红柱石基耐火材料抗铜液侵蚀模拟研究

发布时间：2018-03-01 17:57

本文关键词： 红柱石莫来石化抗侵蚀性能 BP神经网络　出处：《西安建筑科技大学》2015年硕士论文　论文类型：学位论文

【摘要】：冶金工业用熔铜炉炉衬材料所面对的苛刻条件和熔炼工艺要求,使耐火材料必须具备优异的理化性能,其中最重要的就是抗侵蚀性能,它直接决定了炉衬的使用寿命。近年来,红柱石以其优良的抗侵蚀性能和体积稳定性等优点受到广泛关注。因此,探究红柱石基耐火材料对改善炉衬材料抗侵蚀性能的机理,具有重要意义。在此基础上,构造人工神经网络中的BP网络结构,建立原始矿物成分和生产工艺参数与产品抗侵蚀性能之间的关系,将促进材料的研发,使得研究过程高效而有针对性。本文以矾土、红柱石、棕刚玉、活性?-Al2O3微粉、SiO2微粉为原料,硼酸为助熔剂,糊精为黏合剂,研究分别加入不同含量(8%、16%、25%、33%、42%、51%)的红柱石颗粒对化铜感应炉炉衬耐火材料抗侵蚀性能的影响。利用X射线衍射(XRD)检测加入红柱石前后试样的物相组成变化,扫描电子显微镜(SEM)观察和分析试样烧成后的晶体形貌。并在此基础上,通过神经网络的BP算法,建立预测红柱石基耐火材料抗铜液侵蚀性能的模型。结果表明:(1)随着红柱石含量的增加,耐压强度和体积密度的变化趋势相似,都显现出两个峰值。显气孔率和烧后线变化率的变化趋势则恰好相反。当红柱石的添加量为16%时,刚玉和玻璃相的共同作用,使得材料结构变得致密。当红柱石的添加量为51%时,红柱石含量的增加使得一次莫来石化和二次莫来石化生成大量的莫来石,莫来石的网状结构得以加强。(2)当红柱石添加量为25%时,试样的抗侵蚀性能最好。红柱石的一次莫来石化过程形成网状结构,大部分SiO2液相填充气孔,排挤到晶体表面的剩余SiO2液相继续与Al2O3进行二次莫来石化反应,完善莫来石网状结构。(3)耐火材料的制造工艺及其特性决定了耐火材料的性能参数只能是一种统计分布的数据,因此,很难使用解析式精确预测。本文中的红柱石基耐火材料抗铜液侵蚀性能BP网络模型说明了神经网络对于耐火材料性能预测方面的适用性。由此模型得到的化学成分与性能关系也完全符合实际理论。这些结果表明,ANN方法对组成和性能间呈复杂非线性关系的红柱石基耐火材料而言是一种有效的分析工具。
[Abstract]:Requirements of metallurgical industry lining materials of copper smelting furnace in the face of harsh conditions and melting process, the refractory materials must have excellent physical and chemical properties, of which the most important is the anti erosion performance, which directly determines the life of the lining. In recent years, andalusite has received widespread attention because of its excellent corrosion resistance and volume stability and other advantages. Therefore, research of andalusite based refractory to improve the anti erosion mechanism performance of lining material, which is of great significance. Based on the BP network structure of artificial neural network, establish the relationship between the original minerals and production process parameters and product properties of anti erosion, will promote the development of material. The research process of efficient and targeted. In this paper, bauxite, andalusite, corundum, activity? -Al2O3 powder, SiO2 powder as raw material, boric acid as additive, dextrin as adhesive, study respectively The different contents (8%, 16%, 25%, 33%, 42%, 51%) the influence of andalusite particle erosion resistance of copper induction furnace lining refractory. By using X ray diffraction (XRD) detection of samples before and after adding andalusite phase composition changes, scanning electron microscopy (SEM) crystal morphology observation and analysis the sample after sintering. And on this basis, through the BP neural network algorithm, establish the prediction of andalusite based refractory copper liquid anti erosion performance model. The results showed that: (1) with the increase of andalusite content, compressive strength and bulk density change trend is similar, showing two peaks. The change trend of linear change rate porosity and burn after adding to the contrary. Popular pillar is 16%, corundum and glass phase interaction, the material became compact. Adding popular pillar is 51%, the increase of andalusite content makes a mullite And the two time Mullitization of mullite formed a lot of reticular structure of mullite can be strengthened. (2) when the andalusite addition is 25%, the sample is best. The anti erosion properties of andalusite a Mullitization process of the formation of network structure, most of the SiO2 liquid filled pores, exclusion to the remaining SiO2 the surface of the liquid crystal phase to two Mullization reaction with Al2O3, improve the mullite network structure. (3) the manufacturing process of refractory material and its characteristics determine the performance parameters of only refractory material is a kind of statistical distribution of the data, therefore, it is difficult to use analytic accurate prediction of andalusite based refractory copper resistance. In this paper, the liquid material erosion performance of the BP network model to illustrate the applicability of neural network for the prediction of performance of refractory materials. The relationship between chemical composition and properties of this model are also accord with the actual theory. These results show that ANN The method is an effective analytical tool for the complex nonlinear relationship between the composition and the properties of the andalusite based refractory.

【学位授予单位】：西安建筑科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TQ175.1

【相似文献】