MULTIOBJECTIVE DESIGN OPTIMIZATION OF A CENTRIFUGAL COMPRESSOR VOLUTE

Abstract

Centrifugal compressors are used in many technical fields and preferred for small volumetric flows like in turbochargers and small gas turbines. Compressor efficiency and operating range are highly influenced by the volute design. Flow parameters like static pressure recovery and total pressure loss are directly linked to the design parameters. In the last decade, several design studies on centrifugal compressors were performed at the Institute of Propulsion Technology. These studies were mainly driven by the design of impeller and diffuser. This work introduces a multiobjective design approach for centrifugal compressor volutes. DLRs design suite includes the in-house optimization tool AutoOpti which is based on an evolutionary genetic algorithm and TRACE as the RANS flow solver. The commercial software CAESES is used for the volute parametrization. An accurate parametrization is required because of the complex flow state in the volute. In a first sensitivity study it was found that the shape and size of the volute cross-section influenced its efficiency, as the tongue geometry. These parameters are investigated more deeply in the optimization. Finally, this multiobjective design approach for centrifugal compressor volute leads to an improvement in terms of total pressure loss and static pressure recovery.