Objective The Daning-Jixian block on the eastern margin of the Ordos Basin has achieved large-scale production of deep coalbed methane (CBM), with nearly 150 horizontal wells having been put into production. However, with a gradual decrease in formation energy during CBM production, gas wells exhibit declining liquid-carrying capacity. Consequently, liquid accumulation in wellbores has become a major factor affecting deep CBM production. Recovery of deep CBM shows the coexistence of free and desorbed gases, accompanied by significantly varying gas/liquid ratios. Moreover, gas production channels and techniques vary in different production stages. Hence, there is an urgent need to develop a method for liquid accumulation diagnosis and prediction that is suitable for the production characteristics of horizontal wells for deep CBM, aiming to provide a basis for the prevention and control of liquid accumulation and to avoid damage to reservoirs and their productivity caused by liquid accumulation. Methods and Results Using the Reynolds-averaged Navier-Stokes (RANS) κ-ε equation for incompressible viscous fluids and the volume of fluid (VOF) method, as well as the physical simulation experimental results of gas-liquid two-phase flow in a circular tube and an annulus, this study developed a numerical model of gas-liquid two-phase flow in horizontal wells for deep CBM utilizing the Fluent computational fluid dynamics tool and its secondary development. Based on the numerical simulation results, this study plotted the gas-liquid two-phase flow patterns in horizontal wells under different wellbore pressures, different inclinations, along with the circular tube and annulus conditions. Furthermore, this study established the corresponding relationship between the flow pattern and liquid accumulation based on the regularity and pattern evolution of gas-liquid two-phase flow in the production process. The results indicate that bubble and slug flows correspond to the liquid accumulation state. In contrast, churn flow corresponds to the transition state where liquid accumulation will occur, while annular flow corresponds to the state with no or low risk of liquid accumulation. Additionally, the well inclination is directly proportional to the liquid accumulation risk, whereas the pressure is inversely proportional to the risk.

Conclusions The flow pattern chart board-based method for diagnosing liquid accumulation proposed in this study was applied to the horizontal wells for deep CBM in the Daning-Jixian block, providing guidance for proposing the intervention timing, taking control measures in time, with the efficiency of measures having been improved. In subsequent studies, this method will be optimized for intelligent analysis and prediction using artificial intelligence (AI) techniques. This will provide robust technical support for the prediction, prevention, and control of liquid accumulation in CBM wellbores.