Search NASA⌕ Search

SEARCH · Search NASA

Results for “Computer implementation”

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
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Note on two formulations of Crank-Nicolson method for Navier-Stokes equations

Here, we consider two formulations of the Crank-Nicolson (CN) method for the Navier-Stokes equations (NSE). The “natural” way of implementing CN for NSE is formally second order accurate in time for both velocity and pressure, whereas another formulation approximates pressure with only first order accuracy in time. Both versions of the method are applied to the benchmark problem of computing drag and lift in the flow around a cylinder. We show that the presumably more accurate version of the CN can create a solution with nonphysical oscillations and give incorrect predictions for the maximal drag coefficient, whereas the other formulation of the method predicts the drag and lift coefficients more accurately and does not introduce nonphysical oscillations. We locate the source of the issue and suggest several remedies.

Crank-Nicolson↗
DOE OSTI2026

Fantômas unconfined: global QCD fits with Bézier parameterizations

Fantômas is a C++ toolkit for exploring the parametrization dependence of parton distribution functions (PDFs) and other correlator functions in quantum chromodynamics (QCD). Fantômas facilitates the generation of adaptable polynomial parametrizations for PDFs, called metamorphs, to find best-fit PDF solutions and quantify the epistemic uncertainty associated with the parametrizations during their fitting. The method employs Bézier curves as universal approximators for a variety of PDF shapes. Integrated into the xFitter framework for the global QCD analysis, Fantômas provides a foundation for general models of PDFs, while reducing the computational time compared to the approaches utilizing traditional polynomial parametrizations as well as providing an interpretable alternative to neural-network-based models. This paper outlines the structure and practical usage of the Fantômas toolkit, including its inputs, outputs, and implementation within xFitter. It also provides a practical example of using Fantômas for uncertainty quantification as well as the combination of PDF fits into a single ensemble.

Bézier curves↗
DOE OSTI2025

Learning to Simulate Aerosol Dynamics with Graph Neural Networks

Aerosol effects on climate, weather, and air quality depend on characteristics of individual particles, which are tremendously diverse and change in time. Particle-resolved models are the only models able to capture this diversity in particle physiochemical properties, and these models are computationally expensive. As a strategy for accelerating particle-resolved microphysics models, we introduce Graph-based Learning of Aerosol Dynamics (GLAD) and use this model to train a surrogate of the particle-resolved model PartMC-MOSAIC. GLAD implements a Graph Network-based Simulator (GNS), a machine learning framework that has been used to simulate particle-based fluid dynamics models. In GLAD, each particle is represented as a node in a graph, and the evolution of the particle population over time is simulated through learned message passing. Here, we demonstrate our GNS approach on a simple aerosol system that includes condensation of sulfuric acid onto particles composed of sulfate, black carbon, organic carbon, and water. A graph with particles as nodes is constructed, and a graph neural network (GNN) is then trained using the model output from PartMC-MOSAIC. The trained GNN can then be used for simulating and predicting aerosol dynamics over time. Results demonstrate the framework's ability to accurately learn chemical dynamics and generalize across different scenarios, achieving efficient training and prediction times. We evaluate the performance across four scenarios, highlighting the framework's robustness and adaptability in modeling aerosol microphysics and chemistry.

aerosol chemistry dynamics↗
DOE OSTI2025

Tree tensor network hierarchical equations of motion based on time-dependent variational principle for efficient open quantum dynamics in structured thermal environments

