Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “distributed processing”

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 217 records · Page 12

Activities of human RRP6 and structure of the human RRP6 catalytic domain

The eukaryotic RNA exosome is a highly conserved multi-subunit complex that catalyzes degradation and processing of coding and noncoding RNA. A noncatalytic nine-subunit exosome core interacts with Rrp44 and Rrp6, two subunits that possess processive and distributive 3'-to-5' exoribonuclease activity, respectively. While both Rrp6 and Rrp44 are responsible for RNA processing in budding yeast, Rrp6 may play a more prominent role in processing, as it has been demonstrated to be inhibited by stable RNA secondary structure in vitro and because the null allele in budding yeast leads to the buildup of specific structured RNA substrates. Human RRP6, otherwise known as PM/SCL-100 or EXOSC10, shares sequence similarity to budding yeast Rrp6 and is proposed to catalyze 3'-to-5' exoribonuclease activity on a variety of nuclear transcripts including ribosomal RNA subunits, RNA that has been poly-adenylated by TRAMP, as well as other nuclear RNA transcripts destined for processing and/or destruction. To characterize human RRP6, we expressed the full-length enzyme as well as truncation mutants that retain catalytic activity, compared their activities to analogous constructs for Saccharomyces cerevisiae Rrp6, and determined the X-ray structure of a human construct containing the exoribonuclease and HRDC domains that retains catalytic activity. Structural data show that the human active site is more exposed when compared to the yeast structure, and biochemical data suggest that this feature may play a role in the ability of human RRP6 to productively engage and degrade structured RNA substrates more effectively than the analogous budding yeast enzyme.

59 BASIC BIOLOGICAL SCIENCES↗

SolarAPP+ Enhancements and Commercialization (Final Technical Report)

With millions of distributed photovoltaic (PV) systems expected to be installed over the next five years, the permitting and inspection processes of authorities having jurisdiction (AHJs) may become overburdened, causing delays and increased costs for installed systems (Cruce et al. 2022). The central goal of this project was to automate and streamline permitting processes for distributed PV systems and complementary technologies, such as battery storage. Deploying automated permitting has been hypothesized to reduce permit review times, resulting in reduced costs and improved customer experience. This has the potential to expand the PV and PV-plus-storage market nationwide. The National Renewable Energy Laboratory (NREL) and its project partners, including UL Solutions, the Interstate Renewable Energy Council (IREC), the International Code Council (ICC), and more, developed the Solar Automated Permit Processing (SolarAPP+™) software platform to reduce permit review times. NREL and its partners also collaborated with the solar industry, the building safety community, and local governments to develop SolarAPP+.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

Lattice QCD investigation of nucleon transverse momentum-dependent parton distributions

Certain hadronic processes, such as semi-inclusive deep inelastic scattering (SIDIS) and the Drell-Yan process, can be parametrized in terms of transverse momentum-dependent parton distributions (TMDs). These can be defined via matrix elements of a quark bilocal operator containing a staple-shaped Wilson connection, opening the possibility of a lattice QCD calculation of TMDs. A corresponding calculational scheme is developed and explored. On the basis of a parametrization of the relevant matrix elements in terms of invariant amplitudes, the problem can be boosted to a Lorentz frame suitable for the lattice calculation in a simple manner. Results for the Sivers and Boer-Mulders transverse momentum shifts are presented, with a focus on the dependence on both the staple extent and the staple direction.

Engelhardt, Michael↗

Thermal aging effects on crosslinked polyethylene cable insulation with decabromodiphenyl ether flame retardant alternative

