Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Verification study”

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 109 records · Page 6
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

High-Temperature Gas-Cooled Reactors Multiphysics Simulation Demonstration and Code Validation

This study presents a comprehensive benchmarking and verification effort of several thermal-hydraulic and multiphysics capabilities for high-temperature gas-cooled reactor applications. The first part of this effort focuses on the running-in verification of Griffin’s multiphysics capabilities, specifically for simulating the evolution of pebble-bed reactor cores from startup to equilibrium. Since Fiscal Year 2024, improvements and enhancements have been implemented in Griffin, including simplifying the process to specify streamlines and developing the online cross-section generation capability. In the absence of validation data, code-to-code comparisons are conducted with kugelpy, showing good agreement for integral quantities like k-eff predictions and predictions for maximum power density. However, accuracy issues are noted for more detailed quantities like the spatial distribution of fission rate densities which will require further work to address. The second part of this report presents an improved System Analysis Module (SAM) core channel model where the effects of cross flow are considered during the pressurized loss of forced cooling transient, resulting in an improved agreement of the predicted pebble temperature with respect to the predictions from the SAM 2D porous media model. Additionally, the wall channeling effect due to variable porosity at the near wall region of the core is also investigated. Furthermore, to demonstrate Griffin’s online cross-section generation capability, a Multiphysics simulation is performed by coupling Griffin to the SAM core channel model. In the third part of the report, as a part of the Organisation for Economic Co-operation and Development/Nuclear Energy Agency (OECD/NEA) thermal-hydraulic code validation benchmark activity for a high-temperature gas-cooled reactor, the High Temperature Test Facility (HTTF) is investigated first using the NekRS computational fluid dynamics (CFD) code to study the flow mixing phenomenon in the lower plenum of the facility. Then, code-to-code and code-to-data comparisons are performed for Test PG27, which is a pressurized conduction cooldown (PCC) test, using five different codes by six organizations from five countries. The different simulations show good agreements in terms of the general trend but there are differences in some results such as the peak temperatures of different regions and heat removal rate.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2022

FY2021 Improvements to the New CTH Code Verification & Validation Test Suite

Over the past few years, the CTH multiphysics hydrocode has overhauled its software quality and testing processes, implementing current best practices in software quality and building a robust V&V test suite comprised of traditional hydrocode verification problems, including ASC Tri-Lab Test Suite and Enhanced Tri-Lab Test Suite problems, as well as validation problems for some of CTH’s most frequently used equations of state, materials models, and other key capabilities. Substantial progress towards building this new test suite was made in FY19 and FY20. In FY21, the test suite has been expanded to include verification and validation tests of the Steinberg-Guinan-Lund (ST) viscoplastic model and the Johnson Cook (JFRAC) fracture model. Additionally, two new verification tests were added, covering hydrodynamics and high explosive (HE) modeling capabilities: the Kidder Gaussian density problem and the Escape of HE Products (EHEP) problem from the Tri-Lab Test Suite. This report discusses each of these test problems in detail. Verification test results are compared to analytic solutions. Validation test results are compared to experimental data. Wherever possible, convergence or mesh refinement studies are included. Additionally, while implementing the Kidder verification problem, a bug was identified that affects the use of tables to initialize pressure or density in 1D or 2D calculations. A brief discussion of the bug and its fix is included. CTH demonstrates good performance overall on the new test suite problems. Simulation results showed good agreement with analytic solutions for the Kidder problem, with convergence rates ranging between 1.8 and sub-linear, and relatively good agreement for the EHEP problem, though convergence rates for pressure and density were nearly 0. The ST and JFRAC strain rate loading verification tests show good agreement with analytic solutions. Likewise, CTH simulation results show good agreement with experimental validation data, including Taylor rod impact testing, for the materials tested. Future V&V work will focus on adding 2D and 3D versions of existing verification tests as well as adding validation tests of other frequently used capabilities such as other fracture models.

