Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Performance Analysis”

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 325 records · Page 18
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Contextually aware roadside radiation measurement testbed

Here we demonstrate a contextually aware multimodal roadside radiation measurement detection testbed for traffic monitoring applications in nuclear nonproliferation. Many variables in traffic such as vehicle or cargo size, mass, speed, shape, and distance of closest approach can have significant impacts on the radiation measured from a vehicle-transported radiation source. These factors can lead to uncertainties in the analysis of the radiation source, especially for lower-strength radiation sources of interest. Our testbed, known as the Multimodal Measurement System (MMS) uses non-radiation sensors including magnetometers, geophones, radiofrequency receivers, cameras, and LiDAR to extract contextual information about vehicles passing by the system. These contextual data can then be fused with data from radiation measurements to increase the system’s sensitivity and accuracy in nuclear threat detection applications. This work describes the instrumentation of the MMS and its data acquisition pipeline. Furthermore, we describe the pre-analysis performed on the raw multimodal data streams for data fusion, and the high-level machine learning analyses for detection and characterization. The variety of sensors within the MMS provides a valuable testbed that can be used to identify the combinations of contextual sensors that provide the greatest improvements to radiation source detection and characterization within the restrictions for various proliferation detection applications. The MMS is also modular so that additional combinations of sensors can be explored in the future.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2023

Evaluation of longitudinal double-spin asymmetry measurements in semi-inclusive deep-inelastic scattering from the proton for the ECCE detector design

The evaluation of the measurement of double-spin asymmetries for charge-separated pions and kaons produced in deep-inelastic scattering from the proton using the ECCE detector design concept is presented, for the combinations of lepton and hadron beam energies of 5 × 41 GeV 2 and 18 × 275 GeV 2 . The study uses unpolarised simulated data that are processed through a full GEANT simulation of the detector. These data are then reweighted at the parton level with DSSV helicity distributions and DSS fragmentation functions, in order to generate the relevant asymmetries, and subsequently analysed. Furthermore, the performed analysis shows that the ECCE detector concept provides the resolution and acceptance, with a broad coverage in kinematic phase space, needed for a robust extraction of asymmetries. This, in turn, allows for a precise extraction of sea-quark helicity distributions.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2022

Assessing high fidelity multi-component models to facilitate safeguards at Gas Centrifuge Enrichment Plants

We report that the International Atomic Energy Agency (IAEA) inspectors routinely carry out environmental sampling (ES) as a verification method. Collection of environmental swipe samples at various locations in Gas Centrifuge Enrichment Facilities (GCEPs) is an important process in detecting misuse of a declared facility and possibly the existence of undeclared nuclear material. These samples are measured for isotopic composition in uranium containing particles by Thermal Ionization Mass Spectrometry (TIMS) or Inductively Coupled Plasma Mass Spectrometry (ICP-MS). Even though ES is highly effective in detecting the absolute value of enrichments and their deviations from the declared values, it cannot explain the cause of those changes. Several potential explanations can serve as possibilities for particles detected above or other than the declared enrichment. These include normal and non-malicious events such as the design of the enrichment cascades, unintentional failures of the machines, or overshoot during startup of a cascade. It can also be the result of deliberate misuse by the facility operators. The primary objective of this work is to understand how these factors affect the enrichments produced by a cascade and quantify anticipated multi-isotopic concentrations for each case. The following methodology is employed to determine signatures at a particular facility. 1) Utilize a new two-dimensional multi-component diffusion code to obtain centrifuge performance data and use that information to design and perform cascade analysis. Compare and contrast the results with previous 1-D radially averaged solutions from the Pancake code. 2) Design a GCEP cascade with the production goal of 19.75% 235 U for each set of machine data above. Investigate two cascade scenarios that include enrichment of natural uranium (NU) feed to 19.75% 235U in a single cascade compared to a two-step process of NU to 5% and then 5% to 19.75%. 3) Simulate the intentional vs. unintentional off-normal scenarios in the cascades to assess the differences in isotopic concentrations. A non-ideal squared-off cascade model developed at the University of Virginia is used to calculate flow rates and isotopic concentrations of the process gas. The analysis is performed using the Rome machine model operated at 600 m/s rotor speed. The upper and lower bounds of normal and abnormal enrichments in a typical facility are used in conjunction with ES results to understand the root causes of such observations.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗
DOE OSTI2024

