The emergence of digital twin technology offers a promising solution to address the limitations of traditional methods on early diagnosis and accurate propagation analysis of flight ground service delays. However, the application of digital twin technology in the civil aviation domain still stays at the lower maturity of the L2 level, which focuses on physical assets, operational data, and maintenance planning at airports, and failed to achieve the integration of flight ground operation mechanism and real‐time data, making it difficult to realize timely delay diagnosis. The simulation model is also limited to the offline simulation technology, which cannot connect to real‐time data for simulation from intermediate processes. In this work, we developed an advanced L3‐level airport digital twin system for flight ground service processes delay diagnosis and propagation, which focused on real‐time data‐driven simulation models and machine learning applications to meet the timely and precision requirements. First, we used the Unity3D platform to construct static three‐dimensional models of flight ground service objects on the airport cloud server. By parsing these behavioral state interfaces and mapping real‐time dynamic data from the airport sensing and business systems, we achieved accurate visualization of the airport’s dynamic operational processes. Then, a vehicle delay tree–based Bayesian diagnostic model was proposed in the digital twin system to analyze the relationships between multiple flights and service processes, which enables proactive diagnosis of the operation status and provides delay warning information. To improve the accuracy of propagation analysis, we proposed a “breakpoint” simulation method that enables real‐time simulation starting from an intermediate moment, facilitating the inference of flight ground service delays since the early warning moment. In addition, two delay tracing and propagation algorithms were proposed to identify delays and investigate propagation paths. Leveraging real‐time operational information, our approach provides valuable feedback for decision‐making, empowering the airport manager to formulate precise optimization strategies. Experiments on real‐world airport data have validated the effectiveness of our proposed method and provided practical recommendations for airport managers to reduce aircraft delays and improve airport operation efficiency.
Read full abstract