SEARCH · Engineering Papers

Results for “interacting agent systems”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 199 records · Page 11

All sources NASA NTRS DOE OSTI

NASA NTRS1994

Learning procedures from interactive natural language instructions

Despite its ubiquity in human learning, very little work has been done in artificial intelligence on agents that learn from interactive natural language instructions. In this paper, the problem of learning procedures from interactive, situated instruction is examined in which the student is attempting to perform tasks within the instructional domain, and asks for instruction when it is needed. Presented is Instructo-Soar, a system that behaves and learns in response to interactive natural language instructions. Instructo-Soar learns completely new procedures from sequences of instruction, and also learns how to extend its knowledge of previously known procedures to new situations. These learning tasks require both inductive and analytic learning. Instructo-Soar exhibits a multiple execution learning process in which initial learning has a rote, episodic flavor, and later executions allow the initially learned knowledge to be generalized properly.

Huffman, Scott B.↗

NASA NTRS1994

Instructable autonomous agents

In contrast to current intelligent systems, which must be laboriously programmed for each task they are meant to perform, instructable agents can be taught new tasks and associated knowledge. This thesis presents a general theory of learning from tutorial instruction and its use to produce an instructable agent. Tutorial instruction is a particularly powerful form of instruction, because it allows the instructor to communicate whatever kind of knowledge a student needs at whatever point it is needed. To exploit this broad flexibility, however, a tutorable agent must support a full range of interaction with its instructor to learn a full range of knowledge. Thus, unlike most machine learning tasks, which target deep learning of a single kind of knowledge from a single kind of input, tutorability requires a breadth of learning from a broad range of instructional interactions. The theory of learning from tutorial instruction presented here has two parts. First, a computational model of an intelligent agent, the problem space computational model, indicates the types of knowledge that determine an agent's performance, and thus, that should be acquirable via instruction. Second, a learning technique, called situated explanation specifies how the agent learns general knowledge from instruction. The theory is embodied by an implemented agent, Instructo-Soar, built within the Soar architecture. Instructo-Soar is able to learn hierarchies of completely new tasks, to extend task knowledge to apply in new situations, and in fact to acquire every type of knowledge it uses during task performance - control knowledge, knowledge of operators' effects, state inferences, etc. - from interactive natural language instructions. This variety of learning occurs by applying the situated explanation technique to a variety of instructional interactions involving a variety of types of instructions (commands, statements, conditionals, etc.). By taking seriously the requirements of flexible tutorial instruction, Instructo-Soar demonstrates a breadth of interaction and learning capabilities that goes beyond previous instructable systems, such as learning apprentice systems. Instructo-Soar's techniques could form the basis for future 'instructable technologies' that come equipped with basic capabilities, and can be taught by novice users to perform any number of desired tasks.

Huffman, Scott Bradley↗

NASA NTRS2025

Localization of Ad-Hoc Lunar Constellations in Communication Failure Modes for Distributed Spacecraft Autonomy

As Lunar missions increase in complexity, inspired by NASA’s Artemis Program, they will require reliable and sufficient Position, Navigation, and Timing (PNT) capability to support the upcoming Lunar users. The navigation service should also be compatible with the smaller platforms, like CubeSats, being sent by the public and private sectors. A non-dedicated, ad-hoc Lunar navigation constellation can provide PNT services on-demand using the non-dedicated swarm assets. Swarm members cooperatively and autonomously localize themselves with minimal interaction from Earth, freeing up valuable bandwidth and ground segment resources. The autonomous localization of Lunar constellations utilizes neighbor two-way intersatellite link (ISL) measurements in a distributed extended Kalman filter (DEKF) system to minimize operating costs. Because the decentralized Lunar PNT system relies on relay communication amongst the agents, network failures or loss of assets among ad-hoc Lunar constellations may impact localization performance. This study presents an evaluation of localization performance under increasing levels of network degradation. A simulation of an ad-hoc Lunar PNT swarm is augmented to include system faults and the impacts of intermittent and permanent failures on localization performance are evaluated. We investigate three potential causes of network degradation: single spacecraft loss, multiple spacecraft loss, and antenna failure. The numerical assessments from the simulation show that the LPNT system under study, based on an autonomous decentralized concept of operation, is highly robust and resilient to communication failures. Minor faults, such as single spacecraft loss, solar interference, technical malfunctions, message delays, and antenna outages, have minimal impact on state estimation, with only a 4.47% and 3.75% degradation in median position error for assets and a representative ground user, respectively, compared to an ideal communication scenario. However, major faults, such as hardware failures or meteor strikes leading to the loss of multiple spacecrafts, are more concerning. The permanent loss of three spacecraft results in a more severe performance degradation, with median position error increasing by 23.3% for assets and 11.7% for a representative ground user, despite the Lunar PNT system remaining functional.

