A holistic digital twin for service life extension programs

icaf2023 Tracking Number 58

Aircraft are affected by their continuous exposure to the conditions under which they operate. The well-known combination of fatigue, environmental and accidental damages in the airframe leads to a progressive reduction of the capability to carry load. Although the maintenance strategies promoted by the civil and military regulatory frameworks ensure that the degraded structure will withstand limit load at any moment, thus solving the safety implications of ageing, there are still significant economic considerations to be addressed in terms of the growing maintenance costs needed to meet these requirements, including inspections, repairs and replacements. This scenario is particularly demanding in the case of some of the products developed by Airbus Defence & Space, such as CN-235/C-295 tactical aircraft for maritime patrolling missions, in which a strong interaction between fatigue and corrosion appears, combined with a severe usage pattern that evolves throughout the service life depending on the operational needs. Among the options available to implement a Service Life Extension Program (SLEP) for these aircraft, the use of a digital twin (called Fatigue Digital Equivalent, FDE) has been selected in order to determine the condition of the structure by simulating the occurrence, growth and eventual interaction of the different degradation sources. At its core, the FDE is a collection of deterministic and probabilistic models representing all the aspects involved in the airframe's safety including design, manufacturing, maintenance, repair, configuration management and flight operations. The models are then incorporated into a set of holistic analyses used to estimate the remaining useful life of the relevant elements. One of the main challenges of the construction of the FDE is the integration into a single repository of heterogeneous sources of data, including design information created in many cases years ago by multiple teams working in silos, maintenance records generated over the years following different formats and standards, and external in-service inputs provided by the operators. Another relevant activity is the combination of conventional analysis tools with state-of-the-art procedures, such as Machine Learning (ML), in a harmonized process aimed to obtain an accurate assessment of the potential for extension of the structure.