Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “agent based 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 361 records · Page 20
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

A Neural Optimizer With Decision-Focused Learning for Optimal Energy Storage Operation

Here, this article introduces a neural optimizer-based framework for optimizing battery energy storage system (BESS) control for grid services, including demand charge and energy cost reduction. By leveraging decision-focused learning (DFL), the proposed framework ensures seamless integration and adaptation, significantly enhancing control performance. A patch time-series transformer is employed for peak load forecasting, incorporating aleatoric uncertainty quantification to account for forecasting uncertainties within the decision-making process. The framework utilizes a solver-in-the-loop approach to generate optimal BESS actions, which are then used to train the neural optimizer-based agent. By co-optimizing both BESS operational modes and output power within the NN, the system achieves improved performance and robustness. After initial training, the forecasting and control models are jointly fine-tuned to account for forecasting errors, further improving decision precision and efficiency through DFL. Case studies are performed to validate the performance of the framework using multiple real-world datasets, demonstrating superior performance in monthly peak load forecasting compared to state-of-the-art models. In addition, the results are compared against existing decision-making approaches. The results demonstrate a reduction in monthly peak forecasting error by approximately 15% across various performance measures and achieve an optimization gap for BESS operation that is about three times smaller compared to existing methods.

Kim, Hyeonjin [Pacific Northwest National Laborato↗
DOE OSTI2024

A novel digital lifecycle for Material‐Process‐Microstructure‐Performance relationships of thermoplastic olefins foams manufactured via supercritical fluid assisted foam injection molding

Abstract This research significantly enhances the applicability of thermoplastic olefins (TPOs) in the automotive industry using supercritical N 2 as a physical foaming agent, effectively addressing the limitations of traditional chemical agents. It merges experimental results with simulations to establish detailed material‐process‐microstructure‐performance (MP2) relationships, targeting 5–20% weight reductions. This innovative approach labeled digital lifecycle (DLC) helps accurately predict tensile, flexural, and impact properties based on the foam microstructure, along with experimentally demonstrating improved paintability. The study combines process simulations with finite element models to develop a comprehensive digital model for accurately predicting mechanical properties. Our findings demonstrate a strong correlation between simulated and experimental data, with about a 5% error across various weight reduction targets, marking significant improvements over existing analytical models. This research highlights the efficacy of physical foaming agents in TPO enhancement and emphasizes the importance of integrating experimental and simulation methods to capture the underlying foaming mechanism to establish material‐process‐microstructure‐performance (MP2) relationships. Highlights Establishes a material‐process‐microstructure‐performance (MP2) for TPO foams Sustainably produces TPO foams using supercritical (ScF) N 2 with 20% lightweighting Shows enhanced paintability for TPO foam improved surface aesthetics Digital lifecycle (DLC) that predicts both foam microstructure and properties DLC maps process effects & microstructure onto FEA mesh for precise prediction

Engineering↗
DOE OSTI2025

Towards philosophical reasoning with agentic LLMs: Socratic method for scientific assistance

As large language models (LLMs) become central tools in science, improving their reasoning capabilities is critical for meaningful and trustworthy applications. We introduce a Socratic agent for scientific reasoning, implemented through a structured system prompt that guides LLMs via classical principles of inquiry. Unlike typical prompt engineering or retrieval-based methods, our approach leverages definition, analogy, hypothesis elimination, and other Socratic techniques to generate more coherent, critical, and domain-aware responses. We evaluate the agent across diverse scientific domains and benchmark it on the abstraction and reasoning corpus challenge dataset, achieving 97.15% under a fixed prompting protocol and without fine-tuning or external tools. Expert evaluation shows improved reasoning depth, clarity, and adaptability over conventional LLM outputs, suggesting that structured prompting rooted in philosophical reasoning can improve the scientific utility of language models.

