热连轧带钢卷取温度智能优化设定控制模型及工业应用

来源期刊：上海金属2010年第3期

论文作者：张大志 叶海丽 项晓菲 许新超 陈冰

文章页码：51 - 55

关键词：热连轧带钢； 卷取温度； 遗传算法； 神经网络； 智能优化

Key words：Hot Rolled Strip； Coiling Temperature； Genetic Algorithms； Neural Networks； Intelligent Optimization

摘    要：卷取温度是热轧带钢生产中最重要的质量指标之一。为了克服传统卷取温度设定控制模型不够精确等缺陷,根据热连轧带钢生产过程的特点以及高温带钢在层流冷却区域的温度变化特性,采集并分析了大量实测数据,在此基础上,结合改进遗传算法的全局搜索能力和神经网络的非线性拟合能力,建立了基于改进遗传神经网络的带钢卷取温度优化设定控制模型。实际应用证明,该优化模型完全满足在线生产要求:带钢全长卷取温度偏差100%控制在目标值±20℃以内,93%控制在目标值±10℃以内。

Abstract: Coiling temperature is one of the most important targets in hot strip rolling process.The coiling temperature can be controlled by laminar cooling system.A lot of measured data were collected and analyzed through many experiments considering the characteristics of hot strip rolling process and the strip temperature variation in the laminar cooling area,upon these,a new coiling temperature setting control model based on improved genetic algorithms neural network was established to improve the imprecise traditional control model.The industrial application of new model showed the high precision,the temperature control error of the full strip length within control target value ±20℃ was 100%,while the value of ±10℃ was 93%.