Yeji Kim↗

DOE OSTI2023

Agent-based simulation and child protection systems: Rationale, implementation, and verification

Simulation models are an important tool used in health care and other disciplines to support operational research and decision-making. In the child protection literature, simulation models are an under-utilized source of research evidence. Here, in this paper, we describe the rationale for and the development of an agent-based simulation of a child protection system in the US. Using the investigation, prevention service, and placement histories of 600,000 children served in an urban child welfare system, we walk the reader through the development of a prototype known as OSPEDALE. The governing equations built into OSPEDALE probabilistically simulate the onset of investigations. Then, drawing from empirical survival distributions, the governing equations trace the probability of subsequent interactions with the system (recurrence of maltreatment, service referrals, and placement) conditional on the characteristics of children, their assessed risk level, and prior child protection system involvement. As an initial test of OSPEDALE's utility, we compare empirical admission counts with counts generated from OSPEDALE. Though the verification step is admittedly simple, the comparison shows that OSPEDALE replicates the empirical count of new admissions closely enough to justify further investment in OSPEDALE. Management of public child protection systems is increasingly research evidence-dependent. The emphasis on research evidence as a decision-support tool has elevated evidence acquired through randomized clinical trials. Though important, the evidence from clinical trials represents only one type of research evidence. Properly specified, simulation models are another source of evidence with real-world relevance.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2022

Enhanced Oblique Decision Tree Enabled Policy Extraction for Deep Reinforcement Learning in Power System Emergency Control

Deep reinforcement learning (DRL) algorithms have successfully solved many challenging problems in various power system control scenarios. However, their decision-making process is usually regarded as black-boxes. Furthermore, how DRL models interact with human intelligence remains an open problem. Thus, this paper proposes a policy extraction framework to extract a complex DRL model into an explainable policy. This framework includes three parts: 1) DRL training and data generation. We train an agent for a specific control task and generate data, which contains the control policy of the agent. 2) Policy extraction. We propose an information gain rate based weighted oblique decision tree (IGR-WODT) for DRL policy extraction. 3) Policy evaluation. We define three metrics to evaluate the performance of the proposed approach. A case study for the under-voltage load shedding problem shows that the IGR-WODT presents a performance enhancement compared with DRL, weighted oblique decision tree, and univariate decision tree. The proposed policy extraction method could provide an intuitive explanation of the neural network decision-making process to the dispatchers when making final decisions on power grid operation. Also, the resulted rule-based controller could replace the deep neural network-based controller in many field edge devices with limited computing resources, providing comparable performance.

deep reinforcement learning↗

DOE OSTI2022

Diffuse Radiation Forcing Constraints on Gross Primary Productivity and Global Terrestrial Evapotranspiration

Abstract The diffuse radiation fertilization effect—the increase in plant productivity in the presence of higher diffuse radiation ( K ↓,d )—is an important yet understudied aspect of atmosphere‐biosphere interactions and can modify the terrestrial carbon, energy, and water budgets. The K ↓,d fertilization effect links the carbon cycle with clouds and aerosols, all of which are large sources of uncertainties for our current understanding of the Earth system and for future climate projections. Here we establish to what extent observational and modeling uncertainty in sunlight's diffuse fraction ( k d ) affects simulated gross primary productivity (GPP) and terrestrial evapotranspiration ( λE ). We find only 48 eddy covariance sites with simultaneous sufficient measurements of K ↓,d with none in the tropical climate zone, making it difficult to constrain this mechanism globally using observations. Using a land modeling framework based on the latest version of the Community Land Model, we find that global GPP ranges from 114 Pg C year −1 when using k d forcing from the Modern‐Era Retrospective analysis for Research and Applications, version 2 reanalysis to a ∼7% higher value of 122 Pg C year −1 when using the Clouds and the Earth's Radiant Energy System satellite product, with especially strong differences apparent over the tropical region (mean increase ∼9%). The differences in λE , although smaller (−0.4%) due to competing changes in shaded and sunlit leaf transpiration, can be greater than regional impacts of individual forcing agents like aerosols. Our results demonstrate the importance of comprehensively and systematically validating the simulated k d by atmosphere modules as well as the response to differences in k d within land modules across Earth System Models.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2025