Asynchronous GPU-based DEM solver embedded in commercial CFD software with polyhedral mesh support

A novel graphical processing unit-based discrete element method solver is introduced to improve stability, performance, and provide seamless integration into commercial or open-source computational fluid dynamics software. A key innovation is eliminating a need for network communication between solvers, which was previously required for cross-platform coupling. This is accomplished by a direct coupling method that employs dynamic-linked libraries. Furthermore, the solver optimizes memory usage by streamlining the particle-cell search algorithm by eliminating the cells' searching grid. This ensures the solver is compatible with a wide range of mesh types, providing high geometric flexibility. The approach simplifies the simulation process by directly incorporating computational fluid dynamics mesh information into the discrete element method solver. The performance analysis indicates about sixteen times boost in computational speed compared to benchmark central processing unit-based solvers. Finally, the solver's compatibility with polyhedral meshes, a vital advantage for complex geometries, is tested against a referenced study regarding the simulation of an immersed-tube fluidized bed.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2023

Reserve and energy scarcity pricing in United States power markets: A comparative review of principles and practices

Here, errors in forecasting load and renewable-based generation in restructured power systems mean that independent system operators (ISOs) must procure sufficient operating reserves to keep the real-time operation of the system reliable and secure. But when procured reserves turn out to be insufficient in real-time due to the lack of resource capacity or ramp capability, operators often set higher prices for reserves and energy to encourage more supply, and to motivate consumers to decrease usage or shift it to other times. This procedure, which is called scarcity or shortage pricing, is a core feature of U.S. electricity markets. It is receiving increased attention from market designers and stakeholders because scarcity will become more important for spot price formation in the future with the increased penetration of zero-marginal cost renewables, and the shrinking role of fuel costs in setting prices. Scarcity pricing is implemented in various ways by different ISOs. These differences have practical implications for the level of prices and incentives for investment, operations, and demand modification. In this paper, general approaches and specific calculation procedures for reserve and energy scarcity pricing practices and calculations across the seven ISO-based U.S. power markets are reviewed and compared. A consistent terminology is used to facilitate the comparison. Current scarcity pricing practices are grouped into three approaches: (1) imposing an adder after the spot market is run; (2) including stepwise demand curves within market clearing procedures for non-contingency reserve products (e.g., the novel flexiramp product), which tends to yield longer right tails for energy scarcity premium curves; and (3) having stepwise demand curves for traditional contingency reserve products only, which results in shorter right tails in energy scarcity curves. A generic numerical example is presented to highlight the large practical differences among the reserve scarcity pricing approaches and specific implementations. To further investigate factors that contribute the most to demand curves differences among ISOs, a sensitivity analysis is performed. This analysis shows that the largest source of differences among the curves is the scarcity prices assumed in the case of severe scarcity, while the number of steps used and whether flexiramp is considered also yields important differences in scarcity prices. As renewable penetration increases, it will become increasingly crucial to employ administrative demand curves so that spot prices more effectively motivate supply and demand adjustments exactly when and where they are needed. This study shows that the different assumptions yield very different scarcity premiums for reserves and energy, and are likely to provide divergent incentives for resources to respond to shortages. It is concluded that to promote market efficiency, a reserve shortage demand curve should have at least three features: inclusion of the marginal value of reserve products at each shortage level, consideration of the magnitude and probability of supply contingencies, and avoidance of abrupt price discontinuities that can cause excessively volatile market outcomes.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
DOE OSTI2020

TeraChem Cloud: A High-Performance Computing Service for Scalable Distributed GPU-Accelerated Electronic Structure Calculations

