Desenvolvimento e validação de um modelo de balanço populacional para a lixiviação de um concentrado ustulado de zinco em bancada e em escala piloto

Abstract

A Population Balance (PB) model was developed for the leaching of a roasted zinc concentrate (RZC) in sulfuric acid solutions, considering that no previous work has validated with experimental data a PB model for this system. Therefore, leaching experiments were done in order to validate and refine the proposed model. Initially, the chemical and mineralogical composition of the RZC, determined by XRF, XRD, and AAS, indicated the predominance of zinc (65% w/w), mainly as zincite. Next, batch leaching experiments were carried out using typical conditions of industrial processes. The parameters investigated were the stirring speed, the acid concentration, the reaction time, and the molar ratio, which is the molar ratio between the acid and the zincite present in the RZC loaded to the reactor. It was observed that the initial acid concentration affected only the initial leaching rate, while the molar ratio determined the maximum zinc extraction achieved. It was also found that the model correctly described this behavior, which suggests that the model accounted for the major phenomena occurring during the leaching. In the next step, the modeling of continuous leaching was done, considering two different approaches for the modeling of flow patterns in the reactors, the segregated flow, and the perfect mixing. For the first approach, the Residence Time Distribution (RTD) for the reactors of the pilot plant was determined, using positive step type experiments at different stirring speeds, feed rates, and configurations (1 reactor, 2 reactors in series, and 3 reactors in series). It was verified that the behavior of the system approaches the behavior of ideal reactors (continuous stirred-tank reactor). This behavior was favored by higher stirring speeds and lower feed flow rates. Then, continuous leaching tests were carried out in the leaching pilot plant, and the results obtained were compared with the data generated by the model. It was verified that the perfect mixing approach is the best one to represent the flow pattern of the reactors, and the results of the model are very close to those obtained experimentally. Then, using statistical analysis, it was verified the validity and accuracy of the proposed model. A potential application of this model would be predicting the effect of changes in operational parameters (e.g., particle size, distribution polydispersity, and acid concentration) on the reactor performance. Thus, the model provides a robust technique for plant design and control, which may lead to better decision-making and costs reduction.