Development of a microwave-assisted downdraft moving-bed gasifier for continuous processing of lignite and biomass chars

This research illustrates a microwave-assisted downdraft moving-bed gasifier for the first time. Such design enables continuous solid gasification process. An adjustable auger was applied to control the solid removal rate and the gas-solid interaction time. Both lignite and biomass chars were investigated to determine the capability of the current system for low-tar feedstocks with different densities. Here, the presented reactor design was able to operate continuously for 3 hours and 20 minutes under 700 ℃ and atmospheric pressure, with air as the gasifying agent. For yellow pine char, the processing rate could reach 34.1 grams per hour with decent syngas production. The downdraft moving-bed design shows better cold gas and syngas production efficiencies compared to the common fixed-bed design, due to controllable residence time and more homogeneous microwave heating. The limitations of the current design and the direction of novel microwave-assisted chemical reactor design were discussed. This novel reactor design provides a way to improve the efficiency of microwave-assisted gasification process and shows its potential to be incorporated into other established chemical reaction processes as a modular add-on.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2024

Identifying Molecular Scale Interactions of Nucleating Agents with Salt Hydrates for Thermal Energy Storage Applications

Latent heat based thermal energy storage is of interest as a method to mitigate and time-shift thermal load peaks and hence reduce energy demands for heating and cooling buildings. Notably, Glauber’s salt (GS) presents a good choice as a phase change material (PCM) for building applications because it has a melting point near room temperature (i.e., 32.4 °C), a large heat of fusion and energy density, and is low cost (~ $100/ton). However, there are several known limitations to using GS as a PCM, such as incongruent melting, high degree of supercooling, and formation of other hydration states, which renders GS unsuitable for practical use. Inorganic crystals acting as nucleating agents can avoid some of these issues with GS [1]. Understanding nucleation behavior through typical calorimetry experiments offers limited insight into the molecular-scale mechanisms. Here, we report isothermal titration calorimetry [2] to investigate the interactions between sodium sulfate salt and an organic nucleating agent - glycine. Our experiments provide qualitative and quantitative understanding of how glycine interacts with sodium sulfate across various temperatures. The findings offer insights into the nucleation onset temperature (24°C) and underscore the pivotal role of glycine's zwitterionic structure in facilitating nucleation. This preliminary work provides a foundation for subsequent exploration of nucleating agents that would be incorporated in different salt hydrate systems. 1. Purohit, B. K., and V. S. Sistla. "Inorganic salt hydrate for thermal energy storage application: A review." Energy Storage 3, no. 2 (2021): e212. 2. Velázquez-Campoy, Adrián, Hiroyasu Ohtaka, Azin Nezami, Salman Muzammil, and Ernesto Freire. "Isothermal titration calorimetry." Current protocols in cell biology 23, no. 1 (2004): 17-8.

calorimetry, nucleation & growth↗

NASA NTRSDate not supplied

Uncertainty Quantification using Deep Ensembles for Decision Making in Cyber-Physical-Human Systems

In this paper and its companion, Differential Equation Approximation Using Gradient-Boosted Quantile Regression, Robison et al., we examine an approach to quantifying model uncertainty with the aim of increasing the trustworthiness of computational models in human-machine interactions. In Differential Equation Approximation Using Gradient-Boosted Quantile Regression, we focus on gradient-boosted decision trees, while in this one, we give more details about deep ensembles. Uncertainty quantification is crucial for building trustworthy autonomous decision-making agents in human-machine teams. There are two types of uncertainties: aleatoric and epistemic. The former is related to the inherent stochasticity (noise) of the process, whereas the latter is associated with the lack of knowledge or representation capability of models, such as neural networks. By lack of knowledge, we mean the model’s inability to accurately predict outputs for all possible inputs. The aleatory uncertainty can be estimated fairly easily with, for example, filters, whereas epistemic uncertainty is challenging to compute. This paper uses deep ensembles to quantify both aleatory and epistemic uncertainty. It can act as an uncertainty-aware surrogate transition model for decision-making frameworks. "Uncertainty-aware" means that the surrogate transition model should make predictions along with confidence in those predictions. In the context of decision-making, the transition models are ordinary differential equations (ODEs). Since ODEs can be simulated to make one-step or multi-step predictions, a good surrogate model for them should perform reasonably well in both modes. In a multi-step approach, the trajectory sampling method TS∞ was used to propagate uncertainty over multiple steps. The cartpole dynamical system was selected to demonstrate the ability of deep ensembles as good surrogate transition models for decision-making frameworks. The deep ensembles modeled the dynamics of cartpole ODEs and made uncertainty-aware predictions in single-step and multi-step transition modes.

