Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “computational challenge”

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

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

At least 361 records · Page 20

The scientific potential and technological challenges of the High-Luminosity Large Hadron Collider program

Here, we present an overview of the High-Luminosity (HL-LHC) program at the Large Hadron Collider (LHC), its scientific potential and technological challenges for both the accelerator and detectors. The HL-LHC program is expected to start circa 2027 and aims to increase the integrated luminosity delivered by the LHC by an order of magnitude at the collision energy of 14 TeV. This requires upgrades to the injector system, accelerator complex and luminosity levelling. The two experiments, ATLAS and CMS, require substantial upgrades to most of their systems in order to cope with the increased interaction rate, and much higher radiation levels than at the current LHC. We present selected examples based on novel ideas and technologies for applications at a hadron collider. Both experiments will replace their tracking systems. We describe the ATLAS pixel detector upgrade featuring novel tilted modules, and the CMS Outer Tracker upgrade with a new module design enabling use of tracks in the level-1 trigger system. CMS will also install state-of-the-art highly segmented calorimeter endcaps. Finally, we describe new picosecond precision timing detectors of both experiments. In addition, we discuss how the upgrades will enhance the physics performance of the experiments, and solve the computing challenges posed by the expected large data sets. The physics program of the HL-LHC is focused on precision measurements probing the limits of the Standard Model (SM) of particle physics and discovering new physics. We present a selection of studies that have been carried out to motivate the HL-LHC program. A central topic of exploration will be the characterization of the Higgs boson. The large HL-LHC data samples will extend the sensitivity of searches for new particles or new interactions whose existence has been hypothesized in order to explain shortcomings of the SM. Finally, we comment on the nature of large scientific collaborations.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Interactive Gas Chemistry for Enhanced Science Capabilities of the Energy Exascale Earth System Model Version 3

Atmospheric chemistry plays a crucial role in Earth system models (ESMs), controlling atmospheric composition and radiative balance; it is highly interactive with the physical climate, biogeochemical cycles, and human systems. However, it often imposes computational challenges in an ESM. Here we develop a full troposphere‐stratosphere interactive chemistry module for the US Department of Energy's Energy Exascale Earth System Model (E3SM). We intentionally build a streamlined module based on E3SM version 2 that interacts with other components and maintains all of major chemical and chemistry‐climate feedbacks. The module incorporates a new, highly efficient tracer advection scheme; linearization of stratospheric chemistry; and abridged tropospheric chemical mechanism with 28 reactive tracers. This new model, E3SM‐chem, can readily perform century‐long climate simulations of ozone, methane, and nitrous oxide based on emission scenarios as well as provide hourly budgets for the gas‐phase radicals that drive aerosol chemistry. We evaluate E3SM‐chem with an atmosphere‐only simulation as in the recent climate model intercomparison project (CMIP6) finding results similar to the other CMIP6 models. For the present‐day, E3SM‐chem matches the standard measurement metrics for stratospheric and tropospheric ozone, surface air quality, other key reactive gases like carbon monoxide, and the methane lifetime. Overall, E3SM‐chem maintains the climate fidelity of the baseline model while adding at most 20% to the computational cost of the atmosphere model. Hence, interactive chemistry can be a default configuration for long climate simulations at resolutions of 1° or finer, which is crucial for producing self‐consistent chemistry‐climate feedbacks that alter the climate system.

