Abstract

Reservoir characterization relies on the integration of petrophysical and geophysical data in making accurate predictions regarding reservoir properties. Deep learning neural networks have emerged as a powerful tool for such seismic reservoir characterization, particularly in complex geological settings. These networks can offer several advantages over traditional theory-based methods. For instance, Convolutional Neural Networks (CNNs) can yield predictions that are as good as, or even better than, theory-based inversion techniques. Additionally, machine learning techniques have the capability to simultaneously predict multiple reservoir properties, resulting in significant efficiency gains