Track 2: Process Innovation, Circularity and Recovery

Process Innovation, Circularity and Recovery

Process Innovation, Digitalization and Transformation

From Strategy to Results: Implementing Integrated Mine-toProcess Optimization at Toromocho A. Muñoz1*, M. Condori1, P. Carrillo1, W. Valery2, R. Valle2, E. Tabosa2, K. Duffy2, R. Hayashida2, C. Plasencia2 and B. Gonzales2 1Minera Chinalco Perú S.A, Perú (Chinalco), Junin, Perú, (*Presenting author: amunoz@chinalco.com.pe) 2Hatch Pty Ltd (Hatch), Brisbane, Australia. ABSTRACT Minera Chinalco Peru S.A. has implemented an integrated Mine-to-Process optimization program at the Toromocho operation to increase throughput when treating harder than anticipated feed ore. The plant design was based on ore hardness values considerably softer than the current feed. Therefore, the specific energy consumption is higher than design and throughput was constrained below design capacity. The comprehensive Mine-to-Process optimization program, conducted in collaboration with Hatch, developed integrated strategies to maximize production and profitability of the overall operation considering the variable feed ore properties. This requires a structured approach and was implemented in phases, commencing with drill and blast optimization to improve ROM fragmentation, which has a significant impact on comminution circuit performance and capacity. Since 2017, structured blasting initiatives have reduced ROM P80 and increased fines content, providing the foundation for comminution optimization. Subsequent optimization of comminution and flotation identified improvements to leverage these upstream benefits through the value chain. Integrated analysis and modelling identified opportunities to increase throughput by up to 14% across both grinding circuits, but this higher throughput would be detrimental to flotation recovery. However, quick wins identified for the flotation circuit could improve overall copper recovery by approximately 1.5% with additional recovery benefits possible from several major CAPEX initiatives. The implemented quick wins have already delivered measurable gains in throughput, recovery, and operational stability, while the longer term planned upgrades are expected to enable the operation to reach or exceed design capacity. Several aspects of this project have been presented previously (Fangrong et al., 2025, 2023, 2022). However, this paper documents the entire project, presenting the structured methodology, results for each phase and the implementation progress. This illustrates how a comprehensive and structured optimization strategy recognizing and accounting for changing ore characteristics can be effectively translated into sustained operational results. Keywords Mine-to-Process, strategy, optimization, comminution, flotation, throughput, copper, recovery, process, simulations, sustainability. 1. INTRODUCTION Minera Chinalco Perú S.A. (Chinalco) operates the Toromocho open pit copper mine in the Andean region of Junín, Peru at ~4,500 m above sea level. Toromocho is one of the largest copper reserves in Peru and globally, with estimated reserves of 1.52 billion tonnes at an average head grade of approximately 0.48% Cu. The comminution circuit includes a primary gyratory crusher feeding a coarse ore stockpile, which supplies two SABC grinding circuits with different installed capacities (Phase I

and Phase II). Phase I comprises a 40’×26’ SAG mill (28 MW), two pebble crushers and two 28’×44’ ball mills (22 MW each). Phase II, commenced operations in 2020 and comprises a 36’×17’ SAG mill (13.5 MW), one pebble crusher and one 28’×44’ ball mill (22 MW). The target product size for both grinding lines is 80% passing (P80) 195 µm. The grinding product feeds the bulk flotation circuit. Bulk flotation consists of two rougher lines (Phase I and II), which are fed separately by each grinding line, respectively. The rougher concentrate is combined before feeding the bulk cleaner circuit. If the copper concentrate grade exceeds 20% Cu, concentrate from the first cell of the bulk cleaner circuit is bypassed to the final product stream. The remaining bulk cleaner cells' concentrates are sent to a bulk concentrate thickener, then to the regrind circuit, and finally cleaned in the cleaner circuit to produce the final concentrate. Figure 1 shows a simplified flowsheet of the Toromocho circuit. Figure 1 – Toromocho Simplified Comminution and Flotation Circuit Flowsheet (Phase I and Phase II) Toromocho commenced operations in 2013 (with Phase I) and was designed to process a nominal throughput of 5,250 tph. However, the ore hardness values used for plant design are significantly softer than those encountered during operation. This resulted in higher specific energy consumption, negatively impacting plant throughput and preventing the operation from reaching its design capacity. In response to the higher specific energy consumption, Toromocho adopted a proactive improvement strategy rather than accepting the limitations imposed by the original design. Feed size has a significant impact on the performance and capacity of comminution circuits. Therefore, the initial focus was drill and blast optimization tailored to feed ore characteristics (rock structure and strength) to produce finer ROM fragmentation and, consequently, a finer SAG mill feed. This was followed by the extension of optimization efforts downstream to ensure that the comminution circuit could fully capture the benefits of the improved feed size distribution. The structured drill and blast optimization program was developed and implemented at Toromocho in 2017. Key ROM fragmentation indicators including P80 and fines content (% passing 1 inch and % passing ½ inch) and show sustained improvement over time. The ROM P80