42 ENGINEERING↗
DOE OSTI2020

Verification of the sensitivity and uncertainty-based criticality safety validation techniques: ORNL’s SCALE case study

The use of sensitivity and uncertainty-based techniques has been well-developed in the neutronic community for criticality safety applications. Sensitivity techniques calculate the first-order relative variations in keff due to relative cross-section variations by isotope, reaction type, and incident neutron energy. The result is a vector of sensitivity coefficients which is folded with prior cross-section uncertainties to calculate the so-called similarity indices. Similarity indices are integral quantities which serve as useful mathematical measures for assessing the relevance of existing or new experiments to given application conditions, representing design systems for which the models are to be validated. Despite their importance for a wide range of validation activities, the peer-reviewed literature is non-existent on the development of rigorous verification analysis for their calculated values. Thus, this manuscript proposes a simple non-intrusive method to calculate similarity indices, serving as an independent verification tool for their evaluation. The manuscript derives the theoretical basis for the proposed verification analysis. Next, the verification analysis is applied to the ORNL’s SCALE code suite which employs adjoint-based methods to calculate similarity indices. Numerical experiments using a number of well-known benchmark experiments are employed to verify the calculation of the similarity indices.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2020

Code-verification techniques for hypersonic reacting flows in thermochemical nonequilibrium

The study of hypersonic flows and their underlying aerothermochemical reactions is particularly important in the design and analysis of vehicles exiting and reentering Earth's atmosphere. Computational physics codes can be employed to simulate these phenomena; however, code verification of these codes is necessary to certify their credibility. To date, few approaches have been presented for verifying codes that simulate hypersonic flows, especially flows reacting in thermochemical nonequilibrium. In this work, we present our code-verification techniques for verifying the spatial accuracy and thermochemical source term in hypersonic reacting flows in thermochemical nonequilibrium. Additionally, we demonstrate the effectiveness of these techniques on the Sandia Parallel Aerodynamics and Reentry Code (SPARC).

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2026

Benchmark for two-dimensional large scale coherent structures in partially magnetized E × B plasmas—community collaboration & lessons learned

Low-temperature plasmas (LTPs) are essential to both fundamental scientific research and critical industrial applications. As in many areas of science, numerical simulations have become a vital tool for uncovering new physical phenomena and guiding technological development. Code benchmarking remains crucial for verifying implementations and evaluating performance. This work continues the Landmark benchmark initiative, a series specifically designed to support the verification of LTP codes. In this study, seventeen simulation codes from a collaborative community of nineteen international institutions modeled a partially magnetized E × B Penning discharge. The emergence of large scale coherent structures, or rotating plasma spokes, endows this configuration with an enormous range of time scales, making it particularly challenging to simulate. The codes showed excellent agreement on the rotation frequency of the spoke as well as key plasma properties, including time-averaged ion density, plasma potential, and electron temperature profiles. Achieving this level of agreement came with challenges, and we share lessons learned on how to conduct future benchmarking campaigns. Comparing code implementations, computational hardware, and simulation runtimes also revealed interesting trends, which are summarized with the aim of guiding future plasma simulation software development.

benchmarking↗
DOE OSTI2022

Challenges and Possible Solutions in Aeroelastic Modeling for the Distributed Wind Industry

Aeroelastic modeling (AM) is the primary methodology for structural and performance assessment of any wind turbine; it provides an understanding of the impact of design parameters on turbine loading and power response before witnessing it in the field. Despite these advantages, the use of AM in the distributed wind technology (DWT) sector is limited. This article represents a short summary of an in-depth assessment by the authors of the status of AM and its role within the distributed wind technology design standards. The research gathered input and feedback from a large number of U.S. and international stakeholders, reviewed technical strengths and weaknesses of the current edition of the design standards, analyzed the minutes from recent industry workshops and meetings, collected publicly available AM templates, and provided an evaluation of the existing AM codes. Several goals were achieved including providing strategies for the load assessment categorization of turbines based on rotor swept area and archetype, and guidance for AM verification and validation. Recommendations within this study will advance the value and the ease-of-use of AM, thereby allowing the industry to better capitalize this underutilized tool, resulting in a more efficient design process, an easier path to certification, and overall better and more reliable distributed wind technology products.