CPH systems↗

DOE OSTI2023

The interaction of wholesale electricity market structures under futures with decarbonization policy goals: A complexity conundrum

Competitive wholesale electricity markets can help facilitate energy system decarbonization by incentivizing investments in clean energy technologies that meet evolving system needs. We explore market structure impacts on generator operations and deployment by risk-averse, heterogeneous investor firms using the Electricity Markets and Investment Suite - Agent-based Simulation (EMIS-AS) model. Here we apply clean energy targets of 45%-100% by 2035 considering energy, ancillary services, capacity, and clean energy credit products and pricing and eligibility rules. Results highlight a complexity conundrum, whereby finding the "right" market design to achieve decarbonization goals and avoid unintended consequences can be a highly-nuanced, non-incremental challenge. Carefully designed energy-only markets can achieve the same clean energy targets as capacity market structures but with different revenue and profitability outcomes. Operating reserve demand curve-based scarcity pricing can substitute capacity markets for similar deployment outcomes. Carbon pricing alone is most effective at achieving decarbonization levels at low clean energy targets, and clean energy credit markets and carbon pricing are substitutionary at high clean energy targets. Restricting technology participation in capacity and operating reserve markets can impact deployment and operations, even for nonrestricted technologies. Adding an inertia product with fast frequency response yields insufficient provision at high clean energy targets, but work is needed to understand frequency requirements and capabilities.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

NASA NTRS2006

Addressing the Tension Between Strong Perimeter Control an Usability

This paper describes a strong perimeter control system for a general purpose processing system, with the perimeter control system taking significant steps to address usability issues, thus mitigating the tension between strong perimeter protection and usability. A secure front end enforces two-factor authentication for all interactive access to an enclave that contains a large supercomputer and various associated systems, with each requiring their own authentication. Usability is addressed through a design in which the user has to perform two-factor authentication at the secure front end in order to gain access to the enclave, while an agent transparently performs public key authentication as needed to authenticate to specific systems within the enclave. The paper then describes a proxy system that allows users to transfer files into the enclave under script control, when the user is not present to perform two-factor authentication. This uses a pre-authorization approach based on public key technology, which is still strongly tied to both two-factor authentication and strict control over where files can be transferred on the target system. Finally the paper describes an approach to support network applications and systems such as grids or parallel file transfer protocols that require the use of many ports through the perimeter. The paper describes a least privilege approach that dynamically opens ports on a host-specific, if-authorized, as-needed, just-in-time basis.

Hinke, Thomas H.↗

DOE OSTI2020

Role of Protein-Water Interface in the Stacking Interactions of Granum Thylakoid Membranes—As Revealed by the Effects of Hofmeister Salts

The thylakoid membranes of vascular plants are differentiated into stacked granum and unstacked stroma regions. The formation of grana is triggered by the macrodomain formation of photosystem II and light-harvesting complex II (PSII-LHCII) and thus their lateral segregation from the photosystem I—light-harvesting complex I (PSI-LHCI) super-complexes and the ATP-synthase; which is then stabilized by stacking interactions of the adjacent PSII-LHCII enriched regions of the thylakoid membranes. The self-assembly and dynamics of this highly organized membrane system and the nature of forces acting between the PSII-LHCII macrodomains are not well understood. By using circular dichroism (CD) spectroscopy, small-angle neutron scattering (SANS) and transmission electron microscopy (TEM), we investigated the effects of Hofmeister salts on the organization of pigment-protein complexes and on the ultrastructure of thylakoid membranes. We found that the kosmotropic agent (NH 4 ) 2 SO 4 and the Hofmeister-neutral NaCl, up to 2 M concentrations, hardly affected the macro-organization of the protein complexes and the membrane ultrastructure. In contrast, chaotropic salts, NaClO 4 , and NaSCN destroyed the mesoscopic structures, the multilamellar organization of the thylakoid membranes and the chiral macrodomains of the protein complexes but without noticeably affecting the short-range, pigment-pigment excitonic interactions. Comparison of the concentration- and time-dependences of SANS, TEM and CD parameters revealed the main steps of the disassembly of grana in the presence of chaotropes. It begins with a rapid diminishment of the long-range periodic order of the grana membranes, apparently due to an increased stacking disorder of the thylakoid membranes, as reflected by SANS experiments. SANS measurements also allowed discrimination between the cationic and anionic effects—in stacking and disorder, respectively. This step is followed by a somewhat slower disorganization of the TEM ultrastructure, due to the gradual loss of stacked membrane pairs. Occurring last is the stepwise decrease and disappearance of the long-range chiral order of the protein complexes, the rate of which was faster in LHCII-deficient membranes. These data are interpreted in terms of a theory, from our laboratory, according to which Hofmeister salts primarily affect the hydrophylic-hydrophobic interactions of proteins, and the stroma-exposed regions of the intrinsic membrane proteins, in particular—pointing to the role of protein-water interface in the stacking interactions of granum thylakoid membranes.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2024

