Method for fusing information of spatter predictive analyzers for AOD (argon oxygen decarburization) furnaces

Abstract

The invention discloses a method for fusing information of spatter predictive analyzers for AOD (argon oxygen decarburization) furnaces. The method has the advantages that spatter occurred in smelting procedures is used as a research object for low-carbon ferrochromium smelting production procedures of the AOD furnaces, features of in-furnace audio signals, furnace mouth flame images and furnace body vibration signals which are acquired in fields are extracted on the basis of deep analysis on spatter occurrence mechanisms in the smelting procedures, the feature-level signals are fused by the aid of a process for fusing data of BP (back propagation) neural networks on the basis of feature level fusion, accordingly, the most reasonable judgment results can be obtained, occurrence of spatter in the low-carbon ferrochromium smelting production procedures of the AOD furnaces can be accurately predicted by the aid of the method, the alloy yield can be increased, and the production safety can be guaranteed.

Description

technical field [0001] The invention relates to an information fusion method of an AOD furnace splash forecast analyzer. Background technique [0002] Cause Analysis of Spattering in Argon Oxygen Refining [0003] The main cause of AOD furnace spatter is the unbalanced chemical and thermal reactions in the furnace, resulting in a large amount of gas overflowing with a large amount of molten iron. In addition, due to operational errors or changes in the production process, under the violent flushing of a large amount of oxygen and nitrogen, the carbon-oxygen reaction inside the molten pool furnace is extremely unbalanced, and a large amount of CO, CO 2 eruption, resulting in the generation of splashes. [0004] During the AOD furnace smelting process, the following chemical reactions mainly occur in the furnace body, 1: direct reaction of carbon and oxygen C+O→CO↑, 2: indirect reaction of carbon and ferrous oxide C+FeO→CO+Fe 3: various Necessary trace mineral elements reac...

AI Technical Summary

Problems solved by technology

However, due to the limitation of sampling method and analysis time of this method, the accuracy of real-time forecast is greatly reduced. Therefore, most of this method is only used for research and analysis after the accident, and it is difficult to realize real-time control.

Not only that, but the mass spectrometer equipment for laser analysis is expensive. If the furnace capacity is less than 20 tons, it is difficult for ordinary manufacturers to afford the production efficiency. Therefore, it encounters a big capital budget problem in practical application and popularization.

Images