Search NASA⌕ Search

SEARCH · Search NASA

Results for “kernel method”

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 1,369 records · Page 76

Efficient solution of the Wigner–Liouville equation using a spectral decomposition of the force field

The Wigner–Liouville equation is reformulated using a spectral decomposition of the classical force field instead of the potential energy. The latter is shown to simplify the Wigner–Liouville kernel both conceptually and numerically as the spectral force Wigner–Liouville equation avoids the numerical evaluation of the highly oscillatory Wigner kernel which is nonlocal in both position and momentum. The quantum mechanical evolution is instead governed by a term local in space and non-local in momentum, where the non-locality in momentum has only a limited range. An interpretation of the time evolution in terms of two processes is presented; a classical evolution under the influence of the averaged driving field, and a probability-preserving quantum-mechanical generation and annihilation term. Using the inherent stability and reduced complexity, a direct deterministic numerical implementation using Chebyshev and Fourier pseudo-spectral methods is detailed. For the purpose of illustration, we present results for the time-evolution of a one-dimensional resonant tunneling diode driven out of equilibrium.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Fast GPU-Based Generation of Large Graph Networks From Degree Distributions

Synthetically generated, large graph networks serve as useful proxies to real-world networks for many graph-based applications. The ability to generate such networks helps overcome several limitations of real-world networks regarding their number, availability, and access. Here, we present the design, implementation, and performance study of a novel network generator that can produce very large graph networks conforming to any desired degree distribution. The generator is designed and implemented for efficient execution on modern graphics processing units (GPUs). Given an array of desired vertex degrees and number of vertices for each desired degree, our algorithm generates the edges of a random graph that satisfies the input degree distribution. Multiple runtime variants are implemented and tested: 1) a uniform static work assignment using a fixed thread launch scheme, 2) a load-balanced static work assignment also with fixed thread launch but with cost-aware task-to-thread mapping, and 3) a dynamic scheme with multiple GPU kernels asynchronously launched from the CPU. The generation is tested on a range of popular networks such as Twitter and Facebook, representing different scales and skews in degree distributions. Results show that, using our algorithm on a single modern GPU (NVIDIA Volta V100), it is possible to generate large-scale graph networks at rates exceeding 50 billion edges per second for a 69 billion-edge network. GPU profiling confirms high utilization and low branching divergence of our implementation from small to large network sizes. For networks with scattered distributions, we provide a coarsening method that further increases the GPU-based generation speed by up to a factor of 4 on tested input networks with over 45 billion edges.

97 MATHEMATICS AND COMPUTING↗

Preliminary Compendium of the Initial Conditions and Evolution of the Fundamental Interactions of SiC with Ag and Pd

Tri-structural isotropic (TRISO) particles are the fuel of choice for very-high temperature reactors. At the core, these spherical particles consist of an inner fuel kernel of UO{sub 2}, UC{sub 2} or a combination of both, which is coated radially outward by successive layers of low density pyrolytic graphite, an inner high density pyrolytic graphite layer, silicon carbide (SiC) and an outer high density pyrolytic graphite layer. SiC is the main diffusion barrier against the release of fission products due to its hardness and high melting point. Irradiation testing of TRISO particles indicates that trace amounts of metallic fission products can damage or potentially escape through the barrier layer. Surrogate experimental studies have previously been performed by various methods seeking to shed light on and explain the diffusion behavior of Ag and Pd. Although some of these experiments have yielded realistic diffusion coefficients the discrepancy of results between different groups is such that the essential details of the diffusive behavior are still to be fully understood. This work aims to understand the fundamental physical and chemical interactions that take place at the interface of SiC and Ag/Pd, and which give rise to the observed diffusion behavior of these metallic species in TRISO particles. Theoretical simulations of the diffusion of Ag through SiC have been exhaustive but typically yield bulk diffusion coefficients that due to the nature of bulk diffusion phenomena (i.e. slow diffusion) the diffusion coefficients are calculated to be orders of magnitude lower than what experimental observations suggest. This in part is due to the difficulty in simulating realistic models that include various polymorphs of SiC, and the dynamical structural transformations that SiC undergoes in the TRISO environment (which are also to be explained). The result is a particularly challenging problem to solve given that the initial conditions of the interactions of Ag and Pd with SiC are difficult to address, both in theoretical models and surrogate experiments. In order to improve these models the approach taken here is to consider fundamental interactions between Ag and SiC that take place in short time scales by varying the initial environmental conditions, i.e. annealing temperature and time, initial amounts of the metallic species (M) Ag, Pd in SiC/M/SiC multilayers. This work aims to understand the fundamental physical and chemical interactions that take place at the interface of SiC and Ag/Pd, and which give rise to the observed diffusion behavior of these metallic species in TRISO particles. In order to do so we have synthesized SiC/Ag/SiC and SiC/Pd/SiC multilayers of various thicknesses by PLD and characterized them using SEM/EDS. Initial results show that upon annealing Ag agglomerates and ruptures the SiC layer to be released. In the case of Pd a strong chemical interaction is deduced from the slow interlayer diffusion and limited release. Further studies using photoelectron spectroscopy and RHEED will be carried out in order obtain detailed information on the chemical and physical transitions that undergo upon annealing of these systems. (authors)

