Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Computer Modeling”

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 1,369 records · Page 76

A Combined Global and Local Approach to Elucidate Spatial Organization of the Mycobacterial ParB-parS Partition

Combining diverse sets of data at global (size, shape) and local (residue) scales is an emerging trend for elucidating the organization and function of the cellular assemblies. We used such a strategy, combining data from X-ray and neutron scattering with H/D-contrast variation and X-ray footprinting with mass spectrometry, to elucidate the spatial organization of the ParB-parS assembly from Mycobacterium tuberculosis. The ParB-parS participates in plasmid and chromosome segregation and condensation in predivisional bacterial cells. ParB polymerizes around the parS centromere(s) to form a higher-order assembly that serves to recruit cyto-skeletal ParA ATPases and SMC proteins for chromosome segregation. A hybrid model of the ParB-parS was built by combining and correlating computational models with experiment-derived information about size, shape, position of the symmetry axis within the shape, internal topology, DNA-protein interface, exposed surface patches, and prior knowledge. This first view of the ParB-parS leads us to propose how ParB spread on the chromosome to form a larger assembly.

59 BASIC BIOLOGICAL SCIENCES↗

Advances in Grid Computing for the FabrIc for Frontier Experiments Project at Fermialb

The FabrIc for Frontier Experiments (FIFE) project is a major initiative within the Fermilab Scientic Computing Division charged with leading the computing model for Fermilab experiments. Work within the FIFE project creates close collaboration between experimenters and computing professionals to serve high-energy physics experiments of diering size, scope, and physics area. The FIFE project has worked to develop common tools for job submission, certicate management, software and reference data distribution through CVMFS repositories, robust data transfer, job monitoring, and databases for project tracking. Since the projects inception the experiments under the FIFE umbrella have signicantly matured, and present an increasingly complex list of requirements to service providers. To meet these requirements, the FIFE project has been involved in transitioning the Fermilab General Purpose Grid cluster to support a partitionable slot model, expanding the resources available to experiments via the Open Science Grid, assisting with commissioning dedicated high-throughput computing resources for individual experiments, supporting the eorts of the HEP Cloud projects to provision a variety of back end resources, including public clouds and high performance computers, and developing rapid onboarding procedures for new experiments and collaborations. The larger demands also require enhanced job monitoring tools, which the project has developed using such tools as ElasticSearch and Grafana. in helping experiments manage their large-scale production work ows. This group in turn requires a structured service to facilitate smooth management of experiment requests, which FIFE provides in the form of the Production Operations Management Service (POMS). POMS is designed to track and manage requests from the FIFE experiments to run particular work ows, and support troubleshooting and triage in case of problems. Recently a new certicate management infrastructure called Distributed Computing Access with Federated Identities (DCAFI) has been put in place that has eliminated our dependence on a Fermilab-specic third-party Certicate Authority service and better accommodates FIFE collaborators without a Fermilab Kerberos account. DCAFI integrates the existing InCommon federated identity infrastructure, CILogon Basic CA, and a MyProxy service using a new general purpose open source tool. We will discuss the general FIFE onboarding strategy, progress in expanding FIFE experiments presence on the Open Science Grid, new tools for job monitoring, the POMS service, and the DCAFI project.

Herner, K.↗

Multiscale modeling and cinematic visualization of photosynthetic energy conversion processes from electronic to cell scales

We report conversion of sunlight into chemical energy, namely photosynthesis, is the primary energy source of life on Earth. A visualization depicting this process, based on multiscale computational models from electronic to cell scales, is presented in the form of an excerpt from the fulldome show Birth of Planet Earth. This accessible visual narrative shows a lay audience, including children, how the energy of sunlight is captured, converted, and stored through a chain of proteins to power living cells. The visualization is the result of a multi-year collaboration among biophysicists, visualization scientists, and artists, which, in turn, is based on a decade-long experimental-computational collaboration on structural and functional modeling that produced an atomic detail description of a bacterial bioenergetic organelle, the chromatophore. Software advancements necessitated by this project have led to significant performance and feature advances, including hardware-accelerated cinematic ray tracing and instanced visualizations for efficient cell-scale modeling. The energy conversion steps depicted feature an integration of function from electronic to cell levels, spanning nearly 12 orders of magnitude in time scales. This atomic detail description uniquely enables a modern retelling of one of humanity’s earliest stories—the interplay between light and life.