Computational and Systems Biology Advances to Enable Bioagent Agnostic Signatures

Enumerated threat agent lists have long driven biodefense priorities. The global SARS-CoV-2 pandemic demonstrated the limitations of searching for known threat agents as compared to a more agnostic approach. Recent technological advances are enabling agent-agnostic biodefense, especially through the integration of multi-modal observations of host-pathogen interactions directed by a human immunological model. Although well-developed technical assays exist for many aspects of human-pathogen interaction, the analytic methods and pipelines to combine and holistically interpret the results of such assays are immature and require further investments to exploit new technologies. In this manuscript, we discuss potential immunologically based bioagent-agnostic approaches and the computational tool gaps the community should prioritize filling.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2021

Dependency of active pressure and equation of state on stiffness of wall

Autonomous motion and motility are hallmarks of active matter. Active agents, such as biological cells and synthetic colloidal particles, consume internal energy or extract energy from the environment to generate self-propulsion and locomotion. These systems are persistently out of equilibrium due to continuous energy consumption. It is known that pressure is not always a state function for generic active matter. Torque interaction between active constituents and confinement renders the pressure of the system a boundary-dependent property. The mechanical pressure of anisotropic active particles depends on their microscopic interactions with a solid wall. Using self-propelled dumbbells confined by solid walls as a model system, we perform numerical simulations to explore how variations in the wall stiffness influence the mechanical pressure of dry active matter. In contrast to previous findings, we find that mechanical pressure can be independent of the interaction of anisotropic active particles with walls, even in the presence of intrinsic torque interaction. Particularly, the dependency of pressure on the wall stiffness vanishes when the stiffness is above a critical level. In such a limit, the dynamics of dumbbells near the walls are randomized due to the large torque experienced by the dumbbells, leading to the recovery of pressure as a state variable of density.

42 ENGINEERING↗

NASA NTRS2005

Workshop on Radical Agent Concepts (WRAC)

Agent technology, along with autonomous and autonomic computing, has emerged as a major field in computing, and will greatly influence the future development of complex computer-based systems. The area of research is strongly influenced by the Autonomic Computing initiative as well as by developments in biologically-inspired computing, and involves interdisciplinary interaction from those involved in research in social intelligence, psychology, arts, biology, computer science, computer communications and philosophy. This volume includes revised versions of papers presented at the workshop. The workshop was structured so as to allow adequate time for discussion and interaction, to exchange ideas and reflect on the motivations, scientific grounds and practical consequences of the concepts presented. Many of the ideas are truly "radical", and so authors were given time to revise their papers to reflect further thoughts on the ideas presented, and to reflect feedback received at the workshop.