17 WIND ENERGY↗
DOE OSTI2021

Technical analysis of filter testing at the U.S. Department of Energy Filter Test Facility

The U.S. Department of Energy (DOE) technical standard DOE-STD-3020-2015 requires verification testing at the Filter Test Facility (FTF) of high-efficiency particulate air (HEPA) filters to be installed in nuclear facility (Hazard Category 1, 2, 3, and radiological facilities) confinement ventilation systems, or to be installed in habitability systems (e.g., filters for protection of workers required to act for the control and mitigation of emergency situations). Filters in these applications perform a safety function in accident situations, are designated as important to safety, or are necessary for habitability systems. Different FTFs have operated for decades, and results of tests are available to examine the need for and benefits of continued execution of 100 percent QA verification testing. The objective of this study and resulting report was to assess the need for 100 percent QA testing at the current FTF in Maryland, based on an examination of a database of test results from 2013 to August 2020. Filters that require independent testing per DOE-STD-3020-2015 to be installed in nuclear facilities are sent to the FTF before being shipped to the facilities. The FTF examines the documentation (e.g., certificates of compliance, results of particle penetration and airflow resistance testing by the manufacturers, labels affixed to filters and packages, and other requirements specified in purchase orders by the nuclear facility operators), implements visual inspections of the physical state of the filters, and checks the dimensions of frames. Filters that meet the documentation and visual inspection requirements are subjected to particle penetration and airflow resistance testing. In case of rejections, the filters are shipped back to the manufacturers, at the cost of the manufacturers, for replacement. Filters that pass all tests are shipped to the nuclear facilities.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2022

Verification and Validation Activities for the Multi-Fidelity Toolkit

The Multi-Fidelity Toolkit (MFTK) is a simulation tool being developed at Sandia National Laboratories for aerodynamic predictions of compressible flows over a range of physics fidelities and computational speeds. These models include the Reynolds-Averaged-Navier-Stokes (RANS) equations, the Euler equations, and modified Newtonian aerodynamics (MNA) equations, and they can be invoked independently or coupled with hierarchical Kriging to interpolate between high-fidelity simulations using lower-fidelity data. However, as with any new simulation capability, verification and validation are necessary to gather credibility evidence. This work describes formal code- and solution-verification activities as well as model validation with uncertainty considerations. Code verification is performed on the MNA model by comparing with an analytical solution for flat-plate and inclined-plate geometries. Solution-verification activities include grid-refinement studies of HIFiRE-1 wind tunnel measurements, which are used for validation, for all model fidelities. A thorough treatment of the validation comparison with prediction error and validation uncertainty is also presented.

42 ENGINEERING↗
DOE OSTI2026

A total of 19 months of daily weather logging on the US east coast: the WFIP3 event log

The Third Wind Forecast Improvement Project (WFIP3) is a multi-institutional field campaign designed to advance the understanding and prediction of the offshore atmospheric boundary layer along the US east coast. Extending from February 2024 through August 2025, WFIP3 combines long-term coastal and offshore measurements with targeted modeling and forecasting efforts. This data paper presents the WFIP3 event log, a curated record of 578 d of meteorological phenomena and field observations that complements the campaign's extensive high-frequency datasets. The event log provides both manually documented daily weather discussions and automatically derived indicators of atmospheric processes – including low-level jets, wind ramps, extreme wind veer, and weak wind conditions – based on observations from scanning lidars deployed at three coastal and offshore sites. The dataset offers structured metadata, standardized time and site identifiers, and consistent terminology to facilitate its integration with WFIP3's observational and modeling data products. The log supports diverse applications, from model evaluation and forecast verification to the selection of case studies on offshore boundary-layer dynamics. The WFIP3 event log is publicly available through the US Department of Energy's Wind Data Hub, providing the research community with a transparent and enduring contextual reference for the interpretation and use of WFIP3 measurements.

17 WIND ENERGY↗
DOE OSTI2021

Model based co-simulation platform for integrated building system control and design optimization

Both steady-state and dynamic simulations have been widely used by HVAC&R industry to support product/equipment development for decades. Steady-state simulation focuses on the system mass, energy and momentum balance of an equilibrium state. It is based on high-fidelity components models, and thus is suitable for system and component design optimization. Dynamic simulation studies the system transient response and is generally used for controls development and verification. It usually does not require rigorous component models of high accuracy because 1) the commonly used PID control is feedback control whose control performance evaluation doesn’t require high fidelity system/plant model; 2) high-fidelity dynamic model significantly increases the number of equations and variables and creates tremendous challenge for math solver. For supervisory control, transactive control or optimization of an integrated building system, the HVAC&R equipment is often one of the sub-components to be controlled. High-fidelity equipment models are required for accurately evaluating control strategies. In addition, building equipment manufacturers have developed a lot of high-fidelity steady-state equipment/component models per their expertise. Thus, a platform that can integrate OEM high-fidelity steady-state model with dynamic building simulation and/or electric power system & grid simulation to support the development and verification of supervisory control for integrated building systems is necessary. In this study, ORNL’s heat pump design tool (HPDM) is utilized to develop a co-simulation platform for supervisory control and optimization in integrated building systems. It is based on a model that integrates high-fidelity steady-state simulation equipment models with dynamic building simulation. A practical case of using the proposed co-simulation platform to develop and evaluate the supervisory control and optimization is presented and discussed.

