Probability density function of bubble size based reagent dosage predictive control for copper roughing flotation

Jianyong Zhu,Weihua Gui,Chunhua Yang,Honglei Xu,Xiaoli Wang
DOI: https://doi.org/10.1016/j.conengprac.2014.02.021
IF: 4.057
2014-01-01
Control Engineering Practice
Abstract:As an effective measurement indicator of bubble stability, bubble size structure is believed to be closely related to flotation performance in copper roughing flotation. Moreover, reagent dosage has a very important influence on bubble size structure. In this paper, a novel reagent dosage predictive control method based on probability density function (PDF) of bubble size is proposed to implement the indices of roughing circuit. Firstly, the froth images captured in the copper roughing are segmented by using a two-pass watershed algorithm. In order to characterize bubble size structure with non-Gaussian feature, an entropy based B-spline estimator is hence investigated to depict the PDF of the bubble size. Since the weights of B-spline are interrelated and related to the reagent dosage, a multi-output least square support vector machine (MLS-SVM) is applied to depict a dynamic relationship between the weights and the reagent dosage. Finally, an entropy based optimization algorithm is proposed to determine reagent dosage in order to implement tracking control for the PDF of the output bubble size. Experimental results can show the effectiveness of the proposed method.
What problem does this paper attempt to address?