Multi-modal identification method for annealing heating process based on D-S evidence theory

Abstract

The invention provides a multi-modal identification method for an annealing heating process based on a D-S evidence theory. The multi-modal identification method comprises the following steps: determining a production operation modal and a key variable when strip steel is heated in a continuous annealing furnace according to a cold-rolled strip steel continuous annealing production process; determining data distribution characteristics of the key variables, and establishing a state detection model of the key variables; based on the D-S evidence theory, adopting a state detection model to carryout state detection on the key variables; and according to the state and a multi-modal identification strategy, identifying the modal of the cold-rolled strip steel in the continuous annealing and heating process. According to the technical scheme provided by the invention, a computer is used for replacing central control personnel to identify the production state of the strip steel during annealing and heating, so that manpower is reduced, and errors caused by subjectivity of manual identification are reduced; the D-S evidence theory is adopted to solve the limitation caused by a confidenceinterval method threshold, and the generalization ability of model application and the detection accuracy of the actual state of a key variable are improved.

Description

technical field [0001] The invention relates to the technical field of process control, in particular to a multimodal identification method for annealing heating process based on D-S evidence theory. Background technique [0002] With the increasingly fierce homogenization competition of my country's iron and steel enterprises in recent years, the production of high-quality steel has received more and more attention. Among them, cold-rolled strip steel is one of the representatives of high-quality steel products. Continuous annealing process (CAP), that is, the process of heating strip steel in various combinations in a continuous strip annealing furnace, is the most critical process in a cold-rolled strip production line, directly related to the quality of the final strip product. [0003] In order to improve the quality of strip steel, it is necessary to ensure the stability, uniformity and consistency of CAP, so it is necessary to study the modeling, control and optimiza...

AI Technical Summary

Problems solved by technology

[0006] For each key variable, if a single-scale feature such as a trained limit value, threshold or window length is directly used to build a detection model, then the detection model has poor generalization ability and strong limitations

Because the state of key variables in the actual environment is complex and changeable, the detection information of single-scale features has great limitations (it cannot be fully applied to all target production modal information)

Taking plate temperature as an example, when using the same algorithm model with different window lengths or different standard interval boundaries for the production data of the same time period, the detection results are not all the same. It can be seen that selecting different scale features (such as window length etc.) The established training model may have a difference in the discrimination result of the parameter state, which will eventually lead to misjudgment of the production mode

Images