Semi-automatic process control for efficient refurbishment of turbine blades

The refurbishment of turbine blades requires the precise removal of damaged surface coatings. In manufacturing companies, this usually involves time-consuming activities such as detecting residual coatings and adapting process settings to varying turbine blade geometries. The use of automated systems therefore opens up opportunities to improve the efficiency of turbine blade refurbishment processes by replacing manual tasks. This paper presents an conceptual approach for a semi-automatic decoating process of turbine blades that integrates abrasive blasting technology into a closed loop process control system. The experimental setup consists of a robot for guiding the turbine blade in a blast machine and a nozzle system for local removal of residual coatings using abrasive material. Based on image processing the residual coatings of a turbine blade are labeled by a neural network and mapped to a 3D model of the turbine blade, which is used by a software control system to coordinate the decoating process. Using a prototype setup this paper investigates the applicability of the proposed approach and evaluates its feasibility.