Aircraft Lifecycle Digital Twin for Defects Prediction Accuracy Improvement

Abstract

Prediction of defects is essential for Maintenance, Repair and Overhaul organisation (MRO) of aviation industry in order to plan workload, tools and hangar capacity and materials. Standard experience-based approach that uses man-hours rate to scheduled works taking into account aircraft age together with prediction of spare parts requirements on the basis of historical consumption does not provide sufficient accuracy as each airplane has unique operational life cycle and technical condition. It brings to over or under capacity and overstock. This paper describes limitations of current approaches and proposes an approach to modelling of aircraft operational life cycle as a digital twin. Digital twin is a digital copy of a physical object that allows to simulate the behaviour of the object in a real-world environment. With the development of modern IT technologies, it has become a popular tool in high tech and resource-intensive industries such as aviation. This paper presents digital twin of aircraft systems plus operational and maintenance environment as a cloud of data. Here we consider machine learning methods for this digital twin increase prediction and planning precision. Further we will describe the ontology of operational life cycle data based on the analysis of an MRO experience that can be a basis for building of aircraft digital twin.