Particle Splitting: A New Method for SPH Star Formation Simulations

Abstract

We have invented a new algorithm to use with self-gravitating SPH Star Formation codes. The new method is designed to enable SPH simulations to self-regulate their numerical resolution, i.e. the number of SPH particles; the latter is calculated using the Jeans condition (Bate & Burkert 1997) and the local hydrodynamic conditions of the gas. We apply our SPH with Particle Splitting code to cloud-cloud collision simulations. Chapter 2 lists the properties of our standard SPH code. Chapter 3 discusses the efficiency of the standard code as this is applied to simulations of rotating, uniform clouds with m=2 density perturbations. Chapter 4 [astro-ph/0203057] describes the new method and the tests that it has successfully been applied to. It also contains the results of the application of Particle Splitting to the case of rotating clouds as those of Chapter 3, where, with great computational efficiency, we have reproduced the results of FD codes and SPH simulations with large numbers of particles. Chapter 5 gives a detailed account of the cloud-cloud collisions studied, starting from a variety of initial conditions produced by altering the cloud mass, cloud velocity and the collision impact parameter. In the majority of the cases studied, the collisions produced filaments (similar to those observed in ammonia in nearby Star Forming Regions) or networks of filaments; groups of protostellar cores have been produced by fragmentation of the filaments. The accretion rates at these cores are comparable to those of Class 0 objects. Due to time-step constraints the simulations stop early in their evolution. The star formation efficiency of this mechanism is extrapolated in time and is found to be 10-20%.