Copyright C: This paper is declared a work of the U.S. Government and is not subject to copyright protection in the United States SAMPE 2010, Seattle, Paper 398 TESTING AND SIMULATION OF DAMAGE GROWTH AT PLY DROPS IN WIND TURBINE BLADE LAMINATES Pancasatya Agastra and John F. Mandell Department of Chemical and Biological Engineering Montana State University – Bozeman Bozeman, MT 59715 ABSTRACT Wind turbine blades are typified by thick, resin-infused laminates containing ply drops for thickness tapering. A complex structured test coupon has been developed to represent a simplified version of the substructure around ply drops; the coupon is relatively easy to fabricate and test, including tension, reversed and compression loading in fatigue [1,2]. This coupon is used to compare the performance of five types of resin, and several fabrics and ply drop geometries. The effects of these variables on damage growth are quantified for static and fatigue loading, showing significant effects of the resin. Resin types included in this study are polyester (UP), toughened and un-toughened vinyl ester (VE), epoxy (EP) and polydicyclopentadiene (pDCPD). The strain energy release rates for delamination under pure Modes I and II, and mixed-mode have also been determined using established test methods, as have the in-plane static and fatigue properties. Based on the geometry and in-plane and interlaminar properties, a simulation of the damage growth under static loading has been developed using ANSYS finite element modeling. Simulation results are consistent with the experimental damage growth results, and are helpful in relating the performance back to more fundamental properties such as the pure and mixed mode delamination tests, and, in turn, to resin properties. The simulation study also identifies parameters to which ply drop delamination resistance is sensitive, including: variations in elastic constants related to fiber content variations, assumed and actual boundary conditions, ply resin cracking in biax plies and representation of pure and mixed mode delamination data. 1. INTRODUCTION Wind turbine blades are complex composite structures containing a variety of structural details which can compromise blade performance. Local material properties are affected by variations in fiber content and alignment. Geometric features and material discontinuities as in materials transition areas complicate local stress fields, often introducing significant third-dimensional stresses which can lead to ply delamination. Resin selection for blades has generally been focused on performance in relatively simple laminate coupons with reinforcing fabrics and process parameters representative of blade manufacturing by resin infusion. The approach used in this program has been to develop a test coupon which is representative of the more complex structural details in blade construction, where resin dominated failures occur in service, andwhich is still relatively inexpensive to fabricate and test compared to full blade substructure tests [1,2]. This allows the evaluation of multiple resin options in a realistic structural context. A ply drop is a structural detail integrated into the thick laminates of wind turbine blades to provide thickness tapering. Stress concentration arising at this structural detail can lead to ply delamination and loss of structural integrity [3,4]. Ply drops have been the subject of many studies in aerospace composite applications [5-6] as well as wind turbine blades [3, 7-9]. Standard tests cited later give the resistance to ply delamination, GIc and GIIc for pure opening and shearing modes, which are a strong function of the resin toughness [1, 10]. In a composite structure the behavior is more complex than for simple pure mode delamination tests, with the controlling strain energy release rates following mixed opening and shearing modes, and depending on geometry and damage development such as matrix cracking in off-axis plies [1, 3]. While results from the test coupon can be interpreted directly in terms of knockdowns on allowable strains, finite element modeling is required to relate the structural response to more basic in-plane and interlaminar properties, which is the major objective of this paper. Several variables affecting delamination at ply drops for thick fiberglass laminates have been explored in the experimental part of this work: resin and fabric types, ply drop thickness (number of individual plies dropped at a single position), applied load level, and damage growth under tensile, compressive, and reversed-loading in fatigue. Taken together, these represent a broad range of parameters commonly encountered in the wind turbine blade application [1]. Finite element modeling has been used both to design the test coupon and to simulate damage growth. Coupon design, reported in more detail in References 1 and 2, was focused on minimizing strain gradients across the thickness due to bending of the non-symmetric geometry. As indicated in Figure 1 for a case with two ply drops (total of 2.6 mm dropped unidirectional material), strains across the thickness on the thin side vary by less than 10% when grip lateral movement is effectively constrained. The low strain variation across the thickness allows the test results to be incorporated as a design strain knockdown, rather than requiring a fracture mechanics based design strategy.Figure 1. Axial strain distribution (top), and line plots across thickness at indicated axial locations from FEA for a tensile force of 44.5 kN [1, 2]. Simulation of damage growth based on experimental observations of damage geometry and basic materials properties has been carried out for static loading, with fatigue simulations to follow. Validated simulations are a key link in relating more basic material properties and geometric features to damage growth, and have been used successfully for delamination problems in composites [11]. 2. EXPERIMENTATION 2.1 Materials Laminates representing the separate components of the complex coupons were tested in addition to the complex coupon laminates containing ply drops. Three types of laminates were fabricated and tested: 1) Standard laminates. Unidirectional and biaxial laminates were tested to obtain ply input data and as baseline fatigue cases for comparison to the complex coupon results. 2) Delamination laminates. Unidirectional laminates with simulated starter cracks were used in