36 MATERIALS SCIENCE↗

Multitask methods for predicting molecular properties from heterogeneous data

Data generation remains a bottleneck in training surrogate models to predict molecular properties. We demonstrate that multitask Gaussian process regression overcomes this limitation by leveraging both expensive and cheap data sources. In particular, we consider training sets constructed from coupled-cluster (CC) and density functional theory (DFT) data. We report that multitask surrogates can predict at CC-level accuracy with a reduction in data generation cost by over an order of magnitude. Of note, our approach allows the training set to include DFT data generated by a heterogeneous mix of exchange–correlation functionals without imposing any artificial hierarchy on functional accuracy. More generally, the multitask framework can accommodate a wider range of training set structures—including the full disparity between the different levels of fidelity—than existing kernel approaches based on Δ-learning although we show that the accuracy of the two approaches can be similar. Consequently, multitask regression can be a tool for reducing data generation costs even further by opportunistically exploiting existing data sources.

Chemistry↗

Towards a unified nonlocal, peridynamics framework for the coarse-graining of molecular dynamics data with fractures

Molecular dynamics (MD) has served as a powerful tool for designing materials with reduced reliance on laboratory testing. However, the use of MD directly to treat the deformation and failure of materials at the mesoscale is still largely beyond reach. In this work, we propose a learning framework to extract a peridynamics model as a mesoscale continuum surrogate from MD simulated material fracture data sets. Firstly, we develop a novel coarse-graining method, to automatically handle the material fracture and its corresponding discontinuities in the MD displacement data sets. Inspired by the weighted essentially non-oscillatory (WENO) scheme, the key idea lies at an adaptive procedure to automatically choose the locally smoothest stencil, then reconstruct the coarse-grained material displacement field as the piecewise smooth solutions containing discontinuities. Then, based on the coarse-grained MD data, a two-phase optimization-based learning approach is proposed to infer the optimal peridynamics model with damage criterion. In the first phase, we identify the optimal nonlocal kernel function from the data sets without material damage to capture the material stiffness properties. Then, in the second phase, the material damage criterion is learnt as a smoothed step function from the data with fractures. As a result, a peridynamics surrogate is obtained. As a continuum model, our peridynamics surrogate model can be employed in further prediction tasks with different grid resolutions from training, and hence allows for substantial reductions in computational cost compared with MD. We illustrate the efficacy of the proposed approach with several numerical tests for the dynamic crack propagation problem in a single-layer graphene. Our tests show that the proposed data-driven model is robust and generalizable, in the sense that it is capable of modeling the initialization and growth of fractures under discretization and loading settings that are different from the ones used during training.

