Search NASA⌕ Search

SEARCH · Search NASA

Results for “simulation datasets”

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 1,261 records · Page 70
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Use of Machine Learning on PMU Data for Transmission System Fault Analysis

Synchrophasor technology has been used for monitoring, control, and protection of bulk power system for over 10 years. Deployment of phasor measurement units (PMUs) in the USA power system has surpassed 3000 units installed in the transmission substations as stand-alone intelligent electronic devices (IEDs) or as a software add-on to other devices such as digital protective relays (DPRs) or digital fault recorders (DFRs). By now, thousands of terabytes of PMU data may have been captured and stored by various transmission system operators (TSOs) and independent system operators (ISOs). This creates an opportunity to deploy advanced machine learning (ML) techniques to detect and classify faults recorded by PMUs automatically to be used by the system operators for rapid, critical decision-making when manual analysis of the past or unfolding events is not feasible. In this paper we offer a brief background on how the automated fault analysis may be done using DPR and/or DFR data, and compare some of the legacy approaches to the new ML approaches in the context of the system-wide PMU recordings. We then offer insights from developing practical ML solutions that have been applied on field recordings captured by close to 450 PMUs from all three US interconnections (Western, Eastern and ERCOT) over two years (2016-2017). We identify and illustrate ML challenges we addressed: inaccurate data, data with scarce and temporally imprecise fault labels, data recorded by PMUs sparsely located at substations resulting in the fault records taken afar from the ends of the faulted lines, data containing only positive sequence values, and data taken at different voltage levels. We then illustrate the ML model results for fault analysis under different application scenarios. The novelty of this study is not only in the design, implementation, and performance analysis of the ML algorithms, but also in the use of advanced fault modelling and simulation approaches to improve the training results when developing supervised ML models for fault detection and classification. Extensive simulations of faults were conducted on a 14-bus power system to create a training dataset with over 1400 accurately labelled faults. This dataset was applied to enhance the accuracy of fault detection and classification of machine learning-based models trained with small number of labelled faults in large datasets recorded in the grid interconnections ranging from 5,000 to 70,000 buses.

Synchrophasors, Machine Learning, Fault Analysis, ↗
DOE OSTI2025

Data-model files associated with the manuscript "Modeling the Effects of Wetland Restoration on Coastal Hydrology: A Case Study of Elkhorn Slough Watershed, California"

This package contains the data, simulation setups, notebooks and figures used in “Modeling the Effects of Wetland Restoration on Coastal Hydrology: A Case Study of Elkhorn Slough Watershed, California” (Xu et al., 2025). In this study, we selected Elkhorn Slough, a tidal estuary, in California, to investigate the impact of wetland restoration and sea level rise on coastal hydrology using the process-based coastal hydrologic model, Advanced Terrestrial Simulator (ATS), informed by site-specific data. We designed a novel modeling workflow for incorporating wetland restoration features into land cover and soil properties for the model parameterization. The validation results demonstrate a strong agreement between modeled and observed data. We studied the characteristics of coastal watershed hydrology, then focused on the surface water dynamics at two wetland sites within Elkhorn Slough, a reference site and a restored site. Our simulation results indicate that the restored site successfully maintains surface elevation, resulting in reduced surface inundation. We also examined the impact of wetland restoration under expected sea level rise over the next few decades. The low-lying Yampah Marsh, the reference site, is likely to be inundated due to future sea level rise when highest tides arrive; while a higher percentage of Hester Marsh, the restored site, would retain marsh vegetation in coming decades, regardless of tidal conditions. Our study provides important information for examining the outcome of restoration practices that include surface elevation in tidal wetlands under climate changes.Several files can be found from this data package.1. README.md: This file describes the title, journal, co-authors, abstract, repository structure and model version.2. Simulation_Setups.zip: The file contains the model configuration files (XML format) for ATS. 3. Notebooks.zip: The file contains the Jupyter notebooks for generating the pre- and post-restoration meshes and the meshes of future scenarios. 4. Figures.zip: The file contains the figures used in the manuscript.5. Data.zip: The file contains the data used to drive the model simulations, including watershed and wetlands boundaries, mesh files and references to additional datasets (e.g., meteorological forcing, tidal dataset, DEMs, land cover, soil properties). Also, it contains water level observations at the restored wetland.

