论文部分内容阅读
连铸坯的质量控制对提高产品质量和降低生产成本具有重要作用,对生产过程中连铸坯的质量状况进行在线判定和预报已成为很多冶金学者和工程人员关心的热点问题。实际生产过程中,连铸坯质量缺陷包括铸坯表面、内部和形状等各种缺陷类型,且各类缺陷的成因复杂,难以全部采用机理模型进行描述。多元模糊模式识别是基于模糊集理论对具有不确定性和非线性关系的系统能够有效辨识的一种方法。本文将多元模糊模式识别应用于连铸坯质量缺陷的判定系统中,给出了缺陷类型判定的详细步骤和方法。以莱钢特殊钢厂20CrMnTiH齿轮钢连铸大方坯为研究对象,结合冶金理论分析和主成分分析法确定影响连铸坯内部质量的工艺参数,将铸坯无缺陷(合格)、角部裂纹、中间裂纹、中心裂纹以及中心偏析缺陷作为标准模式,并通过统计分析得到不同缺陷模式下各因素的隶属函数,采用最大隶属原则对20CrMnTi连铸坯的质量缺陷类型进行判定,预测准确率为81.82%。结果表明,该方法能够准确地对浇铸过程中连铸坯发生各类缺陷的类型进行预测,在实际生产中具有一定的实用性。
The quality control of continuous casting billet plays an important role in improving the product quality and reducing the production cost. It has become a hot issue that many metallurgists and engineers pay attention to on-line judgment and prediction of the quality of the continuous casting billet in the production process. In the actual production process, the quality defects of continuous casting billet include various types of defects such as surface, internal and shape of slab, and the causes of various types of defects are complex and it is difficult to describe all of them by mechanism models. Multivariate fuzzy pattern recognition is a method based on fuzzy set theory that can effectively identify a system with uncertainties and nonlinearities. In this paper, multivariate fuzzy pattern recognition is applied to the judgment system of slab quality defects, and the detailed steps and methods of judging the defect types are given. Taking 20CrMnTiH gear steel continuous casting bloom of special steel mill of Laiwu Steel as the research object, combining with the metallurgical theory analysis and principal component analysis to determine the process parameters affecting the internal quality of the continuous casting slab, the slab has no defect (pass), corner crack, Middle crack, center crack and center segregation defect were taken as standard models. The membership function of each factor under different defect modes was obtained through statistical analysis. The maximum subordination principle was used to determine the type of quality defects of 20CrMnTi slab. The prediction accuracy was 81.82% . The results show that this method can accurately predict the types of defects occurred in the casting process, and has certain practicability in practical production.