US 12,448,949 B2
Predicting grid frequency
Juan Fernandez Diaz, Zizur Mayor (ES); Manuel Garcia Plaza, Alcala de Henares (ES); Fernando Roque Martinez Mendoza, Burlada (ES); Maciej Marek Niegowski, Olloki (ES); Venkata Swaroop Pappala, Hamburg (DE); Sebastian Trujillo, Hamburg (DE); and Sergey Yashchenko, Hamburg (DE)
Assigned to Siemens Gamesa Renewable Energy Innovation & Technology S.L., Sarriguren (DK)
Appl. No. 18/681,899
Filed by Siemens Gamesa Renewable Energy Innovation & Technology S.L., Sarriguren (DK)
PCT Filed Aug. 9, 2022, PCT No. PCT/EP2022/072315
§ 371(c)(1), (2) Date Feb. 7, 2024,
PCT Pub. No. WO2023/002558, PCT Pub. Date Mar. 2, 2023.
Claims priority of application No. 21382776 (EP), filed on Aug. 23, 2021.
Prior Publication US 2024/0318632 A1, Sep. 26, 2024
Int. Cl. F03D 7/04 (2006.01); F03D 7/02 (2006.01); G01R 19/25 (2006.01)
CPC F03D 7/048 (2013.01) [F03D 7/0284 (2013.01); G01R 19/2513 (2013.01); F05B 2270/1033 (2013.01); F05B 2270/337 (2013.01); F05B 2270/709 (2013.01)] 14 Claims
OG exemplary drawing
 
1. A method of predicting a frequency value of a utility grid to which a wind park is connected and controlling operation of the utility grid and/or the wind park, the method comprising:
obtaining plural utility grid measurement values pertaining to a predetermined time range before and until a present point in time;
obtaining plural wind park measurement values pertaining to the predetermined time range;
feeding the plural utility grid measurement values and the plural wind park measurement values into a recurrent neural network trained to output a predicted frequency value at at least one next point in time, the next point in time being between 0.5 s and 2 s after the present point in time, wherein the plural utility grid measurement values contain a, across the time range, time series of values of the grid frequency as measured at one or more locations and/or a grid voltage as measured at the one or more locations; and
controlling operation of the utility grid and/or the wind park using the predicted frequency value.