数值传热学第四章对流扩散.ppt

第四章 对流扩散方程离散 * Chapter 4 convection and diffusion 4.1 steady 1D situation 4.2 Central,upwind and other schemes（5 basic schemes) 4.3 2D 3D situations 4.4 the outflow boundary 4.5 False diffusion The task is to solve 4.1 Steady 1D situation(CDS) For 1D convection and diffusion problem,the govern equation is Continuity : ρu=constant. Integration over the controlume gives: According to the idea of numerical solution, the value of φ is supposed known at node W P and E,etc,but φ at interface(cell face)is not known.This may result many schemes of estimating φe or φw. For example central difference scheme , upwind scheme and others. A simple-minded formulation Diffusion term CDS: Convection term CDS Result : Where: Discussion: The derivation is based on a linear variation of φ between grid points,and is known as the central-difference scheme(CDS). Let De=Dw=1,and Fe=Fw=4. Then If φE=200 and φW=100,then φP=50; If If φE=100 and φW=200,then φP=250 The unrealistic results arise because we have allowed some of the coefficients to become negative.The CDS is thus satisfactory for only low value of F/D, i.e.,for low Peclet numbers or Reynolds numbers To avoid the unrealistic results,the upwind scheme (UDS)is presented. Principle: the value of φ at an interface is taken as the value of φ at the grid point on the upwind side of that face. The upwind scheme is not as simple-minded as central difference scheme. The unrealistic problem of CDS had puzzled CFD/NHT for many years . The exact solution According to the exact solution,the CDS is correct only at Pe=0 or Pe is very small, i.e. only at low Pe values the φ value at interface may be averaged with two values of each side. With great Pe number,the φ value is more approach to the upwind side node value, So the upwind scheme will not give unrealistic results. And with result of exact solution,the exponential scheme(EDS Spalding 1972; Raithby and Torrance,1974) is presented. F/D is the Peclet number based on the grid interval. E