A genetic-based prognostic method for aerospace electromechanical actuators

Prior awareness of impending failures of primary flight command electromechanical actuators (EMAs) utilizing prognostic algorithms can be extremely useful. Indeed, early detection of the degradation pattern might signal the need to replace the servomechanism before the failure manifests itself. Furthermore, such algorithms frequently use a model-based approach based on a direct comparison of the real (High Fidelity) and monitor (Low Fidelity) systems to discover fault characteristics via optimization methods. The monitor model enables the gathering of accurate and exact data while requiring a minimal amount of processing. This work describes a novel simplified monitor model that accurately reproduces the dynamic response of a typical aerospace EMA. The task of fault detection and identification is carried out by comparing the output signal of the reference system (the high fidelity model) with that acquired from the monitor model. The Genetic Algorithm is then used to optimize the matching between the two signals by iteratively modifying the fault parameters, getting the global minimum of a quadratic error function. Once this is found, the optimization parameters are connected with the assumed progressive failures to assess the system's health. The high-fidelity reference model examined in this study is previously conceptualized, developed, implemented in MATLAB-Simulink and finally experimentally confirmed.