高斯过程对γ射线暴X射线光变曲线的分析

YNAO OpenIR > 星系类星体研究组

	高斯过程对γ射线暴X射线光变曲线的分析
其他题名	Analysis of Gamma-ray burst X-ray light curves by Gaussian process
	宋非凡
学位类型	硕士
导师	毛基荣
	2024-07-01
学位授予单位	中国科学院大学
学位授予地点	北京
学位专业	天体物理
关键词	γ射线暴高斯过程时间序列分析
摘要	光变曲线的准周期震荡行为已经在许多高能天体物理现象中发现，例如在γ射线暴、快速射电暴、磁星巨耀发和X射线双星中。准周期震荡的分析为我们理解这些天体物理对象的动力学行为提供了新途径。在天体物理中常用周期图法分析光变曲线中可能存在的周期信号，然而传统的周期图法主要适用于同方差平稳的高斯数据，对大部分存在明显变化趋势的光变曲线，周期图的方法可能不能做出准确的分析。目前在γ射线暴的前兆辐射、瞬时辐射和X射线余辉中都有可能的准周期震荡信号的发现，这些准周期震荡信号有助于我们理解γ射线暴前身星系统和中心引擎相关的物理性质。因此，本研究专注于短γ射线暴的X射线光变曲线，分析其中可能存在的准周期震荡行为。本研究从Swift第三代目录中选择了15个短γ射线暴样本，并采用高斯过程方法对短γ射线暴样本的X射线光变曲线进行周期分析。相比于传统的周期图方法，高斯过程方法不要求时间序列数据具有同方差性和平稳性，因此γ射线暴X射线光变曲线这样随时间快速衰减的数据，更适合用高斯过程方法进行分析。为了与高斯过程方法的结果对比，Lomb-Scargle周期图方法也应用于周期信号的分析中。此外，本研究还对GRB 050724和GRB 060614的X射线光变曲线进行了更深入的时间分析，即采用了不同的高斯过程模型和加权小波Z变换对方法对这两个样本的X射线光变曲线进行了分析。研究发现，短γ射线暴样本的X射线光变曲线中存在的快速衰减的趋势并不影响高斯过程方法分析的结果，而Lomb-Scargle周期图的结果则显著受到趋势的影响。具体而言，扣除光变曲线中快速衰减的趋势后，高斯过程方法分析得到的功率谱密度，与其直接分析原始光变曲线得到的功率谱密度相比，几乎没有差异。而对于Lomb-Scargle周期图方法，扣除光变曲线中的趋势后，其分析得到的功率谱密度相较于分析原始光变曲线得到的功率谱密度，会变得更加平坦。这一对比说明，Lomb-Scargle周期图方法对趋势的敏感性较高。本研究还对比了不同高斯过程模型、Lomb-Scargle周期图和加权小波Z变换方法分析光变曲线的结果。这一对比不仅为理解各种方法的优劣提供了依据，也为未来研究活动星系核、γ射线暴或其他暂现源光变曲线中的周期信号提供了有价值的参考。对于周期分析的结果，在15个短γ射线暴样本中，仅在GRB 050724的X射线光变曲线中发现了一个可疑的周期信号。这一信号可能的物理起源包括磁星的地壳震荡、γ射线暴喷出物部分落回中子星表面并进行亚稳定核燃烧、或者具有进动的新生磁星的偶极辐射。然而，本研究并未断言GRB 050724的X射线光变曲线中的周期信号是真实存在的。在假设GRB 050724中的周期信号并非由于分析方法或者观测仪器原因引入的假周期信号前提下，本文对GRB 050724的X射线光变曲线中这一周期信号可能的物理起源进行了讨论。
其他摘要	Quasi-periodic oscillations in light curves have been found in many high-energy astrophysical phenomena, such as in Gamma-ray bursts, fast radio bursts, magnetar giant flares and X-ray binaries. Analyzing the quasi-periodic oscillations provides a new way to understand the dynamical behavior of these astrophysical objects, and accurate and reasonable period analysis methods will help to properly search for possible period signals in these sources. The periodogram method is commonly used in astrophysics to analyze the periodic behavior of light curves. However, traditional periodogram method is mainly applicable to homoscedastic stationary Gaussian data, so the periodogram method may not be able to make a accurate analysis for most of the light curves that have significant trends.Possible quasi-periodic oscillatory signals are currently found in the precursor, prompt emission, and X-ray afterglow of Gamma-ray bursts, by which the signals of quasi-periodic oscillations contribute to our understanding of the physical properties associated with the progenitor systems and central engines of the Gamma-ray bursts. Therefore, this study focuses on the X-ray light curves of short Gamma-ray bursts to analyze the possible quasi-periodic oscillatory behavior in them. In this study, 15 short Gamma-ray burst samples are selected form the Swift third catalogue, and the Gaussian process method is used to analyze the possible periodic signals in the short Gamma-ray burst X-ray light curves. Compared with the traditional periodogram method, Gaussian process method does not require the time series data to have homoscedasticity