Parallel computing for incompressible flow using a Nodal-Based method

In this paper, the development of a seamless system for parallel flow analy sis using the "free-mesh method", which is a kind of meshless method, is de scribed. The system consists of two main parts: the computation of the glob al mass, advection, diffusion, gradient, and divergence matrices, and the t ime integration by the decoupled method with respect to velocity and pressu re. This system is quite compatible with the parallel environment because t he matrices are independently computed node-by-node without any node-elemen t connectivity information, and furthermore because the fractional step met hod, with the interpolation functions for velocity and pressure being of eq ual order, is used with the conjugate gradient solver for the time integrat ion. A parallel efficiency of approximately 83% was achieved for a large-sc ale problem with 480,000 degrees of freedom using 16 processors.