Metaheuristic Bio-Inspired Algorithms for Prognostics: Application to on-Board Electromechanical Actuators

Abstract

Metaheuristic bio inspired algorithms are a wide class of optimization algorithms, which recently saw a significant growth due to its effectiveness for the solution of complex problems. In this preliminary work, we assess the performance of two of these algorithms - Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) - for the prognostic analysis of an electro-mechanical flight control actuator, powered by a Brushless DC (BLDC) trapezoidal motor. We focus on the first step of the prognostic process, consisting in an early Fault Detection and Identification (FDI); our model-based strategy consists in using an optimization algorithm to approximate the output of the physical system with a computationally light Monitor Model.