Permeability Prediction and Potential Site Assessment for CO₂ Storage from Core Data and Well-Log Data in Malay Basin...

Area of Science:

• Geological Sciences
• Petroleum Engineering
• Environmental Science

Background:

• Geological formations are crucial for safe carbon dioxide (CO2) storage.
• Accurate site characterization requires detailed analysis of reservoir properties like porosity and permeability.
• Machine learning (ML) offers advanced tools for enhancing reservoir property estimation.

Purpose of the Study:

• To characterize the "X" field reservoir in the Malay Basin for potential carbon dioxide (CO2) storage.
• To evaluate the suitability of geological layers for CO2 sequestration based on key reservoir properties.
• To compare the effectiveness of different machine learning models in predicting reservoir permeability.

Main Methods:

• Utilized well logs and seismic data to determine reservoir properties (porosity, permeability, water saturation, shale volume).
• Employed machine learning models, specifically Naïve Bayes (NB) and multilayer perceptron (MLP), for enhanced permeability estimation and classification.
• Cross-validated reservoir property estimations with core data for accuracy.
• Developed a lithological model and utilized seismic profiling to identify reservoir and caprock zones.

Main Results:

• The multilayer perceptron (MLP) model demonstrated superior performance in permeability prediction, achieving 99% training and 93% testing accuracy.
• Identified the B100 zone as a low-permeability caprock and the D35-1 and D35-2 zones as high-permeability reservoirs suitable for CO2 storage.
• Confirmed the "X" field reservoir's depth (exceeding 1300 m) meets requirements for CO2 storage (1000-1500 m).
• Seismic data confirmed the B100 zone's integrity as a continuous caprock, essential for preventing CO2 migration.

Conclusions:

• The integrated approach combining geological data and machine learning effectively characterized the "X" field for CO2 storage.
• The "X" field in the Malay Basin presents a viable geological site for carbon capture and storage (CCS) initiatives.
• The study highlights the importance of detailed site characterization and advanced modeling techniques in assessing CO2 storage potential.