International Journal of Engineering and Technology

In this paper we propose a method of improvement of operating accuracy of iterative closest point algorithm used for metrology problems solving when determining a location deviation. Compressor blade profiles of a gas turbine engine (GTE) were used as an object for application of the method of deviation determining. It is proposed to formulate the problem of the best alignment in the developed method as a multi-objective problem including criteria of minimum of squared distances, normal vectors differences and depth of camber differences at corresponding points of aligned profiles. Variants of resolving the task using an integral criterion including the above-mentioned were considered. Optimization problems were solved using a quasi-Newton method of sequential quadratic programming. The proposed new method of improvement of the registration algorithm based on geometric features showed greater accuracy in comparison with the discussed methods of optimization of a distance between fitting points, especially if a small quantity of measurement points on the profiles was used.