Decabromodiphenyl ether (decaBDE) has been extensively used as a flame retardant in several applications, including nuclear electrical cable insulation. However, decaBDE has been identified as a persistent, bioaccumulative and toxic (PBT) substance, leading to regulatory scrutiny. The Environmental Protection Agency (EPA) published a regulation on January 6, 2021, aimed at phasing out the manufacturing, processing, and distribution of decaBDE. This rule set a compliance deadline of March 8, 2021, for the manufacture and processing of decaBDE, and an extended deadline of January 6, 2023, for specific applications including wire and cable insulation in nuclear power generation facilities. In response to such regulations, RSCC, a major supplier of safety-related electrical cables and associated products to the U.S. nuclear industry updated the formula of their crosslinked polyethylene (XLPE) insulation to replace the historically used decaBDE flame retardant with an acceptable alternative. This change from the previous decaBDE-containing XLPE prompted interest in comparative performance of the two material formulations, especially with respect to characteristics relevant to safety-related function such as thermal and radiation resistance. RSCC graciously provided samples of wire insulated with the decaBDE-containing XLPE formulation and corresponding wire insulated with XLPE of the new formulation, containing a decaBDE alternative. In this work we compare characteristics of the two formulations and a previously produced commercial version of the RSCC decaBDE-containing XLPE insulation subjected to thermal aging at 150 °C and 165 °C. The comparison was focused on mechanical durability, thermal stability in the oxidative environment, and chemical structures. Briefly, • Tensile elongation at break (EAB) results showed loss of mechanical elasticity with longer aging time, as expected. Aging time dependence of EAB did not differ between the decaBDE-containing and decaBDE-alternative samples. • Subtle differences between the two materials can be detected from Fourier-transform infrared spectroscopy (FTIR) absorbance spectra in the range below 1700 cm -1 , are assumed to be related to decomposition of flame retardant additives during thermal aging. • The oxidation induction time (OIT) data seemed to show that the unaged decaBDE-containing XLPE material is more thermally stable than the unaged decaBDE-alternative material, but the discrepancy in OIT decreased with aging time and the OIT values of the two materials became similar starting with the 4 th day of aging at 165 °C. This thermal aging investigation confirmed that the mechanical durability, a key property monitored for cable qualification, was not significantly affected by the modification of the formulation with a decaBDE alternative flame-retardant system in the investigated thermal aging conditions. Further studies on the same sets of materials exposed to thermal and gamma radiation aging would further inform comparison of the materials safety-related function.

36 MATERIALS SCIENCE↗

Distributed Ledger Technology for Fault Tolerant Distribution Grid Operations

This paper explores the potential of distributed ledger technology (DLT) to improve fault-tolerant grid operations by leveraging its core features as an immutable, decentralized ledger, a distributed, consensus-based agreement process, and a distributed state-replication engine. Distribution power systems deliver electricity to millions of customers; however, they are susceptible to various threats that can result in customer interruptions. These include faults caused by adverse weather conditions, natural disasters, vegetation growth, equipment failure, and malicious attacks. To minimize the effects of these faults, fault-handling approaches rely on network knowledge to isolate affected areas and reconnect unaffected areas, reducing the number of affected customers while maintaining safety. Here, we present a trusted data-sharing architecture that enables independent, distributed actors to reconstruct the pre-fault system state by enabling distributed resources to make appropriate decisions with limited network/system information. Although the process requires some data sharing between switch-delimited areas, the approach limits the amount of private information shared, preserving customers' privacy and business-sensitive information. We include three use cases that form a foundation for third parties to develop functional solutions that can eventually be deployed in the field. The gross error detection method used within switch-delimited areas can identify sensor errors and accurately detect circuit breaker states. The evaluation of possible reconnection while preserving data ownership resulted in a voltage magnitude difference smaller than 0.001% from the OpenDSS power flow solution that has full system knowledge, which is below the expected power flow tolerance. The approach offers a promising opportunity for improving fault-tolerant distribution grid operations.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Measurement of the photon electroproduction cross section at JLAB with the goal of performing a Rosenbluth separation of the DVCS contribution

