The subject of the presented paper is the validation of a design method for HP and IP steam turbine stages. Common design processes have been operating with simplified design methods in order to quickly obtain feasible stage designs. Therefore, inaccuracies due to assumptions in the underlying methods have to be accepted. The focus of this work is to quantify the inaccuracy of a simplified design method compared to 3D Computational Fluid Dynamics (CFD) simulations.
Short computing time is very convenient in preliminary design; therefore, common design methods work with a large degree of simplification. The origin of the presented analysis is a mean line design process, dealing with repeating stage conditions. Two features of the preliminary design are the stage efficiency, based on loss correlations, and the mechanical strength, obtained by using the beam theory. Due to these simplifications, only a few input parameters are necessary to define the primal stage geometry and hence, the optimal design can easily be found. In addition, by using an implemented law to take the radial equilibrium into account, the appropriate twist of the blading can be defined.
However, in comparison to the real radial distribution of flow angles, this method implies inaccuracies, especially in regions of secondary flow. In these regions, twisted blades, developed by using the simplified radial equilibrium, will be exposed to a three-dimensional flow, which is not considered in the design process. The analyzed design cases show that discrepancies at the hub and shroud section do exist, but have minor effects. Even the shroud section, with its thinner leading-edge, is not vulnerable to these unanticipated flow angles.
On-Line Design Method of Compensation Law for Constrained Systems Based on Generalized Singular Value Decomposition of Linear Systems
