Search NASA⌕ Search

SEARCH · Search NASA

Results for “Error Rate Predictions”

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 181 records · Page 10

Phenomenological Constitutive Modeling of High-Temperature Flow Behavior Incorporating Individual and Coupled Effects of Processing Parameters in Super-austenitic Stainless Steel

A novel constitutive model has been developed for predicting flow responses of super-austenitic stainless steel over a wide range of strains (0.05-0.6), temperatures (1173-1423 K) and strain rates (0.001-1 s{sup −1}). Further, the predictability of this new model has been compared with the existing Johnson–Cook (JC) and modified Zerilli–Armstrong (M-ZA) model. The JC model is not befitted for flow prediction as it is found to be exhibiting very high (~ 36%) average absolute error (δ) and low (~ 0.92) correlation coefficient (R). On the contrary, the M-ZA model has demonstrated relatively lower δ (~ 13%) and higher R (~ 0.96) for flow prediction. The incorporation of couplings of processing parameters in M-ZA model has led to exhibit better prediction than JC model. However, the flow analyses of the studied alloy have revealed the additional synergistic influences of strain and strain rate as well as strain, temperature, and strain rate apart from those considered in M-ZA model. Hence, the new phenomenological model has been formulated incorporating all the individual and synergistic effects of processing parameters and a ‘strain-shifting’ parameter. The proposed model predicted the flow behavior of the alloy with much better correlation and generalization than M-ZA model as substantiated by its lower δ (~ 7.9%) and higher R (~ 0.99) of prediction.

36 MATERIALS SCIENCE↗

Evaluating dosimetric accuracy of flattening filter free compensator-based IMRT: Measurements with diode arrays

Purpose: Compensator-based IMRT coupled with the high dose rate flattening filter free (FFF) beams offers an intriguing possibility of delivering an intensity modulated radiation field in just a few seconds. As a first step, the authors evaluate the dosimetric accuracy of the treatment planning system (TPS) FFF beam model with compensators. Methods: A 6 MV FFF beam from a TrueBeam accelerator (Varian Medical Systems, Palo Alto CA) was modeled in PINNACLE TPS (v. 9.0, Philips Radiation Oncology, Fitchburg WI). Flat brass slabs from 0.3 to 7 cm thick and an 18 deg. brass wedge were used to adjust the beam model. A 2D (MAPCHECK) and 3D (ARCCHECK) diode arrays (Sun Nuclear Corp, Melbourne FL), were investigated for use with the compensator FFF beams. Corrections for diode sensitivity caused by the spectral changes in the beam were introduced. Four compensator plans based on the AAPM TG-119 report were developed. A composite ion chamber measurement, beam by beam MAPCHECK measurements, and a composite ARCCHECK measurement were performed. The array results were analyzed with the same thresholds as in TG-119 report--3%/3 mm with global dose normalization--as well as with the more stringent combinations of the gamma analysis criteria. Results: The FFF beam shows a greater variation of the effective attenuation coefficient with brass thickness due to the prevalence of the low energy photons compared to the conventional 6X beam. As a result, a compromise had to be made while trying to achieve dose agreement for a combination of field sizes, brass thicknesses, and measurement depths ({>=}5 cm in water). An agreement of measured and calculated dose to within 1% was observed for brass thicknesses up to 2 cm. For the 3 cm slab, an error of up to 2.8% was noted for the field sizes above 10 x 10 cm{sup 2}, and up to 3.8% for the 5 x 5 cm{sup 2} field. Both diode arrays exhibit a substantial sensitivity drop as the compensator thickness increases, reaching 10% for a 7 cm brass slab. A simple correction based on the brass thickness along the ray was introduced to counteract this effect. Pooled for five profiles, the average ratio of uncorrected and corrected MAPCHECK to ion chamber readings are 0.966 and 1.008, respectively. With the proper correction, all MAPCHECK measurement to calculation comparisons exhibit 100%{gamma}(3%/3 mm) passing rates with global dose-error normalization. For the TG-119-type plans, the average {gamma}(2%/2 mm) passing rate with local normalization is 94% (range 87.8%-98.3%). The lower ARCCHECK{gamma}-analysis passing rates (corrected for diode sensitivity) are predictable based on the observed PDD discrepancies. However, with the 3%/3 mm thresholds and global normalization, the average {gamma}-analysis passing rate is 96.4% (range 89.9%-100%). Conclusions: MAPCHECK analysis demonstrates high passing rates with the stringent {gamma}(2%/2 mm) and local normalization criteria combination. The geometry of the ARCCHECK array creates a stress test for the FFF TPS model because of the shallow depth of the entrance diodes and large air cavity. Hence, the ARCCHECK{gamma}-analysis passing rates are lower than with the MAPCHECK, while still on par with TG-119.

