Identifying Fracture Sweet Spots Using Artificial Neural Network Approach: A Case Study in Najmah/Sargelu Reservoir, Kra Al-Maru Field, West Kuwait

Abstract

AbstractArtificial Neural Network (ANN), defined as intelligence exhibited by computer through emulating the functioning of cerebral neurons, has many applications in today's society. ANN application (Ouahed et al, 2005) has become widespread in petroleum geoscience and reservoir engineering as it is able to generate accurate prediction taking an integrated and un-bias view of multiple inputs using supervised or unsupervised machine learning approaches. Unlike simple clastic sandstone reservoirs, the distribution of fractures in carbonate reservoirs is complicated and heterogeneous due to multiple genetic factors involved that makes high-resolution 3D modeling and simulation challenging. In order to enhance the reservoir characterization and modeling work in fractured reservoirs, an innovative application of Artificial Neural Network is introduced to tackle the disadvantages & shortcomings of existing techniques used in the Oil & Gas industry.This paper discusses the importance of ANN to delineate fracture sweet spots of Nahmah/Sargelu carbonate reservoirs in order to optimize drilling strategy in Kra Al-Maru (KM) field, West Kuwait. ANN works as a simulation of human brain, which comprises of highly interconnected neurons. These neurons process and integrate data from various domains to generate a desirable output. Nowadays, the application of ANN is prevalent in the field of fracture characterization and modelling technology because of its supremacy in data prediction and estimation. An integrated Continuous Fracture Modeling (CFM) workflow is established using ANN technology to understand the relationship between 3D seismic attributes and fracture intensity log derived from borehole image (BHI) data in both qualitative and quantitative evaluation (Pinous et al, 2007). ANN plays a critical role as it transforms the learning from seismic-image log relationship to 3D fracture intensity model. Consequently, the fracture sweet spots in the KM field is delineated to gain comprehensive understanding of fracture distribution in the field. The validity of the work is confirmed by blind test from the recent development well. If the result is not satisfactory, the ANN model is fine-tuned until the result matches in the blind test. The success of the study helps to refine the drilling strategy of the challenging KM Field, while reducing the risks and costs significantly.