One important factor in the durability of polymeric composites is their loss in stiffness over time due to many softening mechanisms, including non-linear Viscoelasticity, viscoplasticity and damage. Damage here refers to all ply-level microstructural changes such as matrix cracking, fiber-matrix debonding and shear yielding. This dissertation uses the theory previously established by Schapery (1999) to develop experimental and data analysis methods for isolating these softening effects.
Schapery’s constitutive theory is first tailored for a continuous fiber composite and evaluated for creep/recovery loading where nonlinear Viscoelasticity, viscoplasticity and damage growth have a significant effect on strain. Numerical methods, implementing a Genetic Algorithm, are developed to fit material parameters in the recovery equations. This method successfully fits simulated recovery data with hereditary damage effects, but was not implemented on real data due to the unusually complex recovery behavior of the material studied.
A method of Acoustic emission monitoring and waveform analysis is developed as a means for tracking two of the primary damage mechanisms in these materials, matrix cracking and fiber/matrix debond. With direct monitoring, the extent of damage in the material does not need to be inferred from its effect on the stress-strain response. Unidirectional 30o, 45o, and 90o coupons of a rubber-toughened carbon/epoxy are monitored in this way for various loading histories. A method of comparing waveforms from different samples is also suggested. An interpretation of the AE data is proposed based on an initial population of existing flaws. Then a cumulative distribution function (CDF) of microcracking is defined and used to study effects of stress history. After developing an idealized model of the material consisting of two Viscoelastic phases, a single loading parameter, which is theoretically independent of loading history and derived from Viscoelastic fracture mechanics, is found to collapse data from all samples and loading histories, thus supporting the theory.
Finally a Damage Effect Study is proposed which identifies the material parameters affected by damage, thereby separating the damage and stress effects on softening. This method is based on vertical shifting of recovery data at different damage states, much like vertical shifting for the effect of stress. Two significant simplifications are found for the material studied: damage does not affect the time scale of the Viscoelastic strain and enters through only one parameter in the transverse strain. Viscoelastic shear strain requires two parameters, however. Also, the elastic component of the modulus is found to increase with increased damage. Results from material testing at fixed damage states indicate a 2-phase Viscoelastic constitutive model may be needed to characterize this particular rubber-toughened composite material.