Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Average Accuracy”

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 307 records · Page 17
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Shape Evolution of C75 Large-Grain Niobium Half-Cells During Cavity Fabrication

The largely anisotropic deformation of large-grain Nb discs during deep drawing into half-cells poses a challenge for achieving a desired shape accuracy. Two 5-cell cavities for the C75 CEBAF cryomodule rework program have been fabricated at Jefferson Lab from large-grain Nb discs directly sliced from an ingot. The shape of the inner surface of eight half-cells has been inspected using a FARO Edge laser scanner during the fabrication process and compared to the reference shape. On average, approximately 63% of the half-cell inner surface was found to be within 0.1 mm of the reference shape and ~90% to be within 0.2 mm, after the final equator machining. Several 5-cell C75 cavities have also been fabricated at Research Instruments, Germany, and measurements of the shape accuracy using a Zeiss 3D coordinate measuring machine gave similar results. One half-cell was measured both at Research Instruments and Jefferson Lab for comparison.

Ciovati, G.↗
DOE OSTI2022

Machine learning for impurity charge-state transition levels in semiconductors from elemental properties using multi-fidelity datasets

Quantifying charge-state transition energy levels of impurities in semiconductors is critical to understanding and engineering their optoelectronic properties for applications ranging from solar photovoltaics to infrared lasers. While these transition levels can be measured and calculated accurately, such efforts are time-consuming and more rapid prediction methods would be beneficial. Here, we significantly reduce the time typically required to predict impurity transition levels using multi-fidelity datasets and a machine learning approach employing features based on elemental properties and impurity positions. We use transition levels obtained from low-fidelity (i.e., local-density approximation or generalized gradient approximation) density functional theory (DFT) calculations, corrected using a recently proposed modified band alignment scheme, which well-approximates transition levels from high-fidelity DFT (i.e., hybrid HSE06). Further, the model fit to the large multi-fidelity database shows improved accuracy compared to the models trained on the more limited high-fidelity values. Crucially, in our approach, when using the multi-fidelity data, high-fidelity values are not required for model training, significantly reducing the computational cost required for training the model. Our machine learning model of transition levels has a root mean squared (mean absolute) error of 0.36 (0.27) eV vs high-fidelity hybrid functional values when averaged over 14 semiconductor systems from the II–VI and III–V families. As a guide for use on other systems, we assessed the model on simulated data to show the expected accuracy level as a function of bandgap for new materials of interest. Finally, we use the model to predict a complete space of impurity charge-state transition levels in all zinc blende III–V and II–VI systems.

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

The influence of tillage and fertilizer on the flux and source of nitrous oxide with reference to atmospheric variation using laser spectroscopy

Nitrous oxide (N 2 O) is the third most important long-lived greenhouse gas and agriculture is the largest source of N 2 O emissions. Curbing N 2 O emissions requires understanding influences on the flux and sources of N 2 O. We measured flux and evaluated microbial sources of N 2 O using site preference ( S P ; the intramolecular distribution of 15 N in N 2 O) in flux chambers from a grassland tilling and agricultural fertilization experiments and atmosphere. We identified values greater than that of the average atmosphere to reflect nitrification and/or fungal denitrification and those lower than atmosphere as increased denitrification. Our spectroscopic approach was based on an extensive calibration with 18 standards that yielded S P accuracy and reproducibility of 0.7 ‰ and 1.0 ‰, respectively, without preconcentration. Chamber samples from the tilling experiment taken ~ monthly over a year showed a wide range in N 2 O flux (0–1.9 g N 2 O-N ha -1 d -1 ) and S P (- 1.8 to 25.1 ‰). Flux and S P were not influenced by tilling but responded to sampling date. Large fluxes occurred in October and May in no-till when soils were warm and moist and during a spring thaw, an event likely representing release of N 2 O accumulated under snow cover. These high fluxes could not be ascribed to a single microbial process as S P differed among chambers. However, the year-long S P and flux data for no-till showed a slight direct relationship suggesting that nitrification increased with flux. The comparative data in till showed an inverse relationship indicating that high flux events are driven by denitrification. Corn ( Zea mays ) showed high fluxes and S P values indicative of nitrification ~ 4 wk after fertilization with subsequent declines in S P indicating denitrification. Although there was no effect of fertilizer treatment on flux or S P in switchgrass ( Panicum virgatum) , high fluxes occurred ~1 month after fertilization. In both treatments, S P was indicative of denitrification in many instances, but evidence of nitrification/fungal denitrification also prevailed. At 2 m atmospheric N 2 O S P had a range of 31.1 ‰ and 14.6 ‰ in the grassland tilling and agricultural fertilization experiments, respectively. These data suggest the influence of soil microbial processes on atmospheric N 2 O and argue against the use of the global average atmospheric S P in isotopic modeling approaches.

