节点文献
基于时间序列数据挖掘的生物氧化提金工艺参数优化
The Optimization of Biological Oxidize Gold Extraction Process Parameters Based on Time Series Data Mining
【作者】 陈飞；
【导师】 南新元；
【作者基本信息】 新疆大学 ， 控制理论与控制工程， 2014， 硕士
【摘要】 生物氧化预处理工艺是解决含砷、硫难处理金矿提金问题的主要方法。但工艺过程受电化学、生物学、物理学和多相混合流体力学等多方面影响，反应极其复杂，通过传统建模方法进行优化研究难度很大，而且大量的工艺过程数据被闲置一旁使隐含的有价信息没有被充分利用，造成了资源的极大浪费。针对上述问题，本文从生物氧化预处理-氰化提金工艺分析出发，深入研究各工艺参数影响作用效果；采用时间序列数据挖掘方法对工艺参数时序数据进行表示，并对ORP关键影响因素与ORP的关联关系进行挖掘；通过智能集成建模方法和挖掘到的关联规则建立ORP趋势预估模型。针对预估模型预估精度低、参数选择粗糙问题，本文引入细菌觅食优化算法(Bacteria Foraging OptimizationAlgorithm, BFOA)对预估模型结构参数进行实时在线优化，优化结果极大地提高了预估模型精度。通过对优化后的预估模型极大值区域的求解得到了工艺参数的最优范围，并依据工艺操作经验验证了结果的正确性和可行性。通过以上研究，本文有效解决了生物氧化预处理-氰化提金工艺参数优化问题，在一定程度上为现场工艺操作提供了数据支撑，为相同领域的优化问题提供了方法依据。
【Abstract】 Biological oxidation pretreatment process is a primary method to solve theproblem of the gold ore containing arsenic and sulpur extraction.However, the processis affected by electrochemistry, biology, physics and hydromechanics. The reactionprocess is complicated and the research of optimization is very difficult throughtraditional modeling approach. Lots of process data that contain valuable informationwere put aside, which cause a waste of resources.In order to solve these problems, this paper studies the influence of variousparameters on process by analysizing the biological oxidation pretreatment processdepply. Time series data mining method is adopted to express process parameters data.The association relationship between ORP and other parameters is mined. ORPtendency predicted model is established through intelligent integrated modelingmethod and the mined association relationship.The model exist drawbacks of lowprecision and rough parameter selection. This paper adopts bacteria foragingoptimization algorithm to optimize the parameters of tendency predicted model online.Optimization results greatly improve the accuracy of model. Optimum range ofprocess parameters are obtained by solving the optimized maxima region of the model.Based on the experience of process, correctness and feasibility of the results isverified.Through the research, the paper effectively solves the parameter optimizationproblem of biological oxidation pretreatment-cyanide gold extraction process. Theresults provide the data support for field process operation and a method for theoptimization problem of the same field.