Show simple item record

dc.contributor.advisorHale, Richard D
dc.contributor.authorVanSkike, William Preston
dc.date.accessioned2022-03-19T17:25:25Z
dc.date.available2022-03-19T17:25:25Z
dc.date.issued2020-12-31
dc.date.submitted2020
dc.identifier.otherhttp://dissertations.umi.com/ku:17442
dc.identifier.urihttp://hdl.handle.net/1808/32635
dc.description.abstractSignificant growth of the wind energy industry over the last two decades has driven rapid evolution of wind turbine technology leading to dramatic increases in wind turbine rotor size and power ratings which continue to stretch the limits for materials and blade structural design. Historically, wind turbines have been analyzed using cross-sectional analysis with large safety factors. Modern approaches to blade design incorporate 2D layered-shell finite element modeling for structural analysis but rely on simplifying assumptions to represent a turbine blade. More recently, high-fidelity modeling including non-linear analysis, solid elements, and higher order elements has been suggested in the literature, resulting in an extensive array of recommended modeling practices without a clear consensus on the appropriate modeling fidelity to analyze a blade or under what circumstances higher fidelity methods are warranted. Meanwhile, the current certification standards allow a wide range of acceptable modeling methods, with minimal details on what constitutes a suitable analysis approach. The objective of this research is to critically assess modeling techniques for structural analysis of wind turbine blades in order to identify the appropriate level of fidelity and corresponding choice of safety factors for structural analysis of wind turbine blades. Additionally, this work seeks to evaluate the variation in modeling accuracy for the range of component thicknesses along the span of a typical blade in order to identify the conditions under which lower fidelity methods are inadequate. A review of recommended modeling approaches from literature in the wind industry has led to the selection of eighteen modeling approaches considering layered-shell and various solid element types, several element formulations and plate theories, smeared and discrete ply representations, linear and nonlinear solutions, and through thickness discretization of the solid models. The reference modeling method selected to evaluate these eighteen techniques is a high-fidelity approach in which a 20-node quadratic hexahedral element is used for each individual sublayer of laminate. The high-fidelity reference method is validated against a related problem in which an exact 3D elasticity solution is available. Each of the eighteen methods are then used to analyze a simplified subcomponent representative of a typical wind turbine blade shell for stresses in the blade skin and spar cap plies in the vicinity of internal substructure and the discrete material transition to the adjacent core panels. Through-thickness strain distributions are compared against the reference model at selected locations to assess the accuracy and limitations of lower fidelity models. Three trade studies are then performed in which the thickness of the main structural components (spar cap, skins, and core) is varied corresponding to relative component thicknesses of typical inboard, midboard, and outboard blade laminates. A fourth trade study is carried out in which the overall cross-section height is varied for a fixed laminate thickness to assess the variation in accuracy of each method for a range of relative laminate-to-airfoil thicknesses. Finally, the midboard trade study is repeated to evaluate the sensitivity of each modeling approach to misalignment in spar cap fiber orientation. The accuracy of the lower fidelity methods is assessed by comparing the predicted load at which fiber failure and inter-fiber failure first occur for each model against the reference approach. Outcomes of these trade studies are used to assess the modeling error in each approach and provide recommendations for analysis of wind turbine blades and appropriate safety factors. For the shell element modeling approaches, three key limitations are identified and discussed throughout this dissertation: linear through thickness variation of in-plane strains, load application at the element midplane, and artificial stiffening due to geometric extension of the webs for mesh compatibility. For fiber failure, shell element modeling errors in the abstracted subcomponent are observed to be up to 15% for the ratios of total laminate thickness to cross-section height from a typical blade. However, the predicted impact on fiber failure for a typical blade is expected to be within 5% when accounting for the number of webs in a typical blade model. As such, the typical partial safety factor for accuracy of analysis of 1.0 is seen to be nonconservative and increasing the safety factor to 1.05 would be advised to account for the observed errors in shell models. For inter-fiber failure, modeling errors exceeding 60% are observed for the inner skin and spar cap due to complex in-plane shear stress distributions in the vicinity of mating structure. Therefore, shell element modeling is not recommended in the vicinity of mating structure or material interfaces. Global shell element models should be supplemented with solid models or subcomponent testing for these regions. Results from the solid modeling methods show several distinct advantages over shell elements, such as a more accurate geometric representation and the ability to represent a nonlinear through thickness variation of in-plane strains (by means of multiple linear elements through thickness or a higher order shape function). However, disadvantages of solid elements are also identified, including limitations in representing the out-of-plane stresses and the overall increase in model complexity. For fiber failure, all solid modeling approaches considered in this work match the reference solution within 0.5%, indicating the current partial safety factors for accuracy of analysis are suitable for solid modeling. For inter-fiber failure, modeling errors for the critical ply failure are within 10% of the reference model for all approaches, with errors decreasing as model fidelity increases. Although these errors are within the typical 1.15 accuracy of analysis partial safety factor for inter-fiber failure, modeling errors in non-critical plies are observed as high as 40% when using three or fewer elements through the thickness. As such, at least five elements are recommended through thickness when using solid elements. Furthermore, when using at least five elements through thickness, modeling errors for inter-fiber failure remain within 10% for all structural components, suggesting a possible reduction in the partial safety factors for accuracy of analysis from 1.15 to 1.10 for solid models.
dc.format.extent301 pages
dc.language.isoen
dc.publisherUniversity of Kansas
dc.rightsCopyright held by the author.
dc.subjectAerospace engineering
dc.subjectMechanical engineering
dc.subjectComposite Materials
dc.subjectFinite Element Analysis
dc.subjectModeling and Simulation
dc.subjectStructural Analysis
dc.subjectWind Energy
dc.subjectWind Turbine
dc.titleComparative Assessment of Finite Element Modeling Methods for Wind Turbine Blades
dc.typeDissertation
dc.contributor.cmtememberArnold, Emily
dc.contributor.cmtememberEwing, Mark
dc.contributor.cmtememberSorem, Robert
dc.contributor.cmtememberTaghavi, Ray
dc.thesis.degreeDisciplineAerospace Engineering
dc.thesis.degreeLevelPh.D.
dc.identifier.orcid0000-0001-5405-4352


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record