Research Of Cross-Sectional Feature Parameter Extraction Of Blade Based On Measurement Data

Aero-engine turbine blade always works under poor environment with high temperature and high pressure. During the production process, complete detection with strict standard must be executed in order to satisfy the quality requirement. The blade profile quality has a decisive impact on the aero-engine performance. However, the complexity and variety of the blade make it difficult to inspect and evaluate the profile, so how to realize the inspection turns to be more and more important.Most of current study focuses on the evaluating of blade feature parameters, such as leading edge, trailing edge, chord line length, max thickness and so on.In this dissertation, a new method for cross-sectional feature parameter extraction of blade from the measurement data is presented. For each feature parameter, the extraction approach is detailed in the paper according to the feature definition. The main contents and contributions of this dissertation are as follows:1) At the stage of the data pre-process, the method for identifying casting vent is proposed based on the monotonicity of the profile data. The method can identify whether the vent exist or not automatically. According to the method of polygon convex, the position of vent is determined and further processing will be adopted.2) Based on polygon convex hull method, the chord line is extracted. The method depends on the geometric character of the blade, consider the pressure surface as the concavest part of the blade surface. During the extract process of convex hull, each concave section is recorded then the pressure surface is picked up which has a maximum distance of first and last point of the concave. The chord line can be got by linking those two points, the length and the angle of chord line are obtained correspondingly.3) The leading and trailing edge is extracted based on the method of maximum radius error. Because the radius of leading and trailing edge is very small, common methods can not satisfy the required precision and robustness.4) A simplification method for computing the maximum thickness of the blade is developed. This method can compute the maximum thickness quickly with proper precision when the blade is divided into convex and concave part roughly.Finally, the software model for feature parameters extraction is developed by using dynamic library technology and visualization tool kit (VTK) based on methods presented in the dissertation. Then, blade 3-deminsional deviation analysis model, thickness analysis model and mean-line extraction function are also well integrated in the software in order to satisfy the requirement of algorithm certification and fulfill the real inspect.