has reduced and fines content increased. These upstream improvements provided the foundation for a broader, integrated optimization effort across the concentrator. Building on this progress, the structured Mine-to-Process program progressed to optimization of comminution and flotation, supported by detailed surveys, ore characterization, modelling, simulation and throughput forecasting. Opportunities were identified to increase the total grinding throughput (from both grinding lines) by up to 14% but with some coarsening of the target product size. Integrated simulations across comminution and flotation indicated the increase in throughput and coarser grinding product size would, for a fixed concentrate grade, reduce overall copper recovery by ~1% if downstream constraints were not addressed. However, flotation debottlenecking and optimization identified quick-win and major CAPEX opportunities. Simulations indicate that the quick-win measures may improve overall copper recovery by ~1.5% while maintaining concentrate grade, with further gains possible via major CAPEX initiatives. Subsequent implementation of the quick-win initiatives has demonstrated the expected benefits. A summary of key activities and events is provided below: • In 2021, a significant increase in throughput (~10% increase compared to 2020) was achieved for the Phase I grinding line following the optimization study carried out in 2020. • Since early 2022, mechanical issues with the Phase I grinding line ball mills have constrained throughput. Chinalco is addressing these issues. • In May 2020, Phase II grinding line commenced operations (mean of 1,890 tph in 2021) • Since then, the throughput of the Phase II grinding line has steadily increased up to 2,220 tph. The design throughput (2,550 tph) has not been reached but should be achievable after implementation of all Hatch recommendations from the Phase II grinding line optimization study. • An audit and debottlenecking of the flotation circuit was carried out in parallel to the Phase II grinding circuit optimization study in May 2023. • Implementation of quick wins and several of the mid to longer-term initiatives has already delivered measurable improvements in throughput, operational stability, and metallurgical performance. This paper documents the structured methodology and practical implementation of the integrated Mine-to-Process strategy at Toromocho, from conceptual opportunity identification to implementation across drill and blast, comminution, and flotation. 2. METHODOLOGY The integrated optimization project at the Toromocho operation was executed using a structured Mine-to-Process methodology implemented progressively across the mining and processing value chain. The structured methodology takes an integrated approach, avoiding optimization of processes in isolation, which can result in suboptimal results for the overall operation. This approach considers the impact of variable feed ore characteristics on each stage (mining, comminution and flotation) as well as the interactions between each stage. The approach combined ore characterization, operational audits, plant surveys, analysis of historical operating data, development of semi-mechanistic and site-specific models, and

integrated simulations, with a strong emphasis on implementation feasibility and sustainability of results. 2.1 Integrated Mine-to-Process Optimization The optimization project was structured into sequential and interlinked stages, each building upon the outcomes of previous work: • Mine and ROM fragmentation optimization, including review and optimization of drill and blast practices. • Phase I grinding circuit assessment and optimization, supported by plant surveys, modelling, and throughput forecasting. • Phase II grinding circuit modelling and optimization, following commissioning of the second grinding line in 2020. • Flotation circuit auditing, modelling, and debottlenecking, to manage the downstream impacts of increased throughput and coarser grinding product size. • Implementation and validation, including prioritization of quick-win initiatives and definition of major CAPEX projects. This staged approach ensured technical consistency and progressive implementation. 2.2 Ore Characterization and ROM Fragmentation Ore hardness and breakage characteristics are key drivers of blasting and comminution performance. Available ore characterization data were consolidated and expanded over time, including Bond Ball Work Index (BWi), SAG Power Index (SPI), Drop Weight Index (DWi), and JK Axb parameters. During early optimization stages, where detailed test work was limited, validated Hatch inhouse comminution relationships were applied to derive representative breakage parameters. This was subsequently complemented by a comprehensive geometallurgical testing program, including SMC and Bond tests across the main alteration domains, confirming the estimated values and improving confidence in the model inputs. The results (Table 1) demonstrate the increased ore hardness compared to the values used for design, and thus the capacity limitations of the comminution circuits due to treating this harder material. Table 1 – Summary of Breakage Test Results Test Unit Design 2014 Samples 2019 Survey 2022 Geomet Testwork Bond Ball Work Index BWi (kWh/t) 12.9 13.3 14.3 13.2 14.2 14.2 SAG Power Index SPI (minutes) 75 129 120 99 - - Drop Weight Index DWi (kW/m3) 6.2* 8.9* 8.5* 7.5* 7.3* 7.0 *Estimated from Hatch in-house relationships. ROM fragmentation is a critical upstream variable influencing SAG mill performance. Optimizing the blast intensity based on ore characteristics (structure and strength) and adjusting crusher gap accordingly, can achieve a finer SAG mill feed and subsequently allow optimization of grinding circuit throughput. This optimization approach has been successfully applied in many

operations through mine-to-process optimization projects (Valery et al., 2019) and can be particularly effective when facing hard or highly variable feed ore characteristics. 2.3 Plant Audits and Surveys The downstream comminution and flotation circuits were then optimized, leveraging the benefits of finer ROM fragmentation from drill and blast optimization. This was conducted in phases as the second grinding line (Phase II) did not commence operations until May 2020. The modelling, simulation and optimization of Phase I grinding circuit was conducted in 2020 using historical operating data and results from surveys conducted previously by Chinalco on the 14th and 22nd of September 2019. A full audit and survey supervised by Hatch was conducted in May 2023 for the optimization of the Phase II grinding circuit and the Cu bulk flotation circuit. The audit and survey conducted in May 2023 also included measurement of blast fragmentation and additional rock characterization for the surveyed ores (Figure 2). This allowed comminution performance to be calibrated and linked with ore characteristics and blast fragmentation results. Figure 2 – Audit and Measurement of ROM Fragmentation (May 2023) As part of the survey, crash-stops and grind-outs were also carried out in the SAG and ball mills to gather key information to calibrate the comminution models. the condition of the lifters/liners and ball size distributions were reviewed. In the SAG mill, the extent of pegging and peening of the grates, the capacity of the pulp lifters, and the presence of slurry pooling were also investigated. Some examples of activities undertaken during the survey are shown in Figure 3. These surveys enabled calibration of site-specific models and ensured alignment between actual plant performance and simulated performance. Figure 3 – Sampling activities and equipment inspections during the survey

