Robust optimization of supersonic ORC nozzle guide vanes

1st International Seminar on Non-Ideal Compressible-Fluid Dynamics for Propulsion & Power

物理学;力学

Bufi, Elio A.^1 ; Cinnella, Paola^2

Laboratoire DynFluid Arts et Metiers ParisTech, Paris (France) Politecnico di Bari, Bari, Italy^1

Laboratoire DynFluid Arts et Metiers ParisTech, Paris, France^2

关键词: CFD based optimizations;    Free-form deformation;    Isentropic efficiency;    Method of characteristics;    Multi-objective fitness function;    Non-dominated sorting algorithms;    Organic Rankine cycles;    Uncertainty quantifications;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/821/1/012014/pdf
DOI  :  10.1088/1742-6596/821/1/012014

学科分类：力学，机械学

来源: IOP

【 摘 要 】

An efficient Robust Optimization (RO) strategy is developed for the design of 2D supersonic Organic Rankine Cycle turbine expanders. The dense gas effects are not-negligible for this application and they are taken into account describing the thermodynamics by means of the Peng-Robinson-Stryjek-Vera equation of state. The design methodology combines an Uncertainty Quantification (UQ) loop based on a Bayesian kriging model of the system response to the uncertain parameters, used to approximate statistics (mean and variance) of the uncertain system output, a CFD solver, and a multi-objective non-dominated sorting algorithm (NSGA), also based on a Kriging surrogate of the multi-objective fitness function, along with an adaptive infill strategy for surrogate enrichment at each generation of the NSGA. The objective functions are the average and variance of the isentropic efficiency. The blade shape is parametrized by means of a Free Form Deformation (FFD) approach. The robust optimal blades are compared to the baseline design (based on the Method of Characteristics) and to a blade obtained by means of a deterministic CFD-based optimization.