97 MATHEMATICS AND COMPUTING↗

Dynamical invariants in systems with and without broken time-reversal symmetry

In the first part of the lectures dynamical invariants in classical mechanics and conventional quantum mechanics will be considered. In particular, we will begin with some remarks on classical mechanics and on quantization in order to establish the theory in the form that will be used later on. Starting from the time-dependent Schroedinger equation, the dynamics of Gaussian wave packets and Ermakov invariants, the time-dependent Green function/Feynman kernel, quantum-classical connections, energetics and Lagrange-Hamilton formalism for quantum uncertainties, momentum space representation and the relation between the Wigner function and the Ermakov invariant will be discussed. The representation of canonical transformations in time-independent and time-dependent quantum mechanics, factorization of the Ermakov invariant and generalized creation/annihilation operators will be studied. Subsequently, the time-independent Schroedinger equation, leading to nonlinear quantum mechanics related to Riccati/Ermakov systems as well as the occurrence of Riccati/Ermakov systems in the treatment of Bose-Einstein condensates via the so-called moment method will be analyzed.In part two, irreversible dynamics of dissipative systems, classical and quantum mechanical descriptions and corresponding invariants will be treated. After some general remarks on classical and quantum mechanics with unitary time-evolution and energy conservation, phenomenological Langevin and Fokker--Planck equations, master equations in classical and quantum mechanics and the system-plus-reservoir approach will be mentioned briefly. Then follows a more detailed discussion of modified Schroedinger equations and, particularly, of a nonlinear Schroedinger equation with complex logarithmic nonlinearity; its properties, solutions, invariants and energetics will be studied. Finally, a comparison with a classical description in expanding coordinates will lead to a non-unitary connection between the logarithmic nonlinear Schroedinger equation, the Caldirola-Kanai approach and the expanding coordinate approach.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

MEASURING THE MASS DISTRIBUTION IN GALAXY CLUSTERS

Cluster mass profiles are tests of models of structure formation. Only two current observational methods of determining the mass profile, gravitational lensing, and the caustic technique are independent of the assumption of dynamical equilibrium. Both techniques enable the determination of the extended mass profile at radii beyond the virial radius. For 19 clusters, we compare the mass profile based on the caustic technique with weak lensing measurements taken from the literature. This comparison offers a test of systematic issues in both techniques. Around the virial radius, the two methods of mass estimation agree to within {approx}30%, consistent with the expected errors in the individual techniques. At small radii, the caustic technique overestimates the mass as expected from numerical simulations. The ratio between the lensing profile and the caustic mass profile at these radii suggests that the weak lensing profiles are a good representation of the true mass profile. At radii larger than the virial radius, the extrapolated Navarro, Frenk and White fit to the lensing mass profile exceeds the caustic mass profile. Contamination of the lensing profile by unrelated structures within the lensing kernel may be an issue in some cases; we highlight the clusters MS0906+11 and A750, superposed along the line of sight, to illustrate the potential seriousness of contamination of the weak lensing signal by these unrelated structures.

79 ASTRONOMY AND ASTROPHYSICS↗

Impact of SPECT corrections on 3D-dosimetry for liver transarterial radioembolization using the patient relative calibration methodology

Purpose: Many centers aim to plan liver transarterial radioembolization (TARE) with dosimetry, even without CT-based attenuation correction (AC), or with unoptimized scatter correction (SC) methods. This work investigates the impact of presence vs absence of such corrections, and limited spatial resolution, on 3D dosimetry for TARE. Methods: Three voxelized phantoms were derived from CT images of real patients with different body sizes. Simulations of {sup 99m}Tc-SPECT projections were performed with the SIMIND code, assuming three activity distributions in the liver: uniform, inside a “liver’s segment,” or distributing multiple uptaking nodules (“nonuniform liver”), with a tumoral liver/healthy parenchyma ratio of 5:1. Projection data were reconstructed by a commercial workstation, with OSEM protocol not specifically optimized for dosimetry (spatial resolution of 12.6 mm), with/without SC (optimized, or with parameters predefined by the manufacturer; dual energy window), and with/without AC. Activity in voxels was calculated by a relative calibration, assuming identical microspheres and {sup 99m}Tc-SPECT counts spatial distribution. 3D dose distributions were calculated by convolution with {sup 90}Y voxel S-values, assuming permanent trapping of microspheres. Cumulative dose-volume histograms in lesions and healthy parenchyma from different reconstructions were compared with those obtained from the reference biodistribution (the “gold standard,” GS), assessing differences for D95%, D70%, and D50% (i.e., minimum value of the absorbed dose to a percentage of the irradiated volume). γ tool analysis with tolerance of 3%/13 mm was used to evaluate the agreement between GS and simulated cases. The influence of deep-breathing was studied, blurring the reference biodistributions with a 3D anisotropic gaussian kernel, and performing the simulations once again. Results: Differences of the dosimetric indicators were noticeable in some cases, always negative for lesions and distributed around zero for parenchyma. Application of AC and SC reduced systematically the differences for lesions by 5%–14% for a liver segment, and by 7%–12% for a nonuniform liver. For parenchyma, the data trend was less clear, but the overall range of variability passed from −10%/40% for a liver segment, and −10%/20% for a nonuniform liver, to −13%/6% in both cases. Applying AC, SC with preset parameters gave similar results to optimized SC, as confirmed by γ tool analysis. Moreover, γ analysis confirmed that solely AC and SC are not sufficient to obtain accurate 3D dose distribution. With breathing, the accuracy worsened severely for all dosimetric indicators, above all for lesions: with AC and optimized SC, −38%/−13% in liver’s segment, −61%/−40% in the nonuniform liver. For parenchyma, D50% resulted always less sensitive to breathing and sub-optimal correction methods (difference overall range: −7%/13%). Conclusions: Reconstruction protocol optimization, AC, SC, PVE and respiratory motion corrections should be implemented to obtain the best possible dosimetric accuracy. On the other side, thanks to the relative calibration, D50% inaccuracy for the healthy parenchyma from absence of AC was less than expected, while the optimization of SC was scarcely influent. The relative calibration therefore allows to perform TARE planning, basing on D50% for the healthy parenchyma, even without AC or with suboptimal corrections, rather than rely on nondosimetric methods.

