Exploring a Knowledge-Based Approach for Predictive Maintenance of Aircraft Engines: Studying Fault Propagation through Spatial and Topological Component Relationships

Predictive maintenance has become a highly favored application in Industry 4.0, particularly in complex systems with requirements for reliability, robustness, and performance. Aircraft engines are among these systems, and several studies have been conducted to try to estimate their remaining lifespan. The C-MAPSS dataset provided by NASA has greatly served the scientific community, and several works based on physical models and data-driven approaches have been proposed.
However, several limitations related to data quality or data availability of failures persist, and integrating domain knowledge can help address these challenges. In this article, we are currently implementing a new approach based on knowledge coupled with qualitative spatial reasoning to study the propagation of faults within system components until complete shutdown. Region Connection Calculus (RCC8) formal model will be used to describe the component relationships, drawing inspiration from the C-MAPSS dataset.