61 RADIATION PROTECTION AND DOSIMETRY↗

A Detailed Comparison of Multidimensional Boltzmann Neutrino Transport Methods in Core-collapse Supernovae

The mechanism driving core-collapse supernovae is sensitive to the interplay between matter and neutrino radiation. However, neutrino radiation transport is very difficult to simulate, and several radiation transport methods of varying levels of approximation are available. In this paper, we carefully compare for the first time in multiple spatial dimensions the discrete ordinates (DO) code of Nagakura, Yamada, and Sumiyoshi and the Monte Carlo (MC) code Sedonu, under the assumptions of a static fluid background, flat spacetime, elastic scattering, and full special relativity. We find remarkably good agreement in all spectral, angular, and fluid interaction quantities, lending confidence to both methods. The DO method excels in determining the heating and cooling rates in the optically thick region. The MC method predicts sharper angular features due to the effectively infinite angular resolution, but struggles to drive down noise in quantities where subtractive cancellation is prevalent, such as the net gain in the protoneutron star and off-diagonal components of the Eddington tensor. We also find that errors in the angular moments of the distribution functions induced by neglecting velocity dependence are subdominant to those from limited momentum-space resolution. We briefly compare directly computed second angular moments to those predicted by popular algebraic two-moment closures, and we find that the errors from the approximate closures are comparable to the difference between the DO and MC methods. Finally, included in this work is an improved Sedonu code, which now implements a fully special relativistic, time-independent version of the grid-agnostic MC random walk approximation.

79 ASTRONOMY AND ASTROPHYSICS↗

Experimental and modeling study on char combustion

In this study, on the basis of experimental verifications with an FTIR online measurement system, theoretical calculations by using the strict continuous-film model were first compared with those by the simple single-film model that is still widely used in mathematical modeling of pulverized coal flames. The results indicated that the single-film model has some significant errors in its predictions of the ignition temperature and the combustion following ignition and hence should have some restrictions on its application. Then an improved char combustion model has been presented, taking into consideration the influence of the finite-rate heterogeneous reduction and oxidation reactions. This model gives the explicit algebraic expressions for the overall rate of combustion, the surface temperature of the particle, and the gas temperature at the flame sheet. Compared with the single-film model, predictions by the present model were in much better agreement with those predicted by the continuous-film model and the experimental data. The novel model is also much easier to be integrated into the comprehensive computing codes for industrial pulverized coal flame than the continuous-film model. 21 refs., 14 figs.

01 COAL, LIGNITE, AND PEAT↗

Multiresolution GPC-Structured Control of a Single-Loop Cold-Flow Chemical Looping Testbed

