SEARCH · Engineering Papers

Results for “representation”

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 901 records · Page 50

All sources NASA NTRS DOE OSTI

DOE OSTI2020

Database development and exploration of process–microstructure relationships using variational autoencoders

The paper demonstrates graphical representation of a large database containing process–microstructure relationships using an unsupervised machine learning algorithm. Correlating microstructural features to processing is an essential first step to answer the difficult problem of process sequence design. Here, a large database of 346,200 orientation distribution functions resulting from a variety of process sequences is constructed, where each sequence comprises up to four stages of tension, compression and rolling along different directions in various permutations. This open-source database is constructed for collaborative development of process design algorithms. The paper demonstrates a novel application of the large database: graphical representation of texture–process relationships. A variational autoencoder is used to reduce the entire database to a two dimensional latent space where variations in processes and properties can be visualized. Using proximity analysis in this latent space, we can quickly unearth multiple process solutions to the problem of texture or property design.

36 MATERIALS SCIENCE↗

DOE OSTI2026

Dense autoencoders, clustering techniques, and semi-supervised learning for HPGe $γ$-spectra

Classifying high-resolution gamma spectra by their isotopic content is an essential task in nuclear forensics and other applications. Traditional analysis methods are often time-intensive, but machine learning (ML) may help analysts quickly process many spectra. Such methods tend to rely on abundant, well-labeled data for training. Historical gamma data exists in various fields but is not uniformly useful for supervised ML due to inconsistent labeling. Here, to address some of these challenges, we present a method to classify and organize unlabeled data from high-purity germanium detectors using an autoencoding neural network (autoencoder). We trained dense autoencoders to compress gamma data into latent representations that enable efficient data characterization. By clustering the encoded spectra or lower-dimensional mappings of them, we identified and removed portions of over-abundant data categories, resulting in a more balanced dataset and improved autoencoder performance. This encoding and clustering pipeline also enabled the organization of spectra into self-consistent categories. Finally, we found that encoded representations showed potential as inputs for semi-supervised learning of nuclide identification (NID) labels, achieving an average F1 score of 0.85 ± 0.03 when mapping encodings to a set of 65 isotope labels.

Autoencoders↗

DOE OSTI2019

A framework for assessment of predictive capability maturity and its application in nuclear thermal hydraulics

Here this work presents a formalized and computerized framework for the assessment of decision regarding the adequacy of a simulation tool for a nuclear reactor application. The decision regarding a code’s adequacy for an application is dependent on the assessment of different attributes that govern verification, validation, and uncertainty quantification of the code. In this work, the focus is on code validation. Therefore, the framework is developed and illustrated from the perspective of decision regarding the validation assessment of a code. Code validation assessment is performed based on the validation test results, data applicability, and process quality assurance factors. The process quality assurance factors warrant the trustworthiness of the evidence and help in checking people and process compliance with respect to the standard requirements. The proposed framework is developed using an argument modeling technique called Goal Structuring Notation (GSN). Goal structuring notation facilitates structural knowledge representation, information abstraction, evidence incorporation, and provides a skeletal structure for quantitative maturity assessment. The decision schema for the development of the decision model is based on the Predictive Capability Maturity Model (PCMM) and Analytic Hierarchy Process (AHP) and formalized using Goal structuring notation. Each decision attribute is formulated as a claim, where the degree of validity of the claim (attribute’s assessment) is expressed using different maturity levels. The GSN representation of the decision model is transformed into a confidence network to provide evidence-based quantitative maturity assessment using the Bayesian network. A metric based on the expected utility of maturity levels, called expected distance metric, is proposed to measure the distance between target maturity and achieved maturity on a scale of zero to one. Expected distance metric helps in comparing the assessment of different attributes and identification of major areas of concern in terms of modeling capability, data needs, and quality of assessment process. The practical application of the framework is demonstrated by a case study on validation assessment of a thermal-hydraulic code for a challenge problem called Departure from Nucleate Boiling (DNB).

