Improving resilience of scientific software through a domain-specific approach
Not Available
SEARCH · Search NASA
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.
Not Available
In-situ analysis on the output data of scientific simulations has been made necessary by ever-growing output data volumes and increasing costs of data movement as supercomputing is moving towards exascale. With hardware accelerators like GPUs becoming increasingly common in high end machines, new opportunities arise to co-locate scientific simulations and online analysis performed on the scientific data generated by the simulations. However, the asynchronous nature of GPGPU programming models and the limited context-switching capabilities on the GPU pose challenges to co-locating the scientific simulation and analysis on the same GPU. This paper dives deeper into these challenges to understand how best to co-locate analysis with scientific simulations on the GPUs in HPC clusters. Specifically, our 'Landrush' approach to GPU sharing proposes a solution that utilizes idle cycles on the GPU to provide an improved time-to-answer, that is, the total time to run the scientific simulation and analysis of the generated data. Landrush is demonstrated with experimental results obtained from leadership high-end applications on ORNL's Titan supercomputer, which show that (i) GPU-based scientific simulations have varying degrees of idle cycles to afford useful analysis task co-location, and (ii) the inability to context switch on the GPU at instruction granularity can be overcome by careful control of the analysis kernel launches and software-controlled early completion of analysis kernel executions. Results show that Landrush is superior in terms of time-to-answer compared to serially running simulations followed by analysis or by relying on the GPU driver and hardwired thread dispatcher to run analysis concurrently on a single GPU.
More than $5 billion in investments in advanced metering infrastructure (AMI) technologies, AMI deployments, as pervasive secondary network voltage monitoring systems, provide opportunities for utility operations and controls. This paper focuses on the considerations for AMI-based tools and techniques as the industry moves toward operationalizing such large data sets. Phase identification is a first such tool. Numerous distribution network analysis, monitoring, and control applications - including volt/volt-ampere reactive control, state estimation, and distribution automation - require accurate phase connectivity information in the system models. The phase connectivity database maintained by utilities is inaccurate because of a significant amount of missing data, restoration activities, and network reconfiguration. Existing phase identification techniques that estimate phase connectivity work well in distribution feeders that have low or no photovoltaic (PV) generation; however, they fail to identify the phases accurately when considerable PV generation is present. This work addresses the phase identification problem in the presence of high PV generation using statistical analysis methods. Further, insights into the AMI data requirements for this application in terms of data window length and resolution are provided using sensitivity analysis performed on an actual distribution feeder model of San Diego Gas & Electric Company. The results of this study show that the phase connectivity, even in the presence of high PV generation, can be accurately identified using statistical analysis of AMI data of 1 day.
We report the in-situ 3D reconstruction of through-going muons in the CUORE experiment, a cryogenic calorimeter array searching for neutrinoless double beta (0vββ) decay, leveraging the segmentation of the detector. Due to the slow time response of the detector, time-of-flight estimation is not feasible. Therefore, the track reconstruction is performed using a multi-objective optimization algorithm that relies on geometrical information from the detector as a whole. We measure the integral flux of cosmic-ray muons underground at the Laboratori Nazionali del Gran Sasso, and find our value to be in good agreement with other experiments that have performed a similar measurement. To our knowledge, this work represents the first demonstration of 3D particle tracking and reconstruction of through-going muons with per-event angular determination in a millikelvin cryogenic detector array. The analysis performed for this work will be critical for validating the muon-related background in CUPID, a next-generation 0vββ experiment, and for follow-up studies on detector response and on delayed products induced by cosmic-ray muons.
Hardware platforms in high performance computing are constantly getting more complex to handle even when considering multicore CPUs alone. Numerous features and configuration options in the hardware and the software environment that are relevant for performance are not even known to most application users or developers. Microbenchmarks, i.e., simple codes that fathom a particular aspect of the hardware, can help to shed light on such issues, but only if they are well understood and if the results can be reconciled with known facts or performance models. The insight gained from microbenchmarks may then be applied to real applications for performance analysis or optimization. In this paper we investigate two modern Intel x86 server CPU architectures in depth: Broadwell EP and Cascade Lake SP. We highlight relevant hardware configuration settings that can have a decisive impact on code performance and show how to properly measure on-chip and off-chip data transfer bandwidths. The new victim L3 cache of Cascade Lake and its advanced replacement policy receive due attention. Finally we use DGEMM, sparse matrix-vector multiplication, and the HPCG benchmark to make a connection to relevant application scenarios.
