A sampling method for blade measurement based on statistical analysis of profile deviations

The complex geometry and inevitable manufacturing errors of aero-engine blade profile bring difficulties to high-efficiency and high-accuracy measurement. Usually, blade profile is measured by coordinate measuring machine (CMM) in mass production. Since the measurement time and cost increase proportionally as the increase of sampling points, it is essential to study a sampling method that can answer two basically different sampling problems: how many blades need to be sampled from mass production, and how many points need to be measured for a blade. This paper calculates the sample size for statistical analysis with sufficient confidence and accuracy. The correlation analysis for profile deviations of cross sections (e.g. rotation parameters) and cross-sectional points (e.g. point to point normal distances) is conducted, and the proposed method is carried out through traversing the neighboring Spearman rank correlation coefficient and noting low correlations between each other. The studied cases of two real blades indicated that the proposed statistical sampling method can improve measurement efficiency and ensure the measurement accuracy of mass-produced blades.