54 ENVIRONMENTAL SCIENCES↗

An Evolutionary Algorithm to Personalize Stool-Based Colorectal Cancer Screening

Fecal immunochemical testing (FIT) is an established method for colorectal cancer (CRC) screening. Measured FIT-concentrations are associated with both present and future risk of CRC, and may be used for personalized screening. However, evaluation of personalized screening is computationally challenging. In this study, a broadly applicable algorithm is presented to efficiently optimize personalized screening policies that prescribe screening intervals and FIT-cutoffs, based on age and FIT-history. We present a mathematical framework for personalized screening policies and a bi-objective evolutionary algorithm that identifies policies with minimal costs and maximal health benefits. The algorithm is combined with an established microsimulation model (MISCAN-Colon), to accurately estimate the costs and benefits of generated policies, without restrictive Markov assumptions. The performance of the algorithm is demonstrated in three experiments. In Experiment 1, a relatively small benchmark problem, the optimal policies were known. The algorithm approached the maximum feasible benefits with a relative difference of 0.007%. Experiment 2 optimized both intervals and cutoffs, Experiment 3 optimized cutoffs only. Optimal policies in both experiments are unknown. Compared to policies recently evaluated for the USPSTF, personalized screening increased health benefits up to 14 and 4.3%, for Experiments 2 and 3, respectively, without adding costs. Generated policies have several features concordant with current screening recommendations. The method presented in this paper is flexible and capable of optimizing personalized screening policies evaluated with computationally-intensive but established simulation models. It can be used to inform screening policies for CRC or other diseases. For CRC, more debate is needed on what features a policy needs to exhibit to make it suitable for implementation in practice.

60 APPLIED LIFE SCIENCES↗

Advanced multiphysics coupling for LWR fuel performance analysis

Even the most basic nuclear fuel analysis is a multiphysics undertaking, as a credible simulation must consider at a minimum coupled heat conduction and mechanical deformation. The need for more realistic fuel modeling under a variety of conditions invariably leads to a desire to include coupling between a more complete set of the physical phenomena influencing fuel behavior, including neutronics, thermal hydraulics, and mechanisms occurring at lower length scales. This paper covers current efforts toward coupled multiphysics LWR fuel modeling in three main areas. The first area covered in this paper concerns thermomechanical coupling. The interaction of these two physics, particularly related to the feedback effect associated with heat transfer and mechanical contact at the fuel/clad gap, provides numerous computational challenges. An outline is provided of an effective approach used to manage the nonlinearities associated with an evolving gap in BISON, a nuclear fuel performance application. A second type of multiphysics coupling described here is that of coupling neutronics with thermomechanical LWR fuel performance. DeCART, a high-fidelity core analysis program based on the method of characteristics, has been coupled to BISON. DeCART provides sub-pin level resolution of the multigroup neutron flux, with resonance treatment, during a depletion or a fast transient simulation. Two-way coupling between these codes was achieved by mapping fission rate density and fast neutron flux fields from DeCART to BISON and the temperature field from BISON to DeCART while employing a Picard iterative algorithm. Finally, the need for multiscale coupling is considered. Fission gas production and evolution significantly impact fuel performance by causing swelling, a reduction in the thermal conductivity, and fission gas release. The mechanisms involved occur at the atomistic and grain scale and are therefore not the domain of a fuel performance code. However, it is possible to use lower length scale models such as those used in the mesoscale MARMOT code to compute average properties, e.g. swelling or thermal conductivity. These may then be used by an engineering-scale model. Examples of this type of multiscale, multiphysics modeling are shown.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

sciCAN: single-cell chromatin accessibility and gene expression data integration via cycle-consistent adversarial network

The boom in single-cell technologies has brought a surge of high dimensional data that come from different sources and represent cellular systems from different views. With advances in these single-cell technologies, integrating single-cell data across modalities arises as a new computational challenge. Here, we present an adversarial approach, sciCAN, to integrate single-cell chromatin accessibility and gene expression data in an unsupervised manner. We benchmarked sciCAN with 5 existing methods in 5 scATAC-seq/scRNA-seq datasets, and we demonstrated that our method dealt with data integration with consistent performance across datasets and better balance of mutual transferring between modalities than the other 5 existing methods. We further applied sciCAN to 10X Multiome data and confirmed that the integrated representation preserves biological relationships within the hematopoietic hierarchy. Finally, we investigated CRISPR-perturbed single-cell K562 ATAC-seq and RNA-seq data to identify cells with related responses to different perturbations in these different modalities.

