摘要
为实现铜转炉渣产出量的及时准确预报,提出应用数据挖掘技术从现场积累的大量生产数据中发掘相关规律。首先应用线性回归技术建立了仅考虑主要影响因素(铜锍含铁量)的粗略预报模型,而后,应用神经网络技术建立了考虑到多个次要影响因素的误差补偿模型,从而改进预报效果。利用实际生产数据对模型进行了仿真测试,仿真结果表明,模型预报效果良好。
To forecast slag weight of copper PS converter, it is proposed to extract some laws from lots of production data using the data mining technology. Firstly, a coarse forecasting model based on only the primary influencing factor (that is the weight of iron in matte) is built using linear regression analysis, then, an error compensating model based on other influencing factors is built to improve the result of forecast. The simulative experiment shows that the model has a good performance, and the idea and approach to build model is feasible and practical.
出处
《化工自动化及仪表》
CAS
2007年第2期17-19,共3页
Control and Instruments in Chemical Industry
基金
国家自然科学基金项目(50374079)
国家博士点基金项目(20030533008)
关键词
过程优化
数据挖掘
铜锍吹炼
渣产出量
process optimization
data mining
copper matte converting
slag weight