Evaluation of effects of operating parameters on combustible material recovery in coking coal flotation process using artificial neural networks

In this research work, the effects of flotation parameters on coking coal flotation combustible material recovery (CMR) were studied by the artificial neural networks (ANNs) method. The input parameters of the network were the pulp solid weight content, pH, collector dosage, frother dosage, conditioning time, flotation retention time, feed ash content, and rotor rotation speed. In order to select the most efficient model for this work, the outputs of different models were compared with each other. A five-layer ANN was found to be optimum with the architecture of 8, 15, 10, and 5 neurons in the input layer, and the first hidden, second hidden, and third hidden layers, respectively, as well one neurons in the output layer. In this work, the training, testing, validating, and data square correlation coefficients (R2) were achieved to be 0.995, 0.999, 0.999, and 0.998, respectively. The sensitivity analysis showed that the rotor speed and the solid weight content had the highest and lowest effects on CMR, respectively. It was verified that the predicted ANN values coincided very well with the experimental results.

Keywords:

Coking Coal , Flotation , Artificial Neural Networks , Back-Propagation Neural Network , Combustible Material Recovery