Precision forecast model consistently predicts throughput to within 0.4%.
Minera Centinela is owned by Antofagasta PLC and processes ore from four open pit deposits: Esperanza, Central Treasury, Northeast Treasury and Mirador, from which it produces copper in cathodes and concentrates, in addition to gold as the main by-product. In 2019 it produced 276,000 tonnes of copper and 222,600 oz of gold from the treatment of approximately 30,000,000 tonnes of ore, accounting for just under 40% of Antofagasta PLC’c group copper production and almost 80% of its gold.
The concentrator’s comminution circuit has what is known as a SABC-B configuration. This comprises a single 40ft x 24ft SAG mill followed by two ball mills in parallel. Pebbles from the SAG mill are crushed and rather than being returned to the SAG mill as in a SABC-A circuit are sent directly to the ball mill circuit.
In 2018 professionals from Antofagasta PLC and the Centinela Mining Company teamed up with JKTech to refine the Centinela comminution circuit throughput forecast model which is used to help formulate the company’s future ore processing plans. In addition it is used to quantify the comminution circuit’s final grind size (P80) from which it predicts the flotation circuit’s response to the different geometallurgical ore types.
The development team went through the following systematic process in refining the model:
- Conducted detailed surveys of the comminution circuit including SMC Tests on the feed ore samples
- Constructed mathematical simulation models using JKSimMet
- Carried out simulations (a total of 53) to predict the throughput response of the SAG mill to changes in ore hardness, feed size, speed, ball level and total filling level. Ore hardness inputs to these simulations were generated by the mine’s geometallurgical model which is based on a very large number of SMC Test results that were obtained from drill core samples of future ore.
- Developed a regression equation using the simulation results that relates grinding circuit throughput to feed size and (SMC Test) ore hardness values.
- Calibrated the regression model using historic (2018) production figures and associated hardness data from the geometallurgical model.
The result from the calibration (data reconciliation) process is shown in Figure 1 and indicates that the model is consistently accurate to within 5% on a monthly basis and on a yearly basis is accurate to a very impressive 0.4%.
Going forward, the highly accurate throughput model that the team has developed will enable the company to make production plans and provide investor guidance of future metals production with a high degree of confidence.
2019, Javier Vargas, Sebastián Tello, Iván Yutronic and Gonzalo Barcos, “Update of the Throughput Model and Ore Processing Plan at the Minera Centinela Concentrator”. Proc. 6th International Seminar on Geometallurgy - Procemin-Geomet, Santiago, Chile.