and stationarity, so data such as the Gamma-ray burst X-ray light curves that decay rapidly with time are more suitable to be analyzed by Gaussian process mothed. In contrast to the results of the Gaussian process method, the Lomb-Scargle periodogram method is also applied to the analysis of periodic signals. In addition, a more in-depth temporal analysis of the X-ray light curves of GRB 050724 and GRB 060614 has been carried out in this study, i.e., different Gaussian process models and weighted wavelet Z-transform method have been used to analyze the X-ray light curves of these two samples. It is found that the trend of rapid decay present in the X-ray light curves of the short γ-ray burst samples does not affect the results analyzed by the Gaussian process method, whereas the results of the Lomb-Scargle periodograms are significantly affected by the trend. Specifically, the power spectral density obtained from the Gaussian process method analysis, after subtracting the rapid decaying trend in the light curve, is almost no different from the power spectral density obtained from its direct analysis of the original light curve. For the Lomb-Scargle periodogram method, the power spectral density obtained from the analysis of the Lomb-Scargle method, after subtracting the trend in the light-variation curve, becomes flatter compared with that obtained by analyzing the original light curve. This comparison shows that the Lomb-Scargle periodogram method is more sensitive to trends. This study also compares the results of different Gaussian process models, Lomb-Scargle periodogram and weighted wavelet Z-transform methods for analyzing the light curve. This comparison not only provides a basis for understanding the advantages and disadvantages of the various methods, but also provides a valuable reference for future studies of the periodic signals in the light curves of active galactic nuclei, Gamma-ray bursts, or other transient sources.For the results of the period analysis, out of a sample of 15 short Gamma-ray bursts, only one possible period signal is found in the X-ray light curve of GRB 050724. The physical origin of this signal could crustal oscillations in magnetars, Gamma-ray burst ejecta partially falling back to the neutron star surface and marginally stable nuclear burning, or dipole radiation from newborn magnetar with precession rotation. However, this study does not state that the periodic signal in the X-ray light curve of GRB 050724 is real. This paper discusses the physical origin of the periodic signal in the X-ray light curve of GRB 050724 under the assumption that the periodic signal in GRB 050724 is not a fake periodic signal due to the analytical method or observational instrument.
学科领域	天文学
学科门类	理学 ; 理学::天文学
页数	0
语种	中文
文献类型	学位论文
条目标识符	http://ir.ynao.ac.cn/handle/114a53/28021
专题	星系类星体研究组
作者单位	中国科学院云南天文台
第一作者单位	中国科学院云南天文台
推荐引用方式 GB/T 7714	宋非凡. 高斯过程对γ射线暴X射线光变曲线的分析[D]. 北京. 中国科学院大学,2024.

条目包含的文件
文件名称/大小	文献类型	版本类型	开放类型	使用许可
高斯过程对γ射线暴X射线光变曲线的分析.（60KB）	学位论文		开放获取	CC BY-NC-SA	浏览请求全文