Title:	Modelling the short-period population of binary stars
DNr:	SNIC 2019/3-666
Project Type:	SNIC Medium Compute
Principal Investigator:	Sofia Feltzing <sofia@astro.lu.se>
Affiliation:	Lunds universitet
Duration:	2019-12-27 – 2020-12-01
Classification:	10305
Keywords:

Abstract

Binary (multiple) stars offer exceptional insight to star formation and evolution, establishing distances across the Galaxy and beyond, providing critical tests in asteroseismology, and furnishing benchmark systems with accurate fundamental stellar properties. Although no one disputes the astrophysical importance of multiple systems, they are often overlooked simply because their nature can be quite effectively concealed in observational data. However, the advances in machine learning, and the unprecedented volume of stellar spectra from modern surveys, enable us to devise new approaches for their discovery and characterisation. This project's goal is: a) to derive the basic stellar parameters of both stars in a binary system for a sample of around 6000 systems, b) use the derived parameters to explore the initial properties of analysed binary systems using a fast binary evolution algorithm BSE. Both parts of the project are computationally intensive. In the first part a Bayesian approach is used to obtain the necessary accuracy and uncertainties of the results involving a Monte Carlo scheme. The second part of the project necessitates a comparison of observed and synthetic properties of the populations of binary stars, for which again a Bayesian procedure will be employed to explore the parameter space of initial distributions of properties of Galactic binary populations.