59 BASIC BIOLOGICAL SCIENCES↗

Accurate direct Eulerian simulation of dynamic elastic-plastic flow

The simulation of dynamic, large strain deformation is an important, difficult, and unsolved computational challenge. Existing Eulerian schemes for dynamic material response are plagued by unresolved issues. We present a new scheme for the first-order system of elasto-plasticity equations in the Eulerian frame. This system has an intrinsic constraint on the inverse deformation gradient. Standard Godunov schemes do not satisfy this constraint. The method of Flux Distributions (FD) was devised to discretely enforce such constraints for numerical schemes with cell-centered variables. We describe a Flux Distribution approach that enforces the inverse deformation gradient constraint. As this approach is new and novel, we do not yet have numerical results to validate our claims. This paper is the first installment of our program to develop this new method.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Quantum Spectral Methods for Differential Equations

Recently developed quantum algorithms address computational challenges in numerical analysis by performing linear algebra in Hilbert space. Such algorithms can produce a quantum state proportional to the solution of a d-dimensional system of linear equations or linear differential equations with complexity poly(logd). While several of these algorithms approximate the solution to within ϵ with complexity poly(log(1/ϵ)), no such algorithm was previously known for differential equations with time-dependent coefficients. In this work, we develop a quantum algorithm for linear ordinary differential equations based on so-called spectral methods, an alternative to finite difference methods that approximates the solution globally. Using this approach, we give a quantum algorithm for time-dependent initial and boundary value problems with complexity poly(logd, log(1/ϵ)).

97 MATHEMATICS AND COMPUTING↗

High-order fully implicit solver for all-speed fluid dynamics: AUSM ride from nearly incompressible variable-density flows to shock dynamics

We introduce a novel Newton–Krylov (NK)-based fully implicit algorithm for solving fluid flows in a wide range of flow conditions—from variable density nearly incompressible to supersonic shock dynamics. The key enabling feature of our all-speed solver is the ability to efficiently solve conservation laws by choosing a set of independent variables that produce a well-conditioned Jacobian matrix for the linear iterations of the global nonlinear iterative solver. In particular, instead of choosing to discretize the conservative variables (density, momentum, total energy), which is traditionally used in Eulerian high-speed compressible fluid dynamics, we demonstrate superior performance by discretizing the primitive variables—pressure–velocity–temperature in the very low-Mach flow limits or density–velocity–temperature/entropy in the shock dynamics range. Moreover, our method allows us to avoid direct inversion of the mass matrix in discrete time derivatives, which is usually an additional source for stiffness, especially pronounced when going to very high-order schemes with non-orthogonal basis functions. Here, we show robust solutions obtained for discontinuous finite element discretization up to seventh-order accuracy. Another important aspect of the solution algorithm is the Advection Upstream Splitting Method (AUSM), adopted to compute numerical fluxes within our reconstructed discontinuous Galerkin (rDG) spatial discretization scheme. The use of the low-Mach modification of the hyperbolic flux operator is found to be necessary for enabling robust simulations of very stiff liquids and metals for Mach numbers below M=10 -5 , which is well known to be very computationally challenging for compressible solvers. We demonstrate that our fully implicit rDG-NK solver with the AUSM + -up flux treatment produces efficient and high-resolution numerical solutions at all speeds, ranging from vanishing Mach numbers to transonic and supersonic, without substantial modifications of the solution procedures. (At high speed, we add limiting and use a simpler preconditioning of the Krylov solver.) Numerical examples include nearly incompressible constant-property flow past a backward-facing step with heat transfer, low-Mach variable-property channel flow of water at supercritical state, phase change and melt pool dynamics for laser spot welding and selective laser melting in additive manufacturing, and Mach 3 flow in a wind tunnel with a step.

97 MATHEMATICS AND COMPUTING↗

A parallel domain decomposition-based implicit method for the Cahn–Hilliard–Cook phase-field equation in 3D

We present a numerical algorithm for simulating the spinodal decomposition described by the three dimensional Cahn–Hilliard–Cook (CHC) equation, which is a fourth-order stochastic partial differential equation with a noise term. The equation is discretized in space and time based on a fully implicit, cell-centered finite difference scheme, with an adaptive time-stepping strategy designed to accelerate the progress to equilibrium. At each time step, a parallel Newton–Krylov–Schwarz algorithm is used to solve the nonlinear system. We discuss various numerical and computational challenges associated with the method. The numerical scheme is validated by a comparison with an explicit scheme of high accuracy (and unreasonably high cost). We present steady state solutions of the CHC equation in two and three dimensions. The effect of the thermal fluctuation on the spinodal decomposition process is studied. We show that the existence of the thermal fluctuation accelerates the spinodal decomposition process and that the final steady morphology is sensitive to the stochastic noise. We also show the evolution of the energies and statistical moments. In terms of the parallel performance, it is found that the implicit domain decomposition approach scales well on supercomputers with a large number of processors.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Framework for optimization of long-term, multi-period investment planning of integrated urban energy systems

