US 12,282,720 B2
Fluid state estimation system, learning device, learning program, estimation device, and estimation program
Kazuya Furuichi, Yokohama (JP); Motoki Irikura, Yokohama (JP); Junichiro Kaji, Yokohama (JP); Tatsuro Hori, Yokohama (JP); Nobuhiro Kakuta, Yokohama (JP); and Ryo Nakayama, Yokohama (JP)
Assigned to CHIYODA CORPORATION, Yokohama (JP)
Filed by CHIYODA CORPORATION, Yokohama (JP)
Filed on Dec. 8, 2023, as Appl. No. 18/533,614.
Application 18/533,614 is a continuation of application No. PCT/JP2022/023180, filed on Jun. 8, 2022.
Claims priority of application No. PCT/JP2021/021900 (WO), filed on Jun. 9, 2021.
Prior Publication US 2024/0273267 A1, Aug. 15, 2024
Int. Cl. G06F 30/28 (2020.01); G06F 30/27 (2020.01)
CPC G06F 30/28 (2020.01) [G06F 30/27 (2020.01)] 8 Claims
OG exemplary drawing
 
1. A fluid state estimation system comprising:
a plurality of components that are connected so that the fluid can enter and exit one another in a plant;
a learning device structured to learn an estimation model for estimating a state of a fluid inside of one of the plurality of components; and
an estimation device structured to estimate the state of the fluid inside of the component of the plurality of components using the estimation model learned by the learning device and inputted to the estimation model by the learning device, wherein
the estimation model is a surrogate model that substitutes a computational fluid dynamics simulator that simulates a two-dimensional distribution or a three-dimensional distribution of one or more physical quantities of a fluid inside the component using computational fluid dynamics;
the estimation model gets a value of an input variable, and outputs a two-dimensional distribution or three-dimensional distribution of the one or more physical quantities as a fluid state information representing the state of the fluid inside of the component,
the learning device includes
a learning data acquirer that acquires, as learning data, a set of the value of the input variable and the value of the fluid state information,
the computational fluid dynamics simulator that simulates the two-dimensional distribution or the three-dimensional distribution of the one or more physical quantities of the fluid inside the component using computational fluid dynamics,
a data assimilator that assimilates a calculated value of fluid state information calculated by the computational fluid dynamics simulator and an actual measurement value of fluid state information measured when a plant is operated, and
a learner that learns the estimation model on the basis of the learning data,
the learning data acquirer acquires, as learning data, the fluid state information generated by the assimilating, by the data assimilator, the calculated value of fluid state information calculated by simulating, by the computational fluid dynamics simulator, a two-dimensional distribution or a three-dimensional distribution of the one or more physical quantities of a fluid inside the component when the input variable is a specific value and an actual measurement value of fluid state information, and
the estimation device includes
a fluid state estimator that inputs a specific value of the input variable to the estimation model to estimate a two-dimensional distribution or a three-dimensional distribution of one or more physical quantities of a fluid inside of the component when the input variable is the specific value.