2.4 Mathematical Modelling and Simulations Site-specific models were developed in JKSimMet for both Phase I and Phase II grinding circuits (Figure 4 and 5). Model calibration was performed using survey data and validated against historical operating performance. To assess the downstream impacts of increased grinding throughput, a site-specific flotation model was developed in JKSimFloat (Figure 6). Integrated grinding and flotation simulations were performed to quantify tradeoffs between throughput, grind size, copper recovery, and concentrate grade. These simulations enabled identification of bottlenecks and evaluation of both operational and capital intensive debottlenecking strategies. Figure 4 – JKSimMet model flowsheet for Phase I Grinding circuit. Figure 5 – JKSimMet Model flowsheet for Phase II Grinding Line SAG Total Load (Exp) 25.0 Total Load (Calc) 25.0 CYC Units in Parallel 40 Pressure, kPa (Exp) 90.0 Pressure, kPa (Calc) 90.0 Ball Mill 2-Prod 7,093.0 7,139.8 0.6 0.6 62.8 62.8 6,580.2 6,625.2 TPH - Solids (Exp) TPH - Solids (Sim) P80 (Exp) P80 (Sim) % Solids (Exp) % Solids (Sim) Volume Flow (Exp) Volume Flow (Sim) Trommel Screen-U/S 4,476.0 4,476.0 4.0 3.5 71.1 70.7 3,319.8 3,353.4 Pebble Cruhser-Prod 653.0 726.9 16.6 19.3 97.6 99.8 235.1 245.4 CYC-O/F 4,476.0 4,475.9 0.2 0.2 29.7 29.5 12,082.2 12,178.5 CYC-U/F 14,186.7 14,279.5 1.1 1.2 80.4 80.4 8,225.8 8,264.1 Cyc Feed 18,662.7 18,755.4 0.8 0.8 57.1 57.0 20,307.9 20,442.6 Ball Mill 1-Prod 7,093.0 7,139.8 0.6 0.6 62.8 62.8 6,580.1 6,625.2 SAG-Prod 5,150.0 5,202.9 8.2 7.3 73.7 73.7 3,562.5 3,598.8 Feed 4,476.0 4,476.0 41.3 41.3 96.0 96.0 1,688.5 1,688.5 Trommel Screen-O/S 653.0 726.9 37.2 38.6 97.6 99.8 235.1 245.4 CYC Units in Parallel 14 Pressure, kPa (Exp) 88.00 Pressure, kPa (Calc) 87.14 Baseline Model Pebble Crusher Power (Calc) 539.92 Trommel1-O/S 429.18 97.31 169.65 46.93 SAG-Prod 2,408.18 73.10 1,771.54 8.25 Ball Mill-Prod 5,490.86 66.26 4,814.68 0.51 Sump-Prod 7,469.84 55.80 8,663.23 0.59 CYC-U/F 5,490.86 80.55 3,344.54 0.82 CYC-O/F 1,978.98 30.12 5,318.69 0.17 Trommel-U/S 1,979.00 62.76 1,901.90 1.45 SAG Feed 1,979.00 96.00 810.03 74.64 Pebble Crusher-Prod 429.18 97.31 169.65 13.71 TPH - Solids (Sim) % Solids (Sim) Volume Flow (Sim) P80 (Sim)

Figure 6 – Base Case Flotation Circuit Model 3. RESULTS AND DISCUSSION The aim of the integrated Mine-to-Process optimization project at Toromocho was to increase production considering the harder than anticipated feed ore characteristics. To maximize overall production, it is necessary to consider the impact on throughput and recovery performance, and as the grinding product size impacts flotation recovery, this is also an important factor. Drill and blast practices were reviewed and progressively optimized (Fangrong et al., 2023b). Since 2017, the ROM P80 was successfully reduced (from an average of 5.5 to 2.5 inches) and the amount of fines increased (% passing 1 inch increased from ~30% to ~50%) (Figure 7).

Figure 7 – Improvement in ROM fragmentation Calibrated site-specific models were used along with site observations and analysis of historical operating data to identify and evaluate opportunities in the grinding and flotation circuits considering the improved ROM fragmentation from drill and blast optimization. The Phase I grinding circuit was investigated first, as Phase II did not commence operation until 2020. The site-specific models were used to simulate possible circuit changes and optimization strategies. The simulations conducted for the Phase I grinding circuit included the following: • Sim 1 – Increase SAG mill pebble ports size and pulp lifter discharge capacity. • Sim 2 – Sim 1 + optimized SAG mill ball charge. • Sim 3 – Sim 2 + reduced the closed side setting (CSS) of both pebble crushers to 11 mm. • Sim 4 – Sim 3 + increased ball mill power. • Sim 5 – Sim 4 + wider primary crusher CSS. The simulations indicated the throughput of the Phase I grinding circuit could be increased up to 14% while maintaining the current product grind size (~P80 197 µm), as can be seen in Figure 8 which shows the cumulative benefits of all recommendations.

Figure 8 – Simulations Results Summary Subsequently, both the Phase II grinding line and flotation circuits were also investigated and the assessment integrated with the previous findings for the Phase I grinding line. Opportunities to increase throughput for the Phase II grinding line included increasing SAG mill speed and power, but this would likely coarsen product size. This coarsening could be mitigated by increasing the ball charge and reducing water addition to the ball mill. Further increases in throughput may be achieved by sending crushed pebbles to the ball mill, shifting load from the SAG to ball mill, but this would also coarsen product size. Thus, opportunities were identified to increase throughput for Phase II grinding line, but product size would coarsen. The combined increase in throughput and coarsening of product size from implementation of recommendations for both grinding lines (i.e. including the previous recommendations for the Phase I grinding line) could have a significant impact on flotation circuit performance. Integrated simulations of the grinding and flotation circuits were conducted to evaluate the overall impact, and these included the following: • Sim 0 – Baseline throughput, ball charge and SAG mill speed for Phase I and II. • Sim 1 – Phase II: Increase SAG mill speed and power. • Sim 2 – Phase II: Sim 1 + Increase in ball mill ball charge. • Sim 3 – Phase II: Sim 2 + Reduce ball mill water addition (increase ball mill slurry density). • Sim 4 – Phase II: Sim 3 + Sending crushed pebbles to ball mill (i.e., “pebble crush forward”). • Sim 5 – Phase I & II: Sim 4 + Phase I opportunities (i.e., optimized SAG mill ball charge + reduce pebble crusher CSS + increased ball mill power). The simulation results are summarized in Figure 9. Combining all opportunities for both grinding lines (Sim 5) could increase throughput by 14% but would coarsen the grinding product size. The combined effect of higher throughput and coarser size could reduce copper flotation recovery by ~1% for the same concentrate copper grade.