In this work, we introduce an efficient method, TTN-HEOM, for exactly calculating the open quantum dynamics for driven quantum systems interacting with highly structured bosonic baths by combining the tree tensor network (TTN) decomposition scheme with the bexcitonic generalization of the numerically exact hierarchical equations of motion (HEOM). The method yields a series of quantum master equations for all core tensors in the TTN that efficiently and accurately capture the open quantum dynamics for non-Markovian environments to all orders in the system–bath interaction. These master equations are constructed based on the time-dependent Dirac–Frenkel variational principle, which isolates the optimal dynamics for the core tensors given the TTN ansatz. The dynamics converges to the HEOM when increasing the rank of the core tensors, a limit in which the TTN ansatz becomes exact. We introduce TENSO, tensor equations for non-Markovian structured open systems, as a general-purpose Python code to propagate the TTN-HEOM dynamics. We implement three general propagators for the coupled master equations: two fixed-rank methods that require a constant memory footprint during the dynamics and one adaptive-rank method with a variable memory footprint controlled by the target level of computational error. We exemplify the utility of these methods by simulating a two-level system coupled to a structured bath containing one Drude–Lorentz component and eight Brownian oscillators, which is beyond what can presently be computed using the standard HEOM. Our results show that the TTN-HEOM is capable of simulating both dephasing and relaxation dynamics of driven quantum systems interacting with structured baths, even those of chemical complexity, with an affordable computational cost.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Design improvements to the SNS ion source and diagnostics

The U.S. Spallation Neutron Source (SNS) is a state-of-the-art neutron scattering facility delivering the world's most intense pulsed-neutron beams to a wide array of instruments which are used to conduct investigations in many fields of science and engineering. The accelerator system is fed by an RF-driven, multicusp, H - ion source which nominally provides pulsed beam currents of 50-60 mA (1ms, 60Hz). This report provides a discussion of ongoing design improvements to the SNS ion source and Low Energy Beam Transport (LEBT) as well as diagnostic upgrades undertaken since the previous ICIS conference. These improvements include (i) simple mechanical modifications to the source outlet aperture which resulted in dramatically increased extracted beam current and comparable or lower emittance at similar beam currents, (ii) design improvements to the LEBT chopper target which will enable full power beam-dumping during physics studies, (iii) refinement of the SNS Allison emittance scanner that has enabled the first reliable LEBT beam measurements at full beam power (65kV, 50-100mA, 1ms, 60Hz) on the SNS ion source test stand and (iv) the implementation of a thermal imaging camera for the monitoring the LEBT electrode temperatures. (v) The design of an advanced Cs system, capable of more efficient Cs utilization with significantly lower Cs losses from the source is also presented. Mechanical details, computational simulations and experimental results are discussed within the context of these improvements.

47 OTHER INSTRUMENTATION↗
DOE OSTI2025

High-Performance Computing Based EMT Simulation: Power Grid with IBRs

Electromagnetic transient (EMT) simulation of power grids with high-fidelity models of inverter-based resources (IBRs) is time-consuming and difficult to scale. The necessity for high-fidelity models of IBRs that incorporate the dynamics of individual inverters within IBRs has been showcased in recent studies. These studies focused on events with partial power reduction in each IBR during a transmission line fault in the power grid. These types of events have been documented in multiple North American Electric Reliability Council (NERC) reports in the past decade. It is imperative then to find solutions to speed-up EMT simulations and scale the size of the region with IBRs studied in EMT simulations. In this paper, a combination of numerical simulation algorithms with high-performance computing techniques are employed in discretization and linear solvers employed in the proposed RE-INTEGRATE EMT simulation platform for power grid with IBRs. For ease of scalability, modular and object-oriented programming is used as these techniques are implemented. Additionally, automation software is developed to convert legacy software codes to the proposed RE-INTEGRATE EMT simulation platform. Thereafter, this platform is evaluated on multi-core central processing units (CPUs). Finally, scale-up tests are performed to showcase the scalability that is possible.

