Fig. 1 CART3 wind turbine multibody model Fig. 2 Example of mode shape animation in Virtual.Lab Motion Fig. 3 detail of the blade multibody modeling Table 2: initial values of correlation indices between numerical and experimental results MODE # EXP/NUM MAC ε 1/1 0.8565 0.0229 2/2 0.9530 0.1735 3/3 0.8603 0.2441 4/4 0.8823 0.0513 5/6 0.9393 0.0157 6/5 0.7543 0.0714 SUM 5.243 0.5789 4. SINGLE-OBJECTIVE OPTIMIZATION RESULTS Firstly, the model updating based on a single-objective optimization is performed. The problem is formulated as the minimization of the objective function in equation (8), without taking into account any weight between the two indices. Assuming that the tower properties listed in [10] are correct, the identification will focus on the structural parameters of the nacelle, the hub and the blade. For the nacelle and the hub, considered as purely rigid bodies and connected rigidly to the remaining of the structure, the parameters considered are the mass and rotational inertia values. For the blades, due to the high number of uncertainties, the stiffness of the beam elements in the flap and lag direction and the inertia properties of each body will be considered as design variable to be modified. In total, 75 structural parameters need to be updated. Due to the dimension of the problem, the optimization will be performed in two consecutive steps: a first one, where the nacelle and hub properties are updated together with the blade stiffness; in the second, the previous parameters will be frozen at their optimal values and the inertia parameters of the bodies of the blade will be considered. It must be noted that despite the blades are 3, 353
RkJQdWJsaXNoZXIy MTMzNzEzMQ==