SEARCH · Engineering Papers
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.
Zero-RK is a software package that simulates chemically reacting systems using sparse, preconditioned, adaptive matrix methods to achieve orders-of-magnitude reduction in simulation time while maintaining accurate results.
Zero-RK is a software package that simulates chemically reacting systems using sparse, preconditioned, adaptive matrix methods to achieve orders-of-magnitude reduction in simulation time while maintaining accurate results
The AMG2023 benchmark solves two diffusion problems with a linear solver preconditioned with algebraic multigrid. The code only contains a driver, a Makefile, and a documentation file. It requires an installation of the open source software library hypre that needs to be downloaded elsewhere and is not included here. Its purpose is to benchmark linear solver performance on high performance computers.
Zero-RK is a software package that simulates chemically reacting systems using sparse, preconditioned, adaptive matrix methods to achieve orders-of-magnitude reduction in simulation time while maintaining accurate results.
Zero-RK is a software package that simulates chemically reacting systems using sparse, preconditioned, adaptive matrix methods to achieve orders-of-magnitude reduction in simulation time while maintaining accurate results.
Kynema-FMB (FKA: Kynema) is an open-source performance portable flexible multibody (FMB) dynamics solver designed for time-domain simulations. While originally tailored for wind turbine structural dynamics, the formulation and implementation are those of a general flexible-multidbody dynamics solver that can readily be applied to a wide range of systems. Kynema was designed with a narrow focus, namely to provide a lightweight, fast, accurate FMD solver for coupling to computational-fluid-dynamics (CFD) codes, especially the CFD codes in the Kynema suite, for fluid-structure-interaction (FSI) simulations. Kynema-FMB is equipped to model systems that can be represented as a collection of beams and rigid bodies that are connected through constraints. Degrees of freedom are defined in the inertial/global frame of reference and include displacements and rotations (formally as rotation matrices, but stored as quaternions). The underlying formulation is built on a Lie-group time integrator designed for index-3 differential-algebraic equations, which is second-order accurate in time (Bruls et al., 2012). Beam models are based on geometrically exact beam theory and are discretized as high-order spectral finite elements similar to those in BeamDyn (Wang et al., 2017). The governing equations for a FMD system like a wind turbine constitute a highly nonlinear system of constrained partial-differential equations. Kynema-FMB uses analytical Jacobians in the nonlinear-system solves in each time step. Linear systems use sparse storage and several third-party sparse-linear-system solvers are enabled. Ill conditioning of linear systems is mitigated with preconditioning described in Bottasso et al, 2008. Kynema-FMB is integrated with a simple open-source controller (ROSCO). There is an application programming interface (API) for coupling to geometry-resolved CFD (like that in Sharma et al., 2023) and actuator-force CFD (like that in Kuhn et al., 2025). In the latter, for actuator-line models, Kynema-FMB includes an internal blade-element solver that depends on user-provided lookup tables for coefficients of lift and drag, i.e., aerodynamic polars. Kynema-FMB is written in C++ and leverages Kokkos and Kokkos-Kernels (KokkosEcosystem) as its performance portability layer enabling simulations on both CPU and GPU systems. The repository is equipped with extensive automated testing at the unit and regression/system levels. The following describes the high-level development objectives conceived for Kynema: *Kynema will follow modern software development best practices, including test-driven development (TDD), version control, hierarchical automated testing, and continuous integration (CI) for a robust development environment. *The core data structures are memory efficient and enable vectorization and parallelization at multiple levels. *Data structures are data-oriented to exploit methods for accelerated computing including high utilization of chip resources (e.g., single instruction multiple data (SIMD) instruction sets) and parallelization using GP-GPUs. *The computational algorithms incorporate robust open-source libraries for mathematical operations, resource allocation, and data management. *The API design considers multiple stakeholder needs and ensure integration with existing and future ecosystems for data science, machine learning, and AI. *Kynema-FMB is written in modern C++ and leverages Kokkos as its performance-portability library with inspiration from the kynema stack.
HPG-MxP is a software package that performs a fixed number of multigrid preconditioned (using a Gauss-Seidel smoother) Generalized minimal residual (PGMRES) iterations in order to solve a possibly nonsymmetric large sparse linear system of equations. It is designed to be a benchmark to measure a computer's performance for sparse linear algebra workloads typical in scientific computing while allowing the use of mixed precision methods. The solution is required to have convergence characteristics and accuracy similar to double precision GMRES. It is based on the High Performance Conjugate Gradient Benchmark (HPCG) which restricts all implementations to use only the IEEE double precision format (FP64). The original implementation (https://github.com/hpg-mxp/hpg-mxp) was written by Ichitaro Yamazaki, Jennifer Loe, Christian Glusa, Sivasankaran Rajamanickam, Piotr Luszczek, and Jack Dongarra. Please refer to that repository for documentation on the original implementation. This version is maintained by the National Center for Computational Sciences at Oak Ridge National Laboratory. It is highly scalable and optimized for Oak Ridge Leadership Computing Facility (OLCF) systems, particularly Frontier.
