[nikn]

Hey,

right now I carry out a grid independence study and as I am not satisfied with my results, I want to ask, if you agree with my procedure. I orientated myself towards this source: http://journaltool.asme.org/Content/JFENumAccuracy.pdf

I had a pretty fine model of a two stage turbine with around 23 Mill. nodes and I use High-res as discretization scheme. Then I made three other simulations with one coarser grid (15 Mill Nodes) and two finer grids (35 Mill and 46 Mill Nodes). Of course I read that normally the grid refinement ratio should be as high as possible and preferably 2, but as you see, it is not possible to make it even finer. On the other hand coarser grids than 15 Mill nodes refused to work.

However I decided to carry out the study with the three finest grids. I decided to consider the dissipation of each rotor and stator as my key value. Now to my questions:

Considering the dissipation I calculated the convergence order of each rotor and stator based on the equations 3a-c in the above-mentioned source. For the first stator and the first rotor my results seem to be fine because I get a value for p which is either 2.1 or 3.2. But on the other hand I get a very high p for stator 2 and rotor 2. (p=29 and p=26). This is far to high, isnt it? Normally p should be even less than 2? Does it maybe mean that my values are very close together and therfore that I am grid independent with all three grids?

In the course of the study I also calculated the cgi according to roache. Could you explain me the gci? And how does it help me to interpret my values?

Thanks a lot for your help
nikn