Search NASA⌕ Search

SEARCH · Search NASA

Results for “benchmark problem”

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 1,459 records · Page 81

Prediction of PFC Plasma Fluxes by Improved Edge/Scrape-off-layer Simulations

Large, localized plasma heat exhaust, subsequent inward transport of eroded impurities, and distribution of tritium to walls comprise one of the most critical class of problems for the development of tokamak fusion reactors. The magnitude and temporal duration of the heat fluxes is controlled by two factors: (1) the plasma power coming into the edge from the core, and (2), the physics processes in the edge/scrape-off-layer (SOL) that distribute the power to Plasma Facing Components (PFCs), both in space and in time. Given that the plasma power is largely determined by the fusion power desired, here I address model development needs for item (2), the distribution of power to PFCs, which naturally carries with it the capability for transport of impurities and tritium. Another key issue not addressed here is the impact of edge plasma transport on the plasma pedestal parameters. The nature of plasma transport in the edge/SOL region has long be differentiated from that in the bulk core, initially because of the larger fluctuation amplitudes that are observed, with density fluctuations relative to the time-average sometimes approaching as high as unity in the SOL. More recent measurements have shown addition effects such as strong intermittency, filamentation, toroidal asymmetry, and large flows [1]. These characteristics have a direct impact on plasma energy and particle fluxes to PFCs and on the flow of impurities in the edge. The theory of the edge/SOL is complicated by the steep gradients, multi-dimensional nature of plasma/neutral variations, and as mentioned above, the strong relative fluctuation levels compared to the core region. Furthermore, the strong interaction of the plasma with neutrals and the associated radiative effects for partially ionized plasma are important, Consequently, theoretical models typically need to be, or should be, more complicated, which may be one reason model development has lagged that in the core (funding being another reason). Simulation models of the edge/SOL can be divided into transport codes, which give the slow evolution of the plasma profiles and fluxes in a complex environment, and turbulence codes, which model the unstable, strongly fluctuating plasma state and thus (ideally) provide transport coefficients to the transport codes; this latter connect has only been made for a very limited set of simulations. The edge transport codes are either 2D fluid or now emerging 4D (2r,2v) kinetic. The turbulence codes are 2D and 3D fluid (only one U.S. 3D code spans the separatrix), and emerging 5D gyrokinetic codes. Many of the transport simulations to model plasma fluxes use ad hoc radial transport coefficients, rendering them more in the class of 'interpretive' simulations, not predictive; this limitation makes fundamental prediction for heat-flux widths and edge/SOL transport in general unobtainable at present. Even if there was a strong connection made between transport and turbulence codes, there is not yet a clear consensus in the community concerning the controlling instabilities and turbulence features, even for fixed plasma profiles. In comparison, the core turbulence simulations/theory/experiment have standard benchmarks (e.g., the Cyclone case), agree on many aspects of the big picture of the dominant ion transport, and have made good progress on electron transport. Again this core/edge difference can at least partially be traced to both the inherent difficulty of the edge problem and the stronger historical support for core simulation development; as the edge issue for high-power devices increases in importance, the funding prior should be change. No theory/simulation can be trusted without substantial experimental validation. The edge/SOL area is physically more accessible, but also more complex (2D-3D variations) and also different plasma parameters than for core diagnostics. Among the valuable diagnostics for transport are IRTV, thermocouples, Langmuir probes (but with maximum incident plasma power limitations), Thomson scattering, and visible. For turbulence diagnostics, we have spectroscopy, beam-emission spectroscopy (BES), again Langmuir probes, microwave backscatter, and gas-puff imaging. This list can be extended, but the point is that while significant diagnostics exist, improvements are clearly needed; for example, the characterization of the turbulence spanning the separatrix is modest except for BES at the midplane (probes have difficulty crossing the separatrix owing to heat-flux limitations). The residual turbulence in H-mode is generally not well characterized because of its lower (but still important?) magnitude.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

Stable and radioactive carbon isotope partitioning in soils and saturated systems: a reactive transport modeling benchmark study

