An energy - based prognost ic f ramew ork to predict evolut ion of damage in composite materials 16

Composite materials are high-performance engineering materials increasingly used by the aerospace industry in part because of their high strength-to-weight ratios. Fatigue damage represents one of the most important sources of concern for in-service performance, which has led to growing research interest in industry and academia. Influenced by a long-standing understanding of the principles of metal fatigue, the initial treatment for fatiguebehavior in compositeswassimilar to metallic structures [1], and asaconsequence, numerousmodelswere formulated asextensionsof metal fatigue theories [2]. However, unlike metals, fatigue damage in composites comprises multiple simultaneous internal fracture mechanisms such as matrix microcracks, delamination, fiber breakage, etc., that ultimately lead to significant changes in the macroscale mechanical properties of the material over its lifespan [3,4]. Thismultivariatedamageprocess leads to uncertainty in assessment of current and futurematerial properties. Thisstemsin part from thematerial heterogeneity and an incomplete knowledge about the physics behind the evolution and interaction of damage mechanisms. Fatigue damage predictions using deterministic models in absence of any ground truth information about the current degradation state are not expected to provide much accurate information about the state of health of the material. However, real-time measurements of the structural performance are now possible through state-of-the-art structural health monitoring (SHM) techniques, and a large amount of responsedatacan bereadily acquired and further analyzed to assessvarious health-related properties of structures. Therefore, a more suitable approach for fatigue damageprediction isthrough theuseof monitored responseof thestructureto updatea given damage model so as to make more accurate predictions that also account for uncertainty. Development of such SHM-based damage prognostics approach is the core objective of this chapter.