Discover the Power of Innovation in Mining
Welcome to my page, where I share my expertise in artificial intelligence, machine learning, optimization, and financial modeling within the mining sector. My work aims to bridge advanced technological solutions with practical applications, enhancing efficiency and fostering innovation in mining operations. Explore my projects and insights to see how today's popular technologies are transforming the industry.
Cut-Off Grade Optimization Using Machine Learning Algorithms
Scope and Goals:
Developed a model for a mining project with a 500,000-ton capacity, optimizing Cut-Off Grade (COG) to maximize Net Present Value (NPV).
Technologies and Methods:
Conducted 10,000 simulations using ore tonnage and grade data. Utilized future price projections and a DeepAR model for grade determination. Linear regression model's R² value was calculated for validation.
Results and Achievements:
The initial maximum NPV was 97 million USD, which was optimized to 117 million USD after COG adjustments.
Plant Capacity Optimization Using Python
Scope and Goals:
Determine the optimal Dense Medium Separation (DMS) capacity for a greenfield mining operation by maximizing Net Present Value (NPV) based on varying plant capacities.
Technologies and Methods:
Conducted analyses on plant capacities ranging from 0.5 Mt to 2 Mt per year, incorporating variables such as mining unit costs, yield values, processing costs, product selling prices, investment costs, and loan repayments. Used Python for mathematical modeling and simulation with 1,000 iterations per capacity value.
Results and Achievements:
Identified a plant capacity of 1.1 Mt as the optimal size, achieving maximum NPV by analyzing different scenarios and variable parameters.
Iron Ore Price Forecasting Using AutoGluon
Scope and Goals:
Forecast iron ore prices using AutoGluon: Time Series Forecasting to provide accurate predictions for better financial and operational planning.
Technologies and Methods:
Utilized AutoGluon for data preprocessing, feature engineering, model training, and evaluation. Employed various models, including ETS, Theta, and DeepAR (Neural Network), with a weighted ensemble model to achieve the best performance. Evaluated models based on Mean Absolute Error (MAE).
Results and Achievements:
The Weighted Ensemble Model, combining predictions from DeepAR and Theta models, achieved the best validation MAE of 15.41, significantly improving forecasting accuracy.
Financial Modeling of Mining Investments
Scope and Goals:
Developed financial models for mining investments to evaluate project viability, plan finances, and manage risks associated with high upfront costs and volatile commodity prices.
Technologies and Methods:
Utilized financial modeling techniques to assess project evaluation, financial planning, and risk management. Created detailed financial plans and strategies to improve performance and reduce risks, using scenario analysis and forecasting.
Results and Achievements:
Enhanced decision-making capabilities for mining projects, providing a comprehensive understanding of financial situations and future risks, ultimately aiding in successful project evaluations.
Advanced AI Models for Mining and Optimization
Optimization Studies in the Mining Sector