Environmental Sciences & Ecology↗
DOE OSTI2022

Convolutional neural networks for intra-hour solar forecasting based on sky image sequences

Accurate and timely solar forecasts play an increasingly critical role in power systems. Compared to longer forecasting timescales, very short-term solar forecasting has lagged behind in both research and practice. In this paper, we propose deep convolutional neural networks (CNNs) to provide operational intra-hour (10-minute-ahead to 60-minute-ahead) solar forecasts. We develop two CNN structures inspired by a widely-used CNN architecture. The CNNs are tailored to our solar forecasting regression tasks and rely solely on sky image sequences. Case studies based on six years of data (over 150,000 data points) demonstrate that the best CNN model has forecast skill scores of 20%-39% over the naive persistence of cloudiness benchmark, even at these very short timescales. The CNNs also have consistently superior performance when compared to shallow machine learning models with meteorological predictors, where the improvement averages around 7%. The sensitivity analyses show that the sky image length, resolution, and weather conditions have impacts on the deep learning model accuracy. In our intra-hour problem with specific setups, two sky images with a 10-minute 128 x 128 resolution yield the most accurate forecasts. Current limitations, future work, and deployment challenges and solutions are also discussed.

14 SOLAR ENERGY↗
DOE OSTI2024

Machine learning models of intermittent operation of RO wellhead water treatment for salinity reduction and nitrate removal

Machine learning models were developed for intermittent multi-mode operation of a wellhead reverse osmosis water purification and desalination system to predict salt passage, nitrate passage, and permeate flux. The models, based on long short-term memory (LSTM) recurrent neural network (RNN) architecture, included an attention mechanism to increase model performance in proximity of the regulatory limit for nitrate. Training and testing of the models for the Startup, Production, Shutdown and Flushing operational modes were based on operational data (consisting of 22 process variables per data sample) acquired every 2–5 s over a six-month period. The significant sets of model input attributes for the different operational modes were assessed via Spearman ranking correlation, Self-Organizing Map (SOM) analysis and feed forward feature selection (FFFS). Although the variability of nitrate passage, salt passage and permeate flux was significant over the four operational modes, prediction performance for the three outcomes were with R2 and Average Absolute Relative Error (AARE) of 0.78–0.95 and 2.96–6.16 %, respectively. Model updates post membrane elements replacement demonstrated similar levels of prediction accuracy. The study results suggest that there is merit in exploring the utility of multi-mode models for sensor fault detection, data imputation, and for potential use in model-predictive control.

Intermittent RO operation↗
DOE OSTI2026

Unsupervised anomaly detection in MeV ultrafast electron diffraction

MeV ultrafast electron diffraction (MUED) is a pump-probe technique used to study the dynamic structural evolution of materials. An ultrashort laser pulse triggers structural changes, which are then probed by an ultrashort relativistic electron beam. To overcome low signal-to-noise ratios, diffraction patterns are averaged over thousands of shots. However, shot-to-shot instabilities in the electron beam can distort individual patterns, introducing uncertainty. Improving MUED accuracy requires detecting and removing these anomalous patterns from large datasets. In this work, we developed a fully unsupervised methodology for the detection of anomalous diffraction patterns. Using a convolutional autoencoder, we calculate the reconstruction mean squared error of the diffraction patterns. Based on the statistical analysis of this error, we provide the user an estimation of the probability that the pattern is normal, which also allows a posterior visual inspection of the images that are difficult to classify. This method has been trained with only 100 diffraction patterns and tested on 1521 patterns, resulting in a false positive rate between 0.2% and 0.4%, with a training time of 10 s per image and a test time of about 1 s per image. Here, the proposed methodology can also be applied to other diffraction techniques in which large datasets are collected that include faulty images due to instrumental instabilities.

