Fracture network Based on Principal Component Analysis and Neural Network– A case study in the Malay Basin

Abstract

Finding oil in the fractured basement rock in South East Asia has been a goal for several decades, but remains a challenge in terms of exploration/production areas in the Malay Basin due to geological complexity including fracture. Thus, the purpose of this study is to delineate fracture network based on the geometrical attributes in order to have better fracture understanding. In this study, the top of the basement acts as the key surface incorporated with the combination of geometrical seismic attributes analysis. The analysis started with data conditioning and seismic interpretation of the key surface. The final steps were conducted by using geometrical seismic attributes, principal component analysis and neural network. Principal component analysis of these four seismic attributes is able to delineate the contribution of each attributes based on eigenvalue with the PC0: 1.3450 (33.63%), PC1:1.0556 (26.39%), PC2:0.9270 (23.17%) and PC3:0.6724 (16.81%). While neural network contributes four main results (i) PC0 (ii) PC0 and PC1 (iii) PC0, PC1 and PC2 (iv) PC0, PC1, PC 2 and PC3. Fracture networks were able to be delineated and geological features that might be overlooked were able to be captured and can be used to guide the fracture network inside the fractured basement.