More than $5 billion in investments in advanced metering infrastructure (AMI) technologies, AMI deployments, as pervasive secondary network voltage monitoring systems, provide opportunities for utility operations and controls. This paper focuses on the considerations for AMI-based tools and techniques as the industry moves toward operationalizing such large data sets. Phase identification is a first such tool. Numerous distribution network analysis, monitoring, and control applications - including volt/volt-ampere reactive control, state estimation, and distribution automation - require accurate phase connectivity information in the system models. The phase connectivity database maintained by utilities is inaccurate because of a significant amount of missing data, restoration activities, and network reconfiguration. Existing phase identification techniques that estimate phase connectivity work well in distribution feeders that have low or no photovoltaic (PV) generation; however, they fail to identify the phases accurately when considerable PV generation is present. This work addresses the phase identification problem in the presence of high PV generation using statistical analysis methods. Further, insights into the AMI data requirements for this application in terms of data window length and resolution are provided using sensitivity analysis performed on an actual distribution feeder model of San Diego Gas & Electric Company. The results of this study show that the phase connectivity, even in the presence of high PV generation, can be accurately identified using statistical analysis of AMI data of 1 day.
The Deep-Burn (DB) concept [ ] focuses on the destruction of transuranic nuclides from used light water reactor (LWR) fuel. These transuranic nuclides are incorporated into tri-isotopic (TRISO) coated fuel particles and used in gas-cooled reactors with the aim of a fractional fuel burnup of 60 to 70% in fissions per initial metal atom (FIMA). This high performance is expected through the use of multiple recirculation passes of the fuel in pebble form without any physical or chemical changes between passes. In particular, the concept does not call for reprocessing of the fuel between passes. In principle, the DB pebble bed concept employs the same reactor designs as the presently envisioned low-enriched uranium core designs, such as the 400 MWth Pebble Bed Modular Reactor (PBMR-400) [ ]. Although it has been shown in the previous Fiscal Year (FY) (2009) that a PuO2 fueled pebble bed reactor concept is viable, achieving a high fuel burnup while remaining within safety-imposed prescribed operational limits for fuel temperature, power peaking, and temperature reactivity feedback coefficients for the entire temperature range, is challenging. The presence of the isotopes 239Pu, 240Pu, and 241Pu that have resonances in the thermal energy range significantly modifies the neutron thermal energy spectrum as compared to a standard, UO2-fueled core. Therefore, the DB pebble bed core exhibits a relatively hard neutron energy spectrum. However, regions within the pebble bed that are near the graphite reflectors experience a locally softer spectrum. This can lead to power and temperature peaking in these regions. Furthermore, a shift of the thermal energy spectrum with increasing temperature can lead to increased absorption in the resonances of the fissile Pu isotopes. This can lead to a positive temperature reactivity coefficient for the graphite moderator under certain operating conditions. Regarding the coated particle performance, the FY 2009 investigations showed that no significant failure is to be expected for the reference fuel particle during normal operation. It was found, however, that the sensitivity of the coating stress to the CO production in the kernel was large. The CO production is expected to be higher in DB fuel than in UO2 fuel, but its exact level has a high uncertainty. Furthermore, in the fuel performance analysis transient conditions were not yet taken into account. The effort of this task in FY 2010 has focused on the optimization of the core to maximize the pebble discharge burnup level, while retaining its inherent safety characteristics. Using generic pebble bed reactor cores, this task will perform physics calculations to evaluate the capabilities of the pebble bed reactor to perform utilization and destruction of LWR used-fuel transuranics. The task will use established benchmarked models, and will introduce modeling advancements appropriate to the nature of the fuel considered (high transuranic [TRU] content and high burn-up). Accomplishments of this work include: •Core analysis of a HTR-MODULE design loaded with Deep-Burn fuel. •Core analysis of a HTR-MODULE design loaded with Deep-Burn fuel and Uranium. •Core analysis of a HTR-MODULE design loaded with Deep-Burn fuel and Modified Open Cycle Components. •Core analysis of a HTR-MODULE design loaded with Deep-Burn fuel and Americium targets.
