Hybrid Modeling for Soft Sensing of Molten Steel Temperature in LF

 Aiming at the limitations of traditional thermal model and intelligent model, a new hybrid model is established for soft sensing of the molten steel temperature in LF. Firstly, a thermal model based on energy conservation is described; and then, an improved intelligent model based on process data is presented by ensemble ELM (extreme learning machine) for predicting the molten steel temperature in LF. Secondly, the self adaptive data fusion is proposed as a hybrid modeling method to combine the thermal model with the intelligent model. The new hybrid model could complement mutual advantage of two models by combination. It can overcome the shortcoming of parameters obtained on line hardly in a thermal model and the disadvantage of lacking the analysis of ladle furnace metallurgical process in an intelligent model. The new hybrid model is applied to a 300 t LF in Baoshan Iron and Steel Co Ltd for predicting the molten steel temperature. The experiments demonstrate that the hybrid model has good generalization performance and high accuracy.     

Steel Temperature Compensating Model With Multi-Factor Coupling Based on Ladle Thermal State

 Combined with the parameters of the production process of a steel factory, numerical simulations for a new ladle from preheating to turnover are conducted using the finite element analysis system software (ANSYS). The measured data proved that the simulated results are reliable. The effects of preheating time, thermal cycling times, and empty package time on steel temperature are calculated, an ideal preheating time is provided, besides, based on the analysis of a single factor and use the nonlinear analysis method, a steel temperature compensating model with diversified coupling factors is proposed, with the largest error of the present coupling model at 1462 ℃, and the errors between actual and target steel temperature in tundish after the model is applied to practical production are basically controlled within ±6 ℃, which can meet the accuracy of the manufacturer and has a practical guiding significance for the production in steelmaking workshops.     

Heat Transfer in Steelmaking Ladle

The heat transfer in a steelmaking ladle was studied. The evaluation of heat transfer of the steel was performed by measuring steel temperature in points including all refining steel process. In the ladle, the temperatures in the refractories and the shell were also measured. To evaluate the thermal profile between the hot and cold faces of the ladle in the slag line position, an experiment which shows the importance of thermal contact resistance was carried out. Higher heat losses in the tapping and the vacuum were verified. The temperature measurements of the ladle indicate distinct thermal profiles in each stage of steel refining. Moreover, as each stage of the process depends on the previous one, the complexity of the ladle thermal control is incremental. So a complete model of heat losses in the ladle is complex.     

Development and application of model for predicting molten metal temperature in continuous casting process

An unsteady mathematical thermal model was developed for predicting the time,molten-steel weight,induction heating power,and temperature changes of the steel from the end of ladle refining to the end of the continuous-casting process of a tundish. The calculations revealed that for a specific strip-casting process,the ladle tonnage should be controlled to about 90 t. If the ladle capacity reaches 130 t,the provision of a 1 500-kW tundish induction heating device is recommended. By comparing the measured and predicted molten-steel temperature values in the Ningbosteel-Baosteel strip casting industrialization demo project( NBS) of a tundish,it was determined that the prediction accuracy of the model could meet the forecasting accuracy requirements for the molten-steel temperature in the tundish during mass production. Simultaneously,the heat flux density on each surface of the tundish was found at about 50 min,which is entirely consistent with the values reported in the related literature,and the tundish had not reached a heat balance during the casting test period. This model can also be applied to calculate the suitable size of a tundish for a specific continuous-casting process,thereby providing a theoretical basis for the design of the continuous-casting tundish.     

Numerical Simulation for Effect of Inlet Cooling Rate on Fluid Flow and Temperature Distribution in Tundish

 The fluid flow in tundish is a non-isothermal process and the temperature variation of stream from teeming ladle dominates the fluid flow and thermal distribution in tundish. A numerical model was established to investigate the effect of inlet cooling rate on fluid flow and temperature distribution in tundish based on a FTSC (Flexible Thin Slab Casting) tundish. The inlet cooling rate varies from 0.5 to 0.25 ℃/min. Under the present calculation conditions, the following conclusions were made. When the stream temperature from teeming ladle drops seriously (for inlet cooling rate of 0.5 ℃/min), there is a “backward flow” at the coming end of casting. The horizontal flow along the free surface turns to flow along the bottom of tundish. The bottom flow shortens the fluid flow route in tundish and deteriorates the removal effect of nonmetallic inclusions from molten steel. Nevertheless, when the inlet cooling rate decreases to 0.25 ℃/min, the horizontal flow is sustained during the whole casting period. The present research provides theoretical directions for temperature control in teeming ladle and continuous casting tundish during production of advanced steels.     