This benchmark provides the first rigorous test of a three-isotope system [ 12 C, 13 C, and 14 C] subject to the combined effects of radioactive decay and both stable equilibrium and kinetic fractionation. We present a series of problems building in complexity based on the cycling of carbon in both organic and inorganic forms. The key components implement (1) equilibrium fractionation between multiple coexisting carbon species as a function of pH, (2) radioactive decay of radiocarbon with associated mass-dependent speciation demonstrating appropriate correction of the Δ 14 C value in agreement with reporting convention, and (3) kinetic stable isotope fractionation due to the oxidation of organic carbon to inorganic forms as a function of time and space in an open, through-flowing system. Participating RTM codes are CrunchTope, ToughReact, Hytec, and The Geochemist’s Workbench. Across all problem levels, simulation results from all RTMs demonstrate good agreement.

58 GEOSCIENCES↗

Low- and high-order accurate boundary conditions: From Stokes to Darcy porous flow modeled with standard and improved Brinkman lattice Boltzmann schemes

The present contribution focuses on the accuracy of reflection-type boundary conditions in the Stokes–Brinkman–Darcy modeling of porous flows solved with the lattice Boltzmann method (LBM), which we operate with the two-relaxation-time (TRT) collision and the Brinkman-force based scheme (BF), called BF-TRT scheme. In parallel, we compare it with the Stokes–Brinkman–Darcy linear finite element method (FEM) where the Dirichlet boundary conditions are enforced on grid vertices. In bulk, both BF-TRT and FEM share the same defect: in their discretization a correction to the modeled Brinkman equation appears, given by the discrete Laplacian of the velocity-proportional resistance force. This correction modifies the effective Brinkman viscosity, playing a crucial role in the triggering of spurious oscillations in the bulk solution. While the exact form of this defect is available in lattice-aligned, straight or diagonal, flows; in arbitrary flow/lattice orientations its approximation is constructed. At boundaries, we verify that such a Brinkman viscosity correction has an even more harmful impact. Already at the first order, it shifts the location of the no-slip wall condition supported by traditional LBM boundary schemes, such as the bounce-back rule. For that reason, this work develops a new class of boundary schemes to prescribe the Dirichlet velocity condition at an arbitrary wall/boundary-node distance and that supports a higher order accuracy in the accommodation of the TRT-Brinkman solutions. For their modeling, we consider the standard BF scheme and its improved version, called IBF; this latter is generalized in this work to suppress or to reduce the viscosity correction in arbitrarily oriented flows. Our framework extends the one- and two-point families of linear and parabolic link-wise boundary schemes, respectively called B-LI and B-MLI, which avoid the interference of the Brinkman viscosity correction in their closure relations. The performance of LBM and FEM is thoroughly evaluated in three benchmark tests, which are run throughout three distinctive permeability regimes. The first configuration is a horizontal porous channel, studied with a symbolic approach, where we construct the exact solutions of FEM and BF/IBF with different boundary schemes. The second problem refers to an inclined porous channel flow, which brings in as new challenge the formation of spurious boundary layers in LBM; that is, numerical artefacts that arise due to a deficient accommodation of the bulk solution by the low-accurate boundary scheme. The third problem considers a porous flow past a periodic square array of solid cylinders, which intensifies the previous two tests with the simulation of a more complex flow pattern. The ensemble of numerical tests provides guidelines on the effect of grid resolution and the TRT free collision parameter over the accuracy and the quality of the velocity field, spanning from Stokes to Darcy permeability regimes. It is shown that, with the use of the high-order accurate boundary schemes, the simple, uniform-mesh-based TRT-LBM formulation can even surpass the accuracy of FEM employing hardworking body-fitted meshes.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

