A new genetic algorithm model-based prognostic approach applied to onboard electrohydraulic servomechanisms

The ever green solution of the electro hydraulic actuator (EHA) applications for the control of modern primary flight commands, justified by the superiority of hydraulic systems in furnishing more efficient solutions for power supplying in a controlled manner, brings us to focus on the need to make the EHA as efficient and reliable as possible. To this purpose, it must be noted that reliability of modern systems is increasingly more based on the valid support of diagnostics and prognostics; in fact, these two are the most robust instruments which mitigate life cycle costs without losing reliability and guarantee, in compliance with regulations, the bases for health management of integrated components, subsystems and systems. Developing a fault detection algorithm able to identify the precursors of EHA faults and their degradation patterns is thus beneficial for anticipating the incoming failure and alerting the maintenance crew so as to properly schedule the servomechanism replacement. About that, this paper proposes a new EHA model-based fault detection and identification method (FDI) that makes use of deterministic and heuristic solvers in order to converge to the actual state of wear of the tested actuator. The proposed FDI algorithm has been tested on three different types of progressive failures (the clogging of the first stage of the flapper-nozzle valve, the rising of friction between spool and sleeve and finally the rising of friction between jack and cylinder): to this purpose, a dedicated simulation test environment was developed. Results showed an adequate robustness and a suitable confidence was gained about its ability to early identify EHA malfunctions with low risk of false alarms or missed failures.