Aircraft Diagnosis Based on Machine Learning Methods

The paper presents a proposal for implementing machine learning methods in the diagnosis and troubleshooting process of the M346 aircraft used by the Polish Armed Forces. Due to the complexity and limitations of traditional documentation and procedures, machine learning techniques offer significant potential to enhance accuracy, reduce downtime and improve operational availability. The approach leverages historical fault data and maintenance records to train models that can recommend troubleshooting actions. The integration of artificial intelligence (AI) into aircraft diagnostics addresses key challenges in the current maintenance paradigm and aims to improve fault isolation, maintenance efficiency, and knowledge retention.