An Accurate Machine for Real-Time Two-Phase Flowmetering in a Laboratory-Scale Flow Loop

Abstract

<para xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> A new flow measurement system for real-time flow computation of a two-phase flow in an oil field is described. The system utilizes an array of ultrasound sensors together with capacitance and conductance sensors to accurately interpret the fluid composition in the whole water-cut range. The flow rate is determined using venturi and differential pressure techniques. A dedicated hierarchical neural network algorithm that relies on the various physical properties of the fluid was implemented and tested in a laboratory-scale flow loop. Experimental results demonstrated that real-time classification within a <formula formulatype="inline"><tex Notation="TeX">$\pm$</tex></formula>5% relative error in volumetric flow regardless of the flow regime or the fluid composition can be achieved. This is an improvement over traditional systems, where weak accuracy is usually obtained either within the 40%–60% water-cut range or in the case of a high water cut. </para>