Mineralogical Based Geomechanical Model Assisted Hydraulic Fracture Design for Fiber Optics Well: Tested in a Tight Gas Sandstone Reservoir

Abstract

Abstract Accelerated gas production within PDO is crucial to meet the growing demand for natural gas. However, this requires effectively extracting gas from complex tight sandstone formations, which poses challenges and necessitates a deep understanding of formation behavior. Hydraulic fracturing has emerged as a key method for reservoir development in PDO. Some tight gas fields in PDO have been developed using vertical wells targeting two different hydrocarbon commingled sandstone units (A, B). To ensure successful hydraulic fracture design, a robust geomechanics model is essential. However, the current challenge lies in the limited ability to characterize mechanical properties across the formation using only sonic log data, which affects stress calculations and hampers the understanding of different mechanical barriers that impede fracture growth. Also, the stress and geomechanical properties of the reservoir are crucial factors influencing fracture design. This paper aims to discuss the application of a mineralogical-based geomechanical model to optimize hydraulic fracturing design and operations in tight gas sandstone reservoirs. A detailed mineralogical analysis of the reservoir rocks was conducted using a standard and existing workflow to identify the mineral composition. Additionally, data analysis was performed by correlating radioactive post-frac tracer results with the resulting minerals (e.g., TH and U). Multi-regression and machine learning algorithms were subsequently employed to predict mineral volume for all the wells studied. Consequently, three mineralogical qualitative curves were generated, providing insights into the dominant expected mineralogy. The paper proposes a new workflow to improve the geomechanical model by incorporating mineralogy and clay content. By adjusting elastic properties based on clay content, the stress model was calibrated using post-frac data. The effectiveness of this proposed method was tested and validated using Fiber Optics and post-frac radioactive tracer results. The results indicate the prevalence of quartz as the main mineral, with an increasing amount of K-feldspar observed in the lower reservoir section based on the Spectral Gamma Ray log. Additionally, the presence of various clay minerals significantly alters the elastic behavior of the reservoir rocks. The study presents low, expected, and high dominant mineral logs, utilizing a combination of multi-regression and Random Forest algorithms to account for the different subsurface characteristics of each reservoir. Notably, a strong correlation was found between the frac barrier and specific uranium and thorium concentrations. The application of this correlation to other wells demonstrated good agreement between the frac barrier determined by radioactive tracer data and the mineral composition of the rocks in the field. Furthermore, the study explores the update of the existing geomechanical model workflow based on the resulting mineral composition. This valuable information provides insights into the mechanical behavior of the rocks, as different minerals exhibit varying mechanical properties and responses to fracturing.