarily yield the best performance. Horizontal scaling introduces additional communication overhead, while vertical scaling is constrained by computation cost and device memory limits. Thus, simply increasing the batch-size leads to diminishing returns: training time and cost decrease initially but eventually plateaus, creating a knee-point in the time/cost vs. batch-size pareto curve. The optimal batch-size therefore depends on the underlying model, data and available compute resources. Large batches also suffer from worse model quality due to the well-known “generalization gap”. In this paper, we present Tula, an online service that automatically optimizes time, cost, and convergence quality for large-batch training of convolutional models. It combines parallel-systems modeling with statistical performance prediction to identify the optimal batchsize. Tula predicts training time and cost within 7.5−14% error across multiple models, and achieves up to 20× overall speedup and improves test accuracy by ≈9% on average over standard large-batch training on various vision tasks, thus successfully mitigating the generalization gap and accelerating training at the same time.

Tyagi, Sahil [ORNL] (ORCID:0009000783144745)↗
DOE OSTI2026

A GPU-based compressible combustion solver for applications exhibiting disparate space and time scales

High-speed chemically active flows pose significant computational challenges due to their disparate space and time scales, with stiff chemistry often dominating simulation time. While modern scientific computing programs achieve exascale performance by leveraging graphics processing units (GPUs), existing GPU-based compressible combustion solvers face critical limitations in memory management, load balancing, and handling the highly localized nature of chemical reactions. To this end, we present a high-performance compressible reacting flow solver built on the AMReX framework and optimized for multi-GPU settings. Here, our approach addresses three GPU performance bottlenecks: memory access patterns through column-major storage optimization, computational workload variability via a bulk-sparse integration strategy for chemical kinetics, and multi-GPU load distribution for adaptive mesh refinement applications. The solver adapts existing matrix-based chemical kinetics formulations to multi-grid contexts. Using representative combustion applications, including 2D and 3D detonations and a 3D jet-in-crossflow configuration, we demonstrate 1.4–5× performance improvements over initial implementations on an in-house cluster of NVIDIA H100 GPUs, and near-ideal weak scaling on the Frontier supercomputer (Oak Ridge Leadership Computing Facility) with up to 1024 AMD Instinct MI250X GPUs. Roofline analysis reveals substantial improvements in arithmetic intensity for both convection (∼ 10 ×) and chemistry (∼ 4 ×) routines, confirming efficient utilization of GPU memory bandwidth and computational resources.

42 ENGINEERING↗
DOE OSTI2021

Can artificial intelligence and data-driven machine learning models match or even replace process-driven hydrologic models for streamflow simulation?: A case study of four watersheds with different hydro-climatic regions across the CONUS

With recent developments in computational techniques, Data-driven Machine Learning Models (DMLs) have shown great potential in simulating streamflow and capturing the rainfall-runoff relationship in given watersheds, which are traditionally fulfilled by Process-based Hydrologic Models (PHMs). There are debates on whether the DMLs can outperform and possibly replace the classical PHMs for streamflow simulation and river forecasting, but no clear conclusions have been made. This study aims to investigate whether the newer DMLs have any potential in further improving the simulation accuracy of classical PHMs, and vice versa. To do this, we compared a few popular PHMs and DMLs over four watersheds across the Continental US (CONUS) that are associated with different input, climate, and regional conditions. A total of five hydrologic models were chosen, including (1) two classical lumped models, i.e., the Sacramento Soil Moisture Accounting (SAC-SMA) and Xinanjiang (XAJ); (2) one modern distributed model, termed Coupled Routing and Excess Storage (CREST); (3) and two DMLs including an Artificial Neural Networks (ANN) and a deep learning model, termed Long Short Term Memory (LSTM). Our results demonstrated that the DMLs still significantly biased when using the baseline input scenario with the PHMs. However, the DMLs fed with delayed input scenarios had great potential and can reach high simulation accuracy. The DMLs, especially the ANN, outperformed other employed models under the rainfall-runoff relationship in which rainfall dominantly drives. Furthermore, the DMLs also showed better performance in the high-flow regime, while the PHMs had a better performance for the low-flow regime, implying both PHMs and DMLs have their own merits and are worthy of joint development. In general, our study indicated a great potential of using DMLs to simulate streamflow, but further studies are still needed to verify the transferability and scalability of DMLs in large-scale experiments, such as the Distributed Model Intercomparison Projects 1&2 conducted by National Weather Services but to compare modern DMLs and PHMs.