60 APPLIED LIFE SCIENCES↗

SU-E-I-21: Dosimetric Characterization and Image Quality Evaluation of the AIRO Mobile CT Scanner

Purpose: The AIRO Mobile CT system was recently introduced which overcomes the limitations from existing CT, CT fluoroscopy, and intraoperative O-arm. With an integrated table and a large diameter bore, the system is suitable for cranial, spine and trauma procedures, making it a highly versatile intraoperative imaging system. This study is to investigate radiation dose and image quality of the AIRO and compared with those from a routine CT scanner. Methods: Radiation dose was measured using a conventional 100mm pencil ionization chamber and CT polymethylmetacrylate (PMMA) body and head phantoms. Image quality was evaluated with a CATPHAN 500 phantom. Spatial resolution, low contrast resolution (CNR), Modulation Transfer Function (MTF), and Normalized Noise Power Spectrum (NNPS) were analyzed. Results: Under identical technique conditions, radiation dose (mGy/mAs) from the AIRO mobile CT system (AIRO) is higher than that from a 64 slice CT scanner. MTFs show that both Soft and Standard filters of the AIRO system lost resolution quickly compared to the Sensation 64 slice CT. With the Standard kernel, the spatial resolutions of the AIRO system are 3lp/cm and 4lp/cm for the body and head FOVs, respectively. NNPSs show low frequency noise due to ring-like artifacts. Due to a higher dose in terms of mGy/mAs at both head and body FOV, CNR of the AIRO system is higher than that of the Siemens scanner. However detectability of the low contrast objects is poorer in the AIRO due to the presence of ring artifacts in the location of the targets. Conclusion: For image guided surgery applications, the AIRO has some advantages over a routine CT scanner due to its versatility, large bore size, and acceptable image quality. Our evaluation of the physical performance helps its future improvements.

60 APPLIED LIFE SCIENCES↗

Robust Multigrid Smoothers for Three Dimensional Elliptic Equations with Strong Anisotropies

We discuss the behavior of several plane relaxation methods as multigrid smoothers for the solution of a discrete anisotropic elliptic model problem on cell-centered grids. The methods compared are plane Jacobi with damping, plane Jacobi with partial damping, plane Gauss-Seidel, plane zebra Gauss-Seidel, and line Gauss-Seidel. Based on numerical experiments and local mode analysis, we compare the smoothing factor of the different methods in the presence of strong anisotropies. A four-color Gauss-Seidel method is found to have the best numerical and architectural properties of the methods considered in the present work. Although alternating direction plane relaxation schemes are simpler and more robust than other approaches, they are not currently used in industrial and production codes because they require the solution of a two-dimensional problem for each plane in each direction. We verify the theoretical predictions of Thole and Trottenberg that an exact solution of each plane is not necessary and that a single two-dimensional multigrid cycle gives the same result as an exact solution, in much less execution time. Parallelization of the two-dimensional multigrid cycles, the kernel of the three-dimensional implicit solver, is also discussed. Alternating-plane smoothers are found to be highly efficient multigrid smoothers for anisotropic elliptic problems.