42 ENGINEERING↗

DOE OSTI2025

A hybrid surrogate modeling framework for the Digital Twin of a Fluoride-salt-cooled High-temperature Reactor (FHR)

While nuclear energy is a non-greenhouse-gas emitting energy source, expensive operational costs due to the high-level of safety requirements decreases their competitiveness in the sustainable energy market. Advanced reactor concepts paired with Digital Twins aim to increase the commercialization gains of nuclear energy by reducing operational costs, increasing reactor reliability and enhancing power generation. To support Digital Twin tasks such as real-time autonomous control, proactive maintenance monitoring or optimizing power demand operations, a fast and accurate virtual representation of the Nuclear Power Plant (NPP) is required. The computational cost of high-fidelity, physics-based models are unsuitable for real-time analysis or scalability. Here, in this work, a hybrid surrogate modeling framework is developed fora Fluoride-salt-cooled High-temperature Reactor (FHR) that leverages physics-inspired models for key reactor components and uses data-driven methods for rapid system state space prediction. The Xenon reactivity feedback model is integrated to inform the surrogate model about the reactor core and the homologous pump theory model is the basis for representing pump degradation. Using a detailed, two dimensional thermal hydraulics model to generate data on the FHR, we train a network of Vectorized Autoregressive Moving-Average with eXogenous input (VARMAX) models to predict the remaining state values. The result is a surrogate model that provides a detailed reactor state representation of 41 system states and a pump degradation analysis. The framework is applied to Load Follows profiles, yielding high accuracy and a speedup that is more than 4000x faster compared to the higher- fidelity thermal hydraulics model, enabling real-time operational intelligence and applications in long horizon predictions. While the surrogate model framework is demonstrated for the particular case of FHR, the hybrid physical/data-driven modeling approach including the network of surrogates and the underlying modularity has the potential to be applied to other physical asset systems.

Digital Twins↗

DOE OSTI2026

Combined effects of horizontal and vertical resolution on reliable turbulence prediction at tidal energy sites: A systematic study in the Salish Sea, WA

Predicting turbulence characteristics with coastal ocean models is essential for tidal energy converter deployment. While large eddy simulation provides a detailed turbulence representation, computational limitations restrict its use to smaller domains. We systematically evaluate whether well-configured coastal models can provide reliable turbulence prediction through progressive refinement of 3D model representation. We implemented four model configurations (Levels 1–4) using terrain-following coordinates, isolating the impacts of horizontal resolution, vertical resolution, and layer distribution. We validated all configurations against field measurements from the Salish Sea, WA. Results show that tidal current velocity predictions remain unchanged regarding model configuration, but turbulence properties are sensitive to resolution refinement. Increasing vertical resolution alone proved insufficient; even with vertical sigma-levels rising from 11 to 41, significant underprediction persisted until finer horizontal resolution better captured bathymetric variations. The Level 4 configuration, incorporating geometric sigma-levels distribution, achieved turbulence prediction skill scores exceeding 0.90. Turbulence closure comparison revealed Mellor- Yamada 2.5 outperformed k-epsilon in TKE prediction (skill scores 0.84–0.94 versus 0.72–0.81) due to better boundary layer parameterization. This study shows that well-configured coastal models effectively bridge the gap between simplified tools and costly high-fidelity modeling, offering the tidal energy industry practical and cost-effective turbulence data at commercially relevant scales.

Marine Energy↗

DOE OSTI2021

Evaluating MPI resource usage summary statistics

The Message Passing Interface (MPI) remains the dominant programming model for scientific applications running on today’s high-performance computing (HPC) systems. This dominance stems from MPI’s powerful semantics for inter-process communication that has enabled scientists to write applications for simulating important physical phenomena. MPI does not, however, specify how messages and synchronization should be carried out. Those details are typically dependent on low-level architecture details and the message characteristics of the application. Therefore, analyzing an application’s MPI resource usage is critical to tuning MPI’s performance on a particular platform. The result of this analysis is typically a discussion of the mean message sizes, queue search lengths and message arrival times for a workload or set of workloads. While a discussion of the arithmetic mean in MPI resource usage might be the most intuitive summary statistic, it is not always the most accurate in terms of representing the underlying data. In this paper, we analyze MPI resource usage for a number of key MPI workloads using an existing MPI trace collector and discrete-event simulator. Our analysis demonstrates that the average, while easy and efficient to calculate, is a useful metric for characterizing latency and bandwidth measurements, but may not be a good representation of application message sizes, match list search depths, or MPI inter-operation times. Additionally, we show that the median and mode are superior choices in many cases. We also observe that the arithmetic mean is not the best representation of central tendency for data that are drawn from distributions that are multi-modal or have heavy tails. Furthermore, the results and analysis of our work provide valuable guidance on how we, as a community, should discuss and analyze MPI resource usage data for scientific applications.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2024