Chemical looping is a near-zero emission process for generating power from coal. It is based on a multi-phase gas-solid flow and has extremely challenging nonlinear, multi-scale dynamics with jumps, producing large dynamic model uncertainty, which renders traditional robust control techniques, such as linear parameter varying H ∞ design, largely inapplicable. This process complexity is addressed in the present work through the temporal and the spatiotemporal multiresolution modeling along with the corresponding model-based control laws. Namely, the nonlinear autoregressive with exogenous input model structure, nonlinear in the wavelet basis, but linear in parameters, is used to identify the dominant temporal chemical looping process dynamics. The control inputs and the wavelet model parameters are calculated by optimizing a quadratic cost function using a gradient descent method. The respective identification and tracking error convergence of the proposed self-tuning identification and control schemes, the latter using the unconstrained generalized predictive control structure, is separately ascertained through the Lyapunov stability theorem. The rate constraint on the control signal in the temporal control law is then imposed and the control topology is augmented by an additional control loop with self-tuning deadbeat controller which uses the spatiotemporal wavelet riser dynamics representation. The novelty of this work is three-fold: (1) developing the self-tuning controller design methodology that consists in embedding the real-time tunable temporal highly nonlinear, but linearly parametrizable, multiresolution system representations into the classical rate-constrained generalized predictive quadratic optimal control structure, (2) augmenting the temporal multiresolution loop by a more complex spatiotemporal multiresolution self-tuning deadbeat control loop, and (3) demonstrating the effectiveness of the proposed methodology in producing fast recursive real-time algorithms for controlling highly uncertain nonlinear multiscale processes. The latter is shown through the data from the implemented temporal and augmented spatiotemporal solutions of a difficult chemical looping cold flow tracking control problem.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Branching Fraction Measurement of B to omega l nu decays

We present a measurement of the B{sup +} {yields} {omega}{ell}{sup +}{nu} branching fraction based on a sample of 467 million B{bar B} pairs recorded by the BABAR detector at the SLAC PEP-II e{sup +}e{sup -} collider. We observe 1041 {+-} 133 signal decays, corresponding to a branching fraction of {Beta}(B{sup +} {yields} {omega}{ell}{sup +}{nu}) = (1.15 {+-} 0.15 {+-} 0.12) x 10{sup -4}, where the first error is statistical and the second is systematic. The dependence of the decay rate on q{sup 2}, the momentum transfer squared to the lepton system, is compared to QCD predictions of the form factors based on a quark model and light-cone sum rules.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Impacts of hybridization and forecast errors on the probabilistic capacity credit of batteries

Battery storage is increasingly identified as being among the least-cost mix of technologies in the evolving U.S. electricity mix. This study explores the marginal capacity credit of batteries using a probabilistic, reliability-based, effective firm capacity method, which we apply for multiple battery power ratings, durations, coupling types, deployment locations, and dispatch profiles within a test system that is based on the Texas Interconnection in the year 2024. We find that the capacity credits for all battery durations depend on their ability to predict the timing of reliability events. Even 1-2 h forecast errors - resulting in early or delayed battery discharging relative to the onset of a reliability event - lead pronounced capacity credit reductions, especially for 4-h duration batteries. Coupling batteries with solar mitigates the uncertainty associated with a shorter-duration battery's availability during reliability events, primarily due to the relatively high solar capacity credit in our test system. Coupled (or hybrid) system designs with oversized solar arrays, the ability to charge the coupled battery with grid energy, and larger batteries lead to the greatest capacity credit benefits of hybridization. We do not see evidence that the hybrid capacity credit exceeds the sum of the separate battery and solar capacity credits.

24 POWER TRANSMISSION AND DISTRIBUTION↗

DARK MATTER SUB-HALO COUNTS VIA STAR STREAM CROSSINGS

Dark matter sub-halos create gaps in the stellar streams orbiting in the halos of galaxies. We evaluate the sub-halo stream crossing integral with the guidance of simulations to find that the linear rate of gap creation, R{sub U}, in a typical cold dark matter (CDM) galactic halo at 100 kpc is R{sub U}{approx_equal}0.0066 M-hat{sub 8}{sup -0.35}kpc{sup -1}Gyr{sup -1}, where M-hat{sub 8}({identical_to} M-hat /10{sup 8} M{sub Sun }) is the minimum mass halo that creates a visible gap. The relation can be recast entirely in terms of observables, as R{sub U}{approx_equal}0.059w{sup -0.85}kpc{sup -1}Gyr{sup -1}, for w in kpc, normalized at 100 kpc. Using published data, the density of gaps is estimated for M31's NW stream and the Milky Way Pal 5 stream, Orphan stream, and Eastern Banded Structure. The estimated rates of gap creation all have errors of 50% or more due to uncertain dynamical ages and the relatively noisy stream density measurements. The gap-rate-width data are in good agreement with the CDM-predicted relation. The high density of gaps in the narrow streams requires a total halo population of 10{sup 5} sub-halos above a minimum mass of 10{sup 5} M{sub Sun }.

