Analysis and Reduction of Errors in an Automated Robot-Based Repair Process Chain

Abstract

Even though modern industrial robots have good repeatabilities, their positioning accuracies are still relatively poor. Moreover, in a complex process chain, involving several handling systems and diverse interdependent tasks, error propagation can make matters worse. In order to achieve the overall desired quality level, intelligent and highly adaptive methods are required to reduce individual errors and remove accuracy couplings as much as possible. This is especially true in high-risk applications, as found in the aviation MRO industry. Because of the difficulty to replicate existing manual MRO accuracy levels, automation in this area is still relatively scarce. For instance the inspection and repair of airplane combustion chamber liners are as yet performed fully manually. In this paper an automated version of the entire liner repair chain is introduced: from robot-guided white light interferometer inspection in a first cell, to part and data transfer to a second robot cell through to the automated repair steps. Particular consideration is given to individual error sources, such as robot and sensor inaccuracies, calibration deviations and the transfer of data between robot cells, as well as error propagation and prevention.