| 摘要 |
Virtual and non-invasive multiphase metering is performed for recognition of multiphase flow regimes of hydrocarbons and other fluids, flow rate, presence of sand, and other multiphase flow parameters. A passive acoustical detector system receives acoustical flow information in the form of acoustic emission signals, and a data processor processes and classifies the acoustical patterns. A statistical signal processing methodology is used. Acoustic models are provided for various flow regimes and flow patterns, using Artificial Intelligence methods including Hidden Markov Models and Artificial Neural Networks along with automated learning procedures. The metering can be used for downhole, top side and surface applications. |