Hinchey, Michael G.↗

DOE OSTI2023

Drug discovery efforts at George Mason University

With over 39,000 students, and research expenditures in excess of $200 million, George Mason University (GMU) is the largest R1 (Carnegie Classification of very high research activity) university in Virginia. Mason scientists have been involved in the discovery and development of novel diagnostics and therapeutics in areas as diverse as infectious diseases and cancer. Below are highlights of the efforts being led by Mason researchers in the drug discovery arena. To enable targeted cellular delivery, and non-biomedical applications, Veneziano and colleagues have developed a synthesis strategy that enables the design of self-assembling DNA nanoparticles (DNA origami) with prescribed shape and size in the 10 to 100 nm range. The nanoparticles can be loaded with molecules of interest such as drugs, proteins and peptides, and are a promising new addition to the drug delivery platforms currently in use. The investigators also recently used the DNA origami nanoparticles to fine tune the spatial presentation of immunogens to study the impact on B cell activation. These studies are an important step towards the rational design of vaccines for a variety of infectious agents. To elucidate the parameters for optimizing the delivery efficiency of lipid nanoparticles (LNPs), Buschmann, Paige and colleagues have devised methods for predicting and experimentally validating the pKa of LNPs based on the structure of the ionizable lipids used to formulate the LNPs. These studies may pave the way for the development of new LNP delivery vehicles that have reduced systemic distribution and improved endosomal release of their cargo post administration. To better understand protein-protein interactions and identify potential drug targets that disrupt such interactions, Luchini and colleagues have developed a methodology that identifies contact points between proteins using small molecule dyes. The dye molecules noncovalently bind to the accessible surfaces of a protein complex with very high affinity, but are excluded from contact regions. When the complex is denatured and digested with trypsin, the exposed regions covered by the dye do not get cleaved by the enzyme, whereas the contact points are digested. The resulting fragments can then be identified using mass spectrometry. The data generated can serve as the basis for designing small molecules and peptides that can disrupt the formation of protein complexes involved in disease processes. For example, using peptides based on the interleukin 1 receptor accessory protein (IL-1RAcP), Luchini, Liotta, Paige and colleagues disrupted the formation of IL-1/IL-R/IL-1RAcP complex and demonstrated that the inhibition of complex formation reduced the inflammatory response to IL-1B. Working on the discovery of novel antimicrobial agents, Bishop, van Hoek and colleagues have discovered a number of antimicrobial peptides from reptiles and other species. DRGN-1, is a synthetic peptide based on a histone H1-derived peptide that they had identified from Komodo Dragon plasma. DRGN-1 was shown to disrupt bacterial biofilms and promote wound healing in an animal model. The peptide, along with others, is being developed and tested in preclinical studies. Other research by van Hoek and colleagues focuses on in silico antimicrobial peptide discovery, screening of small molecules for antibacterial properties, as well as assessment of diffusible signal factors (DFS) as future therapeutics. The above examples provide insight into the cutting-edge studies undertaken by GMU scientists to develop novel methodologies and platform technologies important to drug discovery.

59 BASIC BIOLOGICAL SCIENCES↗

NASA NTRS2005

Towards the Verification of Human-Robot Teams

Human-Agent collaboration is increasingly important. Not only do high-profile activities such as NASA missions to Mars intend to employ such teams, but our everyday activities involving interaction with computational devices falls into this category. In many of these scenarios, we are expected to trust that the agents will do what we expect and that the agents and humans will work together as expected. But how can we be sure? In this paper, we bring together previous work on the verification of multi-agent systems with work on the modelling of human-agent teamwork. Specifically, we target human-robot teamwork. This paper provides an outline of the way we are using formal verification techniques in order to analyse such collaborative activities. A particular application is the analysis of human-robot teams intended for use in future space exploration.

Fisher, Michael↗

DOE OSTI2021

A Novel Multiagent Resource Sharing Algorithm for Control of Advanced Energy Systems

This paper implements a novel resource sharing control strategy on a fuel cell–gas turbine hybrid power system at the National Energy Technology Laboratory’s Hybrid Performance Facility (Hyper). In a fuel cell–gas turbine hybrid power system, the simultaneous interaction of the gas turbine and the fuel cell creates a tightly coupled environment characterized by conflicting dynamics. In this paper, a model-free control approach is applied to solve the tightly coupled control problem posed by this challenging environment. Specifically, this control problem is presented as a resource sharing problem that can be solved using a resource sharing algorithm that is defined based on the distribution construction concept. This algorithm creates computational agents and solves the problem through the distribution and redistribution of shared resources defined as blocks. Furthermore, two agents were created; the first agent (agent 1) controls the gas turbine speed by adjusting the electric load, and the second agent (agent 2) controls the cathode mass flow through the fuel cell using the cold-air bypass valve. A parametric study was performed over the course of 15 experimental tests for both agents 1 and 2 through an evaluation of the responses based on setpoint changes. The algorithm was shown to have behavior comparable to a previously implemented multi-input multioutput state-space controller, which was designed through a model-based control approach. The resource sharing algorithm was able to find stable performance during run-time operations without any prior system knowledge identification on the power plant and without creating models used in traditional control strategies.

25 ENERGY STORAGE↗