U.S. Navy established the Wave Energy Test Site (WETS) in the waters off Marine Corps Base Hawaii, the United States’ first grid-connected site, to the nascent commercial wave energy sector which is critically dependent on available test infrastructure to address challenges with developing Wave Energy Conversion (WEC) devices. WETS consists of test berths at 30m, 60m, and 80m water depths, and can host point absorber and oscillating water column devices to a peak power of 1 MW. HNEI provides key research support to this national effort in the form of environmental monitoring, independent WEC device performance analysis, and critical marine logistical support.
Correction for ‘Pericyclic reaction benchmarks: hierarchical computations targeting CCSDT(Q)/CBS and analysis of DFT performance’ by Pascal Vermeeren et al. , Phys. Chem. Chem. Phys. , 2022, 24 , 18028–18042, https://doi.org/10.1039/D2CP02234F.
Storage systems are complex, including multiple subsystems and components. Sustained operations with top performance require all these subsystems and components working as expected. Having a detailed performance profile helps establishing a baseline. This baseline can be used for easier identification of possible future problems. A systematic bottom-to-top approach, starting with a detailed performance analysis of disks and moving up across layers and subsystems, provides a quantitative breakdown of each component's capabilities and bottlenecks. Coupling these low-level tests with Lustre-level evaluations will present a better understanding of performance expectations under different I/O workloads.
The assessment of resection margins during surgery of oral cavity squamous cell cancer (OCSCC) dramatically impacts the prognosis of the patient as well as the need for adjuvant treatment in the future. Currently there is an unmet need to improve OCSCC surgical margins which appear to be involved in around 45% cases. Intraoperative imaging techniques, magnetic resonance imaging (MRI) and intraoral ultrasound (ioUS), have emerged as promising tools in guiding surgical resection, although the number of studies available on this subject is still low. The aim of this diagnostic test accuracy (DTA) review is to investigate the accuracy of intraoperative imaging in the assessment of OCSCC margins. By using the Cochrane-supported platform Review Manager version 5.4, a systematic search was performed on the online databases MEDLINE-EMBASE-CENTRAL using the keywords “oral cavity cancer, squamous cell carcinoma, tongue cancer, surgical margins, magnetic resonance imaging, intraoperative, intra-oral ultrasound”. Ten papers were identified for full-text analysis. The negative predictive value (cutoff < 5 mm) for ioUS ranged from 0.55 to 0.91, that of MRI ranged from 0.5 to 0.91; accuracy analysis performed on four selected studies showed a sensitivity ranging from 0.07 to 0.75 and specificity ranging from 0.81 to 1. Image guidance allowed for a mean improvement in free margin resection of 35%. IoUS shows comparable accuracy to that of ex vivo MRI for the assessment of close and involved surgical margins, and should be preferred as the more affordable and reproducible technique. Both techniques showed higher diagnostic yield if applied to early OCSCC (T1–T2 stages), and when histology is favorable.
Throughout pyroprocessing efforts at the Hot Fuel Examination Facility (HFEF), insoluble material has accumulated in the Electrorefiner (ER) vessel. The objective of this effort was to analyze the accumulated material to determine its origin. Material was removed from the ER salt bath and distilled to remove excess salt prior to performing analysis. Three samples were sent for chemical and isotopic analysis and underwent an ethyl acetate-bromine dissolution to segregate the oxide fraction from the metal fraction. Actinide concentrations were determined using a quadrupole–inductively coupled plasma–mass spectrometer (Q-ICP-MS). Three additional samples were sent for morphologic and elemental analysis by scanning electron microscopy (SEM) with Energy Dispersive X-ray (EDX) analysis at the Irradiated Materials Characterization Laboratory (IMCL). Analyses suggest that the material is primarily composed of UO2. with a small fraction of metal. This study draws no single conclusion as to the origin of the insoluble material in the ER. The particle sizes and morphologies observed in SEM micrographs indicate that the larger particles observed may be a result of introducing material to the ER that has not been completely reduced in Oxide Reduction (OR) operations, which precede electrorefining. Smaller particles may be due to reactions with oxygen and moisture present in HFEF. This research suggests that microscopic analysis of fuel particles before and after OR operations (including after distillation) as well as the uranium product collected on the cathode in the ER would increase understanding about particle morphology within the pyrochemical process.
