Fault detection, diagnosis, and performance assessment scheme for multiple redundancy aileron actuator

This paper presents a prognostics and health management (PHM)2 scheme for a multiple redundancy aileron actuator (MRAA),3 which includes fault detection, fault diagnosis, and performance assessment. The scheme utilizes the system input, system output, force motor current (FMC),4 and aerodynamic loads for fault detection, diagnosis, and performance assessment. Fault detection is implemented using a two-step radial basis function (RBF) neural network. The first RBF neural network is employed as an observer and generates the residual error, and the second RBF neural network synchronously generates the adaptive threshold. Fault diagnosis is carried out using a system observer and an FMC observer. First, a force motor observer is used to estimate the FMC. Then, the FMC ratio of each channel can be calculated using the estimated FMC and actual FMC. Finally, a fault diagnosis is achieved by comparing the FMC ratios for the channels. For performance assessment, the system observer is adopted to generate the residual error. Then, time-domain features of the residual error are extracted. Finally, the features are input into a pre-trained self-organizing map neural network to realize the performance assessment. The effectiveness of these approaches is demonstrated using several tests at the end of this paper.