Figure 9 Integrating Grinding Circuit Opportunities with Flotation Circuit However, the flotation circuit analysis identified both operational and capital intensive opportunities to manage the impacts of increased throughput. Integrated simulations were conducted to evaluate the benefits of these strategies (Figure 10). Quick-win operational changes, including circuit reconfiguration, improved cell utilization, and full use of installed regrind and flotation capacity, could increase overall copper recovery by approximately 1.5% while maintaining concentrate grade. Thus, more than compensating for the expected loss from increased throughput and coarser grinding product size. Additional recovery gains could be achieved through major CAPEX initiatives, including expansion of cleaner circuit capacity and installation of more efficient flotation technology upstream of the existing cleaner banks. These options were shown to improve recovery under higher throughput scenarios, providing resilience against ore variability and future production increases. Figure 10 Simulation Summary for ‘Quick-win’ Opportunities and Major CAPEX Options In addition, Hatch used data collected during the optimization project to develop a power-based throughput forecast model for the Toromocho comminution circuit. The model calculates the

total circuit specific-energy and SAG mill specific-energy, based on ore hardness properties and the calculated SAG mill feed F80 (linked to the alteration domains). The specific energy is used to determine plant throughput for a given SAG milling and ball milling power. The throughput forecast model was calibrated and then validated using historical production data (March 2019 – September 2019 for calibration September 2019 –February 2020 for validation) and was supported by the mathematical process models developed during the optimization project. Despite limitations in ore characterization data available at the time, the model predicts throughput quite well on a weekly and monthly basis, with errors of 11% and 3.5%, respectively, at the 95% confidence level. Further improvements in model accuracy are expected with additional ore characterization. Accurate throughput forecast models are useful production tools that can assist in improving production reliability and strategic planning to maximize profit over the life-of-mine (LOM). 4. IMPLEMENTATION STATUS Chinalco has commenced implementation which is ongoing as some recommendations take longer to implement in a production environment. A summary of recommendations and implementation status for the Phase II grinding line and flotation circuit is provided in Table 2 and Table 3. Table 2 Summary List of Opportunities and their Status of Implementation for Phase II Grinding Circuit Opportunity Actions Benefits Implementation Status, % Maintain higher SAG mill speed more consistently. Maintain SAG mill speed at 76% C.S. Increase power draw and throughput ~ 6% - some coarsening of product size. Implemented. Increase and maintain ball mill ball charge. Increase and maintain the ball mill ball charge at 35 %vol and operate consistently at mill speeds of about 75 – 76% C.S. Increase power draw (~7%) and reduce P80 to mitigate the coarsening of grinding product size expected at higher throughput. Able to maintain ball mill ball charge at 33%vol. Include finer grinding media Change ball make-up from solely 3” balls to a 25/75 mix of 3”and 2.5” balls. Improve breakage efficiency. Implementing a 60/40 mix of 3”/2.5” balls. Currently running a test with high chrome balls at ball mill #3. Increase slurry residence time in ball mill. Reduce water addition to cyclone underflow. Improve grinding efficiency. Reduce P80 by about 10 µm and improve the current operating work index. Implemented but difficulty controlling the water addition. Shift load from the constrained SAG mill to the ball mill. Redirecting the crushed pebbles to the ball mill. Additional 7% increase in throughput - some coarsening of grinding product P80 (~ 13 - 18 µm). Planned to be implemented in 2028. Improve cyclone performance Maintain the current cyclone feed percent solids (~ 55%) and pressure (~ 90 kPa). To achieve good cyclone classification efficiency. In progress. Currently, it reducing the vortex diameters from 300 to 270 mm and increasing feed percent solids to cyclones from 55 to 62% for improve the circuit efficiency and copper recovery. Improve SAG mill performance. Implement curve discharge design at both SAG mills. • Increase throughput above 3%. Implemented: • New curved design is installed in March 2026. • Implemented: Currently running with 14 grates of 65 mm, and 4 65/70 mm. Reduce slot opening • Increase residence time at SAG mill, reducing slot opening from 65/70 to 65 mm. Increase the service life of the grates to match the shell service life. Increase the feed head outer plate thickness to match service life of grates. Table 3 Summary List of Opportunities and their Status of Implementation for Flotation Circuit Opportunity Actions Benefits Implementation Status, % Quick-Win Opportunities Reducing load in downstream cleaner bulk circuit (Sim A) By-pass first rougher cell concentrate of Phase II to bulk conc. thickener. Reduce the load on cleaner bulk circuit. ~1.3% increase in overall Cu recovery. Under Review.

Opportunity Actions Benefits Implementation Status, % Recovering copper lost to final tailings from Phase I rougher cells (Sim B) Evaluate Phase I rougher cells in terms of impeller/ stator mechanism, cell design, solids suspension, air dispersion and froth launder configuration. Improve the recovery of coarse copper particles in Phase I rougher cells. ~1.3% improvement in overall Cu recovery. Implemented radial launders to all Phase I rougher cells. Currently evaluating the improvement of rotor and stator design, and increasing radial launders at rougher cells phase 2. Fully utilize available regrind capacity (Sim C) Reduce target regrind P80 from 100 µm to ~ 90 µm Liberate locked copper in cleaner feed. ~1.5% increase in overall Cu recovery. Under Review. Increase the cleanerscavenger capacity (Sim D). Operate the existing 4 unused mechanical cells in parallel to the existing cleanerscavenger bank. Improve slow floating particle recovery in cleaner-scavenger and reduce the recirculating load to Phase II roughers. ~0.4% increase in Cu recovery. Implemented but with the additional cells in series to the existing cleaner-scavenger cells. Major CAPEX Required Increase cleaner bulk circuit capacity (Sim E) Installation of alternative flotation technology such as ERIEZ StackCell, Woodgrove SFR and DFR cells and Jameson Cell. Remove fast floating copper bearing minerals earlier to reduce the load on downstream cleaner circuits. ~1.5% increase in Cu recovery. Implemented – additional DFR cells added to circuit at Phase 2. Increasing cleaner column capacity (Sim F) Add extra flotation columns to the cleaner circuit. Increase residence time in cleaner circuit. ~1.6% increase in Cu recovery. Evaluating the possibility of improving cleaner column performance by using the Eriez new cavitation system. 5. CONCLUSION The integrated Mine-to-Process optimization program implemented at Toromocho demonstrates how a structured, system-wide strategy can be effectively translated into sustained operational results. By considering the impact of feed ore characteristics and addressing interactions across mining, comminution, and flotation rather than optimizing individual unit operations in isolation, the program enabled decision making focused on overall value generation (throughput and metal recovery) within real operating constraints. Site-specific modelling and integrated comminution and flotation simulations quantified both the potential benefits and the limitations of identified improvement opportunities. The work identified opportunities to increase grinding throughput by up to ~14% across the two comminution circuits despite increased ore hardness. However, the analyses also showed that comminution driven throughput gains must be evaluated in the context of downstream flotation constraints to avoid unintended recovery penalties (in the order of ~1% at fixed concentrate grade if constraints are not addressed). The integrated approach provided the basis for a balanced optimization strategy that preserves concentrate quality while maximizing overall copper production (throughput and recovery). Consistent efforts ensured the progression from study to implementation through phased execution. Implementation of quick-win and some of the mid to longer-term initiatives, spanning drill and blast practices, operating adjustments across crushing and grinding, and flotation circuit debottlenecking and optimization, have already delivered measurable improvements in throughput, operational stability, and metallurgical performance. Additional recovery gains can be achieved through prioritized CAPEX upgrades to sustain performance at higher throughput conditions, and these are ongoing. The Toromocho case demonstrates the benefits of structured Mine-to-Process optimization underpinned by reliable data, disciplined modelling, integrated simulation assessment, crossfunctional coordination, and structured implementation. Consideration of ore characteristics and the interactions of each stage of the operation ensure improvement of the overall operation by maximizing overall production (throughput and recovery) while maintaining concentrate quality and minimizing costs.

