Lumped parameters multi-fidelity digital twins for prognostics of electromechanical actuators

Abstract

Abstract The growing affirmation of on-board systems based on all-electric secondary power sources is causing a progressive diffusion of electromechanical actuators (EMA) in aerospace applications. As a result, novel prognostic and diagnostic approaches are becoming a critical tool for detecting fault propagation early, preventing EMA performance deterioration, and ensuring acceptable levels of safety and reliability of the system. These approaches often require the development of various types of multiple numerical models capable of simulating the performance of the EMA with different levels of fidelity. In previous publications, the authors already proposed a high-fidelity multi-domain numerical model (HF), capable of accounting for a wide range of physical phenomena and progressive failures in the EMA, and a low-fidelity digital twin (LF). The LF is directly derived from the HF one by reducing the system degrees of freedom, simplifying the EMA control logic, eliminating the static inverter model and the three-phase commutation logic. In this work, the authors propose a new EMA digital twin, called Enhanced Low Fidelity (ELF), that, while still belonging to the simplified types, has particular characteristics that place it at an intermediate level of detail and accuracy between the HF and LF models. While maintaining a low computational cost, the ELF model keeps the original architecture of the three-phase motor and the multidomain approach typical of HF. The comparison of the preliminary results shows a satisfactory consistency between the experimental equipment and the numerical models.