Prediction of phosphate concentrate grade based on artificial neural network modeling

Abstract

In order to determine the grade of phosphate concentrate rapidly and accurately on-line, first, the surface color parameters of primary ore, concentrated and tailing of phosphate rock was collected and extracted based on a self-designing surface color acquisition device. Then these parameters were modeled by artificial neural network. The results showed that the 5-12-1 BP model established by artificial neural network could achieve better prediction results with error less than 5%, which provided theoretical support for realizing the on-line soft measurement of flotation concentrate grade. Keywords: Artificial neural network, Phosphate concentrate, Grade prediction, Color sensor