REFERENCES Fangrong, D., Yang, D., Muñoz, A., Valery, W., Valle, R., Tabosa, E., Duffy, K., Siong, J., Plasencia, C., Gonzales B. (2025) ‘Mine to Flotation Process Optimization at Chinalco Toromocho Operation – Part II’, ProceminGeomet 2025, 6-8 August 2025, Santiago, Chile. Fangrong, D., Yang, D., Muñoz, A., Valery, W., Valle, R., Tabosa, E., Siong, J., Plasencia, C. (2023a) ‘Mine to Flotation Process Optimization at Chinalco Toromocho Operation’, Procemin-Geomet 2023, 4-6 October 2023, Santiago, Chile. Fangrong, D., Yang, D., Muñoz, A., Valery, W., Valle, R., Bonfils, B., Plasencia, C. (2023b) ‘Integrated Mineto-Mill Optimization of Toromocho Operation at Minera Chinalco Peru’ SAG2023 - International Conference on Autogenous and Semiautogenous Grinding Technology, Vancouver, Canada. Morrell, S. & Valery, W. (2022) ‘Measuring and Modelling the Impact of Primary Crusher Stockpiles on Autogenous/Semi-Autogenous Mill Feed Size’, Minerals Engineering, Volume 190. Fangrong, D., Yang, D., Muñoz, A., Valery, W., Valle, R., Plasencia, C. (2022), ‘Optimización Integrada Mina-Planta en Toromocho, Minera Chinalco Perú S.A’, 6to Congreso APC, Lima, Perú. Valery, W., Duffy, K., & Jankovic, A. (2019) Chapter 3.1 - Mine-to-Mill Optimization. In Dunne, R., Kawatra, S., Young, C (Eds.) SME Mineral Processing & Extractive Metallurgy Handbook (pp. 335-346). Englewood, CO: Society for Mining, Metallurgy and Exploration (SME). Kanchibotla, S., Valery, W. & Morrell, S. (1999) ‘Modelling Fines in Blast Fragmentation and its Impact on Crushing and Grinding’, Explo 1999: A Conference on Rock Breaking, Kalgoorlie, Australia.

INTEGRATED ARSENIC REMOVAL AND STABILIZATION FROM MINING STREAMS: COMPARATIVE EVALUATION OF PYROLYSIS AND PRECIPITATION APPROACHES USING THE GLASSLOCK PROCESS™ O. Sanfaçon1, B. Johnson1, J. Tardif1, J.P. Mai1, A. Drouin1, *A. Maltais1, A. Landry1 Dundee Sustainable Technologies Inc., Canada (*Presenting author: amaltais@dundeetechnologies.com) ABSTRACT Arsenic contamination in mining residues and process streams is a critical environmental and operational challenge for the copper and gold industries. This article presents a comparative study of two advanced arsenic removal and stabilization approaches—pyrolysis and precipitation—applied to solid concentrates and acidic process streams, respectively. Both methods utilize the GlassLock Process™ for permanent arsenic immobilization. Laboratory results, process flowsheets and engineering designs are presented, demonstrating the technical and economic viability of these solutions for sustainable mining. KEYWORDS GlassLock Process™, arsenic vitrification, arsenic removal, arsenic stabilization, copper concentrate, arsenic precipitation, enargite. 1. INTRODUCTION The presence of arsenic in copper and gold mining residues and process streams poses significant environmental risks and economic penalties. Traditional management strategies often fail to provide permanent stabilization, leading to long-term liabilities. This study compares two innovative arsenic management approaches—pyrolysis for solid concentrates and iron-based precipitation for acidic streams—each followed by vitrification using the GlassLock Process™. The work integrates laboratory data and process engineering from two major projects with mining partners, highlighting the capability of vitrification to stabilize arsenic from concentrates and process waste.

2. MATERIALS AND METHODS 2.1 Mineralogical Samples & Preparation Materials were sampled and provided to DST by the mining companies involved in the study. A total of two distinct samples. In this study, the following nomenclature is used: 1. EC1 = Enargite flotation concentrate (solids) 2. SB1 = Arsenic-bearing acidic solution from smelter scrubber blowdown. (aqueous) EC1 Sample was thoroughly mixed, dried overnight at 50°C and split using a baffled sample divider to assure representativity; No further grinding was performed. SB1 sample was fully homogeneous upon reception. 2.2 Pyrolysis Test Methodology In a Lindberg tube electric furnace equipped with a 4inches diameter steel tube and dustcollecting filters, pyrolysis tests were performed on the Enargite concentrate (EC1) to remove and capture arsenic as As₂S₃, leaving an arsenic-depleted calcine. Schematic representation of the complete set-up is shown in Figure 1. Tray Electric Lindberg furnace Thermocouple SO2 To ATM Bag filter To on-line gas analyzer Vacuum pump FI Regulator and Flowmeter Screen filter Cooling zone Figure 1 – Pyrolysis work installation with Lindberg furnace and dust-collectors Tests were conducted under SO2 atmosphere on 40g sub-samples at a peak temperature of 650–675°C for 80min, starting when the thermocouple read 600°C. An on-line gas analyzer was linked to the circuit after the furnace to determine oxygen levels post-pyrolysis, indicative of the airtightness of the set-up (<1% for a functioning system). The As₂S₃ by-product collected in the cooling zone and dust filters was oxidated to As2O3, using a similar set-up using a glass tube under air and vacuum. 2.3 Iron-based Arsenic Co-Precipitation Methodology Iron-based co-precipitation was conducted on the scrubber blowdown sample (AS1) to precipitate arsenic as iron-arsenic oxyhydroxide, suitable for vitrification. Tests were performed on sub-samples of 500mL, in a 1L beaker with magnetic stirrer (360 RPM) at room temperature and ambient pressure under a fume hood. A fixed quantity of FeCl3,

