Abstract
In this paper, closed-form solutions for the adhesive stresses in bonded composite single-strap butt joints have been obtained. Two strategies were used for deriving the adhesive peel stress. The solutions are applicable to a butt joint made from different adherend and doubler laminates, as well as the unbalanced single-lap joints. In addition, three-dimensional finite element models of the unit-width composite joints were created for analyzing the adhesive stresses under a plane strain condition. A total of six joint conditions, three joint configurations and each with two layup sequences, were studied. Consistency in the peel stress predictions was obtained from the two theoretical strategies. Good agreement has been achieved between the theoretical and finite element results. The effects of the doubler thickness and laminate layup sequence on the adhesive stress variation can be displayed. The theoretical solution would provide a solid foundation for supporting the practical composite joint assessment.
Introduction
The integration of large carbon-fiber sections in modern aerospace structures, such as the Boeing 787 fuselage, necessitates reliable joining techniques. The single-strap butt joint is a prevalent method, involving two panels (adherends) joined end-to-end via an adhesive-bonded "doubler" plate. A significant engineering challenge arises because these joints are typically "unbalanced"—the adherends and doubler often possess differing thicknesses or laminate layup sequences. This asymmetry induces secondary bending and a complex coupling between shear stress (sliding force) and peel stress (pulling force) within the adhesive layer.
While Finite Element Analysis (FEA) is a standard tool for simulating these stresses, it is computationally intensive and often lacks the transparency needed for rapid design optimization. There is a critical need for precise, analytical closed-form equations capable of handling unbalanced composite geometries.
Methodology
This study derives governing differential equations for the adhesive layer based on beam theory. Two distinct theoretical strategies were developed to solve for the adhesive peel stress:
- Lagrange’s Method: This involves solving a non-homogeneous fourth-order differential equation by combining general and particular solutions.
- Uncoupling Strategy: This approach derives a single, uncoupled sixth-order differential equation for the peel stress, avoiding the "lengthy and sensitive terms" associated with the first strategy.
To validate these mathematical models, high-fidelity 3D Finite Element Models were constructed using 20-node brick elements. The validation suite included six scenarios varying doubler thickness (16, 24, and 32 plies) and laminate layup sequences: S1 ([45/−45/0/90]2s) and S2 ([0/90/45/−45]2s).
Key Insights and Results
The theoretical predictions for both peel and shear stress demonstrated excellent agreement with 3D FEA results across all tested conditions. Several key insights emerged:
- Efficiency of Uncoupling: Strategy 2 was identified as the superior analytical approach due to its mathematical conciseness.
- Impact of Layup Sequence: Fiber orientation significantly influences peak stress. The S2 layup, placing $0^\circ$ fibers on the exterior, resulted in higher bending stiffness and distinct stress profiles compared to the S1 layup, which had approximately 80% of S2's bending stiffness.
- Design Optimization: Increasing the thickness of the doubler plate was shown to effectively reduce peak adhesive stresses, providing a clear design lever for improving joint longevity.
- Critical Stress Points: The highest stresses were consistently located at the inner overlap edges (where the adherends meet), identifying these as the primary sites for potential failure.
Conclusion and Future Work
The closed-form solutions presented here offer a robust and efficient alternative to numerical simulations for the initial design and assessment of composite joints. While the current beam-theory model cannot capture the shear stress drop to zero at the extreme edges—a known limitation of the underlying approximations—it provides high accuracy elsewhere.
Future research will extend these solutions to calculate Strain Energy Release Rates (Mode I and Mode II) for crack growth prediction and explore the optimization of hybrid joints incorporating both adhesives and mechanical fasteners.
Summary updated