Llorente, Ignacio M.↗

Reliable Ignition of LOX-LCH4 Propellants

No cryogenic Reaction Control System (RCS) has ever flown in space. Cryogenic propellants are baselined by HLS and CLPS partners and are the key use case for future ISRU manufactured propellants. A reliable LOX/LCH 4 RCS is an enabling technology for human Lunar and Mars exploration. In previous thermal vacuum (<275 F, <10 torr) testing at NASA GRC Plum Brook, the team uncovered anomalous LOX/LCH4 engine ignition phenomena where flame kernels quenched at these ultra cold hardware temperatures, leading to many pulses where the engines would not light. In 2022, the project was able to recreate the no light condition with an upgraded Frost-Mint test stand at JSC. This project implemented test stand improvements and procured propellants and follow on engine components for a hot fire campaign beginning in November 2024 which will attempt to ignite the RCS engine at ultra cold vacuum conditions by modifying the engine mixture ratio. New modeling techniques were implemented to account for past no lights and are an enabling method to make lunar surface go no go predictions based on hardware conditions. This was enabled by significant improvements to the Frost Mint system, which reduced moisture build up and leaks, increased propellant availability, leading to more attempts at thermal vacuum hot-fires.

Propulsion↗

Co-benefits of global and regional greenhouse gas mitigation for US air quality in 2050

Policies to mitigate greenhouse gas (GHG) emissions will not only slow climate change but can also have ancillary benefits of improved air quality. Here we examine the co-benefits of both global and regional GHG mitigation for US air quality in 2050 at fine resolution, using dynamical downscaling methods, building on a previous global co-benefits study (West et al., 2013). The co-benefits for US air quality are quantified via two mechanisms: through reductions in co-emitted air pollutants from the same sources and by slowing climate change and its influence on air quality, following West et al. (2013). Additionally, we separate the total co-benefits into contributions from domestic GHG mitigation vs. mitigation in foreign countries. We use the Weather Research and Forecasting (WRF) model to dynamically downscale future global climate to the regional scale and the Sparse Matrix Operator Kernel Emissions (SMOKE) program to directly process global anthropogenic emissions to the regional domain, and we provide dynamical boundary conditions from global simulations to the regional Community Multi-scale Air Quality (CMAQ) model. The total co-benefits of global GHG mitigation from the RCP4.5 scenario compared with its reference are estimated to be higher in the eastern US (ranging from 0.6 to 1.0 µg m -3 ) than the west (0–0.4 µg m -3 ) for fine particulate matter (PM 2.5 ), with an average of 0.47 µg m -3 over the US; for O 3 , the total co-benefits are more uniform at 2–5 ppb, with a US average of 3.55 ppb. Comparing the two mechanisms of co-benefits, we find that reductions in co-emitted air pollutants have a much greater influence on both PM 2.5 (96 % of the total co-benefits) and O 3 (89 % of the total) than the second co-benefits mechanism via slowing climate change, consistent with West et al. (2013). GHG mitigation from foreign countries contributes more to the US O 3 reduction (76 % of the total) than that from domestic GHG mitigation only (24 %), highlighting the importance of global methane reductions and the intercontinental transport of air pollutants. For PM 2.5 , the benefits of domestic GHG control are greater (74 % of total). Since foreign contributions to co-benefits can be substantial, with foreign O 3 benefits much larger than those from domestic reductions, previous studies that focus on local or regional co-benefits may greatly underestimate the total co-benefits of global GHG reductions. We conclude that the US can gain significantly greater domestic air quality co-benefits by engaging with other nations to control GHGs.

54 ENVIRONMENTAL SCIENCES↗

Efficient scalable algorithms for hierarchically semiseparable matrices

Hierarchically semiseparable (HSS) matrix algorithms are emerging techniques in constructing the superfast direct solvers for both dense and sparse linear systems. Here, we develope a set of novel parallel algorithms for the key HSS operations that are used for solving large linear systems. These include the parallel rank-revealing QR factorization, the HSS constructions with hierarchical compression, the ULV HSS factorization, and the HSS solutions. The HSS tree based parallelism is fully exploited at the coarse level. The BLACS and ScaLAPACK libraries are used to facilitate the parallel dense kernel operations at the ne-grained level. We have appplied our new parallel HSS-embedded multifrontal solver to the anisotropic Helmholtz equations for seismic imaging, and were able to solve a linear system with 6.4 billion unknowns using 4096 processors, in about 20 minutes. The classical multifrontal solver simply failed due to high demand of memory. To our knowledge, this is the first successful demonstration of employing the HSS algorithms in solving the truly large-scale real-world problems. Our parallel strategies can be easily adapted to the parallelization of the other rank structured methods.