The study of the inner structure of hadrons allows us to understand the nature of the interactions between partons, quarks and gluons, described by Quantum Chromodynamics. The elastic scattering reactions, which have been studied in order to measure the nucleon form factors, are included in this frame. The inelastic scattering reactions are also included in this frame, they allow us to obtain information about the nucleon structure thanks to the development of the parton distribution functions (PDFs). While through elastic scattering we can obtain information about the charge distribution of the nucleon, and hence, about the spatial distribution of the partons, through inelastic scattering we obtain information about the momentum distributions of partons, by employing the PDFs. However, we can study the exclusive inelastic scattering reactions, such as the Deeply Virtual Compton Scattering (DVCS), wich allow us to access to the spatial and momentum distributions simultaneously. This is possible thanks to the generalized parton distributions (GPDs), which allow us to correlate both types of distributions. The process known as DVCS is the easiest way to access the GPDs. This process can be expressed as the scattering of an electron by a proton by means of a virtual photon with the result of the scattered initial particles plus a real photon. We find a process competing with DVCS known as Bethe-Heitler (BH), in which the real photon is radiated by the lepton rather than the quark. Due to the small cross section of DVCS, of the order of nb, in order to conduct these kind of experiments it is necessary to make use of facilities capable of providing high beam intensities. One of these facilities is the Thomas Jefferson National Accelerator Facility , where the experiment JLab E07-007, "Complete Separation of Virtual Photon and π⁰ Electroproduction Observables of Unpolarized Protons", took place during the months of October to December of 2010. The main goal of this experiment is the isolation of the contribution from the term coming form the DVCS from the interference term, resulting from the BH contribution. This isolation is known as "Rosenbluth Separation". The work presented in this thesis focuses on the analysis of the data stored by the electromagnetic calorimeter, employed for the detection of real photons. There is also a a theoretical introduction to the study of the nucleon structure, reviewing the concepts of form factors and parton distributions through elastic and inelastic processes. The computation of the photon leptoproduction cross section is described in detail, as well as the goals of experiment E07-007. This thesis also describes the analysis of the data stored by the electromagnetic calorimeter, with the purpose of obtaining the kinematic variables of the real photons resulting from DVCS reactions. Finally, it describes the selection of events from stored data, the applied cuts to kinematical variables and the background subtraction. Also, the process of extraction of the necessary observables for computing the photon leptoproduction cross section is described, along with the main steps followed to perform the Monte Carlo simulation used in this computation. The resulting cross sections are shown at the end of this thesis.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Membrane cholesterol modulates STEAP2 conformation during dynamic intracellular trafficking processes leading to broad subcellular distribution

Highlights: • STEAP2 distributes broadly into multiple subcellular compartments at steady state. • Cell surface STEAP2 is a minor constituent of the total cellular STEAP2 protein pool. • STEAP2 conformation is modulated within the membrane by its cholesterol content. • STEAP2's metalloreductase activities are suppressed at the plasma membrane. • STEAP2 properties are controlled in a localization-dependent manner by cholesterol. STEAP2 is a member of the Six-Transmembrane Epithelial Antigen of the Prostate (STEAP) protein family that is proposed to function as metalloreductase. While STEAP2 shows a complex subcellular distribution pattern localizing to both secretory and endocytic pathway organelles, how such broad steady-state distribution is maintained is unknown. Similarly, whether STEAP2 undergoes any compartment-specific modulation during intracellular trafficking has not been reported. Leveraging a newly-identified monoclonal antibody that recognizes a conformation-sensitive epitope nested in the second extracellular loop of STEAP2, we demonstrate that the epitope formation was dependent on the cholesterol content of the membrane in which STEAP2 was embedded. Monitoring the STEAP2-dependent internalization of this antibody uncovered STEAP2's rapid internalization from the cell surface and their subsequence trafficking to the Golgi region and endosome-like puncta. Acute inhibition of endocytosis also increased the detectable amount of STEAP2 at the plasma membrane. Collectively, these experiments demonstrate that an intricate balance of membrane flux between the secretory and endocytic pathways underlies the characteristic broad subcellular localization of STEAP2. By using a cell-based assay that detects the metalloreductase functions of cell surface-localizing STEAP4, STEAP2's metalloreductase activities were not detectable, suggesting that its enzymatic function is suppressed at the plasma membrane. The conformational modulation of STEAP2 by the local membrane cholesterol content can therefore serve as a potential mechanism to modulate STEAP2 function in a compartment-restricted manner, by coupling a pre-existing difference in cholesterol content among different cellular membranes to a dynamic trafficking process leading to broad subcellular distribution.