In mid-April 2024, the United Arab Emirates (UAE) and the surrounding Gulf region experienced unprecedented rainfall and catastrophic flooding, causing widespread damage, loss of life, and significant economic costs. During the 3-day period from 15 to 17 April, rainfall in the UAE exceeded 100 mm in the hardest-hit areas, with more than 170% of the typical annual precipitation recorded in just 72 h. The event was associated with exceptionally strong integrated water vapor transport (IVT), driven by a persistent low pressure system and a focused corridor of moisture transport. This environment, combined with favorable dynamical and convective conditions, triggered intense thunderstorms and widespread flooding. This article examines the role of an atmospheric river (AR) in the April 2024 extreme precipitation event, emphasizing the contribution of extreme IVT to preconditioning and amplifying the heavy rainfall. While this event has previously been described primarily in terms of a mesoscale convective system (MCS) and potential vorticity (PV) streamers, here we document and illustrate its close relationship to a record-breaking moisture transport corridor. This AR-based perspective highlights how large-scale moisture-transport frameworks can complement synoptic and mesoscale analyses in understanding extreme rainfall in arid and semiarid regions such as the UAE and the Gulf. The event serves as a stark example of the vulnerabilities faced by arid regions, where atmospheric conditions conducive to extreme flooding may become more frequent in the future.
This study examines two factors impacting initiation of moist deep convection: free-tropospheric environmental relative humidity (φ E ) and horizontal scale of subcloud ascent (R sub ), the latter exerting a dominant control on cumulus cloud width. A simple theoretical model is used to formulate a “scale selection” hypothesis: that a minimum R sub is required for moist convection to go deep, and that this minimum scale decreases with increasing φ E . Specifically, the ratio of R 2 sub to saturation deficit (1 - φ E ) must exceed a certain threshold value that depends on cloud-layer environmental lapse rate. Idealized, large-eddy simulations of moist convection forced by horizontally varying surface fluxes show strong sensitivity of maximum cumulus height to both φ E and R sub consistent with the hypothesis. Increasing R sub by only 300–400 m can lead to a large increase (>5 km) in cloud height. A passive tracer analysis shows that the bulk fractional entrainment rate decreases rapidly with R sub but depends little on φ E . However, buoyancy dilution increases as either R sub or φ E decreases; buoyancy above the level of free convection is rapidly depleted in dry environments when R sub is small. While deep convective initiation occurs with an increase in relative humidity of the near environment from moistening by earlier convection, the importance of this moisture preconditioning is inconclusive as it is accompanied by an increase in R sub . Overall, it is concluded that small changes to R sub driven by external forcing or by convection itself could be a dominant regulator of deep convective initiation.
A nocturnal Amazonian low-level jet (ALLJ) was recently diagnosed using reanalysis data. This work assesses the ability of CESM1.2.2 to reproduce the jet and explores the mechanisms by which the ALLJ influences convection in the Amazon. The coupled CESM simulates the nocturnal ALLJ realistically, while CAM5 does not. A low-level cold air temperature bias in the eastern Amazon exists in CAM5, and thus the ALLJ is weaker than observed. However, a cold SST bias over the equatorial North Atlantic in the coupled model offsets the cold air temperature bias, producing a realistic ALLJ. Climate models significantly underestimate March–May (MAM) precipitation over the eastern Amazon. Here, we ran two sensitivity experiments using the coupled CESM by adding bottom-heavy diabatic heating at noon and midnight for 2.5 h along the coastal Amazon during MAM to mimic the occurrence of shallow precipitating convection. When heating is added during the early afternoon, coastal convection deepens and the ALLJ transports moisture inland from the ocean, preconditioning the environment for deep convective development during the ensuing hours. The increased convection over the eastern Amazon also moderately alleviates the equatorial Atlantic westerly wind bias, leading to deepening of the east Atlantic thermocline in the following months and partially improving the simulated June–August (JJA) Atlantic cold tongue in the coupled model. When heating is added at night, coastal convection does not strengthen as much and the ALLJ transports less moisture. Improvements in the simulated Atlantic winds and SST are negligible. Therefore, diurnal circulations matter to the organization of convection and rain across the Amazon, with impacts over the equatorial Atlantic.
