针对铜闪速熔炼过程工艺指标无法在线检测、过程建模及优化控制困难的问题,研究了基于数据驱动的操作模式优化方法.论文在铜闪速熔炼过程特点分析的基础上,定义了基于数据驱动的操作模式优化的基本概念,提出了基于数据驱动的操作模式优化控制框架,研究了基于数据的冰铜温度、冰铜品位、渣中铁硅比的工艺指标预测模型、炉况的综合评价模型及闪速熔炼过程的操作模式优化.基于大量工业运行数据和炉况评价模型构建优化操作模式库,提出了将模糊C均值聚类与混沌伪并行遗传算法相结合的匹配算法,从优化操作模式库中寻找与当前工况相匹配的最优操作模式,从向实现熔炼过程的优化控制.在铜闪速熔炼生产中的实际应用证明,该方法的有效性.
Considering the difficulties of modeling, online-measurement of technical indexes, and optimal control in copper flash smelting process, a data-driven operational-pattern optimization method is presented. Firstly, the copper flash smelting process is analyzed, basic concepts about data-driven operational-pattern are defined and the frame of data- driven operational pattern optimization is proposed. Secondly, the data-driven prediction models of matte temperature, matter grade and ratio of Fe to SiO2 are established, the overall evaluation model of flash smelter is proposed and operational-pattern optimization for copper flash smelting process is described. Thirdly, an optimized operational-pattern base is constructed based on lots of industrial running data and the overall evaluation model. Then, a matching algorithm combining fuzzy C-means cluster with chaos pseudo parallel genetic Mgorithm is proposed to mine an optimal operational pattern from the optimized operational-pattern base to implement the optimal control of the smelting process. The practical running results in copper flash smelting process show its effectiveness.