Marthi, Phani Ratna Vanamali [ORNL] (ORCID:0000000↗
DOE OSTI2025

Preparing MPICH for exascale

The advent of exascale supercomputers heralds a new era of scientific discovery, yet it introduces significant architectural challenges that must be overcome for MPI applications to fully exploit its potential. Among these challenges is the adoption of heterogeneous architectures, particularly the integration of GPUs to accelerate computation. Additionally, the complexity of multithreaded programming models has also become a critical factor in achieving performance at scale. The efficient utilization of hardware acceleration for communication, provided by modern NICs, is also essential for achieving low latency and high throughput communication in such complex systems. In response to these challenges, the MPICH library, a high-performance and widely used Message Passing Interface (MPI) implementation, has undergone significant enhancements. Here, this paper presents four major contributions that prepare MPICH for the exascale transition. First, we describe a lightweight communication stack that leverages the advanced features of modern NICs to maximize hardware acceleration. Second, our work showcases a highly scalable multithreaded communication model that addresses the complexities of concurrent environments. Third, we introduce GPU-aware communication capabilities that optimize data movement in GPU-integrated systems. Finally, we present a new datatype engine aimed at accelerating the use of MPI derived datatypes on GPUs. These improvements in the MPICH library not only address the immediate needs of exascale computing architectures but also set a foundation for exploiting future innovations in high-performance computing. By embracing these new designs and approaches, MPICH-derived libraries from HPE Cray and Intel were able to achieve real exascale performance on OLCF Frontier and ALCF Aurora respectively.

Guo, Yanfei [Argonne National Laboratory (ANL), Ar↗
DOE OSTI2025

A Block-Structured Adaptive Mesh Framework to Solve Radiation Transfer Equation in Irregular Embedded Geometries

Radiation transport arises in various scientific, industrial, and medical fields, and understanding its effect in applications is needed to make accurate predictions, safety assessments and performance optimizations. Solving the Radiation Transport Equation (RTE) is challenging due to its integro-differential nature, which involves both differential and integral terms. The differential term describes the change in radiation intensity due to absorption and emission, while the integral term accounts for scattering. The accurate modeling of radiation is further complicated in many applications due to the complex, irregular geometries. Various methods exist for solving the RTE, including the zonal, Monte Carlo, spherical harmonics, discrete ordinates, and finite volume methods. Traditional mesh-based approaches, which rely on structured or unstructured meshes, struggle with irregular geometries due to: a) the difficulty of conforming structured grids to irregular domains, b) challenges in enforcing boundary conditions correctly, and c) the additional computational cost of unstructured mesh methods. This work presents a second-order accurate method for solving the RTE in irregular geometries. The radiation intensity is discretized using the finite-volume method in both spatial and angular directions on regular Cartesian grid blocks. Leveraging the block-structured adaptive mesh refinement (AMR) framework provided by AMReX, our method refines the grid locally to reduce spatial discretization error, ensuring a converged numerical solution while minimizing computational costs elsewhere. A two-stage deferred correction approach is employed: First, a first-order discretization on grid blocks is solved using an algebraic multigrid method in HYPRE. Second, a correction term is applied explicitly to achieve second-order accuracy. The correction term is calculated by approximating the radiation flux on cell faces using a Total Variation Diminishing (TVD) scheme. This approach ensures quick convergence of the multigrid method while preserving higher-order accuracy of the numerical solution. Irregular geometries are resolved as embedded boundaries (EB), resulting in both cut cells and regular cells. In cut cells, we modify the fluxes using face fractions and incorporate additional contributions from EB boundary conditions. To ensure higher-order convergence near the EB interface, the correction term is modified by interpolating the radiation intensity to fictitious ghost points. The implementation takes advantage of modern supercomputers by leveraging AMReX’sMPI/X parallelization strategy where X can be MPI or a GPU accelerator including CUDA, HIP and DPC++. We validate our solver using classical test cases, both with and without EB, demonstrating accuracy and efficiency. Additionally, we analyze the impact of adaptive mesh refinement on solution accuracy and computational cost, highlighting the advantages of our approach for high-resolution radiation transport simulations.

computational fluid dynamics (CFD)↗
DOE OSTI2025

Even Higher-Level Synthesis: An Exploration of AI Hardware Accelerators using HLS4ML