Sun, Jian↗
DOE OSTI2024

Geomagnetically Induced Current Field Test on Large Grid-Connected Power Transformers: Analysis, Model Development, and Simulations

Geomagnetic-induced current (GIC) flow in power grids can cause undesirable effects such as transformer overheating, harmonics, higher reactive power demand, etc. Many simulation models have been developed to study these effects, but real-world verification on modern transformer designs is rare. Here, this paper presents the first long-duration GIC field test in the U.S. performed on high-voltage, grid-connected transformers featuring winding clamps and tie rods instead of conventional tie bars. Field measurements were taken to evaluate GIC effects. These measurements also aided in developing and validating thermal and electromagnetic transient (EMT) models of the transformers. During the test, significant current and voltage distortions were observed along with considerable transformer reactive power losses. Analysis of the field measurements showed that the transformers’ hottest spot was at the inner windings, and their k-factors were close to factory test and software default values. Thermal simulations indicated that the transformers would not violate their thermal limits even for a GIC waveform that peaks at about 200 A/phase. EMT simulations revealed that increased transformer loading may reduce GIC-induced reactive power demand and harmonics in certain scenarios. The study also highlighted potential inaccuracies in using the k-factor method to calculate transformer reactive power losses.

EMTDC↗
DOE OSTI2020

Fast and Accurate Core Analysis by the Full-Immersion Pressure-Pulse Decay: Part 1—Theory

Summary Core-scale measurements are considered the ground truth that oil and gas or electric utility companies use to predict the migration of fluids such as oil, natural gas, carbon dioxide, or brine deep underground during their extraction or injection operations. To provide a greater understanding of petrophysical properties of low-permeability geologic formations such as shales and tight gas sandstones, this study introduces a novel core-analysis procedure. The technique follows conventional pressure-pulse-decay permeametry, where the pressure in an inlet chamber adjacent to a cylindrical core plug undergoes a rapid pressurization, the system is shut in, and the pressure reaches a new equilibrium. However, unlike a standard unidirectional pressure-pulse decay, the full-immersion pressure-pulse decay (Hannon 2019) applies a pressure disturbance to the entire outer surface area of the sample. This article covers the numerical simulator designed to model flow through an anisotropic porous sample in this scenario. The model assumes distinct but uniform permeabilities along the radial and axial directions of the cylindrical plug sample. When extracting a plug vertically (or perpendicular to bedding), the permeability along the radial direction associates with the horizontal permeability (i.e., parallel to bedding), whereas in the axial direction, flow occurs perpendicular to bedding (a vertical permeability). Investigations of these model outputs demonstrate an approximately 20-fold decrease in time to complete a full-immersion experiment compared with conventional pressure-pulse decay. Furthermore, the pressure-decay curves resulting from the full-immersion method have slightly different shapes than those resulting from other unidimensional transient methods. These differences begin to demonstrate that, under achievable experimental conditions, the analysis of pressure data from one full-immersion test could enable the simultaneous estimation of the apparent permeabilities parallel and perpendicular to bedding of a cylindrical sample in addition to its porosity. A follow-up article finalizes the proof of this capability with a parameter-estimation procedure and presents experimental verification through a proof-of-concept study. Described in greater detail in that article, the parameter estimator requires multiple forward simulations to analyze the data, which behooves minimizing the compute time for each simulation. By using an alternating direction implicit time-marching scheme and a structured but variably spaced grid, the numerical simulator built for this purpose provides a forward model output with suitable accuracy in approximately 0.5 seconds.

