Evaluación de la respuesta metalúrgica del mineral la española e incorporación de su efecto al modelo de recuperación, minera candelaria.
Loading...
Date
2025
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Universidad de Concepción
Abstract
Este informe evalúa la respuesta metalúrgica del mineral La Española y su impacto en el modelo de recuperación de Minera Candelaria. Se ejecutó una campaña geometalúrgica con 18 compósitos representativos: 7 de Candelaria, 6 de La Española y 5 mezclas. Se realizaron caracterización química y mineralógica, pruebas de conminución (BWi, RWi, SPI, SMC y Ai), flotación rougher por cinética y test de ciclo abierto, además de evaluación reológica.
La Española se caracteriza por una mayor generación de finos y una mineralogía más compleja. La calcopirita presentó menor participación respecto a evaluaciones previas, mientras que la calcosina surge como nuevo portador de cobre. En conminución, el mineral resulta más exigente de lo esperado: con BWi 13,86 kWh/t, SPI 118,8 min, Acb 31,4 y Ai 0.16. Estos valores, cercanos a los de Candelaria, implican mayor demanda energética y desgaste de medios y revestimiento de lo esperado. En flotación rougher, la cinética de La Española es más lenta y converge a 78-86% de recuperación, mientras Candelaria supera 90% en pocos minutos. Las mezclas muestran respuesta intermedia, manteniendo sobre 90% con hasta 10% de La Española en alimentación. El test de ciclo abierto indica una eficiencia de limpieza cercana a 98% con presencia de este mineral. En reología, La Española presenta un yield stress promedio de 2,82 Pa y una viscosidad de 0,011 Pa∙s a 55 %S,superiores a los de Candelaria (0,68 Pa y 0,009 Pa∙s), con un incremento mercado sobre ese nivel.En conjunto, los resultados evidenciaron mayor exigencias energéticas y operacionales en molienda, flotación y manejo de pulpas.
Comparando con una previa campaña en 2018, en 2024 se observaron leyes y recuperaciones menores, atribuibles al control geológico y protocolo. Las muestras 2024 representan los primeros años de alimentación y provienen de zonas someras y meteorizadas, mientras que en 2018 se usaron condiciones no estándar, por lo que se priorizó los resultados 2024 para la estimación del efecto en el modelo.
La caracterización 2024 evidenció un mineral mas duro, abrasivo y con menor liberación de sulfuros flotables, junto a una mayor ganga férrica y menor cobre libre. Estas condiciones explican las recuperaciones rougher de 76-86%, inferiores a las históricas de Candelaria. El modelo de recuperación ajustado por regresión multivariables e incorporando el efecto La Española alcanzó un MAPE (Error porcentual absoluto medio) de 3,9%, adecuado para análisis y tendencias, aunque aún limitado para control operacional. Se recomienda ampliar la base de datos con muestras más profundas, validar el modelo con datos planta, ajustar colectores con foco en Cu soluble y selectividad frente a ganga férrica, y definir lineamientos de mezcla que mantengan recuperación objetivo sin comprometer molienda ni transporte de pulpas.
