System and method for optimizing wind farm performance using unmanned aerial vehicles
A method for optimizing performance of a wind farm having at least one wind turbine includes maneuvering a first unmanned aerial vehicle (UAV) having at least one sensor to a first location near the at least one wind turbine of the wind farm; collecting, via the at least one sensor of the first UAV, data corresponding to one or more wind conditions at the at least one wind turbine; receiving the data corresponding to the one or more wind conditions at the at least one wind turbine via a controller; generating a control action for the at least one wind turbine using the data corresponding to the one or more wind condition at the at least one wind turbine; and implementing, via the controller, the control action.
1 . A method for optimizing performance of a wind farm having at least one wind turbine, the method comprising:
maneuvering a first unmanned aerial vehicle (UAV) having at least one sensor to a first location near the at least one wind turbine of the wind farm;
collecting, via the at least one sensor of the first UAV, data corresponding to one or more wind conditions at the at least one wind turbine;
receiving the data corresponding to the one or more wind conditions at the at least one wind turbine via a controller;
generating, via a machine learning algorithm programmed in the controller, a control action for the at least one wind turbine using the data corresponding to the one or more wind condition at the at least one wind turbine;
implementing, via the controller, the control action;
generating, via the machine learning algorithm, a prioritized list of desired data that instructs a second location to retrieve the data, wherein the second location comprises an area with the greatest uncertainty and highest value for use in improving the machine learning algorithm; and
maneuvering the first UAV to the second location.
2 . The method of claim 1 , wherein the one or more wind conditions comprises at least one of wind speed, wind shear, wind direction, wake, turbulence intensity, temperature, humidity, or air density.
3 . The method of claim 1 , wherein the wind farm comprises a plurality of wind turbines, the plurality of wind turbines comprising the at least one wind turbine,
wherein the control action further comprises at least one of pitching one or more rotor blades of one or more of the plurality of wind turbines, yawing one or more of the plurality of wind turbines, modifying a speed of one or more of the plurality of wind turbines, derating power on one or more of the plurality of wind turbines, or uprating power on one or more of the plurality of wind turbines.
4 . The method of claim 3 , wherein the at least one wind turbine is a first wind turbine, wherein the plurality of wind turbines further comprises a second wind turbine,
wherein the control action further comprises implementing the control action on the first wind turbine so as to cause the second wind turbine to be positioned outside of wake of the first wind turbine.
5 . The method of claim 1 , wherein the controller is further configured to generate a power curve or other model of power performance versus wind conditions for the at least one wind turbine.
6 . The method of claim 1 , further comprising maneuvering the first UAV to a second location around the at least one wind turbine different from the first location.
7 . The method of claim 2 , further comprising continuously sending the data to an archival repository, turbine, or farm level controller such that the data can be used for turbine/farm control as guided by the learned machine learning algorithm, wherein the data is configured to be used for offline training or continuous updating or optimization of the machine learning algorithm using the data.
8 . The method of claim 1 , wherein the machine learning algorithm is further configured to generate a simulated model of the wind farm for estimating or predicting performance of the wind farm or the at least one wind turbine.
9 . The method of claim 1 , further comprising transmitting the data to the controller in real-time.
10 . The method of claim 1 further comprising:
maneuvering a second UAV around the at least one wind turbine; and
collecting additional data relating to the wind farm using a sensor mounted to the second UAV.
11 . The method of claim 1 , wherein the first UAV further comprises an emitter and a receiver mounted thereto, wherein the method further comprises:
emitting, via the emitter, an acoustic signal;
receiving the acoustic signal via the acoustic receiver; and
determining the one or more wind conditions based on changes in the acoustic signal as the acoustic signal moves from the emitter to the receiver.
12 . The method of claim 1 , wherein the method further comprises collecting, via one or more fixed sensors of the wind turbine, data corresponding to one or more wind conditions at the at least one wind turbine, wherein the one or more fixed sensors comprise at least one of LIDAR, sonar, or sensors capable of detecting wind speed, temperature, pressure, humidity, air density, or power generated by the at least one wind turbine.
13 . A system for optimizing performance of a wind farm having at least one wind turbine, the system comprising:
the at least one wind turbine of the wind farm;
a first unmanned aerial vehicle (UAV) having at least one sensor, the sensor configured to collect data corresponding to one or more wind conditions at the at least one wind turbine;
a controller configured to receive the data corresponding to one or more wind conditions at the at least one wind turbine;
wherein the controller generates, via a machine learning algorithm programmed in the controller, a control action for the at least one wind turbine using the data corresponding to the one or more wind condition at the at least one wind turbine, wherein the control action is implemented via the controller,
wherein the controller generates, via the machine learning algorithm, a prioritized list of desired data that instructs a location to retrieve the data, the location comprises an area with the greatest uncertainty and highest value for use in improving the machine learning algorithm, wherein the first UAV is configured to maneuver to the location.
14 . The system of claim 13 , wherein the wind farm comprises a plurality of wind turbines, the plurality of wind turbines comprising the at least one wind turbine,
wherein the control action further comprises at least one of pitching one or more rotor blades of one or more of the plurality of wind turbines, yawing one or more of the plurality of wind turbines, modifying a speed of one or more of the plurality of wind turbines, derating power on one or more of the plurality of wind turbines, or uprating power on one or more of the plurality of wind turbines.
15 . The system of claim 14 , wherein the at least one wind turbine is a first wind turbine, wherein the plurality of wind turbines further comprises a second wind turbine,
wherein the control action further comprises implementing the control action on the first wind turbine so as cause the second wind turbine to be positioned outside of wake of the first wind turbine.
16 . The system of claim 13 , wherein the system comprises one or more fixed sensors provided on the at least one wind turbine, the one or more fixed sensors configured to collect data corresponding to one or more wind conditions at the at least one wind turbine.