Method for forecasting splashing in argon oxygen refined low-carbon ferrochromium production process

Abstract

The invention discloses a method for forecasting splashing in the argon oxygen refined low-carbon ferrochromium production process. Regarding the argon oxygen refined low-carbon ferrochromium production process, splashing generated in the smelting process is used as a research object, on the basis of deeply analyzing the occurrence mechanism of splashing in the smelting process, vibration signals, audio signals and flame image signals inside a furnace in the splashing process are used as characteristic signals for representing splashing, a multi-sensor information fusion technology is utilized to conduct fusion processing on the vibration signals, the audio signals and the flame image signals inside the furnace, and then splashing signals are reconstructed. According to the method, the occurrence of splashing in the argon oxygen refined low-carbon ferrochromium production process is accurately forecasted, the yield of alloy is improved, and production safety is guaranteed.

Description

technical field [0001] The invention relates to a method for predicting splashing in the production process of argon-oxygen refining low-carbon ferrochrome. Background technique [0002] Cause Analysis of Sputtering in Argon Oxygen Refining Production Process [0003] Analyzing the cause of spattering in argon-oxygen refining, it is concluded from the internal chemical reaction mechanism and reaction characteristics that due to the unbalanced carbon-oxygen reaction in the molten pool, a large amount of CO gas is generated instantaneously. When the CO gas volume accumulates to a large amount of energy, it will carry A certain amount of molten iron and slag is ejected from the furnace mouth to produce splashing. In the process of argon-oxygen refining and steelmaking, decarburization mainly focuses on the following two reactions, the direct reaction of carbon and supplied oxygen 1: C+O→CO; the indirect reaction of carbon and ferrous oxide 2: C+FeO=CO+ Fe, the f...

AI Technical Summary

Problems solved by technology

[0005] The gas analysis method proposed by the China Iron and Steel Research Institute provides a way to solve the splashing problem, and has been applied in Pohang, South Korea and Nippon Steel, Japan, and the prediction accuracy has reached 81%. Poor, currently it is only used to analyze the cause of spattering and afterwards, it cannot realize automatic control, and the cost is high. When the furnace body capacity is less than 20 tons, the cost of furnace gas analysis equipment will be much higher than that of smelting equipment, making this method difficult. Realize engineering application

Domestic use of mature sonar slag technology to achieve online detection of slag has achieved certain results, but this method has large interference and poor real-time performance

In order to reduce the splashing in the early stage, Shandong Jinan Iron and Steel Co., Ltd. adopted methods of reducing the amount of ore added, changing the time of ore adding and changing the operation mode of the oxygen lance, but the effect was not good, because the time and amount of addition were determined by experience.

Benxi Iron and Steel Works uses the field data collected by the audio slag control system and cursor tracking to conduct linear regression analysis on the cause of splashing and prevent the occurrence of splashing. Although this method has great advantages over the sonar method and the feeding method, they also have problems. , equipment and operation are too complex

However, due to the serious interference of the smelting site environment and the different interference sources, the audio analysis method has low analysis accuracy and changes with the environment, which has great uncertainty. Moreover, when the signal is detected by this method, the splash has already Occurs slightly, which brings practical difficulties to the subsequent pressure spraying, and cannot fundamentally control the occurrence of splashing

The Ispat Inland steel plant in the United States and the converter workshop of the Trinecke zelezarng steel plant in the Czech Republic use mathematical models to predict slopping, and then prevent slopping by adjusting the process and feeding methods. Analyzing the possibility of splashing without online collection of real-time information during smelting can avoid difficulties such as high cost of furnace gas analysis and difficulty in filtering out audio analysis interference, but this method is difficult for enterprises with large changes in material source and composition Be applicable

Images