The Distribution Transformation Tool (DiTTo) provides a framework for converting between and modifying several different types of distribution model formats . These include commercial formats such as CYME, Synergi, DEW, ePHASORSIM as well as open source formats such as OpenDSS and GridLAB-D. DiTTo consists of an API which represents electrical equipment, configurations, settings and timeseries data. This API acts as an intermediary between different distribution model formats. Several modules for reading and writing the models connect to this API to facilitate easy conversion between formats. Once feeder models are converted into the intermediate API, analysis can be done on the feeder using a network module. The modification module allows changes to be made to the model in the intermediate API. This allows a clear workflow that supports reading a model from one format, performing analysis, making modifications and writing the model to a different format.
Guidance has been requested from the Nuclear Criticality Safety Division (NCSD) regarding processes that involve 520 grams of fissionable material or less. This Level-3 evaluation was conducted and documented in accordance with NCS-AP-004 (Ref. 1), formerly NCS-GUIDE-01. This evaluation is being written as a generic evaluation for all operations that will be able to operate using a 520-gram mass limit. Implementation for specific operations will be performed using a Level 1 CSED, which will confirm and document that this CSED can be used for the specific operation as discussed in NCS-MEMO-17-007 (Ref. 2). This Level 3 CSED updates and supersedes the analysis performed in NCS-TECH-14-014 (Ref. 3).
Concentrating solar thermal (CST) technology has significant potential mainly due to its dispatchability and low cost of storage. However, to compete with other sources, including utility-scale solar PV, its final cost (cents/kWh) still needs to be lowered. Cost reduction can be achieved by improving the system level efficiency of the CST plants through the deployment of advanced power cycles, which operate at high temperatures of ~700°C. However, optical, and thermal losses pose a major challenge to the efficiency of such CST systems. The overall aim of this project is to investigate and de-risk a linear solar receiver concept called an Aerogel Insulated Receiver (AIR) that generates high temperatures (up to 700°C) at a low solar concentration ratio (<100) and with a high collection efficiency (optical × receiver). Our prior work has demonstrated the thermal stability1 and optical and heat-insulating properties2,3 of transparent aerogel insulation at a one-inch scale. The focus of this work is on (1) co-optimization of the geometry of the aerogel tile and receiver enclosure to fit a standard parabolic collector (PTC), (2) scale-up of aerogels into 4-inch tiles while preserving key properties, (3) experimental measurement of receiver heat loss (W/m) in a >70-cm test stand and validation of anticipated receiver performance at high temperatures. Regarding (1), appropriate optical and thermal models for a parabolic trough receiver (PTR) are developed and validated. The geometry of the aerogels and the receiver enclosure are co-optimized to maximize the collection efficiency. The model predicts a 54% collection efficiency at 700°C for an AIR design based on flat aerogels. By combining the collection efficiency with the power block efficiency of supercritical CO 2 cycles, we predict >10% improvements in peak plant efficiency relative to existing line-focusing CST systems. The application of curved plasmonic aerogels is predicted to further increase the collection efficiency to 64%. Regarding (2), we demonstrate the successful development of 6-inch-long refractory aerogel tiles with optical, thermal, and stability characteristics consistent with our prior work. This scale-up requires a transition to a larger ALD station and modifying the ALD process variables such as exposure time and the number of precursor doses. Regarding (3), we design and develop an AIR test stand measuring 3 feet in length. Heat loss performance analysis is carried out using the test stand. The results indicate that aerogel insulation can significantly reduce receiver thermal losses at the high operating temperatures required for next-generation PTRs. The experimental results agree with the heat loss performance predicted by our receiver model. Lastly, we conducted preliminary failure mode and effects (FMEA) and techno-economic (TEA) analyses to identify failure mitigation strategies and commercial opportunities, respectively. Overall, this project identifies key opportunities and challenges in deploying aerogel insulating receivers in next-generation line-focusing CST technologies.
