基于案例推理的球团质量预测模型研究

Research on pellet quality prediction model using case-based reasoning

针对带式焙烧机球团质量的控制问题,提出基于案例推理的球团质量预测模型,建立焙烧性能指标与球团质量指标之间的关系。首先,运用某炼铁厂焙烧球团的生产数据构建初始案例库,采用k-means聚类算法建立索引结构;然后,用最近邻法检索出相似案例,下一步修改重用,利用主动学习策略保存当前案例,更新案例库;最后,用焙烧球团的实际数据进行仿真,验证模型的有效性。与多元回归分析结果对比,基于案例推理模型的预测结果具有较高的精度,可为其进一步的推广应用提供依据。

In view of pellet quality control of the travelling grate,a case-based reasoning model for predicting pellet quality is proposed,and the relationship between the roasting performance index and the pellet quality index is established.First,the initial case library is constructed by using pellet production data of an iron-making plant,and index structure is built by adopting k-means clustering algorithm.Then the nearest neighbor method is used to retrieve similar cases for further modification and reuse,and the active learning strategy is used to save the current case and update the case library.Finally,the actual data of roasted pellets are simulated to verify the validity of the model.Compared with the results of multiple regression analysis,it is found that the prediction results of case-based reasoning model have high accuracy,so as to provide basis for further popularization and application.