In order to achieve stringent greenhouse gas emission reductions, a transition of our entire energy system from fossil to renewable resources needs to be designed. Such an energy transition brings two main challenges: most renewables generate variable electric energy, yet most demand is currently not electric (carrier mismatch) and does not always manifest at the same time as supply (temporal mismatch). Integrating multiple energy infrastructures can address both challenges by using the synergy between different energy carriers; building on existing infrastructure, while allowing a robust and flexible integration of the new. This paper proposes an optimization framework for long-term, multi-period investment planning of urban energy systems in an integrated manner. We formulate it as a mixed-integer linear program, combining a capacitated facility location with a multi-dimensional, capacitated network design problem. It includes generation and network expansion planning as well as interconnections between networks and storage infrastructure for each energy system. It can incorporate pathway effects like techno-economic developments, policy measures, and weather variations. The intended use is to support urban decision makers with long-term investment planning, though it can be tailored to fit other geographical or temporal scales. We demonstrate the model using two cases based on an average city in The Netherlands, which wants to reduce its CO 2 -emissions with 95% by 2050. In the first case, we include explicit carbon-emission constraints to study the effects of the carrier mismatch. In the second case, we implement interannual weather variations to analyze the temporal mismatch. The results give valuable insights into the energy transition design strategy for urban decision makers. They also show the future potential, as well as the computational challenges of the optimization framework.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Electronic and optical properties of titanium nitride bulk and surfaces from first principles calculations

Prediction of the frequency-dependent dielectric function of thin films poses computational challenges, and at the same time experimental characterization by spectroscopic ellipsometry remains difficult to interpret because of changes in stoichiometry and surface morphology, temperature, thickness of the film, or substrate. In this work, we report calculations for titanium nitride (TiN), a promising material for plasmonic applications because of less loss and other practical advantages compared to noble metals. We investigated structural, electronic, and optical properties of stoichiometric bulk TiN, as well as of the TiN(100), TiN(110), and TiN(111) outermost surfaces. Density functional theory (DFT) and many-body GW methods (Green's (G) function-based approximation with screened Coulomb interaction (W)) were used, ranging from G{sub 0}W{sub 0}, GW{sub 0} to partially self-consistent sc-GW{sub 0}, as well as the GW-BSE (Bethe-Salpeter equation) and time-dependent DFT (TDDFT) methods for prediction of the optical properties. Structural parameters and the band structure for bulk TiN were shown to be consistent with previous work. Calculated dielectric functions, plasma frequencies, reflectivity, and the electron energy loss spectrum demonstrated consistency with experiment at the GW{sub 0}-BSE level. Deviations from experimental data are expected due to varying experimental conditions. Comparison of our results to spectroscopic ellipsometry data for realistic nanostructures has shown that although TDDFT may provide a computationally feasible level of theory in evaluation of the dielectric function, application is subject to validation with GW-BSE calculations.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Multi-Block Enhancement for Lagrangian Dendritic Mesh setup in Altair5

Initial mesh setup for an ASC mUlti-physics code at LANL is done using Altair5. Altair5 assumes that the final mesh is composed of logical structured mesh blocks linked together at mesh boundaries to form, ultimately, an unstructured mesh. Within these blocks, meshes may have dendrites, that is, local regions where two zones share common edges (in two-dimensions, or faces in three-dimensions) with a single zone. In many cases, contiguous subsets of the initial set of blocks may have the same material assigned to them, but without smoothing the mesh would form a computationally challenging initial mesh. Some of these blocks may also have zones with nodes on domain boundaries. This paper reports on the implementation of multiblock smoothing in Altair5, which allows for dendrites and for moving boundary nodes. Dendritic nodes are constrained to be located at the average of their neighbor nodes while boundary nodes are constrained to move along the boundary geometry. Two fundamentally different smoothing methods were implemented. First, a variational principle is presented that balances zonal size and distortion via a user selected weighting with constraints imposed using penalty methods for dendritic nodes and Lagrange multipliers for boundary nodes. Second, the Laplace-Beltrami smoother is presented. This is a general elliptic smoother which can easily be modified to give Laplacian and Winslow-Crowley mesh smoothing. Results are shown for several test meshes of interest.