79 ASTRONOMY AND ASTROPHYSICS↗

Quartz c-axis orientation patterns in fracture cement as a measure of fracture opening rate and a validation tool for fracture pattern models

Here, we evaluate a published model for crystal growth patterns in quartz cement in sandstone fractures by comparing crystal fracture-spanning predictions to quartz c-axis orientation distributions measured by electron backscatter diffraction (EBSD) of spanning quartz deposits. Samples from eight subvertical opening-mode fractures in four sandstone formations, the Jurassic– Cretaceous Nikanassin Formation, northwestern Alberta Foothills (Canada), Cretaceous Mesaverde Group (USA; Cozzette Sandstone Member of the Iles Formation), Piceance Basin, Colorado (USA), and upper Jurassic–lower Cretaceous Cotton Valley Group (Taylor sandstone) and overlying Travis Peak Formation, east Texas, have similar quartzose composition and grain size but contain fractures with different temperature histories and opening rates based on fluid inclusion assemblages and burial history. Spherical statistical analysis shows that, in agreement with model predictions, bridging crystals have a preferred orientation with c-axis orientations at a high angle to fracture walls. The second form of validation is for spanning potential that depends on the size of cut substrate grains. Using measured cut substrate grain sizes and c-axis orientations of spanning bridges, we calculated the required orientation for the smallest cut grain to span the maximum gap size and the required orientation of the crystal with the least spanning potential to form overgrowths that span across maximum measured gap sizes. We find that within a 10° error all spanning crystals conform to model predictions. Using crystals with the lowest spanning potential based on crystallographic orientation (c-axis parallel to fracture wall) and a temperature range for fracture opening measured from fluid inclusion assemblages, we calculate maximum fracture opening rates that allow crystals to span. These rates are comparable to those derived independently from fracture temperature histories based on burial history and multiple sequential fluid inclusion assemblages. Results support the R. Lander and S. Laubach model, which predicts that for quartz deposited synchronously with fracture opening, spanning potential, or likelihood of quartz deposits that are thick enough to span between fracture walls, depends on temperature history, fracture opening rate, size of opening increments, and size, mineralogy, and crystallographic orientation of substrates in the fracture wall (transected grains). Results suggest that EBSD maps, which can be more rapidly acquired than measurement of tens to hundreds of fluid inclusion assemblages, can provide a useful measure of relative opening rates within populations of quartz-filled fractures formed under sedimentary basin conditions. Such data are useful for evaluating fracture pattern development models.

36 MATERIALS SCIENCE↗

Impact of size and thermal gradient on supercooling of phase change materials for thermal energy storage

Phase change material based thermal energy storage has many current and potential applications in the heating and cooling of buildings, battery and electronics thermal management, thermal textiles, and dry cooling of power plants. However, connecting lab scale thermal data obtained with differential scanning calorimetry (DSC) to the performance of large-scale practical systems has been a major challenge primarily due to the dependence of supercooling on the size and temperature gradient of the system. Here we show how a phase change material’s supercooling behavior can be characterized experimentally using common lab scale thermal analysis techniques. We then develop a statistics based theoretical model that uses the lab-scale data on small samples to quantitatively predict the supercooling performance for a general thermal energy storage application of any size, including also allowing for the possibility of temperature gradients. Finally, we validate the modeling methodology by comparing to experimental results for solid-solid phase change in neopentyl glycol, which shows how the model successfully predicts the changes in supercooling temperature across a large range of cooling rates (2 orders of magnitude) and volumes (3 orders of magnitude). By accounting for thermal gradients, the model avoids ~2x error incurred by lumped approximations.

