Machine Learning System for Geophysical Reservoir Property Prediction (Phase 2)
Client Need
The client required a solution to automate the processing and analysis of complex geophysical datasets to predict reservoir properties. Traditional manual methods were time-consuming, requiring experts to spend significant time interpreting large-scale data before making geological assessments. The objective was to accelerate this process, enabling geoscientists to focus on critical regions and conduct more accurate evaluations using AI-driven insights.
Challenge
Geophysical data varies significantly across different reservoirs, with inconsistencies in structure, scale, and formatting. Designing a machine learning system that could adapt to different geological contexts while maintaining accuracy posed a challenge. Furthermore, the system needed to provide interpretable visual outputs tailored to geoscientific workflows—accessible even to domain experts with minimal technical or ML experience.
Tech Stack
- Machine Learning Models: Modern boosting algorithms (CatBoost, LightGBM, etc.)
- Data Handling: Domain-aware preprocessing customized per geological region
- Model Training: Parallelized model evaluation and optimization pipeline
- Frontend: ReactJS with Dash, Plotly, and Matplotlib for scientific data visualization
- Backend: Flask server with public APIs for external application integration
- Deployment: Web-based platform for easy access and scalability
Our Solution
Tinhvan Software built a machine learning platform that supports:
- Parallel experimentation with multiple models and data preprocessing techniques
- A streamlined ML pipeline for data ingestion, transformation, training, and evaluation
- Expert-in-the-loop feedback, where geological experts reviewed model outputs and contributed to refining the preprocessing logic and evaluation metrics
- A web application with an intuitive interface that displays domain-specific visualizations, allowing geologists to analyze results without needing to understand ML internals
This enabled non-technical users to make fast, confident decisions based on model-backed predictions—accelerating analysis cycles while preserving domain expertise.
Business Impact
- Reduced geological data analysis time by over 60%
- Improved model accuracy through expert-driven iterative refinement
- Enabled domain experts to conduct advanced modeling without data science knowledge
- Increased productivity of R&D and exploration teams
- Provided a reusable framework for future reservoir prediction projects across different sites
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.