Increasing the fuel discharge burnup of current light water reactors (LWRs) promises reductions in fuel cycle and/or operations costs. By assuming a constant core power density, the economic gain is enabled by better fuel utilization and/or an increased capacity factor. In this effort to investigate greater than 62 MWd/kgU maximum rod average burnup for 110+ kW/l core power density, two core designs have been developed for a standard 17x17, 193 fuel assemblies pressurized water reactor (PWR). The levelized unit cost methodology is employed to evaluate fuel cycle, operation and maintenance, and capital cost impacts and to examine the economic viability of both core design pathways. Core design and optimization are performed using the commercial STUDSVIK code package. Fuel performance analysis is realized in full core configuration via auditing FRAPCON4.1, FAST1.2, and the high-fidelity code BISON. To provide a realistic assessment, the core design process takes into consideration best practices in current PWR core design. It features acceptable performance in terms of various core design constraints on maximum allowable peaking and boron concentration. Gadolinia (Gd2O3) is used as a burnable poison with a maximum of 9 wt% concentration while feeding 89 or 77 fuel assemblies in a 3-batch refueling scheme. Full core fuel performance simulation, which allows for characterization of relevant fuel temperatures, plenum pressures, stresses, and strains, is performed with respect to two bounding burnup levels. Such performance is potentially licensable for the 18-month high burnup core (<68 MWd/kgU peak pin), while it is more challenging for the 24-month high burnup core design pathway (<75 MWd/kgU peak pin). Maximum rod plenum pressure is identified as the most limiting fuel performance parameter. Here, while the scope of the present study focuses on the steady-state plus overpower conditions, the acceptability of the new discharge burnup has to be further assessed by considering uncertainties and impacts under accident scenarios in the future.
This analysis performs validation for the Virtual Environment for Reaction Application’s (VERA’s) neutronics codes MPACT and Shift against measured data from the Seven Percent Critical Experiment (7uPCX) at Sandia National Laboratory. The benchmarking supports the future application of VERA for analysis of high-assay low-enriched uranium (HALEU) fuel in commercial pressurized water reactors (PWRs). VERA demonstrates very good agreement with measured data, and the deterministic code MPACT also agrees very well with higher-fidelity stochastic methods KENO-VI and Shift in both total neutron flux and fission rate distribution comparisons. Despite some significant differences between the 7uPCX and existing commercial PWRs that challenge the neutron cross section data and transport solver methods used by MPACT, all of the results are excellent. The agreement increases the confidence in the applicability and accuracy of VERA for reactor analysis of HALEU fuels in commercial reactors.
The performance of two continuous energy Monte Carlo transport codes, COG11.1 and MCNP6.1, has been compared for the {sup 239}Pu Jezebel Benchmark using cross sections based on the ENDF/B-VII.1 evaluated nuclear data library. The purpose of this study is to determine the length of time it takes each code to complete a series of criticality calculations, using two different methods to model a solid of revolution. A solid of revolution is essential in many fields of engineering and mathematics, and has been a standard feature in computer-aided design and manufacturing since the 1970's. In the field of criticality safety, a solid of revolution can be used to reproduce and simplify complex engineering drawings and systems that would otherwise require excessive use of geometric primitives and computational resources to achieve a comparable level of fidelity. There are two basic methods to model a solid of revolution: Method 1 (COG) specifies a set of x-y pairs or polar coordinates and rotate the curve about the x-axis. Method 2 (COG and MCNP) creates a set of end-to-end conical frustums and/or cylinders. Method 1 is only available in COG and can be implemented using one surface and sector specification. Method 2 can be implemented in both COG and MCNP. This time study was initially conceived as a result of criticality safety analysis performed for a complex system with a high degree of rotational symmetry. Each COG input deck typically took 1-2 days to run in serial mode. COG input decks were then converted to MCNP format and run on the same computer cluster. Instead of expected 1-2 day runtimes, it took MCNP6.1 more than one week to complete each run. Since MCNP typically runs faster than many codes, these lengthy runtimes were considered unusual. This study attempts to investigate the runtime discrepancy between the two codes. The methods and benchmark in this time study were selected to provide a consistent, generalized approach to modeling an arbitrary solid of revolution. While there are many different ways to model the {sup 239}Pu Jezebel Benchmark, the focus of this study is on investigating the practical difference between the two basic methods to model a surface of revolution. Throughout this study, the only code that maintains reasonably low runtimes is COG11.1 (Method 1). Based on the observed performance of each code and method, it is likely that the divergent runtimes are the result of tracking particles across planes that lie between conical frustums, as opposed to defining the object boundary as a single surface. Slight fluctuations in graphed results were also observed, due to random sampling and CPU loading. This was confirmed by re-running input decks and comparing average runtimes to a preliminary curve fit. From the perspective of a criticality safety analyst, the results of this study are expected to contribute towards understanding one of the practical trade-offs between codes and methods as they apply to real-world analysis. Ultimately, it's about choosing the right tool for the job.