54 ENVIRONMENTAL SCIENCES↗
NASA NTRS2015

Comparison of GFED3, QFED2 and FEER1 Biomass Burning Emissions Datasets in a Global Model

Biomass burning contributes about 40% of the global loading of carbonaceous aerosols, significantly affecting air quality and the climate system by modulating solar radiation and cloud properties. However, fire emissions are poorly constrained in models on global and regional levels. In this study, we investigate 3 global biomass burning emission datasets in NASA GEOS5, namely: (1) GFEDv3.1 (Global Fire Emissions Database version 3.1); (2) QFEDv2.4 (Quick Fire Emissions Dataset version 2.4); (3) FEERv1 (Fire Energetics and Emissions Research version 1.0). The simulated aerosol optical depth (AOD), absorption AOD (AAOD), angstrom exponent and surface concentrations of aerosol plumes dominated by fire emissions are evaluated and compared to MODIS, OMI, AERONET, and IMPROVE data over different regions. In general, the spatial patterns of biomass burning emissions from these inventories are similar, although the strength of the emissions can be noticeably different. The emissions estimates from QFED are generally larger than those of FEER, which are in turn larger than those of GFED. AOD simulated with all these 3 databases are lower than the corresponding observations in Southern Africa and South America, two of the major biomass burning regions in the world.

biomass burning↗
DOE OSTI2018

Evaluation of the atmosphere–land–ocean–sea ice interface processes in the Regional Arctic System Model version 1 (RASM1) using local and globally gridded observations

The Regional Arctic System Model version 1 (RASM1) has been developed to provide high-resolution simulations of the Arctic atmosphere–ocean–sea ice–land system. Here, we provide a baseline for the capability of RASM to simulate interface processes by comparing retrospective simulations from RASM1 for 1990–2014 with the Community Earth System Model version 1 (CESM1) and the spread across three recent reanalyses. Evaluations of surface and 2 m air temperature, surface radiative and turbulent fluxes, precipitation, and snow depth in the various models and reanalyses are performed using global and regional datasets and a variety of in situ datasets, including flux towers over land, ship cruises over oceans, and a field experiment over sea ice. These evaluations reveal that RASM1 simulates precipitation that is similar to CESM1, reanalyses, and satellite gauge combined precipitation datasets over all river basins within the RASM domain. Snow depth in RASM is closer to upscaled surface observations over a flatter region than in more mountainous terrain in Alaska. The sea ice–atmosphere interface is well simulated in regards to radiation fluxes, which generally fall within observational uncertainty. RASM1 monthly mean surface temperature and radiation biases are shown to be due to biases in the simulated mean diurnal cycle. At some locations, a minimal monthly mean bias is shown to be due to the compensation of roughly equal but opposite biases between daytime and nighttime, whereas this is not the case at locations where the monthly mean bias is higher in magnitude. These biases are derived from errors in the diurnal cycle of the energy balance (radiative and turbulent flux) components. Therefore, the key to advancing the simulation of SAT and the surface energy budget would be to improve the representation of the diurnal cycle of radiative and turbulent fluxes. The development of RASM2 aims to address these biases. Still, an advantage of RASM1 is that it captures the interannual and interdecadal variability in the climate of the Arctic region, which global models like CESM cannot do.

58 GEOSCIENCES↗
DOE OSTI2014

Color dispersion and Milky-Way-like reddening among type Ia supernovae

Past analyses of Type Ia supernovae have identified an irreducible scatter of 5%-10% in distance, widely attributed to an intrinsic dispersion in luminosity. Another equally valid source of this scatter is intrinsic dispersion in color. Misidentification of the true source of this scatter can bias both the retrieved color-luminosity relation and cosmological parameter measurements. The size of this bias depends on the magnitude of the intrinsic color dispersion relative to the distribution of colors that correlate with distance. We produce a realistic simulation of a misattribution of intrinsic scatter and find a negative bias in the recovered color-luminosity relation, β, of Δβ ≈ –1.0 (∼33%) and a positive bias in the equation of state parameter, w, of Δw ≈ +0.04 (∼4%). We re-analyze current published datasets with the assumption that the distance scatter is predominantly the result of color. Unlike previous analyses, we find that the data are consistent with a Milky-Way-like reddening law (R{sub V} = 3.1) and that a Milky-Way dust model better predicts the asymmetric color-luminosity trends than the conventional luminosity scatter hypothesis. We also determine that accounting for color variation reduces the correlation between various host galaxy properties and Hubble residuals by ∼20%.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2017

