Improved Condition Monitoring of Aircraft Auxiliary Power Unit Based on Transfer Learning

The status of aircraft auxiliary power unit (APU) has direct influence on the safety and reliability of the aircraft. However, only a few on-wing sensors can bring valuable information for condition monitoring, which are not sufficient for formulating effective monitoring model. In this study, one approach is proposed to improve the condition monitoring of the aircraft APU based on transfer learning, which utilizes the abundant simulation data for training the condition monitoring model. It is expected that the feature of simulation data can help to improve APU monitoring result. To verify the effectiveness of this proposed approach, the real data from China Southern Airlines Company Limited Shenyang Maintenance Base and the simulation data from NASA Ames Research Center are utilized.