**1. Introduction**

With roller coaster traditional fuel prices and ever increasing energy demand, wind energy has known significant growth over the last years. To pave the way for higher efficiency and profitability of wind turbines, advances have been made in different aspects related to this technology. One of these has been the increasing size of wind turbines, thus rendering the wind blades gigantic, lighter and more flexible whilst reducing material requirements and cost. This trend towards gigantism increases risks of aeroelastic effects including dire phe‐ nomena like dynamic stall, divergence and flutter. These phenomena are the result of the combined effects of aerodynamic, inertial and elastic forces. In this chapter, we are present‐ ing a qualitative overview followed by analytical and numerical models of these phenomena and their impacts on wind turbine blades with special emphasize on Computational Fluid Dynamics (CFD) methods. As definition suggests, modeling of aeroelastic effects require the simultaneous analysis of aerodynamic solicitations of the wind flow over the blades, their dynamic behavior and the effects on the structure. Transient modeling of each of these char‐ acteristics including fluid-structure interaction requires high level computational capacities. The use of CFD codes in the preprocessing, solving and post processing of aeroelastic prob‐ lems is the most appropriate method to merge the theory with direct aeroelastic applications and achieve required accuracy. The conservation laws of fluid motion and boundary condi‐ tions used in aeroelastic modeling will be tackled from a CFD point of view. To do so, the chapter will focus on the application of finite volume methods to solve Navier-Stokes equa‐ tions with special attention to turbulence closure and boundary condition implementation. Three aeroelastic phenomena with direct application to wind turbine blades are then stud‐ ied using the proposed methods. First, dynamic stall will be used as case study to illustrate

© 2012 Ramdenee et al.; licensee InTech. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2012 Ramdenee et al.; licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

the traditional methodology of CFD aeroelastic modeling: mathematical analysis of the phe‐ nomenon, choice of software, computational domain calibration, mesh optimization and tur‐ bulence and transition model validation. An S809 airfoil will be used to illustrate the phenomenon and the obtained results will be compared to experimental ones. The diver‐ gence will be then studied both analytically and numerically to emphasize CFD capacity to model such a complex phenomenon. To illustrate divergence and related study of eigenval‐ ues, an experimental study conducted at NASA Langley will be analyzed and used for com‐ parison with our numerical modeling. In addition to domain, mesh, turbulence and transition model calibration, this case will be used to illustrate fluid-structure interaction and the way it can be tackled in numerical models. Divergence analysis requires the model‐ ing of flow parameters on one side and the inertial and structural behavior of the blade on the other side. These two models should be simultaneously solved and continuous exchange of data is essential as the fluid behavior affects the structure and vice-versa.

This chapter will conclude with one of the most dangerous and destructive aeroelastic phe‐ nomena – the flutter. Analytical models and CFD tools are applied to model flutter and the results are validated with experiments. This example is used to illustrate the application of aeroelastic modeling to predictive control. The computational requirements for accurate aer‐ oelastic modeling are so important that the calculation time is too large to be applied for real time predictive control. Hence, flutter will be used as an example to show how we can use CFD based offline results to build Laplacian based faster models that can be used for predic‐ tive control. The results of this model will be compared to experiments.
