Experimental validation of a corner stall control methodology using parametrised guide fins

Abstract

Corner separation is known to limit the operability of aeronautical compressors. Dedicated control devices such as guide fins are envisioned to reduce its negative effects. This paper presents a methodology based on RANS (Reynolds-Averaged Navier-Stokes) computations enabling to select guide fins efficient for that purpose. This methodology is applied to a reference case of linear compressor cascade operating at low Mach number (∼0.11). A set of 17 parameters is used to define two design spaces of interest, from which guide fins are generated. From then, an automated process generates and merges an unstructured mesh built around each guide fin with a fixed, structured mesh of reference representing a single channel of the cascade. Finally, RANS results on the resulting hybrid mesh are obtained using the Computational Fluid Dynamics solver elsA. This set up has proven successful in evaluating automatically hundreds of guide fins of various shapes. Several geometries illustrate the diversity of the design space. A selection of guide fins is then evaluated experimentally. Evolutions of the losses downstream of the cascade are compared to their respective RANS predictions, and to the reference case without guide fin. These experimental results validate the implemented methodology and show encouraging results in terms of loss redistribution brought by the control device.