迁钢二炼钢动态运行过程中若干工序基本参数分析与精准控制方法研究

发布时间：2018-04-29 03:17

本文选题：炼钢厂 + 动态运行　；参考：《北京科技大学》2016年博士论文

【摘要】：随着“高效率、低成本洁净钢平台集成技术”和钢铁生产过程“智能制造”的提出,炼钢厂动态运行过程管控越来越受到重视,但对炼钢厂动态运行过程基本参数(物流量、温度和时间因子)的系统研究仍处于前期,特别是对基本参数控制精度的研究还是空白。本文将针对迁钢第二炼钢厂动态运行过程中若干工序基本参数进行分析,分别对物流量、钢水温度和时间的管控以及精准控制下的节能潜力四个方面进行研究。本文分析了迁钢第二炼钢厂的转炉工序、精炼工序、连铸工序、钢包周转、传搁过程的主要工艺参数和控制现状,以及RH精炼、210t LF精炼和传搁过程中钢水温降与物料消耗和时间的关系。由物流量控制现状的解析得知：转炉冶炼过程实现了少渣冶炼,氧气消耗量和利用率相对稳定,但吨钢铁损较大。针对该问题,提出针对每一炉次具体的原料条件,动态调整冶炼辅料加入量的策略,实例计算结果表明,吨钢铁损可以减少4kg/t钢。分别对各工序节点钢水温度的预定和预报进行了管控研究。前者是指在合理的温度制度的基础上,考虑连浇炉次、各工序及其传搁过程时间、钢包状态等因素动态调整各工序节点的控制温度；后者是以上一工序节点的钢水温度为基础,预报下一工序节点的钢水温度,进而连接实现全流程的钢水温度预报。分别结合RH和210t LF两种精炼工艺的特点,运用改进的案例推理方法对钢水温度预报进行研究,该方法可以应用于其他区段,进而实现全流程的钢水温度预报。结合RH精炼已积累了大量生产数据的特点,提出了案例库拆分和优化的措施以提高案例推理方法的预报精度和运算速度,实际数据测试结果表明,基于拆分案例库的案例推理方法的预报精度明显高于单一案例库的预报精度,基于优化案例库的案例推理方法运算时间仅为优化前的44%,运算效率大幅提升。结合210t LF精炼的工艺参数关联性复杂的特点,提出利用贝叶斯网络对生产数据可靠性进行评估,并利用评估结果对案例推理方法进行修正。测试结果表明,修正后的案例推理方法比一般案例推理方法和BP神经网络都有更高的预报精度。在炼钢厂时间因素管控中,分别对确定合理连浇炉数、优化不同浇次间关系和钢包调度水平的评价方法进行研究。在BOF-RH-CC工序产能不匹配的情况下,提出了基于传搁能耗计算,进而对比生产成本的方法确定合理连浇炉数。为了优化不同浇次间关系,在非满负荷生产条件下,提出以钢包运行稳定性为目标,采用遗传算法对已有生产计划进行优化,对迁钢二炼钢2013年某日实际生产计划优化后的结果表明,该日周转钢包离线次数由5次减少到1次,必须投入钢包个数由7个减少到5个,可节约烘烤煤气20400m3,避免转炉对离线钢包所需的钢水温度补偿20℃。对钢包调度水平的研究,提出基于待用时间解析的方法分别评价周转钢包和离线钢包的运行现状,并拟合为钢包利用率用以评价钢包整体调度水平,通过对迁钢二炼钢钢包调度进行实例分析,结果表明钢包利用率可以有效的反映实际钢包调度水平。针对钢水温度和时间因素精准控制下能潜力问题,根据实际操作要求,通过解析生产计划的制定和执行过程,得到二者在现有控制精度下节能潜力的数学模型。结合迁钢二炼钢SPHC-W钢种冶炼过程,将各阶段钢水温度和时间生产数据统计的标准差表征为控制精度,对现有控制水平进行实例研究,结果表明,当各工序节点钢水温度的标准差减小1℃时,计划执行预知温度损失降低0.33℃,当各工序时间标准差减小1 min时,生产计划预知温度损失降低1.76℃C,计划执行预知温度损失降低1.62℃,二者同时减小时表现出可叠加性。
[Abstract]:With the introduction of "high efficiency, low cost clean steel platform integration technology" and "intelligent manufacturing" of steel production process, the control and control of the dynamic operation process of the steelmaking plant has been paid more and more attention. However, the systematic research on the basic parameters of the dynamic process of the steelworks (logistics, temperature and time factors) is still in the early stage, especially for the basic parameters control. The research on the precision of the system is still blank. This paper will analyze the basic parameters of some processes during the dynamic operation of the second refinery, and study the four aspects of the flow, the temperature and time of the molten steel and the energy saving potential under the precision control. This paper analyzes the converter process and the refining process of the second refinery. The main process parameters and control status of the continuous casting process, ladle turnover and transmission process, as well as the relationship between temperature drop of steel water and material consumption and time during the process of RH refining, 210t LF refining and transmission, and the analysis of the current situation of flow control in the converter: the converter smelting process has realized less slag smelting, oxygen consumption and utilization are relatively stable, but tonn steel In view of this problem, the strategy of dynamic adjustment of the amount of smelting auxiliary material is put forward according to the specific raw material conditions of each furnace. The calculation results show that the steel loss of tons can be reduced by 4kg/t steel. The control and control of the temperature of the steel water in each process is studied. The former refers to the basis of the reasonable temperature system. The control temperature of each process node is dynamically adjusted by factors such as continuous casting, each process and its transmission time, ladle state and other factors. The latter is based on the temperature of the steel water of the above working procedure node, and the temperature of the steel water in the next process node is predicted, and then the temperature of the whole process is connected with the two kinds of RH and 210t LF