Wind Turbine Blade Design & Optimization

CAEbridge is assisting with the development of small to mid size wind turbines via CFD simulations. Our initial projects involve the optimization of the turbine rotor blades. Projects to follow will involve the optimal positioning of individual turbines and wind farms in urban, rural and offshore areas.

As a benchmark case, CAEbridge picked a 3 blade design with a rotor diameter of 100m.  Typically these turbines reach their peak power output of 2.5 MegaWatts at a wind speed of 12.5m/s. The rotation rate is kept at about 12 rpm corresponding to a tip speed ratio of 4-5, which is optimal for a 3 blade design. Another parameter affecting the turbine performance is the incoming flow turbulence intensity which is typically around 13% of the mean flow energy.

Our particular turbine geometry was obtained from an online 3D mesh geometry database. While the blade airfoil shapes were quite good qualitatively, their pitch angles were completely off from achieving optimal lift at the above operating conditions. So we developed a computed code that estimates the amount of rotation and twist that needs to be applied to the blades in order to maximize its power output. The very same computer code can be utilized to optimize other parameters such as forces or moments on the overall system.

Once the ideal pitch and twist angles were determined and applied, we built a 1.5 million cell volume grid around the turbine. In this process we were limited by the 32bit computer memory limit; ideally one needs to go up to 3-4 times the number of cells in order to resolve the airfoil boundary layers appropriately. We will address this issue in the next set of runs.

Accounting for the rotation of the blades, we utilized an MRF (Moving Reference Frame) CFD model solving the steady-state (time average) turbulent flow around the turbine. It took about 1.5 days to perform 1000 iterations  on a dual core machine to obtain a reasonably stabilized solution.

Preliminary analysis of the CFD data has shown very reasonable flow structures that correlate well with the reality. Formation of strong helical vortices were detected which are typically visualized by releasing smoke from the blades in experiments. Due to the utilization of the MRF model, the velocity field had to be transformed to the blade reference to detect whether stall occurred through the length of the blades. We did not detect any but the grid resolution may have been too coarse to trigger this. On the other hand we had configured our blade optimization tool to not violate the stall limit at any point.

Having achieved good suction on the aft surface of the blades, we computed the overall torque of the blades and the power generated. The configuration turned out to produce 1.8 MegaWatts which is reasonable but a bit short off the 2.5MegaWatt rating. We believe the intended target can be reached by better grid resolution and a bit more aggressive tuning of the blades.

Blade Pressure and Velocity Vectors at Half Radius

One concern for wind turbines is their environmental impact. In particular, the strong circulation generated in the wake of these machines an affect wildlife adversely. Even with our coarse model, we were able to capture the strong swirl effects and their generation mechanism tied up to the vortex roll up at the tips of the blades. In aviation, the wing tip vortices are suppressed by proper design and installation of winglets, which also benefit the wing lift. Similar ideas should be utilized in wind turbine applications to reduce environmental impact and improve efficiency.