Design optimisation of an impeller with CFD and Meta-Model of optimal Prognosis (MoP)

The CFD method is used to predict the flow and the compressor map. For the optimisation CAE-based parametric optimisation with respect to 15 geometry parameters, based on the primary design is used. The optimisation procedure is divided in two steps: The first one is the sensitivity study combined with the generation of the Metamodel of optimal Prognosis (MoP), where the most relevant input parameters and a quality assurance of the model could be identified. In addition the MoP can be used for optimisation to predict the performance of the compressor in the whole parameter space and to search for optimal designs. The second step, for further design improvement, is an optimisation procedure using additional solver calls, where the ARSM (Adaptive Response Surface Method) algorithm is selected for optimisation. The subdomain of important parameter is defined with respect to the result of the MoP. The whole optimisation strategy is designed to work in large parameter spaces (>10..100) with a minimum number of CFD simulations, “no run too much”; to find an improved design. Nomenclature b blade thickness L blade length m mass flow u circumferential velocity N node Greek Letters β blade angle Θ temperature η efficiency Π pressure ratio Subscripts 1.0 compressor inlet ini initial design 1.1 rotor outlet is isentropic 1.2 diffusor outlet max maximum 2.0 compressor outlet n normalized B blade opt optimised C compressor T total CW compressor wheel