45 MILITARY TECHNOLOGY, WEAPONRY, AND NATIONAL DEF↗

Deep learning of free boundary and Stefan problems

Free boundary problems appear naturally in numerous areas of mathematics, science and engineering. These problems present a great computational challenge because they necessitate numerical methods that can yield an accurate approximation of free boundaries and complex dynamic interfaces. In this work, we propose a multi-network model based on physics-informed neural networks to tackle a general class of forward and inverse free boundary problems called Stefan problems. Specifically, we approximate the unknown solution as well as any moving boundaries by two deep neural networks. Besides, we formulate a new type of inverse Stefan problems that aim to reconstruct the solution and free boundaries directly from sparse and noisy measurements. We demonstrate the effectiveness of our approach in a series of benchmarks spanning different types of Stefan problems, and illustrate how the proposed framework can accurately recover solutions of partial differential equations with moving boundaries and dynamic interfaces. All code and data accompanying this manuscript are publicly available at https://github.com/PredictiveIntelligenceLab/DeepStefan.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

A hybrid adaptive low-Mach number/compressible method: Euler equations

Flows in which the primary features of interest do not rely on high-frequency acoustic effects, but in which long-wavelength acoustics play a nontrivial role, present a computational challenge. Integrating the entire domain with low-Mach-number methods would remove all acoustic wave propagation, while integrating the entire domain with the fully compressible equations can in some cases be prohibitively expensive due to the CFL time step constraint. For example, simulation of thermoacoustic instabilities might require fine resolution of the fluid/chemistry interaction but not require fine resolution of acoustic effects, yet one does not want to neglect the long-wavelength wave propagation and its interaction with the larger domain. The present paper introduces a new multi-level hybrid algorithm to address these types of phenomena. In this new approach, the fully compressible Euler equations are solved on the entire domain, potentially with local refinement, while their low-Mach-number counterparts are solved on subregions of the domain with higher spatial resolution. The finest of the compressible levels communicates inhomogeneous divergence constraints to the coarsest of the low-Mach-number levels, allowing the low-Mach-number levels to retain the long-wavelength acoustics. Here, the performance of the hybrid method is shown for a series of test cases, including results from a simulation of the aeroacoustic propagation generated from a Kelvin-Helmholtz instability in low-Mach-number mixing layers. It is demonstrated that compared to a purely compressible approach, the hybrid method allows time-steps two orders of magnitude larger at the finest level, leading to an overall reduction of the computational time by a factor of 8.

97 MATHEMATICS AND COMPUTING↗

Effects of artificial collisions, filtering, and nonlocal closure approaches on Hermite-based Vlasov–Poisson simulations

Kinetic simulations of collisionless plasmas are computationally challenging due to phase-space mixing and filamentation, resulting in fine-scale velocity structures. This study compares three methods developed to reduce artifacts related to limited velocity resolution in Hermite-based Vlasov–Poisson simulations: artificial collisions, filtering, and nonlocal closure approaches. We evaluate each method's performance in approximating the linear kinetic response function and suppressing recurrence in linear and nonlinear regimes. Numerical simulations of Landau damping demonstrate that artificial collisions, particularly higher orders of the Lenard-Bernstein collisional operator, most effectively recover the correct damping rate across a range of wavenumbers. Moreover, Hou-Li filtering and nonlocal closures underdamp high wavenumber modes in linear simulations, and the Lenard-Bernstein collisional operator overdamps low wavenumber modes in both linear and nonlinear simulations. This study demonstrates that hypercollisions offer a robust approach to kinetic simulations, accurately capturing collisionless dynamics with limited velocity resolution.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Sound at Scale: Characterizing Impacts of Noise Ordinances on the Onshore Wind Energy Technical Potential for the United States

