A deep learning-based automatic slag removal control method for molten iron

Abstract

The invention proposes a deep learning-based automatic slag scraping control method for molten iron. In the present invention, the collected liquid level images of the ladle are extracted, and the slag level is obtained by manual marking, which is used to construct the data set of the deep convolutional neural network; The model is trained and optimized, and the optimized network model is obtained; the factor types are constructed by the factors that affect the molten iron desulfurization and slag, and the factor categories to which the slag level standard belongs are selected by the k-nearest neighbor method, and the depth volume of different factor categories is obtained according to steps 1 to 3. The integrated neural network model; the overall slag level is determined according to the output layer data of the corresponding network model, and the corresponding slag removal action is determined by the system according to the model output. The invention can fully mine on-site image information data and calculate the optimal slag removal process, and the model has strong robustness, can effectively shorten the slag removal time, reduce the waste of molten iron, and improve the production efficiency of the enterprise.

Description

technical field [0001] The invention relates to the technical field of molten iron desulfurization and slag skimming, in particular to a deep learning-based automatic slag skimming control method for molten iron. Background technique [0002] As steel mills pay more and more attention to energy conservation and emission reduction, they need to be efficient and reduce losses in each process section. In the process of hot metal desulfurization and slag removal, there are still hot metal waste and low efficiency. Reducing waste and improving efficiency has become the development goal of all enterprises. In this process, the most important link is the judgment of slag volume grade, slag thickness and slag profile, which determines the final slag removal time and slag removal accuracy. The slag scavenging time and slag scavenging precision of desulfurization and slag scavenging determine the amount and efficiency of molten iron waste in this process section. Therefore, the accur...

AI Technical Summary

Problems solved by technology

The core technology of molten iron slag removal is liquid level image processing and slag volume decision-making. Deep removal will bring out a large amount of molten iron, resulting in waste of molten iron. Less removal will cause sulfur return during converter steelmaking and affect the quality of molten steel.

Images