engleski

Optimisation Process for Turbine Blade Application

An automated process for turbine blade optimisation coupled with CFD is presented in this study. A large number of turbine blade shapes is considered in the optimisation process, thus a large number of corresponding computational meshes would have to be used for CFD calculations. It has been well documented that local mesh resolution and mesh quality is one of principal sources of discretisation error in CFD. For consistent results in optimisation studies it is essential to control the absolute value and distribution of mesh- induced errors, ideally by employing a family of self-similar meshes. This is not easily achieved where the automated optimisation loop involves the use of an external or automatic mesh generator. In this study, a procedure of geometry variation in the CFD mesh is performed using a mesh morphing procedure, where the existing mesh will be automatically deformed to accommodate boundary deformation. Mesh morphing not only simplifies the optimisation loop but also produces a family of self-similar meshes, thus controlling the distribution and magnitude of the discretisation error. Mesh morphing procedure has proven robust and reliable even for large geometrical variations. The optimisation process consists of several steps: parametrization of the blade geometry is done using the proprietary tools agreed with the sponsors of the study. Using the mesh morphing utility, the existing mesh is adapted for the new geometry. Flow in the blade passage is then calculated using OpenFOAM. The results of the simulation are used as criteria in the optimisation algorithm. The workflow and results are presented for several publicly available turbine blades.

Optimisation ; Mesh Morphing ; Geometry Parametrization ; OpenFOAM ; Turbomachinery

nije evidentirano