Utility-scale wind turbines have various sensors mounted on the nacelle, to ensure the power generation is optimised.

How are readings from the wind sensors mounted on wind turbines influenced by the blades' wakes?

And what sort of compensation is done in best practice to adjust for this?

I have some knowledge in this field, coming from several years' experience in drafting and defending patent applications that deal with wind turbines in one form or other. If you're talking about a single wind turbine with a sensor somewhere on the nacelle, downwind of the rotor blades: one way of dealing with the problem of wake-corrupted measurements is to place wind speed and pressure sensors in the spinner or hub, or out in front of the spinner (for example a sonic wind speed sensor on a boom extending outward from the hub). This placement of the sensors can at least ensure that the measured values are not falsified by turbulence caused by the rotor blades. The pressure sensors are used to determine the wind direction. Accurate measurements are important because these are needed to precisely control the yaw drive that turns the hub to face into the wind.

The problem gets more complicated in the case of a wind farm. Here, the turbines in the current "front row" receive the most reliable information from their wind speed and wind pressure sensors. The wind direction information collected from the pressure sensors of the front row turbines can generally be used to also control the yaw drives of the downstream wind turbines. However, the wind speed can be significantly different for the downstream turbines. Some wind parks use sophisticated wake modelling algorithms to continually update the control setpoints for each turbine in order to maximise the park power output.

Hope that helps...

  • 1
    Out of curiosity... in the case of a wind farm, wouldn't it be simplest to use an algorithm similar to what's done for MPPT - move the rotational speed up or down a bit, and see if you get more power out? I imagine there are a number of interrelated variables to track, but once your algorithm can account for those, it seems like you could just employ a "perturb and observe" sort of power tracking. – LShaver Nov 7 '16 at 23:31
  • @LShaver that's a really good question! Please do post it as a new question! – EnergyNumbers Nov 9 '16 at 8:02

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.