Resolving Discrepancies Between Theory and Experiment for the NCN + H Reaction

Flame modeling studies have highlighted the role of branching in the kinetics for the prompt NO switch reaction, NCN + H. In this reaction, there is strong competition between the CH + N2 (R1) and HCN + N (R2) product channels. Increased branching towards reaction R2 promotes the subsequent formation of Fenimore or Prompt NO. Recent direct shock tube studies on this complex reaction conclude that branching to reaction R2 is much more favorable than the predicted branching from theoretical studies. The experimentally predicted prompt NO switch temperature (TS), i.e., the temperature at which k2/(k1+k2) = 0.5, is at 1670 K, in contrast to theory that predicts Ts > 3200 K. The present theory/modeling work was initiated to highlight potential causes for this significant discrepancy between experiment and theory. Our analysis of the simulations reported in the experimental studies indicate that including a fall-off representation for C2H5 radical dissociation (with C2H5 produced from C2H5I dissociation and acting as an H-atom source) has a noticeable influence on the simulations for NCN decay and HCN formation in the shock tube studies. With the inclusion of a theory-based pressure fall-off representation for C2H5 dissociation, we show that this radical persists for much longer timescales even at the high-temperatures in the shock tube studies. This persistence then facilitates the rapid conversion of reactive H-atoms to CH3 radicals at the shock tube conditions via the H + C2H5 → CH3 + CH3 reaction. An updated theoretical analysis for this classic addition-elimination reaction is also provided in this work. The rapid formation of CH3 radicals necessitates the inclusion of additional reactions to adequately simulate the NCN and HCN data. Our simulations conclude that, with the inclusion of these additional reactions, the NCN and HCN profiles can be reasonably well simulated by the most recent theoretical predictions.

Ethyl Radicals↗

DOE OSTI2021

A GPU-based Approach for Path Planning Optimization via Travel Length Reduction

Typically, before constructing an object with an additive manufacturing system, the 3D object must be sent through a process called slicing. Slicing converts a 3D object commonly in the form of an STL file into a set of layers by horizontally intersecting a plane with the object at various heights. At each height, called a layer, multiple 2D polygons can be generated. Each polygon represents a boundary for solid geometry and is called an island. Each island is then comprised of multiple path types in an attempt to optimally fill the polygon. To move between each island and each islands’ paths, travels are inserted. Travels are simply motion by the system to move from one area of construction to another. Travels do not contribute to the construction of the object, and so, are considered wasted motion. In large-scale additive manufacturing, objects can be quite large and the distance between islands can be large as well. As a result, these travels can waste a significant amount of time. Ideally, travels would be as short as possible, however, computing global minimal travel paths is computationally expensive. To combat this problem, researchers at Oak Ridge National Lab developed a GPU-based approach to travel insertion based on a unique factoradic representation. This representation was then utilized by the GPU to solve the Traveling Salesman Problem (TSP). This algorithm was able to compute global minimal travel paths quickly resulting in faster object construction. A general investigation was also carried out to determine when a GPU vs CPU implementation would be beneficial.

Borish, Michael↗

DOE OSTI2019

Margins associated with loss of assured safety for systems with multiple weak links and strong links

Representations for margins associated with loss of assured safety (LOAS) for weak link (WL)/strong link (SL) systems involving multiple time-dependent failure modes are developed. Additionally, the following topics are described: (i) defining properties for WLs and SLs, (ii) background on cumulative distribution functions (CDFs) for link failure time, link property value at link failure, and time at which LOAS occurs, (iii) CDFs for failure time margins defined by (time at which SL system fails) - (time at which WL system fails), (iv) CDFs for SL system property values at LOAS, (v) CDFs for WL/SL property value margins defined by (property value at which SL system fails) - (property value at which WL system fails), and (vi) CDFs for SL property value margins defined by (property value of failing SL at time of SL system failure) - (property value of this SL at time of WL system failure). Included in this presentation is a demonstration of a verification strategy based on defining and approximating the indicated margin results with (i) procedures based on formal integral representations and associated quadrature approximations and (ii) procedures based on algorithms for sampling-based approximations.

