Date of Award


Level of Access Assigned by Author

Campus-Only Thesis

Degree Name

Master of Science (MS)


Civil Engineering


Eric N. Landis

Second Committee Member

Roberto Lopez-Anido

Third Committee Member

Lawrence Thompson


The goal of this research is to understand the fracture characteristics of woven EGlass/ Derakane 8084 vinyl ester composites typical of those used by the US Navy, so that the reliability of composite joints can be understood in greater detail. Fracture characteristics of the woven composite were evaluated using standard fracture mechanics principles and testing procedures. Quasi-static and fatigue loading tests were conducted on small coupons cut from a large panel of laminated composite. These tests simulated mode I (tension/peeling), mode II (in-plane shear), and mixed mode conditions. Primary (co-infused laminate) and secondary (5 layers laminated, cured, and 5 more layers laminated at a later time) composite bonding methods were tested and compared in this analysis. Mode I tests were carried out concurrently as part of this study by others. The results of the mode I tests are included within this thesis. Mode II and mixed mode static tests were carried out within this research, as was mode II fatigue. Among the experimental issues that needed to be addressed was the measurement of crack length. For mode II and mixed mode loading, the modest transparency of the specimens allowed them to be illuminated such that crack length could easily be monitored with a digital video camera. Resistance based crack gages were also used, but were not found to be reliable for mode II loading. Mode I, mode II and mixed mode test results were modeled with resistance curves, which are a plot of critical strain energy release rate (Gc) at different crack lengths. Mode I resistance curves rose sharply and plateaued, while mode II resistance curves increased more gradually. Resistance curve shape in both modes can be attributed to the toughening mechanisms that affect each mode during crack growth: fiber bridging in mode I fracture, and crack arrest due to transverse fiber bundles in mode II fracture. Typical crack initiation values ranged from 279 to 1032 J/m2 in mode I fracture tests, and from 379 to 2422 J/m2 in mode II fracture tests. Mixed mode resistance curves varied due to combined modal effects. When mode I, mode II and mixed mode Gc results were compared, Gc results were generally lower in the mixed mode tests than the pure mode I and mode II tests, while G/c results were generally lower than Gnc results. Mode I and mode II fatigue tests were conducted through a constant displacement test, at 20, 40, 60 and 80 % of the measured Gc from static results. Crack growth of Vi inch was measured using the resistance based crack gages. In all but one case, mode I resisted crack growth for more cycles than mode II fracture. The nature of the displacement control test for mode II fatigue led us to question the conservativeness of the result. To address this concern, additional fatigue tests were run using a nominally constant G. The constant G tests showed little difference at G = 40% of Gnc, but showed significant difference at G = 20% of Gnc