43 PARTICLE ACCELERATORS↗
DOE OSTI2024

Accurate, Uncertainty-Aware Classification of Molecular Chemical Motifs from Multimodal X-ray Absorption Spectroscopy

Accurate classification of molecular chemical motifs from experimental measurement is an important problem in molecular physics, chemistry, and biology. In this work, we present neural network ensemble classifiers for predicting the presence (or lack thereof) of 41 different chemical motifs on small molecules from simulated C, N, and O K-edge X-ray absorption near-edge structure (XANES) spectra. Our classifiers not only achieve class-balanced accuracies of more than 0.95 but also accurately quantify uncertainty. Here, we also show that including multiple XANES modalities improves predictions notably on average, demonstrating a “multimodal advantage” over any single modality. In addition to structure refinement, our approach can be generalized to broad applications with molecular design pipelines.

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

Analytical comparisons of handheld LIBS and XRF devices for rapid quantification of gallium in a plutonium surrogate matrix

This work compares a portable laser-induced breakdown spectroscopy (LIBS) analyzer to a portable X-ray fluorescence (XRF) device for quantification of gallium (Ga) in a plutonium surrogate matrix of cerium (Ce) for the first time. Calibration methods are developed with spectra of Ce–Ga samples from both devices. Here, metrics such as limit of detection (LoD) and mean average percent error (MAPE) are examined to evaluate calibration performance. While the portable LIBS device can yield a nearly instantaneous analytical measurement, its accuracy is hampered by self-absorption. By employing a self-absorption correction and increasing gating delay, LIBS calibrations with errors in the low single percents and LoDs of 0.1% Ga were constructed. The XRF device produces calibrations with superlative sensitivity, yielding LoDs for gallium in the low tens of parts-per-million (ppm), two orders of magnitude lower than the corrected LIBS models. However, a clear trade-off of measurement fidelity is established between the instantaneous analysis of the LIBS device and the minutes-long XRF measurement yielding superior detection limits.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2020

Line curvature algorithm in laser ektacytometry of red blood cells

The iso-intensity line curvature algorithm in laser ektacytometry of red blood cells is investigated by numerical simulation. The algorithm is designed to measure the average deformability, as well as the width and asymmetry of the red blood cell deformability distribution in a blood sample under study. The accuracy and scope of the algorithm are determined. Using a bimodal ensemble as an example, the possibility of determining the fraction of weakly deformable red blood cells in a blood sample by laser ektacytometry is demonstrated. (laser medicine)

60 APPLIED LIFE SCIENCES↗
DOE OSTI2022

Towards Elucidating Structure–Spectra Relationships in Rhamnogalacturonan II: Computational Protocols for Accurate 13C and 1H Shifts for Apiose and Its Borate Esters

Apiose is a naturally occurring, uncommon branched-chain pentose found in plant cell walls as part of the complex polysaccharide Rhamnogalacturonan II (RG-II). The structural elucidation of the three-dimensional structure of RG-II by nuclear magnetic resonance (NMR) spectroscopy is significantly complicated by the ability of apiose to cross-link via borate ester linkages to form RG-II dimers. Here, we developed a computational approach to gain insight into the structure–spectra relationships of apio–borate complexes in an effort to complement experimental assignments of NMR signals in RG-II. Our protocol involved structure optimizations using density functional theory (DFT) followed by isotropic magnetic shielding constant calculations using the gauge-invariant atomic orbital (GIAO) approach to predict chemical shifts. We evaluated the accuracy of 23 different functional–basis set (FBS) combinations with and without implicit solvation for predicting the experimental 1 H and 13 C shifts of a methyl apioside and its three borate derivatives. The computed NMR predictions were evaluated on the basis of the overall shift accuracy, relative shift ordering, and the ability to distinguish between dimers and monomers. We demonstrate that the consideration of implicit solvation during geometry optimizations in addition to the magnetic shielding constant calculations greatly increases the accuracy of NMR chemical shift predictions and can correctly reproduce the ordering of the 13 C shifts and yield predictions that are, on average, within 1.50 ppm for 13 C and 0.12 ppm for 1 H shifts for apio–borate compounds.