Online Adaptive Radiation Therapy

The current paradigm of radiation therapy has the treatment planned on a snapshot dataset of the patient's anatomy taken at the time of simulation. Throughout the course of treatment, this snapshot may vary from initial simulation. Although there is the ability to image patients within the treatment room with technologies such as cone beam computed tomography, the current state of the art is largely limited to rigid-body matching and not accounting for any geometric deformations in the patient's anatomy. A plan that was once attuned to the initial simulation can become suboptimal as the treatment progresses unless improved technologies are brought to bear. Adaptive radiation therapy (ART) is an evolving paradigm that seeks to address this deficiency by accounting for ongoing changes in the patient's anatomy and/or physiology during the course of treatment, affording an increasingly more accurate targeting of disease. ART relies on several components working in concert, namely in-room treatment image guidance, deformable image registration, automatic recontouring, plan evaluation and reoptimization, dose calculation, and quality assurance. Various studies have explored how a putative ART solution would improve the current state of the art of radiation therapy—some centers have even clinically implemented online adaptation. These explorations are reviewed here for a variety of sites.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2019

Irrigation Impact on Water and Energy Cycle During Dry Years Over the United States Using Convection-Permitting WRF and a Dynamical Recycling Model

In this research, an irrigation scheme is implemented in the Weather Research and Forecasting (WRF) model to investigate irrigation impacts over the Continental U.S. (CONUS). Four major irrigated regions and two downwind regions were chosen to understand irrigation impacts over different climate regimes with a focus on irrigation-induced changes on the water and energy cycles. The Dynamic Recycling Model (DRM) is employed to quantify precipitation induced by irrigation and the precipitation recycling ratios over each irrigated region. With the irrigation scheme, WRF improves the simulated precipitation, surface skin temperature, and energy fluxes compared to reference datasets. For the energy cycle, irrigation increases latent heat flux over the irrigated regions along with reduced sensible heat flux. The evaporative cooling effect induced by irrigation leads to a cooler surface and less outgoing longwave radiation at the surface. Irrigation also intensifies the hydrological cycle over the irrigated regions, reflected by the increased precipitation, evapotranspiration, recycling ratio, and moisture export. Downwind regions exhibit increased precipitation and evaporation, decreased moisture flux divergence, and less consistent variations in recycling ratio. The precipitation increases over the irrigated regions can be partly explained by the more unstable low-level conditions, while reduced net moisture export is coincident with the precipitation increases over the downwind regions.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2013

Standardizng Data

To enable the rational design of nanomaterials for improved efficacy and safety, it is critical to understand and exploit the physicochemical properties that drive biological behavior (Morris, 2010). Data mining and computer simulation are essential for deriving information about nanomaterial behavior; however, the datasets needed to support such studies are sparse and stored across a variety of repositories and resources. Schrurs and Lison (2012) have expressed the need for more coherence and structure in the conduct of nanotechnology research. Additionally, the lack of common reporting standards and non-uniformity of information reported have proven to be significant barriers to such data sharing and re-use. The Nanotechnology Working Group (Nano WG), of the US National Institutes of Health National Cancer Informatics Program, has been focused on addressing these barriers. The Nano WG - which includes representatives from over 20 organizations including government agencies, academia, industry, standards organizations, and alliances -has developed ISA-TAB-Nano (Thomas et al, 2013), a general framework for representing and integrating diverse types of data related to the description and characterization of nanomaterials. Recognizing that nanoparticle characterization studies have many of the same challenges as ‘omics-based assays, the Nano WG joined the ISA Commons (Sansone et al., 2012) to leverage lessons learned in ‘omics data sharing. The ISA Commons community brings together 50 collaborators at over 30 scientific organizations around the globe, including regulatory and industrial participants in an increasingly diverse set of life science domains. At the core of the ISA Commons is the ISA metadata tracking framework which forms the basis for the ISA-TAB-Nano extension. The extension of the ISA framework to nanotechnology domain illustrates the power of a synergistic approach that seeks the interoperability of data across multiple research disciplines. To increase adoption, especially in the commercial arena from vendors and manufactures, the ISA-Tab-Nano data-sharing specification has also been submitted for consideration as a standard to the American Society for Testing and Materials (ASTM). Delivering a community-driven specification is only the first phase of the process. To be useful and used, ISA-TAB-Nano must be implemented by tools and databases to assist researchers in reporting their data accordingly, shielding them from unnecessary complexity. Our next step is to extend the open source ISA Software Suite to provide user-oriented tools for the collection, curation, and storage of data compliant with the ISA-TAB-Nano specification. Future work will also focus on the application of the ISA-TAB-Nano format to support emerging standards on minimal information about nanomaterials in biological research (Ostraat et al, 2012; MinChar). ISA-TAB-Nano development is a community-driven effort and we welcome new contributions, collaborations and domain expertise. We invite researchers, software developers, vendors, and other stakeholders to work with us to implement the ISA-Tab-Nano format in their existing systems and research. Likewise, we welcome engagement of regulators, funding agencies, editors, and other policy makers to discuss how this standard can be used to facilitate the sharing and reuse of nanotechnology data across a wide range of disciplines. More information about the ISA-TAB-Nano project can be found online at https://wiki.nci.nih.gov/display/ICR/ISA-TAB-Nano.