42 ENGINEERING↗

DOE OSTI2023

Reliability modeling in a predictive maintenance context: A margin-based approach

Current system reliability methods (typically based on fault trees or reliability block diagrams) can effectively propagate reliability data from the asset to the system level in order to identify system critical points. However, employed asset reliability data are an approximated integral representation of the past industrywide operational experience, and they neglect the present asset health status (available, for example, from online monitoring data and diagnostic assessments) and forecasted health projection (when available from prognostic models). Asset health should be informed solely by that specific asset’s current and historical performance data and should not be an approximated integral representation of the past industrywide operational experience (as currently performed by system reliability models through Bayesian updating processes). Sensor data, diagnostic assessments, and prognostic assessments are in fact not considered in plant reliability models used to inform system engineers on the most critical assets. In addition, the propagation of quantitative health data from the asset to the system level is a challenge given the diverse nature and structure of health data elements (e.g., vibration spectra, temperature readings, expected failure time). Ideally, in a predictive maintenance context, system reliability models should support decision making by propagating available health information from the asset to the system level in order to provide a quantitative snapshot of system health and identify the most critical assets. Here, this paper is directly addressing these two goals by proposing a different approach for reliability modeling that relies on asset diagnostic and prognostic assessments, along with monitoring data to measure asset health. The propagation of health data from the asset to the system level is performed through fault tree models not in probability terms, but in terms of margin where margin is the “distance” between the present status and an undesired event (e.g., failure or unacceptable performance). Through a cause-effect lens, while classical reliability models target the effect associated with asset performance, a margin-based approach focuses on the cause of an undesired asset performance (i.e., its health). Hence, thinking of reliability in terms of margins implies decision-making based on causal reasoning. We will show how fault tree models can be solved using a margin language and how this process can effectively assist system engineers to identify the most critical assets.

97 - MATHEMATICS AND COMPUTING↗

DOE OSTI2021

How structural differences influence cross-model consistency: An electric sector case study

Multiple models are often employed to describe a range of possible outcomes for one or more scenarios, yielding insights into causal relationships and their uncertainties. Electric sector capacity expansion scenarios are a common topic of such efforts due to the economic influence of the electric sector, but model results typically span a broad solution space despite efforts to harmonize input assumptions, making decision implications difficult to discern. This study investigates the relationship between input harmonization and cross-model scenario consistency under disparate electric sector scenarios. We compare cross-model consistency between two state-of-the-art electric sector capacity expansion models (GCAM-USA and ReEDS) for six electric sector scenarios comprising alternate assumptions about fossil fuel resource availability, technology innovation, and long-term economy-wide transitions under four harmonization configurations varying model representations of electricity demand, fuel prices, renewable resources, and capacity retirements. These comparisons reveal that cross-model consistency can vary across scenarios under a given harmonization configuration, suggesting that harmonization efforts must often be scenario-specific if comparable cross-model consistency is desired. Model structural differences can hinder consistency, and the impact of these differences can depend on the scenario. Ultimately, thorough harmonization can reveal insights into cross-model consistency, which can be used to tighten uncertainty bounds and improve the decision-making implications of multi-model activities.

Cohen, Stuart↗

DOE OSTI2021

An overview of data tools for representing and managing building information and performance data

Building information modeling (BIM) has been widely adopted for representing and exchanging building data across disciplines during building design and construction. However, BIM's use in the building operation phase is limited. With the increasing deployment of low-cost sensors and meters, as well as affordable digital storage and computing technologies, growing volumes of data have been collected from buildings, their energy services systems, and occupants. Such data are crucial to help decision makers understand what, how, and when energy is consumed in buildings—a critical step to improving building performance for energy efficiency, demand flexibility, and resilience. However, practical analyses and use of the collected data are very limited due to various reasons, including poor data quality, ad-hoc representation of data, and lack of data science skills. To unlock value from building data, there is a strong need for a toolchain to curate and represent building information and performance data in common standardized terminologies and schemas, to enable interoperability between tools and applications. This study selected and reviewed 24 data tools based on common use cases of data across the building life cycle, from design to construction, commissioning, operation, and retrofits. The selected data tools are grouped into three categories: (1) data dictionary or terminology, (2) data ontology and schemas, and (3) data platforms. The data are grouped into ten typologies covering most types of data collected in buildings. This study resulted in five main findings: (1) most data representation tools can represent their intended data typologies well, such as Green Button for smart meter data and Brick schema for metadata of sensors in buildings and HVAC systems, but none of the tools cover all ten types of data; (2) there is a need for data schemas to represent the basis of design data and metadata of occupant data; (3) standard terminologies such as those defined in BEDES are only adopted in a few data tools; (4) integrating data across various stages in the building life cycle remains a challenge; and (5) most data tools were developed and maintained by different parties for different purposes, their flexibility and interoperability can be improved to support broader use cases. Finally, recommendations for future research on building data tools are provided for the data and buildings community based on the FAIR principles to make data Findable, Accessible, Interoperable, and Reusable.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

