A method for reconstructing a thermal field of a furnace combustor layer based on model explainability

Abstract

The present application relates to the technical field of industrial boiler combustion monitoring and thermal field reconstruction, and discloses a furnace burner layer thermal field reconstruction method based on model interpretability. The method obtains boiler operating parameters, generates the temperature field distribution of the furnace burner layer by using a temperature field prediction model, inputs the temperature field into an interpretation model to obtain corresponding target physical quantity prediction values, and further optimizes the temperature field prediction model based on the deviation between the prediction values and actual measurement values, so that the prediction results meet preset physical constraints, thereby obtaining the reconstructed thermal field. Compared with the prior art, the present application combines a data-driven model with a physical constraint mechanism, realizes fast temperature field prediction, improves the physical consistency and reliability of the prediction results, effectively solves the problems that the traditional temperature measurement method is difficult to realize full-field monitoring, and the numerical simulation method has high calculation cost and poor real-time performance, and is suitable for boiler combustion optimization control and safe operation.

Images