System and method for assessing farm-level performance of a wind farm

1. A method for controlling a wind farm, the method comprising:

operating the wind farm in a first operational mode for a predetermined time period;

identifying one or more pairs of wind turbines having wake interaction during the first operational mode;

generating a pairwise dataset for the one or more pairs of wind turbines;

generating a first wake model based on the pairwise dataset;

operating the wind farm in a second operational mode and collecting operational data during the second operational mode, the second operational mode being operated at a different time period than the first operational mode;

predicting a first farm-level performance parameter for the second operational mode using the first wake model and the operational data collected during the second operational mode;

determining a second farm-level performance parameter during the second operational mode;

assessing farm-level performance of the wind farm by comparing the first and second farm-level performance parameters;

selecting to operate the wind farm in the first operational mode or the second operational mode based on the comparison without toggling between the first and second operational modes; and,

controlling the wind farm based on the selected mode.

2. The method of claim 1 , wherein the first and second farm-level performance parameters comprise a power output for the first and second operational modes, respectively.

3. The method of claim 1 , wherein determining the second farm-level performance parameter further comprises:

collecting operational data from one or more wind turbines in the wind farm during the second operational mode,

determining a power output for each of the wind turbines in the wind farm based on the collected operational data, and

summing the power outputs from each of the wind turbines in the wind farm during the second operational mode to determine a total second power output.

4. The method of claim 3 , wherein predicting the first farm-level performance parameter further comprises:

collecting operational data from one or more wind turbines in the wind farm during the first operational mode, and

building the first wake model based on the collected operational data from the one or more wind turbines to predict the power at one or more downstream turbines.

5. The method of claim 4 , further comprising applying the first wake model to predict power output for the one or more downstream wind turbines during the second operational mode,

summing the power output from the one or more downstream wind turbines, and

determining a total first power output by summing a power output from each of the remaining wind turbines in the wind farm during the second operational mode with the summed power output from the one or more downstream wind turbines.

6. The method of claim 1 , wherein generating the first wake model based on the pairwise datasets further comprises generating the first wake model using Gaussian Kernel Regression.

7. The method of claim 1 , further comprising tuning the first wake model to reduce a prediction error of the first wake model.

8. The method of claim 5 , wherein assessing farm-level performance of the wind farm as a function of the first and second farm-level performance parameters further comprises comparing the total second power output and the total first power output.

9. The method of claim 1 , wherein assessing farm-level performance of the wind farm as a function of the first and second farm-level performance parameters further comprises:

determining a power difference between the total second power output and the total first power output over a predetermined time period,

binning the power difference by farm inflow wind speed and wind direction, and

weighting the binned power differences by a wind distribution.

10. The method of claim 1 , wherein assessing farm-level performance of the wind farm as a function of the first and second farm-level performance parameters further comprises:

generating a second wake model for the second operational mode,

estimating a first power output and a second power output from the first and second wake models, and

comparing the first power output and the second power output so as to verify the farm-level performance.

11. The method of claim 1 , wherein the pairwise dataset for each of the pairs comprises at least one of a relative wind direction between the pairs, the distance between the pair, a turbulence intensity, a power at an upstream wind turbine, a power at a downstream wind turbine, and a power at a corresponding freestream turbine.

12. The method of claim 1 , wherein the operational data comprises at least one of or a combination of the following: nacelle position, power output, torque output, pitch angle, tip speed ratio, rotor speed, yaw angle, thrust, operating state, curtailment demands, geographical information, temperature, pressure, wind turbine location, wind farm location, weather conditions, wind gusts, wind speed, wind direction, wind acceleration, wind turbulence, wind shear, wind veer, or wake.

13. A system for controlling a wind farm, the system comprising:

one or more sensors configured to measure operational data of the wind farm; and

one or more controllers communicatively coupled with the one or more sensors, the one or more controllers configured to perform one or more operations, the one or more operations comprising:

operating the wind farm in a first operational mode for a predetermined time period,

identifying one or more pairs of wind turbines having wake interaction during the first operational mode,

generating a pairwise dataset for the one or more pairs of wind turbines,

generating a wake model based on the pairwise dataset,

operating the wind farm in a second operational mode and collecting operational data during the second operational mode, the second operational mode being operated at a different time period than the first operational mode;

predicting a first farm-level performance parameter for the second operational mode using the first wake model and the operational data collected during the second operational mode;

determining a second farm-level performance parameter during the second operational mode, and

assessing farm-level performance of the wind by comparing the first and second farm-level performance parameters;

selecting to operate the wind farm in the first operational mode or the second operational mode based on the comparison without toggling between the first and second operational modes; and,

controlling the wind farm based on the selected mode.

14. The system of claim 13 , wherein determining the second farm-level performance parameter further comprises:

collecting operational data from one or more wind turbines in the wind farm during the second operational mode,

determining a power output for each of the wind turbines in the wind farm based on the collected operational data, and

summing the power outputs from each of the wind turbines in the wind farm during the second operational mode to determine a total second power output.

15. The system of claim 13 , wherein predicting the first farm-level performance parameter further comprises:

collecting operational data from one or more wind turbines in the wind farm during the first operational mode,

building the first wake model based on the collected operational data from the one or more wind turbines to predict the power at one or more downstream turbines,

applying the first wake model to predict power output for the one or more downstream wind turbines during the second operational mode,

summing the power output from the one or more downstream wind turbines, and

determining a total first power output by summing a power output from each of the remaining wind turbines in the wind farm during the second operational mode with the summed power output from the one or more downstream wind turbines.

16. The system of claim 15 , wherein assessing farm-level performance of the wind farm as a function of the first and second farm-level performance parameters further comprises comparing the total first and second power outputs.

17. The system of claim 13 , wherein assessing farm-level performance of the wind farm as a function of the first and second farm-level performance parameters further comprises:

determining a power difference between the total second power output and the total first power output over a predetermined time period,

binning the power difference by farm inflow wind speed and wind direction, and

weighting the binned power differences by a wind distribution.

18. The system of claim 13 , wherein assessing farm-level performance of the wind farm as a function of the first and second farm-level performance parameters further comprises:

generating a second wake model for the second operational mode,

estimating a first power output and a second power output from the first and second wake models, and

comparing the first power output and the second power output so as to verify the farm-level performance.