DINO: Divergent node cloning for sustained redundancy in HPC

Complexity and scale of next generation HPC systems pose significant challenges in fault resilience methods such that contemporary checkpoint/restart (C/R) methods that address fail-stop behavior may be insufficient. Redundant computing has been proposed as an alternative at extreme scale. Triple redundancy has an advantage over C/R in that it can also detect silent data corruption (SDC) and then correct results via voting. However, current redundant computing approaches do not repair failed or corrupted replicas. Consequently, SDCs can no longer be detected after a replica failure since the system has been degraded to dual redundancy without voting capability. Hence, a job may have to be aborted if voting uncovers mismatching results between the remaining two replicas. And while replicas are logically equivalent, they may have divergent runtime states during job execution, which presents a challenge to simply creating new replicas dynamically. This problem is addressed by, DIvergent NOde cloning (DINO), a redundant execution environment that quickly recovers from hard failures. DINO consists of a novel node cloning service integrated into the MPI runtime system that solves the problem of consolidating divergent states among replicas on-the-fly. With DINO, after degradation to dual redundancy, a good replica can be quickly cloned so that triple redundancy is restored. We present experimental results over 9 NAS Parallel Benchmarks (NPB), Sweep3D and LULESH. Results confirm the applicability of the approach and the correctness of the recovery process and indicate that DINO can recover from failures nearly instantly. The cloning overhead depends on the process image size that needs to be transferred between source and destination of the clone operation and varies between 5.60 to 90.48 s. Here, simulation results with our model show that dual redundancy with DINO recovery always outperforms 2x and surpasses 3x redundancy on up to 1 million nodes. To the best of our knowledge, the design and implementation for repairing failed replicas in redundant MPI computing is unprecedented.

97 MATHEMATICS AND COMPUTING↗

Lensing simulations by Taylor expansion — not so inefficient after all

Cosmic Microwave Background lensing simulation by Taylor expansion has long been considered impractical due to slow convergence, but a recent flat-sky implementation shows that a simple trick eliminates this problem, making Taylor lensing a fast and simple lensing algorithm for the flat sky. Here we generalize the method to the full sky, and study its convergence and performance relative to a commonly used numerical code, Lenspix, with extensive benchmarks of both. Compared to the flat sky case, the method takes a speed hit due to the slow speed of spherical harmonic transforms compared to fast Fourier transforms, resulting in speeds of 1/3 to 2/3 of Lenspix for similar accuracy.

79 ASTRONOMY AND ASTROPHYSICS↗

Pre-conditioned BFGS-based uncertainty quantification in elastic full-waveform inversion

SUMMARY Full-waveform inversion has become an essential technique for mapping geophysical subsurface structures. However, proper uncertainty quantification is often lacking in current applications. In theory, uncertainty quantification is related to the inverse Hessian (or the posterior covariance matrix). Even for common geophysical inverse problems its calculation is beyond the computational and storage capacities of the largest high-performance computing systems. In this study, we amend the Broyden–Fletcher–Goldfarb–Shanno (BFGS) algorithm to perform uncertainty quantification for large-scale applications. For seismic inverse problems, the limited-memory BFGS (L-BFGS) method prevails as the most efficient quasi-Newton method. We aim to augment it further to obtain an approximate inverse Hessian for uncertainty quantification in FWI. To facilitate retrieval of the inverse Hessian, we combine BFGS (essentially a full-history L-BFGS) with randomized singular value decomposition to determine a low-rank approximation of the inverse Hessian. Setting the rank number equal to the number of iterations makes this solution efficient and memory-affordable even for large-scale problems. Furthermore, based on the Gauss–Newton method, we formulate different initial, diagonal Hessian matrices as pre-conditioners for the inverse scheme and compare their performances in elastic FWI applications. We highlight our approach with the elastic Marmousi benchmark model, demonstrating the applicability of pre-conditioned BFGS for large-scale FWI and uncertainty quantification.

58 GEOSCIENCES↗

Toward scalable bound-to-resonance extrapolations for few- and many-body systems

In open quantum many-body systems, the theoretical description of resonant states of many particles strongly coupled to the continuum can be challenging. Such states are commonplace in, for example, exotic nuclei and hadrons, and can reveal important information about the underlying forces at play in these systems. In this work, we demonstrate that the complex-augmented eigenvector continuation (CA-EC) method, originally formulated for the two-body problem with uniform complex scaling, can reliably perform bound-to-resonance extrapolations for genuine three-body resonances having no bound subsystems. Here, we first establish that three-body bound-to-resonance extrapolations are possible by benchmarking different few-body approaches, and we provide arguments to explain how the extrapolation works in the many-body case. We furthermore pave the way towards scalable resonance extrapolations in many-body systems by showing that the CA-EC method also works in the Berggren basis, studying a realistic application using the Gamow shell model.