97 MATHEMATICS AND COMPUTING↗

A sub-pore model for multi-scale reaction–diffusion problems in porous media

Applications of reaction–diffusion systems in porous media pose a challenging problem for computational modeling approaches due to their multi-physics and multi-scale nature. The length scales usually span 3–4 orders of magnitude while physical phenomena involved include heat and mass transfer processes, and surface reactions. In this paper, a novel methodology that accounts for all the length scales and physical phenomena involved in a single framework is described. A length scale based dual approach is proposed – the larger pore channels (macro-pores) are resolved using conventional numerical techniques and a novel ‘sub-pore’ model is used to account for the unresolved pore channels (sub-pores) and the important physics therein. The porous network in the sub-pore system is composed of a fractal-like hierarchical system of straight cylindrical pores. Simplified governing equations for mass and energy transport are solved within the sub-pore system along with a reaction kinetics model to account for surface adsorption. An implicit coupling strategy is used to couple the macro-pore and the sub-pore systems so as to ensure conservation. The developed methodology is then applied to a few test cases and it is established that the proposed framework is necessary for problems where the adsorption time scale is much smaller than (diffusion-limited) or comparable to the diffusion time scale. Furthermore, it is also demonstrated that the framework can be potentially used to model the network of porous channels in its entirety thus significantly reducing computational costs.

36 MATERIALS SCIENCE↗

Ocean Energy Systems Wave Energy Modelling Task: Modelling, Verification and Validation of Wave Energy Converters

The International Energy Agency Technology Collaboration Programme for Ocean Energy Systems (OES) initiated the OES Wave Energy Conversion Modelling Task, which focused on the verification and validation of numerical models for simulating wave energy converters (WECs). The long-term goal is to assess the accuracy of and establish confidence in the use of numerical models used in design as well as power performance assessment of WECs. To establish this confidence, the authors used different existing computational modelling tools to simulate given tasks to identify uncertainties related to simulation methodologies: (i) linear potential flow methods; (ii) weakly nonlinear Froude–Krylov methods; and (iii) fully nonlinear methods (fully nonlinear potential flow and Navier–Stokes models). This article summarizes the code-to-code task and code-to-experiment task that have been performed so far in this project, with a focus on investigating the impact of different levels of nonlinearities in the numerical models. Two different WECs were studied and simulated. The first was a heaving semi-submerged sphere, where free-decay tests and both regular and irregular wave cases were investigated in a code-to-code comparison. The second case was a heaving float corresponding to a physical model tested in a wave tank. We considered radiation, diffraction, and regular wave cases and compared quantities, such as the WEC motion, power output and hydrodynamic loading.

16 TIDAL AND WAVE POWER↗

Feature-based Data Assimilation in Geophysics

Many applications in science require that computational models and data be combined. In a Bayesian framework, this is usually done by denying likelihoods based on the mismatch of model outputs and data. However, matching model outputs and data in this way can be unnecessary or impossible. Using large amounts of steady state data is unnecessary because these data are redundant. It is impossible to assimilate data of complex systems into low-dimensional models, or to assimilate data, collected over long time scales, into chaotic models. One can address these issues by denying likelihoods based on suitable low-dimensional \features" of the data, rather than by the mismatch of model output and \raw" data. Our goal is to contribute to a more fundamental understanding of such a feature-based approach to data assimilation. We present noise models based on perturbed observations to de ne feature-based likelihood functions that lead to computationally efficient solutions. We also explain how a feature-based approach enables us to effectively assimilate selected aspects of data into models, and how this can be interpreted by a reduction in effective dimension. Numerical implementation of these ideas are illustrated in four examples.

97 MATHEMATICS AND COMPUTING↗

Computation of the Seismic Stability of Earth Retaining Structures Final Report CRADA No. TSB-954-94

The purpose of this CRADA project was to evaluate the seismic stability of block retaining wall systems. Retaining wall systems are used extensively in private and commercial developments. This study was designed to develop and demonstrate a computer modeling technology to be used to predict the seismic stability of any block wall system design. The nonlinear finite element computer programs developed at LLNL and employed in the Computational Earthquake Initiative were utilized in this small business CRADA to analyze the seismic stability of the block retaining walls. The unique capability of the LLNL programs to rigorously model frictional contact in a dynamic analysis problem were used in a computer simulation of the dynamic interaction the block waiVsoil systems under seismic excitation.

