In connection with implementing the concept of electric airplanes, it is necessary to ensure the high reliability of electromechanical actuators (EMAs) as important components of aviation systems. The structural composition of an EMA and the types of its faults are considered. Fault diagnosis methods based either on EMA modeling or the analysis of signals received during EMA operation, as well as hybrid methods combining both these approaches, are reviewed. The advantages, disadvantages, and difficulties in applying these approaches are investigated. Special attention is paid to EMA diagnosis methods based on deep learning, which process signals in au-tomatic mode and implement complex fault diagnosis. The concept of aircraft equipment health management (in particular, EMA health management) is presented based on assessing the tech-nical condition and prognosticating the remaining useful life in order to prevent faults before their occurrence. Several hybrid approaches with prognostics are highlighted to solve the aircraft health management problem. Finally, Russian R&D results in the field of machine learning-based aviation health management are reviewed.