Ab initio calculations↗

GPCNeT

Network congestion is one of the biggest problems facing HPC systems today, affecting system throughput, performance, user experience, and reproducibility. Congestion manifests as run-to-run variability due to contention for shared resources (e.g., filesystems) or routes between compute endpoints. Despite its significance, current network benchmarks fail to proxy the real-world network utilization seen on congested systems. We propose a new open-source benchmark suite called the Global Performance and Congestion Network Tests (GPCNeT) to advance the state of the practice in this area. The guiding principles used in designing GPCNeT are described and the methodology employed to maximize its utility is presented. The capabilities of GPCNeT are evaluated by analyzing results from several world's largest HPC systems, including an evaluation of congestion management on a next-generation network. The results show that systems of all technologies and scales are susceptible to congestion and this work motivates the need for congestion control in next-generation networks.

Chunduri, Sudheer↗

AGILE GETRS

AGILE GETRS provides an optimized implementation of DETRS that is faster than vendor optimized libraries for small (< 4000) problem sizes. These small problems sizes are of interest to researchers in the Grid Research Integration and Deployment Center (GRID-C) researching real-time power electronics simulations. In addition to our implementation of DETRS, this code provides a test using Google Benchmark which generates matrices of various sizes, fills them with random numbers, runs GETRS, verifies the solution is correct and reports average runtime and other performance metrics.

Hahn, Steven [Oak Ridge National Laboratory (ORNL)↗

Transport in Stochastic Media with Random Chord Length Distributions

Thermal radiation transport computations in binary Markovian random mixtures rely almost exclusively on the Levermore-Pomraning (LP) model which is obtained by applying a heuristic closure to the ensemble averaged random medium transport equation. The validity of this model has been extensively tested by comparing numerical results over a broad parameter range (material types and mixing parameters) against benchmark solutions in planar geometry. The conditions under which the LP-model provides useful results and when it breaks down are now well established, but work to date has been largely restricted to homogeneous mixing statistics, i.e., the mean chord lengths of both materials are taken to be spatially constant. In recent work, this limitation was relaxed by allowing the mean chord lengths and, in a consistent fashion, the volume fractions in the LP-model to vary spatially and in a follow-up investigation benchmark solutions were obtained by ensemble averaging results over material realizations sampled from a nonhomogeneous Poisson process (NHPP). Numerical experiments in rod geometry with specifically linear and quadratic spatial dependence of chord lengths showed that the material averaged radiation intensities vary nonmonotonically with depth into the medium, in stark contrast to solutions obtained assuming uniform chord lengths. Moreover, depending on the local optical thickness and strength of scattering, the LP-model results showed locally more nuanced deviations from the benchmark solutions than was the case with constant chord lengths. These limited numerical investigations highlight the nontrivial qualitative and quantitative consequences of nonhomogeneous mixing statistics, in particular that closure approximations may not be uniformly valid or invalid over the problem domain.

42 ENGINEERING↗

Quantum-assisted quantum compiling

Compiling quantum algorithms for near-term quantum computers (accounting for connectivity and native gate alphabets) is a major challenge that has received significant attention both by industry and academia. Avoiding the exponential overhead of classical simulation of quantum dynamics will allow compilation of larger algorithms, and a strategy for this is to evaluate an algorithm's cost on a quantum computer. To this end, we propose a variational hybrid quantum-classical algorithm called quantum-assisted quantum compiling (QAQC). In QAQC, we use the overlap between a target unitary U and a trainable unitary V as the cost function to be evaluated on the quantum computer. More precisely, to ensure that QAQC scales well with problem size, our cost involves not only the global overlap T r ( V † U ) but also the local overlaps with respect to individual qubits. We introduce novel short-depth quantum circuits to quantify the terms in our cost function, and we prove that our cost cannot be efficiently approximated with a classical algorithm under reasonable complexity assumptions. We present both gradient-free and gradient-based approaches to minimizing this cost. As a demonstration of QAQC, we compile various one-qubit gates on IBM's and Rigetti's quantum computers into their respective native gate alphabets. Furthermore, we successfully simulate QAQC up to a problem size of 9 qubits, and these simulations highlight both the scalability of our cost function as well as the noise resilience of QAQC. Future applications of QAQC include algorithm depth compression, black-box compiling, noise mitigation, and benchmarking.