based on a ratio of iron to arsenic of 1, was added to AS1 until complete dissolution. MgO was then added to the beaker until reaching pH 6, before stirring the mixture for 6 hours. Afterwards, the mixture was filtered using a Büchner funnel assembly under vacuum and Whatman #3 filter paper. 2.4 Glass Formulation Testing Methodology Both glass products were formulated using ACS grade reagents, following the dry-basis recipes described in Table 1. Table 1 – Dry mix composition Raw Material Units Glass 1 Glass 2 From EC1 Pyrolysis Byproduct From AS1 Precipitation Byproduct As2O3 (Flue dust) %wt. dry 26.6 -- As2O3 (Precipitate) %wt. dry -- 45.6 Hematite %wt. dry 14.5 4.10 Sodium Carbonate %wt. dry 16.0 21.1 Silica Sand %wt. dry 38.9 29.2 Alumine %wt. dry 4.7 -- The glass was melted in a Katanax fluxer at 1,200°C for 3 hours using an alumina crucible with 20g of dry mix per glass. 2.5 Arsenic Glass Environmental Characterization – EPA 1311 The US EPA TCLP (Method 1311) is used to simulate contaminant leaching resulting from solid waste in landfills. The appropriate TCLP extraction fluid for the glass product is diluted glacial acetic acid buffered at pH = 4.93 ± 0.05 with sodium hydroxide. Using crushed and sieved pieces glass of sizes between 1.0mm and 9.3mm. For this project, TCLP leaching tests were performed in a specific L/S ratio of 20:1 for a period of 18 hours in 500mL HDPE bottles. The bottle vessels are rotated at 30 ± 2 RPM on a Rotary Agitation Apparatus as shown in Figure 1. After the leaching step, the mixture is filtrated and analyzed for contaminants. (USEPA, 1992). 2.6 Analytical Methods Base metals determination on solids was achieved by di-acid digestion / Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES) finish (Perkin Elmer Optima 7300DV). Al, Ca, Fe, Mg and Si contents in solids, as well as Arsenic content in glass were determined by sodium alkaline fusion followed by ICP-OES finish. Gold determination was achieved using crucible fusion / Pb cupellation fire assaying / ICP-OES finish. Total sulfur (LECO S230SH) and total carbon (LECO CS230SH) were determined using elemental combustion analyzers. Particle size distribution (PSD) was determined by laser diffraction particle size analyzer (Microtrac).

3. RESULTS 3.1 Pyrolysis Approach on EC1 Results Successful removal of the arsenic was demonstrated using pyrolysis treatment the EC1, achieving 99.6% As removal. Table 2 summarizes the main contents of the concentrate before and after pyrolysis. This approach produced a residue which is depleted in arsenic and has a higher valuable metals grade. Table 2 – Pyrolysis approach characterization and results Sample Sample Type As Cu Stotal Au As removal Mass loss %wt. %wt %wt g/t % % EC1 Enargite Concentrate 15.1 45.5 36.4 6.10 99.6 34,2 EC1-PYR Pyrolysed EC1 0.11 63.0 25.3 15.2 After 400g of concentrate was pyrolyzed, a mass of 60.7g of As₂S₃ was collected in the dust-collectors. The by-product contained 38.6 %wt. As and little to no other base metals contamination (less than 1 %wt. total). This As₂S₃ by-product was then oxidized entirely to form 41.05g of As₂O₃, for which composition is shown in Table 3. This As₂O₃ by-product was then directly amenable to vitrification as described in sections 2.4 and 3.3. Table 3 – Arsenic trioxide (pyrolysis by-product) characterization. Sample Sample Type As content Cu Stotal %wt. (%wt) (%wt) EC1-PYR-AS2O3 Arsenic trioxide from pyrolysis 72.85 ≤40 0.1 3.2 Precipitation Approach on SB1 Results Iron-based precipitation at pH 6 was highly effective for arsenic removal. As shown in Table 4, it allowed to deplete arsenic from the acidic solution. Table 4 – Initial sample and cleaned solution characterization. Sample Sample Type As As removal FeCl addition MgO addition mg/L % g/L g/L SB1 Arsenic-bearing acid 8,620 99.9 18.07 17.0 SB1-Clean Neutralized SB1 7.14

This test allowed to recover 28.1g (dry) of arsenic precipitate. Its characterization is shown in Table 5 below. Table 5 – Arsenic precipitate characterization. Sample Sample Type As content Cu Fe Mg Zn %wt. %wt. %wt %wt %wt SB1-PRE-AS Arsenic precipitate from SB1 16.9 4.39 11.6 13.9 5.65 This precipitate was directly amenable to vitrification as described in the following section. 3.3 GlassLock Vitrification: Product Quality and Environmental Compliance Both processing by-products (EC1 trioxide by-product and SB1 precipitate) were successfully vitrified. Arsenic in glass results and TCLP results are shown in Table 6. Table 6 – Glass formulation results Value Units Glass 1 Glass 2 From EC1 Pyrolysis Byproduct From SB1 Precipitation Byproduct TCLP result mg As/L 3.80 3.02 As in Glass %wt. 17.7 7.96 4. COMPARATIVE DISCUSSION: PYROLYSIS VS. PRECIPITATION 4.1 Process Comparison Table Table 7 – Process Comparison Table Aspect Pyrolysis approach Precipitation approach Feed Type Solid concentrate Acidic process solution Main Reagents (Arsenic capture) SO₂ (pyrolysis), O₂ (oxidation) FeCl₃, MgO As Removal (%) 99.3–99.6 99.9 By-product As₂S₃ (then As₂O₃) Fe-As oxyhydroxide Vitrification Feed As₂O₃ Fe-As precipitate Glass As (%wt) 15–18 8 TCLP As (mg/L) 3.3–3.8 3.02 Metal Recovery Cu, Au, Ag enriched in calcine Not applicable (solution treated) Scalability Proven at pilot scale Designed for full-scale (600 m³/d)