The encapsulation and commoditization of electronic structure arise naturally as interoperability, and the use of nontraditional compute resources (e.g., new hardware accelerators, cloud computing) remains important for the computational chemistry community. Here, we present TERACHEM CLOUD, a high-performance computing service (HPCS) that offers on-demand electronic structure calculations on both traditional HPC clusters and cloud-based hardware. The framework is designed using off-the-shelf web technologies and containerization to be extremely scalable and portable. Within the HPCS model, users can quickly develop new methods and algorithms in an interactive environment on their laptop while allowing TERACHEM CLOUD to distribute ab initio calculations across all available resources. This approach greatly increases the accessibility of hardware accelerators such as graphics processing units (GPUs) and flexibility for the development of new methods as additional electronic structure packages are integrated into the framework as alternative backends. Cost-performance analysis indicates that traditional nodes are the most cost-effective long-term solution, but commercial cloud providers offer cutting-edge hardware with competitive rates for short-term large-scale calculations. We demonstrate the power of the TERACHEM CLOUD framework by carrying out several showcase calculations, including the generation of 300,000 density functional theory energy and gradient evaluations on medium-sized organic molecules and reproducing 300 fs of nonadiabatic dynamics on the B800-B850 antenna complex in LH2, with the latter demonstration using over 50 Tesla V100 GPUs in a commercial cloud environment in 8 h for approximately $1250.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

A Computational Study of RNA Tetraloop Thermodynamics, Including Misfolded States

An important characteristic of RNA folding is the adoption of alternative configurations of similar stability, often referred to as misfolded configurations. These configurations are considered to compete with correctly folded configurations, although their rigorous thermodynamic and structural characterization remains elusive. Tetraloop motifs found in large ribozymes are ideal systems for an atomistically detailed computational quantification of folding free energy landscapes and the structural characterization of their constituent free energy basins, including nonnative states. In this work, we studied a group of closely related 10-mer tetraloops using a combined parallel tempering and metadynamics technique that allows a reliable sampling of the free energy landscapes, requiring only knowledge that the stem folds into a canonical A-RNA configuration. Here we isolated and analyzed unfolded, folded, and misfolded populations that correspond to different free energy basins. We identified a distinct misfolded state that has a stability very close to that of the correctly folded state. This misfolded state contains a predominant population that shares the same structural features across all tetraloops studied here and lacks the noncanonical A-G base pair in its loop portion. Further analysis performed with biased trajectories showed that although this competitive misfolded state is not an essential intermediate, it is visited in most of the transitions from unfolded to correctly folded states. Moreover, the tetraloops can transition from this misfolded state to the correctly folded state without requiring extensive unfolding.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2025

Electronic Structure of the Au Nanoparticle-TiO 2 Heterojunction: Influence of Nanoparticle Size, Shape, Oxygen Vacancies, and Temperature

The electronic structure of the gold nanoparticle-titanium dioxide (AuNP-TiO 2 ) heterojunction plays a critical role for charge transfer and recombination dynamics that underpin its photocatalytic function. However, building a representative model to capture the key physics remains a significant challenge. Here, we investigate the influence of the AuNP size and shape, as well as oxygen vacancy (V O ) defects at the anatase-phase TiO 2 (101) surface and the temperature of the heterojunction, on its interfacial electronic properties. Using density functional theory (DFT), we compare the closed-shell Au 20 and open-shell Au 19 clusters interfaced with pristine and V O defect TiO 2 surfaces. We find that the presence of a V O defect transforms pure TiO 2 from a p-type to an n-type semiconductor, reversing the interfacial band bending from downward to upward. For the heterosystem, density of states (DOS) analysis shows that V O minimally affects the open-shell Au 19 -TiO 2 system, but it significantly alters the closed-shell Au 20 -TiO 2 system, converting it from p-type to n-type at the Γ-point at 0 K. Furthermore, ab initio molecular dynamics (AIMD) simulations at 300 K reveal significant thermal fluctuations in AuNP positions relative to the TiO 2 surface. These fluctuations result in dynamic variations in the gap between the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) in all systems studied: Au 19 NP-Pristine TiO 2 , Au 20 NP-Pristine TiO 2 , Au 19 NP-V O defect TiO 2 , and Au 20 NP-V O defect TiO 2 . Complementary Bader charge analysis performed at both 0 K and finite-temperature AIMD snapshots supports the emergence of an upward band bending and the formation of a Schottky barrier at the V O containing heterojunctions. Notably, we find that at finite temperature, an Au atom can dynamically passivate the V O , leading to a pronounced widening of the HOMO-LUMO gap in the AuNP-V O defect TiO 2 heterojunctions. Our computational findings underscore the pivotal role of the V O defect and thermal effects in modulating interfacial band alignment, electronic states, and HOMO-LUMO gaps, providing insights for designing Au-TiO 2 heterojunctions with tailored electronic properties.

