基于多信息融合与可拓理论的锑浮选工况识别方法  

Performance recognition of antimony flotation based on multi-information fusion and extension theory

作　　者：刘美丽[1] 唐朝晖[1] 王晓丽[1] 桂卫华[1] 杨贵军[1] 

机构地区：[1]中南大学信息科学与工程学院

出　　处：《中南大学学报（自然科学版）》2015年第12期4512-4519,共8页Journal of Central South University:Science and Technology

基　　金：国家自然科学基金资助项目(61171192;61272337);教育部博士点基金资助项目(20130162110013)~~

摘　　要：针对锑浮选泡沫图像特征信息冗余、矿源成分复杂引起工况难以识别的问题,结合泡沫图像特征与入况条件参数,提出一种基于多信息融合与可拓理论的锑浮选工况识别方法。首先,采用敏感性指数与相关性分析提取关键泡沫特征,并融合入矿条件参数,作为浮选工况的评价参数,建立锑浮选工况识别的物元可拓模型,同时运用可拓层次分析法(EAHP)确定各评价参数权重,计算待识别工况与预设工况类别的综合关联度,并以最大关联度为准则,实现锑浮选工况的准确识别。研究结果表明:本方法的识别结果与实际情况吻合良好,为锑浮选工况识别提供了一种行之有效的新方法。Considering performance recognition problem caused by information redundancy of froth image features and complex compositions of crude ore in antimony flotation process, a performance recognition method based on multi-information fusion and extension theory was proposed with the combination of froth image features and condition parameters of crude ore. Firstly, taking the key froth image features extracted by the sensitivity index and correlation analysis and condition parameters of crude ore as the evaluation parameters of flotation performance, a matter-extension model for antimony flotation performance recognition was established. Then, after determining the weight of each parameter by utilizing extension analytic hierarchy process（EAHP）, the comprehensive correlation degrees between unrecognized performance and the preset performance were calculated, and thus, antimony flotation performance could be identified accurately according to the maximum correlation principle. The results show that the identification results agree well with the practical situations, Providing a novel valid performance recognition method for antimony floatation.

关 键 词：锑浮选 工况识别 多信息融合 可拓理论 可拓层次分析法(EAHP) 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]