informatics↗
DOE OSTI2021

Emergence of seasonal delay of tropical rainfall during 1979–2019

Fundamental to Earth’s climate, tropical rainfall marks a dominant annual cycle characterized by its amplitude and phase, representing the range between wet and dry seasons and their onset timing, respectively. Under global warming, previous studies revealed enhanced amplitude of rainfall annual cycle over ocean and delayed phase over land in the tropics, underpinned by solid theories. It is unclear however whether these changes have already emerged in the observational records. Here, we report a robust delay of 3.02 and 3.43 days over the northern tropical land and Sahel respectively during 1979-2018 based on five observation datasets. Furthermore, most of the delay is captured by multiple ensemble simulations driven by external forcings, with greenhouse gases (GHG) more important over the northern tropical land and anthropogenic aerosols (AER) more important over the Sahel. Both GHG and AER delay the annual cycle of regional precipitation similarly by producing a moister and more ocean-like atmosphere, thus increasing its lag in response to seasonal solar forcing. Our findings provide new evidence of the human footprints on Earth’s hydrological cycle at continental and regional scales. As GHG and AER continue to increase and decrease respectively, these seasonal delays will be further amplified in the future.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Global tropical cyclone precipitation scaling with sea surface temperature

Understanding the relationship between tropical cyclone (TC) precipitation and sea surface temperature (SST) is essential for both TC hazard forecasting and projecting how these hazards will change in the future due to climate change. This work untangles how global TC precipitation is impacted by present-day SST variability (known as apparent scaling) and by long-term changes in SST caused by climate change (known as climate scaling). A variety of datasets are used including precipitation and SST observations, realistic climate model simulations, and idealized climate model simulations. The apparent scaling rates depend on precipitation metric; examples shown here have ranges of 6.1 to 9.5% per K versus 5.9 to 9.8% per K for two different metrics. The climate scaling is estimated at about 5% per K, which is slightly less than the atmospheric moisture scaling based on thermodynamic principles of about 7% per K (i.e., the Clausius–Clapeyron scaling). The apparent scaling is greater than the climate scaling, which implies that the relationship between TC precipitation and present-day SST variability should not be used to project the long-term response of TC precipitation to climate change.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2013

A Parallel EM Algorithm for Model-Based Clustering Applied to the Exploration of Large Spatio-Temporal Data

We develop a parallel EM algorithm for multivariate Gaussian mixture models and use it to perform model-based clustering of a large climate data set. Three variants of the EM algorithm are reformulated in parallel and a new variant that is faster is presented. All are implemented using the single program, multiple data (SPMD) programming model, which is able to take advantage of the combined collective memory of large distributed computer architectures to process larger data sets. Displays of the estimated mixture model rather than the data allow us to explore multivariate relationships in a way that scales to arbitrary size data. We study the performance of our methodology on simulated data and apply our methodology to a high resolution climate dataset produced by the community atmosphere model (CAM5). This article has supplementary material online.