36 MATERIALS SCIENCE↗

Numerical study on the Welander oscillatory natural circulation problem using high-order numerical methods

In this study, the classical Welander’s oscillatory natural circulation problem is investigated using high-order numerical methods. As originally studied by Welander, the fluid motion in a differentially heated fluid loop can exhibit stable, weakly instable, and strongly instable modes. A theoretical stability map has also been originally derived from the stability analysis. Numerical results obtained in this paper show very good agreement with Welander’s theoretical derivations. For stable cases, numerical results from both the high-order and low-order numerical methods agree well with the non-dimensional flow rate analytically derived. The high-order numerical methods give much less numerical errors compared to the low-order methods. For stability analysis, the high-order numerical methods could perfectly predict the stability map, while the low-order numerical methods failed to do so. For all theoretically unstable cases, the low-order methods predicted them to be stable. The result obtained in this paper is a strong evidence to show the benefits of using high-order numerical methods over the low-order ones, when they are applied to simulate natural circulation phenomenon that has already gain increasing interests in many future nuclear reactor designs.

42 ENGINEERING↗

Transforming microseismic clouds into near real-time visualization of the growing hydraulic fracture

SUMMARY Microseismic observations during unconventional reservoir stimulation are typically seen as a proxy for clusters of hydraulic fractures and the extent of the stimulated reservoir. Such straightforward interpretation is often misleading and fails to provide a physically reasonable image of the fracturing process. This paper demonstrates the application of a physics-based machine learning algorithm which enables a rapid and accurate fracture mapping from the microseismic data. Our training and validation data set relies on a history-matched geomechanical modelling workflow implemented in GEOS software for the Hydraulic Fracturing Test Site 1 (HFTS-1) project. For this study we augmented the simulated fracture growth through geostatistical modelling of induced seismicity, so that the synthetic microseismic catalogue matches the main statistical properties of the field observations. We formulated the problem of mapping the actual fracture in the clutter of events to parallel common video segmentation workflows: several past video frames (microseismic density snapshots) are passed through a deep convolutional network to classify whether a given voxel is associated with a fracture or intact rock. We found that for accurate fracture mapping, the network’s input and architecture must be augmented to incorporate the fluid injection parameters (pressure, rate, concentration of proppant, and location of the perforation within the cluster). The error rate for the network reached as little as 10 per cent of the fracture area, while a conventional microseismic interpretation approach yielded ∼300 per cent. Our approach also yields must faster predictions than conventional methods (minutes instead of weeks), and could enable engineers to make rapid decisions regarding engineering parameters (pumping rate, viscosity) in real time during stimulation.

58 GEOSCIENCES↗

Predicting neutron damage using TEM with in situ ion irradiation and computer modeling

We have constructed a computer model of irradiation defect production closely coordinated with TEM and in situ ion irradiation of Molybdenum at 80 degrees C over a range of dose, dose rate and foil thickness. We have reexamined our previous ion irradiation data to assign appropriate error and uncertainty based on more recent work. The spatially dependent cascade cluster dynamics model is updated with recent Molecular Dynamics results for cascades in Mo. After a careful assignment of both ion and neutron irradiation dose values in dpa, TEM data are compared for both ion and neutron irradiated Mo from the same source material. Using the computer model of defect formation and evolution based on the in situ ion irradiation of thin foils, the defect microstructure, consisting of densities and sizes of dislocation loops, is predicted for neutron irradiation of bulk material at 80 degrees C and compared with experiment. Reasonable agreement between model prediction and experimental data demonstrates a promising direction in understanding and predicting neutron damage using a closely coordinated program of in situ ion irradiation experiment and computer simulation.

