Abstract

Unconsolidated sandstone reservoirs present significant challenges in hydrocarbon extraction due to their weak mechanical properties, high porosity, and susceptibility to sand production. Hydraulic fracturing and other stimulation techniques are essential for enhancing well productivity in these reservoirs. However, traditional stimulation strategies often fail to achieve optimal results due to formation instability and premature sand production. This study proposes a predictive framework that integrates data-driven analytics, geomechanical modeling, and advanced fracture simulation techniques to optimize stimulation and fracturing operations in unconsolidated sandstone reservoirs. The proposed framework employs artificial intelligence (AI)-driven machine learning algorithms to analyze historical well performance data and predict optimal fracturing parameters. A coupled geomechanical-fluid flow model is incorporated to evaluate formation behavior under different stimulation scenarios. Additionally, proppant transport dynamics and fracture propagation are simulated using computational fluid dynamics (CFD) and discrete element modeling (DEM) techniques to assess fracture conductivity and long-term production sustainability. Field case studies validate the effectiveness of the proposed predictive framework, demonstrating improved fracture design efficiency, reduced proppant settling issues, and enhanced well productivity. The results indicate that integrating AI-based predictive models with geomechanical simulations significantly improves decision-making for fracture design, leading to better reservoir performance and minimized environmental risks. Furthermore, sensitivity analysis highlights the influence of critical parameters such as closure stress, fluid rheology, proppant size distribution, and reservoir heterogeneity on stimulation effectiveness. This framework provides a robust decision-support system for engineers and operators, enabling real-time optimization of fracturing strategies in unconsolidated sandstone reservoirs. By leveraging data analytics and advanced modeling techniques, the study contributes to the development of more efficient and sustainable stimulation approaches, reducing operational costs and mitigating sand production-related challenges. Future research will focus on expanding the framework to incorporate real-time field data for adaptive fracturing control and integrating digital twin technology for enhanced predictive capabilities.