Fault Prediction Technology of Civil Aircraft Based on Qar Data

Abstract

As a quick access recorder (QAR) loaded on the plane, QAR data records hundreds of parameter data produced in the process of flying. But these data are not currently being used effectively. It is becoming increasingly necessary to use QAR data for fault prediction. First, the technical route of the civil aircraft fault prediction is illustrated. Second, four fault prediction methods of civil aircraft based on the QAR data are introduced, including performance prediction method based on curve fitting, time trend prediction based on time series, performance trend prediction method based on nonparametric regression analysis and Predictive Method Based on Improved Grey Model(GM). Third, the implementation of the prediction system is described in detail. Thus it can monitor the running state of the aircraft system and parts in order to timely find out fault symptom information, reasonable maintenance plan, and ensure flight safety. Finally, the temperature parameters of the air conditioning system of Boeing aircraft are predicted by the method proposed in this paper. The prediction results verify the effectiveness of the method.