AI Application for Perforation Interval Prediction Under Limited Well Logging Data Conditions: A Case Study at Te Giac Trang Field

Abstract

The Te Giac Trang (TGT) field, operated by Hoang Long Joint Operating Company (HLJOC), is located in Block 16-1, Cuu Long Basin, approximately 120 km from Vung Tau. The exploitation of the TGT field faces significant challenges due to the lack of well logging data (WL) in some potential reservoir formations such as ULBH and ILBH5.1. This significantly impacts the assessment of reserves and the optimization of exploitation strategies. This paper discusses how to overcome the above limitations by applying artificial intelligence (AI) and machine learning methods to predict important geophysical well logging curves such as Neutron porosity (NEU) and rock density (RHOB) from Gamma Ray (GRN) and resistivity (RD) data. The machine learning model is trained and validated using blind test wells to ensure accuracy, achieving a reliability of up to 70% for wells that have undergone perforation. In addition, the paper also discusses how to use gas composition data to determine the fluid properties of the reservoir, thereby supporting the selection of optimal perforation intervals. Actual results from the TGT field show that this method accurately identifies potential oil reservoirs, contributing to a significant increase in production at an optimized cost. This paper confirms that the application of AI and machine learning can significantly improve oil and gas exploitation in data-limited conditions, revealing new directions in oil field management and operations. The results achieved at the TGT field show the potential for widespread application of this method in the oil and gas industry in Vietnam and worldwide.