Land-atmosphere feedbacks are a critical component of the hydrologic cycle. Vertical profiles of boundary layer temperature and moisture, together with information about the land surface, are used to compute land-atmosphere coupling metrics. Ground based remote sensing platforms, such as the Atmospheric Emitted Radiance Interferometer (AERI), can provide high temporal resolution vertical profiles of temperature and moisture. When co-located with soil moisture, surface flux, and surface meteorological observations, such as at the Atmospheric Radiation Measurement (ARM) Southern Great Plains (SGP) site, it is possible to observe both the terrestrial and atmospheric legs of land-atmosphere feedbacks. In this study, we compare a commonly used coupling metric computed from radiosonde-based data to that obtained from the AERI to characterize the accuracy and uncertainty in the metric derived from the two distinct platforms. This approach demonstrates the AERI’s utility where radiosonde observations are absent in time and/or space. Radiosonde and AERI based observations of the Convective Triggering Potential and Low-Level Humidity Index (CTP-HI low ) were computed during the 1200 UTC observation time and displayed good agreement during both 2017 and 2019 warm seasons. Furthermore, radiosonde and AERI derived metrics diagnosed the same atmospheric preconditioning based upon the CTP-HI low framework a majority of the time. When retrieval uncertainty was considered, even greater agreement was found between radiosonde and AERI derived classification. The AERI’s ability to represent this coupling metric well enabled novel exploration of temporal variability within the overnight period in CTP and HI low . Observations of CTP-HI low computed within a few hours of 1200 UTC were essentially equivalent, however with greater differences in time arose greater differences in CTP and HI low .
Three genuine stratocumulus-to-cumulus transitions sampled during the Cloud System Evolution over the Trades (CSET) campaign are documented. The focus is on Lagrangian evolution of in situ precipitation, thought to exceed radar/lidar retrieved values because of Mie scattering. Two of the three initial stratocumulus cases are pristine [cloud droplet number concentrations ($N_d$) of ~22 cm –3 ] but occupied boundary layers of different depths, while the third is polluted ($N_d$ ~ 225 cm –3 ). Hourly satellite-derived cloud fraction along Lagrangian trajectories indicate that more quickly deepening boundary layers tend to transition faster, into more intense but more occasional precipitation. These transitions begin either in the morning or late afternoon, suggesting that preceding night processes can precondition or delay the inevitable transition. The precipitation shifts toward larger drop sizes throughout the transition as the boundary layers deepen, with aerosol concentrations only diminishing in two of the three cases. Ultraclean ($N_d$ < 1 cm –3 ) cumulus clouds evolved from pristine stratocumulus cloud with unusually high precipitation rates occupying a shallow, well-mixed boundary layer. Results from a simple one-dimensional evaporation model and from radar/lidar retrievals suggest subcloud evaporation likely increases throughout the transition. This, coupled with larger drop sizes capable of lowering the latent cooling profile, facilitates the transition to more surface-driven convection. The coassociation between boundary layer depth and precipitation does not provide definitive conclusions on the isolated effect of precipitation on the pace of the transition. Differences between the initial conditions of the three examples provide opportunities for further modeling studies.
This study explores the radiation resilience of GaN high-electron-mobility transistors (HEMTs) under various preconditioning treatments—pristine, electro-pulsed, electrically overstressed, and 60 Co gamma-irradiated—with a focus on enhancing survivability in high-radiation environments. Key findings emphasize the gate region’s central role in radiation sensitivity, with increased defect concentrations leading to reduced sensitivity—as determined from reduced single-event transient intensities during pulsed-laser studies—and impacting charge-transport dynamics. Notably, the most sensitive region shifted as a function of bias state, with the drain side of the gate being most sensitive under negative gate bias and the source side under unbiased conditions due to two-dimensional electron gas disruptions. These insights clarify the mechanisms influencing radiation sensitivity in microelectronics and highlight the potential of defect and strain engineering to enhance the radiation hardness of GaN HEMTs across various operational states, providing valuable guidance for applications in space, defense, and other radiation-rich environments.