With the rise of artificial intelligence, the popularization of deep learning, and a constantly evolving industry, the demand for flexible and efficient tools has never been greater. As algorithms grow more complex, their runtime and energy consumption increase exponentially. Customized hardware accelerators, long used for specific mathematical operations, remain essential for managing modern applications' computational and power demands. Hardware accelerators can speed up complex computations by orders of magnitude, but their manual design and verification processes are often challenging and time-consuming. High-Level Synthesis (HLS) provides a solution by transforming high-level algorithm descriptions, typically written in C++ or SystemC, into synthesizable RTL suitable for hardware implementation. This approach reduces development time for RTL engineers while offering flexibility beyond what traditional handwritten RTL can provide. We extended this capability to the machine-learning domain with the open-source framework hls4ml, which allows neural networks trained in Python frameworks like Tensorflow or PyTorch to be synthesized into efficient hardware representations for the traditional FPGA and ASIC flows. This breakthrough addresses the growing need for reduced design turnaround and easy verification of ML hardware accelerators with low latency and power efficiency constraints. During this tutorial, we will demonstrate how Python complements HLS by simplifying the ML design process, bridging the gap between software and hardware development. Attendees will explore how we translate neural networks modeled in Python into fixed-point C++ models suitable for HLS workflows. We will dive into strategies like Value-Range Analysis and Quantization-Aware Training, which optimize these designs for deployment and evaluate their accuracy, power consumption, and energy efficiency. To exemplify these concepts, experts from Fermilab will share their experiences applying this technology to high-energy physics experiments, where real-time, low-latency processing is critical. Over the years, Fermilab engineers have demonstrated how deep neural networks, optimized for hardware using hls4ml, can meet the stringent requirements of trigger systems at the CERN Large Hadron Collider. These systems rely on rapid decision-making to process immense data volumes while retaining only the most relevant events for further analysis. The application of hls4ml has also been extended to innovative technologies like smart pixel arrays. These smart pixels integrate ML inference capabilities directly into sensor devices, enabling localized data processing at the pixel level. This approach drastically reduces the need to transmit raw data to external processing units, significantly decreasing power consumption and latency. By embedding neural networks within the pixel architecture, the smart pixels can identify and prioritize relevant data in real time, providing a highly efficient solution for edge computing in scenarios such as particle detectors and imaging systems. Fermilab's work highlights the potential of hardware-accelerated ML in scenarios where both speed and power efficiency are mission-critical. Through this tutorial, attendees will gain valuable insights into the challenges and solutions of deploying ML in hardware. Understanding how HLS and hls4ml streamline the development of neural network-based hardware accelerators is fundamental for the industry's future. Participants will learn how these technologies are shaping the future of AI and scientific computing.

Di Guglielmo, Giuseppe [Fermilab]↗
DOE OSTI2026

PyJMAK: An Open-Source Python Toolkit for Modeling Solid-State Metallurgical Phase Transformations

Accurate prediction of metallurgical phase transformations is an essential basis for autonomous optimization and rapid part qualification. Several methods can be used to estimate the evolution of phase fractions such as JMAK kinetics-based models, phase-field models, thermodynamic models, and data-driven machine learning models. Thermodynamic and phase-field-based methodologies solve multiphysics equations requiring numerous calibration parameters and significant computational resources. As a result, the computation domain is limited to a point or on order of micron-meters. The data-driven models rely on large datasets from experiments and simulations. While the JMAK model only provides information about phase fraction evolution, it can predict this evolution in near real-time using thermal history and thermodynamic data without restriction on the domain. JMAK models have been popularly used by researchers to model phase transformations occuring during additive manufacturing or over arbitrary temperature profiles. Commercial proprietary software such as Abaqus and Ansys or closed-source in-house implementations offer the ability to model JMAK based kinetics to predict phase transformation. However, these software packages are not open-source or freely available for use and development in conjunction with manufacturing machines, sensors, and machine learning algorithms. In addition, the use of the model is restricted by a license token. In contrast, given temperature profiles at multiple points in the domain, this Python-based PyJMAK model can compute phase evolution in parallel due to its stand-alone modular, voxel-based structure, and it can be executed on high-performance computing resources without any license restrictions.

