Using GGI Interfaces

JuPa · September 22, 2014, 06:38

I've searched the forum to read other experiences of GGI interfaces, but I couldn't find anything suitable.

I have a large complicated geometry. I can create large hexahedral areas, and backfill the rest of the volume with tetrahedral elements.

The catch is my mesh will be non-conformal: where the grid on either side of two connected surfaces will not match.

I've read GGI theory in the Ansys guide, and it seems promising.

Does any one have any simulation advice using GGI?

My particular simulation is a multiphase one, involving phase change at the free surface between two continuous fluids.

The obvious answer is to do a sensitivity study to see if anything is effected. But on top of that, does anyone have any advice?

Thank you.

singer1812 · September 22, 2014, 11:56

I use GGIs quite a bit.

Most recently has been were we need quick trades done on a projectile body. We are trading complicated body, canard, and tail features.

I have a baseline large exterior air mesh with a cutout around the projectile. At the projectile level, I have cutouts around the canards and tails.

This allows me to drop in many different types of bodies, canards, and tails, without having to redo the portion in the rest of the regions. Typically the body is hex meshed as they are symmetrical and the other regions are tetrahedral meshes as they can be very complicated.

A few things to keep in mind:
I keep all mesh sizes and growth rates to and from the interface at the same levels.

One unfortunate feature is that there is an interface on the wall region, and cutting across a boundary layer (not the best, but a sin I live with to facilitate quick trades). For this, I match the wall node locations from the hex mesh to the tet-mesh. This creates a smooth wall transition, else you will get an "X" type interface due to wall curvature and node mismatch. I also match the prism layer count and growth rate in tet mesh to the hex mesh. This creates an approximate 1-to-1 interface (not quite 1 to 1 as the prisms might not be as orthoganal as the hexes, but they are close).

Doing this does a really good job for our application as using a unified mesh. We have compared a number of runs and the differences in our items of interest are usually less than 1%.

This is what I have done recently, and it may or may not have bearing on what you are attempting.

September 22, 2014, 06:38	Using GGI Interfaces	#1
JuPa Senior Member Mr CFD Join Date: Jun 2012 Location: Britain Posts: 361 Rep Power: 14	I've searched the forum to read other experiences of GGI interfaces, but I couldn't find anything suitable. I have a large complicated geometry. I can create large hexahedral areas, and backfill the rest of the volume with tetrahedral elements. The catch is my mesh will be non-conformal: where the grid on either side of two connected surfaces will not match. I've read GGI theory in the Ansys guide, and it seems promising. Does any one have any simulation advice using GGI? My particular simulation is a multiphase one, involving phase change at the free surface between two continuous fluids. The obvious answer is to do a sensitivity study to see if anything is effected. But on top of that, does anyone have any advice? Thank you.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[Other] ggi for non-overlaping interfaces (OpenFOAM-2.2)	xinyu	OpenFOAM Meshing & Mesh Conversion	1	January 13, 2017 21:22
Create a GeometricField of a given type on given patch?	philippose	OpenFOAM Programming & Development	4	August 12, 2013 12:41
Frozen Rotor 1:1 Mesh Connection	pharley	CFX	5	January 31, 2013 16:15
Help with ggi interfaces	hardykalyan	OpenFOAM	3	November 6, 2010 14:47
problem with GGI interfaces	strider	CFX	6	May 29, 2006 14:20

September 22, 2014, 11:56		#2
singer1812 Senior Member Edmund Singer P.E. Join Date: Aug 2010 Location: Minneapolis, MN Posts: 511 Rep Power: 20	I use GGIs quite a bit. Most recently has been were we need quick trades done on a projectile body. We are trading complicated body, canard, and tail features. I have a baseline large exterior air mesh with a cutout around the projectile. At the projectile level, I have cutouts around the canards and tails. This allows me to drop in many different types of bodies, canards, and tails, without having to redo the portion in the rest of the regions. Typically the body is hex meshed as they are symmetrical and the other regions are tetrahedral meshes as they can be very complicated. A few things to keep in mind: I keep all mesh sizes and growth rates to and from the interface at the same levels. One unfortunate feature is that there is an interface on the wall region, and cutting across a boundary layer (not the best, but a sin I live with to facilitate quick trades). For this, I match the wall node locations from the hex mesh to the tet-mesh. This creates a smooth wall transition, else you will get an "X" type interface due to wall curvature and node mismatch. I also match the prism layer count and growth rate in tet mesh to the hex mesh. This creates an approximate 1-to-1 interface (not quite 1 to 1 as the prisms might not be as orthoganal as the hexes, but they are close). Doing this does a really good job for our application as using a unified mesh. We have compared a number of runs and the differences in our items of interest are usually less than 1%. This is what I have done recently, and it may or may not have bearing on what you are attempting.