A NEW APPROACH TO THE SIMULATION OF COMPLEX FLOWS
Jean Pierre Boon · 1990
Turbulence is a complicated phenomenon which has remained one of the last mysteries of classical physics. This quotation from a 1969 paper by Elliott Montroll [1] has kept a substantial part of its truth: twenty years later -and one century after Osborne Reynolds established his famous dynamical similarity law (1883)the mystery of turbulence is still far from being fully unveiled. Yet, over the past two decades, theoreticians and experimentalists, physicists and mathematicians, have accomplished considerable progress towards the understanding of complex flow phenomena. This achievement relies to a large extent on the formidable development of modern computational tools [2]. In particular very interesting perspectives have appeared since 1985 for highly parallel computation in fluid dynamics, based on a thoroughly new method: lattice gas hydrodynamics. Two Jaboratories in France (l'Observatoire de l'Universite de Nice, and l'Ecole Norm8le superieure de Paris) and one in the United States (the Center for Nonlinear Studies at Los Alamos) have pioneered this research, which has attracted the interest and the collaborative efforts of several groups in the U.S. and in Europe (Shell Research Laboratorium Amsterdam, Ie Groupe de Physique N on-lineaire de Bruxelles, Politecnica Madrid, Ie Centre de Recherches en Combustion de Marseille, IBM Research Rome, etc ... ). The development of lattice gas hydrodynamics has been technically connected to cellular automata but, as it has been pointed out [3], lattice gases should not be referred to as cellular automata. In many respects, lattice gases have become a field per se with strong interfaces with kinetic theory, algorithmics, computer and supercomputer simulations, and dedicated machine development. Many of these aspects are presented and discussed in two recent volumes devoted to the subject [4,5]. In the present paper we shall restrict ourselves to a review of lattice gas automata and illustrative examples of their application to fluid dynamics.