Rotor grid with interlobe refinement

One of the major factors affecting the performance prediction of twin screw compressors by use of computational fluid dynamics is the accuracy with which the leakage gaps are captured by the discretization methods. The particular attention needs to be dedicated to the interlobe rotor region.


One of the methods of improving the profile accuracy is by increasing the number of grid points on the profile. However, this method faces limitations when it comes to the complex deforming computational domain of the twin screw compressor because the grid quality deteriorates and computational time increases tremendously.


In order to address this problem, an analytical grid distribution procedure that can independently refine the region of high importance i.e. the interlobe space has been developed in SCORGTM. This function uses equidistant distribution of points on the outer computational domain which consists of the casing and rack to calculate distribution on rotors. By this means it is possible to refine mesh in the interlobe gap to allow for a more accurate prediction of leakage flows and reduce number of computational celles in other regions. The total grid size can be controlled by limiting the number of cells in the region not containing interlobe gap and blow-hole areas.


The new grid has regular cell structure with all quadrilateral cells and improved orthogonality.


The non-conformal interface between the rotors was resolved precisely with the interlobe grid refinement.


The results of pressure showed that there was a very small difference between the original grid and the new grid and the internal pressure rise agreed well with the experimental measurement.


Indicated power increased by a small amount and was close to experimental value at all speeds.


There was a big influence of interlobe grid refinement on the prediction of mass flow rate and hence leakage flows. With refined interlobe space, the error between experimental and calculated flow rate reduced significantly.