The application discloses a kind of gas
powder slurry collaborative control method based on multi-
source data fusion, it is related to industrial
process control technical field, it is difficult to solve the technical problem of depth fusion visual image and physical sensor data to carry out gas-
powder-
slurry three-phase collaborative optimization control;Including the following steps: collecting the visual image and physical sensor data of gas-
powder-
slurry three-phase mixed state;With visual image as input, build visual evaluation model, output splash degree
score value, dispersion uniformity
score value and
wetting integrity
score value;Build proportion
decision model, with the fusion
feature vector of fusion quality evaluation index and physical sensor data as the input of multilayer
perception machine, and output gas flow correction amount, powder rate correction amount and
slurry flow correction amount in parallel, when three-way correction amount adjustment direction conflict, with gas
carrying capacity constraint and slurry
absorption saturation constraint as arbitration basis;The correction amount is output to corresponding
actuator to form
closed loop control.