动态纹理建模在硫浮选工况的识别分析  

Flotation state classification applying dynamic textures analysis

作　　者：陈青[1,2] 朱俊宇[1] 唐朝晖[1] 刘金平[1] 桂卫华[1] 

机构地区：[1]中南大学信息科学与工程学院,湖南长沙410083 [2]湖南工业大学计算机与通信学院,湖南株洲412008

出　　处：《计算机与应用化学》2013年第10期1117-1121,共5页Computers and Applied Chemistry

基　　金：国家自然科学重点基金资助项目(61134006);国家自然科学基金资助项目(61071176;61171192;61272337)

摘　　要：泡沫浮选广泛应用于选矿领域,它是一种能够有效提取矿粒的方法。但是,浮选过程存在着大量的影响因素和严重的非线性,这使得浮选过程的优化控制很难实现。因此,为了保证浮选处于最优工况,有必要依据浮选泡沫的表面特征来调整相应的操作变量。本文提出了基于动态纹理建模的方法应用于浮选工况的分类。采用ARMA模型进行动态纹理建模,通过样本学习得到模型参数A,C,Q。对不同类样本模型参数A,C计算其Martin距离,根据最小距离原则来进行分类识别。仿真结果表明:所提出的动态纹理模型能准确地描述动态泡沫,且能有效地检测浮选泡沫状态。Froth flotation is widely used in mineral processing industry, it is an efficient method for extracting ore particles. In the flotation process, however, there are a large number of influencing factors and severe non-linear, which makes the optimization of the flotation process control is difficult to achieve. Therefore, in order to ensure the flotation is in optimum state, it is necessary to adjust the corresponding operating variables based on the surface characteristics of the froth. This paper presents a modeling approach based on dynamic texture used in the classification of flotation states. ARMA model is applied in dynamic texture modeling, and model parameters were obtained by the learning samples A, C, Q. For different types of samples of model parameters A and C, the Martin distance was calculated, and then the principle of the minimum distance was used for classification. Simulation results show that： the dynamic texture model can accurately describe dynamic froth, and can effectively detect flotation froth states.

关 键 词：泡沫浮选 浮选工况 动态纹理 ARMA模型 Martin距离 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]