60 APPLIED LIFE SCIENCES↗

Computational Model of the Mark-IV Electrorefiner: Two-Dimensional Potential and Current Distributions

A computational model of the Mark-IV electrorefiner is currently being developed as a joint project between Idaho National Laboratory, Korea Atomic Energy Research Institute, Seoul National University, and the University of Idaho. As part of this model, the two-dimensional potential and current distributions within the molten salt electrolyte are calculated for U3+ , Zr4+ , and Pu3+ along with the total distributions, using the partial differential equation solver of the commercial Matlab software. The electrical conductivity of the electrolyte solution is shown to depend primarily on the composition of the electrolyte and to average 205 mho/m with a standard deviation of 2.5 × 10-5% throughout the electrorefining process. These distributions show that the highest potential gradients (thus, the highest current) exist directly between the two anodes and cathode. The total, uranium, and plutonium potential gradients are shown to increase throughout the process, with a slight decrease in that of zirconium. The distributions also show small potential gradients and very little current flow in the region far from the operating electrodes.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Simulation of He{sup +} induced afterpulses in PMTs

We report on a Monte Carlo simulation study of afterpulses due to trace gases in EMI 9235QA photomultipliers that are widely used in many luminescence detection systems operating in single photon counting mode. The numerical simulation takes into account the nonuniform electric field distribution and processes including elastic scattering: e + He → e + He, excitation: e + He → e + He{sup ∗}, ionization: e + He → 2e + He{sup +}, elastic scattering: He{sup +} + He → He{sup +} + He, charge transfer: He{sup +} + He → He{sub f} + He{sup +} (He{sub f} indicates a fast neutral) and elastic scattering: He{sub f} + He → He{sub f} + He{sub (f)}. The simulated and the measured time of flight distributions agree well. In addition, the above simulated processes demonstrate mechanisms of the observed series of pulses brought about by a single helium atom ionization.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

SAN-Based Block Polymers as a Platform for Manufacturing Strong Isoporous Membranes

Ultrafiltration (UF) membranes are ubiquitous in water purification and bioprocessing. However, co-designing their mechanical and transport properties remains challenging because of the broad pore size distributions at the surface and within the bulk that result from nonsolvent-induced phase separation (NIPS) – their typical manufacturing process. These distributions influence the hydrodynamic resistance to water flow and the stress concentrations around the pores. Developing advanced UF membranes requires innovative molecular designs that offer control over the surface and bulk pores, as well as the mechanical properties of the load-bearing, polymer. Here, we introduce a platform for designing UF membranes by leveraging solution self-assembly of block polymers and chain architectures with pendant polar groups. The block polymers consist of a poly(styrene-co-acrylonitrile) hydrophobic block, which is known for its strength, and a poly(4-vinyl pyridine) hydrophilic block, which drives solution self-assembly. We focus on a series of block polymers with constant molecular weight, M n ≈ 115 kDa, SAN fraction, 75 wt.%, and varying acrylonitrile content, 0 to 40 mol%, to demonstrate that: (i) RAFT dispersion copolymerization of acrylonitrile and styrene provides a facile route to synthesize strong block polymers, (ii) incorporation of acrylonitrile into the hydrophobic block enhances membrane strength by facilitating chain entanglements and dipole-dipole interactions, and (iii) acrylonitrile alters the balance between membrane permeance and rejection, even when the membranes feature similar surface and bulk pores. Overall, our results provide insights into the molecular design of UF membranes with enhanced mechanical and separation properties, contributing to the development of materials for water and energy technologies.

deformation↗

IDC Reengineering Phase 2 Project Scope

The Comprehensive Nuclear-Test-Ban Treaty Organization (CTBTO) operates the International Data Centre (IDC) to support treaty verification activities by Member States. The IDC collects, stores, processes, and distributes data from seismic, hydroacoustic, infrasound, and radionuclide stations. A significant component of the IDC is the seismic, hydroacoustic, and infrasound (SHI) processing system. The IDC has recognized the need to reengineer the SHI processing system and has defined a three-phase reengineering effort. IDC Reengineering Phase 1 (RP1) started in 2011 with the goal to significantly modernize SHI automatic and interactive data processing software, focusing on enhancements to individual components of the system. Reengineering Phase 2 (RP2), started in 2014, addresses the task of specifying a unified architecture for all SHI software, across processing stages, to pave the way for further software development and sustainment in the future. A key objective of RP2 is an architecture sufficient to provide a basis for a cost estimate for the development or enhancement of the software components and subsystems. This report provides a brief description of the scope of the IDC Reengineering Phase 2 project. It describes the goals and objectives of reengineering, the system definition, and the technical scope of the system.This report expands upon the “Scope for IDC Reengineering Phase 2”, July 2014, ECS-DIS-WGB-43-PTS-MATERIAL-12-ADD-1, produced by the IDC.