LLM reasoning↗
DOE OSTI2020

Mechanisms of Degradation of Toxic Nerve Agents: Quantum-chemical Insight into Interactions of Sarin and Soman with Molybdenum Dioxide

With the urgent need to provide an efficient and reliable protection for people from fatal chemical and biological weapons, our fundamental understanding of how toxins interact with filters is far from complete. The situation is further complicated by natural difficulties of performing experimental measurements with lethal toxins. Unlike experiments, computational modeling offers an attractive and safe yet reliable way of studying behavior of toxic agents on a variety of substrates at a great level of detail. Here, we report DFT-based quantum-chemical calculations of adsorption and decomposition of DMMP, sarin and soman on MoO 2 (011) surface. Our calculations show that MoO 2 strongly adsorbs toxic nerve agents and quickly decomposes them. Decomposition of DMMP on the MoO 2 (011) surface proceeds via the PO-CH 3 bond breaking and a formation of a surface methoxy group. The calculated activation barrier for this reaction is 131.5 kJ mol -1 . Unlike DMMP, decomposition of sarin and soman proceeds via the dealkylation reaction yielding propene and 3,3-dimethyl-1-butene, respectively. Decomposition of sarin requires a remarkably low energy (53.7 kJ mol- 1 ), whereas the similar reaction in soman requires 50 kJ mol -1 more energy. We also make specific predictions to guide Ambient-Pressure X-ray Photoelectron Spectroscopy (APXPS) experiments on sarin interaction with MoO 2 samples. We conclude that MoO 2 serves as an efficient substrate able of degrading toxins.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

Using SWMM for emergency response planning: A case study evaluating biological agent transport under various rainfall scenarios and urban surfaces

To assist in emergency preparedness for a biological agent terrorist attack or accidental pathogen release, potential contaminant levels and migration pathways of spores spread by urban stormwater were evaluated using a Storm Water Management Model (SWMM) of U.S. Coast Guard Base Elizabeth City, North Carolina. The high temporal-spatial resolution SWMM model was built using spore concentrations in stormwater runoff from asphalt, grass, and concrete collected from a point-scale field study. The subsequent modeled contamination scenarios included a notional plume release and point releases mimicking the field study under three rainfall conditions. The rainfall scenarios included a 6-hour natural rainfall event on Dec. 8, 2021 and two design storms (2-year and 100-year events). The observed spore concentrations from asphalt and concrete from the actual field experiment were applied to calibrate the washoff parameters in the SWMM model, using an exponential washoff function. The calibrated washoff coefficient (c 1 ) and exponent (c 2 ) were 0.01 and 1.00 for asphalt, 0.05 and 1.45 for grass, and 2.45 and 1.00 for concrete, respectively. The calibrated SWMM model simulated spore concentrations in runoff at times and magnitudes similar to the field study data. In the point release modeled scenario, the concrete surface generated 55.6% higher average spore concentrations than asphalt. Similarly, in the field experiment, a 175% (p < 0.05) higher average spore concentration in surface runoff was observed from concrete than from asphalt. Here, this study demonstrates how SWMM may be used to evaluate spore washoff from urban surfaces under different precipitation amounts, intensities, and durations, and how visualized spatial migration pathways in stormwater runoff may be used for emergency planning and remediation.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Reaction–Diffusion Coupling Facilitates the Sequential Precipitation of Metal Ions from Battery Feedstock Solutions

Here, the development of new technologies for chemical separations is urgently needed to meet the surging demand for critical materials that has strained resources and caused environmental challenges. Inspired by the classic Liesegang experiment, we demonstrated the separation of critical metal ions based on the coupling of ion diffusion and precipitation kinetics. For this purpose, a model feedstock solution simulating dissolved battery electrodes was placed on top of a hydrogel loaded with a precipitating agent, namely sodium hydroxide. As the lithium, manganese, cobalt, and nickel ions diffused into the gel, a gradient of precipitates formed along the length of the reactor. Elemental analysis of the spatially distributed precipitates showed the enrichment of nickel near the gel-solution interface, followed by the formation of an almost pure (>96%) manganese product further along the reactor. Optimization experiments revealed that a sodium hydroxide concentration of 10 mM and a gel/solution volume ratio of 2:1 favored efficient separations. The robustness of the method was demonstrated in four out of five feedstock compositions of typically used battery cathodes. Our proof-of-concept experiments present a paradigm for critical materials separations that does not require specialty chemicals, binding agents, membranes, or toxic solvents.

