Adaptive Neuro-Fuzzy Inference System (ANFIS) Method to Optimize The Reduction Process of Saprolite Ore Composites in Tube Furnace

Abstract

According to Indonesia Mineral and Coal Law No. 1 in 2014 about the Enhancement of Mineral Value-Added, it is necessary for mining company to process and refine nickel ore domestically to increase its value. In this paper, value of nickel ore is increased by producing a saprolite ore composite which is a mixture of a certain amount of saprolite, coal, sulfate, and bentonite. Then a reduction process of the composite using pyrometallurgical method is designed to find the best combination of the coal ratio, process temperature, process time, and the ratio of additive (Na2SO4) towards the availability of carbon along with processing time and temperature as the primary concern. Then the chemical composition of the saprolite ore composites are analyzed, especially nickel, using X-Ray-Difference Fluorescence (XRF). In order to find the best combination, Adaptive Neuro-Fuzzy Inference System (ANFIS) method is employed to analyze the XRF result due to its ability to reduce the dimension of search space by distributing input information over the network. The objective of this research is to obtain optimal factor combination for reduction process of saprolite ore composites in Tube Furnace by looking at the results of the chemical compositions of Ni which was tested through XRF using ANFIS method. The optimal factor combination is ratio coal 15% with a type of additive Ca2SO4or Composite SB15Ca10P2with temperature 1200 °C and process time 3 hours.