Search NASA⌕ Search

SEARCH · Search NASA

Results for “Simulation Use Case”

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 829 records · Page 46

A Blockchain Case Study for Urban Air Mobility Operational Intent

To realize the potential of Urban Air Mobility (UAM), an assurance of cybersecurity is critical for public acceptance. UAM is a concept that proposes to develop short-range, point-to-point transportation systems in metropolitan areas using vertical takeoff and landing (VTOL) aircraft to overcome increasing surface congestion. The growth in the development of UAM systems, and the associated data exchange and service interactions will be at risk due to numerous types of cybersecurity attacks. The intent of this work is to leverage a permissioned blockchain to simulate secure data exchange and storage for the UAM operational intent use case. In this case study, two vehicle operators are operating in the same airspace. Their intent is to fly vehicles that land at a shared vertiport, securely.

Urban Air Mobility↗

A Blockchain Case Study for Urban Air Mobility Operational Intent

To realize the potential of Urban Air Mobility (UAM), an assurance of cybersecurity is critical for public acceptance. UAM is a concept that proposes to develop short-range, point-to-point transportation systems in metropolitan areas using vertical takeoff and landing (VTOL) aircraft to overcome increasing surface congestion. The growth in the development of UAM systems, and the associated data exchange and service interactions will be at risk due to numerous types of cybersecurity attacks. The intent of this work is to leverage a permissioned blockchain to simulate secure data exchange and storage for the UAM operational intent use case.

UAM↗

A Blockchain Case Study for Urban Air Mobility Operational Intent

To realize the potential of Urban Air Mobility (UAM), an assurance of cybersecurity is critical for public acceptance. UAM is a concept that proposes to develop short-range, point-to-point transportation systems in metropolitan areas using vertical takeoff and landing (VTOL) aircraft to overcome increasing surface congestion. The growth in the development of UAM systems, and the associated data exchange and service interactions will be at risk due to numerous types of cybersecurity attacks. The intent of this work is to leverage a permissioned blockchain to simulate secure data exchange and storage for the UAM operational intent use case.

UAM↗

A bespoke model of Arctic river basins based on hillslope delineation: Model Archive

This dataset is a model archive of the paper A bespoke model of Arctic river basins based on hillslope delineation (in prep), which introduces a watershed decomposition and parameterization method for large scale permafrost hydrology simulation. With this dataset, this study aims to address the research question: whether a computationally efficient hillslope-based modeling framework can reliably simulate discharge at Arctic river-basin scales. This dataset contains model input and output data for five modeling scenarios at a study site located in the Sagavanirktok River basin. The five modeling scenarios include three modeling cases under temperate conditions using full 3D, decomposed 3D, and decomposed 2D modeling strategies; and two modeling cases under actual Arctic conditions with permafrost using full 3D and decomposed 2D modeling strategies. Simulations were performed using the Advanced Terrestrial Simulator (ATS, v1.6 for three temperate scenarios and v1.5 for two Arctic scenarios), a physics-rich integrated surface–subsurface hydrologic model with cryo-hydrology features. For the three temperate models, simulations were conducted for the period of 10/01/1993 - 09/30/2002; and for the two Arctic models, simulations were conducted for the period of 01/01/1994 - 12/31/2002. To facilitate reproducibility of simulations, all datasets are organized hierarchically. The dataset contains: (1) Mesh files (.exo) for full 3D model, decomposed 3D models, and decomposed 2D models, located in huc/190604020802_gauge15906000/mesh/. Mesh files can be visualized through Paraview or read by Python. (2) Climate forcings (.h5) for full 3D model and decomposed 3D/2D models are located in huc/190604020802_gauge15906000/daymet_onePiece/, and huc/190604020802_gauge15906000/vp_pr_revised_daymet_1980_2006_with_wind/ separately. Accessible by Python. (3) Raw measured gage discharge (.csv) from USGS, located in huc/190604020802_gauge15906000/gaged_basin15906000_discharge_usgs/. Accessible by Python. (4) Delineated subdomain raster (.tif) and shape files (.shp), and the final parameterized results (.npy) for decomposed models, located in huc/190604020802_gauge15906000/data_preprocessed-meshing. Accessible by Python. (5) Temperate models are located in nonpermaf_huc190604020802_gauge15906000/, which includes three cases: decomposed 2D models (inside model_0*-hillslope_*), decomposed 3D models (inside model_1*-subcatchment_*), and full 3D model (inside model_2*-onepiece_*). Two step spin-up results (checkpoint_final.h5) are located in model_*1-*_spinup_steadystate and model_*2-*_spinup_cycle, separately, which are used to initialize real transient models. The input files (.xml) and output results (.dat) of the real transient models are located in model_*3-*_transient/. Especially, for two example hillslope models (ID=-11 and 11), additional h5py files are included in model_03-hillslope_transient/hillslope-11/, model_03-hillslope_transient/hillslope11, model_13-subcatchment_transient/subcatchment-11/, model_13-subcatchment_transient/subcatchment/11, respectively, which are used to plot the saturation figure (Figure 5) in the manuscript. Accessible by Python. (6) Arctic models are located in huc190604020802_gauge15906000/, which includes two cases: decomposed 2D models (inside model_04-hillslope_transient), and full 3D model (inside model_05-onepiece_transient_mannp1_ra). Three step spin-up results (checkpoint_final.h5) are located in model_01-column_freezeup/, model_02-column_spinup/, model_03-hillslope_spinup/, respectively, which are used to initialize real 2D transient hillslope models. The input files (.xml) and output results (.dat) of transient 2D hillslope models are located in model_04-hillslope_transient/. The input files (.xml) and output results (.dat) of the full 3D transient model is located in model_05-onepiece_transient_mannp1_ra/. The full 3D transient model is initialized by model_02-column_spinup/. Accessible by Python. (7) The MOSART routed discharge results (.csv) under Arctic conditions is located in huc190604020802_gauge15906000/MOSART/. Accessible by Python. (8) All Python codes (.py) used to parameterize full 3D model to decomposed 2D models are located in script/. These codes fit with watershed workflow (a watershed delineation tool) v1.4 under the branch gaob/v1.4 from https://github.com/gaobhub/watershed-workflow.git.

