Thesis Optimización del circuito de molienda SAG en Minera Los Pelambres mediante modelos fenomenológicos, machine learning y optimización multiobjetivo
Loading...
Date
2026-10-30
Authors
Journal Title
Journal ISSN
Volume Title
Program
Ingeniería Civil Química
Campus
Campus Casa Central Valparaíso
Abstract
La molienda es una de las etapas de mayor consumo energético de una planta concentradora de cobre. En los circuitos con molinos semiautógenos (SAG), además, suele constituir el principal cuello de botella del proceso, ya que limita la cantidad de mineral que la planta puede tratar. Por esta razón, aumentar el tratamiento de un molino SAG, incluso en algunos puntos porcentuales, puede tener un impacto relevante en la producción de cobre fino. Sin embargo, el objetivo operacional no consiste únicamente en procesar más mineral: también se busca utilizar la energía de manera eficiente, entregar un producto con el tamaño adecuado a la molienda secundaria y proteger los revestimientos. El caso de estudio corresponde a la planta concentradora Los Piuquenes de Minera Los Pelambres, ubicada en la Región de Coquimbo. Su circuito de molienda primaria está compuesto por cuatro líneas, cada una equipada con un molino SAG. Los molinos SAG 1, SAG 2 y SAG 3 tienen dimensiones nominales de 36 × 17 pies y motores de aproximadamente 15 MW, mientras que el SAG 4 tiene dimensiones de 38 × 19 pies y un motor de 20 MW. El molino SAG es un sistema dinámico y parcialmente observable. Su interior no puede observarse durante la operación, por lo que el estado de la carga debe inferirse a partir de señales indirectas, como el peso, la potencia, el ruido y los flujos. Mientras la dureza y la granulometría del mineral cambian continuamente, un sistema de control experto ajusta la alimentación, el agua y la velocidad de giro para mantener el circuito estable y operar con el mayor tratamiento posible dentro de una zona segura. Una de las principales dificultades es que el peso medido por la celda de carga no representa directamente la carga interna del molino. Durante cada campaña se pierden varias toneladas de acero debido al desgaste de los revestimientos, los lifters, las parrillas y los medios de molienda. Como la celda mide conjuntamente el molino, sus revestimientos y la carga interna, su lectura disminuye aunque la cantidad de mineral y bolas no cambie. Además, (...).
Grinding is one of the most energy-intensive stages in a copper concentrator plant. In circuits utilizing semi-autogenous (SAG) mills, this stage often acts as the primary process bottleneck, limiting the total amount of ore the plant can process. Consequently, increasing the throughput of a SAG mill—even by just a few percentage points—can have a significant impact on the production of fine copper. However, the operational objective goes beyond simply processing more ore; it also involves using energy efficiently, ensuring the product is ground to the appropriate size for secondary grinding, and protecting the mill liners. This case study focuses on the Los Piuquenes concentrator plant at Minera Los Pelambres, located in the Coquimbo Region. Its primary grinding circuit consists of four lines, each equipped with a SAG mill. SAG mills 1, 2, and 3 have nominal dimensions of 36 × 17 feet and are powered by approximately 15 MW motors, while SAG mill 4 measures 38 × 19 feet and is equipped with a 20 MW motor. A SAG mill is a dynamic, partially observable system. Because its interior cannot be monitored during operation, the state of the mill charge must be inferred from indirect signals, such as weight, power draw, noise, and flow rates. As the hardness and particle size distribution of the ore constantly fluctuate, an expert control system adjusts the feed rate, water addition, and rotational speed to maintain circuit stability and maximize throughput while staying within a safe operating range. A major challenge is that the weight measured by the load cell does not directly represent the actual internal load of the mill. Furthermore, (...).
Grinding is one of the most energy-intensive stages in a copper concentrator plant. In circuits utilizing semi-autogenous (SAG) mills, this stage often acts as the primary process bottleneck, limiting the total amount of ore the plant can process. Consequently, increasing the throughput of a SAG mill—even by just a few percentage points—can have a significant impact on the production of fine copper. However, the operational objective goes beyond simply processing more ore; it also involves using energy efficiently, ensuring the product is ground to the appropriate size for secondary grinding, and protecting the mill liners. This case study focuses on the Los Piuquenes concentrator plant at Minera Los Pelambres, located in the Coquimbo Region. Its primary grinding circuit consists of four lines, each equipped with a SAG mill. SAG mills 1, 2, and 3 have nominal dimensions of 36 × 17 feet and are powered by approximately 15 MW motors, while SAG mill 4 measures 38 × 19 feet and is equipped with a 20 MW motor. A SAG mill is a dynamic, partially observable system. Because its interior cannot be monitored during operation, the state of the mill charge must be inferred from indirect signals, such as weight, power draw, noise, and flow rates. As the hardness and particle size distribution of the ore constantly fluctuate, an expert control system adjusts the feed rate, water addition, and rotational speed to maintain circuit stability and maximize throughput while staying within a safe operating range. A major challenge is that the weight measured by the load cell does not directly represent the actual internal load of the mill. Furthermore, (...).
Description
Keywords
Molienda SAG, Gemelo digital, Optimización operacional, Aprendizaje automático, Sensores virtuales, Eficiencia energética, Consumo específico de energía, Sistema experto, Industria minera
