Wind feedin ts contains values greater than nominal power

[nesnoj]

@SabineHaas As mentioned before, I tested the windpowerlib to find out whether the feedin exceeds the nominal power of a turbine. I used the following script:
https://gist.github.com/nesnoj/b275925d829cd03fce871e2e9d825490

• The max. output is 5,8% larger than nominal power in this case
• I used your example weather data weather.csv from the repo but I encountered the same behaviour with a different dataset
• Similar results with different turbines (I tested E82/3000)

Is this due to an incorrect cp curve?

Original issue: reegis/reegis#9

[nesnoj]

When using power_curve instead of power_coefficient_curve the output exceedance drops to 2,18% which exactly matches the ratio of max. power (2350.25) to the nominal power (2300).

So it seems to be a cp-related issue..

[SabineHaas]

Thank you for reporting and for the example!

• Concerning the power curves it would be interesting how manufacturers set the nominal power of a wind turbine: obviously it's not necessarily the maximum output of the power curve
• I recalculated the cp-curve of your turbine to a power curve using standard air density (see the script here):
  

As you can see the values originating from the cp curve are highly fluctuating around the highest power value in the power curve.

Thus, I'd like to recommend you to calculated with power curves for now.
However, lets keep up the discussion about the cp curves!

[nesnoj]

Thanks for your quick reply and the plot!
That's quite some fluctuation.. I'll use the power curve for now until we have better data.
Cheers

[uvchik]

Does this problem persist or did you solve it by using the power curve.

We could add something like a "force_to_max" method that will force the output to the nominal output if the wind_speed is above a specific value. But this is only needed if there is a power coefficient curve but no power curve. Actually everybody is already free to do this in a post processing.

I cannot gauge the importance of this problem.