Chen, Wei-Chen↗
DOE OSTI2024

Transferring predictions of formation energy across lattices of increasing size*

In this study, we show the transferability of graph convolutional neural network (GCNN) predictions of the formation energy of the nickel-platinum solid solution alloy across atomic structures of increasing sizes. The original dataset was generated with the large-scale atomic/molecular massively parallel simulator using the second nearest-neighbor modified embedded-atom method empirical interatomic potential. Geometry optimization was performed on the initially randomly generated face centered cubic crystal structures and the formation energy has been calculated at each step of the geometry optimization, with configurations spanning the whole compositional range. Using data from various steps of the geometry optimization, we first trained our open-source, scalable implementation of GCNN called HydraGNN on a lattice of 256 atoms, which accounts well for the short-range interactions. Using this data, we predicted the formation energy for lattices of 864 atoms and 2048 atoms, which resulted in lower-than-expected accuracy due to the long-range interactions present in these larger lattices. We accounted for the long-range interactions by including a small amount of training data representative for those two larger sizes, whereupon the predictions of HydraGNN scaled linearly with the size of the lattice. Therefore, our strategy ensured scalability while reducing significantly the computational cost of training on larger lattice sizes.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Echo state network for coarsening dynamics of charge density waves

An echo state network (ESN) is a type of reservoir computer that uses a recurrent neural network with a sparsely connected hidden layer. Compared with other recurrent neural networks, one great advantage of ESN is the simplicity of its training process. Yet, despite the seemingly restricted learnable parameters, ESN has been shown to successfully capture the spatial-temporal dynamics of complex patterns. Here we build an ESN to model the coarsening dynamics of charge-density waves (CDWs) in a semiclassical Holstein model, which exhibits a checkerboard electron density modulation at half-filling stabilized by a commensurate lattice distortion. The inputs to the ESN are local CDW order parameters in a finite neighborhood centered around a given site, while the output is the predicted CDW order of the center site at the next time step. Special care is taken in the design of couplings between hidden layer and input nodes to ensure lattice symmetries are properly incorporated into the ESN model. Since the model predictions depend only on CDW configurations of a finite domain, the ESN is scalable and transferrable in the sense that a model trained on dataset from a small system can be directly applied to dynamical simulations on larger lattices. Furthermore, our work opens avenues for efficient dynamical modeling of pattern formations in functional electron materials.

2-dimensional systems↗
DOE OSTI2021

Event Cause Analysis in Distribution Networks using Synchro Waveform Measurements

This paper presents a machine learning method for event cause analysis to enhance situational awareness in distribution networks. The data streams are captured using time-synchronized high sampling rates synchro waveform measurement units (SWMU). The proposed method is formulated based on a machine learning method, the convolutional neural network (CNN). This method is capable of capturing the spatiotemporal feature of the measurements effectively and perform the event cause analysis. Several events are considered in this paper to encompass a range of possible events in real distribution networks, including capacitor bank switching, transformer energization, fault, and high impedance fault (HIF). The dataset for our study is generated using the real time digital simulator (RTDS) to simulate real-world events. The event cause analysis is performed using only one cycle of the voltage waveforms after the event is detected. The simulation results show the effectiveness of the proposed machine learning-based method compared to the state-of-the-art classifiers.

Niazazari, Iman↗
DOE OSTI2017

Phosphorus in agricultural soils: drivers of its distribution at the global scale

Phosphorus (P) availability in soils limits crop yields in many regions of the world, while excess of soil P triggers aquatic eutrophication in other regions. Numerous processes drive the global spatial distribution of P in agricultural soils, but their relative roles remain unclear. Here, we combined several global datasets describing these drivers with a soil P dynamics model to simulate the distribution of P in agricultural soils and to assess the contributions of the different drivers at the global scale. We analyzed both the labile inorganic P (P ILAB ), a proxy of the pool involved in plant nutrition and the total soil P (P TOT ). We found that the soil biogeochemical background (BIOG) and farming practices (FARM) were the main drivers of the spatial variability in cropland soil P content but that their contribution varied between P TOT vs P ILAB . Indeed, 97% of the P TOT spatial variability could be explained by BIOG, while BIOG and FARM explained 41% and 58% of P ILAB spatial variability, respectively. Other drivers such as climate, soil erosion, atmospheric P deposition and soil buffering capacity made only very small contribution. Lastly, our study is a promising approach to investigate the potential effect of P as a limiting factor for agricultural ecosystems and for global food production. Additionally, we quantified the anthropogenic perturbation of P cycle and demonstrated how the different drivers are combined to explain the global distribution of agricultural soil P.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2015

