Fracture of skin/stiffener intersections in composite wind turbine structures

Douglas Cairns, Daniel Samborsky, Darrin Haugen, John Mandell
1998 1998 ASME Wind Energy Symposium   unpublished
Most wind turbine blades have a stiffening spar running the length of the blade to add rigidity to the airfoil skins. This spar is usually an I-beam or C-channel. The interface between the spar flange and skin surface is often the site of fracture and delamination growth in composite wind turbine blades. Fracture initiates here due to high out-of-plane stresses and stress concentration areas, combined with the low transverse and out-of-plane strength of composite materials. Areas such as the
more » ... reas such as the stiffener flange tip may develop stress singularities due to the geometric mismatch between the flange and skin that cannot be analyzed with standard strength-based criteria. These factors make skin-stiffener detail regions a critical design component in wind turbine blade structures. The goals for this study were to combine experimental testing with finite element analysis (FEA) to establish design guidelines and develop an accurate FEA method for predicting skin-stiffener fracture loads and locations. Experimental fracture toughness tests showed that delamination growth resistance was higher for cracks propagating at a (+45/-45) degree ply interface than for cracks between two (0) degree plies. Increasing the skin bending stiffness and matrix material toughness produced large increases in pull-off loads. Increasing the flange thickness and the adhesive bond-line thickness caused the damage location to change from the web/flange bend region to the flange tip. This was due to the increasing geometric discontinuity at the flange tip, which created high interlaminar stresses. A strength-based failure prediction with FEA results was adequate to predict damage onset in the stiffener specimens in regions without high stress gradients. However, a fracture mechanics approach was necessary to analyze the flange tip region. Good agreement with experimental damage onset loads was obtained by using the one-step virtual crack closure technique (VCCT-1) to calculate strain energy release rate values, which were used with the linear interaction criterion for crack growth to predict propagation loads. An initial crack length of less than 0.2 mm and a crack length to crack extension ratio (a/da) of greater than 20 provided good results for the modeling of damage onset at the flange tip.
doi:10.2514/6.1998-62 fatcat:pdsjnxwcsvbjvljbslmek67fxe