42 ENGINEERING↗

Constitutive Modeling for Hot Working Behavior of SP-700 Titanium Alloy

In this paper, the constitutive modeling and hot deformation behavior of SP-700 titanium alloy was investigated by performing isothermal hot compression tests in the temperature range of 700-950 °C and strain rates of 0.001, 0.1, and 1 s{sup −1}. The flow curves of the alloy indicated that the yield point phenomenon occurs at strain rates of 0.1 and 1 s{sup −1}. The friction, adiabatic heating, and stroke-rate controlling effects were eliminated from the raw data. Then, three types of phenomenological constitutive models, considering the compensation of strain, were used to predict the flow stresses of the alloy. The correlation coefficients (R{sup 2}) of the measured and predicted results for the power-law, hyperbolic-sine law, and dynamic softening models were found to be 0.9790, 0.9883, and 0.9877, respectively. The corresponding average absolute relative errors (AARE) were 6.3, 5.9, and 7.45%, respectively. The flow softening behavior of SP-700 alloy was modeled using the combined hyperbolic-sine law and dynamic softening model approach. The predicted flow stresses were in a good agreement with the experimental data, indicated that the developed models can accurately characterize flow behavior of the alloy. In addition, the strain-rate sensitivity (m) distribution map of the alloy was obtained by plotting the m-value against temperature and strain rate in the form of a contour map. There was the maximum of m-value (0.36) in the temperature ranges of 700-715 and 837-875 °C, and strain rates of 0.001-0.006 s{sup −1}. The microstructure of hot deformed at 700 °C mainly consists of globularized and lamellar α phase, while dynamic globularization was completed at 800 °C. The hot deformation activation energy of the alloy in α/β region (305.5 kJ mol{sup −1}) was higher than that in single-phase β region (165.2 kJ mol{sup −1}) due to the globularization of lamellar α.

36 MATERIALS SCIENCE↗

Spherically expanding flame in silane–hydrogen–nitrous oxide–argon mixtures

The effect of silane addition on the laminar flame speed ($S^0_u$) of flames propagating in hydrogen–nitrous oxide–argon mixtures has been investigated experimentally for the first time using the spherically expanding flame technique in a constant volume combustion chamber. Replacing hydrogen by silane and maintaining the equivalence ratio constant, much higher flame speeds, explosion peak pressures, and pressure rise coefficients were measured. A previously developed detailed reaction model has been updated based on ab initio thermodynamic properties calculations and collision limit violation analysis. The improved reaction model demonstrates encouraging performance in predicting the flame speed, with a mean absolute error below 11%. To explain the effect of silane addition on the flame dynamics, a number of parameters have been calculated including OH and H rate of production, heat release rate per reaction, and sensitivity coefficient on . The dynamics of freely propagating flames in SiH 4 –H 2 –N 2 O–Ar mixtures is essentially controlled by reactions of the H–O–N chemical system: N 2 O+H=N 2 +OH, OH+H 2 =H 2 O+H, and N 2 O(+M)=N 2 +O(+M). Whereas silane addition does not influence much the rate of production of OH, it significantly modifies that of H with a number of pyrolytic chemical pathways of silicon hydrides, such as SiH+H 2 =SiH 2 +H and Si+H 2 =SiH+H, which act as sink of H atom as they proceed in the backward direction. The reactions forming SiO(s) and SiO 2 (s), such as SiO+OH=SiO 2 (s)+H and 2SiO=2SiO(s), are exothermic and significantly contribute to the temperature increase. The adiabatic, constant pressure flame temperature for mixture containing silane is significantly higher, up to several 100’s K. The increase of induced by silane addition seems to be mostly related to the large increase of the flame temperature which leads to higher energy release rate.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Automated vehicle microscopic energy consumption study (AV-Micro): Data collection and model development

