Risk Warning for Aircraft Bleed Air System with Multivariate State Estimation Technique

The engine bleed air system (BAS) is one of the important systems for civil aircraft, and its operating state directly affects the operational safety of the aircraft. The effective risk warning of the BAS is critical to improving aircraft safety and operators' profits; thus, a multivariate state estimation technique with the dynamic process memory matrix is proposed to warn the failure risk of the bleed air system. First, to obtain the optimal estimation value of the observation vector, the memory matrix is formed by searching for the first vectors that are similar to each input observation vector from the healthy data pool that can cover the common working space. Then, the similarity function is defined to quantitatively measure the deviation between the observed vector and the estimated vector, and the amount of risk information contained in each variable is quantified by the analytic hierarchy process. Finally, the dynamic threshold different from the traditional engineering experience threshold is designed, based on the idea of interval estimation. The developed approach is validated on an Airbus A320-series aircraft with quick access recorder data for one year. The results show that the proposed strategy can provide an effective risk warning for the abnormal state of the BAS before a failure occurs.