Petrology Prediction Model – AI for Geoscience (Phase 1)
Client Need
The Vietnam Petroleum Institute sought to enhance geological analysis by utilizing well log data collected under the seafloor, in combination with core sample data. Their objective was to develop a model capable of predicting petrology characteristics of soil layers throughout the drilling depth, thereby supporting more informed exploration and drilling decisions.
Challenge
The primary challenge stemmed from the nature of the available data. While sensor-based well log data was abundant, it was occasionally missing or inaccurate due to harsh environmental conditions. In contrast, core soil samples—although accurate—were sparse and only available at selected intervals. The client needed a solution that could effectively learn from the limited verified samples while generalizing well across noisy, incomplete sensor data.
Tech Stack
- Programming Language: Python
- Machine Learning Models: XGBoost, CatBoost, LGBRegressor, Random Forest
- Data Processing Techniques: IQR, Z-score, Cook’s Distance, StandardScaler, SelectKBest, Model Stacking
Our Solution
Tinhvan Software developed a machine learning-based tool that enables the Vietnam Petroleum Institute to test and evaluate multiple data modeling approaches. The system uses sensor data from well logs as training input and core sample data as ground-truth reference for validation. With robust data preprocessing, outlier handling, and model tuning, the platform supports comparative model evaluation and allows domain experts to refine predictive accuracy through a user-guided interface.
Business Impact
The solution enabled the Vietnam Petroleum Institute to significantly improve subsurface geological forecasting while reducing dependency on costly core sample extraction. By applying AI to historical and real-time data, the Institute gained faster, data-driven insights into soil structure, enhanced exploration planning accuracy, and improved operational efficiency in petroleum research and development.
Case studies
Intelligent Document Processing with AI Agent and OCR on Azure
Tinhvan Software built an Azure-based AI Agent and OCR tool to automate document search and image text extraction, enabling fast, secure, and structured data access. We developed an intelligent AI Agent and OCR pipeline to help enterprises query internal documents using natural language and extract structured data from image files. Built on Azure with infrastructure-as-code, the solution improves knowledge access, streamlines document processing, and enhances efficiency through secure, scalable cloud-native architecture.
Machine Learning System for Geophysical Reservoir Property Prediction (Phase 2)
Tinhvan Software built a machine learning platform to help geologists automate reservoir property analysis. The solution uses boosting models, domain-specific preprocessing, and interactive visualizations—empowering experts to interpret large-scale geophysical data faster and more accurately, with minimal technical overhead.
Petrology Prediction Model – AI for Geoscience (Phase 1)
Petrology Prediction Model, developed by Tinhvan Software for the Vietnam Petroleum Institute, applies machine learning to well log and core sample data, enabling accurate geological insights while reducing reliance on costly manual sampling and accelerating model testing in real-world exploration.