DOE OSTI2021

How Many Trip Requests Could We Support? An Activity-Travel Based Vehicle Scheduling Approach

In a world of ever-changing travel behavior and ever-increasing modal options, is vital to have integrated models that could capture the interactions between supply and demand layers of travel. Addressing this need, we propose three different versions of network representation and mathematical models for the activity-based vehicle routing problem to connect activity-travel graphs of passengers (demand layer) to spatio-temporal networks of vehicles (supply layer). Versions I and II are arc-based, while version III is path-based. In version I, we introduce the concept of activity-travel graphs for passengers. For vehicles, we construct space–time networks and add a new dimension, called “under-service state”, to track the execution status of trip requests at any location and time. In version II, we reduce the complexity of the network structure by eliminating the state dimension and some other modifications in the structure of the passengers’ and vehicles’ network. Although both versions can capture various behavioral constraints of the activity-based vehicle routing problem (e.g., mandatory and optimal activities, duration of activities, chain of activities, preferred starting and ending times of activities), due to the high level of complexity of the network structure, both versions can only solve small-sized problems. To tackle the computational complexity, we propose a path-based network representation in version III, and to make a balance between the disutility of passengers and vehicles, we present a tolled user equilibrium problem. Mathematical models are coded in C and GAMS and implemented on real-world Phoenix regional transportation network with more than 39 million trip requests, which demonstrate the effectiveness of the proposed solution for the original and restricted master problems.

33 ADVANCED PROPULSION SYSTEMS↗

DOE OSTI2020

Development and validation of the cavitation-induced erosion risk assessment tool

This work presents the development of a cavitation-induced erosion risk assessment (CIERA) tool that links multiphase flow simulation predictions with the progress towards material erosion. To develop a robust erosion modeling tool, the cavitation and erosion predictions for pressurized diesel fuel flow within a channel geometry were validated over a range of Reynolds and cavitation number conditions in two different aluminum channel geometries, one featuring a rounded inlet corner and the other with a sharp inlet corner. The multiphase flow development within the channel was modeled using a compressible mixture model, where phase change was represented with the homogeneous relaxation model and the turbulent flow evolution was modeled using a dynamic structure approach for Large Eddy Simulations. To improve representation of the incubation period before material rupture over existing approaches, a physics-based metric was derived based on the cumulative energy absorbed by the solid material from repeated hydrodynamic impacts. When the average peak pressure was related to the incubation period, the incubation period and its sensitivity to changes in flow conditions were found to be overpredicted. In contrast, predictions from CIERA provided a more accurate means to qualitatively and quantitatively predict the influence of flow conditions on the incubation period before material erosion. When the predicted stored energy was related to the solid material properties to estimate the incubation period, multiphase flow simulations demonstrated accurate representation of the sensitivity of erosion severity to changes in flow conditions. The use of CIERA led to quantitative agreement of the predicted incubation period within 5% of the experimentally measured incubation period.

42 ENGINEERING↗

DOE OSTI2026

Knowledge-guided graph machine learning for spatially distributed prediction of daily discharge and nitrogen export dynamics

Spatially distributed prediction of streamflow and nitrogen export dynamics is essential for precision management of agricultural watersheds. While temporal deep learning models such as Long Short-Term Memory (LSTM) have shown strong performance at basin scales, their ability to generalize spatially is limited by insufficient representation of spatial dependencies and flow paths, particularly under data-scarce conditions. To address this gap, we propose HydroGraphNet, a knowledge-guided graph machine learning framework that integrates process-based knowledge and explicit spatial learning into temporal modeling. This framework incorporates directed graph topology to encode watershed connectivity and upstream inflows, with mass balance constraints to improve physical consistency. To enhance generalization in sparsely monitored regions, HydroGraphNet is pretrained on synthetic data generated by the SWAT+ (Soil and Water Assessment Tool Plus) model. We evaluated HydroGraphNet in the Upper Sangamon River Basin (44 HUC-12 subwatersheds, 2001–2020) against two LSTM baselines: a lumped basin-level model and a distributed variant. When benchmarked on SWAT+ simulations in pretraining, HydroGraphNet improved test NSEs by 8.9% (discharge) and 13.7% (NO₃–N load) in temporal extrapolation, and by 27.1% and 34.7% in spatial extrapolation, relative to the Lumped LSTM baseline. After fine-tuning with USGS monitoring data, the model achieved mean test NSE (KGE) scores of 0.768 (0.861) for discharge and 0.626 (0.664) for NO₃–N load, substantially outperforming baselines. Attribution analysis further highlighted the importance of upstream inflow representation and graph-based spatial learning in capturing cross-subwatershed dependencies. The model also reproduced seasonal hydrological and biogeochemical patterns consistent with known processes, demonstrating its robustness and process fidelity for spatially distributed prediction. Altogether, HydroGraphNet advances the integration of physical knowledge and spatially explicit learning in hydrological modeling, offering a generalizable framework for distributed modeling to support spatially targeted water quality management in data-scarce watersheds.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2020

