Algorithm to Construct Line Shaped Feature Point Clouds of Single Value Ruled Blades for Gas Turbines

Abstract

Using the geometric features of surfaces, we proposed an algorithm to construct lineshaped feature point clouds for the reverse design of a blade surface. Blade surfaces in the gas turbines field require high accuracy and have complex shapes. By quickly recognizing the border of the scattered point clouds of blades, the backproject method was used to extract the ruled generatrix vector. By solving the multiobjective optimization problem of border point clouds, the annealing algorithm was applied to match border point clouds nonrigidly. Equally divided points were constructed based on the point clouds registration results to get inerraticorderly lineshaped feature point clouds that are beneficial for generating high quality blade surfaces quickly. The proposed algorithm is very effective for dealing with the point clouds of singlevalue ruled blades. It is also theoretically significant and may be applied for reconstructing the surface of point clouds of singlevalue ruled blade surfaces. Experiments verify the feasibility of the proposed algorithm.