Energy & Fuels↗
DOE OSTI2025

Multimode turbulent flow measurements using magnetic resonance imaging- and laser-based techniques and computational fluid dynamics simulations

We studied the flow field characteristics of a turbulent flow over a regularized cube array with a perpendicular injection flow through the floor between the second and third cubical elements, representing the complex flow interactions of a 3D jet and the wake flows behind cubical obstacles. Four different experimental measurements were performed: two magnetic resonance imaging-based measurements for three-dimensional three-component velocity (MRV) and concentration (MRC) and two laser-based techniques, particle image velocimetry (PIV) and planar laser-induced fluorescence (PLIF), for two-dimensional two-component velocity and concentration measurement, respectively. The mainstream Reynolds number is Re = 15 000⁠, based on the primary inlet velocity U m and channel height D H ⁠, whereas the injector Reynolds number is Re j = 3400⁠, based on the injector velocity U j and the injector's exit width D j ⁠. Numerical simulations were performed for the studied flow configuration of turbulent flow over a regularized cube array using Reynolds-averaged Navier–Stokes (RANS) and large-eddy simulation (LES) approaches. Results obtained from experimental measurements—including MRV, MRC, PIV, and PLIF—as well as RANS and LES simulations are discussed and compared along several horizontal and vertical planes of the studied configuration. In addition, 3D turbulent flow structures, such as leading-edge vortex, horseshoe vortex, and jet shear-layer vortex, and the isosurfaces of scalar concentration successfully revealed by MRV and MRC techniques were found to be in very good agreement with those 3D features extracted from RANS and LES simulations. In conclusion, the high-resolution experimental and numerical database obtained from this study could be useful for validation and verification of numerical codes.

Computational fluid dynamics↗
DOE OSTI2023

Dosimetric delivery validation of dynamically collimated pencil beam scanning proton therapy

Pencil beam scanning (PBS) proton therapy target dose conformity can be improved with energy layer-specific collimation. One such collimator is the dynamic collimation system (DCS), which consists of four nickel trimmer blades that intercept the scanning beam as it approaches the lateral extent of the target. While the dosimetric benefits of the DCS have been demonstrated through computational treatment planning studies, there has yet to be experimental verification of these benefits for composite multi-energy layer fields. The objective of this work is to dosimetrically characterize and experimentally validate the delivery of dynamically collimated proton therapy with the DCS equipped to a clinical PBS system. Approach. Optimized single field, uniform dose treatment plans for 3 × 3 × 3 cm 3 target volumes were generated using Monte Carlo dose calculations with depths ranging from 5 to 15 cm, trimmer-to-surface distances ranging from 5 to 18.15 cm, with and without a 4 cm thick polyethylene range shifter. Treatment plans were then delivered to a water phantom using a prototype DCS and an IBA dedicated nozzle system and measured with a Zebra multilayer ionization chamber, a MatriXX PT ionization chamber array, and Gafchromic™ EBT3 film. Main results. For measurements made within the SOBPs, average 2D gamma pass rates exceeded 98.5% for the MatriXX PT and 96.5% for film at the 2%/2 mm criterion across all measured uncollimated and collimated plans, respectively. For verification of the penumbra width reduction with collimation, film agreed with Monte Carlo with differences within 0.3 mm on average compared to 0.9 mm for the MatriXX PT. Significance. We have experimentally verified the delivery of DCS-collimated fields using a clinical PBS system and commonly available dosimeters and have also identified potential weaknesses for dosimeters subject to steep dose gradients.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2020

Sensitivity of Wind Turbine Array Downstream Effects to the Parameterization Used in WRF

The Weather Research and Forecasting (WRF) Model has been extensively used for wind energy applications, and current releases include a scheme that can be applied to examine the effects of wind turbine arrays on the atmospheric flow and electricity generation from wind turbines. Herein we present a high-resolution simulation using two different wind farm parameterizations: 1) the “Fitch” parameterization that is included in WRF releases and 2) the recently developed Explicit Wake Parameterization (EWP) scheme. We compare the schemes using a single yearlong simulation for a domain centered on the highest density of current turbine deployments in the contiguous United States (Iowa). Pairwise analyses are applied to diagnose the downstream wake effects and impact of wind turbine arrays on near-surface climate conditions. On average, use of the EWP scheme results in small-magnitude wake effects within wind farm arrays and faster recovery of full WT array wakes. This in turn leads to smaller impacts on near-surface climate variables and reduced array–array interactions, which at a systemwide scale lead to summertime capacity factors (i.e., the electrical power produced relative to nameplate installed capacity) that are 2%–3% higher than those from the more commonly applied Fitch parameterization. It is currently not possible to make recommendations with regard to which wind farm parameterization exhibits higher fidelity or to draw inferences with regard to whether the relative performance may vary with prevailing climate conditions and/or wind turbine deployment configuration. However, the sensitivities documented herein to the wind farm parameterization are of sufficient magnitude to potentially influence wind turbine array siting decisions. Thus, our research findings imply high value in undertaking combined long-term high-fidelity observational studies in support of model validation and verification.

