Understanding Shale Gas Hydraulic Fracturing: AI-Powered Automatic Calibration for Fracture Geometries Quantification

Le résumé fourni par la source

ABSTRACT: This study presents an integrated approach to optimize hydraulic fracturing and improve reservoir performance in deep shale gas reservoirs in the Sichuan Basin, China. The methodology combines geological modeling, hydraulic fracture simulations, and AI-driven history matching to quantify the effective fracture geometries and predict long-term production. Using ensembles of reservoir simulations and the Embedded Discrete Fracture Model (EDFM), the study calibrates hydraulic fracture parameters, revealing significant variability in fracture performance. The workflow proposed in this study is applied to a real shale gas well, and the results show that hydraulic fractures created by the completion job have a fracture length of double the effective length of the fracture that actually contributes to production (160 meters versus 80 meters). The study also provides long-term production forecasts and estimated ultimate recovery (EUR) predictions, demonstrating the utility of AI and in evaluating well performance. With multiple realizations of history matching models, the EUR confidence interval of this well is determined to range from 0.1098 to 0.1156 billion cubic meters. This approach offers valuable insights for maximizing hydrocarbon recovery and enhancing the efficiency of unconventional reservoir development strategies such as better designs of completion job and well spacing.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.