Integration of OpenFAST-derived boundary conditions into ANSYS APDL for high-fidelity stress analysis based on 3D finite element models
Im, Seula (2026)
Diplomityö
Im, Seula
2026
School of Energy Systems, Energiatekniikka
Kaikki oikeudet pidätetään.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi-fe20260805115332
https://urn.fi/URN:NBN:fi-fe20260805115332
Tiivistelmä
Wind turbine blades are still maintained using preventive schedules that rely on theoretical load spectra. In practice, however, the actual loads during operation cannot be measured, and the conditions at each site differ enough to make fatigue-life predictions quite uncertain. To reduce that gap, this study uses a deformation-based approach: instead of estimating loads indirectly, the displacement fields from OpenFAST are applied directly to a 3D finite element model of the IEA 22 MW reference blade.
The transfer is done by imposing the OpenFAST bending line onto the spar cap through CERIG kinematic coupling. After verifying that the imposed deformation matches the reference behaviour, the resulting stress and strain fields are evaluated. Several coupling-station layouts were tested, changing the number of stations, their width, and their spanwise position, and the deformation accuracy was checked at four azimuth angles.
A simple layout with three stations covering roughly 10-80% of the blade turned out to be sufficient. It reproduced the flapwise bending line with a normalised RMS error of about 2.3%. The edgewise error stayed low across all layouts, between roughly 1.9% and 2.4%, and did not depend strongly on the station count. What mattered most was how far outboard the last station was placed.
The transfer is done by imposing the OpenFAST bending line onto the spar cap through CERIG kinematic coupling. After verifying that the imposed deformation matches the reference behaviour, the resulting stress and strain fields are evaluated. Several coupling-station layouts were tested, changing the number of stations, their width, and their spanwise position, and the deformation accuracy was checked at four azimuth angles.
A simple layout with three stations covering roughly 10-80% of the blade turned out to be sufficient. It reproduced the flapwise bending line with a normalised RMS error of about 2.3%. The edgewise error stayed low across all layouts, between roughly 1.9% and 2.4%, and did not depend strongly on the station count. What mattered most was how far outboard the last station was placed.
