Degree
Doctor of Philosophy (PhD)
Department
Craft & Hawkins Department of Petroleum Engineering
Document Type
Dissertation
Abstract
The management of gas influxes in a drilling well has been considered one of the most critical topics for maintaining petroleum drilling safety and efficiency. Reliable modeling of gas influx behaviors in a wellbore or a marine riser greatly benefits the improved understanding of a gas kick event and the development of optimized influx management strategies. However, the current studies on transient multiphase flow modeling show varying limitations on both the model rigorousness and the overall numerical efficiency. Meanwhile, the existence of great system uncertainties, such as the unclear influx sizes and distributions in the well/riser, has brought significant challenges in achieving good performances of gas influx modeling.
This dissertation research investigates advanced transient multiphase flow models and the corresponding sub-models for the improved simulation of gas influx behaviors in a well or riser and proposes an improved modeling framework for gas influx management. In addition, the application of Data Assimilation (DA) in combination with advanced modeling is studied to perform real-time gas influx profiling and dynamic system status estimation. A series of full-scale experiments were performed at Louisiana State University (LSU) PERTT Lab facilities to investigate a variety of gas influx management scenarios and to provide high-quality, high-dimensional experimental data for the testing, verification, and validation of the developed numerical modeling and data assimilation methods. An improved numerical simulator based on a modified Drift-flux Model and a series of advanced sub-models was developed to simulate the transient multiphase flow dynamics during influx management. The coupling of the transient heat transfer and time-dependent mass transfer with the multiphase flow equations is investigated. The effect of gas solubility in liquids, gas slippage and migration, and the gas dispersion and suspension behaviors in non-Newtonian fluids (NNFs) are studied and included by developing the physics-based sub-models. The numerical efficiency of the developed models was investigated, and an improved numerical scheme was proposed to extend the applicability of the DFM in real-time well control scenarios. The non-linear Kalman Filter(KF)-based DA methods and their application in real-time system state estimation, gas influx profiling, and online modeling calibration are investigated in this research. An Ensemble Kalman Filter (EnKF)-based algorithm was integrated with the developed numerical simulator to achieve real-time estimation of gas influx size and distribution in both Water-based Mud (WBM) and Oil-based Mud (OBM) systems based on indirect measurement data. The performances of the developed algorithms are tested with experimental data and industry field tests, and the real-time application of the developed algorithms is demonstrated based on a real-time format of Influx Management Envelop (IME).
Results from the numerical simulation have demonstrated the satisfying performance of the developed simulator in predicting single or multiphase phase flow behaviors in the well/riser during various operational scenarios, including influx circulation and gas migration in a closed annulus. Satisfying agreements have been observed by comparing the estimation results with experimental data of different types, including distributed Fiber-Optic Sensing (DFOS) measurements. The proposed data assimilation methods have shown great potential in estimating the gas influx sizes and distributions and reducing overall system uncertainties, especially when noises exist in the measurements. The case studies of the real-time IMEs have demonstrated the significant benefits of implementing the improved modeling and data assimilation methods.
The novelties of this dissertation research are seen by the maximized use of advanced physics-based models and multi-dimensional real-time measurement data. The demonstrated decision-making process using real-time DA-based IME is beneficial in enhancing the safety and efficiency of gas influx handling and can be potentially extended to drilling and well control automation.
Date
4-14-2024
Recommended Citation
Wei, Chen, "Improved Modeling and Application of Data Assimilation in Gas Influx Management during Managed Pressure Drilling" (2024). LSU Doctoral Dissertations. 6392.
https://repository.lsu.edu/gradschool_dissertations/6392
Committee Chair
Chen, Yuanhang