*3.2. Solution and Convergence*

A rudimentary numerical verification of the model was performed in order to check the correctness of the calculation. The tests were performed for *V* = 12 m/s and *TSR* = 5.

During the mesh independence study, the results obtained with the primary grid were compared with a reference grid, with approximately 75% more elements and nodes. Selected variables were compared between these two cases: torque in the axis of rotation (relative error δ ≈ 0.5%) and flapwise bending moment (δ < 0.1%), as well as the axial force (δ ≈ 0.6%). The relative error in the rotor power was of the order of magnitude 10<sup>−</sup>3. The mesh quality was further assessed from the point of view of boundary layer flow solution correctness. The dimensionless distance *y*+ is the standard means for this verification. For turbulence closures incorporating the k-ω model for boundary layer flow, this parameter should be kept at level *y*+ < 3 [23], guaranteeing that the model will be able to correctly depict the velocity profile in the immediate vicinity of the wall. In the considered case, the mesh was inflated from the blade surface to ensure the proper transition of the element size. The differences in *y*+ values for the two meshes turned out to be more evident, with the average value on the blade surface being approximately 10.8 for the primary and 2.2 for reference mesh. For the latter grid, the highest values of *y*+ were identified near the blade tip, thus having only very little influence on the overall rotor performance. On the other hand, for the primary mesh, the local increase in *y*+ was observed in the flow separation region at both sides of the blade, which has a chief influence on the blade performance. This flaw of the primary mesh can significantly deteriorate the model fidelity, especially at lower wind speeds. Followed by this reasoning was the choice of the reference mesh for the actual studies.

The simulation convergence was evaluated based on the attained (normalized) level of residuals. For the momentum equations, this was approximately 3 <sup>×</sup> 10<sup>−</sup>5, for the continuity equation 3 <sup>×</sup> 10<sup>−</sup>6, for turbulent quantities transport equations 7 <sup>×</sup> 10−<sup>5</sup> (k), and for 1 <sup>×</sup> 10−<sup>5</sup> (ω). These values were deemed satisfactory to consider the solution as converged.
