Coupling effect and characterization modeling of iron ore fines mixing and granulating at 0–1 mm
Coupling effect and characterization modeling of iron ore fines mixing and granulating at 0–1 mm
复制标题
DOI：
10.1007/s42243-019-00330-x
复制
发表时间：
2019-11
影响因子：
2.5
通讯作者：
Dai-fei Liu;Xianju Shi;Chaojun Tang;Hai-peng Cao;Jun Li
中科院分区：
文献类型：
--
作者：
Dai-fei Liu;Xianju Shi;Chaojun Tang;Hai-peng Cao;Jun Li
铁矿石的特性是影响制粒的关键因素。以镜铁矿、磁铁精矿和褐铁矿3种矿石为粘结剂，研究了粘结剂的成粒行为和团聚演化过程。根据选择性、成球性和粘附率进行了实验和建模，以代表造粒行为。水的质量分数为（11 ± 1）%，粘附和成核粒径分别为0-1 mm和3-5 mm。实验结果表明，选择性和成球性促进了造粒的发展。镜铁矿的吸水率和褐铁矿的成球性较好。两矿混合存在耦合效应，当磁铁精矿比例大于镜铁矿比例或镜铁矿与褐铁矿混合比例时，耦合效应呈现正效应。三种矿石混合时，耦合效应呈现复杂的叠加状态。采用随机森林算法建立了混合造粒粘附率的表征模型。其输出为粘附率，输入为吸水率、成球指数和配合比。模型参数为957棵树和4个分支，模型的训练和预测误差分别为2.3%和3.7%。模拟结果表明，随机森林模型可以用来描述混合造粒的耦合效应。
Characteristic of iron ore is the essential factor of granulating. Three ores, namely specularite, magnetite concentrate and limonite, were selected as adhesion powder to investigate granulating behavior and evolution process of agglomeration. Experiments and modeling were performed to represent granulating behavior on the basis of selectivity, ballability and adhesion rate. The mass fraction of water and particles size of adhesion and nucleation were set at (11 ± 1)%, 0–1 mm and 3–5 mm, respectively. Experimental results show that selectivity and ballability promote the evolution of granulation. The water absorption rate of specularite and the ballability of limonite are better. The coupling effects exist in two ores mixing and present positive effect when the proportion of magnetite concentrate is greater than that of specularite or specularite and limonite blend. During three ores mixing, the coupling effect presents a complex superposition state. A characterization model of adhesion rate of mixing granulation was established by random forest algorithms. Its output is adhesion rate, and its inputs include water absorption rate, balling index and mixing proportion. The model parameters are 957 trees and four branches, and the training and prediction errors of the model are 2.3% and 3.7%, respectively. Modeling indicates that the random forest model can be used to represent coupling effects of mixing granulation.