58 GEOSCIENCES↗
DOE OSTI2021

Acquire and Test IDAES Framework Components

The cooperation under the Grid Modernization Initiative and the Applied Energy Tri-Laboratory umbrellas requires mutual understanding of capabilities and a common effort to identify gaps existing in the available computational modeling and analysis tools. In this work, we report acquisition and testing of the National Energy Technology Laboratory’s Institute for the Design of Advanced Energy Systems (IDAES) framework, particularly including both electric power systems production cost model simulator Prescient as well as system optimization IDAES itself. Further, we consider the methods in which developments in Prescient and IDAES for grid-energy simulation and optimization may be included in the Framework for Optimization of ResourCes and Economics (FORCE) ecosystem that has been under collaborative development by Idaho National Laboratory, Argonne National Laboratory, and Oak Ridge National Laboratory. We propose two integration methods to introduce Prescient and IDAES into existing FORCE automated workflows. One integration involves using Prescient as a market reduced-order model and IDAES models similar to how existing Modelica models are used. The other integration involves wrapping the Prescient-IDAES dispatch optimization process with the stochastic capacity optimization workflow in FORCE.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2022

Validation of Power Distribution Models using Load Flow Analysis in an ADMS Environment

Electric utilities are facing the need for better monitoring, analysis, and control of their distribution systems. An accurate mathematical model is a key to both the development of cutting-edge, scalable model-based algorithms and the assessment of emerging technologies such as distributed energy resources (DER) for grid planning and operation. However, the constantly evolving nature of power distribution systems poses challenges to maintaining accurate models. In this paper, we propose a novel load flow based approach to validate power distribution models. Networked equipment models described according to the Common Information Model (CIM) standard and a measurement model are used to formulate the distribution load flow problem. First, a system admittance matrix (Ybus) is derived from device-level CIM parameters. Next, the operational parameters (dynamic Ybus and nodal injections) are extracted from the measurement model using sensor configuration and equipment state. An iterative power flow method is then used to compute nodal voltages and branch flows that are compared against the measurement data to find any inconsistencies in the networked equipment model. This approach is implemented within GridAPPS-D, an open-source standards-based platform for advanced distribution management system (ADMS) application development, and demonstrated on the IEEE 13-bus, 123-bus, and 8500-node test feeders.

Common information model, model validation, power ↗
DOE OSTI2022

Composite Power System Reliability with Renewables and Customer Flexibility

Composite Power System Reliability is defined as the computational procedure that quantifies the probability that the power system will perform the function of delivering electric power to customers adequately, on a continuous basis and with an acceptable quality. This definition leaves many details undefined and exemplifies the ambiguity in reliability analysis. The increasing deployment of wind and PV creates additional uncertainties that make reliability analysis a rather complex issue. Because of increased uncertainty the need for composite reliability analysis and utilization of results in power system planning is critical. New approaches are emerging for dealing with these problems from the operational point of view, including demand response programs, tapping on customer and distributed resource flexibility and new control approaches. The key question to be addressed is: how the new operational paradigms affect composite power system reliability. Here, this paper presents the ongoing work of the IEEE Composite System Reliability Task Force of the IEEE PES Reliability, Risk, Probability Application (RRPA) Subcommittee.

24 POWER TRANSMISSION AND DISTRIBUTION↗