25 ENERGY STORAGE↗
DOE OSTI2021

A Modular and Transferable Reinforcement Learning Framework for the Fleet Rebalancing Problem

Mobility on demand (MoD) systems show great promise in realizing flexible and efficient urban transportation. However, significant technical challenges arise from operational decision making associated with MoD vehicle dispatch and fleet rebalancing. For this reason, operators tend to employ simplified algorithms that have been demonstrated to work well in a particular setting. To help bridge the gap between novel and existing methods, we propose a modular framework for fleet rebalancing based on model-free reinforcement learning (RL) that can leverage an existing dispatch method to minimize system cost. In particular, by treating dispatch as part of the environment dynamics, a centralized agent can learn to intermittently direct the dispatcher to reposition free vehicles and mitigate against fleet imbalance. We formulate RL state and action spaces as distributions over a grid partitioning of the operating area, making the framework scalable and avoiding the complexities associated with multiagent RL. Numerical experiments, using real-world trip and network data, demonstrate that RL reduces waiting time by 28% to 38% for the same-day evaluation, 17% to 44% for cross-day evaluation, and 22% to 25% for cross-season evaluation compared with no rebalancing scenarios. This approach has several distinct advantages over baseline methods including: improved system cost; high degree of adaptability to the selected dispatch method; and the ability to perform scale-invariant transfer learning between problem instances with similar vehicle and request distributions.

33 ADVANCED PROPULSION SYSTEMS↗
DOE OSTI2022

Improved Particle Swarm Algorithm Using Rubik’s Cube Topology for Bilevel Building Energy Transaction

Following the rapid growth of distributed energy resources (e.g., renewables, battery), localized peer-to-peer energy transactions are receiving more attention for multiple benefits, such as reducing power loss and stabilizing the main power grid. To promote distributed renewables locally, the local trading price is usually set to be within the external energy purchasing and selling price range. Consequently, building prosumers are motivated to trade energy through a local transaction center. This local energy transaction is modeled in bilevel optimization game. A selfish upper level agent is assumed with the privilege to set the internal energy transaction price with an objective of maximizing its arbitrage profit. Meanwhile, the building prosumers at the lower level will response to this transaction price and make decisions on electricity transaction amount. Therefore, this non-cooperative leader-follower trading game is seeking for equilibrium solutions on the energy transaction amount and prices. Additionally, a uniform local transaction price structure (purchase price equals selling price) is considered here. Aiming at reducing the computational burden from classical Karush–Kuhn–Tucker (KKT) transformation and protecting the private information of each stakeholder (e.g., building), swarm intelligence-based solution approach is employed for upper level agent to generate trading price and coordinate the transactive operations. On one hand, to decrease the chance of premature convergence in global-best topology, Rubik’s Cube topology is proposed in this study based on further improvement of a two-dimensional square lattice model (i.e., one local-best topology-Von Neumann topology). Rotating operation of the cube is introduced to dynamically changing the neighborhood and enhancing information flow at the later searching state. Several groups of experiments are designed to evaluate the performance of proposed Rubik’s Cube topology-based particle swarm algorithm. The results have validated the effectiveness of proposed topology and operators comparing with global-best version PSO and Von Neumann topology-based PSO and its scalability on larger scale applications.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2026

Explainable physics-based constraints on reinforcement learning for accelerator optimization

We present a reinforcement learning (RL) framework for optimizing particle accelerator experiments that builds explainable physics-based constraints on agent behavior. The goal is to increase transparency and trust by letting users verify that the agent’s decision-making process incorporates suitable physics. Our algorithm uses a learnable surrogate function for physical observables, such as energy, and uses them to fine-tune how actions are chosen. This surrogate can be represented by a neural network or by an interpretable sparse dictionary model. We test our algorithm on a range of particle accelerator optimization environments designed to emulate the Continuous Electron Beam Accelerator Facility at Jefferson Lab. By examining the mathematical form of the learned constraint function, we are able to confirm the agent has learned to use the established physics of each environment. In addition, we find that the introduction of a physics-based surrogate enables our RL algorithms to reliably converge for difficult high-dimensional accelerator optimization environments.

explainability↗
DOE OSTI2022

Deep Reinforcement Learning From Demonstrations to Assist Service Restoration in Islanded Microgrids

Microgrids can be operated in island mode during utility grid outages to support service restoration and improve system resilience. To schedule and dispatch distributed energy resources (DERs) in an islanded microgrid, conventional model-based methods rely on accurate distribution network models and lack generalization and adaptability. Data-driven methods are promising for DER coordination but face practical challenges such as potential hazards to microgrids during online training and insufficient online training opportunities due to low outage rates. This paper presents a novel two-stage learning framework that builds on the deep deterministic policy gradient from demonstrations to identify an optimal restoration strategy. At the pre-training stage, imitation learning is applied to equip the control agent with expert experiences to guarantee acceptable initial performance. At the online training stage, action clipping, reward shaping, and expert demonstrations are leveraged to ensure safe exploration while accelerating the training process. In conclusion, the proposed method is illustrated using the IEEE 123-node system and compared with a representative model-based method and the standard deep deterministic policy gradient method to prove solution accuracy and demonstrate increased computational efficiency.