Mora Perez, Carlos [Lawrence Berkeley National Lab↗
DOE OSTI2023

A labeled dataset for building HVAC systems operating in faulted and fault-free states

Abstract Open data is fueling innovation across many fields. In the domain of building science, datasets that can be used to inform the development of operational applications - for example new control algorithms and performance analysis methods - are extremely difficult to come by. This article summarizes the development and content of the largest known public dataset of building system operations in faulted and fault free states. It covers the most common HVAC systems and configurations in commercial buildings, across a range of climates, fault types, and fault severities. The time series points that are contained in the dataset include measurements that are commonly encountered in existing buildings as well as some that are less typical. Simulation tools, experimental test facilities, and in-situ field operation were used to generate the data. To inform more data-hungry algorithms, most of the simulated data cover a year of operation for each fault-severity combination. The data set is a significant expansion of that first published by the lead authors in 2020.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2022

Spectral evidence of solar neighborhood analogs in CALIFA galaxies

We introduce a novel nonparametric method to find solar neighborhood analogs (SNAs) in extragalactic integral field spectroscopic surveys. The main ansatz is that the physical properties of the solar neighborhood (SN) should be encoded in its optical stellar spectrum. We assume that our best estimate of such a spectrum is the one extracted from the analysis performed by the Code for Stellar properties Heuristic Assignment (CoSHA) from the MaStar stellar library. It follows that finding SNAs in other galaxies consist in matching, in a χ 2 sense, the SN reference spectrum across the optical extent of the observed galaxies. We applied this procedure to a selection of CALIFA galaxies, by requiring a close to face-on projection, relative isolation, and non-active galactic nucleus. We explore how the local and global properties of the SNAs (stellar age, metallicity, dust extinction, mass-to-light ratio, stellar surface mass density, star-formation density, and galactocentric distance) and their corresponding host galaxies (morphological type, total stellar mass, star-formation rate, and effective radius) compared with those of the SN and the Milky Way (MW). We find that SNAs are located preferentially in S(B)a–S(B)c galaxies, in a ring-like structure, which radii seem to scale with the galaxy size. Despite the known sources of systematics and errors, most properties present a considerable agreement with the literature on the SN. We conclude that the solar neighborhood is relatively common in our sample of SNAs. Our results warrant a systematic exploration of correlations among the physical properties of the SNAs and their host galaxies. We reckon that our method should inform current models of the galactic habitable zone in our MW and other galaxies.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2025

Accelerating detector simulations with Celeritas: Profiling and performance optimizations

Celeritas is a GPU-optimized Monte Carlo (MC) particle transport code designed to meet the growing computational demands of next-generation high energy physics (HEP) experiments. It provides efficient simulation of electromagnetic (EM) physics processes in complex geometries with magnetic fields, detector hit scoring, and seamless integration into Geant4-driven applications to offload EM physics to GPUs. Recent efforts have focused on performance optimizations and expanding profiling capabilities. This paper presents some key advancements, including the integration of the Perfetto system profiling tool for detailed performance analysis and the development of track-sorting methods to improve computational efficiency.

Lund, Amanda [Argonne National Laboratory (ANL)]↗
DOE OSTI2020

Thermal Insulation for a High Temperature Molten Salt Storage Tank in a CSP Plant

A ternary molten chloride has been suggested as a high temperature (550-720 °C) sensible heat storage medium for a two-tank system. An effective thermal insulation is proposed to provide an opportunity for the fabrication of both the hot and cold storage tanks from lower cost metals, e.g. A240-347H, by maintaining the tank wall below 500 °C. To achieve this, the hot storage tank mainly comprises of inner shell (11.4 cm of a refractory brick as liner (KX-99®), 30 cm of insulation brick) and outer metal shell, where 5 cm of frozen layer of PCM800 (NaCl), and 5 cm of ceramic wool fill the annular space between the shells. Another layer of ceramic wool was introduced at the outer side of the metal shell to keep the temperature lower than 60 °C. ANSYS Fluent was used for the numerical study of the cool down behavior of the system during idle period when tank is quarter-filled. The heat transfer modes were considered as radiation and natural convection heat transfer from the hot molten salt at 720 °C, conduction through the solid layers and convection heat transfer from the outer shell to the environment at 15 °C. The results showed that the proposed insulation system including a frozen layer of PCM800 maintains the temperature at the metal wall lower than 500 °C, and lower than 60°C at the outer side of the tank. Apart from the insulation property of the frozen layer of PCM800, it absorbs any leak of hot molten salt, acting as a barrier between salt and the metal shell. The thermal performance analysis provide an insight into the benefits of the multilayer insulation.

