AI-Based Evaluation of Homogeneous Flow Volume Fractions Independent of Scale Using Capacitance and Photon Sensors

Abstract

Metering fluids is critical in various industries, and researchers have extensively explored factors affecting measurement accuracy. As a result, numerous sensors and methods are developed to precisely measure volume fractions in multi-phase fluids. A significant challenge in multi-phase fluid pipelines is the formation of scale within the pipes. This issue is particularly problematic in the petroleum industry, leading to narrowed internal diameters, corrosion, increased energy consumption, reduced equipment lifespan, and, most crucially, compromised flow measurement accuracy. This paper proposes a non-destructive metering system incorporating an artificial neural network with capacitive and photon attenuation sensors to address this challenge. The system simulates scale thicknesses from 0 mm to 10 mm using COMSOL multiphysics software and calculates counted rays through Beer Lambert equations. The simulation considers a 10% interval of volume variation in each phase, generating 726 data points. The proposed network, with two inputs—measured capacity and counted rays-and three outputs—volume fractions of gas, water, and oil—achieves mean absolute errors of 0.318, 1.531, and 1.614, respectively. These results demonstrate the system’s ability to accurately gauge volume proportions of a three-phase gas-water-oil fluid, regardless of pipeline scale thickness.

Keywords

• Non-destructive metering
• Scale thickness in pipelines
• Multi-phase fluids
• Artificial neural network
• Capacitive sensors
• Gamma-ray attenuation sensor