respectively. The characteristics of refining process, using improved case reasoning method to study the temperature prediction of steel water, this method can be applied to other sections, and then realize the whole process of molten steel temperature prediction. Combined with the characteristics of RH refining has accumulated a large amount of production data, a case base separation and optimization measures are proposed to improve the preview of case reasoning method. The actual data test results show that the prediction accuracy of the CBR Based on the Split case base is obviously higher than that of the single case base. The time of the case based reasoning method based on the optimized case base is only 44% before optimization, and the operation efficiency is greatly improved. Combined with the process parameters Association of 210t LF refining. A Bayesian network is used to evaluate the reliability of production data, and the results of the evaluation are used to modify the case reasoning method. The test results show that the modified CBR has a higher pre reporting precision than the general case reasoning method and the BP neural network. In the time factor control of the steelworks, respectively The evaluation method of determining the number of reasonable pouring furnace, optimizing the relationship between different pouring times and the level of ladle scheduling is studied. In the case of the mismatch of the production capacity of the BOF-RH-CC process, a method is put forward to determine the number of reasonable pouring furnaces based on the calculation of the energy consumption and the comparison of the production cost. Under the condition, the existing production plan was optimized by genetic algorithm. The result of optimization of the existing production plan was made by genetic algorithm. The results showed that the off-line number of off line ladles was reduced from 5 times to 1 times in 2013, and the number of ladles must be reduced from 7 to 5, and the roasted gas 20400m could be saved by 20400m. 3, avoid the converter's temperature compensation for the water temperature required for the off-line ladle at 20. Study on the scheduling level of the ladle, and propose a method to evaluate the running status of the revolving ladle and the off-line ladle respectively based on the analysis of the waiting time, and fit the utilization rate of the ladle to evaluate the overall scheduling of the ladle, and to carry out an example of the transfer of steel ladle for the transfer of steel two. The results show that the ladle utilization rate can effectively reflect the actual ladle scheduling level. According to the precise control of the potential of the temperature and time of the steel, according to the actual operation requirements, the mathematical model of the energy saving potential of the two persons under the existing control precision is obtained by analyzing the process of making and executing the production plan. The model is combined with the moving steel two. The smelting process of steel SPHC-W steel in steelmaking is characterized by the standard deviation of the statistics of the temperature of the steel water and the time production data at various stages. The existing control level is studied by an example. The results show that when the standard deviation of the molten steel temperature of each process is reduced by 1 degrees C, the predicted temperature loss is reduced by 0.33 degrees C, and the standard deviation of each process time is reduced. When 1 min is small, the production plan foresee that the temperature loss will be reduced by 1.76 degrees C, the planned implementation of predicted temperature loss will be reduced by 1.62 degrees, and the two will be superimposed at the same time.

【学位授予单位】：北京科技大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TF345

【参考文献】