In the continuing goal of developing products with better performance at a lower cost, composites are becoming increasingly prevalent in the aerospace industry. Composite structures offer exceptional performance due to their high strength at a low weight. Additionally, one large integrated composite component can replace ten or more traditional metal parts, dramatically reducing manufacturing costs. For the first time in the aviation industry, companies are beginning to use composites for primary load-bearing components. Boeing announced that the new 787 aircraft will be the first airliner to use composite materials in the majority of the aircraft construction.
One of the main failure modes for the skin of an aircraft is buckling. Stringers are often bonded to composite panel skins to help increase the panels buckling strength; however, if there is a defect in the bond between the stringer and the panel, the structure may be weakened. Developing analysis tools that can accurately simulate the behavior of composite materials is critical to the successful design and use of composite structures.
The Cooperative Research Centre for Advanced Composite Structures (CRSACS), based in Australia, performed an experimental test of a skin-stringer panel, studying the strength behavior of a stringer-stiffened composite panel subjected to in-plane shear loading, assuming an imperfection in the bond between the skin and stringer. This imperfection represents the largest undetectable void in the bond. Simulation of the panel test was then performed using ABAQUS v6.5 analysis software.
The information that analysts and designers seek from the simulation of stringer-stiffened panels includes: (1) the total load-carrying capability of the panel, (2) the load at which a crack in the stringer bond line initiates, and (3) the stringer debond and/or crack growth characteristics. In this study, the experimental test results and the simulation data correlated well, indicating that the software was able to predict the behavior of the skin-stringer test panel.
For physical testing, engineers used a flat composite panel with an unbonded region introduced in a portion of the bond between the skin and a single stiffening stringer. The panel and the stringer were co-cured to each other, and were made out of similar composite materials employing similar ply lay-ups. The unbonded region was produced with a release film to impede the adhesive flow and create a void. To ensure that the test would determine reliable part strength, the void was the largest flaw or crack undetectable by inspection techniques.
The particular failure mode under investigation was buckling of the panel under in-plane shear loading. The skin-stringer panel was loaded by applying an enforced displacement at the diagonal corner nodes of the panel. Load range was from 0-200 kN. It was assumed that the initial flaw in the bonded structure would reduce maximum load capability, and also that the crack would grow and further reduce panel strength.
Through the Composites Affordability Initiative (CAI), Boeing developed a finiteelement implementation of the Virtual Crack Closure Technique (VCCT) for simulating crack propagation in composite parts. Subsequently, ABAQUS teamed with Boeing to provide this technology as an add-on feature to ABAQUS/Standard software. VCCT for ABAQUS software has been developed for simulation of fracture and failure analysis, including crack growth in bonded composite structures. This capability allows for the simulation of brittle interfacial crack propagation due to delamination or debonding.
The finite element model was mainly comprised of layered shell elements. The region around the area of the initial flaw was more finely meshed than the remainder of the panel. Mesh tie constraints were used to join the fine mesh region to the coarse mesh region, and to connect the panel to the stringer, away from the crack region. The mesh tie constraint facilitates modeling by automatically connecting meshes of different densities with a surface-based formulation.
The fine mesh areas of the panel and stringer were connected using contact surfaces. Within the contact surfaces, the areas that are initially bonded and unbonded need to be defined. In Figure 1, the initially unbonded area is shown in orange and the initially bonded area is shown in dark blue.
Preparing an ABAQUS model for VCCT analysis is similar to using the ABAQUS/ Standard debond capability based on contact pairs. The surfaces for the region of possible debonding must be created; these surfaces are then used in a contact pair. The interaction properties for this contact pair are then defined to include the fracture parameters, such as the fracture toughness and the mode mixity criteria. Analysis of the skin-stringer test panel utilized the power law model for crack growth:
In the VCCT simulation, the test panel buckled under the imposed displacement. The buckling began before the crack in the panel grew, and subsequent growth of the crack reduced the overall stiffness of the panel. The crack initiated (first node released) at a load of 116 kN and growth of the crack occurred at 146 kN, at which point many nodes released and the crack continued to grow with increasing load.
The simulation results are very close to observations made during the physical test. Three nominally identical panels were tested. Based on audible acoustic emissions, crack initiation and/or growth for the panels occurred at 146 kN, 154 kN, and 175 kN. These load values agreed well with the crack growth predicted in the simulation. Also, the location and shape of the predicted crack growth correlated well with experimental observations.
The graph in Figure 2 shows very good correlation of results between physical testing of the bonded skin-stringer panel and simulation of the physical test. ABAQUS VCCT capabilities successfully modeled the crack growth in the bond between the stringer and the panel, and the software predictions matched the cracking characteristics of the structure.
This work was done by Kyle Indermuehle, Industry Solutions, Aerospace Applications, at ABAQUS, Inc. For more information, visit http://info.ims.ca/5656-122.