BASIC BIOLOGICAL SCIENCES,BIOMASS FUELS↗
DOE OSTI2022

Calibration and Localization of Optically Pumped Magnetometers Using Electromagnetic Coils

In this paper, we propose a method to estimate the position, orientation, and gain of a magnetic field sensor using a set of (large) electromagnetic coils. We apply the method for calibrating an array of optically pumped magnetometers (OPMs) for magnetoencephalography (MEG). We first measure the magnetic fields of the coils at multiple known positions using a well-calibrated triaxial magnetometer, and model these discreetly sampled fields using vector spherical harmonics (VSH) functions. We then localize and calibrate an OPM by minimizing the sum of squared errors between the model signals and the OPM responses to the coil fields. We show that by using homogeneous and first-order gradient fields, the OPM sensor parameters (gain, position, and orientation) can be obtained from a set of linear equations with pseudo-inverses of two matrices. The currents that should be applied to the coils for approximating these low-order field components can be determined based on the VSH models. Computationally simple initial estimates of the OPM sensor parameters follow. As a first test of the method, we placed a fluxgate magnetometer at multiple positions and estimated the RMS position, orientation, and gain errors of the method to be 1.0 mm, 0.2°, and 0.8%, respectively. Lastly, we calibrated a 48-channel OPM array. The accuracy of the OPM calibration was tested by using the OPM array to localize magnetic dipoles in a phantom, which resulted in an average dipole position error of 3.3 mm. The results demonstrate the feasibility of using electromagnetic coils to calibrate and localize OPMs for MEG.

47 OTHER INSTRUMENTATION↗
DOE OSTI2023

IP Protection in TinyML

Tiny machine learning (TinyML) is an essential component of emerging smart microcontrollers (MCUs). How- ever, the protection of the intellectual property (IP) of the model is an increasing concern due to the lack of desktop/server-grade resources on these power-constrained devices. In this paper, we propose STML, a system and algorithm co-design to Secure IP of TinyML on MCUs with ARM TrustZone. Our design jointly optimizes memory utilization and latency while ensuring the security and accuracy of emerging models. We implemented a prototype and benchmarked with 7 models, demonstrating STML reduces 40% of model protection runtime overhead on average.

42 ENGINEERING↗
DOE OSTI2020

Short-depth circuits for efficient expectation-value estimation

The evaluation of expectation values Tr[$\rho \mathcal{O}$] for some pure state $\rho$ and Hermitian operator $\mathcal{O}$ is of central importance in a variety of quantum algorithms. Near-optimal techniques have been developed in the past and require a number of measurements N approaching the Heisenberg limit N = O(1/ ε) as a function of target accuracy ε. The use of quantum phase estimation (QPE) requires, however, long circuit depths C = O(1/ ε) making its implementation difficult on near-term noisy devices. The more direct strategy of operator averaging is usually preferred as it can be performed using N = O(1/ε 2 ) measurements and no additional gates aside from those needed for the state preparation. In this work we use a simple but realistic model to describe the bound state of a neutron and a proton (the deuteron) to show that the latter strategy can require an overly large number of measurements in order to achieve a prefixed relative target accuracy ε r . Further, we propose to overcome this problem using a single step of QPE and classical postprocessing. This approach leads to a circuit depth C = O (ε μ ) (with μ ≥ 0) and to a number of measurements N = O (1/ε 2 + ν ) for 0 < ν ≤ 1 and a much smaller prefactor. We provide detailed descriptions of two implementations of our strategy for ν = 1 and ν ≈ 0.5 and derive appropriate conditions that a particular problem instance has to satisfy in order for our method to provide an advantage.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2021

Daily grass reference evapotranspiration with Meteosat Second Generation shortwave radiation and reference ET products