While the Adaptive Cruise Control (ACC) system in automated vehicles (AVs) is expected to impact transportation energy significantly, existing AV energy consumption models only directly adopt those developed with Human-driven Vehicle (HV) data without even slight adaptation or calibration to accommodate unique AV energy consumption features. This study will investigate how accurately HV data-based models can predict the energy consumption of AVs. Empirical trajectory data and corresponding instantaneous energy consumption rates from both AVs and HVs were collected. We adopted two classical HV data-based models to fit these data. The calibration results indicated that these models yield around 20 30% prediction errors for AVs. To further improve the prediction accuracy, this study designed an AV-Micro model by incorporating components of multiple classic energy consumption models that better capture ACC energy consumption features, including piecewise driving behavior. With this, the AV-Micro model achieves lower than 10% prediction errors. The AV-Micro model’s high consistency across different test runs was verified with statistical significance tests, demonstrating its adaptability in different driving profiles. To confirm the discrepancies between the energy consumption features of AVs and HVs, more statistical significance tests were conducted to show that the AV-Micro model cannot be directly applied to HV data. The findings by calibrated AV-Micro models revealed that AVs consume approximately 80.5–146.4 J more energy than HVs for each meter traveled. Furthermore, the frequency analysis of energy consumption indicates that there is still some room for AVs to improve energy efficiency, particularly given their larger amplitude high-frequency fluctuations.

33 ADVANCED PROPULSION SYSTEMS↗

Development of a River Dynamical Core for E3SM to simulate compound flooding on Exascale-class heterogeneous supercomputers

Flooding events pose significant risk to human life, property, and infrastructure. Physically-consistent quantification of altered flood risks in global models requires hyper-resolution (~1 km) or fine flood simulations using two-dimensional (2D) physics schemes, both of which are unavailable in the current generation Earth System Models. Here, in this work, we have developed the River Dynamical Core (RDycore), which is an open-source, 2D shallow water equation (SWE) library for the U.S. Department of Energy's Energy Exascale Earth System Model (E3SM). RDycore uses PETSc and libCEED libraries that allows it to run efficiently on CPUs and GPUs, as well as select a time-integration algorithm at runtime without requiring any code modifications. RDycore achieves spatial error convergence rates for problems with analytical and manufactured solutions similar to those reported previously in the literature, or consistent with the implemented first-order spatial discretization scheme. RDycore's accuracy in predicting flooding for a well-studied dam break problem is comparable to existing SWE models. For a problem with 471 million grid cells, RDycore achieves a speedup of 6.6x and 7.6x on GPUs compared to CPUs when using 320 compute nodes on DOE's Perlmutter and Frontier supercomputers, respectively. The one-way coupling of the RDycore library within E3SM is demonstrated by performing multiple 5-day flooding simulations during Hurricane Harvey driven by five precipitation datasets. The E3SM--RDycore simulations at 30 m spatial resolution accurately simulate maximum water height during the hurricane when benchmarked against a previously published study and achieve a speedup of 15x (Perlmutter) and 21x (Frontier) on GPUs relative to CPUs. The work presented here is the foundational step in providing hardware and algorithmic portability framework for simulating kilometer-scale river dynamics within E3SM.

Flood Simulation↗

Adaptive Impact-Driven Detection of Silent Data Corruption for HPC Applications

For exascale HPC applications, silent data corruption (SDC) is one of the most dangerous problems because there is no indication that there are errors during the execution. We propose an adaptive impact-driven method that can detect SDCs dynamically. The key contributions are threefold. (1) We carefully characterize 18 real-world HPC applications and discuss the runtime data features, as well as the impact of the SDCs on their execution results. (2) We propose an impact-driven detection model that does not blindly improve the prediction accuracy, but instead detects only influential SDCs to guarantee user-acceptable execution results. (3) Our solution can adapt to dynamic prediction errors based on local runtime data and can automatically tune detection ranges for guaranteeing low false alarms. Experiments show that our detector can detect 80-99.99% of SDCs with a false alarm rate less that 1% of iterations for most cases. The memory cost and detection overhead are reduced to 15% and 6.3%, respectively, for a large majority of applications.

Exascale HPC↗