58 GEOSCIENCES↗

Quantification of the transferability of a designed protein specificity switch reveals extensive epistasis in molecular recognition

Re-engineering protein–protein recognition is an important route to dissecting and controlling complex interaction networks. Experimental approaches have used the strategy of “second-site suppressors,” where a functional interaction is inferred between two proteins if a mutation in one protein can be compensated by a mutation in the second. Mimicking this strategy, computational design has been applied successfully to change protein recognition specificity by predicting such sets of compensatory mutations in protein–protein interfaces. To extend this approach, it would be advantageous to be able to “transplant” existing engineered and experimentally validated specificity changes to other homologous protein–protein complexes. Here, we test this strategy by designing a pair of mutations that modulates peptide recognition specificity in the Syntrophin PDZ domain, confirming the designed interaction biochemically and structurally, and then transplanting the mutations into the context of five related PDZ domain–peptide complexes. We find a wide range of energetic effects of identical mutations in structurally similar positions, revealing a dramatic context dependence (epistasis) of designed mutations in homologous protein–protein interactions. To better understand the structural basis of this context dependence, we apply a structure-based computational model that recapitulates these energetic effects and we use this model to make and validate forward predictions. The context dependence of these mutations is captured by computational predictions, our results both highlight the considerable difficulties in designing protein–protein interactions and provide challenging benchmark cases for the development of improved protein modeling and design methods that accurately account for the context.

59 BASIC BIOLOGICAL SCIENCES↗

Modeling Electrochemical Performance of the Hierarchical Morphology of Precious Group Metal-free Cathode for Polymer Electrolyte Fuel Cell

Here, this paper presents a two-dimensional (2D) computational model of a polymer electrolyte fuel cell (PEFC) with a platinum group metal-free (PGM-free) catalyst cathode that can significantly reduce PEFC costs by eliminating the need for expensive platinum catalysts. Due to their comparatively low volumetric activity, PGM-free cathodes are an order of magnitude thicker than their Pt-based counterpart. The resulting need for greater electrode thickness to achieve sufficient power density requires careful attention to the transport losses across the thicker cathodes. The presented model is used to correlate the composition and morphology of the cathode to PEFC performance. The model is a complete cell, continuum model that includes an advanced agglomerate model for a microstructurally consistent representation of the cathode. A unique feature of the approach is the integration of morphology and transport parameter statistics extracted from nano-scale resolution X-ray computed tomography (nano-CT) imaging of PGM-free cathodes. The model was validated with experimental results of PGM-free cathodes with varying Nafion loading. Lastly, our key findings are a need for increased cathode hydrophobicity and increased ionomer conductivity through either reduced tortuosity or increased bulk conductivity. We further use the model to evaluate targets for the volumetric activity and active site density for future catalysts.

36 MATERIALS SCIENCE↗

A Kinetic Approach to Modeling the Manufacture of High Density Polyurethane Structural Foam: Foaming and Polymerization

This report details the study of PMDI structural foam and the development of a computational model to represent it. This polyurethane has a fast catalyst, such that foaming and polymerization occur simultaneously. The foam is over-packed to twice or more of its free rise density to reach the density of interest. Our approach is to combine model development closely with experiments to discover underlying physical phenomena, to parameterize models, and to validate the models once they have been developed. The model must be able to represent the expansion, filling of molds, curing, and final foam properties. PMDI is a chemically blown foam, where carbon dioxide is produced via the reaction of water and isocyanate. The isocyanate also reacts with polyol in a competing reaction, which produces the polymer. A new kinetic model is developed and implemented, which follows a simplified mathematical formalism that decouples these two reactions. The model predicts the polymerization reaction via condensation chemistry, where vitrification and glass transition temperature evolution must be included to correctly predict the temperature dependence of the kinetics. The foam gas generation kinetics are determined by tracking the molar concentration of both water and carbon dioxide. Understanding the thermal history and loads on the foam due to exothermicity and oven curing is very important to the results, since the kinetics and material properties are all sensitive to temperature. The conservation equations, including the equations of motion, an energy balance, and three rate equations are solved via a stabilized finite element method. We assume generalized-Newtonian rheology that is dependent on the extent of cure, gas fraction, and temperature. The conservation equations are combined with a level set method to determine the location of the free surface over time. Results from the model are compared to experimental flow visualization data and post-test X-ray computed tomography (CT) data for the density. Several geometries are investigated including a mock encapsulation part, two configurations of a mock structural part, and a bar geometry to specifically test the density model. We have found that the model predicts both average density and filling profiles well. However, it under predicts density gradients, especially in the gravity direction. Further model improvements are also discussed.

