Bottom-hole pressure inversion method for nature gas wells based on blowout combustion flame shape parameters

• Published in: Energy (Oxford) 2024-05, Vol.294, p.130673, Article 130673
• Main Authors: Peng, Yudan, Fu, Guangming, Chen, Jiying, Sun, Baojiang, Sun, Xiaohui
• Format: Article
• Language: English
• Subjects: Bottom hole pressure; Fire; Inversion method; Numerical simulation; Well blowout
• ISSN: 0360-5442
• DOI: 10.1016/j.energy.2024.130673

In response to the challenges associated with the difficulty of obtaining critical well control parameters, particularly bottom-hole pressure, during a well blowout fire, a comprehensive computational fluid dynamics (CFD) model for analyzing well blowout fires is developed based on fluid mechanics and combustion theory, considering factors such as gas composition, pressure, wellhead diameter, wind speed, and temperature, and is validated by published experiment date. Then, a CFD simulation database is constructed by conducting a sensitivity analysis of the flame height to investigate the correlations among flow field, operating conditions and flame parameters. Subsequently, an approach incorporates an AutoEncoder (AE) to reduce the dimensionality of the highly correlated raw data, and a Bayesian Optimization-based Radial Basis Function Network (BO-RBFN) to explore the inherent patterns within the established CFD dataset and then establishes a fuzzy relation between immeasurable wellhead flow velocity and well blowout fires operating parameters, achieving an accurate prediction of wellhead flow velocity. Furthermore, utilizing the wellhead gas velocity and the calculated wellhead pressure as the initial conditions, the bottom hole pressure during blowout fire is inversed based on wellbore temperature and pressure model. As a result of this study, a model integrating numerical model, databases, intelligent algorithms is developed for determining bottom-hole pressure in the presence of a blowout fire. And the proposed deep learning-based approach demonstrated exceptional accuracy in predicting wellhead gas flow velocity, with an average absolute difference of less than 2% between validation set and test set. In addition, as the ambient temperature, wellhead pressure, and wellhead diameter increase, the height of the blowout flame increases. In contrast, as the gas temperature and wind speed increase, the height of the blowout flame gradually decreases. Notably, only marginal changes in flame height resulted from alterations in gas composition. This study can quantitatively reveal the cause of well blowout and evaluate the severity of well blowout, and have practical significance in guiding the design and execution of subsequent rescue operations for uncontrolled blowouts fires. •A novel method is developed for determining bottom-hole pressure under well blowout fire.•An accurate prediction of wellhead flow velocity is achieved based on intelligence algorithms