Du, Yan↗
DOE OSTI2023

Understanding racial and ethnic disparities in COVID-19 outcomes using an agent-based simulation [Poster]

Infectious disease outbreaks pose a major threat to public health, economic stability, and human security. In order to mitigate disease impacts, we need to understand drivers that contribute to its spread. Latinx, Black, and American Indian racial and ethnic groups experienced disproportionate health outcomes during the COVID-19 pandemic. Current modeling approaches often fail to capture the high level of heterogeneity present in the population needed to measure the drivers of these disparities.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Optimizing Patient-Specific Medication Regimen Policies Using Wearable Sensors in Parkinson’s Disease

Effective treatment of Parkinson’s disease (PD) is a continual challenge for healthcare providers, and providers can benefit from leveraging emerging technologies to supplement traditional clinic care. We develop a data-driven reinforcement learning (RL) framework to optimize PD medication regimens through wearable sensors. We leverage a data set of n = 26 PD patients who wore wrist-mounted movement trackers for two separate six-day periods. Using these data, we first build and validate a simulation model of how individual patients’ movement symptoms respond to medication administration. We then pair this simulation model with an on-policy RL algorithm that recommends optimal medication types, timing, and dosages during the day while incorporating human-in-the-loop considerations on medication administration. The results show that the RL-prescribed medication regimens outperform physicians’ medication regimens, despite physicians having access to the same data as the RL agent. To validate our results, we assess our wearable-based RL medication regimens using n = 399 PD patients from the Parkinson’s Progression Markers Initiative data set. We show that the wearable-based RL medication regimens would lead to significant symptom improvement for these patients, even more so than training RL policies directly from this data set. In doing so, we show that RL models from even small data sets of wearable data can offer novel, generalizable clinical insights and medication strategies, which may outperform those derived from larger data sets without wearable data.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2023

A Simplified Approach Based on Cellular Automata for Describing Direct Reduced Iron Production in Different Reducing Conditions

A quick computation approach based on cellular automata is developed and implemented to describe the reduction of iron ore pellets by a mixture of reducing agents featured by different H 2 /CO ratios. The evolution of oxygen concentration inside the pellet is followed from the beginning to the end of contact between reducing agent and pellet. The variation of thermal state of pellets and gas mixture is computed based on their initial temperature, considering the heat involved and the convective heat exchange between pellet and gas mixture. The use of cellular automata and finite‐difference method to solve the diffusion equation point out the absence of any diffusion coefficient value, allowing to make the model fit the experimental trial, because the problem is that it is not ruled just by diffusion but also by the concentration variation of reducing agent inside the pellet due to porosity increasing during reduction. The updating of the reducing agents concentration implies a sharp decrease of oxygen concentration that the cellular automata model considers. The developed model is able to provide the in‐line control of reduction process and could be used to adjust the chemical concentration and temperature of injected reducing agents.

Metallurgy & Metallurgical Engineering↗
DOE OSTI2026

SetGo: Metadata Readiness for Scientific AI Datasets

Scientific datasets intended for AI use require both computational readiness for model training and metadata readiness for discovery, sharing, and reuse. The Readiness Engine for Data Integration (REDI) addresses computational readiness, but no corresponding tool evaluates whether a dataset’s metadata are sufficiently complete, governed, and standards-compliant for publication and agent-based consumption. Existing FAIR assessors operate only on published repository records, and no single system covers FAIR compliance, licensing, provenance, governance, reproducibility, and catalog readiness together. We present SetGo, an open-source Python toolkit that assesses and repairs metadata readiness across these six dimensions before a dataset is published or archived. Applied to four scientific corpora, SetGo surfaces deficiencies that general-purpose tools do not detect: ERA5 climate metadata scores 4% on ACDD 1.3 compliance; materials datasets fail OPTIMADE species-definition requirements; and PDB-derived proteomics data carries licensing terms incompatible with standard SPDX identifiers. Guided enrichment raises overall FAIR scores from 52–57% to 81–91%, and a single setgo publish command pushes to Hugging Face Hub, CKAN, or OpenMetadata with ML Commons Croissant 1.0 metadata sidecars. To support interactive and automated workflows, SetGo integrates with coding agents powered by large language models (LLMs) through a /setgo skill that enables natural-language execution of the full assess–enrich–publish loop, with user involvement limited to supplying missing metadata values.