36 MATERIALS SCIENCE↗

A model for the generation of localized transient [Na{sup +}] elevations in vascular smooth muscle

We present a stochastic computational model to study the mechanism of signaling between a source and a target ionic transporter, both localized on the plasma membrane (PM). In general this requires a nanometer-scale cytoplasmic space, or nanodomain, between the PM and a peripheral organelle to reflect ions back towards the PM. Specifically we investigate the coupling between Na{sup +} entry via the transient receptor potential canonical channel 6 (TRPC6) and the Na{sup +}/Ca{sup 2+} exchanger (NCX), a process which is essential for reloading the sarcoplasmic reticulum (SR) via the sarco/endoplasmic reticulum Ca{sup 2+}ATPase (SERCA) and maintaining Ca{sup 2+} oscillations in activated vascular smooth muscle. Having previously modeled the flow of Ca{sup 2+} between reverse NCX and SERCA during SR refilling, this quantitative approach now allows us to model the upstream linkage of Na{sup +} entry through TRPC6 to reversal of NCX. We have implemented a random walk (RW) Monte Carlo (MC) model with simulations mimicking a diffusion process originating at the TRPC6 within PM-SR junctions. The model calculates the average Na{sup +} in the nanospace and also produces profiles as a function of distance from the source. Our results highlight the necessity of a strategic juxtaposition of the relevant ion translocators as well as other physical structures within the nanospaces to permit adequate Na{sup +} build-up to initiate NCX reversal and Ca{sup 2+} influx to refill the SR.

60 APPLIED LIFE SCIENCES↗

Building thermal dynamics modeling with deep transfer learning using a large residential smart thermostat dataset

Understanding thermal dynamics and obtaining the computational model of residential buildings enable its scaled application in energy retrofits, control optimization and decarbonization. In this paper, we present a deep learning approach to model building thermal dynamics with smart thermostat data collected from residential buildings, with the goal to investigate model generalizability. In the first stage, we developed and compared different Deep Learning architectures including Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM) models and CNN-LSTM to predict indoor air temperature in a multi-step time horizon. In the second stage, we implemented a Transfer Learning (TL) process, which aims to improve the prediction performance on a new set of buildings (targets), exploiting the knowledge of related or similar buildings (sources). Different TL strategies and source model identification methods were investigated. The study showed that the CNN-LSTM performed the best among the architectures compared, with an average Mean Absolute Error (MAE) of 0.26 °C for one-hour-ahead (twelve 5-min future steps) predictions. Furthermore, the results showed that freezing the LSTM layer and fine-tuning the other layers of the CNN-LSTM achieved the best performance among four TL strategies, which further improved the performance with respect to a machine learning approach by 10%, and proving the effectiveness and generalizability of the proposed approach. A comparison of three different source model identification methods showed that randomly selecting source models constrained by similar building characteristics can provide good TL performance while retaining simplicity comparing with other quantitative source identification methods.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Unified Wind-Wave Growth and Spectrum Functions for All Water Depths: Field Observations and Model Results

Abstract Wind-wave development is governed by the fetch- or duration-limited growth principle that is expressed as a pair of similarity functions relating the dimensionless elevation variance (wave energy) and spectral peak frequency to fetch or duration. Combining the pair of similarity functions, the fetch or duration variable can be removed to form a dimensionless function of elevation variance and spectral peak frequency, which is interpreted as the wave energy evolution with wave age. The relationship is initially developed for quasi-neural stability and quasi-steady wind forcing conditions. Further analyses show that the same fetch, duration, and wave-age similarity functions are applicable to unsteady wind forcing conditions, including rapidly accelerating and decelerating mountain gap wind episodes and tropical cyclone (TC) wind fields. Here it is shown that with the dimensionless frequency converted to dimensionless wavenumber using the surface wave dispersion relationship, the same similarity function is applicable in all water depths. Field data collected in shallow to deep waters and mild to TC wind conditions and synthetic data generated by spectrum model computations are assembled to illustrate the applicability. For the simulation work, the finite-depth wind-wave spectrum model and its shoaling function are formulated for variable spectral slopes. Given wind speed, wave age, and water depth, the measured and spectrum-computed significant wave heights and the associated growth parameters are in good agreement in forcing conditions from mild to TC winds and in all depths from deep ocean to shallow lake. Significance Statement This paper presents a growth function and spectrum model to describe wind-wave development in all water depths. Their applicability covers a wide range of wind forcing conditions including steady, accelerating, decelerating, and tropical cyclone events. Support for the unified spectrum model and growth function is presented with field observations and numerical computations.

