摘要
针对迁钢公司2160热轧在实际生产中遇到的带钢头尾宽度超差问题，分析其过程控制系统中短行程控制模型的设定及在线自适应方法，并通过模型仿真给出各人工配置参数对神经网络形式曲线形状的影响，据此来调整人工配置参数和自适应参数。实践表明，细化人工配置参数层别和调整在线自适应参数可成功降低带钢头尾宽度与本体宽度的偏差，提高带钢头尾宽度控制精度。
During hot strip roiling, short stroke control technology is necessary to avoid the fish tail and get rectangular ends. For the problem of width excess of strip head and tail ends in 2160 hot strip mill, the set-up and adaption principles of its SSC model were analyzed, and the influence of each human configuration parameter to the neural network SSC curve was evaluated by emulation, according to which the human configuration parameters and adaption parameters were adjusted in practice. The result showed that these measures can reduc strip head and tail width deviation effectively.
出处
《首钢科技》
2009年第3期30-33,共4页
Shougang Science and Technology
关键词
短行程控制
仿真
神经网络
自适应
short stroke control, emulation, neural network, adaption
