Design of Radial Turbines Operating in the Organic Rankine Cycle Using Optimization and CFD

Angie Lizeth Espinosa Sarmiento (Universidade Federal de Itajuba)

CHALLENGE - This thesis work presents a methodology for the aerothermodynamic design of radial turbines used in subcritical Organic Rankine Cycles (ORC). The proposed method integrates CFD techniques, construction of response surfaces using Radial Basis Function (RBF) and genetic algorithms. 

SOLUTION - The first step is a one-dimensional analysis where the thermodynamic properties of the working fluids, in this case R245fa and R123, are calculated from previously validated models of real gases contained in the REFPROP© program. Computational Fluid Dynamics (CFD) simulations are then used to determine the main performance characteristics of radial turbines over a wide range of operating conditions. Finally, in order to achieve the highest possible efficiency at the design point, an optimization procedure in optimization platform modeFRONTIER was applied to the rotor of the radial turbine operating with R245fa as the working fluid. NSGA II genetic algorithm was chosen to solve the optimization problem.








BENEFITS - The methodology developed is shown to be an appropriate tool for the design of radial turbines, allowing effective definition of some geometric parameters which are absent in the one-dimensional model, thereby improving performance and reducing computational cost. When analyzing the optimized CFD project, in relation to the non-optimized CFD project, an increase of 1.9% to 3% is observed in total efficiency and power. These results indicate that the optimization methodology is effective, since the increase obtained in radial turbine efficiency is considered significant in the study of this type of turbomachinery.