摘要:
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, 从而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了该方法的有效性.
Abstract:
Considering the difficulties of modeling,online-measurement of technical indexes, and optimal control incopper flash smelting process, a data-driven operational-patternoptimization method is presented. Firstly, the copper flash smeltingprocess is analyzed, basic concepts about data-drivenoperational-pattern are defined and the frame of data-drivenoperational pattern optimization is proposed. Secondly, thedata-driven prediction models of matte temperature, matter grade andratio of Fe to SiO2 are established, the overall evaluation modelof flash smelter is proposed and operational-pattern optimizationfor copper flash smelting process is described. Thirdly, anoptimized operational-pattern base is constructed based on lots ofindustrial running data and the overall evaluation model. Then, amatching algorithm combining fuzzy C-means cluster with chaos pseudoparallel genetic algorithm is proposed to mine an optimaloperational pattern from the optimized operational-pattern base toimplement the optimal control of the smelting process. The practicalrunning results in copper flash smelting process show itseffectiveness.
点击查看大图
下载:
