Revealing the Mechanism of Magnetic Domain Formation by Topological Data Analysis

arxiv(2022)

引用 0|浏览5
暂无评分
摘要
Understanding pattern formation processes with long-range interactions that result in nonuniform and nonperiodic structures is important for revealing the physical properties of widely observed complex systems in nature. In many cases, pattern formation processes are measured on the basis of spatially coarse-grained features of the smallest elements, that is, the pixel space of a grayscale image. By binarizing the data using a common threshold, one can analyze such grayscale image data using geometrical features, such as the number of domains (Betti number) or domain size. Because such topological features are indicators for evaluating all regions uniformly, it is difficult to determine the nature of each region in a nonuniform structure. Moreover, binarization leads to loss of valuable information such as fluctuations in the domain. Here, we show that the analysis method based on persistent homology provides effective knowledge to understand such pattern formation processes. Specifically, we analyze the pattern formation process in ferromagnetic materials under a rapidly sweeping external magnetic field and discover not only existing types of pattern formation dynamics but also novel ones. On the basis of results of topological data analysis, we also found a candidate reduction model to describe the mechanics of these novel magnetic domain formations. The analysis procedure with persistent homology presented in this paper provides one format applicable to extracting scientific knowledge from a wide range of pattern formation processes with long-range interactions.
更多
查看译文
关键词
magnetic domain formation,topological data
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要