41 EE - Solar Energy Technologies Office (EE-4S)↗
DOE OSTI2021

The five line-of-sight neutron time-of-flight (nToF) suite on the National Ignition Facility (NIF)

Measurement of the neutron spectrum from inertial confinement fusion implosions is one of the primary diagnostics of implosion performance. Analysis of the spectrum gives access to quantities such as neutron yield, hot-spot velocity, apparent ion temperature, and compressed fuel ρr through measurement of the down-scatter ratio. On the National Ignition Facility, the neutron time-of-flight suite has been upgraded to include five independent, collimated lines of sight, each comprising a high dynamic range bibenzyl/diphenylacetylene-stilbene scintillator [R. Hatarik et al., Plasma Fusion Res. 9, 4404104 (2014)] and high-speed fused silica Cherenkov detectors [A. S. Moore et al., Rev. Sci. Instrum. 89, 10I120 (2018)].

Moore, A. S. (ORCID:0000000221258882)↗
DOE OSTI2021

The design of a photoionization front experiment using the Z-Machine as a driving source and estimated measurements

Radiation-driven heat fronts are present in the early universe during reionization, the circumstellar medium of supernovae, and in high-energy-density physics experiments. Dedicated experiments to observe and diagnose the behavior of these types of heat fronts can improve our understanding of these phenomena. A simulation study of photoionization fronts using the HELIOS-CR radiation hydrodynamics code provides an experimental design for the Z-Machine at Sandia National Laboratory using a measurement-calibrated input radiation flux to drive the photoionization front. The simulations use detailed atomic physics and non-diffusive radiation transport in 1D to determine an optimal gas pressure of 0.75 atm for an experiment in N gas as well as the effects of increasing the thickness of the window that seals the gas cell. Post-processing of these simulations demonstrates that ratios of atomic rate coefficients place the heat front in a physics regime where photoionization dominates the energy deposition. Here, to see the sensitivity of the simulations to changes in the model and spatial grid, this analysis performed resolution, atomic model detail, and radiation transport angular grid studies showing less than 10% deviation from the nominal model for increased complexity, when possible. An effort to emulate 3D geometric effects on the radiation flux using an artificial attenuation scheme has shown that, even for conservative estimates of the flux, simulations still produce a photoionization front. Estimations of a streaked, visible spectroscopy measurement using SPECT3D showed that line emission measurements are present early in time and that later in time thermal emission should become dominant.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2021

First principles reactive simulation for equation of state prediction

The high cost of density functional theory (DFT) has hitherto limited the ab initio prediction of the equation of state (EOS). In this article, we employ a combination of large scale computing, advanced simulation techniques, and smart data science strategies to provide an unprecedented ab initio performance analysis of the high explosive pentaerythritol tetranitrate (PETN). Comparison to both experiment and thermochemical predictions reveals important quantitative limitations of DFT for EOS prediction and thus the assessment of high explosives. In particular, we find that DFT predicts the energy of PETN detonation products to be systematically too high relative to the unreacted neat crystalline material, resulting in an underprediction of the detonation velocity, pressure, and temperature at the Chapman–Jouguet state. The energetic bias can be partially accounted for by high-level electronic structure calculations of the product molecules. Furthermore we demonstrate a modeling strategy for mapping chemical composition across a wide parameter space with limited numerical data, the results of which suggest additional molecular species to consider in thermochemical modeling.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

A perspective on ordered vacancy compound and parent chalcopyrite thin film absorbers for photoelectrochemical water splitting

Chalcopyrites could fill the gap between the low-cost, poor-efficiency single junction metal oxide photoelectrochemical (PEC) water splitting cells and the high efficiency, yet costly III–V tandems. In this Perspective, we first review the key barriers that must be addressed by the community to enable economical chalcopyrite-based PEC water splitting. Then, we highlight how theoretical modeling can be used to identify promising ordered vacancy compound absorbers with improved energetics compared to their chalcopyrite parents. Finally, we describe how advanced spectroscopic analysis performed on chalcopyrite photocathodes after PEC testing uncovered a new passivation layer candidate for prolonged durability.