4.2 Process Flow Diagrams Below are Process Flow Diagrams (PFDs) illustrating the recovery of arsenic following the pyrometallurgical treatment of ore. The PFDs demonstrate that arsenic can be recovered in several forms. Arsenic may be recovered as a solid using a dust collector, as shown in Figure 2; it may be recovered as a slurry with a wet electrostatic precipitator, as shown in Figure 3; or it may be recovered dissolved in an acidic liquid within a scrubber, as shown in Figure 4. Subsequently, arsenic recovered as either a solid or a slurry is directly conveyed to the GlassLock process to produce an inert, non-leachable glass, as depicted in Figure 5. FLUIDIZING AIR OFF GAS FLUIDIZED BED PYROLYSIS POST-COMBUSTION CHAMBER AIR SO2 (g) As2O3 (s) As2S2 (g) ARSENIC-FREE CONCENTRATE COOLING SCREW As2O3 (s) O/F U/F 2 3 TO MARKET / LEACHING PROCESS 5.3 % As CONCENTRATE CYCLONE ARSENIC BAG HOUSE 1 TO VITRIFICATION PLANT COOLING WATER 4 TO ATMOSPHERE GYPSUM TO LANDFILL LIME SLURRY TANK LIME SILO OFFLOAD TRUCK SO2 SCRUBBER GYPSUM RETENTION TANK TO EFFLUENTS PROCESS WATER WATER WASTE HEAT BOILER OFF GAS OFF GAS CYCLONE CYCLONE QUENCH TO BAGHOUSE WATER GAS SPRAY COOLER As2O3 (s) STEAM SO2 (g) As2O3 (g) SO2 (g) As2O3 (g) Dust Trap Dust Trap Figure 2 – Integrated flowsheet: pyrolysis + arsenic recovery with dust collector

FLUIDIZING AIR OFF GAS FLUIDIZED BED PYROLYSIS POST-COMBUSTION CHAMBER AIR SO2 (g) As2O3 (s) As2S2 (g) ARSENIC-FREE CONCENTRATE COOLING SCREW As2O3 (s) O/F U/F 3 TO MARKET / LEACHING PROCESS 5.3 % As CONCENTRATE CYCLONE 1 TO VITRIFICATION PLANT COOLING WATER 4 TO ATMOSPHERE GYPSUM TO LANDFILL LIME SLURRY TANK LIME SILO OFFLOAD TRUCK SO2 SCRUBBER GYPSUM RETENTION TANK TO EFFLUENTS PROCESS WATER WATER WASTE HEAT BOILER OFF GAS OFF GAS CYCLONE CYCLONE TO BAGHOUSE WATER GAS SPRAY COOLER As2O3 (s) STEAM SO2 (g) As2O3 (g) SO2 (g) As2O3 (g) Water WESP ARSENIC SLURRY TANK FILTER PRESS FILTER PRESS As2O3 (s) 2 BLEED Dust Trap Dust Trap Figure 3 – Integrated flowsheet: pyrolysis + arsenic recovery with wet electrostatic precipitator FLUIDIZING AIR OFF GAS FLUIDIZED BED PYROLYSIS POST-COMBUSTION CHAMBER AIR SO2 (g) As2O3 (s) As2S2 (g) ARSENIC-FREE CONCENTRATE COOLING SCREW O/F U/F 2 TO MARKET / LEACHING PROCESS 5.3 % As CONCENTRATE CYCLONE 1 COOLING WATER TO ATMOSPHERE GYPSUM TO LANDFILL LIME SLURRY TANK LIME SILO OFFLOAD TRUCK SO2 SCRUBBER GYPSUM RETENTION TANK TO EFFLUENTS PROCESS WATER WATER WASTE HEAT BOILER OFF GAS OFF GAS CYCLONE CYCLONE TO BAGHOUSE WATER GAS SPRAY COOLER As2O3 (s) STEAM SO2 (g) As2O3 (g) SO2 (g) As2O3 (g) ARSENIC SCRUBBER QUENCH SO2 (g) Dust Trap SCRUBBER BLOWDOWN TO ARSENIC PRECIPITATION 3 TO VITRIFICATION PLANT WEAK ACID 7-8% H2SO4 4 Dust Trap Figure 4 – Integrated flowsheet: pyrolysis + arsenic recovery with scrubber

FILTER PRESS GYPSUM AGITATED TANK FILTRATE COMPRESSED AIR ARSENIC SCRUBBER BLOWDOWN CaOH2 OFFLOAD TRUCK WEIGHING BIN HOLDING TANK COMPRESSED AIR TO As PRECIPITATION CIRCUIT ARSENIC FROM ACID REGENERATION CIRCUIT Figure 6 – Integrated flowsheet: acid solution concentration reduction FILTER PRESS TO VITRIFICATION PLANT PRECIPITATION SLURRY TANK FILTRATE HOLDING TANK COMPRESSED AIR As-CONTAMINATED ACID FeCl3 OFFLOAD TRUCK WEIGHING BIN P-92 DILUTION UNIT HOLDING TANK FLOCULANT DOSING SYSTEM STATIC MIXER CALIBRATION COLUMN MgO OFFLOAD TRUCK TREATED SOLUTION TO EFFLUENT THICKENER OVERFLOW P-87 WEIGHING BIN FLOCULANT CONCENTRATE SOLUTION WATER HOLDING TANK COMPRESSED AIR Figure 7 – Integrated flowsheet: arsenic precipitation