Hwang, Paul A.↗

SURVEY OF COUPLING SCHEMES IN TRADITIONAL COUPLED NEUTRONICS AND THERMAL-HYDRAULICS CODES

The design and the analysis of nuclear power plants (NPPs) require computational codes to predict the behavior of the NPP nuclear components and other systems (i.e., reactor core, primary coolant system, emergency core cooling system, etc.). Coupled calculations are essential to the conduct of deterministic safety assessments. Inasmuch as the physical phenomena that govern the performance of a nuclear reactor are always present simultaneously, ideally computational modeling of a nuclear reactor should include coupled codes that represent all of the active physical phenomena. Such multi-physics codes are under development at several institutions and are expected to become operational in the future. However, in the interim, integrated codes that incorporate modeling capabilities for two to three physical phenomena will remain useful. For example, in the conduct of safety analyses, of paramount importance are codes that couple neutronics and thermal-hydraulics, especially transient codes. Other code systems of importance to safety analyses are those that couple primary system thermal-hydraulics to fission product chemistry, neutronics to fuel performance, containment behavior and structural mechanics to thermal-hydraulics, etc. This paper surveys the methods used traditionally in the coupling of neutronic and thermal-hydraulics codes. The neutron kinetics codes are used for computing the space-time evolution of the neutron flux and, hence, of the power distribution. The thermal-hydraulics codes, which compute mass, momentum and energy transfers, model the coolant flow and the temperature distribution. These codes can be used to compute the neutronic behavior and the thermal-hydraulic states separately. However, the need to account with fidelity for the dynamic feedback between the two sets of properties (via temperature and density effects on the cross section inputs into the neutronics codes) and the requirement to model realistically the transient response of nuclear power plants and to assess the associated emergency systems and procedures imply the necessity of modeling the neutronic and thermal-hydraulics simultaneously within a coupled code system. The focus of this paper is a comparison of the methods by which the coupling between neutron kinetics and thermal-hydraulics treatments has been traditionally achieved in various code systems. As discussed in the last section, the modern approaches to multi-physics code development are beyond the scope of this paper.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

SIPMaP: A Tool for Modeling Irregular Parallel Computations in the Super Instruction Architecture

Performance modeling is becoming an increasingly important part of the parallel application development process, particularly for expensive computations that will be run on very high-end systems where resources are scarce. We describe a performance modeling tool SIPMaP (Super Instruction Processor Modeling and Prediction) developed for the Super-Instruction Architecture (SIA). The SIA is designed for applications where the dominant data structures are large multi-dimensional arrays and it comprises a DSL, the Super-Instruction Assembly Language (SIAL) that supports expressing algorithms in terms of blocks (tiles), and its runtime system Super Instruction Processor (SIP) that manages distribution and disk storage of the arrays. SIPMaP generates performance models from the SIAL source code. In comparison with many applications where useful performance models have been developed and reported, these programs are irregular and have other difficult to model characteristics such as extensive overlapping of communication and computation.

Computer Science↗

Radiometric characterization process for locating radioactivity hold-up and measuring non-uniformly distributed sources - 15342

