CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Main CFD Forum (https://www.cfd-online.com/Forums/main/)
-   -   Second order Upwind (https://www.cfd-online.com/Forums/main/12831-second-order-upwind.html)

Mike January 16, 2007 15:43

Second order Upwind
 
Hi,

Could anyone help me find an article about second order upwind. Is it Quick method or something different.

Best

Bren January 16, 2007 18:15

Re: Second order Upwind
 
It'sa type of finite differencing stencil. You should be able to learn about it in any finite difference textbook.

The idea is that finite difference stencils originate from Taylor series. By expanding a function, 'f' say, we can develop finite difference approximations to derivatives.

Upstream differencing means that the finite difference operators are calculated by using points that are up-stream.

An example is the first derivative of f with respect to x (where dx is the grid spacing). The derivative at the node 'i' can be written f_x(central)=(1/2)*(f(i+1)-f(i-1))/dx f_x(upstream)=(f(i+1)-f(i))/dx

Other operators are available; the order of the operator comes from the truncation error in the Taylor series.

Hope this is of some help, Bren

ag January 17, 2007 08:35

Re: Second order Upwind
 
QUICK is one type of higher-order upwind, and refers to the reconstruction used to generate face values of the field variables (although it can also be used on the flux variables directly). Such reconstructions are typically used in conjunction with Riemann type solvers, although you can also find higher order upwind reconstruction used in incompressible solvers. One can also formally develop upwind stencils following Bren's outline. Try googling for higher-order upwind CFD.

Valdemir G. Ferreira January 22, 2007 03:05

Re: Second order Upwind
 
Dear Mike,

Please, see Adaptative QUICKEST scheme by VG Ferreira et al. in Communications in Numerical Methods in Engineering (In Press).

VGF


All times are GMT -4. The time now is 04:01.