基于遗传算法的加热炉炉内钢温软测量模型

GA-based soft sensing model for billet temperature in reheating furnace

针对加热炉炉内钢温的在线测量问题,本文从总括热吸收率法钢温模型中总括吸收率参数的确定问题入手,尝试利用遗传算法在解决最优化问题方面的突出特点,获得全局最优的总括吸收率,进而得到炉内钢温的软测量模型。该模型有机结合了机理和统计建模方法的优点,在取得大量易获得的过程数据的前提下,可以得到较高的测量精度。采用双蓄热步进式棒材加热炉现场运行数据的仿真结果证明了该方法的有效性。

Aiming at the on-line billet temperature measurement in reheating furnace, this paper makes an effort to get total heat exchange factor from the billet model based on genetic algorithm method that has good feature in optimization. The soft sensing model of the billet temperature was established. This soft sensing model has the advantages of physical and statistical modeling methods. Higher measurement accuracy can be achieved if lots of process data can be easily acquired. The simulation based on the field running data of double-regenerative walking beam reheating furnace shows the validity of this method.