Worn area modeling for automating the repair of turbine blades

Abstract

The current repair process for worn-out turbine blades and blisks is very manual, and experience based and hence is error-prone. It requires sophisticated CNC-driven (Computer Numerical Control) laser equipment to replace manual welding for the repair. This paper presents an innovative strategy that can lead to the automation of laser welding and cladding. The project makes use of reverse engineering techniques to capture the geometric shape of the worn area by digitized point cloud and nominal geometry. The core software technologies include four aspects: (i) point-to-surface-best-fitting technology that puts the point cloud coordinate system nominal CAD (Computer Aided Design) coordinate system, (ii) a procedure to extract a worn boundary that automatically separates the worn area and the undamaged area, (iii) geometric model representation for the worn area, and (iv) generation of an STL file to drive the CNC laser machine. The tool has been implemented using UG API (Unigraphics Application Programming Interface). Some experimental results have shown that the presented strategy is efficient for repair automation.