97 MATHEMATICS AND COMPUTING↗

Implementation of Programmatic Quality and the Impact on Safety

The purpose of this paper is to discuss the implementation of a programmatic quality assurance discipline within the International Space Station Program and the resulting impact on safety. NASA culture has continued to stress safety at the expense of quality when both are extremely important and both can equally influence the success or failure of a Program or Mission. Although safety was heavily criticized in the media after Col~imbiaa, strong case can be made that it was the failure of quality processes and quality assurance in all processes that eventually led to the Columbia accident. Consequently, it is possible to have good quality processes without safety, but it is impossible to have good safety processes without quality. The ISS Program quality assurance function was analyzed as representative of the long-term manned missions that are consistent with the President s Vision for Space Exploration. Background topics are as follows: The quality assurance organizational structure within the ISS Program and the interrelationships between various internal and external organizations. ISS Program quality roles and responsibilities with respect to internal Program Offices and other external organizations such as the Shuttle Program, JSC Directorates, NASA Headquarters, NASA Contractors, other NASA Centers, and International Partner/participants will be addressed. A detailed analysis of implemented quality assurance responsibilities and functions with respect to NASA Headquarters, the JSC S&MA Directorate, and the ISS Program will be presented. Discussions topics are as follows: A comparison of quality and safety resources in terms of staffing, training, experience, and certifications. A benchmark assessment of the lessons learned from the Columbia Accident Investigation (CAB) Report (and follow-up reports and assessments), NASA Benchmarking, and traditional quality assurance activities against ISS quality procedures and practices. The lack of a coherent operational and sustaining quality assurance strategy for long-term manned space flight. An analysis of the ISS waiver processes and the Problem Reporting and Corrective Action (PRACA) process implemented as quality functions. Impact of current ISS Program procedures and practices with regards to operational safety and risk A discussion regarding a "defense-in-depth" approach to quality functions will be provided to address the issue of "integration vs independence" with respect to the roles of Programs, NASA Centers, and NASA Headquarters. Generic recommendations are offered to address the inadequacies identified in the implementation of ISS quality assurance. A reassessment by the NASA community regarding the importance of a "quality culture" as a component within a larger "safety culture" will generate a more effective and value-added functionality that will ultimately enhance safety.