SU-E-T-657: Quantitative Assessment of Plan Robustness for Helical Tomotherapy for Head and Neck Cancer Radiotherapy

Introduction: Geometric uncertainties in daily patient setup can lead to variations in the planned dose, especially when using highly conformal techniques such as helical Tomotherapy. To account for the potential effect of geometric uncertainty, our clinical practice is to expand critical structures by 3mm expansion into planning risk volumes (PRV). The PRV concept assumes the spatial dose cloud is insensitive to patient positioning. However, no tools currently exist to determine if a Tomotherapy plan is robust to the effects of daily setup variation. We objectively quantified the impact of geometric uncertainties on the 3D doses to critical normal tissues during helical Tomotherapy. Methods: Using a Matlab-based program created and validated by Accuray (Madison, WI), the planned Tomotherapy delivery sinogram recalculated dose on shifted CT datasets. Ten head and neck patients were selected for analysis. To simulate setup uncertainty, the patient anatomy was shifted ±3mm in the longitudinal, lateral and vertical axes. For each potential shift, the recalculated doses to various critical normal tissues were compared to the doses delivered to the PRV in the original plan Results: 18 shifted scenarios created from Tomotherapy plans for three patients with head and neck cancers were analyzed. For all simulated setup errors, the maximum doses to the brainstem, spinal cord, parotids and cochlea were no greater than 0.6Gy of the respective original PRV maximum. Despite 3mm setup shifts, the minimum dose delivered to 95% of the CTVs and PTVs were always within 0.4Gy of the original plan. Conclusions: For head and neck sites treated with Tomotherapy, the use of a 3mm PRV expansion provide a reasonable estimate of the dosimetric effects of 3mm setup uncertainties. Similarly, target coverage appears minimally effected by a 3mm setup uncertainty. Data from a larger number of patients will be presented. Future work will include other anatomical sites.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2026

Forecasting generative amplification

Generative networks are perfect tools to enhance the speed and precision of LHC simulations. Especially when generating events beyond the size of the training dataset, it is important to understand their statistical precision. We present two complementary methods to estimate the amplification factor without large holdout datasets. Averaging amplification uses Bayesian networks or ensembling to estimate amplification from the precision of integrals over given phase-space volumes. Differential amplification uses hypothesis testing to quantify amplification without any resolution loss. Applied to state-of-the-art event generators, both methods indicate that amplification is already possible in specific regions of phase space.

Bahl, Henning [Heidelberg Univ. (Germany)] (ORCID:↗
DOE OSTI2024

Graph Neural Networks for Parameterized Quantum Circuits Expressibility Estimation (Rev.1)

Parameterized quantum circuits (PQCs) are fundamental to quantum machine learning (QML), quantum optimization, and variational quantum algorithms (VQAs). The expressibility of PQCs is a measure that determines their capability to harness the full potential of the quantum state space. It is thus a crucial guidepost to know when selecting a particular PQC ansatz. However, the existing technique for expressibility computation through statistical estimation requires a large number of samples, which poses significant challenges due to time and computational resource constraints. This paper introduces a novel approach for expressibility estimation of PQCs using Graph Neural Networks (GNNs). We demonstrate the predictive power of our GNN model with a dataset consisting of 25,000 samples from the noiseless IBM QASM Simulator and 12,000 samples from three distinct noisy quantum backends. The model accurately estimates expressibility, with root mean square errors (RMSE) of 0.05 and 0.06 for the noiseless and noisy backends, respectively. We compare our model’s predictions with reference circuits from Sim et al. and IBM Qiskit’s hardwareefficient ansatz sets to further evaluate our model’s performance. Our experimental evaluation in noiseless and noisy scenarios reveals a close alignment with ground truth expressibility values, highlighting the model’s efficacy. Moreover, our model exhibits promising extrapolation capabilities, predicting expressibility values with low RMSE for out-of-range qubit circuits trained solely on only up to 5-qubit circuit sets. This work thus provides a reliable means of efficiently evaluating the expressibility of diverse PQCs on noiseless simulators and hardware.

97 MATHEMATICS AND COMPUTING↗