Dynamic Identification of Actuator Frequency Response using Wavelets

Current techniques for monitoring actuator performance are inferential, that is, implied via indirect system diagnostics. Results of such methods are neither easily nor accurately interpreted into metrics that can monitor precision and stability of the actuator response. The approach proposed herein identifies the frequency response of an actuator with the Wavelet Eigensystem Realization Algorithm (WERA) method. This proposed technique, originally developed by STI for aeroservoelastic system identification, will use the input and output signals from the actuator, thus offering direct information on its status relative to design specifications and maintenance limits. The technique employs discrete wavelet transforms to compute the impulse response (Markov parameters) of the estimated actuator system. This is then used in the Eigensystem Realization Algorithm to compute the discretized state-space matrices. Metrics are generated assessing the validity of the identified system. Computation is conducted in near real time. The application of the WERA technique will first be assessed via simulation analysis conducted using nominal and off-nominal model representations for propulsion and flight control actuators. A laboratory demonstration of the system identification capabilities will be conducted with exemplar hardware in the loop to evaluate the feasibility of a WERA-based software approach to actuator health monitoring.