US 12,435,700 B2
Wind turbine performance determination and control
Zijun Zhang, Kowloon (HK); and Luoxiao Yang, Kowloon (HK)
Assigned to City University of Hong Kong, Kowloon (HK)
Filed by City University of Hong Kong, Kowloon (HK)
Filed on Nov. 14, 2022, as Appl. No. 17/986,036.
Claims priority of provisional application 63/289,306, filed on Dec. 14, 2021.
Prior Publication US 2023/0184222 A1, Jun. 15, 2023
Int. Cl. F03D 17/00 (2016.01)
CPC F03D 17/006 (2023.08) [F05B 2270/32 (2013.01); F05B 2270/335 (2013.01); F05B 2270/709 (2013.01); G06T 2207/20084 (2013.01)] 23 Claims
OG exemplary drawing
 
1. A computer-implemented method for determining wind turbine performance and controlling operation of a wind turbine, comprising:
generating a digital image based on operation data of the wind turbine, the operation data including wind speed data and associated power output data;
processing the digital image using a data-driven model based on a convolutional neural network to obtain processed digital image; the convolutional neural network having an encoder-decoder architecture; the data-driven model which is based on the convolutional neural network being developed according to the following steps:
generating wind turbine operation data images and associated wind turbine power curve images based on a mathematical function; and
applying the wind turbine operation data images and associated wind turbine power curve images generated based on the mathematical function to the convolutional neural network to train the convolutional neural network and in turn to develop the data-driven model;
processing the processed digital image to determine a representation of a wind turbine power curve associated with operation of the wind turbine;
comparing the wind turbine power curve with a reference wind turbine power curve associated with the wind turbine; and
controlling the operation of the wind turbine by adjusting one or more operation parameters of the wind turbine based on the comparison.