Numerical strategies for capturing rotating stall in fan

In spite of advances in computational capabilities, most numerical studies for rotating stall use relatively simple models, and there are no established methods to simulate rotating stall accurately. In this study, four numerical approaches for obtaining rotating stall are considered. These cover a range of interface boundary conditions, stall triggering methods, and different algorithms for changing operating conditions. The findings are as follows. First, placing a mixing plane behind the fan causes some non-physical phenomena during stall inception and it should be avoided. The whole-annulus computational domain without initial disturbance simulates the post-stall phenomena well but it has some shortcomings for modelling the stall inception. Asymmetric disturbance, caused by one mis-staggered blade, is successful in both triggering the rotating stall and representing the stall phenomena itself. The slow adjustment of the downstream boundary conditions through on-the-fly nozzle area change can remedy numerical overshoots in performance curves, but it delays stall inception process and increases total computational time significantly. This article contains detailed results on stall inception and development for each methodology.