TO ATM BATCH PLANT BAGHOUSE TO DRYER FURNACE BAGHOUSE TO DUST BIN E-22 MIXER BRIQUETTER Weighing Bin #1 COMPRESSED AIR NATURAL GAS OFF GAS DRYER NATURAL GAS ARSENIC GLASS TO STORAGE OXYGEN FURNACE OFF GAS DRY BRIQUETTES E-27 FURNACE HOOD DUST BIN FROM BAGHOUSES SILICA TAILINGS SILO SUPPLY HEMATITE TAILINGS SILO SUPPLY Weighing Bin #2 Weighing Bin #3 Weighing Bin #4 FURNACE BAGHOUSE OFF GAS SODIUM CARBONATE SILO SUPPLY WATER FROM ARSENIC PRECIPITATION CIRCUIT 1 2 3 4 5 Figure 8 – Integrated flowsheet: GlassLock 4.3 Environmental and Economic Considerations Both approaches produce a stable, non-hazardous glass, eliminating long-term arsenic liability. Pyrolysis is best suited for solid residues with high arsenic and valuable metals, while precipitation is ideal for large-volume, low-solid, arsenic-rich process streams. Both processes are compatible with existing mining infrastructure and can be integrated into closure or ongoing operations. 5. CONCLUSIONS Both pyrolysis and precipitation, when coupled with the GlassLock Process™, achieve >99% arsenic removal and produce environmentally compliant, stable glass products. Pyrolysis is optimal for solid, arsenic-rich concentrates, enabling further metal recovery. Precipitation is highly effective for treating large volumes of arsenic-laden process solutions. The GlassLock Process™ is scalable and provides a permanent solution to arsenic management in mining.

REFERENCES DST-54182-LaboratoryTestProgram-REPORT-Feb2025_V1.pdf. DST-54162- Lab. & Eng. Program_FINAL REPORT_SEPT 2025_Rev1.pdf. Method 1311, Revision 0, July 1992, Final Update I to the Third Edition of the Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, EPA publication SW-846. Additional references as cited in the original reports.

DIGITAL TRANSFORMATION OF THE MINING DRILLING PROCESS: MULTIBRAND DATA INTEGRATION, REAL-TIME MONITORING, TARGETED REPORTING, AND ARTIFICIAL INTELLIGENCE APPLICATION L.A. Messa1, F.J. Guerra2 1OT-TICAR Department Cuajone Mine – Southern Peru Copper Corporation ABSTRACT The growing global demand for strategic minerals required by the energy transition challenges the mining industry to deliver resources faster, with greater operational intelligence and under increasingly stringent standards of responsibility and transparency. Drilling represents a critical leverage point within the mining value chain, as it directly influences fragmentation quality, downstream comminution energy consumption, and overall operational stability. Yet in many large-scale operations, high-resolution control system data remains underutilized due to interoperability barriers, technological heterogeneity, and field connectivity constraints. This paper presents the design and implementation of an interoperable and scalable digital model for production drilling in a large-scale copper operation, integrating heterogeneous controllers and multi-vendor fleets through open industrial standards and an edge-computing architecture incorporating store-and-forward mechanisms to ensure data integrity during communication disruptions. The proposed architecture enables real-time acquisition of operational variables, centralized storage, online monitoring, per-hole traceability, and advanced analytics focused on productivity and energy efficiency. A locally deployed AI-assisted query layer is introduced to enable natural-language interaction with historical and real-time operational data while maintaining industrial cybersecurity requirements. Results demonstrate improved data availability, reduced decision latency, and enhanced ability to correlate specific energy per drilled meter with geological domains and operational performance. The principal contribution is a replicable framework that transforms raw control signals into actionable operational intelligence, strengthening data governance, technical efficiency, and operational transparency. Such capabilities constitute essential enablers for a more productive, predictable, and responsible mining sector capable of meeting the global challenge of supplying critical minerals. KEYWORDS Digital Mining, Drilling Optimization, Industrial Interoperability, Real-Time Operational Intelligence, Edge Computing, Artificial Intelligence Applications

1. INTRODUCTION The accelerating global demand for critical minerals, essential for electrification, renewable energy systems, and large-scale decarbonization, poses a structural challenge to the mining industry: how to expand production while simultaneously advancing efficiency, transparency, and environmental stewardship. Copper and other strategic minerals stand at the core of the energy transition. Yet increasing their supply is no longer driven solely by new discoveries or capital-intensive expansion projects. It increasingly depends on the industry’s capacity to reduce operational variability, enhance process predictability, and transform the vast volumes of operational data generated across the value chain into actionable intelligence. Within this context, production drilling constitutes one of the most influential technical leverage points in open-pit mining operations. Drilling performance directly shapes blast outcomes, determines rock fragmentation quality, influences downstream comminution energy intensity, and affects the overall stability and cost structure of the operation. Apparent minor deviations in parameters such as penetration rate, pulldown force, or torque can cascade through subsequent stages, amplifying energy consumption, decreasing predictability, and generating compounding efficiency losses across the production chain. Consequently, leading mining operations worldwide are intensifying their focus on productivity optimization, cost discipline, environmental performance, and operational resilience. Digital transformation has emerged as a strategic enabler of these objectives. Among all production processes, drilling holds strategic relevance because it fundamentally determines: • Blast design effectiveness • Fragmentation distribution and quality • Loading and hauling efficiency • Crusher throughput and specific energy consumption Despite its importance, drilling operations in many large-scale mines still rely on partial automation and manual reporting. Operational data generated by modern drilling rigs often remain underutilized due to: • Heterogeneous equipment manufacturers • Proprietary communication protocols • Lack of unified connectivity infrastructure • Limited interoperability between control and enterprise systems This paper presents a replicable model that transforms operational data into structured intelligence, strengthening productivity, predictability, and accountability in the supply of critical minerals in the face of current global challenges that transforms drilling into a connected, datadriven process aligned with Mining 4.0 principles. 2. STATE OF THE ART IN DIGITAL DRILLING The adoption of digital technologies in mining has advanced rapidly over the past decade, driven by the need to improve productivity, safety, energy efficiency, and sustainability. Within the

Made with FlippingBook

RkJQdWJsaXNoZXIy MTM0Mzk2