Prabhune, Bhagya [Oak Ridge National Laboratory (O↗
DOE OSTI2024

Multistart algorithm for identifying all optima of nonconvex stochastic functions

Here, we propose a multistart algorithm to identify all local minima of a constrained, nonconvex stochastic optimization problem. The algorithm uniformly samples points in the domain and then starts a local stochastic optimization run from any point that is the "probabilistically best" point in its neighborhood. Under certain conditions, our algorithm is shown to asymptotically identify all local optima with high probability; this holds even though our algorithm is shown to almost surely start only finitely many local stochastic optimization runs. We demonstrate the performance of an implementation of our algorithm on nonconvex stochastic optimization problems, including identifying optimal variational parameters for the quantum approximate optimization algorithm.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Implementation of the Glued Sphere Discrete Element Model for Non-Spherical Particles in MFiX Software

To enhance solver capabilities, simulation flexibility and model validation within the MFiX software, the U.S. Department of Energy (DOE) is funding efforts to develop and integrate the glued-sphere discrete element method into the latest version of MFiX as a dedicated computational module. The glued-sphere discrete element method is a numerical technique to depict the behavior of non-spherical particles in granular flows or particulate systems by representing them as a collection of component spheres. These spheres are bonded together to approximate the shape and mechanical/chemical properties of a more complex particle. The method effectively reuses the existing sphere-sphere collision algorithm, interphase momentum and heat transfer calculations utilized in the traditional discrete element method, extending these capabilities to non-spherical particles. Additionally, this method explicitly resolves intra-particle temperature and species distributions. The MFiX glued-sphere computational module includes tools for generating glued sphere configurations, a dedicated solver, and visualization capabilities in post-processing. More specifically within the computational module, collision detection and calculations were first performed on component spheres and then mapped onto non-spherical particles. The linear spring-dashpot model was utilized to simulate the sphere-sphere interactions.

Ke, Renjie↗
DOE OSTI2025

A fault-tolerant neutral-atom architecture for universal quantum computation

Quantum error correction (QEC) is essential for the realization of large-scale quantum computers. However, owing to the complexity of operating on the encoded ‘logical’ qubits, understanding the physical principles for building fault-tolerant quantum devices and combining them into efficient architectures is an outstanding scientific challenge. Here we use reconfigurable arrays of up to 448 neutral atoms to implement the key elements of a universal, fault-tolerant quantum processing architecture and experimentally explore their underlying working mechanisms. We first use surface codes to study how repeated QEC suppresses errors, demonstrating 2.14(13)x below-threshold performance in a four-round characterization circuit by leveraging atom loss detection and machine learning decoding. We then investigate logical entanglement using transversal gates and lattice surgery and extend it to universal logic through transversal teleportation with three-dimensional [[15,1,3]] codes, enabling arbitrary-angle synthesis with polylogarithmic overhead. Finally, we develop mid-circuit qubit reuse16, increasing experimental cycle rates by two orders of magnitude and enabling deep-circuit protocols with dozens of logical qubits and hundreds of logical teleportations with [[7,1,3]] and high-rate [[16,6,4]] codes while maintaining constant internal entropy. Our experiments show key principles for efficient architecture design, involving the interplay between quantum logic and entropy removal, judiciously using physical entanglement in logic gates and magic state generation, and leveraging teleportations for universality and physical qubit reset. These results establish foundations for scalable, universal error-corrected processing and its practical implementation in neutral atom systems.

atomic and molecular physics↗
DOE OSTI2025

NREL's Journey with HPC in the Cloud and Hybrid Computing

This is a planned lightning talk at the NLIT Summit 2025 conference. This would serve as somewhat of a progress update to the presentation I gave at re:Invent 2024 back in November which can be seen here: https://www.youtube.com/watch?t=2133&v=NMq3kL9qObU&feature=youtu.be (my section begins at the included timestamp value). This presentation discusses our usage of Cloud-hosted HPC systems, and in what circumstances they benefit our researchers strategically. We have been making incremental progress in this area since that recording, so for this presentation I would include our latest experiences and observations as we are beginning to implement a hybrid HPC solution. We're in the midst of a cross-team effort of implementing a prototype hybridization solution which would allow users to strategically burst jobs to the cloud. In this talk for NLIT, I would detail lessons-learned, non-starters, architecture diagrams, and other implementation details that may benefit those interested as we continue our experimentation. Our prototype may not be complete by the time of this presentation, but even in the discovery phase of our anticipated design we've discovered a lot of information from others who have worked on hybrid solutions that are worth sharing.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Shaping the Future of Self-Driving Autonomous Laboratories Workshop

