| Ferrochrome alloy is a ferroalloy with chromium and iron as the main components,which is widely used in chemical production,industrial casting and other industries,and is also an important alloy additive for stainless steel smelting.Argon Oxygen Decarburization Method is a smelting method utilizing AOD furnace as smelting equipment,which is one of the main methods for smelting low carbon ferrochrome alloy,and is also one of the refining methods for producing stainless steel and other alloy steels at present.In the process of smelting ferrochrome alloy,the content of various elements has a great impact on the properties of the alloy,especially the carbon content,which is the key element composition that needs to be controlled in smelting.The smelting of ferrochrome alloy is a physicochemical reaction that occurs in a high temperature environment,usually by blowing oxygen into the initial molten liquid in the converter to promote the oxidation reaction of the elements in the molten liquid in the furnace,and thus play a role in changing the composition of the molten liquid in the furnace and increasing the temperature.In the actual smelting production process,the main core lies in the end point control,which usually requires experienced staff to observe the state of the flame and spark at the furnace mouth,and calculate the end point carbon content according to the initial load in the furnace and the total amount of oxygen consumed by the furnace.This method has high requirements for practitioners and great uncertainty.In recent years,with the development of in-furnace inspection technology,a variety of smelting process control methods have emerged.However,due to the fact that the smelting process involves a large number of physicochemical reactions and complex mechanisms,in order to avoid the occurrence of safety accidents such as spattering,the conservative manual empirical method is still the mainstream smelting control method,which has seriously hindered the process research and performance improvement of the AOD furnace refining process.In this paper,the control of AOD furnace refining process is studied,and the main research contents and results are as follows:(1)The dynamic model of decarburization rate and temperature in the refining process of ferrochrome alloy is studied,and the process mechanism and physicochemical reaction mechanism of the smelting process are investigated in detail.On the basis of the existing refining model,the decarburization law at each stage of smelting is further analyzed,and a model of the whole smelting mechanism is established,which provides a basis for the study of process control methods.(2)A study on the end-point control method of low-carbon minted ferroalloys based on improved internal mold control was carried out.Firstly,the internal mode control was analyzed,and the Anti-Windup design was carried out to deal with the actuator windup phenomenon that occurs when it is directly applied.Secondly,in order to shorten the smelting time,expert control and anti-windup internal model control(EC-AWIMC)method is proposed to control in stages with different filter time constants.Finally,the adaptive fuzzy control and anti-windup internal model control(AFC-AWIMC)method is investigated to further improve the smelting speed by adjusting the current filter time constants with the actual oxygen supply rate,which is proved to shorten the smelting time and improve the temperature control capability of the smelting process through actual smelting.(3)In view of the lack of detection technology in the furnace and the fact that most prediction algorithms can only predict the carbon temperature data at the smelting end point,the Similarity Criteria-Long Short Term Memory(SC-LSTM)Method is proposed to establish the dynamic prediction model of AOD furnace after blowing.The similarity criterion is used to effectively excavate the temporal characteristics between the oven samples,and the correlation between the end carbon temperature and the input characteristics is obtained by the weighted criterion,which can better measure the similarity with the characteristics of time series samples.The adaptive evaluation function for fuzzy clustering to determine the number of categories is constructed,making the clustering effect of the historical library samples more reasonable.(4)Based on the existing AOD furnace control system automation equipment and testing equipment,the control method and prediction method proposed in this paper are integrated with the original system,and the smelting process control software is further developed to provide a method of integrating the refining control system on top of the existing AOD furnace control system.Through the on-line measurement of the system,the hit rate of the prediction of end-point carbon content reaches 20%,40%,92%,and 95%respectively with the allowable errors of ±0.005%,±0.01%,±0.015% and ±0.02%,and the hit rate of the prediction of end-point temperature reaches 31%,80%,93%,and 96%respectively with the allowable errors of ±5℃,±10℃,±15℃,and ±20℃. |
PDF Full Text Request