Here we assess the accuracy of estimating daily grass reference evapotranspiration (PM-ET o ) using daily shortwave radiation (R s ) and reference evapotranspiration (ET REF ) products provided by the Meteosat Second Generation (MSG) geostationary satellite delivered by the Satellite Applications Facility on Land Surface Analysis (LSA-SAF) framework. The accuracy of using reanalysis ERA5 shortwave radiation data (R s ERA5 ) provided by the European Center for Medium-Range Weather Forecasts (ECMWF) is also evaluated. The assessments were performed using observed weather variables at 37 weather stations distributed across continental Portugal, where climate conditions range from semi-arid to humid, and 12 weather stations located in Azores islands, characterized by humid, windy and often cloudy conditions. This study&#x27;s use of data from a variety of climate conditions contributed to a unique and innovative assessment of the usability of LSA-SAF and ERA5 products for ET o estimation. The first assessment focused on comparing LSA-SAF estimates of R s (R s LSA-SAF ) against ground stations (R s ground ). The results showed a good matching between the two R s data sets for continental Portugal but a tendency for R s LSA-SAF to under-estimate R s ground in the cloudy islands of Azores. ET o values computed using R s LSA-SAF data and observed temperature, humidity and wind speed (ET o LSA-SAF ) were then compared with PM-ET o estimates with ground-based data, which were used as benchmark; input data of temperature and humidity needed for PM-ET o were quality checked for surface aridity effects. It was observed that ET o LSA-SAF is strongly correlated with PM-ET o (R 2 > 0.97) for most locations in continental Portugal, with regression coefficient of a linear regression forced to the origin ranging between 0.95 and 1.05, mean root mean square error (RMSE) of 0.13 mm d –1 , and Nash and Sutcliff efficiency of modeling (EF) above 0.95. For most Azores locations, ET o LSA-SAF over-estimated PM-ET o . This is likely a consequence of the high spatio-temporal heterogeneity of weather conditions that occur in these oceanic islands together with the different footprints of satellite (averaged over the pixel) and station observations. Reanalysis ERA5 shortwave radiation data presented similar behavior to the LSA-SAF products, however with slightly lower accuracy. The daily LSA-SAF ET REF product (ET REF LSA-SAF ) was assessed and results have shown a good accuracy of this product, with acceptable RMSE and high EF values, for continental Portugal but a low accuracy for the Azores islands. A simplified bias correction approach was shown to improve both ET o derived from the LSA-SAF products, namely for Azores stations, which seem to be representative of smaller areas. The use of the FAO-PM temperature approach (PMT) was also assessed using the R s LSA-SAF and R s ERA5 data, which showed a superiority of the LSA-SAF product for ET o estimations (ET o PMT LSA-SAF ). No significant differences (p < 0.05) were observed in terms of the median value of the RMSE when adopting ET o PMT and ET REF LSA-SAF . Differently, results showed that using the R s LSA-SAF in the PMT approach (ET o PMT LSA-SAF ) produces significantly better RMSE results than ET o PMT and ET REF LSA-SAF . Overall, the performed assessment allows concluding that the use of R s LSA-SAF, and to a lesser extent the use of the R s ERA5 , highly improves the accuracy of computation of ET o when R s observations are not available, including when only temperature data are accessible. The use of the ET REF LSA-SAF product is a good alternative when observed weather data are not available.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Numerical modeling of chemical looping oxidative dehydrogenation of ethane in parallel packed beds

Chemical looping oxidative dehydrogenation (CL-ODH) of ethane has the potential to be a highly efficient alternative to steam cracking for ethylene production. Accurate reactor modeling is of critical importance to efficiently scale up and optimize this new technology. This study reports a one-dimensional, heterogeneous packed bed model to simulate the CL-ODH of ethane to ethylene with a Na 2 MoO 4 -promoted CaTi 0.1 Mn 0.9 O 3 redox catalyst. Here, the overall reaction kinetics was well-described by coupling the gas-phase steam cracking of ethane with the reduction kinetics of the redox catalyst by H 2 and C 2 H 4 . The impact of H 2 on the formation rate of CO 2 byproduct from C 2 H 4 conversion was also thoroughly investigated to validate the applicability of the kinetic model under operational environments. The temperature variation within the different CL-ODH steps and the temperature distribution along the bed were also carefully considered. The accuracy of the model was validated by experiments conducted in a large lab-scale packed bed reactor (200 g catalyst loading), with an average deviation of 2.8% in terms of ethane conversion and ethylene yield. The model was subsequently used to optimize the operating parameters of the CL-ODH reactor, indicating that up to 63.7% single-pass C 2 +olefin yield can be achieved with the current redox catalyst bed whereas further optimization of the redox catalyst to inhibit C 2 H 4 activation can result in 69.4% single-pass C 2 +yield while maintaining low CO 2 selectivity.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2026

Single-Particle Isotopic Analysis Using the Liquid Sampling-Atmospheric Pressure Glow Discharge Microplasma Coupled to an Orbitrap Mass Spectrometer