A Compensation Model of Continuous Temperature Measurement for Molten Steel in Tundish

 A compensation model has been proposed to reduce errors caused by the immersion depth of the sensor and the time lag of continuous temperature measurement for molten steel in tundish, which is based on the limited data fitting method and data fusion technology. According to the heat transfer analysis of sensor, the thermal model has been bulit to determine the temperature variation function.The parameters of the compensation model are recognized by generic algorithm, which combines the determine function, the molten steel mass in the ladle and pouring time. The processing of error compensation is divided into three stages: tracking, holding and compensation. When the processing is stable, the measured temperature error is small, and the measured temperature is regarded as accurate value and tracked. For the end of pouring stage of the ladle, the temperature error is caused by the immersion depth of the sensor, and the measured temperature before sharp decreasing is considered as real temperature and held. For the temperature increasing stage after ladle changed, the measured temperature is compensated online.The application results show that the error between the compensation temperatures and the actual ones have been decreased to ±2 ℃, and the time lag could be shortened from 3-5 min to 40 s by applying this model.     

Effect of Calcium Treatment on Non-Metallic Inclusions in Ultra-Low Oxygen Steel Refined by High Basicity High Al2O3 Slag

 The influence of calcium treatment on non-metallic inclusions had been studied when control technology of refining top slag in ladle furnace was used in ultra-low oxygen steelmaking. A sufficient amount aluminium was added to experimental heats for final deoxidizing during BOF tapping, and the refining top slag with strong reducibility, high basicity and high Al2O3 in ladle furnace was used to produce ultra-low oxygen steel and the transformation of non-metallic inclusions in molten steel was compared by calcium treatment and no calcium treatment. The results show that the transformation of Al2O3→MgO·Al2O3 spinel→CaO-MgO-Al2O3 complex inclusions has been completed for aluminum deoxidation products and calcium treatment to molten steel is unnecessary when using the control technology of ladle furnace refining top slag to produce ultra-low oxygen steel, and the complex inclusions are liquid at the temperature of steelmaking and easily removable to obtain very high cleanliness steel by flotation. Furthermore, the problems of nozzle clogging in casting operations do not happen and the remaining oxide inclusions in steel are the relatively lower melting point complex inclusions.     

Hybrid Neural Network Model for RH Vacuum Refining Process Control

A hybrid neural network model, in which RH process (theoretical) model is combined organically with neural network (NN) and case-base reasoning (CBR), was established.The CBR method was used to select the operation mode and the RH operational guide parameters for different steel grades according to the initial conditions of molten steel, and a three-layer BP neural network was adopted to deal with nonlinear factors for improving and compensating the limitations of technological model for RH process control and end-point prediction. It was verified that the hybrid neural network is effective for improving the precision and calculation efficiency of the model.     

Final Temperature Prediction Model of Molten Steel in RH-TOP Refining Process for IF Steel Production

 In order to precisely control the final temperature of molten steel in RH (Ruhrstahl Heraeus)-TOP blowing refining, the final temperature prediction models of molten steel in RH-TOP blowing refining process for Interstitial Free (IF) steel production were established under the condition of oxygen blowing and non-oxygen blowing respectively. The results show that the beginning molten steel temperature of refining and the amount of added scrap were influential factors, the baking temperature in vacuum chamber was a factor that had small influence. When the model was operated, the hitting probability was above 95% (under the condition of both oxygen blowing and non-oxygen blowing) of prediction deviation of ±10 ℃. The accuracy is analyzed.     