EARTH SCIENCE > CRYOSPHERE↗

Reduced fast electron transport in shock-heated plasma in multilayer targets due to self-generated magnetic fields

Fast electron transport has been studied in cold solid density CH, cold CH foam (200 mg / cm 3 ), and CH plasma (40 eV 30 mg/cm 3 ) targets—the latter created by shocking the CH foam with a 1.2 kJ long pulse laser and allowing it to expand. The fast electrons were produced using the OMEGA EP laser pulse (800 J, 8 ps) incident on a Au flat target. With the CH plasma, the fluence of fast electrons reaching a Cu foil at the far side of the transport was reduced significantly (25 × weaker peak K $α$ emission). Particle-in-cell simulations using the osiris code modeled fast electron transport in the unshocked foam and plasma cases assuming fixed ionization and including source generation, transport in Au and CH layers, Coulomb collisions, and refluxing. Simulations indicate two main mechanisms which alter electron energy transport through the target between the foam and plasma cases, both due to the magnetic field: a collimating field in the CH region, caused by the resistivity of the return current and more prevalent in the foam; and an insulating field at the Au-CH interface, present only with the plasma.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Process Anomaly Detection for Sparsely Labeled Events in Nuclear Power Plants

An essential aspect of online monitoring, subtle anomaly detection increases the detection lead time for equipment failure and enables a nuclear power plant (NPP) to mitigate unexpected partial or full outages, resulting in significant cost saving to the plant. Once an anomaly is detected by plant staff, its cause and severity are investigated. Because the vast majority of anomalies require some level of investigation, including some that require time-consuming examination, before they are passed over to the engineering organization for further analysis, plants are often equipped with tools to assist the staff in performing anomaly detection. Those tools operate as a black box and are often based on statistical methods that establish sensor correlations using preconfigured mathematical models and flag correlation deviations as anomalies. Due to the number of anomalies detected at a given NPP on a daily basis, a significant number of flagged anomalies usually await examination for days or weeks. A primary cause of this backlog is that the methods used by the tools generate many false positives. Though this is usually attributed to oversensitive model settings due to very narrow normal operation bands, it can also be associated with the model development being inadequate for the process being monitored, or with missing model inputs that could have explained misclassified positives. The performance of anomaly detection tools impacts their plant acceptance and utilization, especially when the effort to address false positives generated by the tool depletes the value or cost saved by using that tool. Thus, means to advance anomaly detection performance have been investigated by the Department of Energy’s Light Water Reactor Sustainability program. Previous and ongoing efforts have targeted unsupervised machine-learning (ML) methods, which do not require the labeling of any data fed into the ML model. By contrast, in supervised anomaly detection methods, every data point is labeled as either a normal or abnormal process condition, and the model is trained to replicate the classification process. Supervised methods usually outperform unsupervised methods, due to the added value in differentiating normal from anomalous states of the monitored process. An NPP’s corrective action program requires it to track and document, via a dedicated report, the resolution of any issues that occur within the plant. Once created, each report is reviewed by a plant screening committee, and several classifications and decisions are made. Recently, a collaborating NPP developed an artificial intelligence and ML-based classifier to categorize a condition report (CR) into classes that can serve to label the data as normal or anomalous. Applying CRs as labels represents a semi-supervised use case. Semi-supervised ML assumes that labels exist for some data points (i.e., labeled anomalies, in this case) but not for the rest. In this effort, semi-supervised ML methods were used to fuse data from CRs with anomaly detection methods in order to test the hypothesis that partially labeled anomalies would improve the accuracy of the anomaly detection methods. Specifically, two methods were used. The first is the deep Semi-supervised Anomaly Detection (deep SAD) method, which can handle labels ranging from fully unsupervised to fully supervised cases. The second is a newly designed ML method developed specifically for this effort and referred to as the high-order feature (HOF)-based method. To evaluate these two methods in controlled environments, synthetic data generators were developed and used. The first datasets used a spring-mass-damper (SMD) system simulator commonly found in mechanical engineering references. This was used to create two use cases: a one- and a three-mass system. Anomalies were introduced by changing the spring and damper coefficients while the system was actuated by random forces. The second datasets used the commercial Dymola-Modelica software to build a simplified nuclear reactor model. Anomalies were added in the form of corrupted sensor readings and/or control commands. The deep SAD method was tested using the SMD system, while the HOF method was tested using both datasets. Application of the deep SAD semi-supervised ML method demonstrated that labels can generate increased confidence in detecting true anomalies. This helped increase the number of true positives and decrease the number of false negatives—something that would aid in addressing the backlog of possible anomalies. Application of the HOF method demonstrated that labels can aid in down selecting from a candidate set of features to a more optimal subset in order to better differentiate between normal and anomalous conditions.