• diffuser
• turbulent

The objective of the present work is twofold: first, the accuracy at different RANS models in predicting the separated flow in asymmetric diffuser are investigated. Then the capabilities of a passive method for the control of boundary-layer separation in the considered diffuser configuration are appraised.
The passive control method consists in modifying the geometry of the diffuser walls using appropriately-shaped cavities.\\
The configuration of the diffuser is characterized by an expansion angle of 10 degrees and ER=4.7. The inlet flow is fully developed and the Reynolds number is 17500 (based on the inlet velocity and the height of the inlet channel).
A preliminary mesh sensitivity analysis, for three different turbulence models, is performed for the flow inside the diffuser without cavities. Based on this analysis the SST $\kappa_{}$-$\omega_{}$ model is chosen for the optimization procedure of the control shaped cavities.\\
Two different configurations have been studied: the first with a single bump-shaped cavity and the second one with two subsequent bump-shaped cavities.
An optimization has been carried out for each of these configurations.
The performance of the diffuser has been evaluated in terms of both pressure recovery coefficient and the dissipation function.
The passive control improves the performance of the diffuser in both the configurations, leading to an increase of the mean pressure recovery coefficient, by reducing the extent of the separation region.\\
Finally the effect of the Reynolds numbers has been examined. The results show a decreasing of the size of the separation zone with decreasing Reynolds number.
The optimum configurations obtained for the passive control, at the reference Reynolds, have been tested also for the different considered Reynolds numbers. Improvements of the efficiency are obtained in all cases.