Over the past ten years, a new method for the computer
simulation of fluids has been developed: the lattice gas model [5].
Instead of considering a large number of individual molecules, the
molecular dynamics approach, a much smaller number of fluid `particles'
are considered.
A fluid `particle' is a
large group of molecules which although much larger than a molecule is
still considerably smaller then the smallest length scale of the
simulation. This reduces the amount of data
which needs to be stored since large simulations can be performed using
less than one million `particles'.
This is justified on the grounds that the
macroscopic properties do not depend directly on the microscopic
behaviour of the fluid. This can be seen in low Mach flows where,
provided the Reynolds number is the same, experiments carried out in a
water tank and a wind tunnel produce the same results. These two fluids
have different microscopic structures, but they both exhibit the same
macroscopic features. In a lattice gas model the
`particles' are restricted to move on the links of a
regular underlying grid and the motion
evolves in discrete time-steps. The conservation laws are
incorporated into update rules which are applied at each discrete time.
A lattice gas model in which the state of the fluid needs to be known only at the lattice sites and only at discrete times can run much faster on a computer than a molecular dynamics simulation. The lattice gas model has another big advantage over molecular simulation since all the collisions occur at the same time. This is a particular advantage if the simulation is being run on a parallel computer. These two time saving advantages of the lattice gas model allow simulations of a significantly large scale to be performed.