A novel approach to facility radiometric characterization for non-uniformly distributed radiation sources has been developed by ANTECH and is currently being employed at the Argonne National Laboratory. It is being applied to the measurement of a wide range of objects and the interiors and contents of laboratories and buildings. The integrated non destructive assay measurement process involves the use of a scanning spectroscopic gamma ray camera as well as high purity Germanium detectors with different collimators and hence differing detector fields of view to measure and quantify non-uniformly distributed sources of radioactivity in laboratory equipment, glove boxes, storage tanks, pipework and general laboratory objects and process components. The purpose of the radiometric characterization is to quantify radiation sources and radioactive contamination levels in facilities containing nuclear material and to ensure compliance with facility characterization requirements and for decontamination and decommissioning planning and execution. The integrated measurement process also involves automated data capture and processing software and a database with which both measurement instrument operators and subject matter expert analysts can interact. An appropriate measuring instrument (gamma camera, scintillation detector, high purity Germanium detector system, neutron detector or alpha-beta counter) records the measurement data and the measurement site is recorded in a photograph. All of the data relevant to the measurement (or set of measurements for a particular object) are recorded in the database including measured gamma ray spectra, equipment description and relevant calibration, measurement site photograph and measurement descriptive data. Once data analysis protocols are established for specific geometries and measurement situations they can be applied to the measured data in a partially automated manner, reducing analysis and transcription errors. In addition to measurements and data analysis the process also utilises radiometric modelling computer codes including MicroShield and Monte Carlo Neutrons Photons in order to perform model calculations where appropriate. In a typical measurement, the ANTECH RadSearch spectroscopic gamma camera is utilized to find areas of concentrated activity so that these areas can be measured and calculated separately from areas with diffuse contamination. The use of the gamma camera in scanning mode helps to ensure that the assay does not overestimate the detected radiation source, which is a frequent consequence of using standard far field geometry and attenuation analysis codes that assume uniformity of both density and radioactive source distribution. Once the hot spots are identified, high purity Germanium detector measurements are made using various collimators with wide or narrow fields of view as required. These measurements are then processed through a semi-automatic calculation process assuming an appropriate source configuration as identified by RadSearch. These source configurations include point sources, line sources or surface sources. The Generalized Geometry Hold-up algorithms are employed to determine an estimate of the activity of the various sources that are identified in a measurement. The ANTECH integrated measurement process provides an automated calculation procedure where the analyst can select the source type, geometry and appropriate input data, and then the software calculates the activity by radionuclide for all sources identified in the measurement. The measurement and analysis process is controlled under the ANTECH ISO9001: 2008 Quality Assurance System and managed by a site specific Quality Assurance Plan. Both Argonne National Laboratory and Department of Energy technical reviewers have vetted the Integrated Measurement and Analysis Process. The application of the measurement and analysis process is described and measurement examples are provided. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Reduced‐Order Modeling of Energetic Materials Using Physics‐Aware Recurrent Convolutional Neural Networks in a Latent Space (LatentPARC)

Physics-aware deep learning (PADL) has gained popularity for use in spatiotemporal dynamics simulations, such as those in computational modeling of energetic materials (EM). We show that the challenge PADL methods face while learning complex field evolution problems can be simplified and accelerated by decoupling it into two tasks: learning complex geometric features in evolving fields and modeling dynamics over these features in a lower-dimensional feature space. We build upon our previous work on physics-aware recurrent convolutional neural networks (PARC). PARC embeds knowledge of underlying physics into its neural network architecture for more robust and accurate prediction of evolving physical fields. PARC was shown to effectively learn complex nonlinear features such as the formation of hotspots and coupled shock fronts in various initiation scenarios of EMs, as a function of microstructures, serving effectively as a microstructure-aware burn model. Here, we further accelerate PARC and reduce its computational cost by projecting the original dynamics onto a lower-dimensional invariant manifold, or “latent space.” The projected latent representation encodes the complex geometry of evolving fields (e.g., temperature and pressure) in a set of data-driven features. The reduced dimension of this latent space allows us to learn the dynamics during the initiation of EM with a lighter and more efficient model. We observe a significant decrease in training and inference time while maintaining results comparable to PARC at inference. This work takes steps towards enabling rapid prediction of EM thermomechanics at larger scales and characterization of EM structure–property–performance linkages at a full application scale.

Mathematics and Computing↗

Making machine learning a useful tool in the accelerated discovery of transition metal complexes

Abstract As machine learning (ML) has matured, it has opened a new frontier in theoretical and computational chemistry by offering the promise of simultaneous paradigm shifts in accuracy and efficiency. Nowhere is this advance more needed, but also more challenging to achieve, than in the discovery of open‐shell transition metal complexes. Here, localized d or f electrons exhibit variable bonding that is challenging to capture even with the most computationally demanding methods. Thus, despite great promise, clear obstacles remain in constructing ML models that can supplement or even replace explicit electronic structure calculations. In this article, I outline the recent advances in building ML models in transition metal chemistry, including the ability to approach sub‐kcal/mol accuracy on a range of properties with tailored representations, to discover and enumerate complexes in large chemical spaces, and to reveal opportunities for design through analysis of feature importance. I discuss unique considerations that have been essential to enabling ML in open‐shell transition metal chemistry, including (a) the relationship of data set size/diversity, model complexity, and representation choice, (b) the importance of quantitative assessments of both theory and model domain of applicability, and (c) the need to enable autonomous generation of reliable, large data sets both for ML model training and in active learning or discovery contexts. Finally, I summarize the next steps toward making ML a mainstream tool in the accelerated discovery of transition metal complexes. This article is categorized under: Electronic Structure Theory > Density Functional Theory Software > Molecular Modeling Computer and Information Science > Chemoinformatics

Kulik, Heather J.↗