合金收得率影响因素与预测模型

Factors Influencing Alloy Yield and Prediction Model

脱氧合金化是转炉冶炼工艺的最后一步,钢液内合金成分含量控制是否精确直接影响着精炼工序的冶炼难度与冶炼周期。合金收得率是转炉合金工人配加合金时的重要参考标准,关于合金元素收得率判断的准确性直接影响着钢水成分稳定性与生产成本。该文通过基于聚类的热卡填充与相关分析,完成了对收集数据部分缺失的修复、历史收得率的计算及收得率影响因素的分析。在三次清洗数据之后,该文对剩余相对完整的218组数据进行K-means聚类分析,取同类均值代替插补,依据条件相似度完成了转炉终点Mn含量缺失数据的填充。其次通过查阅文献,得出收得率计算公式。最后针对潜在的变量因素共线性特征,应用相关分析求解相关性系数,得到影响C元素收得率的主要因素为:转炉终点C含量等;影响Mn元素收得率的主要因素为:转炉终点Mn含量等。最后利用Xgboost回归预测求出合金收得率与影响因素的关系。该文对所建立的模型性能进行了研究,对模型预测结果进行了测试,证实模型运行准确可靠,能够为生产冶炼作出指导。

Deoxidization alloying is the last step of converter smelting process.The accuracy of alloy content control in molten steel directly affects the smelting difficulty and smelting cycle of refining process.The yield of alloy is an important reference standard for converter alloying workers.The accuracy of determining the yield of alloying elements directly affects the stability of molten steel composition and production cost.Through the hot card filling and correlation analysis based on clustering,this paper completes the repair of the missing part of the collected data,the calculation of the historical collection rate and the analysis of the influencing factors of the collection rate.After three times of data cleaning,k-means clustering analysis was carried out on the remaining relatively complete 218 sets of data in this paper,the homogeneous mean was taken to replace the interpolation,and the missing data of Mn content at converter endpoint was filled according to the conditional similarity.Secondly,through literature review,the formula for calculating the yield was obtained.Finally,based on the collinearity of potential variables,correlation analysis was applied to solve the correlation coefficient,and the main factors influencing the yield of C element were obtained as follows:C content at the end of converter;The main factors influencing the yield of Mn are the content of Mn at the end of converter.Finally,the relationship between the yield of the alloy and the influencing factors is obtained by using Xgboost regression prediction.In this paper,the performance of the model is studied,and the prediction results are tested,which proves that the model is accurate and reliable,and can provide guidance for smelting production.