数据驱动的铜锍吹炼造渣过程混合动态模型.docVIP

 数据驱动的铜锍吹炼造渣过程混合动态模 型# 王雅琳，桂卫华，阳春华* 5 10 15 20 25 30 （中南大学信息科学与工程学院，长沙，410083） 摘要：针对具有强动态变化特征的铜锍吹炼造渣过程，提出一种数据驱动的混合动态模型建 立方法。混合模型包括基于反应动力学原理的非线性动力学模型和基于生产数据的动力学系 数修正因子优化模型。考虑到优化模型计算量大，构造规模小但具代表性的典型样本集，并 提出混合智能算法，以确保寻优过程能有效获得最佳动力学系数修正因子，从而使修正的动 力学模型真实反映造渣过程的动态变化。用某厂 PS 转炉的实际生产数据进行实例仿真，混 合动态模型的最大相对预测误差小于 5%。模型有效描述了造渣过程铜锍组份及温度随时间 变化的过程。 关键词：铜锍吹炼造渣过程；混合动态模型；数据驱动；混合智能算法 中图分类号：TF355; TP391.9 Data-driven hybrid dynamic model for slag making stage of matte converting process WANG Yalin, GUI Weihua, YANG Chunhua (School of Information Science and Engineering, Central South University, Changsha 410083, China) Abstract: For the slag-making stage of the matte converting process with strong dynamic characteristics, a data-driven hybrid dynamic model was proposed which includes a nonlinear reaction kinetic model based on metallurgical reaction kinetics and an optimization model for correction factor of its kinetic coefficient based on production data. Taking into account the large amount of calculation for the optimization model, a typical set with small-scale and representative sample was constructed and a hybrid intelligent algorithm was proposed, which ensure that the optimal correction factor can be obtained effectively during searching process and the dynamic changes of the slag-making process can be truly reflected by the amended kinetic model. Contrasting with the real value of PS converter in a smeltery, the maximum relative error of the simulation result of the dynamic model proposed is less than 5%, which show that the model effectively describes the dynamic process of the components and the temperature of matte with time changes in the slag-making process. Key words: the slag-making stage of matte converting process; hybrid dynamic model; 35 40 data-driven; hybrid intelligent algorithm 0 引言 铜锍吹炼造渣过程是对熔融铜锍除铁造渣的冶金过程，其质量控制以人工操作为主，工 艺指标波动大，过程难以优化运行，能源资源浪费严重。为实现优质、增效和节能降耗，开 展该过程的优化控制研究日趋重要。 实现优化控制，首先需要建立过程数学模型。Goto 和 Chen 等人[1-3]基于热力学原理建 立的铜锍吹炼造渣过程数学模型，为静态平衡计算模型，无法