胡念三 1刘克稻 2张鑫宇 3冯文菲 1邱杨 1朱浩3
作者信息
- 1. 中国十九冶集团有限公司,四川 成都 610000
- 2. 桂林理工大学 机械与控制工程学院,广西 桂林 541004
- 3. 四川大学 建筑与环境学院,四川 成都 610065
- 折叠
摘要
Abstract
The overall blast furnace sliding transport is an innovative construction method in the furnace renovation process.Prior to the sliding transport,accurate prediction of the pushing performance is crucial for ensuring the smooth operation of the blast furnace pushing system(BFPS).This paper develops a predictive model based on the K-Nearest Neighbors(KNN)regression approach to forecast the maximum friction and stress during the furnace sliding process for an efficient and cost-effective solution.Firstly,based on the transformation project of Yugang blast furnace,a simulation model of the blast furnace displacement system is built.Through analysis of actual working conditions,it was found that the height difference and gaps between track segments are key factors affecting the maximum friction and stress during sliding.Therefore,multiple simulation experiments with different combinations of track height differences and gap parameters were designed,and through simulation analysis obtained a training sample set.Then,a prediction model for the displacement performance of blast furnace displacement system based on KNN regression was proposed.To improve the performance of the model,Bayesian optimization search was also used to optimize the hyperparameters of the KNN regression algorithm.Finally,the prediction results of the KNN model were compared with three classic machine learning algorithms,including support vector regression(SVR)and random forests(RF),to verify the performance advantages of the KNN model.The results show that the prediction model based on KNN regression algorithm has high accuracy in predicting the maximum friction force and stress during the sliding process,and ensures the consistency of the two types of prediction results under the same operating conditions;In comparative analysis,KNN regression algorithm has better accuracy and stability than the other three algorithms in this task.Compared with the other three methods,the KNN model not only responds quickly,but also has strong practicality and reliability,which can provide effective support for practical engineering applications.关键词
高炉推移/推移性能预测/KNN回归/贝叶斯优化Key words
blast furnace pushing system/pushing performance prediction/KNN regression/Bayesian optimi-zation分类
建筑与水利
