Advanced hybrid algorithm for estimation of circularity deviation in coordinate metrology

Abstract

Abstract In the field of precision mechanics such as automobiles or aeronautics, the estimation of defects in the shape of mechanical parts is a very common task, and the declaration of conformity is mainly based on the precision of this value. This research introduces an innovative hybrid optimization methodology that enhances the precision of estimation of circularity defects. The method proposed combines the genetic algorithm (GA) and the interior point method (IPM), this approach overcomes the limitations of traditional methods like least squares and orthogonal distance regression, which rely on initial estimates parameters to ensure convergence. The GA-IPM hybrid eliminates the need for such initial guesses, leading to faster convergence and more accurate results. To control a mechanical part, a cloud of points is preleved by using a coordinate measuring machines, after treatment of those datasets the estimation of circularity defects is calculated using the proposed algorithm GA-IPM. To validate our approach a comparative study is carried out according to the standard ISO 10 360–6, which consists of comparing the results (substitute geometry parameters) obtained with those of National Institute of Standards and Technology considered as reference values. Several examples have been processed, circle totally or partially measured in order to verify the robustness of the algorithm in terms of convergence and calculation precision, the circularity defect is subsequently calculated on the basis of these parameters. The comparison shows that our method estimates the parameters and the circularity defect well without needing a vector of initial estimates parameters. Given the results obtained, this hybrid approach represents a contribution to the estimation of the circularity deviation, thus offering industries a reliable, precise method to ensure the declaration of conformity of mechanical parts.