Questions regarding the best practise for generating a turbine grid, including: What is the recommended cell size increase between rotating and stationary domains? Is it dependent on the interface used at this location (frozen, stage or transient)? Also do the radial grid nodes need to match 1:1 when using transient solution?
|
There are general rules of thumb to follow regardless of the type of Multiple Frame of Reference (MFR) Interface you select. 1. The nodes on either side of the interface do not need to match 1:1 as we are using the General Grid Connection (GGI) technology. However, it is wise to have similar distributions hub to shroud and circumferentially on either side of the interface as this will ensure increased accuracy of the interpolation of the fluxes from one side to the other. For example, it is not sensible to map fluxes from 50 face elements onto 1 face element as you would lose any gradients that might pass over the 50 face elements! 2. You should have at least 100,000 nodes per blade passage irrespective of whether or not it is stationary or rotating. 3. If there are large flow gradients and fluctuations between components eg wakes and much interaction of components, then it it wise to ensure you refine the grid sufficiently in the streamwise direction. This is true of the radial and circumferential directions too if there are large flow gradients. Another rule of thumb is to have at least one node per circumferential degree. 4. If you are experiencing different results with different grid densities, this may be expected as with the coarser grid, the flow gradients are smeared out or diffused. You should undertake a grid density study to determine the appropriate mesh size. |
||
|