Wilkinson, Sean [ORNL] (ORCID:0000000214437479)↗
DOE OSTI2020

Integrated Dynamic Facade Control with an Agent-based Architecture for Commercial Buildings

Dynamic façades have significant technical potential to minimize heating, cooling, and lighting energy use and peak electric demand in the perimeter zone of commercial buildings, but the performance of these systems is reliant on being able to balance complex trade-offs between solar control, daylight admission, comfort, and view over the life of the installation. As the context for controllable energy-efficiency technologies grows more complex with the increased use of intermittent renewable energy resources on the grid, it has become increasingly important to look ahead towards more advanced approaches to integrated systems control in order to achieve optimum life-cycle performance at a lower cost. This study examines the feasibility of a model predictive control system for low-cost autonomous dynamic façades. A system architecture designed around lightweight, simple agents is proposed. The architecture accommodates whole building and grid level demands through its modular, hierarchical approach. Automatically-generated models for computing window heat gains, daylight illuminance, and discomfort glare are described. The open source Modelica and JModelica software tools were used to determine the optimum state of control given inputs of window heat gains and lighting loads for a 24-hour optimization horizon. Penalty functions for glare and view/ daylight quality were implemented as constraints. The control system was tested on a low-power controller (1.4 GHz single core with 2 GB of RAM) to evaluate feasibility. The target platform is a low-cost ($35/unit) embedded controller with 1.2 GHz dual-core cpu and 1 GB of RAM. Configuration and commissioning of the curtainwall unit was designed to be largely plug and play with minimal inputs required by the manufacturer through a web-based user interface. An example application was used to demonstrate optimal control of a three-zone electrochromic window for a south-facing zone. The overall approach was deemed to be promising. Further engineering is required to enable scalable, turnkey solutions.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2020

Multi-Agent Safe Policy Learning for Power Management of Networked Microgrids

This paper presents a supervised multi-agent safe policy learning (SMAS-PL) method for optimal power management of networked microgrids (MGs) in distribution systems. While unconstrained reinforcement learning (RL) algorithms are black-box decision models that could fail to satisfy grid operational constraints, our proposed method considers AC power flow equations and other operational limits. Accordingly, the training process employs the gradient information of operational constraints to ensure that the optimal control policy functions generate safe and feasible decisions. Furthermore, we have developed a distributed consensus-based optimization approach to train the agents’ policy functions while maintaining MGs’ privacy and data ownership boundaries. After training, the learned optimal policy functions can be safely used by the MGs to dispatch their local resources, without the need to solve a complex optimization problem from scratch. Lastly, numerical experiments have been devised to verify the performance of the proposed method.

, policy gradient↗
DOE OSTI2021

Unbundling Smart Meter Services Through Spatio-Temporal Decomposition Agents in DER-rich Environment

Smart meters and the advanced metering infrastructure (AMI) facilitate distribution system operators (DSOs) to gather information on energy consumption at the customer level. With the increasing penetration of building-level intermittent distributed energy resources (DERs) behind the meter, DER information is not available to DSOs. At the same time, smart meter enables users to participate in grid, with real-time information. Information for behind the meter is needed by user to coordinate building level assets for maximum benefits. The concept of unbundled smart meter (USM) needs agents to decompose smart meter measurements to provide service to DSO as well as customers. In this paper, we propose a Spatio-Temporal Decomposition Agent (STDA) for USM based on Artificial Intelligence (AI). STDA can help users optimize their energy usage, help DSO to utilize building assets for the grid operation. The energy usage strategy developed by STDA is suitable for different users, and can be customized by deep learning (DL) models according to the different energy consumption habits of each user. The power prediction performance results of various DL models and evaluation using a set of data from a Hawaii utility is presented. Furthermore, STDA integration with Home Energy management Systems (HEMS) to manage resources is presented and validated. STDA pre-processes the measurements before model training, and provides the spatio-temporal decomposed forecasting.

42 ENGINEERING↗