Unitary quantum lattice simulations for Maxwell equations in vacuum and in dielectric media

Utilizing the similarity between the spinor representation of the Dirac and the Maxwell equations that has been recognized since the early days of relativistic quantum mechanics, a quantum lattice algorithm (QLA) representation of unitary collision-stream operators of Maxwell's equations is derived for both homogeneous and inhomogeneous media. A second-order accurate 4-spinor scheme is developed and tested successfully for two-dimensional (2-D) propagation of a Gaussian pulse in a uniform medium whereas for normal (1-D) incidence of an electromagnetic Gaussian wave packet onto a dielectric interface requires 8-component spinors because of the coupling between the two electromagnetic polarizations. In particular, the well-known phase change, field amplitudes and profile widths are recovered by the QLA asymptotic profiles without the imposition of electromagnetic boundary conditions at the interface. The QLA simulations yield the time-dependent electromagnetic fields as the wave packet enters and straddles the dielectric boundary. QLA involves unitary interleaved non-commuting collision and streaming operators that can be coded onto a quantum computer: the non-commutation being the very reason why one perturbatively recovers the Maxwell equations.

Physics↗

DOE OSTI2021

Koopman wavefunctions and Clebsch variables in Vlasov–Maxwell kinetic theory

Motivated by recent discussions on the possible role of quantum computation in plasma simulations, here, we present different approaches to Koopman's Hilbert-space formulation of classical mechanics in the context of Vlasov–Maxwell kinetic theory. The celebrated Koopman–von Neumann construction is provided with two different Hamiltonian structures: one is canonical and recovers the usual Clebsch representation of the Vlasov density, the other is non-canonical and appears to overcome certain issues emerging in the canonical formalism. Furthermore, the canonical structure is restored for a variant of the Koopman–von Neumann construction that carries a different phase dynamics. Going back to van Hove's prequantum theory, the corresponding Koopman–van Hove equation provides an alternative Clebsch representation which is then coupled to the electromagnetic fields. Finally, the role of gauge transformations in the new context is discussed in detail.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

DOE OSTI2022

Data-driven model for divertor plasma detachment prediction

We present a fast and accurate data-driven surrogate model for divertor plasma detachment prediction leveraging the latent feature space concept in machine learning research. Our approach involves constructing and training two neural networks: an autoencoder that finds a proper latent space representation (LSR) of plasma state by compressing the multi-modal diagnostic measurements and a forward model using multi-layer perception (MLP) that projects a set of plasma control parameters to its corresponding LSR. By combining the forward model and the decoder network from autoencoder, this new data-driven surrogate model is able to predict a consistent set of diagnostic measurements based on a few plasma control parameters. In order to ensure that the crucial detachment physics is correctly captured, highly efficient 1D UEDGE model is used to generate training and validation data in this study. The benchmark between the data-driven surrogate model and UEDGE simulations shows that our surrogate model is capable of providing accurate detachment prediction (usually within a few per cent relative error margin) but with at least four orders of magnitude speed-up, indicating that performance-wise, it has the potential to facilitate integrated tokamak design and plasma control. Comparing with the widely used two-point model and/or two-point model formatting, the new data-driven model features additional detachment front prediction and can be easily extended to incorporate richer physics. This study demonstrates that the complicated divertor and scrape-off-layer plasma state has a low-dimensional representation in latent space. Understanding plasma dynamics in latent space and utilising this knowledge could open a new path for plasma control in magnetic fusion energy research.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