The "Shaping the Future of Self-Driving Autonomous Laboratories" workshop, held in Denver on November 7-8, 2024, brought together leading experts from materials science and computing to address the growing need to revolutionize scientific research through AI-driven autonomous laboratories. The workshop identified critical challenges, including the integration of heterogeneous data, development of AI systems that understand fundamental physical principles, and comprehensive safety protocols. Key recommendations emerged around developing universal laboratory equipment interfaces, implementing automated metadata collection systems, and creating hybrid AI approaches that combine data-driven learning with scientific principles. The workshop emphasized maintaining human oversight while leveraging automation, transforming scientific education to prepare the next generation of researchers, and establishing a national consortium leveraging DOE facilities as anchors for broader collaboration with academia and industry. Participants stressed the urgency of addressing the growing disconnect between human decision-making timescales and modern instrumentation capabilities, highlighting the need for strategic automation while preserving essential human insight and oversight in the research process.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Feasibility study for test rig assessments of fish passage conditions in a Kaplan turbine

The assessment of fish passage conditions in hydroelectric turbines consists of identifying and quantifying physical magnitudes leading to increased risks of injury of fish passing through turbines in operation. Such assessments are usually carried out either with the use of computer-based methods during design or with field testing of live fish and sensors passing through prototypes. A method in between consists of test rig experimentation, which is critical for testing fish-focused design concepts and offers the opportunity for implementing the most effective design measures for improved fish survivability. However, fish-related assessments in test rigs are not sufficiently documented for industrial applications. This work presents the main findings of an experimental campaign to quantify fish-related hydraulic magnitudes in a physical model of a Kaplan turbine in a commercial test rig. Two operating conditions were tested by releasing miniaturized autonomous sensor devices (termed Sensor Fish Mini) at the turbine intake flow, passing them through the runner in motion and recovering them at the draft tube exit. During passage, time series of acceleration, absolute pressure and rotational velocity were recorded. The recordings were then interpreted to determine the magnitude and likely location of hydraulic stressors hazardous to fish. The statistical tests on the reported measurements indicated that low pressure, collisions on the runner and rotations in the draft tube were not different between the two tested operating points. On the other hand, pressure drop and collision rates on the distributor differed considerably as a function of net head. The outcomes of this investigation showed that test rig evaluations of fish-related properties with Sensor Fish Mini can contribute to an evidence-based development of turbine geometries designed for providing safer passage conditions. Future work will investigate the scaling of test rig measurements to hydraulically equivalent magnitudes in the prototype and their biological consequences.

13 HYDRO ENERGY↗
DOE OSTI2025

Identifying High Ionic Conductivity Compositions of Ionic Liquid Electrolytes Using Features of the Solvation Environment