01 COAL, LIGNITE, AND PEAT↗

Scalable Hyper-parameter Estimation for Gaussian Process Based Time Series Analysis

Gaussian process (GP) is increasingly becoming popular as a kernel machine learning tool for non-parametric data analysis. Recently, GP has been applied to model non-linear dependencies in time series data. GP based analysis can be used to solve problems of time series prediction, forecasting, missing data imputation, change point detection, anomaly detection, etc. But the use of GP to handle massive scientific time series data sets has been limited, owing to its expensive computational complexity. The primary bottleneck is the handling of the covariance matrix whose size is quadratic in the length of the time series. In this paper we propose a scalable method that exploit the special structure of the covariance matrix for hyper-parameter estimation in GP based learning. The proposed method allows estimation of hyper parameters associated with GP in quadratic time, which is an order of magnitude improvement over standard methods with cubic complexity. Moreover, the proposed method does not require explicit computation of the covariance matrix and hence has memory requirement linear to the length of the time series as opposed to the quadratic memory requirement of standard methods. To further improve the computational complexity of the proposed method, we provide a parallel version to concurrently estimate the log likelihood for a set of time series which is the key step in the hyper-parameter estimation. Performance results on a multi-core system show that our proposed method provides significant speedups as high as 1000, even when running in serial mode, while maintaining a small memory footprint. The parallel version exploits the natural parallelization potential of the serial algorithm and is shown to perform significantly better than the serial faster algorithm, with speedups as high as 10.

Chandola, Varun↗

QuadConv: Quadrature-based convolutions with applications to non-uniform PDE data compression

We present a new convolution layer for deep learning architectures which we call QuadConv — an approximation to continuous convolution via quadrature. Our operator is developed explicitly for use on non-uniform, mesh-based data, and accomplishes this by learning a continuous kernel that can be sampled at arbitrary locations. Moreover, the construction of our operator admits an efficient implementation which we detail and construct. As an experimental validation of our operator, we consider the task of compressing partial differential equation (PDE) simulation data from fixed meshes. Here, we show that QuadConv can match the performance of standard discrete convolutions on uniform grid data by comparing a QuadConv autoencoder (QCAE) to a standard convolutional autoencoder (CAE). Further, we show that the QCAE can maintain this accuracy even on non-uniform data. In both cases, QuadConv also outperforms alternative unstructured convolution methods such as graph convolution.

Compression↗

Learning the right channel in multimodal imaging: automated experiment in piezoresponse force microscopy

We report the development and experimental implementation of the automated experiment workflows for the identification of the best predictive channel for a phenomenon of interest in spectroscopic measurements. The approach is based on the combination of ensembled deep kernel learning for probabilistic predictions and a basic reinforcement learning policy for channel selection. It allows the identification of which of the available observational channels, sampled sequentially, are most predictive of selected behaviors, and hence have the strongest correlations. We implement this approach for multimodal imaging in piezoresponse force microscopy (PFM), with the behaviors of interest manifesting in piezoresponse spectroscopy. We illustrate the best predictive channel for polarization-voltage hysteresis loop and frequency-voltage hysteresis loop areas is amplitude in the model samples. The same workflow and code are applicable for any multimodal imaging and local characterization methods.

36 MATERIALS SCIENCE↗

Using indium tin oxide material to implement the imaging of microwave plasma ignition process

In this paper, a method is introduced to get global observation of microwave plasma ignition process at high pressure. A microwave resonator was designed with an indium tin oxide coated glass at bottom. Microwave plasma ignition was implemented in methane and air mixture at 10 bars by a 2 ms-3 kW-2.45 GHz microwave pulse, and the high speed images of the ignition process were obtained. The images visually proved that microwave plasma ignition could lead to a multi-point ignition. The system may also be applied to obtain Schlieren images, which is commonly used to observe the development of flame kernel in an ignition process.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