Simulation study of the optimal distribution of cyanide in a gold leaching circuit

Abstract

The mineral industry has been using cyanidation to recover gold from ores for more than a century; however, a systematic study of the best reactant addition strategy in a cascade of agitated leaching tanks is not available in the open literature. A phenomenological mathematical model of the gold cyanidation process, calibrated with a set of industrial data from an Australian plant, together with an economic performance index is used to analyze this problem. The simulated results show that the best compromise between the two antagonistic effects, cyanide consumption and gold recovery, which are both function of cyanide concentrations, leads to a reagent distribution that depends on the leaching and cyanide consumption kinetics, pulp feed characteristics, and economic factors such as the gold market value. For the specific studied plant, in the operating range of low cyanide consumption and fast gold dissolution, all the cyanide must be added in the first tank; however, in the operating conditions of high cyanide consumption, cyanide has to be distributed in the first, second and third tanks.

Dynamic simulation of the carbon-in-pulp and carbon-in-leach processes

“…The first one explains the constant rate of cyanide consumption by copper and iron minerals and the second explains the effect of cyanide consumption proportional to the free cyanide content, such as the losses by hydrolysis to hydrogen cyanide and ammonia and the oxidation to cyanate or cyanogen; the rate of consumption is thus (Adams, 1990;de Andrade Lima, 2001;de Andrade Lima et al, 2002;de Andrade Lima and Hodouin, 2006):…”

“…The equations that describe the CIL and only leaching process are calibrated with the data from a plant that treats about 400 t/h of ore coming from twelve deposits located in the Eastern Goldfields Province, Yilgarn Craton, Western Australia (Roberts and Elias, 1990;Bax, 1996;de Andrade Lima, 2001;de Andrade Lima and Hodouin, 2006). The ore contains basalt and small amounts of marcassite (FeS 2 ), pyrite (FeS 2 ), chalcopyrite (CuFeS 2 ), galena (PbS), sphalerite (ZnS), gersdorffite (NiAsS), arsenopyrite (FeAsS), and pyrrhotite (Fe 1-x S), but the reactive sulfides content is very low (Roberts and Elias, 1990;Bax, 1996).…”

“…Concentration of the solids was also given for each sample. Every two hours, samples of the pulp were taken and filtered to measure the concentration of carbon in the tanks and their gold content was measured by fire assay and atomic absorption spectrophotometry (Bax, 1996;de Andrade Lima, 2001;de Andrade Lima and Hodouin, 2006). The time evolution of the ore flow rates was measured at the entrance of the comminution-leaching section of the plant, but not at the leaching-adsorption circuit entrance; therefore in this study estimated values are used, by assuming that the comminution circuit behaves as a second-order system with a time constant of 0.37 hours.…”

Real time optimization based on a serial hybrid model for gold cyanidation leaching process

“…Wadsworth et al (2000) proposed a mechanistic model in which the surface reactions were thought to control the gold dissolution rate kinetically. Afterward De Andrade Lima and Hodouin (2006) proposed a mechanistic model of GCLP that consisted of mass conservation equations of gold in the solid phase, gold in the liquid phase and cyanide in the liquid phase. Therein the kinetic models of gold dissolution and cyanide consumption were identified with a set of industrial data from an Australian plant.…”

“…It was shown that the best configuration was not identical for different gold content and cyanide concentration. A systematic study of the optimal distribution of cyanide in a cascade of leaching tanks was performed by De Andrade Lima and Hodouin (2006). However, the above optimization problems are only the simulation studies based on a fixed mechanistic model, which are difficult to be applied in industrial practice.…”