Model-based prognostic health-management algorithms for the freeplay identification in electromechanical flight control actuators

The paper deals with the development of a modelbased prognostic algorithm for the freeplay identification in electromechanical flight actuators. The basic idea underlying the work is that the freeplay implies the onset of limit cycle oscillations in the actuator position response, with amplitudes and frequencies that depend on the freeplay size, so that the limit cycle characterization can provide an estimation of the freeplay. To develop the algorithm, a number of prognostic models are generated from a high-fidelity model of an electromechanical actuator for primary flight controls, in which increasing values of freeplay are imposed. The limit cycles obtained by each prognostic model are characterized, weighed against uncertainties, and collected to create a reference database. The algorithm, designed to be used during maintenance, collects the time histories of the actuator sensors’ during position tracking tests capable of inducing freeplay-related limit cycles, and operates a signals’ treatment (FFT, normalization, amplification, filtering) aiming to amplify the limit cycle content with respect to both high-frequency disturbances and low-frequency dynamics. The amplitude and the frequency of the limit cycle oscillation are then compared with the expected values of the prognostic database, up to generate a freeplay estimation and a Remaining-Useful-Life prediction.