CFD Driven Optimization of Hydraulic Turbine Draft Tubes using Surrogate Models

Draft tube designs have to a large extend been based on intuition and on the experience of the design engineer. In a close future, CFD simulations coupled with optimisation algorithms will assist in the search for an optimal technical solution. Such a shape optimisation technique to redesign an existing draft tube is presented in this paper. By this method, a design can be predicted in terms of a predefined objective function, here the pressure recovery factor. The optimisation is performed with the Response Surface Method (RSM) implemented on the commercial code iSIGHT7.0, while the CFD simulations are carried out with the commercial code CFX4.4. The boundary conditions are based on detailed experimental data and the turbulence is modelled with the standard kturbulence model. With this set-up and with two different parameterisations, Adapted and Profile Design respectively, an optimal geometry of the ERCOFTAC Turbine99 draft tube can be predicted.