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Navier-Stokes simulation of wind-tunnel flow using LU-ADI factorization algorithm

The three dimensional Navier-Stokes solution code using the LU-ADI factorization algorithm was employed to simulate the workshop test cases of transonic flow past a wing model in a wind tunnel and in free air. The effect of the tunnel walls is well demonstrated by the present simulations. An Amdahl 1200 supercomputer having 128 Mbytes main memory was used for these computations.

Obayashi, Shigeru↗

Navier-Stokes simulation of wind-tunnel flow using LU-ADI factorization algorithm

The three dimensional Navier-Stokes solution code using the LU-ADI factorization algorithm was employed to simulate the workshop test cases of transonic flow past a wing model in a wind tunnel and in free air. The effect of the tunnel walls is well demonstrated by the present simulations. An Amdahl 1200 supercomputer having 128 Mbytes main memory was used for these computations.

Obayashi, Shigeru↗

Launch Control System Software Development System Automation Testing

The Spaceport Command and Control System (SCCS) is the National Aeronautics and Space Administration's (NASA) launch control system for the Orion capsule and Space Launch System, the next generation manned rocket currently in development. This system requires high quality testing that will measure and test the capabilities of the system. For the past two years, the Exploration and Operations Division at Kennedy Space Center (KSC) has assigned a group including interns and full-time engineers to develop automated tests to save the project time and money. The team worked on automating the testing process for the SCCS GUI that would use streamed simulated data from the testing servers to produce data, plots, statuses, etc. to the GUI. The software used to develop automated tests included an automated testing framework and an automation library. The automated testing framework has a tabular-style syntax, which means the functionality of a line of code must have the appropriate number of tabs for the line to function as intended. The header section contains either paths to custom resources or the names of libraries being used. The automation library contains functionality to automate anything that appears on a desired screen with the use of image recognition software to detect and control GUI components. The data section contains any data values strictly created for the current testing file. The body section holds the tests that are being run. The function section can include any number of functions that may be used by the current testing file or any other file that resources it. The resources and body section are required for all test files; the data and function sections can be left empty if the data values and functions being used are from a resourced library or another file. To help equip the automation team with better tools, the Project Lead of the Automated Testing Team, Jason Kapusta, assigned the task to install and train an optical character recognition (OCR) tool to Brandon Echols, a fellow intern, and I. The purpose of the OCR tool is to analyze an image and find the coordinates of any group of text. Some issues that arose while installing the OCR tool included the absence of certain libraries needed to train the tool and an outdated software version. We eventually resolved the issues and successfully installed the OCR tool. Training the tool required many images and different fonts and sizes, but in the end the tool learned to accurately decipher the text in the images and their coordinates. The OCR tool produced a file that contained significant metadata for each section of text, but only the text and coordinates of the text was required for our purpose. The team made a script to parse the information we wanted from the OCR file to a different file that would be used by automation functions within the automated framework. Since a majority of development and testing for the automated test cases for the GUI in question has been done using live simulated data on the workstations at the Launch Control Center (LCC), a large amount of progress has been made. As of this writing, about 60% of all of automated testing has been implemented. Additionally, the OCR tool will help make our automated tests more robust due to the tool's text recognition being highly scalable to different text fonts and text sizes. Soon we will have the whole test system automated, allowing for more full-time engineers working on development projects.

