VIRTUAL FLOW METER USING DATA DRIVEN DIVERSE ENSEMBLE LEARNING AND PHYSICS BASED MULTIPHASE FLOW SIMULATOR

Production well testing is one of a major routine in the upstream oil and gas operation. The operation is executed traditionally using test separator vessel equipped with
individual oil, gas and water flow meters. For more compact installation, Multiphase Flow Meter (MPFM) is used. Manual well testing using test separator could be interrupted by bad weather, while measurement verification of MPFM is not yet
available. This research explored two independent methods of Virtual Flow Meter (VFM) i.e. using data-driven using Diverse Ensemble Learning Neural Network (DELNN) and Transient Multiphase Flow Simulator (TMFS) to provide measurement
verification for the physical MPFM. The research developed and implemented those two VFM models in an offshore oil production platform with 2 oil producing wells flowing through a single MPFM.