钢铁生产和材料质量异常监视与诊断研究

发布时间：2018-04-23 18:33

本文选题：支持向量机 + 分布估计算法　；参考：《东北大学》2014年硕士论文

【摘要】：在钢铁生产过程中,高炉炼铁是通过料面燃烧生成铁水,料面的分布情况直接影响铁水质量,从而影响钢材质量。实际生产中,由于料面的物质形态以及高炉高温生产的特性,很难直接观测到料面分布情况,从而难以掌握高炉炉况对铁水质量的影响。因此,研究高炉料面异常情况的监视和诊断方法对提高铁水质量,降低废材量,提高材料质量具有非常重要的意义。钢铁企业一般以客户合同组织生产时,生产的钢铁材料都匹配给相关的客户合同。但在实际生产中,钢铁材料质量不合格会使原来有合同的材料变成无主合同材料,从而增加了原料库库存量,提高了企业成本,甚至影响企业的正常生产。因此,分析钢铁材料异常原因,并对材料异常情况进行监视及诊断,对提高产品合格率,降低无主合同材料量具有非常重要的意义。本文基于数据解析的思想,分别对高炉料面异常诊断问题和钢铁材料质量异常诊断问题,设计了支持向量机方法对异常情况进行监视与诊断。针对高炉料面异常问题,在诊断方法中提出了在线支持向量机的思想。针对钢铁材料质量异常问题,在诊断方法中引入分布估计算法对支持向量机进行改进,以提升支持向量机算法性能。本文主要研究内容包括以下几个部分：(1)以高炉生产过程为背景,研究了高炉料面异常诊断问题。通过高炉的结构特征、高炉炼铁特点和高炉内温度分布与高炉料面之间的关系,对高炉料面异常情况进行分析。针对高炉炼铁生产的密闭性、高温性与实时性的特点,提出了在线支持向量机的方法对高炉的料面异常进行诊断。利用采集的高炉实时数据训练支持向量机算法建立数学模型,并根据建立的数学模型进行高炉异常料面的在线诊断研究。试验结果表明,在线支持向量机对高炉的料面异常情况的诊断比较准确。(2)针对钢铁材料异常诊断问题,分析了有主合同材料转变成无主合同材料的影响因素,采用支持向量机算法建立了钢铁材料质量异常情况诊断数学模型,模型中采用分布估计算法对支持向量机参数进行优化。实验结果表明,利用诊断数学模型对钢铁材料异常问题的分析比较准确。(3)以实际钢铁生产过程为背景,基于提出的分布估计支持向量机方法,设计开发了钢铁材料质量异常监视与诊断子系统。该系统具有操作简单、灵活等特点,有现货原因分析、午会管理和特殊合同管理三大主要部分,最终为钢铁企业的决策提供数据支持。
[Abstract]:In the process of iron and steel production, blast furnace ironmaking produces hot metal through burning of the feed surface, and the distribution of the material surface directly affects the quality of hot metal and thus the quality of steel. In actual production, it is difficult to observe the distribution of material surface directly because of the material form of material surface and the characteristics of high temperature production of blast furnace, so it is difficult to understand the influence of blast furnace condition on the quality of hot metal. Therefore, it is of great significance to study the monitoring and diagnosis methods of the abnormal condition on the blast furnace surface for improving the quality of hot metal, reducing the quantity of waste materials and improving the quality of materials. When iron and steel enterprises organize production by customer contract, the steel materials are matched to the relevant customer contract. However, in actual production, the unqualified quality of iron and steel materials will make the original contract materials become non-main contract materials, thus increasing the stock of raw materials, increasing the cost of enterprises, and even affecting the normal production of enterprises. Therefore, it is of great significance to analyze the causes of abnormal materials of iron and steel, and to monitor and diagnose the abnormal situation of materials, which is of great significance to improve the qualified rate of products and reduce the quantity of materials without main contract. Based on the idea of data analysis, this paper designs the support vector machine (SVM) method to monitor and diagnose the abnormal situation of blast furnace surface and iron and steel material quality, respectively. The idea of online support vector machine (SVM) is put forward in the diagnosis method for the abnormal surface of blast furnace. In order to improve the performance of support vector machine (SVM), the distribution estimation algorithm is introduced to improve the quality of iron and steel materials. The main contents of this paper are as follows: (1) based on the background of blast furnace production process, the problem of abnormal diagnosis of blast furnace surface is studied. Based on the structural characteristics of blast furnace, the ironmaking characteristics of blast furnace and the relationship between the temperature distribution of blast furnace and the surface of blast furnace, the abnormal situation of the surface of blast furnace is analyzed. Aiming at the characteristics of sealing, high temperature and real time of blast furnace ironmaking, a method of on-line support vector machine (SVM) is put forward to diagnose the abnormal surface of blast furnace. The mathematical model is established by training the real time data of blast furnace with support vector machine algorithm, and the on-line diagnosis of abnormal material surface of blast furnace is studied according to the established mathematical model. The experimental results show that the on-line support vector machine is more accurate in the diagnosis of abnormal material surface of blast furnace. Aiming at the problem of abnormal diagnosis of iron and steel materials, the factors influencing the conversion of the main contract materials into the non-main contract materials are analyzed. Support vector machine (SVM) algorithm is used to establish a mathematical model for the diagnosis of abnormal quality of iron and steel materials. In the model, the parameters of SVM are optimized by using the distribution estimation algorithm. The experimental results show that the diagnostic mathematical model is used to analyze the abnormal problems of iron and steel materials accurately. (3) based on the actual steel production process, the support vector machine (SVM) method for distribution estimation is proposed. An anomaly monitoring and diagnosis subsystem for iron and steel materials is designed and developed. The system is characterized by simple operation, flexible operation, spot cause analysis, afternoon meeting management and special contract management, and finally provides data support for iron and steel enterprises' decision-making.
【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F426.31

【相似文献】