Compressor and Turbine Multidisciplinary Design for Highly Efficient Micro-gas Turbine

• Published in: Journal of thermal science 2018-06, Vol.27 (3), p.259-269
• Main Authors: Barsi, Dario, Perrone, Andrea, Qu, Yonglei, Ratto, Luca, Ricci, Gianluca, Sergeev, Vitaliy, Zunino, Pietro
• Format: Article
• Language: English
• Subjects: Aerodynamics; Artificial neural networks; Centrifugal compressors; Classical and Continuum Physics; Computational fluid dynamics; Design optimization; Engineering Fluid Dynamics; Engineering Thermodynamics; Exhaust gases; Finite element method; Gas temperature; Gas turbine engines; Heat and Mass Transfer; Inflow; Mathematical analysis; Mesh generation; Multidisciplinary design optimization; Neural networks; Physics; Physics and Astronomy; Reynolds averaged Navier-Stokes method; Solvers; Turbines; Turbomachinery
• ISSN: 1003-2169; 1993-033X
• DOI: 10.1007/s11630-018-1007-2

Multidisciplinary design optimization (MDO) is widely employed to enhance turbomachinery components efficiency. The aim of this work is to describe a complete tool for the aero-mechanical design of a radial inflow turbine and a centrifugal compressor. The high rotational speed of such machines and the high exhaust gas temperature (only for the turbine) expose blades to really high stresses and therefore the aerodynamics design has to be coupled with the mechanical one through an integrated procedure. The described approach employs a fully 3D Reynolds Averaged Navier-Stokes (RANS) solver for the aerodynamics and an open source Finite Element Analysis (FEA) solver for the mechanical integrity assessment. Due to the high computational cost of both these two solvers, a meta model, such as an artificial neural network (ANN), is used to speed up the optimization design process. The interaction between two codes, the mesh generation and the post processing of the results are achieved via in-house developed scripting modules. The obtained results are widely presented and discussed.