Have been using TASCflow for sometime and this is the first time I have had this error message - any suggestions?

The problem is MFR, 700k nodes, one-to-one periodicity, with a blower wheel in a housing. The inlet interface is stage and the blower outlet to housing interface is frozen rotor. It is a multiblock grid with combined block structure using tascu2s.

The error message is as follows:

Reading the GCI file for region information:

Notice: General Grid Interface variables initialized from nodal variables.

The first suggestion was a parameter of the sliding_interface_eta_model = 1 or 2, neither setting helped as they felt it was a projection/runtime error.

Thanks for any help, Erich

ERROR: Sliding Interface Edge List Dis-connected Internal Error: Consult an expert, SRFEDG! JEDGE= 524 NEDGES= 5249

STOP due to error : General Failure

(1). One of the principle to minimize the error is to keep things very simple. (2). But, you are taking a very compex approach, Moving Reference Frame, very large number of cells, wheel, housing, multi-block, tascu2s,... (3). So, take the first step to reduce the number of cells. A 700k mesh will take at least 700Meg of RAM, I think. Reduce it to say 50k cells. And look at the mesh first. I am just guessing that you probably have a poor mesh, or meshes not properly connected. At 50k, you can do the debugging much more efficient. (4). I think, tascu2s is a software to convert a un-structured mesh into structured mesh. And you are giving the code a very hard time also. Tascflow is a multi-block structured code. I can only say that your mesh is too complicated for the code to handle automatically.

Thanks, John, will experiment with fewer nodes;however, I think I found my problem in that a solid region was next to the interface (the ggi interface allows this while the stage or frozen does not). I had put a shaft in the center of the blower to eliminate any coalescent boundaries and it terminated at the stage interface. Going to remesh. You are right on with the memory - have over a gig so not a problem, but see your point of simplification. I use tascu2s only to reduce the number of grids into fewer blocks for ease in post processing only. My minimum skew was 29 degrees and maximum aspect around 40.

(1).Thank you very much for giving me the opportunity to exercise my brain. (2). By question-and-answer in the obvious areas using common sense, it is possible to zoom in the problem area. The doctor does not actually heal the patient, it is the patient's body which does the healing by itself. The doctor only help the patient to focus on the problem area. (3). I am running a 3-D case with the skew angle 5.9 degree. It is taking too much time to refine the mesh, so I just let it go. (ideally, the user shouldn't worry about the mesh at all.)