This report evaluates the metallurgical response of ore from La Española and its impact on Minera Candelaria’s recovery model. A geometallurgical campaign was carried put with 18 representative composites: 7 from Candelaria, 6 from La Española, and 5 blends. Chemical and mineralogical characterization, comminution tests (BWi, RWi, SPI, SMC and Ai), rougher flotation kinetics an open-cycle tests, as well as rheological evaluations were conducted. La Española is characterized by higher fines generation and a more complex mineralogy. Chalcopyrite showed lower participation compared to previous evaluations, while chalcocite emerges as a new significant copper carrier. In comminution, the ore proved more demanding than expected, with a BWi of 13.86 kWh/t, SPI of 118.8 min, Axb of 31.4, and Ai of 0.16. These values, closer to those of Candelaria, imply higher energy demand and greater wear on grinding media and liners. In rougher flotation, La Española exhibits slower kinetics, converging to 78–86% recovery, while Candelaria exceeds 90% within a few minutes. Blends show intermediate responses, maintaining recoveries above 90% with up to 10% La Española in the feed. The open-cycle test indicates a cleaning efficiency near 98% in the presence of this ore. In rheology, La Española shows an average yield stress of 2.82 Pa and a viscosity of 0.011 Pa·s at 55% solids, higher than Candelaria’s (0.68 Pa and 0.009 Pa·s), with a sharp increase above this level. Overall, the results revealed greater energy and operational demands in grinding, flotation, and pulp handling. Compared with a previous 2018 campaign, the 2024 tests showed lower grades and recoveries, attributed to geological control and protocol differences. The 2024 samples represent the first years of processing and come from shallow, weathered zones, while 2018 tests used non-standard laboratory conditions; therefore, only 2024 results were considered for the model evaluation. The 2024 characterization revealed a harder, more abrasive ore with lower liberation of floatable sulfides, increased iron-rich gangue, and less free copper. These conditions explain the lower rougher recoveries (76–86%) compared to Candelaria’s historical values. The recovery model, adjusted through multivariable regression and incorporating the La Española effect, achieved a Mean Absolute Percentage Error (MAPE) of 3.9%, suitable for analysis and trend evaluation though still limited for operational control. It is recommended to expand the dataset with deeper samples, validate the model with plant data, adjust collector strategies focusing on soluble Cu and selectivity against iron gangue, and define blending guidelines to maintain target recovery without compromising grinding or pulp transport.
This report evaluates the metallurgical response of ore from La Española and its impact on Minera Candelaria’s recovery model. A geometallurgical campaign was carried put with 18 representative composites: 7 from Candelaria, 6 from La Española, and 5 blends. Chemical and mineralogical characterization, comminution tests (BWi, RWi, SPI, SMC and Ai), rougher flotation kinetics an open-cycle tests, as well as rheological evaluations were conducted. La Española is characterized by higher fines generation and a more complex mineralogy. Chalcopyrite showed lower participation compared to previous evaluations, while chalcocite emerges as a new significant copper carrier. In comminution, the ore proved more demanding than expected, with a BWi of 13.86 kWh/t, SPI of 118.8 min, Axb of 31.4, and Ai of 0.16. These values, closer to those of Candelaria, imply higher energy demand and greater wear on grinding media and liners. In rougher flotation, La Española exhibits slower kinetics, converging to 78–86% recovery, while Candelaria exceeds 90% within a few minutes. Blends show intermediate responses, maintaining recoveries above 90% with up to 10% La Española in the feed. The open-cycle test indicates a cleaning efficiency near 98% in the presence of this ore. In rheology, La Española shows an average yield stress of 2.82 Pa and a viscosity of 0.011 Pa·s at 55% solids, higher than Candelaria’s (0.68 Pa and 0.009 Pa·s), with a sharp increase above this level. Overall, the results revealed greater energy and operational demands in grinding, flotation, and pulp handling. Compared with a previous 2018 campaign, the 2024 tests showed lower grades and recoveries, attributed to geological control and protocol differences. The 2024 samples represent the first years of processing and come from shallow, weathered zones, while 2018 tests used non-standard laboratory conditions; therefore, only 2024 results were considered for the model evaluation. The 2024 characterization revealed a harder, more abrasive ore with lower liberation of floatable sulfides, increased iron-rich gangue, and less free copper. These conditions explain the lower rougher recoveries (76–86%) compared to Candelaria’s historical values. The recovery model, adjusted through multivariable regression and incorporating the La Española effect, achieved a Mean Absolute Percentage Error (MAPE) of 3.9%, suitable for analysis and trend evaluation though still limited for operational control. It is recommended to expand the dataset with deeper samples, validate the model with plant data, adjust collector strategies focusing on soluble Cu and selectivity against iron gangue, and define blending guidelines to maintain target recovery without compromising grinding or pulp transport.
Description
Tesis presentada para optar al título de Ingeniero/a Civil Metalúrgico/a.
Keywords
Metalurgia, Mineralogía