The Tank Closure Cesium Removal (TCCR) system uses ion exchange columns filled with crystalline silicotitanate (CST) media to process radioactive waste solutions for the removal of ¹³⁷Cs. TCCR currently focuses on dissolving Savannah River Site (SRS) radioactive tank waste (primarily sodium saltcake solids) within Tank 10H followed by at-tank ion exchange column treatment. Two supernate batches from Tank 10H have been processed through the TCCR unit and processing of a third batch is expected soon. After processing of this supernate batch, plans are to replace the CST columns and process dissolved salt solution from Tank 9H through Tank 10H and then through the new CST columns installed in the TCCR unit. The new columns are expected to contain either a media similar to an archived CST batch (IE-911) or the R9120-B CST media used in the current TCCR columns (the two materials are fundamentally the same; just different specifications, product names, and preconditioning steps). Samples of Tank 9H dissolved saltcake were received at the Savannah River National Laboratory (SRNL) and characterized. The Tank 9H supernate contained a high sodium concentration (~9.6 M Na⁺) and will require dilution to near 6 M [Na⁺] prior to processing through the TCCR unit. The cesium concentration in Tank 9H is currently significantly higher than was observed with Tank 10H. Three dilutions of the Tank 9H supernate were conducted to mimic possible dilutions that could be conducted in the tank farm prior to TCCR processing using inhibited water, sodium hydroxide, and sodium nitrate solutions. Dilution #1 was prepared by diluting the Tank 9H supernate by a factor of 1.6 with inhibited water, Dilution #2 was prepared by diluting the Tank 9H supernate by a factor of 3.7 with a mixed sodium hydroxide/sodium nitrate diluent, and Dilution #3 was prepared by also diluting by a factor of 3.7 but with sodium hydroxide only. All three dilutions have similar Na⁺ concentrations (~6 M), but Dilutions #2 and #3 contain significantly less cesium. Major and key minor components of the diluted Tank 9H supernate samples are provided in Table ES-1. Batch contact equilibrium tests were conducted with the Tank 9H dilutions and the two different CST media batches being considered for use in the new TCCR columns. Results are summarized in Table ES-2 and compared to ZAM model predictions. The highest cesium distribution coefficient and percent removal were observed with Dilution #3. IE-911 CST (an archived CST media batch) was more effective at removing cesium than the more recently prepared R9120-B CST media, though both media samples removed >88% of the cesium and the differences may not be statistically different considering the overall uncertainty. Based on the results, maximum cesium loadings were calculated for each CST media type and Tank 9H dilution. In general, maximum cesium loadings from dilutions of this supernate batch are quite high (approaching 0.1 mmol total Cs⁺/g CST for Dilution #1). The highest calculated maximum ¹³⁷Cs loading for the Tank 9H dilutions using the ZAM model with input of the tank compositions and batch contact results was 207 Ci/kg CST (Dilution #1 with IE-911 CST). In all cases, higher maximum cesium loading values were predicted for IE-911 CST versus R9120-B, although the differences varied considerably between the dilutions. The maximum loading value for IE-911 CST with Tank 9H Dilution #1 was only 7% higher than the maximum loading for R9120-B. The maximum loading value for IE-911 CST with Dilution #3 was 24% higher than the maximum loading for R9120-B. Dilution #2 was intermediate between these values. Note that these maximum loading values are the theoretical calculated values, and actual operating conditions will cause differences. A CST binder dilution (correction) factor near 0.7 (relative to pure powder CST) was required for each batch contact test with IE-911 engineered CST using the three Tank 9H dilutions. Correction factors calculated for R9120-B CST ranged from 0.56 to 0.66 for the three dilutions. Following the batch contact tests, the CST samples were isolated from the Tank 9H solution, washed, dried, and digested in acid following established procedures. The analysis results are provided in Table ES-3. Total cesium loading values determined by CST digestion were similar to the calculated loading values based on solution analysis (Table ES-2) for all samples. In addition, the CST was observed to load calcium (R9120-B sample only), iron, strontium, lead, uranium, and plutonium after contact with the waste supernate, as has been observed previously.
This report compares GPU and CPU performance of structured and unstructured hypre solvers on Lassen and Summit. We applied the solvers to various problems of different sizes varying the number of nodes and processes. We provide the individual results for setup phase and solve phase, since preconditioned solvers are often used in different settings. While for a single system solve the total time is determined by the sum of setup and solve time, this can change for different scenarios. For example, in time dependent problems the preconditioner might have to be setup only occasionally or possibly even only once. In that situation solve times will dominate. Thus, since often better GPU speedups can be attained in the solve phase, this will lead to generally better speedups.
We can use formal methods to verify that systems compose correctly without the possibility of incorrect behavior. This means exhaustively checking that System A’s postconditions agree with System B’s preconditions. If so, it is safe to compose. Normal Setting: Every computational path must be accounted for and checked. Verification cost (time) is multiplicative across systems.
We can use formal methods to verify that systems compose correctly without the possibility of incorrect behavior. This means exhaustively checking that System A’s postconditions agree with System B’s preconditions. If so, it is safe to compose.
An optimization-based strategy for solving nonlinear mechanics problems is proposed. In contrast to typical nonlinear equation solver algorithms that aim to find zeros in the residual force function, we minimize an energy (or energy-like) function to encourage solutions which are locally stable equilibria. These smooth and potentially non-convex objective functions are minimized using a preconditioned conjugate-gradient trust-region algorithm. Contact is formulated as an inequality constrained minimization problem, and is solved with an augmented Lagrangian algorithm. Friction is included in the approach via a regularized quasi-potential energy, and other dissipative behavior is included through the use of variational constitutive updates. Finally, to accelerate convergence rates for the Lagrange multipliers, we propose a novel multiplier update algorithm utilizing the Fischer-Burmeister function, and demonstrate super-linear solver convergence for some applications.