摘要
煤粉喷吹量调节系统是一个动态、随机干扰严重的多输入单输出非线性对象,传统煤粉喷吹量控制技术存在精度低、自适应能力小的缺陷,济钢基于模糊神经网络的学习功能和处理不确定情况的能力,开发了喷煤量精确控制技术,该技术实施后,喷煤量相对误差最大值由15%以上降低到2%以下,相对误差标准偏差由2.83%变化到0.52%,高炉技术指标得以改善。
PCI（pulverized coal injection） control system is dynamic with multiple inputs and singular output and thus with strong disturbance, nonlinear. The traditional control is of low precision and capability of self-adaptation. By use of neural network, an accurate PCI control system was developed and has been applied at Ji＇nan Iron and Steel Co.. The results show that, the maximum related error of PCI rate was decreased from above 15% to below 2%, the standard error of relative error was reduced from 2.83% to 0.52%, the technical and economic indexes of BF were improved.
出处
《钢铁》
CAS
CSCD
北大核心
2006年第6期9-12,共4页
Iron and Steel
关键词
高炉
粉煤喷吹
精确控制
模糊神经网络
blast furnace
pulverized coal injection
accurate control
fuzzy neural network
