PSI - Issue 8

F. Cianetti et al. / Procedia Structural Integrity 8 (2018) 56–66 Author name / Structural IntegrityProcedia 00 (2017) 000 – 000 principal model axis are), normalized with respect to the modal d isplacement of the free end ∅ . That is, for the i th node, the following applies: ∅ = ∅ ∅ ⁄ . The matrix instead represents a matrix, function of z (vertical) axis coordinates of the various nodes, normalized with respect to component L full length, ̅ = ⁄ , and of size w ×5 . For the generic node i , the corresponding row of the matrix is represented by the following line: = [ 2 3 4 5 6 ] (13) 63 8

Fig. 6. FE model of WindTurbinest ructure (tower, nacelleand support ing t ransverse beam). Shaped and wireframe representat ions

4. Modal model tuning

To verify that the FE model was correctly realized and eventually to tune it on the experimental behavior , a modal identification (Rao (1990), Me irovitch (2010)) was done on the wind turbine by experimental tests (hammer load condition and accelerometric measures). A first comparison was realized between the experimental results and those obtained by a modal superposition harmon ic analysis realized by a consolidated state-space modeling

0.04

Exp. Hammer Num. State-Space Hammer Num. MBS Wind Shot

0.035

0.03

0.01 Output FFT Amplitude [m/s 2 ] 0.015 0.02 0.025

0.005

10 1

10 2

Frequency [Hz]

Fig. 7. Modal ident ificat ion. Result s comparison among experimental and numerical models (state -space and mult ibody)