EPFL, Swiss Plasma Center (SPC), CH-1015 Lausanne, Switzerland.
MIT, Plasma Science and Fusion Center (PSFC), Cambridge, Massachusetts 02139, USA.
Rev Sci Instrum. 2023 Mar 1;94(3):033512. doi: 10.1063/5.0133506.
Filamentary structures, also known as blobs, are a prominent feature of turbulence and transport at the edge of magnetically confined plasmas. They cause cross-field particle and energy transport and are, therefore, of interest in tokamak physics and, more generally, nuclear fusion research. Several experimental techniques have been developed to study their properties. Among these, measurements are routinely performed with stationary probes, passive imaging, and, in more recent years, Gas Puff Imaging (GPI). In this work, we present different analysis techniques developed and used on 2D data from the suite of GPI diagnostics in the Tokamak à Configuration Variable, featuring different temporal and spatial resolutions. Although specifically developed to be used on GPI data, these techniques can be employed to analyze 2D turbulence data presenting intermittent, coherent structures. We focus on size, velocity, and appearance frequency evaluation with, among other methods, conditional averaging sampling, individual structure tracking, and a recently developed machine learning algorithm. We describe in detail the implementation of these techniques, compare them against each other, and comment on the scenarios to which these techniques are best applied and on the requirements that the data must fulfill in order to yield meaningful results.
丝状结构，也称为斑点，是磁约束等离子体边缘湍流和输运的一个突出特征。它们导致了跨场的粒子和能量输运，因此在托卡马克物理学中，更普遍地在核聚变研究中引起了关注。已经开发了几种实验技术来研究它们的性质。其中，常用的方法有使用固定探针、被动成像，以及近年来的气体脉冲成像（GPI）。在这项工作中，我们介绍了为 GPI 诊断套件的二维数据开发并使用的不同分析技术，这些技术具有不同的时间和空间分辨率。尽管这些技术是专门为 GPI 数据开发的，但它们也可以用于分析呈现间歇性、相干结构的二维湍流数据。我们专注于评估大小、速度和出现频率，方法包括条件平均采样、单个结构跟踪以及最近开发的机器学习算法。我们详细描述了这些技术的实现，对它们进行了比较，并评论了这些技术最适用的场景以及数据必须满足的要求，以便产生有意义的结果。