Isotope ratio (IR) determinations on individual particles can provide valuable information relative to the sample, including processing and dating, which could be valuable to geological and nuclear material analysis communities. In comparison to “bulk” measurements, in which all particle information is lost and homogenized within a sample, “particle” measurements allow for fingerprinting and can provide unique insights related to the sample’s nano- and micro- compositions. Furthermore, the complex nature of particles, with potential isobaric interferences from either the matrix or other elements within the same particle, poses a significant analytical challenge for conventional mass spectrometry platforms employed for elemental analysis due to their limited mass resolution. Presented here is the first demonstration of an ultrahigh resolution method for single particle (SP) isotopic analysis using the liquid sampling-atmospheric pressure glow discharge (LS-APGD) microplasma ionization source coupled to an Orbitrap mass spectrometer and FTMS Booster X2T acquisition/processing system. For proof of concept, well-characterized CeO 2 microparticles with a nominal diameter of 1.0 μm were analyzed at a mass resolution of ∼330,000. Important instrumental parameters were optimized to enable the detection of single particles. The 142 Ce/ 140 Ce ratio of individual particles and the population average were determined and compared to an SP inductively coupled plasma time-of-flight mass spectrometry (ICP-TOF-MS) analysis. The isotopic accuracy and precision were determined by two methods and found to be in good agreement with the SP-ICP-TOF-MS method, with the linear regression slope method providing a more accurate ratio, showing a ∼1.4% relative difference from the ratio determined from a particle digest. The encouraging performance of the method was further supported by a determined detection limit of 2.9 fg ( 142 Ce). The developed method is anticipated to overcome many of the challenges posed by isobaric interferences encountered in conventional mass spectrometry techniques when analyzing real-world samples for particle isotopic composition.

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

Real-time elemental analysis of liquids for process monitoring using laser-induced breakdown spectroscopy with a liquid wheel sampling approach

This article presents an engineered sampling system that used a rotating wheel to form a thin liquid layer, permitting the use of laser-induced breakdown spectroscopy (LIBS) for in situ, real-time elemental impurity quantification during liquid processing. The sampling approach was demonstrated on eight elements from across the periodic table (Na, Al, K, Ca, Ti, Sr, Mo, and Yb). Univariate and multivariate calibrations were presented for each element. The average value for percent root mean square errors of cross-validation for the multivariate models was 3.64%, highlighting the method's strong prediction accuracy. Additionally, the limits of detection for each analyte were estimated from their univariate models: Na = 0.0532, Al = 18.5, K = 0.105, Ca = 0.273, Ti = 67.7, Sr = 0.640, Mo = 22.4, and Yb = 22.9 μg mL –1 . Finally, a test in which multivariate models were used to monitor a liquid system for 80 min was performed to investigate the real-time monitoring capabilities of this liquid LIBS sampling approach. Rigorous measurements were performed to effectively predict the absence and concentrations of multiple analytes as they were spiked and diluted. This demonstration showed the feasibility of using LIBS for real-time liquid quantification models with estimated precision ≤ 8.1%. Finally, the limitations of this approach and potential future improvements are discussed.

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

OrbNet Denali: A machine learning potential for biological and organic chemistry with semi-empirical cost and DFT accuracy

We present OrbNet Denali, a machine learning model for an electronic structure that is designed as a drop-in replacement for groundstate density functional theory (DFT) energy calculations. The model is a message-passing graph neural network that uses symmetryadapted atomic orbital features from a low-cost quantum calculation to predict the energy of a molecule. OrbNet Denali is trained on a vast dataset of 2.3 × 10 6 DFT calculations on molecules and geometries. This dataset covers the most common elements in biochemistry and organic chemistry (H, Li, B, C, N, O, F, Na, Mg, Si, P, S, Cl, K, Ca, Br, and I) and charged molecules. OrbNet Denali is demonstrated on several well-established benchmark datasets, and we find that it provides accuracy that is on par with modern DFT methods while offering a speedup of up to three orders of magnitude. For the GMTKN55 benchmark set, OrbNet Denali achieves WTMAD-1 and WTMAD-2 scores of 7.19 and 9.84, on par with modern DFT functionals. For several GMTKN55 subsets, which contain chemical problems that are not present in the training set, OrbNet Denali produces a mean absolute error comparable to those of DFT methods. For the Hutchison conformer benchmark set, OrbNet Denali has a median correlation coefficient of R 2 = 0.90 compared to the reference DLPNO-CCSD(T) calculation and R 2 = 0.97 compared to the method used to generate the training data (ωB97X-D3/def2- TZVP), exceeding the performance of any other method with a similar cost. Similarly, the model reaches chemical accuracy for noncovalent interactions in the S66x10 dataset. For torsional profiles, OrbNet Denali reproduces the torsion profiles of ωB97X-D3/def2-TZVP with an average mean absolute error of 0.12 kcal/mol for the potential energy surfaces of the diverse fragments in the TorsionNet500 dataset.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