A Metaheuristics-Based Algorithm To Optimize The Fatigue Spectra Of Mechanical Parts

Abstract

Design usage spectra for fatigue evaluation contain occurrences per hour of each maneuver but no sequential order is imposed. However, when applying fatigue cycle counting methodologies (such as rainflow) associated to Maneuver-to-Maneuver (MTM) spectrum, the order of maneuvers has a preponderant influence. To define a sequential order, mission scenarios based on operator's usage may not cover all possible flight operations and may not be in line with design spectrums found in certification guidance materials. On the other hand, sequences maximizing load amplitude (and then, fatigue damage) may be physically unrealistic and too penalizing. This paper proposes the use of Ant Colony Optimization (a metaheuristics-based algorithm) on a MTM spectrum to generate sequences respecting simultaneously the constraints of occurrence (given by the spectrum) and physical realism (given by maneuver pairwise chaining rules).