A Multiphase Flow Model for Dual-Gas Coproduction Wells Considering Gas-Liquid Countercurrent Flow

Summary Dual-gas coproduction, as a synergistic method for developing deep coalbed methane (CBM) and conventional natural gas reservoirs, has demonstrated significant potential in enhancing both single-well productivity and recovery efficiency. However, the multiphase flow behavior within the wellbore is complex, particularly at the interface between reservoirs, where gas-liquid countercurrent flow frequently occurs. This phenomenon severely impairs wellbore flow performance and gas production effectiveness. Existing research on multiphase flow mechanisms in dual-gas coproduction wells remains insufficient, and there is currently a lack of flow models capable of accurately capturing gas-liquid countercurrent behavior, posing a key challenge to further advancement of coproduction technologies. To address this issue, we investigated the wellbore structure and production characteristics of dual-gas coproduction wells and developed a full-scale laboratory wellbore experimental apparatus to simulate gas-liquid countercurrent flow. Using high-speed imaging and pressure sensors, we captured gas-liquid distribution, flow regimes, and pressure variations during countercurrent conditions. Three representative flow regimes were systematically identified—cocurrent flow, countercurrent flow induced by liquid loading, and countercurrent flow triggered by large-amplitude wave rupture—clarifying the transitional mechanisms and instability features of multiphase flow in dual-gas wells. Based on these observations, a novel transition boundary model was established, integrating liquid loading and wave rupture theory, which reduces prediction errors for critical gas velocity and pressure to within ±5%. Furthermore, a countercurrent-adapted multiphase flow model was developed, achieving a mean relative error of only 0.02 in pressure drop predictions. Compared with mainstream multiphase flow models (Beggs-Brill, Ansari, and Duns-Ros), the proposed model reduces residuals and standard deviations by more than 30%, thereby offering quantitatively validated improvements in predictive performance. These advancements provide explicit novelty in bridging experimental characterization with theoretical modeling, delivering strong theoretical support and practical guidance for optimizing wellbore structures, deliquification strategies, and production stability in dual-gas coproduction wells.