08 HYDROGEN↗
DOE OSTI2021

NENCI-2021. I. A large benchmark database of non-equilibrium non-covalent interactions emphasizing close intermolecular contacts

In this work, we present NENCI-2021, a benchmark database of ~8000 Non-Equilibirum Non-Covalent Interaction energies for a large and diverse selection of intermolecular complexes of biological and chemical relevance. To meet the growing demand for large and high-quality quantum mechanical data in the chemical sciences, NENCI-2021 starts with the 101 molecular dimers in the widely used S66 and S101 databases and extends the scope of these works by (i) including 40 cation–π and anion–π complexes, a fundamentally important class of non-covalent interactions that are found throughout nature and pose a substantial challenge to theory, and (ii) systematically sampling all 141 intermolecular potential energy surfaces (PESs) by simultaneously varying the intermolecular distance and intermolecular angle in each dimer. Designed with an emphasis on close contacts, the complexes in NENCI-2021 were generated by sampling seven intermolecular distances along each PES (ranging from 0.7× to 1.1× the equilibrium separation) and nine intermolecular angles per distance (five for each ion–π complex), yielding an extensive database of 7763 benchmark intermolecular interaction energies (E int ) obtained at the coupled-cluster with singles, doubles, and perturbative triples/complete basis set [CCSD(T)/CBS] level of theory. The E int values in NENCI-2021 span a total of 225.3 kcal/mol, ranging from -38.5 to +186.8 kcal/mol, with a mean (median) E int value of -1.06 kcal/mol (-2.39 kcal/mol). In addition, a wide range of intermolecular atom-pair distances are also present in NENCI-2021, where close intermolecular contacts involving atoms that are located within the so-called van der Waals envelope are prevalent—these interactions, in particular, pose an enormous challenge for molecular modeling and are observed in many important chemical and biological systems. A detailed symmetry-adapted perturbation theory (SAPT)- based energy decomposition analysis also confirms the diverse and comprehensive nature of the intermolecular binding motifs present in NENCI-2021, which now includes a significant number of primarily induction-bound dimers (e.g., cation–π complexes). NENCI-2021 thus spans all regions of the SAPT ternary diagram, thereby warranting a new four-category classification scheme that includes complexes primarily bound by electrostatics (3499), induction (700), dispersion (1372), or mixtures thereof (2192). A critical error analysis performed on a representative set of intermolecular complexes in NENCI-2021 demonstrates that the E int values provided herein have an average error of ±0.1 kcal/mol, even for complexes with strongly repulsive E int values, and maximum errors of ±0.2–0.3 kcal/mol (i.e., ~±1.0 kJ/mol) for the most challenging cases. For these reasons, we expect that NENCI-2021 will play an important role in the testing, training, and development of next-generation classical and polarizable force fields, density functional theory approximations, wavefunction theory methods, and machine learning based intra- and inter-molecular potentials.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2022

Representations and strategies for transferable machine learning improve model performance in chemical discovery

Strategies for machine-learning (ML)-accelerated discovery that are general across material composition spaces are essential, but demonstrations of ML have been primarily limited to narrow composition variations. By addressing the scarcity of data in promising regions of chemical space for challenging targets such as open-shell transition-metal complexes, general representations and transferable ML models that leverage known relationships in existing data will accelerate discovery. Over a large set (~1000) of isovalent transition-metal complexes, we quantify evident relationships for different properties (i.e., spin-splitting and ligand dissociation) between rows of the Periodic Table (i.e., 3d/4d metals and 2p/3p ligands). We demonstrate an extension to the graph-based revised autocorrelation (RAC) representation (i.e., eRAC) that incorporates the group number alongside the nuclear charge heuristic that otherwise overestimates dissimilarity of isovalent complexes. To address the common challenge of discovery in a new space where data are limited, we introduce a transfer learning approach in which we seed models trained on a large amount of data from one row of the Periodic Table with a small number of data points from the additional row. We demonstrate the synergistic value of the eRACs alongside this transfer learning strategy to consistently improve model performance. Analysis of these models highlights how the approach succeeds by reordering the distances between complexes to be more consistent with the Periodic Table, a property we expect to be broadly useful for other material domains.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