PhD Thesis Title:
Wet Gas Flow Measurement Through Data-Driven Modelling Using Venturi Tube
This PhD research addressed a long-standing challenge in flow metrology — achieving accurate measurement of wet-gas flow using differential-pressure devices such as Venturi tube. The work comprised two complementary strands: computational fluid dynamics (CFD) investigations to understand the underlying flow physics and pressure distribution mechanisms, and data-driven modelling to develop intelligent predictive frameworks for flow-rate estimation. Advanced machine-learning and deep-learning architectures were employed to capture nonlinear sensor–flow relationships, while uncertainty quantification (UQ) techniques were applied to evaluate the robustness and confidence of the predictions. In addition, transfer learning (TL) was explored to improve model adaptability across different metering facilities and operating regimes. Together, these methodologies establish a transparent, explainable, and metrologically traceable framework for wet-gas flow measurement. The outcomes contribute to the development of next-generation intelligent Venturi meters, supporting the digital transformation of flow measurement and enabling improved diagnostics for both traditional oil-and-gas operations and emerging hydrogen and CO₂ transport applications.
As a recipient of the WCSIM Postgraduate Award Scholar 2025, I wish to express once again my sincere appreciation for your support, which significantly contributed to the progress and visibility of my research. I have recently published a paper in the Measurement (Elsevier) journal acknowledging WCSIM’s support:
- Title: Machine learning-driven multiphase flow prediction for wet gas: a temporal data perspective incorporating fluid property analysis and explainable AI
- DOI: https://doi.org/10.1016/j.measurement.2025.119077