Improving estimation accuracy of metallurgical performance of industrial flotation process by using hybrid genetic algorithm – artificial neural network (GA-ANN)
More details
Hide details
Mining and Metallurgical Engineering Department, Amirkabir University of Technology
Faculty of Engineering, University of Kashan
Sarcheshmeh Copper Concentration Plant
Omid Salmani Nuri   

Amirkabir University of Technology (Tehran Polytechnic), 424 Hafez Ave, 15875-4413 Tehran, Iran
Publication date: 2016-12-01
Physicochem. Probl. Miner. Process. 2017;53(1):366–378
In this study, a back propagation feed forward neural network, with two hidden layers (10:10:10:4), was applied to predict Cu grade and recovery in industrial flotation plant based on pH, chemical reagents dosage, size percentage of feed passing 75 µm, moisture content in feed, solid ratio, and grade of copper, molybdenum, and iron in feed. Modeling is performed basing on 92 data sets under different operating conditions. A back propagation training was carried out with initial weights randomly mode that may lead to trapping artificial neural network (ANN) into the local minima and converging slowly. So, the genetic algorithm (GA) is combined with ANN for improving the performance of the ANN by optimizing the initial weights of ANN. The results reveal that the GA-ANN model outperforms ANN model for predicting of the metallurgical performance. The hybrid GA-ANN based prediction method, as used in this paper, can be further employed as a reliable and accurate method, in the metallurgical performance prediction.