17 WIND ENERGY↗
DOE OSTI2025

Building a new multiphysics workflow in MOOSE: application to tritium migration, trapping and advection in TMAP8

Fusion devices are anticipated to produce and consume several kilograms of tritium per year. This rare fuel resource is both highly mobile and radioactive, making tracking inventories a priority for operation and safety. The fusion safety program at the Idaho National Laboratory has been developing the Tritium Migration and Analysis Program (TMAP), of which the latest version is a MOOSE-based application. TMAP8 is verified against its predecessors and possesses additional multi-dimensional tritium migration modeling capabilities. As we extend its capabilities towards both whole device (in multiple dimensions) and whole plant (with multiple components) simulations, the syntax of inputs must become compact, descriptive, compatible with quality assurance processes, and as error-proof as achievable. The new Physics system developed MOOSE can set up equations and instantiating them on plant components. The system permits the automatic definition of complex discretization with a consistency between object parameters achieved programmatically. The Physics system can currently instantiate the equations for heat conduction and Navier Stokes weakly compressible flow. In MOOSE-terms, it automates the definition of kernels, boundary conditions, and several core and helper materials and fields. As part of this effort, Physics classes were developed for tritium migration, trapping and advection within either a multi-dimensional Navier Stokes fluid dynamics simulation, or a 1D thermal hydraulics piping system. In this presentation, we will showcase the new syntax, its application to several verification and validation cases which were already studied using the classical TMAP8 syntax, and a demonstration of the new coupling capabilities for the migration of tritium into blanket coolant channels and the subsequent advection into the coolant loop.

70 - PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2022

Power and Communications Hardware-in-the-Loop CPS Architecture and Platform for DER Monitoring and Control Applications: Preprint

The rapid growth of distributed energy resources (DERs) has prompted increasing interest in the monitoring and control of DERs through hybrid smart grid communications. The deployment of communications and computation has transformed the traditional physical power grid into a smart cyber-physical system (CPS). To fully understand the interdependency between physical grid and cyber netowrks, this study designed a power and communications hardware-in-the-loop (PCommHIL) CPS architecture, which enables the flexible verification of DER monitoring and control with hybrid communications architectures and Internet protocols. Design, development and case study of a PCommHIL testbed for the DER coordination are discussed in detail, and the proposed platform integrates DER devices, Advanced Metering Infrastructures (AMIs), and a suite of hybrid communications networks for distribution automation applications. Case study on DER situational awareness and Volt-Var control validates the efficacy of this proposed PCommHIL platform with hybrid communications designs. Results show that the HAN communication technologies play a critical role in hybrid designs and it is the bottleneck for DER applications. High performance communication technologies are highly recommended to be applied in the HAN for enhanced monitoring and real-time control of DERs.

AMIs↗
DOE OSTI2021

Learning Universal Authorship Representations

This code contains all the utilities required to reproduce the results of our EMNLP 2021 paper "Learning Universal Authorship Representations". It contains the utilities required for download the datasets, training our model, and performing all evaluations necessary for reproducing the results in the paper. Here's the abstract of our work: Determining whether two documents were composed by the same author, also known as authorship verification, has traditionally been tackled using statistical methods. Recently, authorship representations learned using neural networks have been found to outperform alternatives, particularly in large-scale settings involving hundreds of thousands of authors. But do such representations learned in a particular domain transfer to other domains? Or are these representations inherently entangled with domain-specific features? To study these questions, we conduct the first large-scale study of cross-domain transfer for authorship verification considering zero-shot transfers involving three disparate domains: Amazon reviews, fanfiction short stories, and Reddit comments. We find that although a surprising degree of transfer is possible between certain do- mains, it is not so successful between others. We examine properties of these domains that influence generalization and propose simple but effective methods to improve transfer.

Rivera Soto, RafaelA↗