Automation↗

Steady-State Gyrokinetics Transport Code (SSGKT), A Scientific Application Partnership with the Framework Application for Core-Edge Transport Simulations, Final Report

This project initiated the development of TGYRO - a steady-state Gyrokinetic transport code (SSGKT) that integrates micro-scale GYRO turbulence simulations into a framework for practical multi-scale simulation of conventional tokamaks as well as future reactors. Using a lightweight master transport code, multiple independent (each massively parallel) gyrokinetic simulations are coordinated. The capability to evolve profiles using the TGLF model was also added to TGYRO and represents a more typical use-case for TGYRO. The goal of the project was to develop a steady-state Gyrokinetic transport code (SSGKT) that integrates micro-scale gyrokinetic turbulence simulations into a framework for practical multi-scale simulation of a burning plasma core ? the International Thermonuclear Experimental Reactor (ITER) in particular. This multi-scale simulation capability will be used to predict the performance (the fusion energy gain, Q) given the H-mode pedestal temperature and density. At present, projections of this type rely on transport models like GLF23, which are based on rather approximate fits to the results of linear and nonlinear simulations. Our goal is to make these performance projections with precise nonlinear gyrokinetic simulations. The method of approach is to use a lightweight master transport code to coordinate multiple independent (each massively parallel) gyrokinetic simulations using the GYRO code. This project targets the practical multi-scale simulation of a reactor core plasma in order to predict the core temperature and density profiles given the H-mode pedestal temperature and density. A master transport code will provide feedback to O(16) independent gyrokinetic simulations (each massively parallel). A successful feedback scheme offers a novel approach to predictive modeling of an important national and international problem. Success in this area of fusion simulations will allow US scientists to direct the research path of ITER over the next two decades. The design of an efficient feedback algorithm is a serious numerical challenge. Although the power source and transport balance coding in the master are standard, it is nontrivial to design a feedback loop that can cope with outputs that are both intermittent and extremely expensive. A prototypical feedback scheme has already been successfully demonstrated for a single global GYRO simulation, although the robustness and efficiency are likely far from optimal. Once the transport feedback scheme is perfected, it could, in principle, be embedded into any of the more elaborate transport codes (ONETWO, TRANSP, and CORSICA), or adopted by other FSP-related multi-scale projects.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

SU-F-J-201: Validation Study of Proton Radiography Against CT Data for Quantitative Imaging of Anatomical Changes in Head and Neck Patients

Purpose: In clinical pencil-beam-scanned (PBS) proton therapy, the advantage of the characteristic sharp dose fall-off after the Bragg Peak (BP) becomes a disadvantage if the BP positions of a plan’s constituent pencil beams are shifted, eg.due to anatomical changes. Thus, for fractionated PBS proton therapy, accurate knowledge of the water equivalent path length (WEPL) of the traversed anatomy is critical. In this work we investigate the feasibility of using 2D proton range maps (proton radiography, PR) with the active-scanning gantry at PSI. Methods: We simulated our approach using Monte Carlo methods (MC) to simulate proton beam interactions in patients using clinical imaging data. We selected six head and neck cases having significant anatomical changes detected in per-treatment CTs.PRs (two at 0°/90°) were generated from MC simulations of low-dose pencil beams at 230MeV. Each beam’s residual depth-dose was propagated through the patient geometry (from CT) and detected on exiting the patient anatomy in an ideal depth-resolved detector (eg. range telescope). Firstly, to validate the technique, proton radiographs were compared to the ground truth, which was the WEPL from ray-tracing in the patient CT at the pencil beam location. Secondly, WEPL difference maps (per-treatment – planning imaging timepoints) were then generated to locate the anatomical changes, both in the CT (ground truth) and in the PRs. Binomial classification was performed to evaluate the efficacy of the technique relative to CT. Results: Over the projections simulated over all six patients, 70%, 79% and 95% of the grid points agreed with the ground truth proton range to within ±0.5%, ±1%, and ±3% respectively. The sensitivity, specificity, precision and accuracy were high (mean±1σ, 83±8%, 87±13%, 95±10%, 83±7% respectively). Conclusion: We show that proton-based radiographic images can accurately monitor patient positioning and in vivo range verification, while providing equivalent WEPL information to a CT scan, with the advantage of a much lower imaging dose.

