*3.3. Fluid Flow Simulation Results*

The first aspect of the performed results postprocessing was an analysis of the power output. In Figure 4 simulation (squares) and experiment (circles) results are compared for the same working points (*V* and *TSR*). A systemic error is noticeable, with the numerical results being approximately 120–140 W lower than experimental ones. While the current study is mostly dedicated to FSI investigation, these discrepancies may result in underestimation of actual aerodynamic loads and need addressing, for example, as a guideline for future works.

**Figure 4.** Analyzed cases and power obtained at different wind speeds and Tip-Speed Ratio *(TSR)*.

An important factor influencing the results are the simplifications of the numerical model. Firstly, the simulation performed in a steady-state mode may not entirely adequately depict the complex helical tip vortex wake structure, leading to modification in the induced velocities. This is further altered by the frozen rotor scheme–the most closely depicting the real flow, yet still simplifying it. For the full resolution of these flow phenomena, an unsteady model would be required. However, this was deemed prohibitively expensive in terms of computational resources and will be examined in the future, possibly with a two-way FSI. Another sources of error may be the size of the simulation domains (especially the one encompassing the rotor), and the fact that just 1/3 of the problem is being resolved. While this was done according to the standard procedures, these elements may influence the pressure fields, especially in the places where high gradients occur. The influence of this aspect was previously observed, also for rotor actuator models [22] and may be also connected with data transmission through interfaces.

In all, the obtained results may lead to a conclusion that the numerical model has the tendency to underestimate the aerodynamic loads. However, the performed structural assessment (Chapter 4) shows that the estimated stresses are well within the material flexural strength limit, and the deformations are very small, so even if a safety factor is taken into account, these observations will remain valid.

On a more general note, it is also visible how the flow is influenced by the relatively low Reynolds number phenomena: at lower wind speeds the rotor needs to turn at higher *TSR* in order to remain close to performance peaks. This becomes less evident as the wind speed increases, at which point, in experimental investigations, aeroelastic effects may start to play an important role.

In addition to cases mimicking experimental results, the simulation was also performed for design *TSR* = 5 at *V* = 8.4 m/s (rated wind speed) and 12 m/s (diamonds in Figure 4). This was done to assess the loads acting on the blade and provide data for further FSI simulation. The pressure fields on the blade are visible in Figure 5, providing input data for mechanical analysis. Special interest must be paid to the tip region, as in there the visibly high-pressure gradients between the two sides of the blade contribute to increased mechanical loads.

**Figure 5.** Pressure coefficient distribution at blade pressure (top) and suction (bottom) side (*V* = 12 m/s, *TSR* = 5).

Lastly, a general overview of the flow around the blade is visible in Figure 6. No boundary layer separation is visible whatsoever, suggesting that the wind turbine operates in pre-stall conditions. Quite visibly, the flow speed increases along with radius and it is fair to say that near the blade tip it is almost 3 times higher than at the bottom.

**Figure 6.** Normalized velocity (*Vref* = 12 m/s, *TSR* = 5) contours and vectors (left) and pressure coefficient contours (right) at *r*/*R* equal to (**a**) 0.25, (**b**) 0.5, (**c**) 0.75, and (**d**) 0.9875.
