摘要
针对现有层流冷却过程带钢温度模型缺乏换热系数、带钢定位、带钢卷取温度计算的有效方法这一问题,提出了由冷却单元阀门开闭状态模型、带钢冷却单元定位模型、不同换热方式下的带钢温度模型组成的带钢卷取温度动态模型,将案例推理、规则推理、神经网络等相结合,提出了规则与数据驱动的模型参数智能辨识方法.采用某钢厂实际生产运行数据对所提出的带钢卷取温度动态模型进行了实验研究,实验结果表明本文提出的方法能够有效提高带钢卷取温度模型的精度.
The existing cooling process models lack the methods to compute the heat transfer parameter and the position that strip reaches and cannot be used to compute the strip coiling temperature directly. So a strip coiling temperature model is proposed, which consists of the status of cooling unit valves calculating model, the strip segment tracking model, and the top surface temperature model under different heat transfer conditions. What is more, a rule and data driven hybrid intelligent identification algorithm is developed combining the case-based reasoning, rule-reasoning with the neural network. The tests using real industrial data of a steel plant have been conducted and indicated that the proposed strip coiling temperature model has made a great contribution to the prediction precision of the strip coiling temperature during the laminar cooling process.
出处
《自动化学报》
EI
CSCD
北大核心
2012年第11期1861-1869,共9页
Acta Automatica Sinica
基金
国家重点基础研究发展计划(973计划)(2009CB320601)
国家自然科学基金(61104084)
创新引智计划(111计划)(B08015)
住建部科学技术计划项目(2012-K7-19)资助~~
关键词
层流冷却
参数辨识
规则驱动
数据驱动
卷取温度
Laminar cooling, parameter identification, rule-driven, data-driven, coiling temperature