On-line Ladle Lining Temperature Estimation by Using Bounded Jacobian Nonlinear Observer

The knowledge of transient temperature of the ladle wall is a key factor in optimizing energy consumption in steelmaking process.The transient temperature needs to be estimated.A nonlinear lumped parameter model was used to model the thermal dynamics of the ladle.Then,the bounded Jacobian nonlinear observer was utilized to esti-mate the temperature.With this method,the estimation model became a closed-loop model and the observer gains were obtained by solving linear matrix inequalities and simply implemented to the system.Comparison between the simulation and recorded data at a participating steel plant in Thailand showed that the nonlinear observer accurately estimated the temperature of the ladle lining.This estimated temperature was very useful in determining suitable tap-ping temperature for energy conservation and steel quality.     

转炉-精炼-连铸过程钢中氧的控制

结合近年来的文献和笔者的研究工作,概要论述了转炉—精炼—连铸过程中钢洁净度(以总氧含量T[O]表示)的控制及夹杂物对产品质量的影响。提高钢的洁净度应从产生夹杂物的源头抓起,尽可能降低转炉终点氧含量。根据生产统计数据,建立了转炉终点氧预报模型。介绍了硅镇静钢、硅铝镇静钢、铝镇静钢三种脱氧模式及脱氧产物的控制方法。采用钢包精炼方法把夹杂物消灭在钢水进入结晶器之前是获得“干净”钢水的关键。介绍了RH、LF、中间包钢水总氧预报模型。介绍了在连铸过程中防止钢水再污染和进一步去除夹杂物的措施。     

新型纳米级微孔隔热材料在炼钢生产中的应用

唐山中厚板材有限公司针对炼钢生产中钢水热量损失过大问题,在转炉、钢包和中间包上,系统采用了新型WDS纳米级微孔隔热材料,提高了耐火衬的蓄热量,在不影响耐火材料使用寿命的情况下,转炉的终点温度提高了7℃,从转炉、钢包到LF炉过程温降速度减小了0.3℃/min,采取降低LF炉出站温度5℃的方式,中间包后期钢水温度仍然能够满足连铸工艺需要,达到了节能降耗的目的,可以产生综合经济效益1480万元/a。     

LF精炼终点钢水温度灰箱预报模型

LF精炼工序在炼钢过程起着调节温度的关键作用,准确预报LF精炼终点钢水温度对实际生产有重要意义.传统的LF精炼预报模型包括机理模型与黑箱模型.机理预报模型能够体现各工艺因素对终点钢水温度的影响,但由于LF精炼传热机理研究尚不完善,依靠机理模型预报终点钢水温度,难以达到预期效果;黑箱预报模型能够准确预报终点钢水温度,但不能反映精炼过程各工艺因素对钢水温度的影响,尤其当生产工艺条件发生改变时,黑箱模型在应用上会受到限制.本文以方大特钢LF精炼炉为研究对象,建立一种机理预报模型与黑箱预报模型(BP神经网络预报模型)相结合的LF精炼终点钢水温度灰箱预报模型.该模型既能反映各工艺因素对终点钢水温度的影响,又能准确预测终点钢水温度,其终点钢水温度预测误差在±5℃以内的命中率可以达到95%以上.      

氩站钢包精炼技术开发及应用

玉溪新兴钢铁有限公司针对无钢包精炼炉设备的不利条件,开展了氩站钢包精炼技术开发及应用。通过对钢包造渣工艺、夹杂物处理技术等进行优化,对氩站设备进行改进,钢水及铸坯质量基本达到LF精炼工艺水平,中包连浇炉数达18炉以上,冶炼成分合格率达98．10％,铸坯综合合格率达99．78％,满足了品种钢生产需求,成功开发出08A1、KNS、K510L等新产品,提高了钢材产品市场竞争力。     

