The mass-flow rate that flows through wellhead/downhole chokes is typically used to measure oil/gas production. Published mass-flow rate prediction models are proposed based on wellhead choke data; however, because of the different flow behavior between vertical and horizontal tubes, unacceptable errors may occur when used on downhole choke wells. These models can be improved by three additional factors; flow patterns, gravity, and gas expansion coefficient. In this study, 180 pressure, mass-flow rate, and flow pattern data points were measured with four different size chokes on a vertical tube. According to the flow behavior analysis, a new comprehensive model was proposed to improve the accuracy of multiphase mass-flow rate prediction for downhole chokes. The experimental results show that the downhole choke redistributes the gas and liquid phase, causing the flow patterns downstream of the choke to change. Additionally, pressure data analysis results show that the pressure difference between the P_u (pressure upstream of the chokes) and P_d (pressure downstream of the chokes) can be used to identify the flow pattern transitions upstream of the chokes. The statistical error results of the new model coupled with various slip models reveal that the Simpson (RMSE = 0.0018 Kg/s, R² = 0.99) and Schuller (RMSE = 0.002 Kg/s, R² = 0.97) models achieve the highest accuracy prediction for the intermittent and continuous flow, respectively. Furthermore, the critical pressure ratio measured on a vertical tube can be accurately predicted by the new comprehensive model. Compared to other models, the new comprehensive model is the most accurate, yielding the best mass-flow rate predictions for the dataset; slug (RMSE = 0.0018 Kg/s and R² = 0.98), churn (RMSE = 0.0018 Kg/s and R² = 0.99), and annular flow conditions (RMSE = 0.002 Kg/s and R² = 0.97).

流经井口/井下节流阀的质量流量通常用于测量石油/天然气产量。基于井口节流数据提出已发布的质量流量预测模型；然而，由于垂直管和水平管之间的流动行为不同，在用于井下节流井时可能会出现不可接受的错误。这些模型可以通过三个额外的因素进行改进；流型、重力和气体膨胀系数。在这项研究中，使用垂直管上的四个不同尺寸的节流阀测量了 180 个压力、质量流量和流型数据点。根据流动行为分析，提出了一种新的综合模型，以提高井下节流阀多相质量流量预测的准确性。实验结果表明，井下节流阀重新分配了气相和液相，导致节流阀下游的流型发生了变化。此外，压力数据分析结果表明，P_u（扼流圈上游的压力）和 P _d（扼流圈下游的压力）可用于识别扼流圈上游的流型转变。新模型结合各种滑移模型的统计误差结果表明，Simpson (RMSE = 0.0018 Kg/s, R ²  = 0.99) 和 Schuller (RMSE = 0.002 Kg/s, R ²  = 0.97) 模型达到了最高的精度分别预测间歇流和连续流。此外，新的综合模型可以准确预测垂直管上测量的临界压力比。与其他模型相比，新的综合模型是最准确的，为数据集提供了最好的质量流量预测；段塞 (RMSE = 0.0018 Kg/s 和 R ² = 0.98)、流失（RMSE = 0.0018 Kg/s 和 R ²  = 0.99）和环流条件（RMSE = 0.002 Kg/s 和 R ²  = 0.97）。