58 GEOSCIENCES↗

Towards Flexible Exascale Stream Processing System Simulation

Stream processing is an important emerging computational model for performing complex operations on and across multi-source, high-volume, unpredictable dataflows. We present Flow, a platform for parallel and distributed stream processing system simulation that provides a flexible modeling environment for analyzing stream processing applications. The Flow stream processing system simulator is a high-performance, scalable simulator that automatically parallelizes chunks of the model space and incurs near-zero synchronization overhead for acyclic stream application graphs. We show promising parallel and distributed event rates exceeding 149 million events per second on a cluster with 512 processor cores.

99 GENERAL AND MISCELLANEOUS↗

Production of diverse PET probes with limited resources: 24 18 F-labeled compounds prepared with a single radiosynthesizer

Significance Molecular imaging with PET can provide a dynamic, whole-body picture of the rate of biological processes or distribution of biological targets by tracking the distribution of radiolabeled molecules or particles in the body over time. Continual efforts to develop new PET probes are expanding the variety of processes and targets that can be visualized, facilitating basic research, drug development, and patient care. However, access to these probes at all stages of their development is hindered by high costs arising, in large part, from the significant resources that are typically dedicated to production of a single probe. Emerging technologies with increased synthesis flexibility are allowing increased probe diversity with fewer resources and could significantly increase access to new molecular imaging agents.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

THE MASS DISTRIBUTION OF STELLAR-MASS BLACK HOLES

We perform a Bayesian analysis of the mass distribution of stellar-mass black holes using the observed masses of 15 low-mass X-ray binary systems undergoing Roche lobe overflow and 5 high-mass, wind-fed X-ray binary systems. Using Markov Chain Monte Carlo calculations, we model the mass distribution both parametrically-as a power law, exponential, Gaussian, combination of two Gaussians, or log-normal distribution-and non-parametrically-as histograms with varying numbers of bins. We provide confidence bounds on the shape of the mass distribution in the context of each model and compare the models with each other by calculating their relative Bayesian evidence as supported by the measurements, taking into account the number of degrees of freedom of each model. The mass distribution of the low-mass systems is best fit by a power law, while the distribution of the combined sample is best fit by the exponential model. This difference indicates that the low-mass subsample is not consistent with being drawn from the distribution of the combined population. We examine the existence of a 'gap' between the most massive neutron stars and the least massive black holes by considering the value, M{sub 1%}, of the 1% quantile from each black hole mass distribution as the lower bound of black hole masses. Our analysis generates posterior distributions for M{sub 1%}; the best model (the power law) fitted to the low-mass systems has a distribution of lower bounds with M{sub 1%}>4.3 M{sub sun} with 90% confidence, while the best model (the exponential) fitted to all 20 systems has M{sub 1%}>4.5 M{sub sun} with 90% confidence. We conclude that our sample of black hole masses provides strong evidence of a gap between the maximum neutron star mass and the lower bound on black hole masses. Our results on the low-mass sample are in qualitative agreement with those of Ozel et al., although our broad model selection analysis more reliably reveals the best-fit quantitative description of the underlying mass distribution. The results on the combined sample of low- and high-mass systems are in qualitative agreement with Fryer and Kalogera, although the presence of a mass gap remains theoretically unexplained.

79 ASTRONOMY AND ASTROPHYSICS↗

High Performance Computing to Quantify the Evolution of Microscopic Concentration Gradients During Flash Processing

