摘要
应用BP神经网络技术分别建立混匀矿烧结基础特性预报模型和烧结矿质量预报模型。采用Visual C++2010与Matlab 2008混合编程的方式,开发了烧结配矿专家系统软件。结果表明:同化性温度、液相流动性指数和粘结相强度的预报命中率分别为90%、83.3%和90%,成品率、转鼓指数和低温还原粉化RDI_(+3.15)的预报命中率分别为90%、90%和85%。通过专家系统研究了褐铁矿配比对混匀矿烧结基础特性和烧结矿质量的影响,随褐铁矿配比的增加,烧结矿的成品率、转鼓指数和低温还原粉化RDI_(+3.15)先升高后降低。当混匀矿的同化温度、液相流动性指数和粘结相强度分别为1 259℃、1.22和1 727 N时,烧结矿的质量指标最优。
The BP neural network technology is used to establish the prediction model of sintering basic characteristics and the prediction model of sinter quality. Using C ＋ ＋ Visual 2010 and Matlab 2008 mixed programming, the expert system software was developed. Results show： The prediction hit rate of the assimilation temperature, the liquid phase mobility index and the bond strength were 90% , 83.3 % and 90% , respectively. The prediction bit rate of production yield, tumbler index,low temperature reduction degradation RDI＋3.15 were 90%, 90% and 85%, respectively. It has been researched by expert system that the influence of the ratio of limonite on the sintering basic characteristics and sinter quality. With the increasing proportion of limonite, sinter yield, tumbler index and low temperature reduction degradation RDI＋3.15 first increased and then decreased. When the assimilation temperature, the liquid phase mobility index and the bond strength were 1 259 ℃, 1.22 and 1727 N, respectively. The quality index of sinter was the best.
出处
《烧结球团》
北大核心
2017年第2期10-14,24,共6页
Sintering and Pelletizing
基金
河北省自然科学基金-钢铁联合基金资助项目(E2012209014)
关键词
烧结基础特性
烧结矿质量
BP神经网络
专家系统
basic sintering characteristics
sinter quality
BP neural network
expert system
