Description

Design of high efficiency turbines for high speed applications like turbochargers, rocket boosters, cryogenic engines etc. is the need of the hour. Conventionally design of radial turbines is done using non-dimensional charts like specific speed vs specific diameter. These charts are made for best efficiency points and uses the principal of similitude. Though this method gives good results, there is a need for a more robust method of design which puts lesser constraints on the designer. One such method is to use genetic algorithms in sink with CFD simulations. Parametric study is conducted with every iteration being fed to the GA-solver. Emerging geometry is then 3d printed and tested experimentally for performance on a standard rig.