Recent surveys have documented the rapid rise of sound ordinances across state and county jurisdictions, which has become crucial for wind energy siting. However, the lack of information on ordinances and computational challenges in turbine sound modeling create uncertainties regarding how evolving policies may affect resource potential and clean energy objectives. Therefore, we develop an approach to evaluate wind turbine sound profiles at millions of locations across the U.S. and translate them into setback distances for every residential structure. Compared to a baseline reference scenario, we find a 7% reduction in the national wind energy capacity potential when accounting for counties with existing sound ordinances. Additionally, when expanding the surveyed sound ordinances nationwide, we observe a potential loss of 53% of the national wind capacity under the most stringent ordinances, with a disproportionate share of this lost capacity coming from high-quality and low-cost wind resource. This work reveals that neglecting sound ordinances results in a significant overestimation of wind resource potential and highlights the important trade-offs between increased wind energy deployment to meet target decarbonization goals and the social/environmental impacts of this deployment that must be considered.

aeroacoustics↗

Full 3D Kinetic Modeling and Quantification of Positive Streamer Evolution in an Azimuthally Swept Pin-to-Plane Wedge Geometry

Cathode-directed streamer evolution in near atmospheric air is modeled in 3D pin-to-plane geometries using a 3D kinetic Particle-In-Cell (PIC) code that simulates particle-particle collisions via the Direct Simulation Monte Carlo (DSMC) method. Due to the computational challenges associated with a complete 360° volumetric domain, a practical alternative was achieved using a wedge domain and a range of azimuthal angles was explored (5°, 15°, 30°, and 45°) to study possible effects on the streamer growth and propagation due to the finite wedge angle. A DC voltage of 6 kV is administered to a hemispherical anode of radius 100 μm, with a planar cathode held at ground potential, generating an over-volted state with an electric field of 4 MV/m across a 1500 μm gap. The domain is seeded with an initial ion and electron density of 10 18 m -3 at 1 eV temperature confined to a spherical region of radius 100 μm centered at the tip of the anode. The air chemistry model [1] includes standard Townsend breakdown mechanisms (electron-neutral elastic, excitation, ionization, attachment, and detachment collision chemistry and secondary electron emission) as well as streamer mechanisms (photoionization and ion-neutral collisions) via tracking excited state neutrals which can then either quench via collisions or spontaneously emit a photon based on specific Einstein-A coefficients [2, 3]. In this work, positive streamer dynamics are formally quantified for each wedge angle in terms of electron velocity and density as temporal functions of coordinates r, Φ, and z. Applying a random plasma seed for each simulation, particles of interest are tracked with near femtosecond temporal resolution out to 1.4 ns and spatially binned. This process is repeated six times and results are averaged. Prior 2D studies have shown that the reduced electric field, E/n, can significantly impact streamer evolution [4]. We extend the analysis to 3D wedge geometries, to limit computational costs, and examine the wedge angle’s effect on streamer branching, propagation, and velocity. Results indicate that the smallest wedge angle that produced an acceptably converged solution is 30°. The potential effects that a mesh, when under-resolved with respect to the Debye length, can impart on streamer dynamics and numerical heating were not investigated, and we explicitly state here that the smallest cell size was approximately 10 times the minimum λ D in the streamer channel at late times. This constraint on cell size was the result of computational limitations on total mesh count.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

A Decomposition-Based Learn-To-Optimize Approach with Feasibility Layer Assistance for Sub-Hourly Unit Commitment

Sub-hourly unit commitment (UC) with 15-min intervals is gaining significant attention as a way to respond rapidly to the fluctuations in electricity supply and demand introduced by renewable resources. However, the increased temporal resolution and complex inter-temporal dependencies pose substantial computational challenges for traditional optimization methods. To this end, this paper explores a decomposition-based learn-to-optimize approach. Building on recent advances in machine learning, our method revisits the long- overlooked Lagrangian relaxation framework, which is a classical decomposition technique that enables tractable subproblem solving. These smaller subproblems are inherently well-suited for machine learning, as their reduced dimensionality and structural regularity allow predictive models to efficiently learn and generalize solution patterns. We thus propose a generic predictive model, which embeds Gated Recurrent Units (GRUs) and Attention in the encoder-decoder structure, and integrate a rule-based feasibility layer to capture temporal dependencies, reduce training effort, and improve feasibility w.r.t. unit-level constraints. Our method has been validated on the IEEE 118-bus system, demonstrating promising performance in solving sub-hourly UC problems efficiently and feasibly.

97 MATHEMATICS AND COMPUTING↗