LF热态渣循环利用技术

 LF热态渣的循环利用可减少废渣排放,降低对环境的危害。对LF热态循环渣的脱硫能力及可回收性进行了分析,热态循环渣返回LF炉和转炉参与冶金反应后,可大幅降低渣料消耗,LF炉每罐回收热态循环渣1～1.5 t,平均节省石灰及其他助溶剂用量5 kg/t（钢）,转炉每罐回收热态循环渣3～5 t,渣料消耗平均降低10～15 kg/t（钢）。采用热态循环渣配加石灰的LF炉造渣制度后,在相同的处理时间内,处理终点钢水中硫质量分数与常规处理几乎相同,同时节省了能源消耗,但必须考虑对钢水增硅、增锰的影响。热态循环渣返回转炉后导致入炉铁水温度低及吹炼过程渣量较大,因此转炉吹炼全程以低枪位操作更为适宜。在不影响生产组织的情况下,热态渣以返回转炉循环利用为最佳途径。     

新钢包热状态对钢水温度的影响

 以210t未加绝热层的无碳新钢包为研究对象,建立数值模拟模型进行计算,并利用实际生产数据进行验证。计算结果显示新钢包蓄热到第10周期达到热饱和状态;新钢包阶段由于蓄热所造成的钢水最大温降为183℃,据此建立了新钢包阶段的钢水温度补偿模型。实际应用结果表明,该温度补偿模型的应用提高了中间包的钢水温度命中率,对炼钢厂的生产具有一定指导意义。     

BOF→LF→CC流程钢水温度预报

针对炼钢连铸流程的工艺特点和生产数据,建立了基于BP神经网络的"BOF→LF→CC"流程钢水温度预报模型。通过相关性分析筛选模型变量,利用五数概括法筛选数据,采用LM优化算法改进BP神经网络,利用生产数据对模型进行了训练和测试。并用Java语言开发了钢水温度预报模型的程序,在某钢厂进行了应用。结果表明,各区段钢水温度预报模型的预报命中率基本可以满足生产的要求。     

中间包温度优化工艺实践

根据本厂连铸中间包温度不稳定变化情况,总结了影响中间包温度波动的重要因素,主要从给连铸提供钢 水的精炼工序出发,通过快速造渣工艺改进和钢包工艺优化,稳定LF炉终点温度的精确控制,并以此来稳定和提 高连铸中间包温度命中率。     

提高齿轮钢8620RH( K)末端淬火内控符合率的工艺实践

齿轮钢8620RH(K)(/%:0.18～0.22C,0.17～0.26Si,0.70～0.90Mn,≤0.025P,0.015～0.025S,0.4～0.6Cr,0.4～0.7Ni,0.15～0.25Mo,0.015～0.045Al,≤0.010 0N)的冶炼流程为60 t BOF-LF-VD-300 mm×360mm铸坯。分析了工艺改进前BOF终点残余元素,出钢量,精炼合金调整量,VD增N方式和铸坯C偏析,得出BOF出钢量不稳定、中间包钢水过热度高、内控成分合格率低,铸坯C偏析指数高是导致末端淬火试验AHRC值4～5带宽符合率偏低的主要因素。通过稳定转炉装入量和出钢量,LF吹氮气增氮,空置红钢包加盖,控制中间包钢水过热度15～25℃,优化连铸电磁搅拌参数等工艺措施,使8620RH(K)钢末端淬火内控符合率由原<60%提高至82.93%。     

基于IPSO-RELM转炉冶炼终点锰含量预测模型

分析了影响转炉冶炼终点钢水中锰含量的因素, 针对基于BP神经网络算法的转炉冶炼终点锰含量预测模型存在的收敛速度慢, 预测精度低等问题, 提出了一种基于极限学习机(ELM) 算法建模的新思路, 并引入正则化以及改进粒子群优化算法(IPSO), 建立了基于改进粒子群算法优化的正则化极限学习机(IPSO-RELM) 的转炉终点锰含量预测模型; 应用国内某炼钢厂转炉实际生产数据对模型进行训练和验证, 并与基于BP、ELM和RELM算法的三类模型进行比较.结果表明, 采用IPSO-RELM方法构建的模型, 锰含量预测误差在±0. 025%范围内的命中率达到94%, 均方误差为2. 18×10-8, 拟合优度R2为0. 72, 上述三项指标均显著优于其他三类模型, 此外, 该模型还具有良好的泛化能力, 对于转炉实际冶炼过程具有一定的指导意义.