Development of an Integrated Framework for Optimizing Mine Drainage Systems: Utilizing Numerical Modeling, Decision-Making and Machine Learning Applications

Mine drainage remains a critical challenge in ensuring the safety and sustainability of mining operations, as it is often complicated by complex subsurface flow behaviors and mechanical stress interactions. This study proposes an integrated three-phase framework for analyzing and optimizing drainage systems at the Angouran lead–zinc mine. In the first phase, the hydro-mechanical behavior of the rock mass was simulated using UDEC software, demonstrating that increased normal stress reduces fracture aperture and permeability. The simulated pore pressure (۴.۵×۱۰⁵ Pa) closely matched the field measurements (۴.۴×۱۰⁵ Pa), with only a ۲.۲% deviation. In the second phase, a multi-criteria decision-making approach using the Analytic Hierarchy Process (AHP) and input from ۳۲ domain experts identified Q۴ (very high quality) and Q۲ (medium quality) indicators as the most influential criteria. In the third phase, three machine learning models—linear regression, polynomial regression, and artificial neural networks (ANNs)—were trained on piezometric data to predict water discharge. The ANN model outperformed the other models, achieving an R² of ۰.۹۴ and RMSE of ۰.۱۸, effectively capturing the nonlinear dynamics of groundwater flow within the mine. The findings highlight that the integration of numerical modeling, expert-based decision analysis, and AI-driven prediction provides a robust and innovative approach for designing and managing mine dewatering systems, with potential applicability to other complex hydrogeological environments.

کلیدواژه ها:

Mine Drainage Management ، Numerical Modeling ، Machine learning ، Decision-Making Frameworks ، Sustainable Water Treatment