EVALUASI TERINTEGRASI DARI ELEKTROFASIES HINGGA PREDIKSI PERMEABILITAS DENGAN PENDEKATAN MACHINE LEARNING: APLIKASI PADA RESERVOIR KARBONAT

• Main Author: Gafar Karim, Abdul
• Format: Theses
• Language: Indonesia
• Subjects: Pertambangan dan operasi berkaitan
• Online Access: https://digilib.itb.ac.id/gdl/view/56328
• Institution: Institut Teknologi Bandung

This study aims to determine the distribution of electrofacies and permeability in the well using a machine learning approach. The results of this permeability calculation are then compared with the results of the rock typing approach (hydraulic flow unit). The proposed method utilizes a machine learning approach based on clustering electrofacies as an additional feature in the calculation of permeability. The use of the electrofacies augmentation feature aims to keep the approach taken within the logic of knowledge of the petroleum engineering domain expert. Machine learning algorithms (AdaBoost, gradient boosting, random forest, neural network, SVM and linear regression) are also evaluated comprehensively in the evaluation process carried out. The results on the permeability estimation through the rock typing approach and the evaluation of feature augmentation made by Bestagini still do not give better results than using the proposed feature augmentation using additional parameter in the form of clustering electrofacies. With the AdaBoost algorithm, the permeability prediction results provide high accuracy and an increase in accuracy of 5%. (from 93% to 98%). The proposed approach using machine learning and feature augmentation in the form of electrofacies is able to identify electrofacies very well and facilitate distribution at uncored wells & intervals. This approach also provides permeability prediction results with higher accuracy in a relatively short time.