Modelling nearby FGK Population I stars: A new form of estimating stellar param...

• Autor(es): Fernandes, João Manuel ; Vaz, A. Ismael F. ; Vicente, Luís Nunes
• Data: 2008
• Identificador Persistente: http://hdl.handle.net/10316/11217
• Origem: Estudo Geral - Universidade de Coimbra
• Assunto(s): Optimization; Global derivative-free optimization; Direct search; Particle swarm; Stars; Fundamental parameters; Hertzsprung-Russell (HR); Planetary systems
• Tipo de Documento: Preprint
• Idioma: Inglês

Descrição

Modelling a single star by means of theoretical stellar evolutionary tracks is a nontrivial problem due to the large number of unknowns compared to the number of observations. Currently, stellar age and mass are estimated using interpo- lations in grids of stellar models and/or isochrones assuming ad-hoc approximations for the mixing length parameter and the metal to helium enrichment, normally scaled on the solar values. This paper presents a new method to model the FGK main sequence of stars of Population I, capable of simultaneously estimating the fol- lowing stellar parameters: mass, age, helium and metals abundance, mixing length parameter, and overshooting. Our approach is based on the application of a global optimization method (PSwarm) to solve the inverse, simulation-based optimization models of nding the values for the stellar parameters that better match the given observations. The evaluation of the tting function requires the simulation of a stellar evolutionary code. In these models, the helium and the mixing length are not scaled on the Sun but, together with the overshooting, considered free optimization parameters. The optimization algorithm used by PSwarm is a rigorous direct search method enriched by a popu- lation based heuristic (particle swarm) to improve the ability to search for a global optimizer. We develop a public-domain computational tool to interface the global optimiza- tion solver and the stellar evolutionary code. We test our method using the Sun and ve FGK ctitious stars and then apply our methodology to about 135 detailed spectroscopic analysed stars, including 74 planet host stars. We present and discuss the stellar parameters estimated for each star in the context of previous works. The impact of the results on stellar evolutionary studies is brie y discussed.