Modeling of bubble surface area flux in an industrial rougher column using artificial neural network and statistical techniques

Soft computing-based modeling of flotation processes – A review

“…Artificial neural network-based model Copper Aldrich et al (1995Aldrich et al ( , 1997, Moolman et al (1995aMoolman et al ( , 1995bMoolman et al ( , 1995cMoolman et al ( , 1995d, Hales et al (1999), Çilek (2002), Saghatoleslami et al (2004), Massinaei and Doostmohammadi (2010), Nakhaei et al (2012, Nakhaei and Irannajad (2013a,b), Massinaei et al (2014), Jahedsaravani et al (2014Jahedsaravani et al ( , 2015 and Hosseini et al (2015) Molybdenum Nakhaei et al (2012 and Nakhaei and Irannajad (2013b) Platinum group metals Aldrich et al (1995Aldrich et al ( , 1997, Moolman et al (1996), Marais (2010) Scheiner et al (1996), Karr and Scheiner (2000), and Al-Thyabat (2008, 2009) Coal Kalyani et al (2008), Jorjani et al (2008Jorjani et al ( , 2009 regarding the simulation of phosphate flotation in Florida. Scheiner et al (1996) developed six different models for the simulation of phosphate flotation circuit, namely the first principle model, standard multivariate regression model, partial least squares model, neural network model, fuzzy-genetic model and rate constant model.…”

“…A paper presented by Massinaei and Doostmohammadi (2010) has been of particular interest provided it is one of the very few cases where the ANN model is used not directly for prediction or simulation of metallurgical performance, but for modeling the bubble surface area flux as the measure of gas dispersion in the cell (rougher column of a Cu minerals flotation circuit). The model took into account three input variables (superficial gas velocity, slurry density (solids%) and frother dosage).…”

“…The results provided a deep insight into the effect of operating variables, particularly slurry density -which is more complicated than the others -on gas dispersion parameters in an industrial column in rougher circuit. It was finally noted that by replacing slurry density (solids%) with slurry viscosity and including the particle size as an independent variable, a more general model, with the ability to predict bubble surface area flux in industrial columns, could be achieved Massinaei and Doostmohammadi (2010). had proposed two ANN models for predicting the copper content (network architecture [3-13-6-1]) and molybdenum content (network architecture [4-4-8-1]) in the copper concentrate.…”

“…Comparison between (a) statistical and (b) ANN method in predicting the bubble surface area flux (S b )(Massinaei and Doostmohammadi, 2010).…”

Prediction of Gas Holdup in a Flotation Column by Artificial Neural Network

“…In this regard, an Artificial Neural Network (ANN) has been successfully applied in many mineral-processing applications [2][3][4][5][6][7][8][9]. ANN has been applied to estimate Cu and Mo grade and recovery during pilot plant flotation column concentrate [7].…”