Abstract
As the aerospace industry continues its shift towards digital manufacturing, an increased use of inline sensors and data collection is creating an opportunity to further optimize the production process. Measurements from individual parts make it possible to adapt each process to current conditions rather than running all processes nominally. This is sometimes described as individualized production as opposed to traditional mass production. The concept of a digital twin for manufacturing has recently gained more attention as a promising method for individualized production. A digital twin collects data from a real environment to create a virtual copy of a physical phenomenon, which can be used to predict how its real counterpart is going to behave. The approach has been proposed for a manufacturing environment where it would be used to predict the outcome of a production process. This could prove particularly useful for fabrication processes, a method used for making aero engine parts by joining large assemblies of smaller parts through welding. This paper presents functionalities that can be used to implement a digital twin in a high precision fabrication process, outlining different approaches for data collection, data analysis, and adaptive process adjustments. An example is shown where physical measurements are used to improve the predictive capabilities of a welding simulation in order to enable more accurate process adjustments.