Binary mixtures of ionic liquids with molecular solvents are gaining interest in electrochemical applications due to the improvement in their performance over neat ionic liquids. Dilution with suitable molecular solvents can reduce the viscosity and facilitate faster diffusion of ions, thereby yielding substantially higher ionic conductivity than that for a pure ionic liquid. Although viscosity and diffusion coefficients typically behave as monotonic functions of concentration, ionic conductivity often passes through a peak value at an optimum molar ratio of the molecular solvent to the ionic liquid. The ionic conductivity maximum is generally explained in terms of a balance between the ease of charge transport and the concentration of the charge carriers. In this work, fluctuation in the local environment surrounding an ion is invoked as a plausible explanation for the ionic conductivity mechanism with a binary mixture of 1-ethyl-3-methylimidazolium tetrafluoroborate and ethylene glycol as an example. The magnitude of the dynamism in the local environment is captured by measuring the spatial and temporal features of the solvation environment. Standard deviation in the number of ions in the solvation environment serves as a spatial feature, while the cage correlation lifetimes for oppositely charged ions within the first solvation shell serve as a temporal feature. Large standard deviations in the cluster ion population and short cage correlation lifetimes are indicators of highly dynamic ionic environment at the molecular level and consequently yield high ionic conductivity. Such compositions were found to be in good agreement with the optimum ionic liquid mole fractions obtained through experimental measurement. Short cage correlation lifetimes enable the identification of optimum mixture compositions using simulation trajectories significantly shorter than those required to implement the Nernst–Einstein or Einstein formalisms for calculating ionic conductivity. We validated the applicability of this approach across force fields and in six ionic liquid-molecular solvent electrolytes formed with combination of cations, anions, and solvents. We offer a computationally efficient approach of screening ionic liquid-molecular solvent binary mixture electrolytes to identify molar ratios that yield high ionic conductivity.

25 ENERGY STORAGE↗
DOE OSTI2025

Baseflow Identification via Explainable AI With Kolmogorov‐Arnold Networks

Abstract Hydrological models often involve constitutive laws that may not be optimal in every application. We propose to replace such laws with the Kolmogorov‐Arnold networks (KANs), a class of neural networks designed to identify symbolic expressions. We demonstrate KAN's potential on the problem of baseflow identification, a notoriously challenging task plagued by significant uncertainty. KAN‐derived functional dependencies of the baseflow components on the aridity index outperform their original counterparts; they demonstrate that water availability, rather than potential evapotranspiration, drives baseflow by constraining actual evapotranspiration under arid conditions. On a test set, they increase the Nash‐Sutcliffe efficiency (NSE) by 65%, decrease the root mean squared error by 29%, and increase the Kling‐Gupta efficiency by 34%. This superior performance is achieved while reducing the number of fitting parameters from three to two. Next, we use data from 378 catchments across the continental United States to refine the water‐balance equation at the mean‐annual scale. The KAN‐derived equations based on the refined water balance outperform both the current aridity index model, with up to a 105% increase in NSE, and the KAN‐derived equations based on the original water balance. While the performance of our model and tree‐based machine learning methods is similar, KANs offer the advantage of simplicity and transparency and require no specific software or computational tools. This case study focuses on the aridity index formulation, but the approach is flexible and transferable to other hydrological processes. Plain Language Summary Equations used in hydrologic model are often suboptimal, resulting in reduced prediction accuracy and efficiency. We implemented Kolmogorov‐Arnold networks (KAN), a machine learning algorithm for deriving symbolic formulations, to estimate groundwater recharge and showed that it outperforms an existing state‐of‐the‐art semi‐empirical formulation. In hydrology, Nash‐Sutcliffe efficiency (NSE), root mean squared error (RMSE), and Kling‐Gupta efficiency (KGE) are commonly used to evaluate model performance. Higher NSE and KGE values indicate better performance, while lower RMSE values are preferable. Our results show that NSE increased by 71%, RMSE decreased by 32%, and KGE improved by 25%. In addition, KAN identifies an optimal functional form and can be used to derive new analytical formulas using the prior knowledge. The KAN‐inspired equation outperformed the original formulation and reduced the fitting parameters. Furthermore, we refined the water‐balance equation at the mean‐annual scale and showed that, based on the new water‐balance equation, KAN can derive new formulations that are superior to the original aridity index formulations (up to 105% increase in NSE) and KAN‐derived equations based on the original water balance. These findings highlight the significant potential of KAN to advance the scientific understanding of a wide range of hydrologic processes. Key Points Kolmogorov‐Arnold networks (KANs) enhance interpretability of machine‐learned hydrological models KAN‐derived symbolic formulations outperform state‐of‐the‐art semi‐empirical aridity indices KAN‐identified functional form yields an analytical index with fewer fitting parameters and improved performance

baseflow↗