Staggered grid
From ThermalFluidsPedia
m 

Line 1:  Line 1:  
+  {{Comp Method for Forced Convection Category}}  
[[Image:Fig4.23.pngthumb400 pxalt=Staggered grid: (a) control volume for all other variables, (b) control volume for u, and (c) control volume for v. Figure 1: Staggered grid: (a) control volume for all other variables, (b) control volume for u, and (c) control volume for v.]]  [[Image:Fig4.23.pngthumb400 pxalt=Staggered grid: (a) control volume for all other variables, (b) control volume for u, and (c) control volume for v. Figure 1: Staggered grid: (a) control volume for all other variables, (b) control volume for u, and (c) control volume for v.]]  
Current revision as of 16:15, 2 June 2010
Computational methodologies for forced convection 
To overcome the checkerboard pressure field and develop an effective algorithm for the pressure field, one can use a staggered grid system (Patankar and Spalding, 1972; Patankar, 1980) that stores the pressure and all other variables on the main grid but calculates the velocity at the face of the control volume (see Fig. 1). The control volume for the velocity component in the xdirection, u, is staggered from the control volume for all other variables to the right direction by half a grid. Similarly, the control volume for v is staggered up by a half grid. The discretized equations for u and v will be obtained by integrating the momentum equation in their control volumes shown in Fig. 1 (b) and (c), respectively. The discretization equations in the staggered grid system for all variables except velocity are the same as those presented in the preceding subsection. Since velocity is defined at the face of the control volume, the grid Peclet numbers can be directly calculated from the velocity component, i.e., no assumption on the velocity profile between the grid points is needed. For the momentum equation in the xdirection, we will need to integrate eq. (4.287) for the control volumes of u shown in Fig. 1 (b). The result can be expressed as

where the first term on the right hand side represents the summation of all the neighbor points of e. The effect of pressure has been separated from the source term, and the pressures at P and E were used to calculate the velocity ue. For a twodimensional problem, the area on which the pressure difference acts on is A_{e} = Δy. Similarly, the discretization equation for v can be obtained by integrating eq.(4.287) for the control volume of vn shown in Fig. 1(c), i.e.,

where the area on which pressure difference acts on is A_{n} = Δx for a twodimensional problem. For a threedimensional problem, eqs(4.292) and (4.293) are still valid except the neighbor points from the top and bottom should be included in the first term on the righthand side. The areas on which the pressure difference act on should be modified to A_{e} = ΔyΔz and A_{n} = ΔxΔz. In addition to eqs. (4.292) and (4.293), another equation for the velocity component in the zdirection, wt, is also needed, i.e.,

which is obtained by integrating the momentum equation in the zdirection for the control volume of wt that are staggered to the positive zdirection by a half grid. The area on which the pressure difference, p_{P} − p_{T}, acts on is A_{t} = ΔxΔy.