|For several years my research group, headed
by Prof. G. S. Dulikravich, has been heavily involved in the design and
optimization of gas turbines blades. I developed a code for the automatic
generation of turbine airfoil shapes based on 10 design parameters. This
code was intended to provide a parameterization with a low number of design
variables, a wide range in shapes, and robustness. Such features are very
important when genetic algorithms are applied to shape optimization problems.
A graphical depiction of the parameterization is show is below.
One unique feature of this parameterization is that it is formulated in terms of quantities known by airfoil designers, such as exit/inlet flow angle and throat distance. This makes it easier for an experienced turbine designer to set meaningful limits on the variables and restrict the optimization code's search area to promising regions.
Click on the images below to see a larger picture
Interactive generation of turbine airfoil sections with coolant passages
Viewing the resulting 3D blade with coolant passages
If the user is satisfied with the blade geometry, the code produces a high quality triangular surface mesh for the geometry and saves it to a file. From this file a 3D volume tetrahedral mesh is created with a commercial grid generator and the configuration is run with our transonic finite volume NSE CFD code.
Click on the image below to see a larger picture