针铁矿法沉铁过程亚铁离子浓度预测

摘要

The reactor outlet ferrous iron concentration of iron precipitation process by goethite is obtained by off-line analysis with a long time delay, which leads to difficult real-time control for removing iron process. It is significant to research the reactor outlet ferrous iron concentration on-line prediction method for this problem. On the basis of analyzing the chemical reaction mechanism in the iron precipitation process, by integrating the influence of copper ions to the reaction process and the feature of continuous stirred tank reactor (CSTR), the mechanism model for the iron precipitation process is established in this paper. The double particle swarm optimization (DPSO) algorithm based on information exchange is proposed to determine the parameters in the mechanism model, and the error compensation model for the mechanism model is built with least squares support vector machine (LS-SVM). The ferrous iron concentration integrated predicting model is established by the mode of parallel connection with compensation. The application results demonstrate that the integrated prediction model can effectively reflect the variation tendency of ferrous iron concentration, which may give a foundation for the optimization control in the procedure of iron precipitation by goethite.%针对针铁矿法沉铁过程出口亚铁离子浓度离线化验获得,存在很大滞后性,难以实现沉铁过程实时控制的问题,研究反应器出口亚铁离子浓度在线预测方法。本文在分析沉铁过程化学反应机理的基础上,考虑铜离子对反应过程的影响,结合连续搅拌反应器(Continuous stirred tank reactor, CSTR)特性,建立了针铁矿法沉铁过程的机理模型,并提出了基于信息交换的双粒子群搜索算法(Double particle swarm optimization, DPSO)优化选择机理模型的参数,构建基于最小二乘支持向量机(Least squares support vector machine, LS-SVM)的机理模型输出误差的补偿模型,采用并联补集成方式建立了亚铁离子浓度的集成预测模型。工业现场数据验证了所建模型能有效地反映亚铁离子浓度的变化趋势,为针铁矿法沉铁过程的优化控制奠定了基础。