60 APPLIED LIFE SCIENCES↗

Divergence Reduction in Monte Carlo Neutron Transport with On-GPU Asynchronous Scheduling

While Monte Carlo Neutron Transport (MCNT) is near-embarrasingly parallel, the effectively unpredictable lifetime of neutrons can lead to divergence when MCNT is evaluated on GPUs. Divergence is the phenomenon of adjacent threads in a warp executing different control flow paths; on GPUS, it reduces performance because each work group may only execute one path at a time. The process of Thread Data Remapping (TDR) resolves these discrepancies by moving data across hardware such that data in the same warp will be processed through similar paths. A common issue among prior implementations of TDR is the synchronous nature of its remapping and processing cycles, which exhaustively sort data produced by prior processing passes and exhaustively evaluate the sorted data. In another work, we defined a method of remapping data through an asynchronous scheduler which allows for work to be stored in shared memory and deferred arbitrarily until that work is a viable option for low-divergence evaluation. This article surveys a wider set of cases, with the goal of characterizing performance trends across a more comprehensive set of parameters. These parameters include cross sections of scattering/capturing/fission, use of implicit capture, source neutron counts, simulation time spans, and tuned memory allocations. Across these cases, we have recorded minimum and average execution times, as well as a heuristically tuned near-optimal memory allocation size for both synchronous and asynchronous scheduling. Across the collected data, it is shown that the asynchronous method is faster and more memory efficient in the majority of cases, and that it requires less tuning to achieve competitive performance.

Computer Science↗

Sensitivity of the interannual variability of mineral aerosol simulations to meteorological forcing dataset

Abstract. Interannual variability in desert dust is widely observed and simulated, yet the sensitivity of these desert dust simulations to a particular meteorological dataset, as well as a particular model construction, is not well known. Here we use version 4 of the Community Atmospheric Model (CAM4) with the Community Earth System Model (CESM) to simulate dust forced by three different reanalysis meteorological datasets for the period 1990–2005. We then contrast the results of these simulations with dust simulated using online winds dynamically generated from sea surface temperatures, as well as with simulations conducted using other modeling frameworks but the same meteorological forcings, in order to determine the sensitivity of climate model output to the specific reanalysis dataset used. For the seven cases considered in our study, the different model configurations are able to simulate the annual mean of the global dust cycle, seasonality and interannual variability approximately equally well (or poorly) at the limited observational sites available. Overall, aerosol dust-source strength has remained fairly constant during the time period from 1990 to 2005, although there is strong seasonal and some interannual variability simulated in the models and seen in the observations over this time period. Model interannual variability comparisons to observations, as well as comparisons between models, suggest that interannual variability in dust is still difficult to simulate accurately, with averaged correlation coefficients of 0.1 to 0.6. Because of the large variability, at least 1 year of observations at most sites are needed to correctly observe the mean, but in some regions, particularly the remote oceans of the Southern Hemisphere, where interannual variability may be larger than in the Northern Hemisphere, 2–3 years of data are likely to be needed.

99 GENERAL AND MISCELLANEOUS↗

USM3D Simulations for the Third Sonic Boom Prediction Workshop (SBPW-3)

The NASA USM3D flow solver was used to compute test cases for the Third AIAA Sonic Boom Prediction Workshop (SBPW3). The test cases include a nearfield biconvex shock-plume interaction wind tunnel model, and the C608 Low Boom Flight Demonstrator. Numerical simulations were conducted on the mixed-element and tetrahedral grids provided by the workshop committee, as well as a family of grids generated by an in-house approach for sonic boom analyses known as BoomGrid. The near-field pressure signatures were extracted, propagated to the ground and the perceived loudness levels on the ground were computed. The USM3D near-field pressure signatures, corresponding ground signatures, and loudness levels on the ground are compared with that of mean values from other workshop participants. The effect of three flux-limiters on the accuracy of near-field pressure signatures prediction was investigated and results are compared with that of mean values from other workshop participants. The effect of using wall function grids in accurately predicting near-field pressure signatures was also evaluated.

