Quantitative Description of Damage Evolution and Property Degradation of Fiber Reinforced Composite Materials

摘要

Fiber reinforced composite materials are materials are increasingly being usedrnin many applications. However, characterization and prediction of their long termrnbehavior is still an active area of research. Composite inherently heterogeneous andrngo through a complex process of material state changes. Their functional liferndepends upon characterizing and predicting evolution of local details (e.g.rndistributed damage initiation, accumulation, and interaction) which affect globalrnproperty (strength and stiffness) degradation and eventual failure. The lack ofrnunderstanding of those local changes often results in empiricism, limitingrninnovation in the use of composite materials. Importantly, the intermediate, precatastrophicrnstage of damage development when isolated defect sites accumulaternand begin to form incipient fracture paths is especially difficult to characterize. Thisrnproposed research aims at improving material level understanding of the formationrnof fracture path. In this paper, broadband dielectric spectroscopy (BbDS) has beenrnutilized to capture material state changes due to degradation in composite materials.rnThis multi-physical response of composite materials from BbDS provides importantrnfeatures which relate to degradation behavior and associated loss of propertiesrn(strength and stiffness) due different applied conditions. In this study, multiphysicalrndescription of material state change has been discussed in two examplerncases: i) stages of incipient fracture path formation due to cyclic bending, and ii)rnstiffness evolution under tensile fatigue loading. A 3D X-ray microscope has beenrnused to visualize and hence validate incipient fracture path formation. Validatedrnresults from these test cases show that degradation process, and evolution ofrnproperties can be characterized with a high sensitivity using multi-physical responsernof composite material. These results can potentially help enhance predictivernmodeling. Details of experimental methods, and results are included in the paper.