[cgasdxing]

I'm doing research on POD-Galerkin projection reduced order model. It is reported in some paper that the ROM discretization schemes are same as in full order model.

But when I write the Matlab program, the advection term can only be fixed format (eg. center difference, or forward difference). It is difficult to make reduced coefficient of advection term vary with the full order local velocity filed. Unless I recalculate this coefficient from full order velocity filed at each ROM step.

Is there any way to implement the convection characteristics in ROM ?

Thank you for your help.