Supersonics↗

Simulating and benchmarking neutron total scattering instrumentation from inception of events to reduced and fitted data

In the design and realization of modern neutron scattering instrumentation, particularly when designing beamline concepts from the ground up, it is desirable to fully benchmark against realistically simulated data. This is especially true for total scattering beamlines, where the future deliverable data is to be analysed in both reciprocal- and real-space representations, and needs must be carefully balanced to ensure sufficient range, resolution and flux of the instrument. An approach to optimize the design of neutron scattering instrumentation via a workflow including ray-tracing simulations, event-based data reduction, heuristic analysis and fitting against realistically simulated spectra is demonstrated here. The case of the DISCOVER beamline concept at the Spallation Neutron Source is used as an example. The results of the calculations are benchmarked through simulation of existing instrumentation and subsequent direct comparison with measured data. Finally, on the basis of the validated models, the ability to explore design characteristics for future beamline concepts or future instrument improvements is demonstrated through the examples of detector tube size and detector layout.

47 OTHER INSTRUMENTATION↗

Numerical simulation of a disk-shaped electron accelerating electrostatic probe

This paper presents a method utilizing the numerical plasma simulation particle-in-cell technique for studying electron accelerating probe characteristics. A probe configuration other than the commonly used plane, cylindrical, or spherical case will be studied; the probe configuration used is that of a disk. Cylindrical geometry is used in the simulation model. One of the advantages of a numerical simulation is the ease with which the details of the system can be examined. Therefore, in addition to the current-voltage characteristics of the static probe, other results are presented such as the time evolution of the plasma properties and studies of the sheath region surrounding the probe.

Nonnast, J. H.↗

Optimization of Turbine Blade Design for Reusable Launch Vehicles

To facilitate design optimization of turbine blade shape for reusable launching vehicles, appropriate techniques need to be developed to process and estimate the characteristics of the design variables and the response of the output with respect to the variations of the design variables. The purpose of this report is to offer insight into developing appropriate techniques for supporting such design and optimization needs. Neural network and polynomial-based techniques are applied to process aerodynamic data obtained from computational simulations for flows around a two-dimensional airfoil and a generic three- dimensional wing/blade. For the two-dimensional airfoil, a two-layered radial-basis network is designed and trained. The performances of two different design functions for radial-basis networks, one based on the accuracy requirement, whereas the other one based on the limit on the network size. While the number of neurons needed to satisfactorily reproduce the information depends on the size of the data, the neural network technique is shown to be more accurate for large data set (up to 765 simulations have been used) than the polynomial-based response surface method. For the three-dimensional wing/blade case, smaller aerodynamic data sets (between 9 to 25 simulations) are considered, and both the neural network and the polynomial-based response surface techniques improve their performance as the data size increases. It is found while the relative performance of two different network types, a radial-basis network and a back-propagation network, depends on the number of input data, the number of iterations required for radial-basis network is less than that for the back-propagation network.

Shyy, Wei↗

Enhancing Autonomous Control of Microreactors Using Multi-Agent Reinforcement Learning

In order for microreactors to be economically competitive, operation costs will need to be minimized through some degree of autonomous control. Previous work has demonstrated the effectiveness of reinforcement learning (RL) for load-following control in a drum-controlled microreactor. This study extends that work by exploring the potential of RL to independently control each of the reactor’s drums. We compare a single-agent RL approach with a multi-agent RL (MARL) framework, testing them for generalization across different load-following power profiles and control timescales, and for robustness in cases of randomly disabled control drums. Since the point kinetics simulation environment used in this study cannot resolve spatial effects, we assume that in the absence of spatially localized disturbances, optimal drum movements should be symmetrical. We demonstrate that single-agent RL is able to achieve accurate performance only when symmetric actions are ignored; otherwise, it fails to train a useful controller. Meanwhile, the MARL framework performs symmetric actions by design and trains a robust, accurate agent, as evidenced by mean absolute errors in power matching of 0.41% for the training power profile, 0.68% for a profile with half the drums disabled, and 0.21% for a profile on a realistic load-following time horizon.

22 - GENERAL STUDIES OF NUCLEAR REACTORS↗