Huls, Dale Thomas↗

Speed and Memory Improvements to MCNP6 Delayed-Gamma Line Treatment

The delayed-gamma (DG) line-emission (line) sampling was first introduced into MCNPX in 2005. At the time, storage demands for problems with large numbers of DG lines was a major concern. In order to accommodate large amounts of line data while keeping the memory footprint of such a problem from exceeding memory limits, complicated logic was included to sift through various storage options. In mid-2015 concerns arose over the adequacy of the MCNP6 DG source option (inherited from MCNPX) with line sampling for modeling high-fidelity problems. One indication that there was a problem came when studying the MCNP6 produced DG spectra of U-238 and its progeny using high-fidelity spectroscopic tallies when compared to DG line amplitudes directly from ENDF/B-VII.1 emission data. Despite producing a correct integral number of gammas, the comparison to ENDF data showed a broadening in the energy lines and a reduction in the amplitude of more prominent lines. It was later found that the cause of this broadening was one of the memory-saving techniques used to store DG line data when problems required many DG lines. Specifically, the DG lines were integrated into fine energy bins for sampling. When a DG line was to be sampled, a random number would be used to choose which energy bin to sample based on emission probability. Once chosen, a second random number was used to perform nearest-line sampling in that bin. The integration of lines led to an oversampling of low-probability lines, especially when in close proximity to high-probability lines. This treatment also caused a similar issue with the time-dependent behavior of DG line emission which occurred when low-probability lines were sampled near high probability lines with the former decay constant being used. The effect on time-dependent gamma production was demonstrated in analytic benchmarks performed by Weldon et al.. To address these issues a new gamma line storage and sampling treatment was introduced in 2015. This new treatment, stores gamma emission data in a line-by-line CDF which is sampled directly, thereby resolving both the line-broadening issue, and the time-dependence issue as was demonstrated by Tutt el al.. In addition, the new DG sampling treatment was written to use dynamically dimensioned storage arrays as opposed to large fixed-dimensioned arrays, as in the former treatment, and also makes more efficient use of line data that is stored by reducing or eliminating repeated data. As a result, storage arrays are generated only for DG data relevant to the problem being executed. The result should be a reduction in both memory requirements and runtime. However, these reductions could not be realized until the logic for the old treatment could be removed. This paper provides an update on recent work which has focused on the removal of the old DG sampling logic and attempts to quantify the memory reduction and speed up of selected problems.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Data and Analytics to Inform Energy Retrofit of High Performance Buildings

Buildings consume more than one-third of the world?s primary energy. Reducing energy use in buildings with energy efficient technologies is feasible and also driven by energy policies such as energy benchmarking, disclosure, rating, and labeling in both the developed and developing countries. Current energy retrofits focus on the existing building stocks, especially older buildings, but the growing number of new high performance buildings built around the world raises a question that how these buildings perform and whether there are retrofit opportunities to further reduce their energy use. This is a new and unique problem for the building industry. Traditional energy audit or analysis methods are inadequate to look deep into the energy use of the high performance buildings. This study aims to tackle this problem with a new holistic approach powered by building performance data and analytics. First, three types of measured data are introduced, including the time series energy use, building systems operating conditions, and indoor and outdoor environmental parameters. An energy data model based on the ISO Standard 12655 is used to represent the energy use in buildings in a three-level hierarchy. Secondly, a suite of analytics were proposed to analyze energy use and to identify retrofit measures for high performance buildings. The data-driven analytics are based on monitored data at short time intervals, and cover three levels of analysis ? energy profiling, benchmarking and diagnostics. Thirdly, the analytics were applied to a high performance building in California to analyze its energy use and identify retrofit opportunities, including: (1) analyzing patterns of major energy end-use categories at various time scales, (2) benchmarking the whole building total energy use as well as major end-uses against its peers, (3) benchmarking the power usage effectiveness for the data center, which is the largest electricity consumer in this building, and (4) diagnosing HVAC equipment using detailed time-series operating data. Finally, a few energy efficiency measures were identified for retrofit, and their energy savings were estimated to be 20percent of the whole-building electricity consumption. Based on the analyses, the building manager took a few steps to improve the operation of fans, chillers, and data centers, which will lead to actual energy savings. This study demonstrated that there are energy retrofit opportunities for high performance buildings and detailed measured building performance data and analytics can help identify and estimate energy savings and to inform the decision making during the retrofit process. Challenges of data collection and analytics were also discussed to shape best practice of retrofitting high performance buildings.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Efficient sensitivity analysis of the thermal profile in powder bed fusion of metals using hypercomplex automatic differentiation finite element method

