Conference Agenda

In recent years, the Peruvian insurance sector has undergone significant transformation driven by digitalization and the increasing demand for more personalized services. However, micro and small enterprises (MYPEs) in the industry have faced limitations in policy management, reflected in a high rate of defects in the renewal process and a rise in customer attrition. This issue reduced operational efficiency and policyholder loyalty, directly affecting business competitiveness. In this context, the present research proposed an improvement model based on the Random Forest method and the Total Quality Management (TQM) philosophy under the PDCA approach. The model aimed to predict the probability of non-renewal and implement corrective actions oriented toward customer retention. Recent studies have demonstrated that combining predictive models with continuous improvement strategies can increase customer retention by 7% to 15%, while simultaneously optimizing service quality and managerial decision-making. It was concluded that the integration of predictive analytics and quality management constitutes an effective and scalable strategy to optimize service continuity in Peruvian insurance companies.

One of the main problems of Total Productive Maintenance (TPM) is its superficial implementation, focused on basic routines rather than on data analysis, resulting in limited improvements in equipment availability and Overall Equipment Effectiveness (OEE), and a strong dependence on the human factor. For this reason, a systematic literature review of 70 studies was conducted, using Scopus, Web of Science, SciELO, Dialnet, and Google Scholar as primary databases, in order to rigorously analyze how the integration of Industry 4.0 technologies, such as Artificial Intelligence, IoT, and digital twins, strengthens Total Productive Maintenance (TPM) through the measurement, prediction, and optimization of equipment performance. Using the PRISMA methodology, recent scientific evidence from 2020 to 2025 is synthesized, demonstrating that the incorporation of data analytics, neural networks, digital twins, and IoT enables a transition from reactive maintenance to predictive and intelligent maintenance. The study identifies critical gaps related to standardization, organizational change management, and the scarcity of failure data, providing clear guidelines for future implementations and applied research lines. Finally, key aspects that should be strengthened and those requiring improvement are discussed in order to achieve an effective integration of AI, sensor-based systems, and risk models that support more accurate, culturally viable, and sustainable asset management over time.

Gold mining is a fundamental activity for the Peruvian economy; however, the progressive depletion of high-grade ore forces mining companies to seek innovative alternatives to maximize the extraction of residual metal from previously exploited resources. One such alternative is the Injection Leaching (IL) method, a technique based on the pressurized injection of cyanide solution directly into exhausted heap leach pads to enhance gold recovery. This research evaluated the implementation of Injection Leaching as a re-leaching strategy in the Newmont Yanacocha Mining Unit, specifically in the Carachugo 10 (CA10), La Quinua 8 (LQ8) and La Quinua 1-7 (LQ1-7) leach pads. The study was carried out through a pilot test supported by geophysical monitoring using electrical resistivity, geodetection systems and nuclear magnetic resonance (NMR), which allowed understanding the distribution and behavior of cyanide solution inside the heap. The results showed a rapid increase in the gold concentration in the pregnant solution, stabilizing between 0.5 and 0.6 ppm in CA10 (Figure 15), similar to fresh ore conditions. Additionally, the project generated approximately 350,000 additional ounces of gold, extending the mine’s productive life. Therefore, Injection Leaching proved to be a technically, economically, and environmentally viable method for the recovery of remnant gold, overcoming the limitations of conventional drip and sprinkler leaching systems.

Abstract- Determining the optimal replacement timing for high-CAPEX mining equipment under operational uncertainty remains a strategic challenge in capital-intensive industries. While deterministic life cycle cost (LCC) models provide baseline estimates, they fail to capture the stochastic nature of maintenance degradation patterns across different operational regimes. This research develops a multi-scenario stochastic framework integrating Monte Carlo simulation (30,000 realizations) with Equivalent Uniform Annual Cost (EUAC) optimization to quantify the economic life of ultra-class haul trucks under three operational contexts: World-Class operations (maintenance degradation rate g = 8% ± 2%), Industry Standard (g = 10% ± 2%), and Severe Conditions (g = 12% ± 2%). The methodology employs geometric gradient modeling with structural constraint g_max = 15% based on physical wear limits documented in tribological studies. Results demonstrate that the technical driver (g) exhibits dominant influence over the financial driver (TMAR): a 2-percentage-point increase in g contracts economic life by 25% (from 16 to 12 years), translating to earlier CAPEX reinvestment of $3.5M per unit. Under Severe Conditions, optimal replacement converges to 11 years. Sensitivity analysis via Power BI visual analytics reveals that maintenance strategy transitions yield higher economic leverage than financial parameter adjustments. The contribution establishes a Decision Support System (DSS) architecture that operationalizes uncertainty into actionable replacement policies.

High energy consumption during the comminution stage is one of the main challenges to the operational efficiency of Peruvian mining. In this context, this research focused on optimizing the comminution circuit at Compañía Minera Antamina S.A., located in the district of San Marcos, province of Huari, region of Áncash, Peru. The purpose of the study was to theoretically analyze the reduction in energy consumption and the operational benefits derived from the implementation of emerging technologies, such as machine learning, advanced process control, and high-pressure grinding (HPGR). The research was applied, descriptive-comparative in nature, and employed a non-experimental, cross-sectional design. It was based on a documentary analysis of technical reports and recent specialized literature (2019–2024). The results showed that modernizing the grinding circuit reduced specific energy consumption from 14 to 10.5 kWh/t, equivalent to an approximate 25% savings, and increased throughput from 2,750 to 4,400 tons per hour. These improvements demonstrated more efficient use of installed power and a reduction in the process's environmental footprint. In conclusion, optimizing the comminution circuit at Antamina represents an effective strategy for reducing energy costs, improving productivity, and strengthening the sustainability of the mining operation.

Keywords- Comminution, energy efficiency, machine learning, Antamina, sustainability.

This study analyzes unproductive time in the hauling process of an open-pit mine using a supervised machine learning approach based on simple linear regression to optimize production per shift. The dataset initially consisted of 10,000 operational records collected from 55 Volvo FMX 540 haul trucks operating at the Summa Gold mining unit in Peru; after exploratory data analysis and outlier removal, 9,000 records were retained for model development. A quantitative and applied correlational methodology was adopted, using 90% of the data for training and 10% for validation. The linear regression model relates effective operating hours to production per shift and achieved a correlation coefficient of 0.998 and a coefficient of determination (R²) of 0.997. The results demonstrate a strong relationship between operational availability and hauled tonnage, allowing reliable prediction of shift production and supporting operational decision-making. The proposed model contributes to reducing unproductive time and enables the operation to reach a production target of 30,000 tons per shift, improving efficiency in small-scale open-pit mining operations.