Font Size: a A A

Predictive Control Of Slurry PH For Antimony Rougher Flotation Process

Posted on:2015-07-25Degree:MasterType:Thesis
Country:ChinaCandidate:Z J ZengFull Text:PDF
GTID:2298330431999326Subject:Control Science and Engineering
Abstract/Summary:PDF Full Text Request
In the mineral flotation process, the pH value directly affects the ion composition of pulp, the activity of flotation reagents, as well as the flotability of minerals. The optimal flotation indexes in various ore flotation production can be obtained only in the appropriate pH environment. And the pH control process has the characteristics of large time-delay and strong non-linearity, in addition, the online inspection equipments used for the pH test are of big error and poor stability, which brings great challenges to the pH value control. In the practical production, the operating workers control the pulp pH value usually by off-line assaying the pH value and manually adjusting the reagents-addition amount, which has poor real-time performance, low accuracy, large reagents consumption, and leads to frequent fluctuations of working conditions. Therefore, the real-time accurate measurement and control of slurry pH value is of great significance to make the flotation production indexes stable.A suitable pH value of the slurry is the key to efficient froth flotation, and the froth surface features indicate the slurry pH directly. Since the control of slurry pH is characterized with strong nonlinearities, long time delays and difficult online measurement, the Principal Component Analysis is combined with the important index of variable projection to select the pH-associated sensitive image features, then a soft sensor model-multi-model LSSVM (least squares support vector machine) is introduced, which is based on affinity propagation clustering (AP). Then, an online support vector regression (OSVR) control strategy of slurry pH is presented based on differential evolution (DE). The proposed methods firstly construct an offline prediction model of slurry pH, and then correct the prediction model online using its online learning capabilities. With respect to each sampling point, the DE optimization method is used to solve the decision variables, so as to achieve the real-time control of slurry pH value. The industrial application results show that the proposed control strategy can stabilize the slurry pH value, reduce the chemical consumption, and improve the flotation efficiency.
Keywords/Search Tags:differential evolution, affinity propagation clustering, online support vector regression, predictive control for PH value
PDF Full Text Request
Related items