High-Precision Modeling of Turbine Blade from Cross-Section Data

The shape of leading/trailing edge is very important to the performance of the turbine blade. Generally, the leading/trailing edge is a small section of the elliptic arc. In order to model the blade accurately, we propose here a new idea of handling the difficult situation where the noised data span a small section of the elliptic arc. This new fitting algorithm based on the fitting ellipses subject to the constraint of length of semi-minor axis is presented to extract the important parameters of leading edge and trailing edge at first. Validation with experimental results demonstrates the accuracy and the robustness of the parameters extraction method. The discrete data points of cross-section are interpolated with G2 continuous rational Bezier curves. Leading edge and trailing edge are described as rational Bezier curves precisely. Parametric grid distortion directly influences blades machining quality. The reasons of blade parametric grid distortion in modeling process are analyzed, and the key methods to rectify parametric grid distortion are discussed in details, such as cross-section curve discretization and a method for nearly arc length parameterization based on the trigonometric function.