The basic principle of the multiblock method is to split a computational domain into several smaller ones. It is frequently used to deal with complex geometry and in parallel computing. The basic multiblock method is described including the way of splitting and the interface conditions in matching solutions between adjacent subdomains. Then we describe how to use the multiblock method to do parallel computing and how to analyze stability, convergence, conservation and uniqueness. Finally, an overview on stability, accuracy, convergence, conservation and uniqueness of a multiblock method is given.

In the generalized characteristic coordinate system(GCCS) proposed by the authors elsewhere, the frame moves at a speed which is a linear combination of the convective speed and the sound speed.This coordinate system is applied to the computation of the one-dimensional Euler equation and the advantages of GCCS is demonstrated in capturing expansive waves and shock waves.

Through a theoretical study,it is found that the incompressible slip flow model for flows in a microchannel with vanishing Reynolds numbers is unstable when the Knudsen number is larger than 1/9.The effect of the finite Reynolds numbers on the stability is considered.Results show that the critical Knudsen number for instability is an increasing function of the Reynolds numbers.

An automatic Nonet-Cartesian grid method is presented together with Euler solutions of flows around complicated geometries. Grids are generated automatically by the recursive subdivision of a single cell into nine subcells for isotropic Nonet-Cartesian grids and into three subcells independently in each direction for anisotropic Nonet-Cartesian grids, encompassing the entire flow domain. The grid generation method is applied here to steady inviscid shocked flow computation. Results using this approach demonstrate that this method provides a simple and accurate procedure for solving flow problems with shock waves.

Recently Wu Z N and Zou H proposed an overlapping grid method.The suitability of applying the method to parallel computation of steady and unsteady compressible inviscid flows with three-point block-tridiagonal implicit schemes was addressed in the paper [1],and an easily usable interface treatment was constructed and analyzed for both steady and unsteady problems.This method is here applied to the computation of N-S equations.In both steady and unsteady cases a good absolute parallel efficiency has been still demonstrated for bidimensional subsonic and transonic flow computations.