AI FOR LEACHATE DATA PROCESSING AND BARREL TEST GEOCHEMISTRY *L. Barba 1, G. Mallén 2, M. Téllez 3, C. Tueros4, L. Casadiegos5 1 CEOl/ Water & Environmental Services Perú, Peru, 2 Senior Specialist in Hydrochemistry and Isotopy/ Water & Environmental Services Perú, Peru 3 COO/ Water & Environmental Services Perú, Peru 4 Specialist in Hidrogeology and Data Science/ Water & Environmental Services Perú, Peru 5 Specialist in Hidrogeology and Hydrochemistry/ Water & Environmental Services Perú, Peru (*Presenting author: luisa.casadiegos@wes.com.pe) ABSTRACT In mining operations, barrel-based geochemical tests enable the simulation of long-term leaching processes and the assessment of acid rock drainage (ARD) generation potential and metal leaching from waste rock, tailings, and heap materials. However, prolonged monitoring produces large volumes of heterogeneous data, whose manual processing is time-consuming, fragmented, and prone to error, thereby limiting their timely use in environmental and operational decision-making. This paper presents a technology-driven solution based on artificial intelligence for the automated processing, analysis, and prediction of geochemical data. The proposed solution integrates automated ETL workflows, data ingestion agents, and analytical and predictive models, enabling data standardization, anomaly detection, and forecasting of the geochemical behavior of mining materials. Implemented in a real operational environment with more than 15 years of historical data, the platform reduced processing and reporting time by over 60%, shortening cycles from weeks to hours. Furthermore, it enabled early identification of imbalances between sulfide and neutralizing components, improving the assessment of acid rock drainage risk and optimizing the planning of mitigation measures. The results demonstrate that automation and the application of predictive models not only enhance operational efficiency but also transform geochemical management toward a predictive and preventive approach. This approach is scalable and replicable across different mining contexts, contributing to more efficient and sustainable environmental management. KEYWORDS Data Analytics, Geochemistry, Leachate, Acid Rock Drainage (ARD), ETL, Bots, DecisionMaking
RkJQdWJsaXNoZXIy MTM0Mzk2