During the Flash process, the cross section of a plain-carbon or a low-alloy steel is austenitized through rapid heating and transformed on rapid cooling to a predominantly martensite + bainite structure with small amounts of retained austenite. Unlike conventional heat treating, homogeneity is intentionally avoided during Flash processing of steels. The Flash process assembly consists of a pair of rolls that transfer the steel sheets through the heating and cooling stage of the thermal cycle. The initial microstructure of the steel consists of ferrite (body-centered cubic iron) + carbide ((Fe,X)mCn) mixture. The heating rate through the peak temperature is a function of temperature and reaches a peak of about 300-400°C/s and the cooling rate has a maximum value of 3,000-4,000°C/s. The on-heating phase transformations include carbide dissolution, austenite (face-centered cubic iron) nucleation and growth, and diffusion of carbon and other substitutional elements in the steel. The on-cooling phase transformations include formation of martensite (body-centered tetragonal phase containing supersaturated solute) and bainite (ferrite plates with or without fine carbides). In this project, the focus is on Fe-C-Cr steels that are currently Flash processed for armor applications. The modeling effort proposed here will help optimize the Flash thermal cycle for these low alloy steels to achieve the target performance, which is an ongoing effort at SFP Works. A significant feature of Flash processed Fe-C-Cr steels is the presence of scatter in the through-thickness in the sheet. The variability in hardness results from a variability in the bainite + martensite microstructure that is sensitive to the local chemical concentration of C and Cr. Such a chemical inhomogeneity is intentionally obtained in the Flash process. Although such a microstructural gradient is presumably responsible for the exceptional properties of the Flash processed steel, it is very important to quantify the gradients as a function of Flash variabilities in processing parameters and the input microstructure. Understanding the mechanistic pathway that leads to microstructural gradients could be ground-breaking and instrumental for achieving better process control and optimized microstructural state to meet application-specific strength-ductility requirements. Since the final microstructure depends on setting up precise solute concentration gradients through a rapid heating process, and transforming these regions into various phases, it is important to understand how small changes in steel chemistry, input microstructure (carbide size and distribution), and process variables (Flash thermal cycle) will impact the solute concentration gradients.

97 MATHEMATICS AND COMPUTING↗

Supervised Learning for Distribution Secondary Systems Modeling: Improving Solar Interconnection Processes

The current interconnection process and hosting capacity analysis for distributed energy resources (DERs), such as photovoltaics (PV) and battery energy storage systems, are based on analyzing grid network constraints (voltage and thermal) using only medium-voltage distribution network models. This is because most utilities do not have secondary low-voltage system models that connect service transformers and residential customers. This is important because in many cases the main impact of interconnecting DERs could occur on the low-voltage distribution systems. This paper proposes a supervised learning method to approximate local secondary models to improve the interconnection process. The proposed supervised learning method includes a decision tree model that predicts the secondary topology and a logistic regression model that predicts conductor types. The case studies demonstrate the benefits of including secondary low-voltage circuits in the interconnection process. We report the proposed modeling methodology is readily scalable and thus can reduce the cost and effort of PV interconnection for the industry and stakeholders.

14 SOLAR ENERGY↗

State Energy Offices’ Engagement in Electric Distribution Planning to Meet State Policy Goals

NASEO and Berkeley Lab released a new publication on State Energy Offices’ Engagement in Electric Distribution Planning to Meet State Policy Goals. State and Territory Energy Offices develop plans, programs, policies, and projects that have a substantial impact on electric distribution systems. They also can participate in distribution system planning (DSP) processes to help ensure that utilities – consumer- and investor-owned – meet the state’s future energy needs. This paper recognizes the wide spectrum of roles that State Energy Offices can play in DSP processes, including planning for distributed energy resources and grid modernization. It highlights various examples of non-regulatory activities by State Energy Offices including planning, conducting studies, convening stakeholder processes, and implementing programs that inform and contribute to distribution system planning. It also provides examples of State Energy Offices’ engagement in proceedings before their respective public utility commissions. As State Energy Offices face myriad challenges associated with meeting state policy goals, preparing for anticipated rates of distributed energy resource deployment, addressing concerns regarding grid reliability and resilience, and recommending and making long-term investment decisions, the examples provided through this guide can serve as a resource in navigating those challenges.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Dynamic Data Repartitioning for Load-Balanced Parallel Particle Tracing

We present a novel dynamic load-balancing algorithm based on data repartitioning for parallel particle tracing in flow visualization. Instead of static data assignment, we dynamically repartition the data into blocks and reassign the blocks to processes to balance the workload distribution among the processes. Block repartitioning is performed based on a dynamic workload estimation method that predicts the workload in the flow field on the fly as the input. In our approach, we allow data duplication in the repartitioning, enabling the same data blocks to be assigned to multiple processes. Load balance is achieved by regularly exchanging the blocks (together with the particles in the blocks) among processes according to the output of the data repartitioning. Compared with other load-balancing algorithms, our approach does not need any preprocessing on the raw data and does not require any dedicated process for work scheduling, while it has the capability to balance uneven workload efficiently. Results show improved load balance and high efficiency of our method on tracing particles in both steady and unsteady flow.

data repartitioning↗