Multiphase flowrate measurement with time series sensing data and sequential model

Abstract

Accurate multiphase flowrate measurement is challenging but crucially important in energy industry to monitor the production processes. Machine learning has recently emerged as a promising method to estimate the multiphase flowrate based on different flow meters. In this paper, we propose a Convolutional Neural Network (CNN) combined with Long-Short Term Memory (LSTM) model to estimate the mass liquid flowrate of oil/gas/water three-phase flow based on the Venturi tube. The range of the estimated mass flowrate of the liquid phase varies from 92.1 to 10000 kg/h. We collect time series sensing data from Venturi tube installed in a pilot-scale multiphase flow facility and utilize single-phase flowmeters to acquire reference data before mixing. The experimental results suggest the proposed CNN-LSTM model is able to effectively deal with the time series sensing data from Venturi tube and achieve acceptable liquid flowrate estimation under different flow conditions.