US 12,013,119 B2
Method and assembly for controlling an internal combustion engine having multiple burners
Hans-Gerd Brummel, Berlin (DE); Uwe Pfeifer, Berlin (DE); and Volkmar Sterzing, Neubiberg (DE)
Assigned to SIEMENS ENERGY GLOBAL GMBH & CO. KG, Munich (DE)
Appl. No. 17/602,398
Filed by SIEMENS ENERGY GLOBAL GMBH & CO. KG, Munich (DE)
PCT Filed Mar. 19, 2020, PCT No. PCT/EP2020/057579
§ 371(c)(1), (2) Date Oct. 8, 2021,
PCT Pub. No. WO2020/212067, PCT Pub. Date Oct. 22, 2020.
Claims priority of application No. 19169491 (EP), filed on Apr. 16, 2019.
Prior Publication US 2022/0252264 A1, Aug. 11, 2022
Int. Cl. F23N 5/00 (2006.01); F02D 41/14 (2006.01)
CPC F23N 5/003 (2013.01) [F02D 41/1405 (2013.01); F23N 2223/04 (2020.01); F23N 2223/40 (2020.01); F23N 2223/48 (2020.01); F23N 2237/02 (2020.01); F23N 2241/20 (2020.01); F23N 2900/05006 (2013.01)] 13 Claims
OG exemplary drawing
 
1. A method for controlling a combustion engine with multiple burners, the method comprising:
a) recording burner-specific combustion measurement data for a respective burner, each respective burner being assigned a burner identifier that identifies the respective burner;
b) recording performance measurement data of the combustion engine which is used to determine a performance value;
c) training a machine learning model using the burner-specific combustion measurement data, the assigned burner identifiers and the performance measurement data, to generate burner-specific control data that optimize the performance value when the multiple burners are actuated using the burner-specific control data; and
d) outputting the burner-specific control data generated by the trained machine learning model to actuate the multiple burners;
wherein the machine learning model comprises multiple submodels which are specific to respective burners and which each generate the burner-specific control data for the respective burners.
 
13. A computer program product, comprising a computer readable hardware storage device having computer readable program code stored therein, said program code executable by a processor of a computer system to implement a method comprising:
a) recording burner-specific combustion measurement data for a respective burner, each respective burner being assigned a burner identifier that identifies the respective burner;
b) recording performance measurement data of the combustion engine which is used to determine a performance value;
c) training a machine learning model using the burner-specific combustion measurement data, the assigned burner identifiers and the performance measurement data, to generate burner-specific control data that optimize the performance value when the multiple burners are actuated using the burner-specific control data; and
d) outputting the burner-specific control data generated by the trained machine learning model to actuate the multiple burners;
wherein the machine learning model comprises multiple submodels which are specific to respective burners and which each generate the burner-specific control data for the respective burners.