Rapid cyclic temperature fluctuation occurring in powder bed fusion of metals using a laser beam (PBF-LB/M) influences the formation of flaws in printed parts. Consequently, there is a pressing need to enhance the quality of printed parts by developing innovative methodologies that can predict thermal histories and help uncover the intricate relationships between process parameters and thermal profiles. Sensitivity Analysis (SA) emerges as an essential tool for this, offering the potential for process optimization and enhanced quality control. Nonetheless, conventional SA methodologies often incur in excessive computational costs and potential numerical approximation errors. Here, to address this technical challenge, we present a novel method for SA that integrates the HYPercomplex-based Automatic Differentiation (HYPAD) technique with transient thermal simulations conducted via the finite element method (FEM). Leveraging this methodology, we efficiently and accurately perform SA for PBF-LB/M processes in a post-processing step. Compared to traditional methods like Finite Differences (FD), HYPAD-FEM required 96 % less computational time for obtaining sensitivities for 22 process parameters, under a comparative study conducted within the context of the 2018–02 AM benchmark of the National Institute of Standards and Technology. In summary, HYPAD-FEM offers superior efficiency and accuracy in SA over conventional methods, delivering the best sensitivity of a model without the need for step-size selection and problem or parameter-based implementations.

36 MATERIALS SCIENCE↗

Combustion machine learning: Principles, progress and prospects

Progress in combustion science and engineering has led to the generation of large amounts of data from large-scale simulations, high-resolution experiments, and sensors. This corpus of data offers enormous opportunities for extracting new knowledge and insights—if harnessed effectively. Machine learning (ML) techniques have demonstrated remarkable success in data analytics, thus offering a new paradigm for data-intense analyses and scientific investigations through combustion machine learning (CombML). While data-driven methods are utilized in various combustion areas, recent advances in algorithmic developments, the accessibility of open-source software libraries, the availability of computational resources, and the abundance of data have together rendered ML techniques ubiquitous in scientific analysis and engineering. This article examines ML techniques for applications in combustion science and engineering. Starting with a review of sources of data, data-driven techniques, and concepts, we examine supervised, unsupervised, and semi-supervised ML methods. Various combustion examples are considered to illustrate and to evaluate these methods. Next, we review past and recent applications of ML approaches to problems in combustion, spanning fundamental combustion investigations, propulsion and energy-conversion systems, and fire and explosion hazards. Challenges unique to CombML are discussed and further opportunities are identified, focusing on interpretability, uncertainty quantification, robustness, consistency, creation and curation of benchmark data, and the augmentation of ML methods with prior combustion-domain knowledge.

33 ADVANCED PROPULSION SYSTEMS↗

Charged vanadium-benzene multidecker clusters: DFT and quantum Monte Carlo study

Using explicitly correlated fixed-node quantum Monte Carlo and density functional theory (DFT) methods, we study electronic properties, ground-state multiplets, ionization potentials, electron affinities, and low-energy fragmentation channels of charged half-sandwich and multidecker vanadium-benzene systems with up to 3 vanadium atoms, including both anions and cations. It is shown that, particularly in anions, electronic correlations play a crucial role; these effects are not systematically captured with any commonly used DFT functionals such as gradient corrected, hybrids, and range-separated hybrids. On the other hand, tightly bound cations can be described qualitatively by DFT. A comparison of DFT and quantum Monte Carlo provides an in-depth understanding of the electronic structure and properties of these correlated systems. The calculations also serve as a benchmark study of 3d molecular anions that require a balanced many-body description of correlations at both short- and long-range distances.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Fast and Efficient Stochastic Optimization for Analytic Continuation

In this analytic continuation of imaginary-time quantum Monte Carlo data to extract real-frequency spectra remains a key problem in connecting theory with experiment. Here we present a fast and efficient stochastic optimization method (FESOM) as a more accessible variant of the stochastic optimization method introduced by Mishchenko et al. [Phys. Rev. B 62, 6317 (2000)], and we benchmark the resulting spectra with those obtained by the standard maximum entropy method for three representative test cases, including data taken from studies of the two-dimensional Hubbard model. Genearally, we find that our FESOM approach yields spectra similar to the maximum entropy results. In particular, while the maximum entropy method yields superior results when the quality of the data is strong, we find that FESOM is able to resolve fine structure with more detail when the quality of the data is poor. In addition, because of its stochastic nature, the method provides detailed information on the frequency-dependent uncertainty of the resulting spectra, while the maximum entropy method does so only for the spectral weight integrated over a finite frequency region. Therefore, we believe that this variant of the stochastic optimization approach provides a viable alternative to the routinely used maximum entropy method, especially for data of poor quality.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