Approach towards process digitalisation and the integration of digital twins into aircraft maintenance shop floor procedures

Kurzfassung

As an ongoing field of research, a lot of work went into the development of digital twins as an elementary part of the aviation industry’s digital transformation. (Meyer et al. 2020) These are often asset-based and therefore tailored to an individual part, component or system which they accompany through their complete lifecycle. With respect to the aircraft maintenance, these digital twins are utilized to gain data driven insight and derive corresponding services, like condition monitoring, predictive analysis etc. Besides that, another focus of the maintenance digitalization consists of the digitization of manual-based and paper heavy MRO-procedures (Uhlmann and Otto 2014). Widely unconsidered so far is how shop floor level processes have to be redesigned to meet digital twins interface requirements and enable and highly interconnected physical maintenance procedures. To address this gap, we identified the digitized version of a manual-based maintenance process on the example of an aircraft battery servicing as a reference. The implementation has the quality to serve as a blue print for several different manual-based maintenance processes. By adding a layer of programmatic abstraction onto this identification, we generate a generic solution to expose it to different existing und upcoming digital and mechanic interfaces, e.g. ERP, Smart Tooling, digital twins and XR. Additionally, a well described and easily accessible interface can function as a benchmark for upcoming digital twin efforts and enables the transferability to many other shop floor operations. The result of this project is a user-centered (Abras et.al. 2004) digitized workspace prototype, with a highly flexible and modular process design. A focus of the project is the application of multi-domain, AI-based assistance technologies (e.g. OCR, Object detection or NLP) to relieve the mechanic. Due to this, the mechanic is assisted where it is beneficial and not forced to use assumedly hindering systems. By consideration of clean architecture and open source, this solution is highly convertible to other maintenance processes and independent from used techniques. In this realization, pdf based manual data gets converted into a convenient DICOM11 inspired format, from which application units control display, measurement and visual result conditioning. The backend design reduces the data management expense, which comes easily with standardized data formats. The result indicates that the generic linking of digitized shop floor processes to digital twin supports process transformation and optimization efforts.