Failure prediction of composite bonded joints by analyzing strain homogeneity

Open Access
Author:
Collins, James Grayson
Graduate Program:
Materials Science and Engineering
Degree:
Doctor of Philosophy
Document Type:
Dissertation
Date of Defense:
November 11, 2011
Committee Members:
  • Christopher Muhlstein, Committee Chair
  • Charles Bakis, Committee Member
  • Jim Runt, Committee Member
  • David Green, Committee Member
Keywords:
  • failure prediction
  • digital image correlation
  • bonded adhesive joints
  • strain development
Abstract:
Composite joints and structures are used xtensively in applications where a high strength to weight ratio is required, such as the frame in aircraft wings. However, strain fields in the composite sections and load transfer between them and other bonded material is complicated. Models have been unable to capture the strain complexity, limiting the scope and accuracy of descriptive and predictive models. Additionally, models for strain development in these structures have not been verifieded with experimental observations. Experimental determination of these complex strain fields would allow for failure prediction and improved joint design. In this work, double lap strap (DLS) specimens with EA 9394 epoxy adhesive were monotonically loaded in tension. Deformations in the adhesive regions of the specimens were converted to strain maps using digital image correlation. The resulting measurement of the shear modulus (2.12 GPa) was significantly higher than the manufacturer's reported value of 1.46 GPa and the neat epoxy value of 1.44 GPa. Even though the joint behaved linear-elastically, the strain fields were complex, exhibiting locally positive and negative regions of shear strain that were not predicted by linear-elastic nite element models. The size and distribution of the hots spots were characterized by a homogeneity index, He. The index increases with increasing applied stress for all specimens, but at a given fraction of failure stress, He is higher for higher strength specimens. He is a robust predictor of a specimen's performance between 30% and 90% of a specimen's strength. This non-destructive, predictive evaluation of joint quality can be used as a tool to improve joint design and manufacturing