Multi-objective aerodynamic optimisation of a radial-inflow turbine

Optimisation, robustness and reliability analyses have an increasing importance in turbomachinery industry engineering. With continuing progress in computational simulations, computational-based optimisation has proven to be a useful tool in reducing the design process duration and expense. This paper describes an optimisation procedure to modify the geometry of a 7 kW R245Fa radial-inflow turbine from a concentrated solar resource of 150oC with a pressure ratio of 3.7 to improve aerodynamic efficiency and satisfy manufacturing constraints. The procedure integrates the parameterisation of the turbine blade geometry, multi-objective optimisation, and 3D CFD analysis. ANSYS-BladeGen was applied to create the 3D geometry of the flow passage carefully examining the proposed design against the baseline geometrical data. Generating the required computational mesh with ANSYS-TurboGrid followed by grid refinement, CFD simulations are then performed with ANSYS-CFX in which three-dimensional Reynolds-Averaged Navier-Stokes equations are solved subject to appropriate boundary conditions and real gas properties (RGP) - generating a required table from REFPROP - was used. Considering a steady state solution, a high resolution for both Advection Schemes and Turbulence Numeric were applied resulting in higher accuracy while it reasonably increases the computational cost. OptiSlang Dynardo was used to conduct a multi-objective optimisation and identify the most relevant input parameters in order to reduce the numerical effort for the optimisation algorithm. Implementing evolutionary algorithm resulted in a Pareto front to choose a nominal design for a subsequent reliability analysis and define previously unknown feasible design space boundaries.