Construction Of Stellar Synthetic Spectrum Library In Parallel Computing Environment

The vast universe has her infinite charm. The ultimate goal of astronomy is to provide a simple theory to describe the entire universe. In order to advance astronomy, LAMOST was brought out by astronomers in China. Spherical spectrum contains many important physical information, through which human can determine chemical compositions of objects qualitatively or quantitatively. Stellar atmosphere physical parameters, which contain effective temperature, surface gravity, chemical abundance, are important factors in the difference of stellar spectra. Stellar atmosphere physical parameters automatic measurement is an important research in the LAMOST automatic data processing.In recent decades, people have figured out multiple methods to acquire stellar atmosphere physical parameters from mid-resolution spectra, which includes Template Matching, Artificial Neural Networks (ANN), Elements Line Index Method, Kernel Regression Method, etc, while Template Matching is the most commonly used. This method, one of the Nearest Neighbor Methods, is based on pre-existing spectrum libraries, in which the stellar atmosphere physical parameters of every spectrum have been already acquired by traditional methods. And the stellar atmosphere physical parameters of the test data are acquired by being matched to pre-existing spectrum libraries and their similarity are described by some similarity measures functions.With the development of astronomical data processing technology, various methods to acquire stellar atmosphere physical parameters are gradually maturing. Except for algorithm researching, matching spectra libraries can also be improved to improve the accuracy of measurement of stellar atmosphere physical parameters. Because of its calculation involves large and complex computing, parameters of large step are used to access to a wide range of parameter distribution. To meet the requirement of LAMOST Stellar Parameter Pipeline, we rebuild the Kurucz spectra library in a smaller step in order to satisfy the need for stellar atmosphere physical parameters computing. It can also be used in other related topics. The main research work is as follows:1.Study methods for stellar atmosphere physical parameters measurement and stellar atmosphere model theory, and study the Kurucz stellar atmosphere spectra library by its model computing and spectrum synthesizing which include ODF computing, Rosseland file computing, ATLAS9model computing and SPECTRUM programs.2. Study MPI parallel computing, and high resolution spectrum (R=50000) and low resolution spectrum (R=2000) are computed using large-scale computer cluster in National Astronomical Observatories of Chinese Academy of Science corresponding to ATLAS9stellar atmosphere model data which was released by Fiorella Castelli in2003.3. Base on the ATLAS9stellar atmosphere model data released by Fiorella Castelli, the Kurucz stellar atmosphere spectrum library is computed in a smaller step using large-scale computer cluster in National Astronomical Observatories of Chinese Academy of Science, totally40,000models, and all of the models are normal stellar models. Corresponding spectra are computed in high resolution (R=50000) and low resolution (R=2000). Correlation Analysis is done to testify the correctness of the produced synthetic spectra data library.