An Automated Data-Driven Workflow for Identifying Fractured Horizontal Well Sweet Spots in Shale Reservoirs

Abstract

ABSTRACT The unique characteristics of shale reservoirs, including micro-nanopores, fast lithofacies fluctuations, ultra-tight reservoirs, and complex hydrocarbon occurrence states, make sweet spot evaluation for shale different from conventional tight oil reservoirs. Fortunately, data science techniques have proven effective in managing the large volume of data collected during shale reservoir development. This study presents a comprehensive sweet spot evaluation workflow in shale plays using a machine learning approach based on two key features: fracability and mobility. Quantitative shale fracability and mobility parameters are calculated from conventional logging data using optimal assessment indicators. Additionally, the study proposes a modified comprehensive sweet spot evaluation model based on the independent weight coefficient method. A deep clustering algorithm based on autoencoder (AE) is trained to classify the partition pattern of the shale sweet spot and different classes of shale reservoir characteristics. The established models exhibit higher agreement rates and better reservoir classification results compared to traditional models. The modified ML model successfully uncovers nonlinear relationships in shale reservoir characterization and provides more reliable sweet spot evaluation results, which further facilitates automatic identification and search of shale reservoir sweet spots. INTRODUCTION Shale oil and gas resources have gained increasing attention in recent years due to the shift in exploration focus from conventional to unconventional resources (Ghosh et al., 2018; B. Liu et al., 2019, 2020; Luo et al., 2018, 2020). As an unconventional hydrocarbon resource, shale oil has become crucial to a diversified global energy pattern. Unlike the North American shale oil deposited in marine environments usually, China's continental shale oil possesses multiple types, wide distribution, and significant development potential, providing an essential supplement and replacement energy source for stable and increased oil production in oilfields. The shale oil reservoir is characterized by rapid lithofacies changes, large physical property differences, and strong heterogeneity, necessitating extensive volume transformation to achieve effective utilization and development benefits.